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
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
6b45c83f-cf81-4084-b3c8-0d897ebe9268 | relational-features-in-fine-grained-opinion | null | null | https://aclanthology.org/J13-3002 | https://aclanthology.org/J13-3002.pdf | Relational Features in Fine-Grained Opinion Analysis | null | ['ro', 'Richard Johansson', 'Aless Moschitti'] | 2013-01-01 | null | null | null | cl-2013-1 | ['fine-grained-opinion-analysis'] | ['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] | [-7.2009735107421875, 3.7909207344055176] |
ad5c893c-bdc4-4f9d-aa04-5ec11b70458a | improved-financial-forecasting-via-quantum | 2306.12965 | null | https://arxiv.org/abs/2306.12965v1 | https://arxiv.org/pdf/2306.12965v1.pdf | Improved Financial Forecasting via Quantum Machine Learning | Quantum algorithms have the potential to enhance machine learning across a variety of domains and applications. In this work, we show how quantum machine learning can be used to improve financial forecasting. First, we use classical and quantum Determinantal Point Processes to enhance Random Forest models for churn prediction, improving precision by almost 6%. Second, we design quantum neural network architectures with orthogonal and compound layers for credit risk assessment, which match classical performance with significantly fewer parameters. Our results demonstrate that leveraging quantum ideas can effectively enhance the performance of machine learning, both today as quantum-inspired classical ML solutions, and even more in the future, with the advent of better quantum hardware. | ['Samurai Brito', 'André J. Ferreira-Martins', 'Iordanis Kerenidis', 'Natansh Mathur', 'Skander Kazdaghli', 'Sohum Thakkar'] | 2023-05-31 | null | null | null | null | ['point-processes'] | ['methodology'] | [ 1.90490093e-02 -7.58804334e-03 -3.17283154e-01 -2.74984151e-01
-7.98556387e-01 -4.75690931e-01 5.50282359e-01 1.96527168e-01
-2.98481613e-01 6.09365046e-01 -1.03996255e-01 -7.69341886e-01
1.91983670e-01 -1.26040900e+00 -5.32885134e-01 -6.16497576e-01
7.20408978e-03 6.09656036e-01 -1.59742739e-02 -5.72589457e-01
6.96753979e-01 2.42476448e-01 -8.90462935e-01 4.90343809e-01
8.03839624e-01 1.03376579e+00 -7.67385364e-01 7.69193769e-01
2.72656411e-01 9.88624215e-01 -1.56324618e-02 -9.38689053e-01
3.70695472e-01 -1.93677142e-01 -6.96335375e-01 -9.31023657e-01
2.66656101e-01 -6.91125870e-01 -7.97471464e-01 1.02898169e+00
2.73947835e-01 1.15258200e-02 6.42190337e-01 -3.95199746e-01
-1.10980082e+00 7.52215624e-01 -2.16639474e-01 3.86255115e-01
-2.16600165e-01 4.20226723e-01 1.80574667e+00 -4.36424673e-01
3.26652139e-01 7.99059451e-01 5.98107040e-01 3.37151021e-01
-1.38100421e+00 -8.44039559e-01 -1.03430474e+00 7.31944919e-01
-8.36510241e-01 -2.79722303e-01 4.72489417e-01 -1.55782625e-01
1.43294716e+00 -2.66978770e-01 4.96145725e-01 7.44537592e-01
6.92749679e-01 6.20178580e-01 1.14582443e+00 -7.05848753e-01
2.10774347e-01 -2.91345157e-02 5.61291397e-01 8.12170506e-01
3.98386359e-01 1.07225108e+00 -9.23288584e-01 -3.77287000e-01
5.32562435e-01 -4.66312952e-02 3.13436359e-01 7.89855644e-02
-9.58494067e-01 1.46764278e+00 5.45787394e-01 1.59460828e-01
-3.13503146e-01 6.62839890e-01 1.76021472e-01 2.36383408e-01
3.38114917e-01 9.64478135e-01 -4.27012712e-01 -2.76793540e-01
-7.51261294e-01 3.70259970e-01 7.91286409e-01 3.13113958e-01
8.45025063e-01 8.21778849e-02 1.80454627e-01 -6.06505536e-02
1.25365704e-01 8.85245502e-01 3.11841190e-01 -1.25906587e+00
3.20691228e-01 -5.55248745e-03 8.75802338e-02 -4.89790678e-01
-5.66663504e-01 -7.19209254e-01 -7.32949197e-01 5.31943962e-02
3.82471025e-01 -1.38968512e-01 -5.26604056e-01 1.08412766e+00
-2.27719724e-01 3.90082121e-01 3.22152078e-01 5.77758193e-01
1.30857095e-01 5.57020485e-01 -1.43297166e-01 1.31470948e-01
1.24144840e+00 -7.97036052e-01 -2.74155229e-01 -1.44560471e-01
1.16561985e+00 -6.71186447e-01 3.75795960e-01 6.66016579e-01
-6.85769975e-01 -1.30285069e-01 -1.26565742e+00 8.23288932e-02
-1.01496547e-01 -3.74320835e-01 1.56134784e+00 1.35482311e+00
-6.77350879e-01 1.35812902e+00 -1.01260543e+00 3.88463497e-01
5.76061785e-01 4.32670146e-01 2.28652641e-01 9.15814191e-02
-1.79646325e+00 9.68285203e-01 4.67130542e-01 -2.06816241e-01
-2.43189707e-01 -4.61696684e-01 -2.25073040e-01 3.94540936e-01
-7.80088603e-02 -8.25437009e-01 1.60899222e+00 -3.16752017e-01
-1.70131850e+00 3.57635796e-01 1.16642296e-01 -1.27417648e+00
-2.73903251e-01 -2.68601537e-01 -4.98861223e-01 1.69998661e-01
-3.57516974e-01 2.53228515e-01 5.80158174e-01 1.18784346e-01
-6.05972946e-01 -6.10348582e-01 -2.37489328e-01 -9.72472653e-02
-1.48853451e-01 -1.59736603e-01 4.46710438e-01 -9.10519212e-02
2.40229025e-01 -1.19750726e+00 -3.78858417e-01 -9.36217368e-01
1.34547397e-01 -2.78529108e-01 1.92800947e-02 -4.68258947e-01
9.50097978e-01 -1.67329276e+00 -2.58540094e-01 4.33213621e-01
4.36680257e-01 2.55013436e-01 1.68816745e-01 3.16132873e-01
3.07730496e-01 7.90505484e-02 2.09601298e-02 2.77244687e-01
1.10437490e-01 2.16062628e-02 -5.26638448e-01 1.97376728e-01
2.61430144e-01 1.53752208e+00 -6.24385178e-01 1.10448331e-01
-1.80683449e-01 4.14692471e-03 -8.20886075e-01 -5.23554385e-01
-2.15741605e-01 1.63693354e-01 -4.58188564e-01 5.11082113e-01
5.12400031e-01 -5.67860663e-01 1.51324689e-01 2.97302485e-01
1.62378281e-01 8.98560464e-01 -8.16340744e-01 1.09032571e+00
-3.04409534e-01 7.05519736e-01 -5.32061458e-01 -7.03141510e-01
7.13577628e-01 7.61214048e-02 1.11513630e-01 -1.02947640e+00
1.58899084e-01 5.77192068e-01 3.84277374e-01 -3.89878382e-03
8.36931169e-01 -7.03163385e-01 -3.58658433e-02 9.45289552e-01
9.97996703e-02 -1.32487997e-01 -1.21659406e-01 1.27231717e-01
1.06049573e+00 -6.40706643e-02 1.19274622e-02 1.70131147e-01
8.04963410e-02 -7.45344758e-02 2.56176770e-01 1.02344513e+00
-5.55117846e-01 -1.76716000e-01 4.79208946e-01 -6.43790364e-01
-1.32320642e+00 -1.19234228e+00 -5.13085902e-01 1.13614511e+00
-7.36435875e-02 -3.15385342e-01 -3.90547574e-01 -2.17242181e-01
3.26505184e-01 1.08091366e+00 -1.95219815e-01 -3.89529407e-01
-5.45002043e-01 -1.52003872e+00 8.96551728e-01 5.30452013e-01
5.43668866e-01 -8.09355497e-01 -3.76066118e-01 2.66656965e-01
3.51545066e-01 -1.02429080e+00 1.73017681e-01 6.70006692e-01
-1.42590690e+00 -6.04979455e-01 -2.62449890e-01 -2.63885736e-01
-3.38111311e-01 -1.00233443e-01 1.01030552e+00 -2.84185350e-01
-1.98298097e-02 -3.43156606e-01 -3.21912169e-01 -1.79011196e-01
-7.47464538e-01 3.39354277e-01 2.15357855e-01 -4.11951870e-01
9.58447576e-01 -3.99208635e-01 -5.09486079e-01 -3.28793496e-01
-5.29882431e-01 -1.68320388e-01 9.97967482e-01 1.11564684e+00
5.45106642e-03 -1.06342271e-01 4.75636631e-01 -1.08837509e+00
5.78425527e-01 -1.64148241e-01 -7.14994788e-01 1.03281260e-01
-1.26343668e+00 8.91966462e-01 6.38435006e-01 2.28893220e-01
-8.19605947e-01 -4.41044480e-01 -8.18271413e-02 1.44924462e-01
1.06173955e-01 7.05826402e-01 7.98679650e-01 -3.84279788e-01
1.22596586e+00 -3.80897745e-02 -1.44395038e-01 -1.32920504e-01
6.74385130e-01 7.32678294e-01 2.46772304e-01 -3.70053023e-01
7.64405310e-01 2.78088659e-01 7.09487319e-01 -5.23361683e-01
-1.01238990e+00 -3.63499373e-01 -7.51289606e-01 4.15644020e-01
9.55043793e-01 -8.23040783e-01 -8.69487584e-01 3.79873335e-01
-9.03566360e-01 -1.57988816e-02 9.17664841e-02 1.00526035e+00
-4.29967880e-01 3.50642651e-01 -1.39567542e+00 -7.91213334e-01
-6.80259228e-01 -7.71733463e-01 5.30653775e-01 5.21806657e-01
3.29677939e-01 -8.62751961e-01 2.74025649e-01 6.68996096e-01
2.79849678e-01 -2.85083950e-01 1.09105539e+00 -7.70906925e-01
-1.20546746e+00 -4.80149001e-01 -3.80521923e-01 4.30053115e-01
-6.71780288e-01 -2.89753288e-01 -1.20624900e+00 6.01112731e-02
-3.83331738e-02 -4.02497798e-01 1.31853294e+00 2.14371219e-01
3.56500000e-01 -2.31605824e-02 5.38808256e-02 6.57256722e-01
1.10559702e+00 1.11489542e-01 6.41660988e-01 4.83445585e-01
4.47993726e-01 2.17652936e-02 1.86665267e-01 3.17380011e-01
2.25147411e-01 4.01554316e-01 6.67246133e-02 6.21904731e-01
5.09329498e-01 -2.59447508e-02 4.19338554e-01 9.04977441e-01
-4.84090894e-01 5.21451116e-01 -1.19059455e+00 -3.80621493e-01
-1.51702178e+00 -1.44508862e+00 -4.80658978e-01 2.23872209e+00
6.33596659e-01 4.96995836e-01 -1.57333612e-01 -2.93411642e-01
1.94357693e-01 -2.46577412e-02 -5.70661366e-01 -8.68627429e-01
-1.16766304e-01 1.12439263e+00 1.22381055e+00 4.02908295e-01
-1.19882953e+00 1.15972006e+00 6.83294153e+00 6.64941370e-01
-9.75053251e-01 4.01157558e-01 6.52715862e-01 8.81184414e-02
-2.86325395e-01 4.65773523e-01 -8.54943573e-01 1.38198547e-02
1.81485057e+00 -1.06663942e-01 8.87801170e-01 6.61038697e-01
-4.03284550e-01 -3.63318734e-02 -9.68758285e-01 8.95410776e-01
-2.42804885e-01 -1.72300458e+00 -2.82458156e-01 3.74902397e-01
1.05082178e+00 7.12581754e-01 4.84481484e-01 6.66735649e-01
3.09668332e-01 -1.12734854e+00 2.83908546e-01 5.48804462e-01
4.55801666e-01 -1.22906625e+00 1.12550223e+00 3.40164423e-01
-4.35619414e-01 -2.82798082e-01 -6.35193348e-01 -5.81011057e-01
4.46163230e-02 4.31370735e-01 -9.89670932e-01 5.32370150e-01
3.15085649e-01 4.10210043e-01 -6.07564628e-01 7.71499753e-01
-2.82101899e-01 9.98769283e-01 -5.18822074e-01 -5.98041117e-01
3.98100674e-01 -5.73114276e-01 8.39519501e-02 7.21667111e-01
2.33082414e-01 3.33196104e-01 -4.15482849e-01 1.07685137e+00
-2.06228405e-01 3.18987705e-02 -3.78082246e-01 -6.03998959e-01
1.30977124e-01 9.83754635e-01 -4.10896242e-01 -2.71315485e-01
-3.85228693e-01 9.32761371e-01 1.85719818e-01 3.71314362e-02
-7.81034768e-01 -5.86705446e-01 3.17862540e-01 -2.87591755e-01
2.48850986e-01 -6.50881350e-01 -8.59741151e-01 -1.53178310e+00
-6.14210546e-01 -6.89057052e-01 6.82979450e-02 -3.45447600e-01
-1.18610644e+00 5.99431135e-02 -7.88984835e-01 -7.40758300e-01
-5.20309031e-01 -1.15399575e+00 -5.12708306e-01 9.45796609e-01
-1.30752027e+00 -7.30359018e-01 5.38754225e-01 -1.91912383e-01
-5.11236846e-01 -5.43083310e-01 1.07972717e+00 8.43683407e-02
-3.59226942e-01 5.07679045e-01 1.15674806e+00 1.92335010e-01
6.34616554e-01 -1.25690746e+00 9.33598220e-01 5.74928284e-01
9.96990919e-01 6.15964234e-01 5.71592391e-01 -4.73011494e-01
-1.44180703e+00 -5.42696416e-01 7.68782914e-01 -8.60872746e-01
1.05690122e+00 -1.05458550e-01 -5.85515261e-01 4.52108622e-01
-1.52332544e-01 -4.06863838e-01 1.01830721e+00 8.79800260e-01
-9.28875029e-01 7.69025879e-03 -7.10632980e-01 2.22737208e-01
2.01948225e-01 -1.17260885e+00 -5.30890942e-01 5.75549483e-01
7.44006455e-01 -2.54481621e-02 -9.48604107e-01 2.63694853e-01
9.40423846e-01 -1.16809559e+00 7.36344099e-01 -9.37648773e-01
6.08661950e-01 4.27170932e-01 -2.82201678e-01 -1.00172925e+00
-4.55330521e-01 -8.72672737e-01 -4.01328355e-01 3.07953715e-01
7.54407287e-01 -8.76692891e-01 1.35251844e+00 4.68448699e-01
2.79331088e-01 -4.97322887e-01 -1.13316011e+00 -7.60744274e-01
9.08590853e-01 -7.04876900e-01 3.99206817e-01 6.45266235e-01
3.54894370e-01 8.28119218e-01 -3.42284709e-01 -6.14126399e-03
9.02366638e-01 3.51217926e-01 2.40358159e-01 -1.19804430e+00
-9.57086384e-01 -7.95922101e-01 -5.80470502e-01 -8.94030809e-01
-6.05597273e-02 -1.29433215e+00 -4.59658206e-01 -6.70154929e-01
2.35793322e-01 -3.60993803e-01 -5.91540992e-01 -1.36907056e-01
-1.06016971e-01 3.08318645e-01 2.85140663e-01 4.99956608e-01
-4.34410542e-01 4.91879821e-01 9.92954791e-01 3.86516936e-02
-1.19780004e-02 3.62382084e-01 -3.85156333e-01 4.70122963e-01
9.33865607e-01 -4.27180529e-01 4.68565404e-01 -4.44402546e-02
8.86609077e-01 1.07031532e-01 2.66964197e-01 -1.10730934e+00
3.98259088e-02 8.91482979e-02 4.92336035e-01 -2.55611420e-01
1.83553547e-01 7.85407722e-02 -6.46523237e-01 9.05165851e-01
-2.42162168e-01 -2.10101396e-01 -2.29276583e-01 7.20400512e-01
1.45260781e-01 -6.23752773e-01 7.99576521e-01 6.32172972e-02
-4.12774891e-01 5.08766808e-02 -1.57392174e-01 -1.36707529e-01
7.96380818e-01 4.38726246e-01 -7.28358030e-01 -3.98979247e-01
-4.71895128e-01 -1.85886055e-01 3.66959602e-01 -6.90662861e-02
3.13061655e-01 -9.87002015e-01 -4.69077498e-01 2.16635376e-01
-1.84836864e-01 -8.07472229e-01 1.58477157e-01 8.37286294e-01
-1.00698125e+00 1.31308889e+00 -1.69764474e-01 -2.51728982e-01
-4.49403644e-01 3.99837941e-01 3.51366818e-01 -6.26159668e-01
-2.43835852e-01 8.51769149e-01 -3.30733955e-01 -3.55138004e-01
-3.35177690e-01 -2.80829728e-01 5.51969446e-02 -1.47209108e-01
3.75598967e-01 5.08468747e-01 4.59801741e-02 -2.52773285e-01
4.51969244e-02 2.26928622e-01 -1.92853600e-01 -3.30491453e-01
1.31476104e+00 3.67354423e-01 -1.29231602e-01 5.17641664e-01
8.88957322e-01 -1.99419171e-01 -7.25516021e-01 -3.17723274e-01
5.17808855e-01 7.78896585e-02 4.25989360e-01 -6.92567527e-01
-3.45170319e-01 1.40524733e+00 6.54040098e-01 2.58262098e-01
4.73538756e-01 -1.08123340e-01 1.50761282e+00 1.29939151e+00
7.11700499e-01 -1.17379653e+00 -1.19588271e-01 9.15535688e-01
-4.28069741e-01 -1.32060719e+00 7.86760747e-02 1.99526578e-01
-5.03492355e-01 1.67961442e+00 -1.38592914e-01 -4.03599024e-01
6.46108985e-01 -2.01239884e-01 -2.58452743e-01 -2.24377409e-01
-9.85312819e-01 -1.46409169e-01 2.12885216e-01 -2.50290055e-03
4.98956472e-01 6.40018165e-01 -1.74855754e-01 5.77349067e-01
-3.96963328e-01 -2.39128005e-02 8.38518739e-01 6.54068351e-01
-1.02468598e+00 -1.45444298e+00 -3.69133562e-01 9.04746711e-01
-6.50011659e-01 -7.30886519e-01 1.20211216e-02 1.49930254e-01
-1.99037105e-01 7.52217114e-01 -7.60253966e-02 -7.41326153e-01
-3.81255418e-01 7.01871634e-01 6.85201824e-01 -5.97921491e-01
-3.82063210e-01 -2.03156129e-01 5.61138056e-02 -3.15509975e-01
8.25990438e-02 -8.14120352e-01 -1.30532706e+00 -8.02974105e-01
-7.84512758e-01 6.93649203e-02 6.95880890e-01 1.03136969e+00
3.79859805e-01 1.28596783e-01 4.72896010e-01 -3.86247844e-01
-1.68659878e+00 -8.30630839e-01 -8.42328191e-01 -2.42972016e-01
-1.77283786e-04 -3.33915889e-01 -2.18013138e-01 -3.78395349e-01] | [5.581636428833008, 4.970762729644775] |
cf0e5f7b-f116-4133-951b-787c7bb36881 | everything-at-once-multi-modal-fusion | 2112.04446 | null | https://arxiv.org/abs/2112.04446v2 | https://arxiv.org/pdf/2112.04446v2.pdf | Everything at Once -- Multi-modal Fusion Transformer for Video Retrieval | Multi-modal learning from video data has seen increased attention recently as it allows to train semantically meaningful embeddings without human annotation enabling tasks like zero-shot retrieval and classification. In this work, we present a multi-modal, modality agnostic fusion transformer approach that learns to exchange information between multiple modalities, such as video, audio, and text, and integrate them into a joined multi-modal representation to obtain an embedding that aggregates multi-modal temporal information. We propose to train the system with a combinatorial loss on everything at once, single modalities as well as pairs of modalities, explicitly leaving out any add-ons such as position or modality encoding. At test time, the resulting model can process and fuse any number of input modalities. Moreover, the implicit properties of the transformer allow to process inputs of different lengths. To evaluate the proposed approach, we train the model on the large scale HowTo100M dataset and evaluate the resulting embedding space on four challenging benchmark datasets obtaining state-of-the-art results in zero-shot video retrieval and zero-shot video action localization. | ['Hilde Kuehne', 'James Glass', 'David Harwath', 'Rogerio Feris', 'Brian Kingsbury', 'Samuel Thomas', 'Andrew Rouditchenko', 'Brian Chen', 'Nina Shvetsova'] | 2021-12-08 | null | null | null | null | ['action-localization'] | ['computer-vision'] | [ 2.87416697e-01 -1.51242554e-01 -1.92126170e-01 -2.82566279e-01
-1.37963247e+00 -8.19684625e-01 7.60677278e-01 3.99001002e-01
-7.50500858e-01 5.01035631e-01 3.92226994e-01 2.20747009e-01
-2.81442285e-01 -5.01151562e-01 -9.10330951e-01 -6.39259458e-01
-1.46937311e-01 4.82357413e-01 5.26892245e-01 4.87032980e-02
-1.60128176e-01 1.92097172e-01 -1.80410826e+00 7.56304026e-01
2.01514453e-01 1.29535460e+00 8.58073309e-03 7.68208683e-01
1.62878647e-01 8.94681275e-01 -1.84774041e-01 -4.93714511e-01
1.59658179e-01 -1.72309265e-01 -9.29308414e-01 1.34144314e-02
5.61496973e-01 -4.91250783e-01 -6.10949457e-01 6.70743823e-01
6.05863214e-01 3.18898797e-01 5.56673944e-01 -1.32438684e+00
-6.31777883e-01 5.73264241e-01 -2.17304826e-01 1.39847010e-01
7.22549021e-01 1.63330778e-01 1.29126966e+00 -9.32418168e-01
7.89186895e-01 1.29397047e+00 4.45701331e-01 4.19406831e-01
-1.19498718e+00 -1.54009283e-01 3.20070200e-02 5.47758996e-01
-1.20819497e+00 -5.05009711e-01 5.88311315e-01 -3.96658540e-01
9.79614258e-01 1.37811914e-01 5.36864221e-01 1.26066828e+00
-6.20450564e-02 1.01169503e+00 6.66263044e-01 -3.90171766e-01
9.68931019e-02 -2.76608989e-02 -5.75222261e-03 7.13446140e-01
-3.99858981e-01 -1.35785937e-01 -9.11236882e-01 -1.54811934e-01
2.78472900e-01 4.17053103e-01 -2.23721892e-01 -6.15721643e-01
-1.56037843e+00 6.67901099e-01 4.46503639e-01 3.85994971e-01
-3.49482298e-01 3.63282233e-01 8.50544274e-01 5.33230662e-01
3.14431429e-01 1.39510751e-01 -3.48622113e-01 -3.29696894e-01
-7.50742793e-01 5.97921573e-03 4.62990254e-01 7.51261055e-01
7.93090284e-01 -3.49558711e-01 -4.46623564e-01 8.17237377e-01
1.82390466e-01 4.75851595e-01 6.23462796e-01 -1.05863106e+00
6.02928996e-01 5.24830580e-01 -4.20728140e-02 -6.57228887e-01
-2.60258645e-01 3.54040980e-01 -5.96084595e-01 -5.28631546e-02
2.47591332e-01 1.30990729e-01 -9.59563494e-01 2.04547977e+00
2.78473705e-01 4.87112582e-01 1.71707913e-01 9.47341561e-01
7.41162777e-01 6.42634928e-01 1.05967999e-01 1.06369406e-01
1.43017733e+00 -1.03324127e+00 -5.62628508e-01 7.26253688e-02
7.14339375e-01 -6.27054870e-01 8.19910288e-01 1.18530355e-01
-1.24326372e+00 -5.34401476e-01 -8.81134450e-01 -3.48925054e-01
-7.18954861e-01 3.72035964e-03 3.27402890e-01 1.82485431e-02
-1.03709900e+00 5.71397364e-01 -8.73540103e-01 -5.38278580e-01
1.91856384e-01 2.68141240e-01 -1.00009131e+00 -4.16017145e-01
-1.41775715e+00 9.00846004e-01 6.35818958e-01 -1.86563954e-01
-1.17202675e+00 -6.66922092e-01 -1.05677128e+00 1.54469177e-01
5.03139079e-01 -8.14957023e-01 9.51874495e-01 -9.46085811e-01
-1.23953271e+00 8.02796781e-01 -9.01991874e-02 -4.43014711e-01
3.79735112e-01 -2.52070636e-01 -4.06611234e-01 8.08899283e-01
-2.09898427e-02 8.59294772e-01 1.10415447e+00 -9.15160835e-01
-6.89356327e-01 -5.16647577e-01 5.72578132e-01 2.27492824e-01
-7.79486358e-01 -7.59294108e-02 -5.19120872e-01 -4.29693788e-01
-1.07593358e-01 -9.86113727e-01 3.92479479e-01 4.05908793e-01
-6.00732341e-02 -2.01456890e-01 1.04829395e+00 -5.52797854e-01
1.00101590e+00 -2.45977998e+00 9.03789103e-01 -8.95680860e-02
1.19197302e-01 2.68512629e-02 -5.75050414e-01 6.66793108e-01
-1.01005159e-01 -1.64881453e-01 -1.32327467e-01 -7.70801187e-01
3.18008155e-01 3.64223808e-01 -7.48679563e-02 4.64366317e-01
2.80907482e-01 1.01711476e+00 -1.01442683e+00 -5.55449426e-01
4.76353317e-01 7.93468416e-01 -5.30951619e-01 2.82688826e-01
-5.27025759e-02 1.51933700e-01 -1.08587250e-01 8.03860724e-01
1.71316266e-01 -1.79465637e-01 2.07538884e-02 -5.26799679e-01
1.73018202e-01 -1.65873319e-01 -1.14242435e+00 2.45966768e+00
-6.93556488e-01 6.20302975e-01 -1.37256369e-01 -1.00880527e+00
2.79961437e-01 7.80353546e-01 6.83028340e-01 -5.41997373e-01
1.30547822e-01 1.29381016e-01 -4.43242252e-01 -5.91110945e-01
4.32290673e-01 -8.58829096e-02 -4.24040228e-01 4.67090100e-01
9.47384000e-01 2.56406039e-01 2.45590657e-01 3.50957543e-01
1.25293410e+00 1.21403180e-01 9.61714387e-02 2.54418224e-01
6.32526100e-01 -4.92826074e-01 9.61774737e-02 5.18491864e-01
-2.04219986e-02 6.23792708e-01 3.95510793e-01 -2.01915383e-01
-9.44515347e-01 -1.29043913e+00 2.84712940e-01 1.52234960e+00
2.07330421e-01 -4.79970962e-01 -4.03773129e-01 -7.15167165e-01
5.20452037e-02 1.91634998e-01 -8.63633811e-01 -4.62346107e-01
-2.78984994e-01 -3.22163738e-02 7.09215462e-01 6.44113779e-01
2.00572938e-01 -8.71007442e-01 -7.15857387e-01 3.40485238e-02
-4.64899898e-01 -1.22485042e+00 -6.30156815e-01 2.27318063e-01
-6.11991227e-01 -1.09250307e+00 -8.00257206e-01 -7.17344046e-01
2.82491654e-01 1.41109601e-01 8.78395498e-01 -2.82704324e-01
-4.10026699e-01 9.99481738e-01 -5.92882633e-01 2.62195230e-01
-1.39054403e-01 -1.16025638e-02 2.26757899e-02 5.10518312e-01
1.88303456e-01 -6.68862641e-01 -4.10794646e-01 9.68843028e-02
-1.48048770e+00 -2.14273393e-01 4.70700651e-01 9.76234972e-01
5.82135439e-01 -2.64311045e-01 4.40116167e-01 -3.79613608e-01
3.60591292e-01 -5.29036939e-01 -1.63316518e-01 6.81103408e-01
1.35915026e-01 3.50064903e-01 4.86789912e-01 -6.59003079e-01
-7.23106027e-01 1.68740347e-01 1.62430018e-01 -1.13422334e+00
-1.56606257e-01 5.25111258e-01 -2.45411307e-01 -1.02227442e-02
3.40034902e-01 2.84056813e-01 -1.37538075e-01 -4.94608641e-01
9.29759681e-01 4.52826947e-01 6.53437316e-01 -5.19751370e-01
6.37683749e-01 6.31539583e-01 -1.19077824e-02 -7.01396525e-01
-6.09011114e-01 -7.13665903e-01 -7.94617534e-01 -2.62276530e-01
1.13846636e+00 -9.96367335e-01 -6.57139659e-01 3.92006695e-01
-1.24661684e+00 -8.92145280e-03 -5.54632246e-01 6.17864370e-01
-7.57558048e-01 2.98861355e-01 -6.32854760e-01 -4.61842656e-01
-1.28084093e-01 -1.11109173e+00 1.47144115e+00 -8.22844654e-02
-7.26645663e-02 -1.01887941e+00 8.39155540e-02 1.92259058e-01
2.24303275e-01 1.99226305e-01 7.87743330e-01 -7.56667554e-01
-5.81689417e-01 -3.84682059e-01 -1.38314053e-01 3.25586945e-01
-1.41124390e-02 -1.92079470e-01 -1.01476955e+00 -4.14395571e-01
-4.06609356e-01 -8.73016298e-01 1.14719486e+00 -9.90660265e-02
1.00341749e+00 -1.28450215e-01 -1.57658517e-01 5.39549947e-01
1.45762658e+00 -2.51307756e-01 4.94912654e-01 1.60631046e-01
5.72579682e-01 3.54123801e-01 4.99746174e-01 5.39021909e-01
3.87781978e-01 7.41021752e-01 5.60608804e-01 2.63953358e-01
-2.12101452e-02 -3.27739835e-01 3.87132913e-01 7.23439693e-01
-2.94712167e-02 -4.62920576e-01 -4.45376933e-01 7.13902593e-01
-1.95330215e+00 -1.24769866e+00 5.70925176e-01 2.21236753e+00
6.57900095e-01 -1.45737499e-01 1.65656939e-01 2.07283556e-01
4.73273277e-01 3.78606588e-01 -4.01267171e-01 -1.74046829e-01
-1.31271705e-01 1.77900478e-01 2.58823723e-01 4.50567752e-01
-1.41249633e+00 6.77429557e-01 5.46448040e+00 6.91870511e-01
-1.26964676e+00 4.80437666e-01 -8.00542980e-02 -4.34005708e-01
-3.79363388e-01 -1.35062829e-01 -2.97897667e-01 3.33829939e-01
9.39609170e-01 -1.02078475e-01 6.31540000e-01 4.76422250e-01
-2.54338622e-01 -2.13974100e-02 -1.53195322e+00 1.13809574e+00
4.75640029e-01 -1.22475374e+00 2.57526934e-01 -2.86410213e-01
4.56498563e-01 5.33470958e-02 3.70214656e-02 4.12573367e-01
-1.40792042e-01 -7.72023022e-01 8.96399736e-01 8.03302646e-01
9.12149489e-01 -4.51986879e-01 5.41097522e-01 1.87200159e-01
-1.43790066e+00 -3.27662021e-01 8.82548094e-03 3.07466060e-01
5.06511331e-01 1.04476131e-01 -2.92268991e-01 9.34006989e-01
5.43230653e-01 9.51505005e-01 -6.77235723e-01 9.45145905e-01
9.96261239e-02 4.33355421e-02 -4.98575836e-01 4.69760537e-01
2.97677994e-01 1.81061924e-01 5.23403883e-01 1.13243735e+00
5.49644172e-01 -1.20124780e-01 2.16452494e-01 2.82954395e-01
-3.27660114e-01 -1.65014237e-01 -7.44904280e-01 -3.19906622e-01
3.88811052e-01 1.08268201e+00 -3.52632195e-01 -4.43177044e-01
-7.62140512e-01 1.38913441e+00 3.83688897e-01 4.40181375e-01
-1.01361990e+00 -2.41297916e-01 7.16767073e-01 -1.89856037e-01
5.02263188e-01 -1.35778755e-01 3.37759435e-01 -1.37917578e+00
2.18884483e-01 -5.78275859e-01 8.10044050e-01 -8.96934927e-01
-1.21971691e+00 6.12085104e-01 2.01638520e-01 -1.38762069e+00
-4.23926294e-01 -4.51669902e-01 -1.98105618e-01 5.81262887e-01
-1.50241315e+00 -1.51998615e+00 -1.81687862e-01 1.09438968e+00
4.89656240e-01 -1.51909143e-01 1.01823294e+00 7.94116437e-01
-1.76839545e-01 6.70203149e-01 -6.40908675e-03 4.94481362e-02
9.96607602e-01 -1.06260681e+00 -3.50571364e-01 5.70244312e-01
3.02406579e-01 3.98443222e-01 3.51536036e-01 -3.34668383e-02
-1.66858733e+00 -9.73262727e-01 7.51288474e-01 -4.94701385e-01
9.49572682e-01 -2.48589799e-01 -7.88794994e-01 7.12509573e-01
3.24853748e-01 4.53442216e-01 7.08535254e-01 -4.96302880e-02
-8.22975218e-01 -2.10203424e-01 -9.08364356e-01 1.47102565e-01
9.22910929e-01 -1.11099768e+00 -7.88763165e-01 1.93729997e-01
9.63956356e-01 -1.68623567e-01 -1.08673489e+00 3.32917184e-01
7.55557477e-01 -8.22478056e-01 1.12051558e+00 -8.27722132e-01
5.69632709e-01 -2.60452896e-01 -6.19348586e-01 -1.24158037e+00
3.06579582e-02 -3.03747118e-01 -4.40093875e-01 1.04760742e+00
3.56899232e-01 -4.59433734e-01 1.62278995e-01 3.53822201e-01
-8.98266509e-02 -5.53664088e-01 -1.35578716e+00 -6.90264821e-01
-2.86504239e-01 -3.42674613e-01 2.96607852e-01 7.35582709e-01
1.32572070e-01 4.39723223e-01 -6.62522793e-01 1.76288426e-01
4.89306569e-01 2.48310059e-01 4.92055684e-01 -1.04427338e+00
-4.73062277e-01 -1.52977630e-01 -8.74369323e-01 -8.98496509e-01
3.50741386e-01 -9.19996023e-01 -9.32660252e-02 -1.38033235e+00
3.34558427e-01 1.73587412e-01 -7.04201162e-01 7.65392959e-01
7.14161173e-02 5.08926749e-01 5.33159196e-01 5.38332723e-02
-1.18845892e+00 6.86923742e-01 8.95205915e-01 -3.97795796e-01
1.86980993e-01 -5.64800978e-01 -1.59224883e-01 4.80272859e-01
3.02487195e-01 -3.34516555e-01 -5.91325462e-01 -6.82964325e-01
5.53260483e-02 4.23702240e-01 7.60521352e-01 -1.04348230e+00
2.78851777e-01 9.09368023e-02 1.22652620e-01 -3.29780310e-01
9.78319645e-01 -1.17372155e+00 1.88485831e-01 -4.63365801e-02
-5.96662819e-01 -1.21458665e-01 1.24884024e-01 7.66929388e-01
-6.28993869e-01 -8.15786272e-02 6.38685346e-01 4.59249541e-02
-9.88507152e-01 4.16518450e-01 1.70265101e-02 -1.66755952e-02
1.26167250e+00 7.10193291e-02 -3.40652138e-01 -3.44495893e-01
-1.24032700e+00 2.71212697e-01 5.32025278e-01 7.36050069e-01
6.19248450e-01 -1.70395350e+00 -3.66392553e-01 2.12920625e-02
5.62163889e-01 -4.82809812e-01 6.24733746e-01 9.54979658e-01
-3.52337435e-02 3.57690483e-01 -3.20067346e-01 -6.78450108e-01
-1.31373346e+00 8.66522610e-01 1.16252571e-01 -2.52952695e-01
-4.41544950e-01 5.57599604e-01 -7.41336495e-02 -3.43232363e-01
2.58980930e-01 -2.17398822e-01 -3.39776874e-02 5.64662993e-01
5.90291679e-01 2.58954674e-01 -1.05764709e-01 -8.95702660e-01
-4.51920152e-01 6.03723884e-01 2.47270271e-01 -3.56072664e-01
1.27589869e+00 -2.13204965e-01 -4.25798260e-02 8.18801939e-01
1.88093340e+00 -4.99548495e-01 -1.14624393e+00 -4.76987988e-01
-1.47613063e-01 -5.73072135e-01 -9.91698354e-02 -5.07350087e-01
-9.75740910e-01 1.10206866e+00 7.14729607e-01 1.89949110e-01
1.21471381e+00 3.41759533e-01 9.12537158e-01 4.47201908e-01
3.19329768e-01 -1.04605818e+00 4.51203883e-01 4.58043247e-01
7.41886914e-01 -1.25440347e+00 -3.79737586e-01 1.41244128e-01
-6.29368544e-01 1.08364666e+00 1.65732175e-01 6.39789477e-02
6.50860250e-01 -1.29016265e-02 -2.70182371e-01 -8.50669444e-02
-9.95771289e-01 -5.29164135e-01 3.60703856e-01 4.28939044e-01
1.13284580e-01 -7.29036853e-02 -2.49713976e-02 3.72441828e-01
5.46553910e-01 2.54717797e-01 5.85775636e-02 1.09642851e+00
-2.75464207e-01 -1.14514267e+00 -6.05921932e-02 3.19527656e-01
-2.55407274e-01 -5.33758290e-03 -1.62501752e-01 5.09537101e-01
9.85724330e-02 6.33023798e-01 -2.12259442e-02 -4.45391983e-01
4.37898129e-01 4.48066354e-01 5.52132666e-01 -4.16458756e-01
-2.64840364e-01 -1.27172515e-01 -5.52764609e-02 -8.50240350e-01
-8.06734383e-01 -6.95195794e-01 -9.27671134e-01 4.82694209e-02
-1.21050648e-01 1.41528137e-02 4.83182341e-01 1.04387426e+00
4.02485281e-01 5.79759359e-01 4.18386221e-01 -1.05258858e+00
-5.99418104e-01 -8.79520118e-01 -4.36379433e-01 8.54893744e-01
5.57798624e-01 -8.08090270e-01 -3.80882412e-01 2.44547769e-01] | [10.24343204498291, 1.122999906539917] |
2f0bc557-856a-4a28-9b24-bf06d58787a9 | interactive-combinatorial-bandits-balancing | 2207.03091 | null | https://arxiv.org/abs/2207.03091v2 | https://arxiv.org/pdf/2207.03091v2.pdf | Online SuBmodular + SuPermodular (BP) Maximization with Bandit Feedback | We investigate non-modular function maximization in an online setting with $m$ users. The optimizer maintains a set $S_q$ for each user $q \in \{1, \ldots, m\}$. At round $i$, a user with unknown utility $h_q$ arrives; the optimizer selects a new item to add to $S_q$, and receives a noisy marginal gain. The goal is to minimize regret compared to an $\alpha$-approximation to the optimal full-knowledge selection (i.e., $\alpha$-regret). Prior works study this problem under a submodularity assumption for all $h_q$. However, this is not ideally amenable to applications, e.g., movie recommendations, that involve complementarity between items, where e.g., watching the first movie in a series enhances the impression of watching the sequels. Hence, we consider objectives $h_q$, called \textit{BP functions}, that decompose into the sum of monotone submodular $f_q$ and supermodular $g_q$; here, $g_q$ naturally models complementarity. Under different feedback assumptions, we develop UCB-style algorithms that use Nystrom sampling for computational efficiency. For these, we provide sublinear $\alpha$-regret guarantees for $\alpha = 1/\kappa_{f} [1 - e^{-(1 - \kappa^g) \kappa_{f}} ]$, and $\alpha = \min\{1 - \kappa_f/e, 1 - \kappa^g\}$; here, $\kappa_f, \kappa^g$ are submodular and supermodular curvatures. Furthermore, we provide similar $\alpha$-regret guarantees for functions that are almost submodular where $\alpha$ is parameterized by the submodularity ratio of the objective functions. We numerically validate our algorithms for movie recommendation on the MovieLens dataset and selection of training subsets for classification tasks. | ['Jeff Bilmes', 'Maryam Fazel', 'Lillian J Ratliff', 'Omid Sadeghi', 'Adhyyan Narang'] | 2022-07-07 | null | null | null | null | ['movie-recommendation'] | ['miscellaneous'] | [-1.74252898e-01 2.79781729e-01 -5.50514460e-01 -4.17987794e-01
-8.94877076e-01 -7.47467041e-01 -6.64353371e-01 5.99421300e-02
-5.33316016e-01 8.07553411e-01 -1.92804009e-01 -3.73855621e-01
-8.00302744e-01 -9.14941847e-01 -9.88496423e-01 -7.70442307e-01
-6.51338518e-01 4.24346507e-01 -4.68615323e-01 -3.77068847e-01
1.31327748e-01 3.53555717e-02 -1.28423989e+00 -1.36519670e-02
9.77447033e-01 1.58164513e+00 1.70101658e-01 5.46056569e-01
2.01481953e-01 5.15912592e-01 -2.93107420e-01 -7.47997522e-01
9.44321990e-01 -5.43458402e-01 -6.57656491e-01 2.24407330e-01
2.55629450e-01 -3.46636057e-01 -1.36268720e-01 1.30458522e+00
1.42735630e-01 5.37163615e-01 5.43701231e-01 -1.15351570e+00
-6.99480951e-01 7.22707212e-01 -7.53695905e-01 5.10092638e-02
1.57053337e-01 -4.19751406e-02 1.76401448e+00 -6.07751131e-01
4.75725442e-01 9.28515673e-01 2.49038011e-01 4.87344623e-01
-1.24844313e+00 -7.98961401e-01 5.15284657e-01 -2.55799949e-01
-1.24673307e+00 -5.18644415e-03 4.51243699e-01 -3.91166061e-01
5.60598373e-01 6.52719140e-01 3.18631262e-01 -1.94403362e-02
-5.90684526e-02 7.90967405e-01 8.98452997e-01 -1.26340881e-01
3.45631987e-01 4.22163963e-01 1.80139720e-01 8.10294390e-01
2.45716810e-01 -8.60331655e-02 -5.76727331e-01 -1.48752198e-01
7.79063642e-01 1.06613927e-01 -4.55328673e-01 -1.09292179e-01
-5.38631976e-01 1.10140252e+00 3.50012869e-01 -1.15061611e-01
-4.43722785e-01 3.16417873e-01 -1.52441397e-01 9.84067440e-01
4.08879459e-01 6.14451289e-01 -7.02465534e-01 8.26755166e-03
-9.37108874e-01 4.20613736e-01 8.84732008e-01 1.33762872e+00
8.54347706e-01 -1.37051642e-01 -1.92197580e-02 8.23443234e-01
1.77437529e-01 7.00230300e-01 -2.24569336e-01 -1.53465593e+00
8.60887170e-01 5.63327491e-01 5.79945326e-01 -8.03399622e-01
-5.17843068e-01 -4.66569394e-01 -7.95154810e-01 6.69144467e-02
6.25083983e-01 -4.18536097e-01 -3.41424912e-01 2.14619040e+00
1.34825081e-01 -5.80903471e-01 -4.58437622e-01 1.23767066e+00
3.35409045e-01 6.67135298e-01 -4.11635160e-01 -1.11832595e+00
1.12606883e+00 -6.44463301e-01 -4.27708954e-01 -8.83954540e-02
4.82081294e-01 -6.70142412e-01 1.19568205e+00 7.18916893e-01
-1.79713500e+00 -4.59184907e-02 -7.93661773e-01 2.47811839e-01
1.22287735e-01 -2.85784602e-01 7.25625992e-01 9.56299305e-01
-1.02709496e+00 4.98028845e-01 -2.00961456e-01 2.35590771e-01
3.24373156e-01 8.17377746e-01 -3.33140157e-02 -2.80267090e-01
-9.68464434e-01 2.77216822e-01 -1.14068024e-01 -2.00165108e-01
-7.52212286e-01 -7.22823679e-01 -4.29641038e-01 1.96589559e-01
9.23346937e-01 -5.62475979e-01 1.12069237e+00 -1.03065741e+00
-1.13218522e+00 7.27842689e-01 -9.58273336e-02 -2.07537219e-01
3.38371217e-01 7.41165951e-02 -3.71324718e-02 -2.09491178e-02
1.49341479e-01 2.49465629e-01 8.02593291e-01 -1.07768476e+00
-8.33820701e-01 -7.85404444e-01 8.06124508e-01 4.48345125e-01
-4.72749740e-01 8.61230716e-02 -3.53280336e-01 -4.43613499e-01
2.65713036e-01 -1.02422154e+00 -5.54604292e-01 -1.22887075e-01
-2.23136410e-01 2.04973109e-03 -1.25844374e-01 -5.71348727e-01
1.57844412e+00 -1.98302257e+00 3.75796199e-01 7.50199556e-01
2.64096677e-01 -4.06832248e-01 -5.80714606e-02 4.03670281e-01
3.27335685e-01 3.82624209e-01 -1.46846980e-01 -3.78402442e-01
2.08654001e-01 9.76706669e-02 1.75885099e-03 5.63525319e-01
-7.67261446e-01 5.11166334e-01 -5.00301719e-01 1.75188363e-01
-4.61492002e-01 -2.54451007e-01 -9.96861637e-01 2.02343836e-02
-5.93359292e-01 5.50232790e-02 -4.60389167e-01 7.13141441e-01
8.67814243e-01 -5.97166240e-01 4.11656171e-01 2.97340512e-01
-6.89579919e-02 4.21461044e-03 -1.72608888e+00 1.40114725e+00
-4.49382812e-01 -1.10773750e-01 1.00961077e+00 -1.19621265e+00
5.44965148e-01 1.15721673e-01 8.70371819e-01 -5.94251990e-01
3.05008471e-01 2.24242270e-01 -1.80004016e-01 -2.10315108e-01
5.69976091e-01 -5.32283247e-01 -2.14111894e-01 8.64351749e-01
-1.20683923e-01 8.69422555e-02 3.51375222e-01 4.19112593e-01
1.04369378e+00 -5.87572038e-01 3.65782157e-02 -5.83085060e-01
2.67461509e-01 -2.28270516e-01 7.23662853e-01 1.06209850e+00
5.73289245e-02 3.81895959e-01 9.57453728e-01 -4.23031347e-03
-8.36716235e-01 -1.05051100e+00 -8.90189316e-03 1.73614132e+00
6.11449838e-01 -2.41587251e-01 -6.54699206e-01 -5.52804410e-01
3.77384365e-01 6.45629942e-01 -6.04134440e-01 1.27072588e-01
-3.07385534e-01 -1.00654674e+00 -1.59541041e-01 3.35413516e-01
2.62321860e-01 -7.49590039e-01 -9.35446396e-02 2.21913293e-01
-1.47051394e-01 -4.31217939e-01 -1.18764889e+00 2.19407529e-01
-9.31100130e-01 -8.31499934e-01 -5.74175417e-01 -4.43847507e-01
8.61175954e-01 4.59534347e-01 1.05268121e+00 -6.30506277e-02
-9.95959714e-02 4.59244341e-01 -3.39704633e-01 -1.92163870e-01
2.96448171e-01 -2.16776222e-01 2.91849554e-01 2.79769525e-02
-5.71353137e-02 -4.71482068e-01 -1.10300708e+00 6.53331399e-01
-1.06144333e+00 -5.70116639e-01 2.86441237e-01 6.98959112e-01
8.87692988e-01 1.14672706e-01 6.85541153e-01 -1.03881717e+00
6.70419931e-01 -7.85699010e-01 -8.64452422e-01 1.72708124e-01
-1.00705111e+00 -5.19658439e-02 9.93138850e-01 -3.48490477e-01
-6.80405915e-01 -4.11981761e-01 2.44765934e-02 -2.86306351e-01
6.23652041e-01 6.68321252e-01 -1.27414629e-01 5.70485182e-02
6.44539952e-01 -4.26208461e-03 1.24455858e-02 -4.69646901e-01
6.26873314e-01 4.42703575e-01 2.84624584e-02 -6.51979506e-01
4.51632172e-01 5.53050995e-01 -1.05027817e-02 -4.76982266e-01
-9.63718951e-01 -4.48072523e-01 2.17416525e-01 -6.51740953e-02
3.70051593e-01 -7.36947060e-01 -1.44861960e+00 -2.39012510e-01
-5.10575175e-01 -2.56148100e-01 -6.80649102e-01 5.63023746e-01
-8.34885359e-01 1.09609589e-01 -4.41956609e-01 -1.21770012e+00
-3.35810602e-01 -9.03318942e-01 3.62135291e-01 2.72318155e-01
3.07720363e-01 -6.27699673e-01 -2.91184992e-01 8.99955690e-01
3.80182385e-01 -2.90582001e-01 8.57585967e-01 -1.92610979e-01
-9.93418157e-01 -1.94464937e-01 -2.26222485e-01 6.51755214e-01
-1.93072259e-01 -6.00211203e-01 -3.38656306e-01 -7.12769806e-01
-7.74596166e-03 -3.88702512e-01 7.29605079e-01 7.80854940e-01
1.28182912e+00 -1.12622178e+00 1.26511246e-01 6.83847845e-01
1.39648914e+00 4.41233933e-01 2.17964694e-01 -1.22087508e-01
1.16904445e-01 3.94623637e-01 7.54999399e-01 1.20019424e+00
3.89542043e-01 4.79434848e-01 6.76528454e-01 3.78693819e-01
7.20327377e-01 1.15266919e-01 4.59691882e-01 5.44691920e-01
-2.07832962e-01 -3.95992041e-01 -1.04194134e-01 4.59090203e-01
-1.89842415e+00 -7.78839648e-01 2.15544462e-01 2.83685589e+00
8.62527788e-01 1.34933546e-01 5.57146907e-01 -2.41178706e-01
4.81314421e-01 1.06450003e-02 -8.16769958e-01 -5.76294243e-01
-2.87690729e-01 5.83654463e-01 9.73428845e-01 6.71982944e-01
-7.40821540e-01 5.56183040e-01 4.47493649e+00 9.49605942e-01
-7.62096286e-01 3.05050343e-01 8.54788840e-01 -1.11492968e+00
-7.24552095e-01 6.19040728e-02 -7.83144414e-01 6.46520019e-01
5.84902883e-01 -3.00112754e-01 1.16388130e+00 1.00045478e+00
3.30712348e-01 -3.11910510e-01 -1.01853132e+00 9.23793256e-01
-3.86870429e-02 -1.30311239e+00 -6.26458108e-01 3.96658450e-01
1.24324381e+00 -2.71707028e-01 5.10990441e-01 1.77842692e-01
6.82243884e-01 -1.11867964e+00 5.62247574e-01 7.13816285e-02
1.03402889e+00 -1.01558375e+00 3.97898048e-01 5.83690524e-01
-1.09131813e+00 -6.60246551e-01 -4.91254985e-01 -2.35664189e-01
3.92141193e-01 7.48949289e-01 -2.31583223e-01 4.66858327e-01
8.87323320e-01 2.01308057e-02 3.88425678e-01 7.22901583e-01
8.58494416e-02 3.36440802e-01 -6.04631901e-01 -1.25194088e-01
1.51644558e-01 -7.02117205e-01 4.88339156e-01 7.28431165e-01
3.36631149e-01 7.99085855e-01 3.70898396e-01 6.26656950e-01
-7.52966225e-01 6.94201410e-01 -1.72134176e-01 -5.50548732e-03
3.86519611e-01 1.06131613e+00 -4.44985420e-01 5.44688031e-02
-2.99803913e-01 7.33152032e-01 2.65795290e-01 3.24166745e-01
-7.58357704e-01 -2.50905216e-01 1.11408365e+00 4.32504296e-01
3.52947980e-01 -1.69502601e-01 -6.21135175e-01 -1.08698928e+00
2.97205001e-01 -7.92209864e-01 9.31105614e-01 -1.72226951e-01
-1.29270864e+00 7.35963210e-02 -2.01976806e-01 -8.68187070e-01
1.36032745e-01 -3.75902027e-01 -1.87037408e-01 7.75601327e-01
-1.01989293e+00 -4.51972634e-01 1.82305366e-01 7.90246964e-01
1.84588015e-01 -2.80073658e-02 4.23590809e-01 4.30247873e-01
-3.02881062e-01 1.03669059e+00 7.11563349e-01 -4.44971859e-01
3.30384135e-01 -1.18545938e+00 -6.02355063e-01 5.44959128e-01
-1.42121449e-01 7.06711948e-01 5.91930151e-01 -3.60924125e-01
-1.64370298e+00 -8.17211866e-01 5.78796148e-01 -6.91912472e-02
5.04522920e-01 -3.68187100e-01 -4.19612288e-01 6.97721481e-01
-2.97013462e-01 -4.01545800e-02 9.42441463e-01 3.18213195e-01
-2.31777623e-01 -5.98745108e-01 -1.62242448e+00 5.62695086e-01
1.42224300e+00 -1.89584956e-01 3.74488205e-01 5.63839793e-01
8.30865562e-01 -1.05585702e-01 -1.28546059e+00 2.83032387e-01
6.27496779e-01 -9.13120389e-01 7.67839372e-01 -8.04172039e-01
2.87605226e-01 4.85994518e-02 -6.93627059e-01 -9.81872499e-01
-4.19693202e-01 -1.00970721e+00 -1.48115277e-01 6.90399885e-01
8.58523190e-01 -4.90336180e-01 1.10480642e+00 9.41566110e-01
9.37079936e-02 -1.06814504e+00 -1.27414978e+00 -6.86351061e-01
4.19079989e-01 -4.27764058e-01 2.43035361e-01 6.79497063e-01
2.69975305e-01 4.20557223e-02 -6.91817939e-01 -4.82622860e-03
6.52324975e-01 4.38961983e-01 4.30647194e-01 -7.70343184e-01
-1.04507363e+00 -5.52474320e-01 3.24143231e-01 -1.55884802e+00
-3.72784883e-01 -8.31678629e-01 -2.12623104e-01 -1.17593598e+00
5.10107458e-01 -9.94476795e-01 -3.80071580e-01 2.50342339e-01
1.03992395e-01 5.89637570e-02 4.56134588e-01 1.20447829e-01
-8.75404000e-01 3.14167202e-01 1.45991933e+00 -4.05813456e-02
-5.14695942e-01 4.15876299e-01 -1.31899679e+00 3.62170994e-01
4.76504862e-01 -3.10950905e-01 -5.72365165e-01 -3.67436707e-01
8.83106351e-01 8.83229613e-01 -2.03823313e-01 -1.37819380e-01
1.19287580e-01 -7.54590452e-01 -2.15502784e-01 -2.54478633e-01
5.08682132e-01 -5.86291969e-01 7.87125006e-02 1.36335865e-01
-5.25945485e-01 -6.90523908e-02 -5.09590447e-01 5.52200258e-01
1.77222371e-01 -3.40940505e-01 9.83539402e-01 -3.39370728e-01
9.76347998e-02 6.84767127e-01 -7.07872882e-02 3.20668906e-01
1.07266748e+00 2.67781541e-02 -2.59492427e-01 -9.74908292e-01
-8.24149549e-01 6.02208257e-01 3.59950215e-01 -5.96562475e-02
4.19396281e-01 -9.96353149e-01 -5.91713905e-01 -1.88525677e-01
-8.01077858e-02 1.33273736e-01 6.36757970e-01 9.91596699e-01
-1.02191754e-01 2.04616487e-01 2.43085638e-01 -1.10595107e-01
-8.82718682e-01 6.14289880e-01 2.32071698e-01 -5.15442848e-01
3.59126806e-01 1.47944486e+00 2.60005087e-01 -2.02932701e-01
3.85063738e-01 -2.06161618e-01 2.66516060e-01 1.55568466e-01
3.24905753e-01 7.54869401e-01 -4.22765575e-02 -1.25876904e-01
-1.89871654e-01 9.79784429e-02 -2.12769777e-01 -4.85985786e-01
1.14341486e+00 -3.82706761e-01 -3.26055884e-01 5.91154769e-03
1.22266841e+00 1.05329163e-01 -1.18180120e+00 -3.32451224e-01
-5.16800880e-01 -6.62036359e-01 -3.16459835e-01 -8.63952875e-01
-1.36367154e+00 3.78560096e-01 3.34861130e-01 5.38122773e-01
1.26520491e+00 2.32877448e-01 6.84147000e-01 2.70859867e-01
7.34392643e-01 -1.47670531e+00 1.36999920e-01 3.60144109e-01
7.76054025e-01 -1.16125107e+00 -5.95762879e-02 -2.43129164e-01
-6.67642415e-01 6.55767083e-01 6.18969560e-01 -9.71547663e-02
9.82791781e-01 -3.82647999e-02 -4.56838429e-01 -8.35253224e-02
-7.37215281e-01 -1.43113419e-01 1.76831931e-01 -2.56182045e-01
3.98088545e-01 5.21674454e-01 -9.34009910e-01 1.33092380e+00
-2.93186069e-01 -6.03210554e-02 3.98563832e-01 7.39083111e-01
-7.02377737e-01 -1.12865865e+00 -2.59790599e-01 1.12355137e+00
-7.81005561e-01 -6.83489293e-02 -7.99589828e-02 2.76763380e-01
2.87230581e-01 1.27613091e+00 -1.44513980e-01 -3.78442436e-01
2.69767493e-01 -2.75702894e-01 5.50812244e-01 -5.12896836e-01
-7.32730687e-01 2.58763433e-01 -6.43000901e-02 -4.66796726e-01
2.77414899e-02 -6.32855475e-01 -1.30166566e+00 -7.19590068e-01
-4.04581904e-01 3.85330766e-01 6.23091340e-01 6.87291205e-01
1.55402675e-01 -2.15920508e-01 1.19900453e+00 -2.18061388e-01
-1.09531379e+00 -6.95463717e-01 -1.33751667e+00 5.77183068e-01
-6.70184642e-02 -5.05594909e-01 -5.14940202e-01 -2.14704975e-01] | [4.873674392700195, 3.622356653213501] |
483de7c6-f9e4-4c9f-b4aa-af6dd88a57e3 | continuous-sign-language-recognition-via | 2207.00928 | null | https://arxiv.org/abs/2207.00928v1 | https://arxiv.org/pdf/2207.00928v1.pdf | Continuous Sign Language Recognition via Temporal Super-Resolution Network | Aiming at the problem that the spatial-temporal hierarchical continuous sign language recognition model based on deep learning has a large amount of computation, which limits the real-time application of the model, this paper proposes a temporal super-resolution network(TSRNet). The data is reconstructed into a dense feature sequence to reduce the overall model computation while keeping the final recognition accuracy loss to a minimum. The continuous sign language recognition model(CSLR) via TSRNet mainly consists of three parts: frame-level feature extraction, time series feature extraction and TSRNet, where TSRNet is located between frame-level feature extraction and time-series feature extraction, which mainly includes two branches: detail descriptor and rough descriptor. The sparse frame-level features are fused through the features obtained by the two designed branches as the reconstructed dense frame-level feature sequence, and the connectionist temporal classification(CTC) loss is used for training and optimization after the time-series feature extraction part. To better recover semantic-level information, the overall model is trained with the self-generating adversarial training method proposed in this paper to reduce the model error rate. The training method regards the TSRNet as the generator, and the frame-level processing part and the temporal processing part as the discriminator. In addition, in order to unify the evaluation criteria of model accuracy loss under different benchmarks, this paper proposes word error rate deviation(WERD), which takes the error rate between the estimated word error rate (WER) and the reference WER obtained by the reconstructed frame-level feature sequence and the complete original frame-level feature sequence as the WERD. Experiments on two large-scale sign language datasets demonstrate the effectiveness of the proposed model. | ['Quan Gan', 'Fei Yuan', 'Jing Li', 'Qidan Zhu'] | 2022-07-03 | null | null | null | null | ['sign-language-recognition'] | ['computer-vision'] | [ 4.52395201e-01 -3.18923920e-01 -1.24808662e-01 -1.50832206e-01
-1.10756302e+00 2.19484121e-01 4.17545617e-01 -7.81680822e-01
-7.99111426e-01 5.18766046e-01 2.77739614e-01 2.79973984e-01
-1.21802554e-01 -7.24080503e-01 -3.94000381e-01 -1.11825538e+00
1.42666250e-01 -2.28872478e-01 2.78647393e-01 -1.82069018e-01
2.27266788e-01 5.24061561e-01 -1.88097775e+00 3.69282722e-01
1.01536691e+00 1.34662652e+00 2.08503962e-01 3.09164405e-01
-1.99948609e-01 9.89249349e-01 -7.14791119e-01 -1.25386342e-01
3.52080524e-01 -6.85582578e-01 -2.74106711e-01 -5.19193746e-02
1.16170712e-01 -5.51700294e-01 -8.36056173e-01 1.03880346e+00
9.26514566e-01 3.19772810e-01 2.68059939e-01 -1.08963168e+00
-3.67083341e-01 3.77242602e-02 -2.69546390e-01 7.04450905e-02
1.11470819e-01 4.33584929e-01 6.34814739e-01 -9.16633844e-01
7.41465032e-01 1.28401721e+00 4.40796852e-01 8.02465022e-01
-7.29077399e-01 -7.68134415e-01 3.21271308e-02 6.34864330e-01
-1.59273911e+00 -3.10469627e-01 8.42554629e-01 -2.14274168e-01
7.58657217e-01 4.63706963e-02 7.29121804e-01 1.13787568e+00
2.54429549e-01 8.57902288e-01 1.16987514e+00 -4.40222979e-01
1.23262331e-01 -3.09581667e-01 3.01977410e-03 6.29964411e-01
-1.95115551e-01 6.47239864e-01 -5.81008315e-01 2.25526690e-01
8.93731833e-01 -1.28671765e-01 -4.21806842e-01 1.94958746e-01
-9.18665588e-01 5.85878134e-01 4.05439138e-01 5.25105596e-01
-4.17148978e-01 8.88427161e-03 5.69616675e-01 3.29270422e-01
1.56718120e-01 -3.58216017e-01 -2.03660458e-01 -1.64216727e-01
-9.07311738e-01 -1.76855382e-02 2.73950487e-01 6.94203436e-01
2.03309000e-01 4.21078652e-01 -5.25353551e-01 7.92873204e-01
2.52666533e-01 5.13479590e-01 1.06692815e+00 -7.79814184e-01
4.45008278e-01 4.56946373e-01 -1.00546412e-01 -7.83721328e-01
-4.00044248e-02 -4.88654286e-01 -9.29503679e-01 3.67510378e-01
4.54789788e-01 6.57105893e-02 -1.07405877e+00 1.82639587e+00
1.42315134e-01 3.09474230e-01 1.39294446e-01 1.11851466e+00
8.29569757e-01 8.00492346e-01 1.82842016e-01 -5.40956438e-01
1.30225921e+00 -7.29116619e-01 -8.28178227e-01 2.75747985e-01
4.34233397e-01 -5.85952759e-01 9.35883462e-01 2.89920270e-01
-1.00096679e+00 -8.19878757e-01 -9.83805656e-01 -2.72814780e-01
-1.48865119e-01 5.26756942e-01 1.24459863e-01 2.94327468e-01
-6.36922836e-01 4.20291483e-01 -9.14793551e-01 -4.81836721e-02
3.48493576e-01 2.47811690e-01 -3.12240899e-01 -2.26259336e-01
-1.30232203e+00 9.90722835e-01 2.60485291e-01 4.87213671e-01
-6.37434602e-01 -5.00370681e-01 -7.90869236e-01 -6.87001944e-02
-4.71594632e-02 -4.39987808e-01 7.57653356e-01 -1.02612901e+00
-1.84063673e+00 6.30202711e-01 -3.70020926e-01 -3.91709119e-01
7.36675262e-01 1.62763998e-01 -6.31015003e-01 1.73522651e-01
-1.33737922e-01 4.32451308e-01 1.16081965e+00 -7.98853457e-01
-8.32530916e-01 -3.23020160e-01 -2.91807115e-01 8.20958465e-02
-1.66384473e-01 2.24589288e-01 -3.87986064e-01 -9.97787356e-01
2.68374532e-01 -6.51938617e-01 5.85450679e-02 1.28269419e-01
1.69876382e-01 -3.28008592e-01 9.47077453e-01 -1.19043577e+00
1.15814912e+00 -2.46369267e+00 1.07866414e-01 2.01198876e-01
-1.77718684e-01 4.32928652e-01 -4.34339374e-01 -9.76989120e-02
-1.72216073e-01 -3.53000939e-01 -3.52915466e-01 -1.60201550e-01
-2.33172312e-01 2.11754650e-01 -3.34954262e-01 5.21247685e-01
2.62839645e-01 9.03947294e-01 -8.02752137e-01 -5.46693385e-01
4.73886907e-01 9.11075771e-01 -2.75429070e-01 1.55129239e-01
6.26039803e-02 5.89236915e-01 -5.87514341e-01 6.07124805e-01
5.92004120e-01 2.19689921e-01 -3.30628455e-01 -4.70302582e-01
-1.12402119e-01 8.05412903e-02 -1.09372532e+00 1.73071814e+00
-5.45170009e-01 5.65367281e-01 -1.47019833e-01 -8.88245225e-01
1.08659625e+00 3.17823082e-01 7.38125801e-01 -1.26699567e+00
3.61411273e-01 3.80149394e-01 -8.01659226e-02 -6.11996293e-01
7.86385164e-02 -2.20027536e-01 -2.81609502e-02 -4.41979244e-02
1.00278609e-01 1.73513919e-01 -7.21696839e-02 -2.55791306e-01
1.08178449e+00 3.69486034e-01 4.88146655e-02 3.39209408e-01
9.34587002e-01 -2.41312563e-01 7.69070327e-01 9.57241356e-02
-3.29101175e-01 4.02375013e-01 2.32278779e-01 -2.56161094e-01
-1.05486095e+00 -9.40059602e-01 -1.81577653e-01 6.44401133e-01
3.01205188e-01 3.90859647e-03 -5.59542537e-01 -5.64644158e-01
-2.69715697e-01 7.23718941e-01 -4.14856076e-01 -5.30965090e-01
-7.83533156e-01 -3.91121149e-01 7.53340662e-01 4.28595066e-01
9.75364506e-01 -1.25232029e+00 -6.84008420e-01 1.52129576e-01
-3.44016135e-01 -1.17519236e+00 -7.04709649e-01 -3.94454986e-01
-7.23418415e-01 -8.67181718e-01 -9.23274279e-01 -8.73998106e-01
6.62763357e-01 -1.82026416e-01 1.57529339e-01 -6.62490353e-02
-4.98287708e-01 8.38036165e-02 -6.31755829e-01 9.65889245e-02
-3.30624372e-01 -5.19193351e-01 -1.89171776e-01 5.64118385e-01
3.22460830e-01 -5.03239453e-01 -5.47482312e-01 2.59831160e-01
-9.28262770e-01 -5.97476996e-02 6.68914378e-01 1.34113503e+00
6.34023845e-01 1.04897484e-01 4.31203932e-01 1.65025249e-01
2.89746135e-01 1.01627655e-01 -7.10245788e-01 1.28290132e-01
-3.24251801e-01 9.67949554e-02 6.45699918e-01 -6.95349455e-01
-1.20778501e+00 -3.29186022e-02 -2.40473181e-01 -6.85743928e-01
1.22650035e-01 3.07872713e-01 -4.03824121e-01 -1.43351272e-01
3.42020512e-01 1.00573981e+00 2.71205992e-01 -4.02358472e-01
2.45718181e-01 8.01060021e-01 6.81240857e-01 -8.15985426e-02
7.89755225e-01 5.29159248e-01 8.24285951e-03 -7.96220601e-01
-5.19854546e-01 -2.17926502e-01 -4.11652923e-01 -3.03143382e-01
8.57666254e-01 -8.12581658e-01 -7.50161231e-01 8.00217271e-01
-1.19147301e+00 1.28150843e-02 -7.11446166e-01 9.65662062e-01
-7.36645341e-01 4.83325809e-01 -5.92799127e-01 -8.36211920e-01
-3.56343925e-01 -1.20292485e+00 1.04076934e+00 1.50817469e-01
2.83730179e-01 -4.18938577e-01 -3.22959989e-01 3.86069149e-01
3.28950316e-01 4.36720192e-01 7.30723500e-01 -3.12775403e-01
-7.55051911e-01 -4.45484459e-01 -3.22772533e-01 9.11382556e-01
-9.49010327e-02 -3.15289736e-01 -8.67550373e-01 -3.35192680e-01
3.22849602e-01 -1.97443873e-01 1.03361583e+00 5.13568401e-01
1.16117764e+00 -3.10234785e-01 -3.72885424e-03 7.88061082e-01
1.50490558e+00 5.20547628e-01 1.19883204e+00 8.35734755e-02
5.06409228e-01 3.76677513e-01 8.20502579e-01 2.95866996e-01
1.26298249e-01 8.45115185e-01 -1.47622582e-02 2.10871652e-01
-6.65588796e-01 -4.02419388e-01 7.75419831e-01 1.04140568e+00
-2.39843309e-01 -2.63872612e-02 -3.51173818e-01 4.49358404e-01
-1.73734820e+00 -1.42584181e+00 3.53732198e-01 2.37502480e+00
6.78478062e-01 -7.61448145e-02 -2.14449018e-01 4.62333441e-01
8.48279536e-01 2.48098552e-01 -7.62357056e-01 -2.39046775e-02
-3.45296919e-01 3.21334809e-01 3.07763666e-01 5.07105768e-01
-7.51034081e-01 8.09015393e-01 4.29802847e+00 1.43951190e+00
-1.45760608e+00 2.77753979e-01 1.49988919e-01 -2.22910970e-01
1.82901755e-01 -1.38214171e-01 -6.55332446e-01 6.82153761e-01
8.77782524e-01 -1.90133676e-01 6.01723373e-01 5.91526747e-01
5.46961069e-01 9.34480354e-02 -7.39284217e-01 1.28319776e+00
1.56140581e-01 -9.49303567e-01 2.66858965e-01 9.74590853e-02
2.70760924e-01 -2.52580065e-02 4.86983284e-02 4.77853507e-01
-1.87727705e-01 -9.28046644e-01 7.94237912e-01 8.83739591e-01
1.44757318e+00 -5.98648012e-01 7.01810420e-01 3.51279795e-01
-1.57408166e+00 -3.12399656e-01 -1.09619185e-01 2.35319868e-01
3.08310926e-01 4.13553238e-01 -2.14478537e-01 6.28911972e-01
4.05055434e-01 6.94453537e-01 -3.46904015e-03 9.86394763e-01
-1.49595514e-01 3.89076084e-01 -4.17236805e-01 1.27814323e-01
3.32195684e-02 -1.67578831e-01 8.17479253e-01 9.09184039e-01
3.94408315e-01 2.39470392e-01 -1.92532446e-02 6.98627174e-01
1.80248424e-01 -9.83037334e-03 -2.64071435e-01 1.58893183e-01
3.75698894e-01 7.71096051e-01 -2.83584148e-01 -2.44494513e-01
-2.71833926e-01 1.11304677e+00 -1.25052750e-01 4.36792761e-01
-8.96960855e-01 -7.24671006e-01 4.27408963e-01 -8.87878239e-02
4.10864532e-01 -1.70806438e-01 -2.18254879e-01 -1.29875517e+00
6.16249681e-01 -6.40422106e-01 2.55660325e-01 -6.89432144e-01
-1.12781763e+00 6.41669691e-01 -2.24563703e-01 -1.95717275e+00
-2.83566058e-01 -4.25513983e-01 -2.57628530e-01 1.07748044e+00
-1.70046043e+00 -1.27161419e+00 -4.05570298e-01 9.15098846e-01
7.21810043e-01 -4.04936254e-01 6.40700638e-01 5.91532826e-01
-4.72784281e-01 1.07079256e+00 1.06266864e-01 4.31970984e-01
2.20710948e-01 -3.94530892e-01 -5.60281910e-02 9.52392399e-01
-3.27779412e-01 8.64854380e-02 2.09192276e-01 -5.55739403e-01
-1.15850222e+00 -1.24408460e+00 1.03548503e+00 8.25395659e-02
4.37121183e-01 3.05598136e-02 -6.98265433e-01 9.73451287e-02
-6.79383397e-01 4.02576625e-01 -1.42816585e-02 -9.30715501e-01
-3.76412809e-01 -3.72573882e-01 -1.42924786e+00 5.33733487e-01
1.18484545e+00 -7.74525821e-01 -6.83263481e-01 7.67533630e-02
7.54172623e-01 -3.87371898e-01 -8.87326419e-01 6.31244063e-01
9.34333861e-01 -5.76439738e-01 8.01579535e-01 -2.46328443e-01
4.51514393e-01 -3.75255108e-01 -3.02049994e-01 -8.82496476e-01
-1.20053170e-02 -3.40935439e-01 -1.52098686e-01 1.14863026e+00
-4.34021018e-02 -6.88013315e-01 4.98275459e-01 3.13515991e-01
-9.09268856e-02 -8.02233934e-01 -1.70855236e+00 -1.15645826e+00
-2.69285232e-01 -5.48381746e-01 3.37909579e-01 5.26805162e-01
-2.57115394e-01 -5.61632365e-02 -2.96860695e-01 9.54050422e-02
8.24624836e-01 -1.25340179e-01 3.63672644e-01 -8.88281524e-01
-4.72403206e-02 -4.05978829e-01 -7.16415346e-01 -9.99073088e-01
2.74939060e-01 -7.98353851e-01 1.90820411e-01 -1.37008715e+00
-2.75125235e-01 -1.79567546e-01 -5.03962040e-01 2.14144558e-01
2.06953302e-01 -4.11750264e-02 3.31612319e-01 2.71659076e-01
-3.91112193e-02 1.03074884e+00 1.69464934e+00 -2.58247823e-01
-3.84647757e-01 4.41054581e-03 8.16458538e-02 6.07456267e-01
3.51335287e-01 -3.86578798e-01 -2.55815476e-01 -1.13067254e-01
-7.88914919e-01 4.04529065e-01 6.56044066e-01 -1.08552933e+00
3.42456102e-01 -5.20375222e-02 1.72613829e-01 -4.84269887e-01
7.12458134e-01 -9.67051089e-01 -1.25873640e-01 7.85697520e-01
-4.02909070e-01 -4.77630526e-01 -2.30037440e-02 5.03204584e-01
-4.84576851e-01 1.65224567e-01 1.20266283e+00 1.08374640e-01
-8.67454708e-01 4.30197388e-01 -8.70683417e-02 -2.10348040e-01
1.02880383e+00 -5.99631250e-01 -1.46364287e-01 -1.77375853e-01
-8.00832689e-01 -1.06128886e-01 -6.77879378e-02 5.60664117e-01
1.07481527e+00 -1.48694527e+00 -9.59024429e-01 5.68672001e-01
5.40477931e-02 -2.17803389e-01 6.29998565e-01 7.85611928e-01
-3.15112531e-01 3.62202942e-01 -2.53639847e-01 -3.89277935e-01
-1.05381167e+00 3.92866254e-01 6.16331220e-01 -3.76149625e-01
-9.69685972e-01 7.56645143e-01 -1.37667224e-01 1.10381516e-02
4.19711173e-01 -3.97018969e-01 -1.82261959e-01 7.33177969e-03
6.49761796e-01 4.85064656e-01 -4.66687754e-02 -1.01642895e+00
-2.57377833e-01 8.85052502e-01 1.58170775e-01 -3.99650425e-01
1.15566504e+00 -4.43182439e-02 -2.26638317e-02 1.92178503e-01
1.48661351e+00 -3.23869377e-01 -1.25418365e+00 -4.64184612e-01
-4.26715314e-01 -5.46115935e-01 3.91557544e-01 -8.29233885e-01
-1.11929154e+00 7.57621706e-01 1.14688814e+00 -5.55599988e-01
1.61959791e+00 -3.12994510e-01 1.34538257e+00 2.38398034e-02
6.45566404e-01 -1.09051836e+00 -1.85583770e-01 5.18499076e-01
1.15742362e+00 -7.21092761e-01 -3.16065133e-01 -5.63737601e-02
-3.90421003e-01 9.41225231e-01 4.48927820e-01 -4.88336265e-01
6.55986845e-01 1.43035039e-01 -4.41073067e-02 3.06771755e-01
-4.86848861e-01 -2.96538621e-01 3.71147513e-01 5.89482784e-01
-8.28388855e-02 -1.53161541e-01 -8.18278313e-01 8.11032712e-01
-1.51365459e-01 5.81543863e-01 4.80536968e-02 5.59602380e-01
-1.98659122e-01 -9.60538745e-01 -2.92938769e-01 3.47336560e-01
-6.25064299e-02 1.05884345e-02 2.65469670e-01 5.78399718e-01
5.15449166e-01 7.57375956e-01 1.02118133e-02 -6.92242920e-01
5.83652020e-01 1.60048865e-02 5.08312821e-01 -6.74472898e-02
-3.08982402e-01 -5.67817949e-02 -1.61195546e-01 -8.38725030e-01
-3.67929161e-01 -5.20099998e-01 -1.48546982e+00 -3.29285040e-02
-3.23653787e-01 -1.55854225e-01 6.20238900e-01 1.18198299e+00
7.88884386e-02 5.30566216e-01 8.67922127e-01 -8.04848552e-01
-7.32235193e-01 -9.39219892e-01 -5.51618934e-01 4.67942208e-01
5.96096873e-01 -6.14659727e-01 -4.32556331e-01 2.01714680e-01] | [9.21805477142334, -6.464583396911621] |
a92447fc-988a-496c-be09-811c69807c5c | unsupervised-neural-machine-translation-1 | 1810.12703 | null | http://arxiv.org/abs/1810.12703v1 | http://arxiv.org/pdf/1810.12703v1.pdf | Unsupervised Neural Machine Translation Initialized by Unsupervised Statistical Machine Translation | Recent work achieved remarkable results in training neural machine
translation (NMT) systems in a fully unsupervised way, with new and dedicated
architectures that rely on monolingual corpora only. In this work, we propose
to define unsupervised NMT (UNMT) as NMT trained with the supervision of
synthetic bilingual data. Our approach straightforwardly enables the use of
state-of-the-art architectures proposed for supervised NMT by replacing
human-made bilingual data with synthetic bilingual data for training. We
propose to initialize the training of UNMT with synthetic bilingual data
generated by unsupervised statistical machine translation (USMT). The UNMT
system is then incrementally improved using back-translation. Our preliminary
experiments show that our approach achieves a new state-of-the-art for
unsupervised machine translation on the WMT16 German--English news translation
task, for both translation directions. | ['Atsushi Fujita', 'Benjamin Marie'] | 2018-10-30 | null | null | null | null | ['unsupervised-machine-translation'] | ['natural-language-processing'] | [ 5.44077396e-01 4.09347892e-01 -4.70274836e-01 -5.09997010e-01
-1.16610348e+00 -5.28351605e-01 1.13841069e+00 -4.36244756e-01
-5.27759016e-01 1.11166191e+00 1.93491995e-01 -9.14058745e-01
6.98667228e-01 -4.58712041e-01 -1.13418591e+00 -3.24065149e-01
5.40943325e-01 1.32074511e+00 -4.20863986e-01 -6.07907176e-01
-2.54558772e-01 -8.78491811e-03 -7.43222892e-01 6.76405311e-01
1.23752403e+00 2.11605579e-01 1.50657520e-01 5.69229662e-01
-1.76838264e-01 3.03983748e-01 -3.86309713e-01 -6.83887124e-01
6.57210588e-01 -1.07846642e+00 -9.64038432e-01 1.32945016e-01
5.12242973e-01 -1.80365816e-01 -8.30975473e-02 9.19358373e-01
4.80998278e-01 -2.96431333e-01 6.42912090e-01 -6.97909355e-01
-9.20974493e-01 1.18587816e+00 -2.59444714e-01 3.38726724e-03
-4.73373048e-02 1.55859053e-01 7.52698481e-01 -1.32371008e+00
1.06420326e+00 1.09487712e+00 5.08643210e-01 8.36480737e-01
-1.66479957e+00 -4.01534945e-01 -2.96992689e-01 -2.58680936e-02
-9.93387878e-01 -7.87170351e-01 4.22596425e-01 -2.53889591e-01
1.49781978e+00 1.69307902e-01 3.67466867e-01 1.49920964e+00
2.56609857e-01 8.05882752e-01 1.52044785e+00 -1.13442683e+00
-7.78882718e-03 2.97605366e-01 -3.91848922e-01 4.58984762e-01
3.47151072e-04 2.89786875e-01 -5.16204119e-01 1.98711574e-01
6.90995872e-01 -5.86139500e-01 2.59936929e-01 -1.72822371e-01
-1.92857873e+00 7.93062568e-01 7.68292416e-03 6.21694684e-01
-4.30420399e-01 -6.55949265e-02 4.58079308e-01 1.11923206e+00
1.03545237e+00 6.25795722e-01 -7.55375981e-01 -8.46376792e-02
-1.19197905e+00 -2.42914945e-01 8.00310552e-01 1.20709181e+00
8.45256865e-01 3.53903383e-01 -1.00212924e-01 1.04559481e+00
-1.13028161e-01 8.79298270e-01 9.04025078e-01 -5.97010672e-01
9.62741792e-01 2.34017745e-01 3.48565984e-03 -9.37227979e-02
-2.55677365e-02 -5.27948260e-01 -8.51688564e-01 -1.18462592e-01
2.39397034e-01 -3.69222313e-01 -1.23348355e+00 1.75058973e+00
-3.03760301e-02 -3.27810615e-01 6.47094071e-01 6.63298011e-01
4.05576348e-01 7.60466337e-01 -2.37466067e-01 -4.92161870e-01
7.39905298e-01 -1.53979671e+00 -7.19534755e-01 -3.24952662e-01
9.73664224e-01 -1.19278073e+00 1.02887821e+00 4.66826074e-02
-1.27001870e+00 -6.66407049e-01 -9.18527722e-01 5.12345741e-03
-4.73626882e-01 6.54359519e-01 3.56095731e-01 4.95956868e-01
-1.31281888e+00 6.84525132e-01 -1.06659925e+00 -7.73767710e-01
6.01935834e-02 5.69114625e-01 -5.75154722e-01 -4.59346808e-02
-1.29685402e+00 1.43941641e+00 4.67574328e-01 1.28724068e-01
-1.16354632e+00 -7.07326382e-02 -6.70350015e-01 -4.99263912e-01
4.39127237e-02 -9.86760855e-01 1.64499831e+00 -1.71349633e+00
-2.09557223e+00 1.11797190e+00 -3.60464394e-01 -7.45535970e-01
7.62730539e-01 -2.62179673e-01 -3.83488625e-01 -1.51223883e-01
3.79999489e-01 8.66390705e-01 8.15086663e-01 -1.10370255e+00
-2.43508473e-01 -5.97369857e-03 -3.96961272e-01 2.97109187e-01
-2.92558670e-01 3.78307045e-01 -1.69400543e-01 -7.93243349e-01
6.04985049e-03 -1.22022152e+00 -4.12838399e-01 -8.59244406e-01
-5.21869898e-01 3.51726525e-02 3.85591090e-01 -8.84195566e-01
7.73393095e-01 -1.55410886e+00 6.83376729e-01 -1.93123803e-01
-2.77822137e-01 4.41139758e-01 -6.04657471e-01 6.79103136e-01
-1.39612690e-01 2.72848289e-02 -5.69876194e-01 -7.98682690e-01
-4.98734005e-02 5.46152174e-01 -3.77057672e-01 1.20770514e-01
4.02007371e-01 1.44878936e+00 -9.32049274e-01 -3.88958901e-01
-6.59620017e-02 2.61820201e-02 -2.47450277e-01 2.79042035e-01
-4.98639166e-01 1.07721436e+00 -1.07720815e-01 4.02990997e-01
4.90163565e-01 8.01530387e-03 5.72331429e-01 2.60025054e-01
-2.69225091e-01 7.13131607e-01 -2.92109370e-01 2.27073669e+00
-6.58691823e-01 7.23646522e-01 -2.89593250e-01 -1.04092741e+00
9.47854936e-01 6.47878528e-01 1.29002839e-01 -6.95025086e-01
1.36240229e-01 1.06363642e+00 2.20736414e-01 -2.98530877e-01
3.92277122e-01 -2.54734367e-01 1.30324602e-01 9.40089226e-01
7.02220201e-01 -1.59848005e-01 4.13570344e-01 -1.16948821e-01
6.39862001e-01 7.29446590e-01 1.37446165e-01 -5.71677923e-01
2.71088630e-01 4.83128786e-01 2.62619108e-01 5.86138189e-01
4.74794924e-01 4.66103166e-01 -7.15225637e-02 -5.60349464e-01
-1.74936783e+00 -1.06947267e+00 1.42760381e-01 1.12410450e+00
-6.08347058e-01 -3.04752976e-01 -1.21666646e+00 -1.02338850e+00
-6.18023634e-01 8.63195956e-01 -5.03655970e-01 9.80740786e-02
-1.20486760e+00 -9.84096169e-01 7.01209128e-01 2.32018143e-01
3.90656114e-01 -1.12308490e+00 4.76289950e-02 4.41624582e-01
-6.06615007e-01 -1.19018209e+00 -4.65664357e-01 4.72706169e-01
-1.35888565e+00 -2.82041788e-01 -9.54801559e-01 -1.17839944e+00
9.93910968e-01 4.21023481e-02 1.38198423e+00 -3.58347774e-01
4.75328445e-01 -3.08045030e-01 -3.07199806e-01 -2.87955552e-01
-1.38075888e+00 7.73332357e-01 4.82558608e-01 2.74878740e-02
3.97621453e-01 -7.64278710e-01 -4.21337001e-02 3.69848549e-01
-8.54700565e-01 7.06522942e-01 1.10152197e+00 1.11565769e+00
6.15247428e-01 -7.93152034e-01 5.66074371e-01 -1.12035787e+00
6.29460633e-01 -1.30326957e-01 -4.03115243e-01 3.75808716e-01
-7.98786521e-01 3.93391460e-01 9.34144974e-01 -6.43679082e-01
-9.07495618e-01 -9.54393744e-02 -4.59010415e-02 -3.25956255e-01
-1.77438229e-01 4.88921762e-01 -2.42490396e-02 3.09078265e-02
1.01577389e+00 5.79769790e-01 -1.17841363e-01 -6.72098577e-01
6.56303525e-01 9.30607557e-01 4.33593303e-01 -8.12119722e-01
9.96843159e-01 1.23807833e-01 -2.17325553e-01 -3.27767342e-01
-6.20277166e-01 6.11458570e-02 -1.30770767e+00 1.45928919e-01
7.57984281e-01 -9.16028559e-01 5.56927681e-01 3.73010695e-01
-1.37796748e+00 -7.47929513e-01 -2.55552590e-01 7.45314240e-01
-8.50427866e-01 1.40048072e-01 -9.63205934e-01 -2.56533206e-01
-6.95509195e-01 -1.31445575e+00 1.10550368e+00 -4.42558497e-01
-2.90991038e-01 -1.04603028e+00 5.45961499e-01 4.41680819e-01
6.41278028e-01 -1.57484710e-01 8.36015821e-01 -8.37449491e-01
-4.82085496e-01 1.19862705e-02 1.30508602e-01 6.57169759e-01
2.21799910e-01 -4.97945458e-01 -7.39583910e-01 -4.17439848e-01
6.87933713e-02 -4.76733923e-01 7.33193636e-01 5.11290412e-03
8.62720683e-02 -4.45637822e-01 -1.41104802e-01 6.92867458e-01
1.19804597e+00 -1.16339549e-01 5.28592944e-01 5.24824500e-01
5.88287413e-01 4.24848467e-01 4.45081800e-01 -4.83291537e-01
1.44857422e-01 6.95830584e-01 -1.73595697e-01 -5.83687484e-01
-2.14964092e-01 -3.79324734e-01 8.32957745e-01 1.70645404e+00
-4.23607796e-01 -5.26091754e-02 -9.53716695e-01 6.17034435e-01
-1.97121847e+00 -6.19148314e-01 -3.86772119e-02 2.03134060e+00
1.29752195e+00 9.55305919e-02 -1.21090002e-01 -4.34239715e-01
6.61893249e-01 -2.17611283e-01 -1.74109250e-01 -8.03071380e-01
-3.42733562e-01 5.78322172e-01 5.73257506e-01 6.04326844e-01
-8.91751528e-01 1.66819966e+00 6.70159197e+00 7.45125234e-01
-1.26835895e+00 7.46964455e-01 6.19612932e-01 1.03555024e-01
-3.28898609e-01 2.26936623e-01 -5.92728257e-01 1.45069554e-01
1.49601245e+00 1.04579031e-01 7.66304612e-01 4.39171195e-01
2.45699719e-01 4.28246677e-01 -1.28698659e+00 4.20785695e-01
1.70991436e-01 -1.33293784e+00 4.38844442e-01 3.95686217e-02
1.49640763e+00 7.09078431e-01 1.13148540e-02 3.85140032e-01
4.91811693e-01 -8.37797880e-01 6.29092336e-01 8.39181319e-02
1.26597774e+00 -3.99037421e-01 7.25182652e-01 5.79038203e-01
-5.51538169e-01 5.35702705e-01 -4.81930554e-01 -2.71433797e-02
1.13970794e-01 4.11104649e-01 -1.26284540e+00 9.38357472e-01
1.61106586e-02 7.13008702e-01 -3.72529685e-01 1.84349775e-01
-6.22469842e-01 8.08984101e-01 -3.06738079e-01 2.29854003e-01
4.89056438e-01 -4.93679732e-01 5.86622059e-01 1.47990823e+00
5.28376698e-01 -6.08486772e-01 1.97146863e-01 6.61161959e-01
-4.35959697e-01 5.23635566e-01 -8.07441056e-01 -2.95372307e-01
-1.19678378e-01 9.16170657e-01 -5.28959870e-01 -7.71823287e-01
-3.67375851e-01 1.48086059e+00 4.86018360e-01 5.62895834e-01
-5.46657979e-01 8.87122899e-02 1.50937691e-01 -1.06248692e-01
1.13018632e-01 -4.69564468e-01 -3.45556080e-01 -1.76489449e+00
1.56328693e-01 -1.32906699e+00 -2.27840185e-01 -6.41972244e-01
-1.09971309e+00 1.33201206e+00 -1.82256266e-01 -1.48755789e+00
-7.80658126e-01 -5.68087399e-01 -3.65458220e-01 1.13280559e+00
-1.30902314e+00 -1.62411988e+00 5.43169439e-01 3.05833191e-01
9.68315423e-01 -7.15052783e-01 1.28323948e+00 2.00788185e-01
-2.06063032e-01 7.25878716e-01 5.56461632e-01 2.80133665e-01
1.10758579e+00 -1.13245261e+00 1.27323747e+00 1.15360904e+00
6.47355616e-01 7.10363746e-01 6.89955533e-01 -7.61224210e-01
-1.19423771e+00 -1.26208043e+00 1.66442454e+00 -8.23041320e-01
7.14112759e-01 -6.69369936e-01 -3.59438658e-01 1.13777077e+00
9.56412613e-01 -5.01649797e-01 5.95210671e-01 5.77551201e-02
-4.03338134e-01 1.73582375e-01 -7.15246320e-01 7.81646013e-01
1.02125275e+00 -6.99724793e-01 -8.64779830e-01 7.55160868e-01
9.22489166e-01 -2.87115306e-01 -7.75219440e-01 5.13959110e-01
3.21346909e-01 -3.72072756e-01 5.73864400e-01 -9.30206060e-01
6.40181780e-01 -2.52105117e-01 -1.61316484e-01 -1.88382590e+00
9.26679652e-03 -1.07566404e+00 2.05333889e-01 7.80593097e-01
1.24023557e+00 -7.76130915e-01 4.93788004e-01 -3.33510756e-01
-5.24801433e-01 -5.02921999e-01 -1.08876801e+00 -8.60856056e-01
5.54248631e-01 -1.13000892e-01 4.68082279e-01 1.26661670e+00
9.95723158e-02 8.99131060e-01 -7.24008858e-01 -2.87943095e-01
4.22953755e-01 2.27415830e-01 1.02404821e+00 -7.50746608e-01
-6.08536959e-01 -2.41383776e-01 1.63655698e-01 -1.13146305e+00
2.26418287e-01 -1.38353300e+00 1.13798849e-01 -1.51192832e+00
3.37342083e-01 1.04151219e-01 -1.20746046e-01 5.46531618e-01
1.89931002e-02 6.81272447e-01 1.36072775e-02 6.73832238e-01
-3.13172996e-01 4.25373018e-01 1.43835068e+00 -3.39712888e-01
-9.30509642e-02 -1.11284509e-01 -2.02299103e-01 3.30109417e-01
9.46704090e-01 -7.16378629e-01 -1.26205549e-01 -1.20909810e+00
8.85005742e-02 -1.38733760e-01 -2.13691533e-01 -7.47875810e-01
-1.03445344e-01 -2.19187811e-02 2.19516709e-01 -3.18708628e-01
1.89242065e-02 -4.78063256e-01 1.39796242e-01 3.69592518e-01
-4.70131308e-01 5.68041742e-01 1.76539794e-01 -2.58255424e-03
-3.36417794e-01 3.29801179e-02 5.89322627e-01 -4.65081125e-01
-7.84420967e-02 -2.59044580e-03 -5.62343299e-01 -2.55337358e-01
3.78498614e-01 2.00767085e-01 -2.56890446e-01 -1.72940016e-01
-8.94033968e-01 -1.07643954e-01 6.30833745e-01 4.66407478e-01
1.37903258e-01 -1.43261027e+00 -1.41693306e+00 4.26710606e-01
1.54392913e-01 -4.13992614e-01 -4.81337935e-01 1.11350834e+00
-4.38169003e-01 7.62176096e-01 -4.18168217e-01 -6.39482081e-01
-8.27784002e-01 4.95666623e-01 2.82257259e-01 -7.40069330e-01
-3.39271843e-01 3.75434488e-01 -1.96808338e-01 -1.19059694e+00
-4.29887652e-01 -2.25578040e-01 3.09254438e-01 -4.23642069e-01
4.97232154e-02 -1.74758822e-01 4.66259420e-01 -8.28089774e-01
2.18393862e-01 3.73482049e-01 -1.88949496e-01 -8.52762282e-01
1.16804469e+00 3.48927849e-03 -4.56800550e-01 5.01671910e-01
1.00595748e+00 1.32508799e-01 -5.41956902e-01 -4.65998679e-01
1.06663577e-01 6.12608977e-02 -4.38335121e-01 -1.18138695e+00
-5.20854115e-01 9.43134427e-01 3.79974812e-01 -3.20911914e-01
9.53535497e-01 -1.68784901e-01 9.85370934e-01 8.62370789e-01
7.35591590e-01 -1.15660453e+00 -4.03766692e-01 9.08203840e-01
6.47214711e-01 -1.39250135e+00 -4.29214746e-01 -4.44731340e-02
-4.40624684e-01 1.20025766e+00 3.50446820e-01 -3.45403813e-02
4.69566062e-02 1.77986354e-01 6.83057010e-01 3.25694442e-01
-8.54042292e-01 -1.68157801e-01 3.41233611e-01 4.09962773e-01
7.18301594e-01 3.91348839e-01 -7.68704951e-01 8.93981680e-02
-4.01395708e-01 -8.37197620e-03 1.73425049e-01 8.61437619e-01
-1.67994767e-01 -2.00583577e+00 -2.33232036e-01 7.18581378e-02
-5.24401963e-01 -7.32554436e-01 -8.80627871e-01 6.91315234e-01
-4.70537785e-03 8.41747940e-01 -2.06118926e-01 -5.26897252e-01
8.02131370e-02 5.91940463e-01 7.47714579e-01 -8.58426094e-01
-8.12964380e-01 4.31732953e-01 5.33060849e-01 -1.83683291e-01
-5.96315503e-01 -6.01363361e-01 -6.87958002e-01 4.36562188e-02
-1.36575311e-01 5.05412877e-01 9.58579540e-01 1.22551012e+00
3.85027736e-01 2.08348125e-01 6.62002921e-01 -9.26807344e-01
-6.26825094e-01 -1.79094434e+00 3.61953884e-01 2.40144327e-01
1.26849979e-01 -1.91323142e-02 -5.63866980e-02 4.71668035e-01] | [11.558069229125977, 10.332286834716797] |
f50c3784-cdfb-4a98-9705-73d546518343 | nutribullets-hybrid-multi-document-health | 2104.03465 | null | https://arxiv.org/abs/2104.03465v1 | https://arxiv.org/pdf/2104.03465v1.pdf | Nutribullets Hybrid: Multi-document Health Summarization | We present a method for generating comparative summaries that highlights similarities and contradictions in input documents. The key challenge in creating such summaries is the lack of large parallel training data required for training typical summarization systems. To this end, we introduce a hybrid generation approach inspired by traditional concept-to-text systems. To enable accurate comparison between different sources, the model first learns to extract pertinent relations from input documents. The content planning component uses deterministic operators to aggregate these relations after identifying a subset for inclusion into a summary. The surface realization component lexicalizes this information using a text-infilling language model. By separately modeling content selection and realization, we can effectively train them with limited annotations. We implemented and tested the model in the domain of nutrition and health -- rife with inconsistencies. Compared to conventional methods, our framework leads to more faithful, relevant and aggregation-sensitive summarization -- while being equally fluent. | ['Regina Barzilay', 'Tao Lei', 'Lili Yu', 'Darsh J Shah'] | 2021-04-08 | null | null | null | null | ['text-infilling'] | ['natural-language-processing'] | [ 7.31404483e-01 9.70894098e-01 -2.83686191e-01 -2.61942148e-01
-1.25958335e+00 -5.90140522e-01 7.90204883e-01 1.12450314e+00
-1.61394924e-01 1.05986500e+00 1.11259711e+00 -1.39661238e-01
-1.01448938e-01 -7.79007792e-01 -5.43592930e-01 -2.48880032e-02
1.73749000e-01 9.15244341e-01 1.26791432e-01 -5.08206606e-01
5.59006333e-01 -5.63765131e-02 -1.42060602e+00 8.12518358e-01
1.47437644e+00 3.25436532e-01 3.22849840e-01 8.11511934e-01
-5.69430888e-01 9.78535295e-01 -9.98688638e-01 -4.38550651e-01
-2.36131519e-01 -1.03611803e+00 -1.12940276e+00 1.05094671e-01
3.18657994e-01 -2.92183191e-01 3.36936831e-01 8.11281919e-01
5.27135193e-01 4.38866653e-02 9.09384966e-01 -8.61119747e-01
-3.94615412e-01 1.46930754e+00 -2.25629494e-01 -5.21064028e-02
8.81639600e-01 -1.91973239e-01 1.10531235e+00 -5.03663242e-01
1.07425225e+00 1.25725067e+00 5.72542787e-01 5.91738880e-01
-1.35144448e+00 1.99163146e-02 2.57856548e-01 -2.74241865e-01
-7.95380235e-01 -7.16971636e-01 4.95393515e-01 -2.89071649e-01
1.36075425e+00 7.07326591e-01 8.69703054e-01 7.72509933e-01
2.19960451e-01 8.48697603e-01 4.51943398e-01 -7.80138493e-01
2.28634819e-01 1.82260007e-01 1.75174013e-01 4.00472909e-01
6.14308000e-01 -6.18929684e-01 -7.33088493e-01 -2.30284214e-01
1.26698911e-01 -4.79621202e-01 -2.16421783e-01 1.45990580e-01
-1.20511222e+00 7.10227609e-01 -7.45471045e-02 2.77598888e-01
-7.58303881e-01 -5.21067195e-02 5.74900091e-01 1.87552586e-01
7.43375897e-01 9.43240404e-01 -2.46097028e-01 1.12885796e-01
-1.50465667e+00 7.26265550e-01 1.30786312e+00 1.27971566e+00
4.14113432e-01 -2.46926099e-01 -7.42576838e-01 6.57286048e-01
9.93159339e-02 2.96505094e-01 3.78752440e-01 -9.06513333e-01
7.52078652e-01 7.82052159e-01 1.71163291e-01 -9.82997954e-01
-3.71626079e-01 -4.90990579e-01 -6.53289616e-01 -2.24948794e-01
-7.43128285e-02 -2.68494755e-01 -6.47657394e-01 1.47719073e+00
2.27877662e-01 -5.05607665e-01 4.28989381e-01 2.73188651e-01
1.03472984e+00 7.46688008e-01 1.30779609e-01 -6.59256518e-01
1.43623340e+00 -8.11579883e-01 -1.05004597e+00 -2.11578861e-01
8.06572676e-01 -8.35106492e-01 7.33814955e-01 2.99890995e-01
-1.66966319e+00 -2.20386758e-01 -1.13455951e+00 -4.36721981e-01
-3.60471785e-01 1.37730867e-01 3.18368018e-01 2.21421972e-01
-1.15306187e+00 8.29825342e-01 -4.72110629e-01 -5.24842441e-01
2.12599710e-01 1.29281566e-01 -1.43433660e-01 8.19063466e-03
-1.01529837e+00 1.01390529e+00 9.71933365e-01 -3.16278547e-01
-2.72346258e-01 -8.44760180e-01 -9.56397593e-01 2.74419039e-01
5.17293274e-01 -1.35743022e+00 1.70027363e+00 -6.42941892e-01
-1.43782699e+00 5.26454628e-01 -3.20331961e-01 -6.46614015e-01
5.51579893e-01 -4.19627577e-01 -1.30127326e-01 3.00032973e-01
5.24157465e-01 8.72558594e-01 3.93190116e-01 -1.29280698e+00
-7.70981669e-01 -3.91469188e-02 -1.28770977e-01 5.93454659e-01
1.01226121e-01 1.93061709e-01 -2.68771708e-01 -7.93190181e-01
-1.31085023e-01 -3.36378336e-01 -3.70850116e-01 -6.73729777e-01
-1.08109307e+00 -2.91396976e-01 2.90628135e-01 -9.52162862e-01
1.71657109e+00 -1.39696753e+00 2.87920684e-01 1.03407785e-01
9.73160863e-02 6.65229484e-02 -1.27246350e-01 1.08920777e+00
1.35925204e-01 4.78430539e-01 -4.51702237e-01 -4.16763097e-01
9.08697322e-02 5.22523634e-02 -5.04849970e-01 -3.72942537e-01
4.53269064e-01 8.77606392e-01 -1.22691011e+00 -8.76447439e-01
-1.74639165e-01 1.16342455e-01 -7.63052166e-01 2.57163584e-01
-9.70505893e-01 1.94314197e-01 -4.41574514e-01 2.72190928e-01
2.36633092e-01 4.39147046e-03 4.14337993e-01 -2.12640449e-01
-2.93031126e-01 8.93391788e-01 -1.08913922e+00 1.95980394e+00
-3.60557556e-01 3.57486606e-01 -3.38208973e-01 -8.33105624e-01
7.95937181e-01 3.68404716e-01 2.74273813e-01 -5.40112019e-01
-1.16827823e-02 3.09762537e-01 -2.68008500e-01 -6.47133708e-01
1.14620876e+00 -4.16967385e-02 -3.81940991e-01 6.91015601e-01
1.64754495e-01 -6.94349051e-01 9.66485083e-01 7.83488452e-01
8.68852019e-01 2.80067533e-01 6.79045558e-01 -3.43275756e-01
3.85876030e-01 6.21840775e-01 3.22519273e-01 8.32538247e-01
7.89213777e-01 6.20573997e-01 8.01848710e-01 -9.04982463e-02
-9.53669548e-01 -5.70540905e-01 3.67887855e-01 8.44162405e-01
-2.48384818e-01 -1.14570010e+00 -1.03679287e+00 -6.25479639e-01
-2.28147328e-01 1.47455299e+00 -4.52609599e-01 -6.22900985e-02
-7.91971862e-01 -5.09098768e-01 3.91121447e-01 2.95803070e-01
1.24955878e-01 -1.15454769e+00 -9.41351593e-01 5.82401633e-01
-6.41567767e-01 -7.30117798e-01 -4.69866186e-01 1.03472494e-01
-8.61820340e-01 -9.76806462e-01 -4.38058257e-01 -6.84247196e-01
6.46005929e-01 -2.03235418e-01 1.53894961e+00 3.45102772e-02
-1.01539448e-01 2.22066894e-01 -3.91583711e-01 -9.09945488e-01
-1.16968966e+00 3.74130160e-01 -4.68966365e-01 -6.89500034e-01
-6.88623339e-02 -3.37813795e-01 -3.52309108e-01 -4.64489281e-01
-1.18690002e+00 6.06900513e-01 5.97073317e-01 6.77331150e-01
5.94135642e-01 -8.66103619e-02 9.04665649e-01 -1.33657539e+00
1.29133940e+00 -5.01143992e-01 -1.43978372e-01 6.69226706e-01
-5.96944809e-01 4.26787436e-01 7.22176969e-01 -1.56951964e-01
-1.29510474e+00 -2.08542690e-01 -6.30383715e-02 6.73166156e-01
-7.12082861e-03 9.68902230e-01 -6.37442172e-02 1.00474024e+00
8.22783530e-01 1.09554455e-01 -9.22310203e-02 -4.02656972e-01
8.57015610e-01 5.30291557e-01 6.82461977e-01 -6.36338174e-01
2.17540532e-01 1.49979433e-02 -3.06466013e-01 -8.69868219e-01
-7.34129608e-01 -2.82214820e-01 -6.36073351e-01 3.51518132e-02
6.22110546e-01 -7.97211707e-01 -6.53333962e-02 -6.66036829e-02
-1.44886220e+00 -2.47186556e-01 -9.75077331e-01 2.63566285e-01
-5.62704444e-01 3.83051157e-01 -4.40683186e-01 -6.62383318e-01
-9.01273489e-01 -6.10903561e-01 1.24611962e+00 3.33990216e-01
-9.80306804e-01 -8.47501695e-01 3.33333910e-01 6.40999898e-02
3.95451039e-01 4.72502261e-01 1.17646503e+00 -1.05457616e+00
-2.34367967e-01 -5.99400625e-02 1.35266393e-01 -3.73430029e-02
2.56620318e-01 1.75886646e-01 -4.96246576e-01 2.45591566e-01
-1.73952475e-01 -3.37074965e-01 8.83896053e-01 4.13595945e-01
6.57385707e-01 -9.39818263e-01 -4.36061561e-01 -1.04759276e-01
1.05604386e+00 1.10770524e-01 4.76313531e-01 2.01772019e-01
4.00214702e-01 1.07545018e+00 5.47997236e-01 4.89970803e-01
6.13671482e-01 4.26605046e-01 -1.70150355e-01 -1.20487481e-01
-3.80387753e-01 -5.82017899e-01 1.91362984e-02 9.41595972e-01
2.41514474e-01 -6.32960796e-01 -6.63855255e-01 7.87137449e-01
-1.99141729e+00 -1.08777022e+00 -5.82756773e-02 1.81167078e+00
1.40795696e+00 1.22273982e-01 2.60520726e-01 1.08904071e-01
3.90834928e-01 2.18577646e-02 -1.81322485e-01 -6.22180223e-01
-7.77899697e-02 2.20483392e-01 -2.24424340e-02 6.51860237e-01
-6.83918238e-01 9.77392197e-01 6.28292990e+00 5.91148734e-01
-5.72992504e-01 -2.99743295e-01 6.07830346e-01 -1.67122364e-01
-8.71058166e-01 1.64341912e-01 -7.19506621e-01 1.73684299e-01
1.04421759e+00 -6.46771491e-01 -2.01432258e-01 3.51366997e-01
5.42644978e-01 -5.23362994e-01 -1.26473618e+00 2.24618524e-01
2.95015097e-01 -1.79063594e+00 6.04188800e-01 -3.65896732e-01
7.41652787e-01 -4.84358996e-01 -5.56318402e-01 8.99512768e-02
5.43944836e-01 -9.15812194e-01 1.05349946e+00 6.61032498e-01
8.12851846e-01 -7.62955070e-01 6.93572342e-01 3.91414285e-01
-8.20166707e-01 2.85214424e-01 -2.93112159e-01 5.62369227e-02
4.81293589e-01 7.12580144e-01 -1.10475564e+00 1.20399117e+00
6.67640939e-02 5.10953367e-01 -4.66291130e-01 9.32095289e-01
-4.04987633e-01 3.39297384e-01 -2.46604636e-01 -1.46647930e-01
-6.70786053e-02 -1.95672616e-01 6.10642374e-01 1.75710893e+00
4.06488717e-01 1.99766815e-01 2.39291802e-01 8.15972447e-01
-1.68827564e-01 5.32973886e-01 -5.74604154e-01 -1.35941938e-01
4.48247015e-01 9.97574866e-01 -8.59519243e-01 -8.97213757e-01
3.32976244e-02 8.54516745e-01 1.64631188e-01 1.57342151e-01
-3.50077748e-01 -5.11534810e-01 -8.26435015e-02 7.04892278e-02
5.14788404e-02 1.62821904e-01 -4.95516241e-01 -1.06718075e+00
2.18050089e-02 -1.20303464e+00 5.89085400e-01 -6.24743104e-01
-8.31043839e-01 7.46479332e-01 5.57331443e-01 -7.70664513e-01
-8.43070924e-01 2.34681115e-01 -8.84302258e-01 8.73116255e-01
-1.19510365e+00 -1.17748797e+00 1.02385089e-01 -2.42477972e-02
9.10579145e-01 1.30400345e-01 9.32920456e-01 -2.61906505e-01
-4.73205537e-01 1.55376747e-01 -2.45096847e-01 -2.78859556e-01
6.78139627e-01 -1.56534159e+00 7.07787275e-01 1.04484630e+00
3.28910053e-02 8.37296367e-01 1.11386645e+00 -1.18239617e+00
-8.12083721e-01 -1.04182959e+00 1.71464789e+00 -2.41407663e-01
3.20380837e-01 -1.49616361e-01 -9.50503767e-01 4.49658126e-01
9.34254646e-01 -1.02329218e+00 8.18616569e-01 -9.02868286e-02
-8.34208280e-02 1.09416299e-01 -9.90553141e-01 7.55033255e-01
9.99054611e-01 -1.51277110e-01 -1.18070257e+00 5.83634555e-01
9.75011706e-01 -4.63656187e-01 -6.76308215e-01 9.43777561e-02
3.67403179e-01 -7.61829019e-01 6.48248017e-01 -7.43471682e-01
8.59379768e-01 -2.31574565e-01 1.60496369e-01 -1.67772245e+00
-7.64009282e-02 -9.86222625e-01 -5.38640283e-02 1.61518478e+00
7.97366500e-01 -3.15453142e-01 3.23716700e-01 6.25308692e-01
-4.80721802e-01 -5.74620545e-01 -3.31794620e-01 -3.73801351e-01
6.66236207e-02 -3.71947847e-02 7.50152409e-01 6.22104168e-01
6.17557824e-01 8.04904103e-01 -1.71042293e-01 -3.20970237e-01
3.46858531e-01 2.37752646e-01 6.62646353e-01 -1.17665565e+00
-2.56597430e-01 -6.21215761e-01 4.87703294e-01 -7.11804271e-01
-1.08209826e-01 -9.49199498e-01 2.29271263e-01 -2.50521803e+00
2.88334399e-01 -1.32716388e-01 3.72413069e-01 5.12720644e-01
-3.79880846e-01 -3.64862740e-01 2.33271942e-01 1.39007062e-01
-6.42962635e-01 3.07874262e-01 9.32417154e-01 -1.18322328e-01
-7.50405908e-01 -8.96119550e-02 -1.28696775e+00 4.76675779e-01
9.86494184e-01 -5.84331870e-01 -7.43852019e-01 -2.48045996e-01
4.16206270e-01 3.20179731e-01 -2.34534875e-01 -7.91962504e-01
3.76357615e-01 -1.38641700e-01 3.00672382e-01 -9.62980866e-01
-2.93406069e-01 -1.45203084e-01 3.67732704e-01 5.45196176e-01
-9.68706548e-01 3.17209870e-01 3.42045575e-01 2.35945016e-01
-1.40415773e-01 -4.86783594e-01 2.00142443e-01 -4.27237809e-01
-1.60957322e-01 -3.01172674e-01 -5.85309029e-01 3.88648570e-01
4.99049246e-01 1.18008405e-02 -6.45609319e-01 -4.31565493e-01
-3.20783526e-01 4.04772192e-01 3.94872606e-01 2.02999756e-01
5.02570987e-01 -8.65921617e-01 -1.20875752e+00 -9.33307037e-02
7.50570819e-02 4.07352090e-01 3.80703472e-02 4.70266193e-01
-6.92159653e-01 6.15804672e-01 -1.52856663e-01 -1.18222453e-01
-1.15773439e+00 3.97007346e-01 -3.59699279e-02 -7.48217046e-01
-7.43186891e-01 4.10818368e-01 -1.18674368e-01 -1.72128037e-01
7.70115629e-02 -7.20235944e-01 -5.97927988e-01 5.03987491e-01
8.22231889e-01 4.00261670e-01 2.59821087e-01 -3.28313589e-01
-1.03756994e-01 6.81730062e-02 -2.00045601e-01 -5.20813346e-01
1.45078027e+00 -2.21079424e-01 -3.20585757e-01 2.81042814e-01
6.62979782e-01 3.49069357e-01 -8.03394794e-01 -6.93864599e-02
6.29578948e-01 1.90180302e-01 -3.53116244e-01 -1.11422455e+00
-2.00898796e-01 3.35873991e-01 -4.27353978e-01 4.99923795e-01
1.04266715e+00 -2.57389974e-02 6.92769349e-01 4.66783494e-01
-2.05050021e-01 -1.22058988e+00 -2.19810288e-02 3.68486494e-01
1.26118398e+00 -7.24445283e-01 4.60173011e-01 -4.92963552e-01
-8.55112195e-01 1.22647822e+00 3.32148224e-01 2.21197829e-01
-2.11308599e-02 2.76546448e-01 -7.68012553e-02 -3.75428975e-01
-1.02883220e+00 -1.80796504e-01 5.14520764e-01 5.99764645e-01
7.26523519e-01 -7.32821003e-02 -7.78232276e-01 4.66232628e-01
-5.35327375e-01 -4.52341773e-02 8.72936606e-01 1.10511696e+00
-6.27426505e-01 -1.32904840e+00 -1.59177780e-01 6.60852492e-01
-5.22973597e-01 -3.40825707e-01 -7.86764562e-01 6.27135992e-01
-1.01862833e-01 1.26014256e+00 -2.83086319e-02 8.77956450e-02
6.08891368e-01 2.59719938e-01 4.63612586e-01 -1.15646660e+00
-1.02137148e+00 3.72199893e-01 8.88961136e-01 -2.77656138e-01
-5.32878578e-01 -7.08507597e-01 -1.33778048e+00 -4.74588387e-02
-2.12958574e-01 6.02714539e-01 5.65689445e-01 8.65208864e-01
4.99720663e-01 8.68726850e-01 9.98894796e-02 -7.49561608e-01
-5.07450819e-01 -9.58050966e-01 -1.21013530e-01 3.14925641e-01
1.63992643e-01 3.66977677e-02 2.10318908e-01 3.50299329e-01] | [12.352731704711914, 9.336812973022461] |
9cc0cd68-2da9-4914-aa63-26b5bc0421bc | instance-shadow-detection | 1911.07034 | null | https://arxiv.org/abs/1911.07034v2 | https://arxiv.org/pdf/1911.07034v2.pdf | Instance Shadow Detection | Instance shadow detection is a brand new problem, aiming to find shadow instances paired with object instances. To approach it, we first prepare a new dataset called SOBA, named after Shadow-OBject Association, with 3,623 pairs of shadow and object instances in 1,000 photos, each with individual labeled masks. Second, we design LISA, named after Light-guided Instance Shadow-object Association, an end-to-end framework to automatically predict the shadow and object instances, together with the shadow-object associations and light direction. Then, we pair up the predicted shadow and object instances, and match them with the predicted shadow-object associations to generate the final results. In our evaluations, we formulate a new metric named the shadow-object average precision to measure the performance of our results. Further, we conducted various experiments and demonstrate our method's applicability on light direction estimation and photo editing. | ['Xiao-Wei Hu', 'Chi-Wing Fu', 'Pheng-Ann Heng', 'Tianyu Wang', 'Qiong Wang'] | 2019-11-16 | instance-shadow-detection-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Wang_Instance_Shadow_Detection_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Wang_Instance_Shadow_Detection_CVPR_2020_paper.pdf | cvpr-2020-6 | ['shadow-detection'] | ['computer-vision'] | [ 7.12175190e-01 1.22111514e-01 2.09439173e-01 -8.76038909e-01
-5.61551690e-01 -3.88339281e-01 6.26244366e-01 -4.46670294e-01
-2.29841582e-02 7.40955949e-01 3.90350297e-02 1.29130810e-01
1.26387671e-01 -5.94274580e-01 -7.68853366e-01 -6.70175076e-01
3.02335799e-01 5.00668526e-01 7.10661769e-01 3.79679948e-01
5.46378851e-01 6.44130170e-01 -1.65761673e+00 3.40350538e-01
9.73391831e-01 1.29267263e+00 6.21626616e-01 6.44949377e-01
-1.93607852e-01 6.74004614e-01 -6.20633841e-01 -3.42866957e-01
4.61418271e-01 -3.70956451e-01 -4.21424001e-01 3.54047090e-01
1.07522130e+00 -4.17070270e-01 6.08789325e-02 8.27762127e-01
6.28012240e-01 1.95074037e-01 4.30037290e-01 -1.51876533e+00
-6.40003562e-01 8.96749273e-02 -8.28050017e-01 4.21913713e-02
4.28787529e-01 5.35132587e-01 8.32663119e-01 -1.55939817e+00
6.98667467e-01 1.34562922e+00 5.44840157e-01 3.51740092e-01
-1.08467472e+00 -7.50588655e-01 2.66491652e-01 3.74797732e-01
-1.27833438e+00 -7.26303041e-01 6.78317368e-01 -1.02028921e-01
5.17464876e-01 4.94668752e-01 6.21726930e-01 7.15846241e-01
1.27675468e-02 1.02722287e+00 1.66401410e+00 -3.42968166e-01
2.03125462e-01 3.58291388e-01 1.88679695e-01 8.62315178e-01
-2.08440833e-02 1.86732009e-01 -7.19820619e-01 -1.21802345e-01
3.32990676e-01 -4.85214218e-02 -4.43587333e-01 -2.33110145e-01
-1.47163498e+00 5.71269058e-02 5.16026497e-01 -3.82308543e-01
-2.85350144e-01 1.81858435e-01 -1.66845888e-01 -2.78865457e-01
3.17388862e-01 1.73023582e-01 -3.97193521e-01 4.00170237e-01
-6.10022604e-01 3.18349987e-01 5.30915201e-01 1.43577838e+00
1.00634944e+00 -4.47686940e-01 -9.24844682e-01 8.25643897e-01
1.49769247e-01 9.62249279e-01 -1.23964153e-01 -9.34662879e-01
2.91969568e-01 7.43725896e-01 4.34214741e-01 -8.68556440e-01
-4.04742897e-01 -7.01403469e-02 -4.94061232e-01 1.94584563e-01
2.03217760e-01 1.35428011e-01 -1.16650057e+00 1.44518840e+00
6.66320741e-01 6.70960665e-01 -7.22207874e-03 1.08389735e+00
9.95370507e-01 6.20990455e-01 -1.11230455e-01 -4.16774273e-01
1.40744901e+00 -1.41076005e+00 -6.47190988e-01 -5.18631220e-01
-2.69653229e-03 -1.17309058e+00 1.53645182e+00 2.64028788e-01
-1.01069403e+00 -5.92532098e-01 -7.17233956e-01 -2.05988020e-01
-3.26312721e-01 6.15091980e-01 7.12974072e-01 3.06580603e-01
-5.66885591e-01 2.61803329e-01 -1.68652713e-01 -2.11007744e-01
7.18719482e-01 8.92307311e-02 1.91294312e-01 -2.51735479e-01
-5.55020630e-01 7.93768883e-01 2.43922293e-01 2.42297724e-01
-1.15029120e+00 -9.41644371e-01 -4.00312662e-01 -1.10583238e-01
7.76098549e-01 -6.51892006e-01 1.24276388e+00 -8.02524507e-01
-1.32687831e+00 9.15981114e-01 -6.19661748e-01 3.85416113e-03
3.70569289e-01 -2.06158400e-01 -5.61590195e-01 -4.07987922e-01
2.09115833e-01 6.79665625e-01 1.02587628e+00 -1.92104399e+00
-1.26095295e+00 -3.35623443e-01 1.91150442e-01 3.16217005e-01
-5.18597253e-02 -2.49918140e-02 -8.18681419e-01 -3.87823313e-01
2.75937498e-01 -9.35594738e-01 -9.32980105e-02 1.50138631e-01
-8.41264725e-01 -3.12591374e-01 1.05543923e+00 -4.48136419e-01
9.84924793e-01 -2.31659007e+00 -2.71321982e-01 2.57307976e-01
2.07082570e-01 6.78105056e-02 -1.42860144e-01 -2.62712855e-02
1.99931741e-01 -5.60145199e-01 -4.15759653e-01 -5.32974720e-01
2.11875532e-02 2.34606043e-01 -5.29427469e-01 2.34080583e-01
2.65477151e-02 7.41879046e-01 -1.07607508e+00 -5.82991898e-01
3.37338299e-01 1.76314980e-01 -4.42608781e-02 5.41575968e-01
-5.41259110e-01 2.69866616e-01 -3.86419922e-01 1.05937672e+00
1.08411551e+00 -2.00062752e-01 -5.91713898e-02 -7.19768107e-01
-1.56593099e-01 -1.92408621e-01 -1.43512928e+00 1.20823848e+00
-5.96805453e-01 9.13867891e-01 -1.85432762e-01 1.43031672e-01
9.49735701e-01 -3.17251831e-01 4.31691498e-01 -6.48337305e-01
2.18923148e-02 1.06455550e-01 -3.79332632e-01 -6.16682231e-01
6.83035851e-01 1.55223638e-01 3.38911146e-01 4.06064212e-01
-4.44244564e-01 -2.90849984e-01 1.38895111e-05 1.52413860e-01
8.98671567e-01 1.83214068e-01 2.72719145e-01 -6.62464127e-02
5.01418531e-01 -8.35344419e-02 4.03456211e-01 6.22301459e-01
-4.73130755e-02 8.67483497e-01 6.21971749e-02 -4.19868499e-01
-5.87714970e-01 -1.32464170e+00 -3.39646041e-01 1.19276369e+00
7.44611561e-01 -1.00435339e-01 -5.88259161e-01 -9.16938007e-01
2.31516212e-01 1.11841178e+00 -6.37102783e-01 2.00052083e-01
-4.07846004e-01 -7.06866741e-01 -7.28866309e-02 4.18693423e-01
5.65699339e-01 -1.36930144e+00 -8.47671509e-01 -1.07488297e-01
-1.05830386e-01 -1.34526908e+00 -8.84534895e-01 -9.76214930e-02
-1.05862163e-01 -1.37850833e+00 -5.99524677e-01 -5.53524196e-01
8.79705906e-01 7.61573493e-01 1.37106586e+00 9.87273902e-02
-6.02639437e-01 3.36274564e-01 -1.75805807e-01 -8.18319678e-01
-4.80581485e-02 -4.84118104e-01 -2.11097561e-02 5.00372410e-01
4.13376354e-02 -3.50496560e-01 -9.65844750e-01 6.78474307e-01
-7.32798100e-01 4.09813106e-01 8.98961782e-01 3.78999084e-01
8.59839082e-01 -2.22600758e-01 1.16910040e-01 -1.04408884e+00
3.16673696e-01 1.15960822e-01 -9.47713077e-01 6.73450768e-01
-8.28864038e-01 -2.45932952e-01 2.79926121e-01 -2.72212833e-01
-1.37026322e+00 5.08868217e-01 5.71941614e-01 -5.42311788e-01
-1.08084463e-01 -2.36561224e-01 -3.74950886e-01 -1.59297928e-01
7.41453230e-01 1.83938652e-01 -5.82401633e-01 -3.50099146e-01
7.10927069e-01 6.20524526e-01 8.14450979e-01 -4.44740415e-01
1.04241514e+00 8.16318214e-01 2.94929724e-02 -6.33494437e-01
-1.47438312e+00 -6.68961167e-01 -4.58939970e-01 -4.95538235e-01
7.68053293e-01 -5.19152224e-01 -7.64304578e-01 4.77159858e-01
-1.43473375e+00 -4.23030496e-01 -3.72544974e-01 2.15463385e-01
-4.93769616e-01 1.53874993e-01 2.82306314e-01 -9.69219327e-01
-2.90930629e-01 -1.04772937e+00 1.49773288e+00 4.84083176e-01
3.46126795e-01 -2.28448957e-01 -7.02394173e-02 2.86202222e-01
3.21614780e-02 -6.05715672e-03 6.97923362e-01 -1.44615188e-01
-1.04260361e+00 2.65428871e-02 -9.68149304e-01 3.85064423e-01
2.19363034e-01 1.23527208e-02 -1.28678715e+00 -1.42962299e-03
-5.26187718e-01 4.07093950e-02 6.65681541e-01 2.91404903e-01
1.68404603e+00 -1.64573491e-01 -5.86806774e-01 5.59767783e-01
1.44867694e+00 1.03808582e-01 6.20634139e-01 -7.26455748e-02
9.57894742e-01 5.01467347e-01 1.22962594e+00 6.28162742e-01
3.37500989e-01 8.97938728e-01 4.12070900e-01 -2.60239899e-01
-6.54228985e-01 -8.45574290e-02 1.48825049e-01 1.47070363e-01
-1.71104431e-01 -3.87418360e-01 -7.03562677e-01 2.39926964e-01
-1.76936853e+00 -8.09283674e-01 -5.46072721e-01 2.18145823e+00
7.72002816e-01 3.84727190e-03 -5.73289581e-02 -2.95007676e-01
7.28164494e-01 1.09624766e-01 -7.81568408e-01 3.40428025e-01
-1.30120352e-01 -2.42147289e-04 7.39063382e-01 5.75634837e-01
-7.78141320e-01 1.26679897e+00 6.04864788e+00 7.92508960e-01
-7.07666278e-01 -3.49617377e-02 3.87630194e-01 -6.99119195e-02
-3.35736215e-01 1.50967181e-01 -1.11933267e+00 5.11094451e-01
1.26582056e-01 5.96819818e-02 8.01036239e-01 8.62178624e-01
3.28813910e-01 -5.44788539e-01 -1.36254668e+00 9.90516067e-01
3.66358191e-01 -1.17685485e+00 -9.10513848e-02 -1.26407787e-01
9.63440776e-01 -4.22959179e-01 -1.69042066e-01 6.23234399e-02
9.07901302e-02 -6.39719188e-01 8.00363660e-01 9.98406351e-01
8.89445484e-01 -2.50764549e-01 3.83824736e-01 8.68443474e-02
-1.24559569e+00 -1.45279020e-01 -4.28176999e-01 3.58282834e-01
2.79603958e-01 8.00773740e-01 -1.44322217e+00 3.95087659e-01
5.41787922e-01 4.36446428e-01 -7.68762052e-01 1.42658877e+00
-2.99125195e-01 2.87982017e-01 -2.09363341e-01 -1.08702257e-01
-2.90434420e-01 -2.17571884e-01 5.24502039e-01 1.08868277e+00
5.94502762e-02 4.06353384e-01 2.42278799e-01 9.86219823e-01
-1.73454136e-01 -6.69742934e-03 -9.45950896e-02 4.34507787e-01
8.02020609e-01 1.60983729e+00 -9.02329147e-01 -7.42901146e-01
-2.47404590e-01 1.20858431e+00 1.17258877e-01 4.51941162e-01
-1.38455963e+00 -3.27323467e-01 4.44905043e-01 1.72104716e-01
-7.96939339e-03 2.19630480e-01 -3.63730252e-01 -6.23847246e-01
2.64084637e-01 -5.74012935e-01 8.33306536e-02 -1.56291592e+00
-1.41471016e+00 3.08804065e-01 1.04566462e-01 -1.21073067e+00
4.03127253e-01 -5.76218486e-01 -5.89154363e-01 7.22008228e-01
-1.47197866e+00 -1.34783471e+00 -1.13945913e+00 5.75230658e-01
8.61603439e-01 5.11844717e-02 4.97103035e-01 2.73782402e-01
-4.77874011e-01 4.31485206e-01 -2.05615303e-03 -2.39761963e-01
8.65134656e-01 -1.26334143e+00 3.89427394e-01 6.50607228e-01
3.03370297e-01 1.12555683e-01 7.29253113e-01 -5.99208891e-01
-1.38289118e+00 -1.30490494e+00 6.58394635e-01 -8.45646083e-01
1.67360768e-01 -3.28115851e-01 -5.91878772e-01 5.10527670e-01
-7.86515549e-02 3.26876998e-01 -9.27941129e-02 3.68712842e-02
-9.37792659e-02 -5.44146538e-01 -1.12792206e+00 8.33740592e-01
1.57588041e+00 -3.36837173e-01 -4.43146080e-01 1.04223967e+00
7.84464359e-01 -7.87334323e-01 -3.43649328e-01 5.68681180e-01
5.55514574e-01 -1.17975163e+00 1.21325529e+00 -2.32278958e-01
3.16818953e-01 -6.40770674e-01 -2.71385521e-01 -1.01235056e+00
-1.39935464e-01 -3.05530638e-01 -1.73789158e-01 1.38089502e+00
4.65063423e-01 -4.69934762e-01 7.73533046e-01 5.56130886e-01
-5.65802515e-01 -1.00835145e+00 -5.06229103e-01 -7.08703816e-01
-1.05530536e+00 -3.54388803e-01 9.20804858e-01 6.56568229e-01
-8.60960066e-01 1.86386108e-01 -3.37352484e-01 5.81558287e-01
7.84691572e-01 7.69377232e-01 1.31377351e+00 -9.44933116e-01
-1.80204839e-01 -2.15200722e-01 3.26439552e-02 -9.20849085e-01
4.97068726e-02 -5.66571414e-01 6.83124602e-01 -1.73050010e+00
5.30074894e-01 -7.74166763e-01 -1.77931577e-01 5.36463320e-01
-5.47278047e-01 4.20209110e-01 1.66978687e-01 2.38440230e-01
-8.08520973e-01 5.13581991e-01 1.46754491e+00 -3.32664326e-02
-4.29304749e-01 3.02935421e-01 -3.57758105e-01 8.37606192e-01
3.72883528e-01 -2.94299066e-01 -3.30667973e-01 -2.94091970e-01
-2.57818639e-01 -2.04665467e-01 7.79303372e-01 -1.24663150e+00
-3.13424319e-02 -7.70705104e-01 4.71374631e-01 -1.01312304e+00
6.67427063e-01 -1.01631188e+00 1.33150831e-01 1.30726337e-01
-2.36510351e-01 -4.44975972e-01 -1.83671206e-01 7.25055695e-01
4.28489238e-01 -3.16403210e-02 8.33035767e-01 3.91871482e-02
-1.00981557e+00 5.33889413e-01 4.95362252e-01 -4.47273813e-02
1.44002557e+00 -4.49232340e-01 -4.14737672e-01 1.01845093e-01
-5.21421313e-01 1.50033981e-01 3.96800637e-01 4.26106215e-01
7.25967467e-01 -1.25562286e+00 -5.05323529e-01 2.21160099e-01
5.40282428e-01 1.89171255e-01 1.96106389e-01 6.86938703e-01
-1.30849645e-01 1.51140541e-01 4.26867194e-02 -7.19386756e-01
-1.61713183e+00 5.08444190e-01 2.07935750e-01 4.98596013e-01
-6.24952197e-01 9.51835573e-01 6.18593395e-01 -1.68408766e-01
3.99312168e-01 -5.48789084e-01 2.64928192e-01 -2.46751651e-01
5.63227475e-01 6.19382441e-01 -1.35322334e-02 -2.46748850e-01
-4.05163378e-01 5.81675410e-01 3.11330169e-01 9.73840430e-02
9.76186872e-01 -2.68372089e-01 -2.31808007e-01 3.25947285e-01
7.66584873e-01 4.56396371e-01 -1.44089997e+00 -4.17701513e-01
-2.65605986e-01 -1.15375304e+00 -1.53351575e-01 -1.27495480e+00
-8.76610100e-01 3.63204628e-01 9.10218537e-01 -6.40307888e-02
1.20189047e+00 2.55319655e-01 7.56218553e-01 4.31994617e-01
3.68336856e-01 -1.12260640e+00 2.11222082e-01 2.39122152e-01
1.12214494e+00 -1.40301442e+00 2.20091641e-01 -9.49942350e-01
-6.98346138e-01 6.92454636e-01 1.07941067e+00 2.46559441e-01
4.77270901e-01 2.27447599e-01 3.25964317e-02 -4.10091132e-01
-5.26979208e-01 -4.78953749e-01 5.28484344e-01 6.99632049e-01
-1.47514120e-01 2.79003531e-01 -6.05679825e-02 3.23592097e-01
-1.59592509e-01 4.42611985e-02 1.71257228e-01 5.73831320e-01
-7.58202910e-01 -7.89990604e-01 -6.03470027e-01 6.63679779e-01
3.43836606e-01 -2.08998397e-01 -6.44409359e-01 4.65444773e-01
4.81773823e-01 7.90230155e-01 1.55055197e-03 -4.04152781e-01
6.46725297e-01 -2.36532539e-01 5.09747386e-01 -6.89966738e-01
-2.96608452e-02 -2.84635127e-01 2.84448005e-02 -7.17921853e-01
-2.25468829e-01 -4.85867620e-01 -1.31924760e+00 -4.99768890e-02
-4.73530591e-01 -3.59051824e-01 6.43653750e-01 8.27241182e-01
3.44550014e-01 5.59429824e-01 9.67281044e-01 -1.10051954e+00
-2.99706429e-01 -8.99190247e-01 -4.26634490e-01 5.02556026e-01
1.58993348e-01 -8.61255646e-01 -1.74294516e-01 2.50211090e-01] | [10.852727890014648, -4.107909679412842] |
e1fa0ea8-5b1d-428c-bfec-c5f960e0f42c | weakly-supervised-3d-multi-person-pose | 2211.16951 | null | https://arxiv.org/abs/2211.16951v1 | https://arxiv.org/pdf/2211.16951v1.pdf | Weakly Supervised 3D Multi-person Pose Estimation for Large-scale Scenes based on Monocular Camera and Single LiDAR | Depth estimation is usually ill-posed and ambiguous for monocular camera-based 3D multi-person pose estimation. Since LiDAR can capture accurate depth information in long-range scenes, it can benefit both the global localization of individuals and the 3D pose estimation by providing rich geometry features. Motivated by this, we propose a monocular camera and single LiDAR-based method for 3D multi-person pose estimation in large-scale scenes, which is easy to deploy and insensitive to light. Specifically, we design an effective fusion strategy to take advantage of multi-modal input data, including images and point cloud, and make full use of temporal information to guide the network to learn natural and coherent human motions. Without relying on any 3D pose annotations, our method exploits the inherent geometry constraints of point cloud for self-supervision and utilizes 2D keypoints on images for weak supervision. Extensive experiments on public datasets and our newly collected dataset demonstrate the superiority and generalization capability of our proposed method. | ['Yuexin Ma', 'Jingyi Yu', 'Jingya Wang', 'Lan Xu', 'Juze Zhang', 'Yiming Ren', 'Yiteng Xu', 'Peishan Cong'] | 2022-11-30 | null | null | null | null | ['3d-pose-estimation', '3d-multi-person-pose-estimation', 'multi-person-pose-estimation'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [-3.11679453e-01 -5.45102596e-01 -1.94629401e-01 -4.68476832e-01
-4.43745285e-01 -4.71403062e-01 1.95917577e-01 -3.51948619e-01
-5.38656354e-01 6.96964145e-01 2.20358044e-01 4.59486842e-01
-7.62799904e-02 -6.90174162e-01 -4.08323109e-01 -5.53790331e-01
1.57815039e-01 7.64039159e-01 2.04939499e-01 -6.97953850e-02
-1.88324302e-01 8.22792590e-01 -1.36313486e+00 -4.02108133e-01
6.87452018e-01 8.05078804e-01 9.08395201e-02 4.37585533e-01
2.82454640e-01 2.34290481e-01 -3.71151209e-01 -3.28084081e-01
4.73194927e-01 1.36753216e-01 -1.89104334e-01 5.49639761e-01
9.77054477e-01 -8.04443538e-01 -6.50169909e-01 8.54605854e-01
9.13363039e-01 8.30086246e-02 3.22620153e-01 -1.27813673e+00
-2.96740979e-01 -2.02544793e-01 -8.95331085e-01 1.08094193e-01
1.02242947e+00 3.54772419e-01 6.30639434e-01 -8.86744797e-01
5.77794909e-01 1.55081594e+00 8.72466505e-01 5.73307276e-01
-8.28528345e-01 -7.69351840e-01 3.91427249e-01 1.25734732e-01
-1.85352421e+00 -3.09523851e-01 1.02245212e+00 -3.01522255e-01
6.50485814e-01 3.08141578e-04 1.08087218e+00 1.20350587e+00
-2.05260023e-01 8.26333225e-01 9.23763871e-01 -1.18209004e-01
-1.77807733e-01 1.51285110e-02 -1.26149192e-01 1.00105071e+00
5.74392617e-01 2.72573739e-01 -7.81171679e-01 -1.33631423e-01
1.17380202e+00 6.26396537e-01 -2.79094011e-01 -7.43775129e-01
-1.16709888e+00 5.69121838e-01 7.32638299e-01 -1.88612387e-01
-3.40191394e-01 1.36576906e-01 1.15184002e-01 -1.23001426e-01
3.16734135e-01 -1.52073130e-01 -2.55054325e-01 -1.37329539e-02
-6.83810055e-01 5.06213665e-01 2.57619172e-01 1.18283832e+00
7.91520774e-01 -1.74465388e-01 1.32847384e-01 5.39078891e-01
3.57241094e-01 1.16212130e+00 1.65146053e-01 -1.04652703e+00
7.88289368e-01 9.30792511e-01 2.86449999e-01 -1.17854452e+00
-5.37798762e-01 -3.09363633e-01 -8.85421932e-01 -2.15130761e-01
4.75318611e-01 -2.25349218e-01 -5.90559542e-01 1.59686446e+00
7.46543825e-01 1.19525112e-01 -4.03185725e-01 1.34173810e+00
7.71359444e-01 3.79408486e-02 -3.18534315e-01 5.34652732e-02
1.20681357e+00 -6.58864737e-01 -2.99123943e-01 -5.57195663e-01
2.52956569e-01 -2.32738823e-01 7.89876401e-01 2.39611551e-01
-8.58186841e-01 -7.31786609e-01 -9.57839727e-01 -9.07385945e-02
-6.23933077e-02 2.99602121e-01 7.41885841e-01 6.24737084e-01
-6.82567775e-01 1.85649484e-01 -9.02714431e-01 -4.69351560e-01
4.23459440e-01 4.93720382e-01 -6.83454275e-01 -5.22946119e-01
-9.60956633e-01 6.05694354e-01 2.09759340e-01 3.81750464e-01
-6.19728923e-01 -3.78021628e-01 -1.09258842e+00 -4.02278483e-01
5.26322961e-01 -1.22577810e+00 7.83590019e-01 -3.80598381e-02
-1.16320813e+00 8.52047861e-01 -3.54126215e-01 -1.17098533e-01
8.62909317e-01 -7.04300880e-01 -4.65373211e-02 6.66232049e-01
3.70674431e-01 8.56270909e-01 6.70101762e-01 -1.14077508e+00
-6.50173664e-01 -1.19009984e+00 1.03147969e-01 6.53921723e-01
-4.62810874e-01 -3.66087526e-01 -7.13295758e-01 -2.44329825e-01
5.87055504e-01 -1.10876799e+00 -3.42736989e-01 4.30110663e-01
-6.00089252e-01 -1.11486107e-01 7.70326078e-01 -4.78406340e-01
5.67976952e-01 -1.77222061e+00 2.10914955e-01 8.94778967e-02
3.09208095e-01 -5.99358864e-02 1.80697531e-01 1.77889049e-01
4.08575535e-01 -3.35291862e-01 1.06895097e-01 -7.15054214e-01
-9.35468748e-02 1.87539995e-01 1.28094226e-01 8.41139913e-01
2.47741789e-02 9.21425343e-01 -8.23673725e-01 -8.65579009e-01
5.70446491e-01 8.10334206e-01 -5.28853595e-01 2.53429860e-01
2.02370733e-01 7.88173378e-01 -7.63650477e-01 1.12326586e+00
7.73360431e-01 -1.72929019e-01 -3.30176681e-01 -1.69143260e-01
3.14178802e-02 -9.25269499e-02 -1.56497490e+00 2.23670721e+00
-7.01029897e-02 3.71015817e-02 -1.35887079e-02 -5.52991211e-01
8.54722917e-01 2.31472492e-01 5.49184084e-01 -4.31568891e-01
1.16137294e-02 -2.21378002e-02 -7.40295589e-01 -3.39722842e-01
2.90387392e-01 -1.92236714e-02 -1.79184020e-01 1.60977408e-01
-6.77338541e-02 -4.96026054e-02 -1.38338044e-01 3.44934240e-02
7.93215752e-01 4.52986121e-01 3.61483991e-01 2.49382496e-01
6.44712567e-01 -2.12211877e-01 8.42725277e-01 6.05950654e-01
-4.97307181e-01 5.62018573e-01 -3.03278714e-01 -6.84743106e-01
-8.74800861e-01 -1.33541405e+00 2.75605526e-02 6.03630424e-01
4.87868488e-01 -3.44527036e-01 -2.98676819e-01 -5.32461584e-01
3.08399439e-01 -1.69291660e-01 -2.82353133e-01 7.76218176e-02
-7.70379186e-01 -5.25610864e-01 5.07496834e-01 7.03186214e-01
9.80375826e-01 -4.26348537e-01 -7.27255762e-01 -1.37469739e-01
-5.58404803e-01 -1.58922434e+00 -6.41662002e-01 -5.31312525e-01
-9.16635990e-01 -1.10125268e+00 -8.45598340e-01 -4.21267688e-01
7.04707265e-01 8.76244903e-01 7.84278929e-01 2.78493948e-02
-4.03223485e-01 7.30906963e-01 -3.03707225e-03 -1.90260217e-01
6.07829392e-01 -7.12545291e-02 6.01821542e-01 -1.00928396e-01
6.80946708e-01 -8.71661961e-01 -7.36030102e-01 4.33231473e-01
-7.18715461e-03 -1.10399574e-01 6.23305023e-01 5.93303382e-01
5.49443066e-01 1.70626402e-01 1.37873888e-01 -3.36935729e-01
6.34085163e-02 4.38649058e-02 -5.48283458e-01 -4.24661040e-02
-2.13845998e-01 -4.73587960e-01 1.24109536e-01 -3.85093331e-01
-9.12241459e-01 5.90937078e-01 2.18556836e-01 -8.26868355e-01
-4.15745735e-01 5.68309836e-02 -4.54007924e-01 -3.33694607e-01
5.27715981e-01 2.29959697e-01 -4.16607857e-02 -4.96441782e-01
2.87375510e-01 5.63367724e-01 8.95442903e-01 -6.08309686e-01
1.27967834e+00 1.07958794e+00 2.86110997e-01 -8.65998387e-01
-1.02662599e+00 -8.87747943e-01 -1.44151139e+00 -3.06510866e-01
9.61044073e-01 -1.73078775e+00 -1.12759304e+00 5.17766178e-01
-1.18687522e+00 4.08585399e-01 1.01984508e-01 6.38447702e-01
-3.31214130e-01 7.37405360e-01 -6.26206577e-01 -1.13068652e+00
-2.60270506e-01 -8.82521451e-01 1.58616650e+00 2.63274878e-01
1.23303346e-06 -8.38927150e-01 -1.16474517e-01 7.70220697e-01
-3.10369343e-01 5.96145153e-01 -9.81617942e-02 3.07358522e-02
-9.25336063e-01 -5.39506912e-01 -2.08417520e-01 -1.25278354e-01
1.67983875e-01 -5.23931086e-01 -9.85816777e-01 -5.63421249e-01
-5.78286834e-02 -4.51671839e-01 6.92369699e-01 4.26710993e-01
7.62821555e-01 7.77283162e-02 -4.97353226e-01 8.33723962e-01
1.16992879e+00 -4.92752165e-01 -3.88162062e-02 2.86013544e-01
1.36922157e+00 7.03685284e-01 7.80257881e-01 8.34303856e-01
1.06115913e+00 8.76580417e-01 3.27157110e-01 2.62820143e-02
1.29423454e-01 -6.53146148e-01 3.27975869e-01 5.24589717e-01
-3.69435102e-01 3.64679039e-01 -7.57448733e-01 3.56866539e-01
-1.80991030e+00 -1.06840968e+00 4.72976714e-02 2.27915025e+00
4.60062504e-01 1.47516266e-01 4.27705973e-01 6.51751608e-02
7.95061350e-01 6.20743521e-02 -6.84162974e-01 8.00165713e-01
-2.09234521e-01 -4.16958421e-01 5.16583502e-01 3.06800306e-01
-1.24250209e+00 8.47667038e-01 5.53974724e+00 3.35521787e-01
-6.66204154e-01 -1.20627709e-01 -1.11218132e-01 -4.10632014e-01
2.61294618e-02 -2.10027441e-01 -1.29941523e+00 4.05164003e-01
2.83419311e-01 2.37929448e-01 1.52681068e-01 8.51971567e-01
2.57793456e-01 -1.59713328e-01 -1.23705685e+00 1.65867233e+00
2.43369907e-01 -9.04167235e-01 -9.87930298e-02 3.03261280e-01
5.78156233e-01 -2.06048712e-01 -1.08954571e-01 9.15346388e-03
1.40215889e-01 -7.98492491e-01 5.88719547e-01 5.28185487e-01
6.24746978e-01 -9.11153018e-01 7.11804926e-01 8.75821292e-01
-1.65479004e+00 -2.16638714e-01 -6.39819980e-01 -4.49215084e-01
3.47106248e-01 6.04528010e-01 -7.10242212e-01 6.96866214e-01
9.80558515e-01 1.16093719e+00 -7.27973223e-01 9.42843258e-01
-3.64656299e-01 -1.32549465e-01 -6.25227809e-01 2.07330793e-01
-1.24807104e-01 -2.32166350e-01 5.98794162e-01 7.18691111e-01
3.12810630e-01 3.87892395e-01 9.44844484e-01 6.79139018e-01
3.06652606e-01 -2.55025208e-01 -7.68037796e-01 3.85362417e-01
8.01048994e-01 1.15401900e+00 -3.96359175e-01 -1.70290321e-01
-4.79886264e-01 1.12952328e+00 3.33873510e-01 2.48029649e-01
-6.60900950e-01 7.39312619e-02 7.48975575e-01 1.54146865e-01
3.19542885e-01 -6.87661707e-01 -8.20466056e-02 -1.67134392e+00
3.96207392e-01 -5.57350457e-01 4.87526119e-01 -8.60300839e-01
-1.34991395e+00 3.28403831e-01 2.25541100e-01 -1.52021122e+00
-3.10590804e-01 -5.49221873e-01 -2.40966678e-01 8.38310480e-01
-1.58180451e+00 -1.71439123e+00 -8.89258981e-01 1.09303880e+00
2.22317070e-01 -1.64959803e-01 4.70889479e-01 3.26959252e-01
-4.99561727e-01 4.46944118e-01 -6.38745368e-01 4.24711376e-01
8.00593972e-01 -1.08706081e+00 3.24678838e-01 9.35758352e-01
8.81304219e-02 7.57510424e-01 4.64086205e-01 -1.00307930e+00
-1.68302870e+00 -1.00867498e+00 7.06162214e-01 -9.85292614e-01
1.56556740e-01 -4.50352609e-01 -4.12414014e-01 7.61482239e-01
-6.89957261e-01 8.13876837e-02 6.05653286e-01 2.31916979e-01
-3.72165769e-01 -2.01690465e-01 -1.20095038e+00 4.28985476e-01
1.56302190e+00 -5.65493822e-01 -6.28016174e-01 4.05991763e-01
7.73847938e-01 -7.74127901e-01 -7.26085961e-01 4.07045543e-01
5.22179186e-01 -9.95688796e-01 1.67918968e+00 -8.63989443e-02
-8.28050151e-02 -4.19699460e-01 -3.51059884e-01 -8.21542084e-01
-3.22397053e-01 -3.41991812e-01 -2.06418931e-01 1.02244508e+00
-4.18747485e-01 -4.86361116e-01 1.24645877e+00 5.31361103e-01
1.40961915e-01 -3.05756360e-01 -1.06695986e+00 -7.22476006e-01
-3.36044699e-01 -3.80428463e-01 7.15731859e-01 5.63561440e-01
-3.74586076e-01 3.69546652e-01 -8.42306137e-01 7.18999922e-01
1.30287576e+00 3.67644966e-01 1.35753381e+00 -1.63654149e+00
-9.51439738e-02 1.84600234e-01 -8.13157380e-01 -1.62514842e+00
6.01633973e-02 -4.37064052e-01 -1.77958205e-01 -1.22310102e+00
2.81404585e-01 -2.07077414e-01 1.72572181e-01 2.99010396e-01
-2.71877557e-01 4.69911218e-01 3.34567994e-01 4.94199455e-01
-7.33546376e-01 7.72124648e-01 1.31861055e+00 -9.78407077e-03
-3.62238325e-02 2.71422863e-01 -6.22608006e-01 1.00665557e+00
1.90484211e-01 -1.44387364e-01 -5.04983664e-01 -5.36840677e-01
-8.43840688e-02 2.05221698e-01 1.03310013e+00 -1.26456976e+00
4.08314258e-01 -3.17357719e-01 1.06616151e+00 -1.16513658e+00
1.00702691e+00 -9.07786131e-01 1.33670336e-02 4.79324579e-01
2.12744921e-01 1.54388353e-01 -2.51673758e-01 8.31435561e-01
6.61802068e-02 3.10147434e-01 5.63938379e-01 -5.64952672e-01
-8.64812195e-01 1.00254488e+00 5.59640169e-01 -1.19300723e-01
9.50835884e-01 -6.04697764e-01 -1.69970393e-02 -4.93904263e-01
-4.81616229e-01 6.18167698e-01 8.27412009e-01 3.75455260e-01
9.71245527e-01 -1.63966525e+00 -5.81252098e-01 3.71731758e-01
1.57736108e-01 4.30270374e-01 4.55780029e-01 7.24093735e-01
-3.54976624e-01 5.99558175e-01 -2.32432544e-01 -1.15064633e+00
-1.36740839e+00 3.68631601e-01 3.27027589e-01 8.83221477e-02
-8.22006583e-01 8.37749422e-01 1.10615030e-01 -8.02901506e-01
1.57175213e-01 -1.35803418e-02 -1.04117922e-01 -1.95462629e-01
7.31629491e-01 3.69262278e-01 -3.82404357e-01 -9.88972366e-01
-6.76309168e-01 1.21490061e+00 1.56107470e-01 -1.74381465e-01
1.20024192e+00 -7.24730074e-01 1.43870145e-01 4.87065554e-01
9.96426642e-01 -3.31701525e-02 -1.59712362e+00 -6.36023462e-01
-4.57619399e-01 -9.83917534e-01 -1.51186436e-01 -2.93406695e-01
-9.50439274e-01 1.09085119e+00 4.55590487e-01 -3.74599874e-01
9.75796461e-01 -1.30537376e-01 9.55931604e-01 6.37023091e-01
9.52882648e-01 -1.03558576e+00 4.12292093e-01 1.71358839e-01
5.32139242e-01 -1.52893066e+00 4.81134444e-01 -7.73140550e-01
-5.07125318e-01 9.06194806e-01 9.60911930e-01 -1.08009331e-01
3.57747555e-01 -1.50255099e-01 5.77497482e-02 -1.31301016e-01
-3.19739401e-01 -3.22284609e-01 2.62952685e-01 1.08174503e+00
-1.69098884e-01 3.58218364e-02 4.57952261e-01 3.04750770e-01
-4.90802109e-01 -1.92462485e-02 1.57105938e-01 9.09969509e-01
-5.20034254e-01 -7.98454046e-01 -6.76217914e-01 -6.83758967e-03
2.54442226e-02 3.73111248e-01 -4.18282092e-01 8.90882850e-01
2.65618593e-01 8.40667248e-01 -1.08225793e-01 -3.90071750e-01
4.35467571e-01 -2.05299377e-01 7.31399000e-01 -5.48359096e-01
-3.19131128e-02 9.25517231e-02 -1.71808794e-01 -7.20086217e-01
-5.92809260e-01 -1.07063627e+00 -1.00325155e+00 -5.61472893e-01
-8.10089484e-02 -1.73659354e-01 3.61038387e-01 1.15792310e+00
2.72309273e-01 1.79427788e-02 5.88073254e-01 -1.25322402e+00
-6.47246063e-01 -6.86942518e-01 -6.16423845e-01 5.01640797e-01
5.41347027e-01 -1.09629309e+00 -2.20830068e-01 -1.51685700e-01] | [7.038650035858154, -1.0043952465057373] |
8cf11527-7bf2-4c5c-b804-bbd742c69f5e | global-local-bidirectional-reasoning-for | 2003.12971 | null | https://arxiv.org/abs/2003.12971v1 | https://arxiv.org/pdf/2003.12971v1.pdf | Global-Local Bidirectional Reasoning for Unsupervised Representation Learning of 3D Point Clouds | Local and global patterns of an object are closely related. Although each part of an object is incomplete, the underlying attributes about the object are shared among all parts, which makes reasoning the whole object from a single part possible. We hypothesize that a powerful representation of a 3D object should model the attributes that are shared between parts and the whole object, and distinguishable from other objects. Based on this hypothesis, we propose to learn point cloud representation by bidirectional reasoning between the local structures at different abstraction hierarchies and the global shape without human supervision. Experimental results on various benchmark datasets demonstrate the unsupervisedly learned representation is even better than supervised representation in discriminative power, generalization ability, and robustness. We show that unsupervisedly trained point cloud models can outperform their supervised counterparts on downstream classification tasks. Most notably, by simply increasing the channel width of an SSG PointNet++, our unsupervised model surpasses the state-of-the-art supervised methods on both synthetic and real-world 3D object classification datasets. We expect our observations to offer a new perspective on learning better representation from data structures instead of human annotations for point cloud understanding. | ['Jie zhou', 'Yongming Rao', 'Jiwen Lu'] | 2020-03-29 | global-local-bidirectional-reasoning-for-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Rao_Global-Local_Bidirectional_Reasoning_for_Unsupervised_Representation_Learning_of_3D_Point_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Rao_Global-Local_Bidirectional_Reasoning_for_Unsupervised_Representation_Learning_of_3D_Point_CVPR_2020_paper.pdf | cvpr-2020-6 | ['3d-object-classification'] | ['computer-vision'] | [-0.13834001 0.50113213 -0.2596282 -0.5263697 -0.46187413 -0.706431
0.7510849 0.48690104 0.293899 0.13003351 -0.06048245 -0.05789443
-0.21074972 -0.94525766 -1.0433865 -0.5212901 -0.06582409 1.0326111
0.6386391 -0.14219336 0.26267 1.0236537 -1.7777061 0.39682662
0.49090657 1.2801071 0.21697703 0.06931079 -0.44396305 0.6536664
-0.20844817 -0.20688094 0.5096882 0.2509418 -0.94426984 0.5204965
0.97337556 -0.03330566 -0.2380736 1.0369548 -0.13155809 -0.21362981
0.88492 -1.4166656 -0.7918903 0.34170714 -0.38164443 -0.31296402
0.28039598 0.05693749 1.4180709 -0.9264651 0.82219076 1.348023
0.69618136 0.30379227 -1.4773811 -0.38202292 0.5904259 0.07452816
-1.4251643 -0.0380307 0.83475924 -0.6226894 1.0653455 0.04320486
0.63141465 0.7631373 -0.07642678 0.9366578 0.8512408 -0.01470241
0.27802545 0.07063302 0.34202078 0.79066133 0.46204096 -0.18081212
-0.49343374 -0.26107308 0.9584128 0.4810734 -0.02296137 -1.2727867
-1.2527536 0.5140456 0.94302005 0.1407728 -0.31973565 0.2597871
0.05855168 0.2463097 0.4254448 0.5096766 -0.7542745 0.1989778
-0.5310316 0.2564541 0.79593694 1.6635044 1.2195237 -0.3984957
0.31498897 0.49102318 0.3793137 0.51140916 -0.04162928 -0.946243
0.34563 1.4118452 -0.09790912 -0.88733125 -0.2961861 -0.6180647
-0.7629991 0.5237519 0.21219765 0.69287837 -0.9871352 1.3931108
0.21479179 0.1904972 -0.12150963 0.81072336 0.86064446 0.45160624
-0.07037347 0.5022194 1.362288 -0.6107953 0.03448799 -0.16546877
0.65534216 -0.2799958 0.84376705 0.2651069 -1.030677 -0.81210357
-1.0980604 -0.31729388 -0.58152735 -0.05217579 0.8170201 0.17160887
-0.79681545 0.65026677 -0.9900864 -0.56598157 0.77990305 0.4043607
-0.75399816 -0.03958729 -0.29700124 0.81709176 0.4761282 -0.2915736
-0.82609457 -1.0000571 -0.8608481 0.11677078 0.37442696 -0.9941539
1.0805172 -0.59278345 -0.9519715 1.2572343 -0.04856187 -0.30449852
0.22831842 -0.13105883 0.06817574 0.17484617 0.16471288 0.9774756
0.7996379 -1.7915765 -0.74188405 -0.86148125 0.1907315 0.07792234
0.2411276 -0.6293937 -0.4311844 -0.241803 0.9243714 -0.9838708
-0.02007725 0.59678227 -0.4075171 -0.55039644 1.1359642 0.06711651
0.23954687 -2.3523185 0.2535694 0.35839254 0.5350296 -0.19158752
-0.03732276 0.3876037 -0.16372046 0.13282436 -0.25733516 -0.38622898
0.24988711 0.70612395 -0.73401475 0.53085953 0.42259085 0.9684489
-0.9479946 -0.28613824 0.27982906 0.2803776 -0.68643445 0.13332818
-0.53478384 0.33214852 -0.63463414 0.8016356 0.7193946 -0.6621547
-0.06740744 -0.16141821 0.16829558 0.58412087 -1.1285198 2.141007
-0.30046555 0.32877958 -0.1809346 -1.0164976 1.2570593 0.06615817
0.4985763 -0.3236205 -0.25210434 0.21960331 -0.1975028 -0.20362954
0.20048143 -0.02998409 0.07935032 0.3195023 0.40699586 -0.6651408
-0.24059717 0.37520516 0.93429637 0.38008112 0.16843347 -0.3378147
0.25314227 0.01993304 0.34257057 0.8910568 0.06455448 0.7539264
0.55409074 -0.6208269 -0.9876013 -1.4652401 -0.29791376 0.78647184
0.7163758 -0.75263727 -0.12067841 -0.9852323 0.6085811 0.47845727
-0.63815975 -0.02447472 -0.4082676 0.04820607 0.22713235 0.92337924
0.26425833 -0.72274864 -0.5047871 -0.1142858 0.17569281 -1.3205144
0.30153874 0.29790503 -1.3145306 -1.098992 -0.14719999 -0.84059215
0.91179085 0.48066887 1.3073494 0.5333049 -0.04264801 0.6462833
-0.2997678 -0.5347053 -0.15051372 0.2006714 0.02550949 -0.06162727
0.5222583 -0.86584485 -0.30302188 0.46563888 -0.6787945 0.02919764
0.67072093 0.36780387 0.9417095 -0.08664932 0.04541642 -0.8725932
-0.10522595 -0.51108205 -0.43839335 0.13656507 -0.29581428 0.2672736
0.3531387 -0.20702742 -0.7617658 0.16969158 0.20660132 -0.8032394
-0.7015257 0.2228642 -0.36463597 0.05519275 0.5376034 0.1629164
0.0226692 -0.8451221 0.5547532 0.20577693 0.5751488 -0.94554424
1.0638455 1.0607425 0.46935064 -0.82164085 -0.8459869 -0.6820325
-1.2316629 0.05118593 0.7987201 -1.0433958 -0.63752073 0.01828456
-1.4149834 0.06780858 -0.60390544 0.14079174 -0.8454421 0.17421511
-0.4485541 -0.31656814 0.3519375 -1.04512 1.6303807 -0.2686255
-0.23640925 -0.8554843 -0.20992775 0.17055279 -0.07477186 0.17191665
1.2754204 -0.8522355 -1.2142615 -0.2237742 -0.38593894 0.2300854
0.21231882 0.02769189 -0.927015 -0.10979528 0.11353584 -0.20320699
0.81749666 -0.09204306 1.4406782 -0.01464222 -0.6691461 0.571224
1.388613 -0.27275547 0.47220534 0.11157022 0.91275454 0.69280785
0.35939476 0.19403595 0.28294167 0.47183132 0.95025337 0.05250522
-0.05488937 -0.57869625 -0.06245515 0.53206545 -0.32387176 0.12146498
-1.1926918 0.42950708 -1.9592502 -0.7880468 -0.33620217 1.8565841
0.23319456 0.42938015 -0.03898746 0.06488045 0.4253642 -0.04626412
-0.52470523 0.09100423 -0.1432365 0.05460989 0.41307324 0.12496082
-0.9823985 0.8310221 6.140005 0.38891718 -0.82681346 -0.21685067
0.2502377 0.3276073 -0.37179282 0.3813139 -0.89133835 -0.01912703
0.3128488 0.10690709 -0.05496128 1.1434813 -0.49485537 0.42234176
-1.9886918 1.060924 -0.00788229 -1.5311276 0.5537728 0.35220337
0.65552765 0.1974353 -0.0539492 0.23254237 0.32893878 -0.94834
1.0053447 0.71665543 0.3064841 -0.38537312 0.5760031 0.66066855
-1.2902803 0.10242667 -0.6546955 -0.18508512 -0.21937451 0.38991812
-0.8501845 0.8254844 0.7639442 1.2996112 -0.7576548 0.90909743
-0.3712619 0.24383269 -0.51987576 0.23644939 0.29357672 -0.14158402
0.73467004 0.87173736 0.07186491 0.22585021 0.2380906 1.4026018
-0.13221875 -0.238437 -0.9797494 0.10776742 0.3765096 0.9823704
-0.7150279 -0.4502177 -0.59036106 0.5492912 0.5932816 0.20489794
-0.42522326 -0.01396438 0.89874166 0.295876 0.67251074 -0.5185738
-0.6503879 -1.1153114 0.18334016 -0.4687998 0.3457247 -1.091452
-1.7109045 0.32321376 0.24080466 -1.7229531 0.14334129 -1.0948997
-0.45333567 0.6020662 -1.4009144 -1.3586773 -0.41368493 0.522487
0.44880736 -0.11782344 0.82284933 -0.2733044 0.20904927 -0.04727119
-0.38697496 0.14969699 0.27327168 -1.3906331 0.4664474 0.27954847
0.6299346 0.8394863 0.35791215 -0.5333682 -1.553557 -1.015802
0.6120394 -1.0266854 0.6485854 -0.7365727 -1.2566477 1.0014741
-0.15941826 0.36613795 0.47696367 0.4514268 -0.96829396 -0.0519262
-0.91115314 0.5217794 1.6360313 -0.73720986 -1.0682496 0.3178449
0.8168746 -0.37391722 -1.0809283 0.6401862 0.27524626 -1.0353492
1.2217585 -0.9418499 0.57869583 -0.49446166 -0.5814157 -1.0577257
-0.5750165 -0.0483474 -0.2695033 0.9462338 0.41088477 -0.5775662
0.9785987 0.5769384 -0.38550997 -0.74840534 -0.7437317 -0.98452663
0.14508095 -0.6093733 1.044359 1.0704894 -0.16298014 0.3602834
0.45288673 0.7063403 0.81067324 0.6868235 1.0878351 -1.8399667
-0.05854927 -0.5005757 -1.2916658 -1.5372492 0.43000424 -1.3391628
-0.22003634 -1.5985416 0.06075565 -0.74675727 -0.27943388 0.7008406
0.25105092 0.03390249 0.31809962 0.51229036 -0.6042143 0.5880908
1.333919 -0.5295424 0.11032882 -0.08520904 -0.5787902 0.97248423
0.40132356 -0.4890049 -0.37301615 -0.75939804 0.21159294 -0.34545153
0.8133193 -0.98197025 0.43421808 -0.02228642 0.45735922 -0.913285
0.5987627 -1.3469424 0.10002298 0.16736972 -0.31316385 -0.23942317
0.11943498 0.7250725 -0.19147113 0.01778872 0.48980236 -0.35256338
-0.8004226 0.6069847 0.2711806 -0.29891887 0.8320728 -0.5043525
-0.3331452 -0.08637597 -0.8695312 0.19190148 0.90000975 0.6584819
0.72772527 -1.3004739 -0.5454815 0.58628327 0.67718416 0.6080618
-0.17278576 0.4630024 -0.34703347 0.43589452 -0.23691057 -1.398103
-0.9700491 0.63015467 0.2744117 0.36646155 -1.1381611 0.6623393
0.655875 -0.704064 0.28048396 -0.76341355 0.2018764 -0.34133542
0.01425639 0.09617761 0.18782951 -0.709757 -0.43216413 0.90005463
-0.07749175 0.37583846 1.483122 0.1144656 -0.27241164 0.8184315
1.2008383 -0.21378286 -1.246525 -0.56118655 0.19209631 -0.6454543
-0.16692269 -0.41660583 -0.8649039 0.9247959 0.13278915 0.48182464
0.54293185 0.8591994 0.18746682 0.74206454 0.8057141 -0.5152904
0.09393562 0.5122906 1.0374384 -1.2714574 0.16551937 -0.87727106
-0.49343893 1.1347766 0.6308973 -0.59385794 0.86785156 -0.07773256
-0.30715564 -0.81494933 -0.9778722 -0.22923937 0.53047746 0.7090952
-0.01498908 0.03666411 0.6348374 0.4311162 -0.3271446 -0.42684683
0.03888322 1.0503626 -0.44430068 -0.983718 -0.16218987 0.36489457
0.08418217 0.38712695 -0.83032143 1.0983675 0.25449017 0.42871875
0.5177313 -0.20513415 0.5361091 0.09780201 0.76044893 -1.0065823
-0.12905662 -0.26588696 -0.19650847 -0.66069645 -0.5908692 -0.63106257
-1.5078187 0.0114113 -0.2209551 -0.06445333 0.69346523 0.99640894
0.5944144 0.22506477 0.4306349 -1.0368319 -0.40941784 -0.56433195
-0.73425347 0.80575234 0.34250924 -1.035609 -0.4127963 0.05504032] | [7.9995012283325195, -3.2598462104797363] |
641fbe11-5211-4c54-bf5c-c8f83908fdf7 | coqa-a-conversational-question-answering | 1808.07042 | null | http://arxiv.org/abs/1808.07042v2 | http://arxiv.org/pdf/1808.07042v2.pdf | CoQA: A Conversational Question Answering Challenge | Humans gather information by engaging in conversations involving a series of
interconnected questions and answers. For machines to assist in information
gathering, it is therefore essential to enable them to answer conversational
questions. We introduce CoQA, a novel dataset for building Conversational
Question Answering systems. Our dataset contains 127k questions with answers,
obtained from 8k conversations about text passages from seven diverse domains.
The questions are conversational, and the answers are free-form text with their
corresponding evidence highlighted in the passage. We analyze CoQA in depth and
show that conversational questions have challenging phenomena not present in
existing reading comprehension datasets, e.g., coreference and pragmatic
reasoning. We evaluate strong conversational and reading comprehension models
on CoQA. The best system obtains an F1 score of 65.4%, which is 23.4 points
behind human performance (88.8%), indicating there is ample room for
improvement. We launch CoQA as a challenge to the community at
http://stanfordnlp.github.io/coqa/ | ['Siva Reddy', 'Danqi Chen', 'Christopher D. Manning'] | 2018-08-21 | coqa-a-conversational-question-answering-1 | https://aclanthology.org/Q19-1016 | https://aclanthology.org/Q19-1016.pdf | tacl-2019-3 | ['generative-question-answering'] | ['natural-language-processing'] | [ 3.92808728e-02 5.36431491e-01 2.15988442e-01 -4.79119003e-01
-1.20652092e+00 -1.00368667e+00 8.34486485e-01 3.39888602e-01
-2.66396672e-01 9.79792774e-01 1.14010775e+00 -5.69480002e-01
-1.05204113e-01 -6.00277781e-01 -5.02877772e-01 -1.89462706e-01
1.41956687e-01 9.76308644e-01 3.91991973e-01 -8.70898843e-01
3.89930606e-01 -3.07368577e-01 -1.16957891e+00 1.02369809e+00
1.16776681e+00 4.94380385e-01 2.35086322e-01 1.19087660e+00
-5.61397374e-01 1.44140708e+00 -8.39312255e-01 -7.85672307e-01
-3.94904137e-01 -7.15749562e-01 -1.97006214e+00 -2.61168420e-01
4.19261843e-01 -4.32026297e-01 -1.36772811e-01 5.38120568e-01
1.88093245e-01 2.25084871e-01 5.46604514e-01 -1.06969166e+00
-5.06927133e-01 7.84881651e-01 5.45545593e-02 5.22173047e-01
1.18538165e+00 2.42362618e-01 1.57680404e+00 -6.16391182e-01
6.53033018e-01 1.42827284e+00 3.04817677e-01 8.18923950e-01
-8.85922372e-01 -2.05684602e-01 -1.79263234e-01 4.82669950e-01
-4.64599222e-01 -6.46586478e-01 5.22399306e-01 -2.82234758e-01
1.32619536e+00 5.93718529e-01 3.46679449e-01 1.15837467e+00
-7.55786821e-02 1.11259484e+00 8.64264548e-01 -4.43384498e-01
1.39335217e-02 -2.50858925e-02 6.91406667e-01 4.04341042e-01
-2.79295623e-01 -5.28093994e-01 -6.87897742e-01 -3.79221290e-01
1.03398688e-01 -4.68484879e-01 -5.57378113e-01 1.22847922e-01
-1.34305966e+00 9.51060951e-01 4.13923025e-01 3.08162779e-01
-4.43178058e-01 -2.00857997e-01 3.16807628e-01 7.35483170e-01
6.91090748e-02 9.05226231e-01 -6.35485172e-01 -7.91432977e-01
3.34723406e-02 6.54532731e-01 1.69349933e+00 8.26821744e-01
4.62351531e-01 -9.17163551e-01 -6.77529871e-02 1.21846282e+00
1.54582560e-01 6.12582326e-01 2.45373040e-01 -1.65038764e+00
9.27964926e-01 7.40384579e-01 4.73053753e-01 -9.93529677e-01
-4.06478584e-01 1.17270328e-01 -3.90585810e-01 -7.23374605e-01
9.59647894e-01 -4.63024884e-01 9.90818255e-03 1.63317096e+00
3.02644551e-01 -5.50313413e-01 6.08563066e-01 6.98786139e-01
1.39976346e+00 8.80916893e-01 8.89928173e-03 -1.02155387e-01
1.88599217e+00 -1.14614213e+00 -8.63293886e-01 -3.92480344e-01
7.51752377e-01 -9.69400764e-01 1.29109442e+00 1.74274147e-01
-1.27179217e+00 -1.74295530e-01 -5.32208443e-01 -5.09102702e-01
8.43554549e-03 -3.96470964e-01 3.22129875e-01 1.37809813e-01
-9.58646417e-01 -7.46822059e-02 -2.87257224e-01 -6.61278605e-01
1.73879974e-02 -9.73102078e-02 -2.10247397e-01 -2.54881024e-01
-1.50956643e+00 1.01779544e+00 1.80711932e-02 -1.18683092e-01
-6.07230186e-01 -4.85309452e-01 -5.90931296e-01 9.61005539e-02
5.65358043e-01 -6.40144646e-01 2.16733551e+00 -5.48581600e-01
-1.44223309e+00 9.01758671e-01 -6.16208792e-01 -4.51244295e-01
3.37155193e-01 -5.85349143e-01 -3.29523116e-01 6.02178752e-01
2.78037012e-01 6.41280472e-01 1.61107779e-01 -1.00480568e+00
-7.88531542e-01 -1.72896609e-01 7.43113637e-01 4.98274833e-01
6.49541989e-02 1.04691558e-01 -1.95913017e-01 1.77141145e-01
-6.49410337e-02 -7.05114782e-01 2.11873025e-01 -4.85837013e-01
-3.12024117e-01 -9.00238633e-01 5.50130785e-01 -1.05580866e+00
9.63688791e-01 -1.69055581e+00 9.39881578e-02 -3.39269370e-01
4.25079107e-01 5.97248934e-02 -1.99041054e-01 1.10812008e+00
4.86520588e-01 5.26136868e-02 -2.76663244e-01 -5.42412288e-02
1.44810587e-01 2.69824952e-01 -4.49668914e-01 -1.62295133e-01
3.20690215e-01 1.20214784e+00 -1.20916378e+00 -3.75051290e-01
-8.28118771e-02 1.71882566e-02 -5.89713156e-01 7.51491129e-01
-8.42952192e-01 5.41881144e-01 -6.82071328e-01 3.04224014e-01
2.68900990e-01 -6.36332750e-01 6.41908050e-02 3.04765016e-01
1.05679937e-01 1.27768016e+00 -3.21589589e-01 1.56389463e+00
-4.83347923e-01 9.34727192e-01 2.57148862e-01 -5.63985825e-01
7.48815954e-01 4.34243977e-01 -8.41736868e-02 -8.08769166e-01
2.79884599e-03 3.90008576e-02 2.99087554e-01 -9.31467175e-01
6.22342527e-01 5.03836088e-02 -2.91064918e-01 9.27141428e-01
-1.42500803e-01 -4.40424502e-01 3.21137518e-01 8.52156639e-01
1.39934158e+00 -6.04289949e-01 2.56249249e-01 -3.61455142e-01
7.94610798e-01 5.03838122e-01 -4.95204926e-02 8.75480413e-01
-2.97126442e-01 2.27792665e-01 9.31953311e-01 -1.86443806e-01
-7.12241173e-01 -1.00379395e+00 9.28641781e-02 1.26992643e+00
1.10221077e-02 -4.80570465e-01 -8.91100407e-01 -6.48201823e-01
-2.06426486e-01 1.03301954e+00 -4.04451758e-01 2.86044747e-01
-8.06520998e-01 -1.49928525e-01 6.25973761e-01 3.55583936e-01
8.59573066e-01 -1.42932105e+00 -3.79442245e-01 3.71683806e-01
-1.41410911e+00 -1.31944978e+00 -3.19019467e-01 -2.39008829e-01
-5.89791656e-01 -1.40945101e+00 -5.17162025e-01 -8.04430246e-01
5.68526648e-02 4.23624665e-01 1.73559272e+00 3.54859263e-01
2.54756689e-01 7.49139905e-01 -8.24181080e-01 -3.91543478e-01
-7.79113173e-01 3.38310421e-01 -6.06753051e-01 -4.67158347e-01
8.00938129e-01 -4.11204010e-01 -7.09982812e-01 5.54229498e-01
-4.06402797e-01 1.59428567e-02 1.87828854e-01 8.99274111e-01
-3.15271586e-01 -6.81240618e-01 1.01243842e+00 -8.87342155e-01
1.26138091e+00 -8.45154405e-01 -9.64628756e-02 3.60764563e-01
7.03506842e-02 -7.53511786e-02 4.66025412e-01 1.42476618e-01
-1.40310800e+00 -6.72365010e-01 -5.50891519e-01 8.40974629e-01
-3.90127629e-01 4.31758583e-01 -6.95780292e-02 6.39982998e-01
9.26108003e-01 2.18924507e-02 2.48279154e-01 -3.52847308e-01
4.83240277e-01 1.07040608e+00 5.72486937e-01 -8.52905154e-01
1.95559904e-01 2.08021939e-01 -8.15801680e-01 -1.20624721e+00
-1.23825300e+00 -8.91041934e-01 -2.28630006e-01 -3.01671118e-01
9.85115290e-01 -7.13796556e-01 -1.19217503e+00 3.09239209e-01
-1.51433051e+00 -5.99569559e-01 -1.91031490e-02 1.92651510e-01
-6.38988972e-01 5.27755797e-01 -8.85789812e-01 -6.32639706e-01
-4.68620867e-01 -7.92571723e-01 6.62312686e-01 3.59450400e-01
-1.03380609e+00 -1.03973436e+00 1.48263738e-01 1.37537336e+00
4.34330672e-01 -3.14396590e-01 9.97712195e-01 -1.08674371e+00
-4.70912695e-01 2.01222956e-01 -1.98424727e-01 1.78161740e-01
7.45562986e-02 -3.57735783e-01 -9.81387556e-01 7.93226734e-02
9.95155275e-02 -9.35660362e-01 4.96262252e-01 -7.49041364e-02
6.53498828e-01 -4.42388326e-01 -3.35556176e-03 -4.24729258e-01
5.57374299e-01 1.63281649e-01 5.74199140e-01 1.91543609e-01
8.27563480e-02 1.12031448e+00 4.35981721e-01 1.07970245e-01
1.11650956e+00 2.55930394e-01 1.61104038e-01 5.85482478e-01
-3.11396960e-02 -2.94374377e-01 2.85624743e-01 1.29301941e+00
1.78673059e-01 -6.12487495e-01 -1.36381042e+00 8.84625137e-01
-1.85905564e+00 -1.12385654e+00 -5.53449631e-01 1.55199909e+00
1.07783723e+00 2.97894739e-02 1.08017974e-01 -1.93480447e-01
4.30449784e-01 1.94045767e-01 -4.18488979e-01 -6.19215250e-01
-7.28781596e-02 5.41907661e-02 -5.00493050e-01 1.14181793e+00
-6.58833802e-01 8.04231405e-01 5.98154640e+00 2.08778277e-01
-2.77258247e-01 -2.65876930e-02 4.10724878e-01 3.54987860e-01
-5.97291887e-01 -9.35281347e-03 -5.26045978e-01 1.82290912e-01
1.19087577e+00 -2.66370267e-01 3.48073691e-01 4.19374377e-01
4.77198586e-02 -5.47162116e-01 -1.14563954e+00 5.30142486e-01
1.05790600e-01 -1.23849905e+00 3.49183120e-02 -3.44918668e-01
5.05169928e-01 1.43347621e-01 -4.82692480e-01 6.06756449e-01
6.98315918e-01 -1.08296168e+00 1.40903925e-03 3.91003072e-01
9.02508870e-02 -4.56500590e-01 9.04679835e-01 7.92060077e-01
-6.01190627e-01 -5.65669648e-02 -2.25821987e-01 -5.22848368e-01
4.84666556e-01 1.27048880e-01 -1.16132796e+00 2.27618545e-01
6.88582897e-01 4.23712194e-01 -3.13055307e-01 6.93112671e-01
-5.66986442e-01 9.96861339e-01 -3.17567915e-01 -7.99928069e-01
3.59951526e-01 -6.97096735e-02 6.30514681e-01 1.22401619e+00
-3.10815006e-01 8.15549970e-01 -1.24776088e-01 5.38549781e-01
-3.58643800e-01 1.19329505e-01 -3.65777463e-01 -1.79680660e-01
7.21013308e-01 9.01953042e-01 -1.57201096e-01 -4.06563938e-01
-5.50923407e-01 8.18366230e-01 5.45541167e-01 2.50912577e-01
-3.61085624e-01 -5.00449181e-01 6.60091758e-01 -2.01670945e-01
-4.96603064e-02 -1.70614183e-01 -9.29564610e-03 -1.13853419e+00
2.71075010e-01 -1.46735001e+00 6.30897939e-01 -9.36775625e-01
-1.59563470e+00 5.04900932e-01 -1.35215521e-01 -5.22451043e-01
-7.92752624e-01 -3.86729568e-01 -6.45628154e-01 7.83367395e-01
-1.35067451e+00 -6.63134098e-01 -6.30308926e-01 6.31536901e-01
8.85629952e-01 1.72497824e-01 9.31338072e-01 3.91437532e-03
2.33023893e-02 2.16317222e-01 -1.66271672e-01 4.90578443e-01
8.29720557e-01 -1.32825482e+00 6.20095193e-01 1.43844366e-01
-1.62256472e-02 7.63228178e-01 7.97586083e-01 -3.53145212e-01
-1.37349963e+00 -3.85546803e-01 1.52268040e+00 -1.16110754e+00
8.81512165e-01 -1.90659165e-01 -1.28001571e+00 8.60701025e-01
1.04032850e+00 -9.62308764e-01 9.89972174e-01 5.32282054e-01
-4.39122617e-01 2.66060203e-01 -9.75277781e-01 6.98944211e-01
9.13986921e-01 -8.12162161e-01 -1.65519404e+00 5.28407753e-01
1.11575711e+00 -3.40734571e-01 -7.51045287e-01 6.73598796e-02
2.87709415e-01 -1.02328146e+00 7.33418345e-01 -9.01078761e-01
7.84643352e-01 8.31478089e-02 -2.17251256e-01 -1.32322454e+00
8.55991542e-02 -7.31561482e-01 -1.64908320e-01 1.12007403e+00
8.04539382e-01 -6.98257565e-01 4.28399533e-01 8.51851463e-01
-1.36638790e-01 -4.03877974e-01 -7.96920121e-01 -1.45567551e-01
5.37454724e-01 -3.42296332e-01 5.48875034e-01 8.72742712e-01
8.09781432e-01 1.09682429e+00 1.05905652e-01 -5.14182225e-02
2.89338410e-01 2.39521846e-01 1.02847803e+00 -1.07648408e+00
-1.88166544e-01 -3.49633574e-01 3.64721239e-01 -1.85317087e+00
1.90948695e-01 -6.61343575e-01 2.06017449e-01 -1.92067564e+00
2.69889981e-01 -7.84076471e-03 4.73073035e-01 2.00999126e-01
-3.07821363e-01 -3.78684968e-01 5.81732392e-02 2.18889803e-01
-8.92146289e-01 5.49064875e-01 1.54083395e+00 -2.00002521e-01
-1.10161237e-01 -3.80479023e-02 -9.82178748e-01 6.17300928e-01
1.05410075e+00 6.14917874e-02 -4.48348254e-01 -6.35477066e-01
3.08140218e-01 5.81342340e-01 2.72544652e-01 -6.39307022e-01
5.02161682e-01 -1.47529244e-01 -2.01942742e-01 -5.98389328e-01
5.22941113e-01 -2.46256113e-01 -6.45427823e-01 3.20607901e-01
-9.13988292e-01 1.50745660e-01 1.47813544e-01 4.02015209e-01
-4.29846168e-01 -2.12543905e-01 3.55470538e-01 -3.29119086e-01
-6.09845400e-01 -5.53850651e-01 -6.78126633e-01 9.42010343e-01
5.21992981e-01 4.46664035e-01 -1.07003546e+00 -1.20163751e+00
-6.59678698e-01 9.89940345e-01 -4.68788669e-02 5.68773031e-01
5.05254984e-01 -6.67856991e-01 -1.09757686e+00 -4.34090555e-01
3.71774077e-01 1.67743303e-02 3.49318802e-01 5.56552708e-01
-4.68349218e-01 8.54926944e-01 1.41984690e-02 -5.75579524e-01
-1.22433674e+00 -7.63030797e-02 3.03842217e-01 -1.58392787e-01
-5.11977017e-01 9.81413424e-01 -1.91124342e-02 -8.68722320e-01
8.99255946e-02 -3.63480270e-01 -5.71420431e-01 2.10008904e-01
9.02114928e-01 3.23820740e-01 -7.12484270e-02 -2.34591126e-01
-1.82514846e-01 1.02029271e-01 -3.00597608e-01 -3.03976327e-01
1.01663554e+00 -4.72962707e-01 -4.29074824e-01 5.01311898e-01
1.12632310e+00 1.07841492e-01 -7.63181746e-01 -4.34814245e-01
3.26571614e-01 -1.83177039e-01 -6.36849761e-01 -1.32110393e+00
-6.91671446e-02 8.96707296e-01 -3.42129618e-01 7.41195321e-01
5.15200913e-01 6.87108696e-01 1.24271190e+00 1.11217737e+00
2.66096264e-01 -8.73712122e-01 4.51850116e-01 1.15486372e+00
1.36384273e+00 -1.41247332e+00 -4.15019006e-01 -3.79248977e-01
-8.72312903e-01 9.04892325e-01 7.20609307e-01 1.34455130e-01
2.15290368e-01 -1.83218315e-01 5.27016640e-01 -5.64286768e-01
-1.37069118e+00 -1.74195915e-01 1.23746298e-01 3.12757552e-01
5.77450812e-01 8.75771642e-02 -3.62632930e-01 5.27932644e-01
-7.46117890e-01 -3.47184449e-01 6.75948977e-01 7.14761496e-01
-7.93488324e-01 -8.32041979e-01 -1.75675616e-01 4.60641861e-01
-2.56212145e-01 -7.66562968e-02 -9.82102156e-01 7.19133556e-01
-8.58782172e-01 1.84779894e+00 1.41502187e-01 -4.54298109e-02
4.75100428e-01 3.76810074e-01 1.72047004e-01 -5.13333738e-01
-8.09642613e-01 -6.72497809e-01 1.00440145e+00 -4.60897446e-01
-4.16951060e-01 -8.32767963e-01 -1.37028134e+00 -5.61321974e-01
-1.40949368e-01 9.59753752e-01 2.28019163e-01 1.03958011e+00
4.41319972e-01 1.46128684e-01 2.97164708e-01 1.56043589e-01
-5.99898636e-01 -1.20733583e+00 2.69269615e-01 4.89930272e-01
6.09087527e-01 -7.44591579e-02 -5.44374287e-01 -1.30932152e-01] | [12.073373794555664, 8.012422561645508] |
32b96701-ed2a-4aa9-9af7-4c20d4cac93e | sentiment-analysis-itas-complicated | null | null | https://aclanthology.org/N18-1171 | https://aclanthology.org/N18-1171.pdf | Sentiment Analysis: It's Complicated! | Sentiment analysis is used as a proxy to measure human emotion, where the objective is to categorize text according to some predefined notion of sentiment. Sentiment analysis datasets are typically constructed with gold-standard sentiment labels, assigned based on the results of manual annotations. When working with such annotations, it is common for dataset constructors to discard {``}noisy{''} or {``}controversial{''} data where there is significant disagreement on the proper label. In datasets constructed for the purpose of Twitter sentiment analysis (TSA), these controversial examples can compose over 30{\%} of the originally annotated data. We argue that the removal of such data is a problematic trend because, when performing real-time sentiment classification of short-text, an automated system cannot know a priori which samples would fall into this category of disputed sentiment. We therefore propose the notion of a {``}complicated{''} class of sentiment to categorize such text, and argue that its inclusion in the short-text sentiment analysis framework will improve the quality of automated sentiment analysis systems as they are implemented in real-world settings. We motivate this argument by building and analyzing a new publicly available TSA dataset of over 7,000 tweets annotated with 5x coverage, named MTSA. Our analysis of classifier performance over our dataset offers insights into sentiment analysis dataset and model design, how current techniques would perform in the real world, and how researchers should handle difficult data. | ['Rohit Verma', 'Robert Belfer', 'Bh', 'Lal', 'Jeremy Georges-Filteau', 'Kian Kenyon-Dean', 'Scott Fujimoto', 'Shruti eri', 'Nirmal Kanagasabai', 'Eisha Ahmed', 'Derek Ruths', 'Auguste e', 'Roman Sarrazingendron', 'Christopher Glasz', 'Barleen Kaur'] | 2018-06-01 | null | null | null | naacl-2018-6 | ['twitter-sentiment-analysis'] | ['natural-language-processing'] | [ 2.83848435e-01 2.06940070e-01 1.70429293e-02 -8.24204385e-01
-7.30576515e-01 -1.02096820e+00 5.40267646e-01 9.32896376e-01
-6.99020982e-01 4.97733563e-01 4.87206906e-01 -5.26957631e-01
3.19719426e-02 -7.76692569e-01 -3.88773382e-01 -5.75372219e-01
6.33124828e-01 4.19488996e-01 -8.90306681e-02 -6.30525470e-01
3.62902164e-01 -7.27307275e-02 -1.76746762e+00 5.80132961e-01
3.61847967e-01 1.28813481e+00 -2.97089696e-01 2.91639596e-01
-4.08661813e-01 9.49374020e-01 -1.05236149e+00 -8.89834344e-01
8.99853185e-02 -1.96103856e-01 -1.05607188e+00 -1.33658340e-02
-6.82029501e-02 2.61088550e-01 7.29013681e-01 1.00590134e+00
2.30132639e-01 -1.07990585e-01 5.50643086e-01 -1.19946182e+00
-1.77546501e-01 8.32739472e-01 -2.88799554e-01 -2.43598558e-02
3.38406652e-01 -1.35601178e-01 1.45240748e+00 -5.22219658e-01
6.80978298e-01 6.42549992e-01 8.78385723e-01 2.16078356e-01
-8.93243730e-01 -6.03639603e-01 2.42781252e-01 -4.30544108e-01
-1.04992652e+00 -3.73366624e-01 5.66892982e-01 -6.97813630e-01
5.98924398e-01 6.85404420e-01 6.02322221e-01 1.11359572e+00
-8.14368762e-03 5.77535152e-01 1.42079723e+00 -4.47028965e-01
4.29785043e-01 7.12035120e-01 4.43177611e-01 -2.04149894e-02
4.20912951e-01 -6.12839103e-01 -5.42919993e-01 -4.11688268e-01
-4.70221132e-01 -8.86202976e-02 -3.00695729e-02 1.39399335e-01
-8.58468652e-01 1.03510022e+00 1.14540346e-02 4.00403500e-01
-2.53245443e-01 -2.89457351e-01 8.42008173e-01 3.79249066e-01
9.16035235e-01 6.99542701e-01 -1.07553089e+00 -4.10494775e-01
-1.03159165e+00 2.47397631e-01 9.36963260e-01 6.94097877e-01
6.46919012e-01 -4.26316142e-01 2.46127725e-01 9.36306715e-01
3.33156794e-01 3.35094720e-01 7.04475462e-01 -6.58170283e-01
2.55873144e-01 1.06097245e+00 4.49974626e-01 -1.06952155e+00
-6.73984408e-01 -2.34611899e-01 -3.78494114e-01 -3.76488678e-02
5.96381128e-01 -4.44769055e-01 -5.15421867e-01 1.46666741e+00
2.42370889e-01 -6.64178550e-01 2.00715095e-01 7.18150198e-01
9.26355600e-01 3.19574535e-01 1.64093062e-01 -2.37121433e-01
1.66668582e+00 -2.35593691e-01 -5.87526917e-01 -2.39699468e-01
1.19142032e+00 -9.69249606e-01 1.33894849e+00 6.55336857e-01
-5.32695830e-01 -1.32796124e-01 -8.46164823e-01 7.86209255e-02
-9.18748200e-01 -2.03339905e-01 7.18939900e-01 1.14100218e+00
-6.68749809e-01 3.47312301e-01 -3.60892206e-01 -3.40773284e-01
3.92522693e-01 2.14093924e-01 -3.57662410e-01 4.48601604e-01
-1.30909371e+00 6.89561486e-01 6.05380209e-03 3.99407521e-02
-4.34276536e-02 -4.91665393e-01 -6.86900914e-01 -1.73885837e-01
3.38009596e-01 -9.98524204e-02 1.33588803e+00 -1.45231271e+00
-9.10365343e-01 1.35354364e+00 -1.78084418e-01 -1.90050632e-01
3.58395904e-01 1.34766148e-02 -5.33686757e-01 -2.82995015e-01
4.52120125e-01 6.87740669e-02 5.39632618e-01 -1.37719440e+00
-1.03993785e+00 -4.60892230e-01 3.55040431e-01 -3.41566056e-02
-4.01677430e-01 4.04619485e-01 6.12334944e-02 -5.59920907e-01
2.00143587e-02 -1.15500379e+00 -2.00462654e-01 -6.55081391e-01
-4.85473692e-01 -3.08399469e-01 7.61582375e-01 -1.94232091e-01
1.38826561e+00 -2.03840661e+00 -6.44559205e-01 4.19615746e-01
6.62708953e-02 9.46593061e-02 5.10188699e-01 5.50622463e-01
-6.39371276e-02 5.80944359e-01 -6.36328906e-02 -3.99731755e-01
2.85347462e-01 1.01338953e-01 -7.24133968e-01 3.59509140e-01
-1.60257488e-01 3.90380442e-01 -9.31393504e-01 -4.03644949e-01
-1.11225814e-01 1.39297575e-01 -4.75582957e-01 -1.16005346e-01
-2.09832996e-01 3.41385305e-01 -6.09038174e-01 6.80853069e-01
2.96199501e-01 -2.76660979e-01 3.43635887e-01 -2.15807676e-01
-3.36483955e-01 5.46155989e-01 -1.04380941e+00 9.62372661e-01
-3.33084077e-01 7.35613763e-01 6.12238422e-02 -8.50874901e-01
1.02092457e+00 2.46955737e-01 6.06991768e-01 -3.30882311e-01
7.73055434e-01 2.60477334e-01 -1.69840783e-01 -3.20912778e-01
9.80395436e-01 -4.23462749e-01 -8.15337718e-01 7.19961107e-01
-3.60129178e-01 -5.45460582e-01 2.96524674e-01 1.25315592e-01
1.09589183e+00 -2.50697434e-01 8.34556818e-02 -4.37286377e-01
4.64029580e-01 4.85367864e-01 3.98783803e-01 5.50212324e-01
-3.40899765e-01 6.17411137e-01 1.06845653e+00 -4.91083294e-01
-8.68158281e-01 -2.88704574e-01 -3.54912907e-01 1.49866462e+00
-1.20170258e-01 -8.38461101e-01 -6.57653272e-01 -8.63153398e-01
-1.85507074e-01 6.95071936e-01 -9.42446351e-01 1.90675452e-01
5.20220734e-02 -1.18968928e+00 5.86702168e-01 9.35556069e-02
7.34016672e-02 -1.02368426e+00 -8.12923193e-01 1.50248244e-01
-4.42045569e-01 -1.00741410e+00 7.84860775e-02 6.49650574e-01
-9.28443968e-02 -1.20819187e+00 -8.79459754e-02 -4.06660408e-01
7.21149445e-01 4.30290774e-03 1.20513916e+00 2.12709159e-01
4.16291952e-01 5.92060164e-02 -9.63269830e-01 -8.87597978e-01
-4.75936055e-01 1.55322313e-01 -1.15278579e-01 1.77512348e-01
8.54119301e-01 -6.09700084e-02 -6.05624020e-01 4.50395644e-01
-1.09920418e+00 -2.94643700e-01 -1.18126132e-01 6.18937373e-01
3.83051336e-01 2.69691944e-01 6.90724373e-01 -1.62795651e+00
9.41038191e-01 -6.06633246e-01 -2.82653958e-01 -1.56487152e-01
-7.15539157e-01 -4.11697596e-01 8.08855295e-01 5.79663087e-03
-8.76977861e-01 -2.43404388e-01 -4.66912925e-01 2.23346278e-01
-2.25856602e-01 8.97262752e-01 5.96889406e-02 3.32237303e-01
8.23743343e-01 -4.01541412e-01 -2.86373228e-01 -1.43599793e-01
-4.99086827e-03 1.15259182e+00 1.07431114e-01 -5.44755280e-01
3.33535433e-01 7.38374770e-01 -3.04414272e-01 -4.89939511e-01
-1.55085719e+00 -7.65505016e-01 -3.75878811e-01 -2.17884779e-01
5.97992122e-01 -9.63897586e-01 -8.57863843e-01 2.39886254e-01
-5.27949333e-01 -2.90997237e-01 -4.47809160e-01 1.23782955e-01
-2.94935852e-01 1.59111828e-01 -2.09183246e-01 -9.21242595e-01
-4.35524344e-01 -1.26847780e+00 1.12369847e+00 1.00785136e-01
-1.24999809e+00 -1.01500142e+00 -1.66631900e-02 7.90886879e-01
2.24150017e-01 4.42537576e-01 6.09543800e-01 -1.19138122e+00
6.49114668e-01 -7.31023490e-01 1.37372941e-01 4.81398553e-01
1.29938766e-01 4.67023075e-01 -1.12849963e+00 -1.08633146e-01
8.03525969e-02 -5.78489006e-01 5.56483984e-01 6.81072474e-02
1.02568698e+00 -2.69605041e-01 -9.12651047e-02 2.50723306e-03
1.23056293e+00 4.37424555e-02 2.78135836e-01 7.28019595e-01
3.20419043e-01 1.17450202e+00 9.90388215e-01 6.62916422e-01
6.26174092e-01 3.07511032e-01 3.59008282e-01 1.38013378e-01
7.05225348e-01 1.66600682e-02 4.42879707e-01 7.74995625e-01
2.26727381e-01 -4.71740097e-01 -1.06994963e+00 6.22062504e-01
-1.70327115e+00 -7.97036946e-01 -4.97709423e-01 1.92398834e+00
9.73378360e-01 7.21995592e-01 9.13019702e-02 6.95802629e-01
5.50765932e-01 2.07715094e-01 1.08602354e-02 -8.74280751e-01
-2.92881727e-01 2.40781710e-01 5.13405681e-01 2.23075822e-01
-1.16634429e+00 7.65381694e-01 5.62744617e+00 5.95682204e-01
-1.29639208e+00 2.85869185e-02 9.65733886e-01 -1.07316904e-01
-4.50334609e-01 2.26511002e-01 -8.28528285e-01 7.18015075e-01
1.01893079e+00 4.05368581e-02 -3.41312319e-01 8.92549038e-01
3.89429092e-01 -4.85501975e-01 -7.14596510e-01 5.60334384e-01
1.29219349e-02 -1.19008183e+00 -2.88199812e-01 -9.64009166e-02
6.69300079e-01 -7.77427666e-03 7.28315860e-02 4.12199110e-01
6.10781312e-01 -9.04282629e-01 1.13070989e+00 6.69124648e-02
7.08448648e-01 -7.08128154e-01 1.18636203e+00 2.87786603e-01
-6.82257712e-01 -2.46848091e-02 5.62286191e-02 -3.30597013e-01
7.14289621e-02 9.11388695e-01 -6.68439388e-01 2.88913727e-01
1.06767559e+00 5.88318229e-01 -6.84341252e-01 3.24445307e-01
-1.00273386e-01 9.36333299e-01 -3.46301436e-01 -3.35566133e-01
3.72679949e-01 -1.67538017e-01 1.45361021e-01 1.34309423e+00
7.76367933e-02 3.24127302e-02 -7.60345673e-03 1.03982076e-01
-1.79988757e-01 4.05350596e-01 -3.16834927e-01 -9.55616385e-02
2.33139187e-01 1.57108331e+00 -1.38048899e+00 -3.36664915e-01
-4.74371016e-01 2.71854401e-01 -7.82228485e-02 -3.06926165e-02
-5.77369034e-01 -3.75795037e-01 5.28792083e-01 2.05960661e-01
-4.25965190e-02 3.62469554e-01 -7.74345219e-01 -1.06561983e+00
2.60302033e-02 -9.95245457e-01 5.75550199e-01 -7.32585788e-01
-1.35539627e+00 6.55737877e-01 -3.06588262e-01 -1.18052351e+00
6.25230297e-02 -7.60411561e-01 -3.05124342e-01 5.65011084e-01
-1.17279172e+00 -9.88948286e-01 -3.21573198e-01 2.87026644e-01
1.70665503e-01 2.13244259e-01 8.18304002e-01 1.68646082e-01
-3.17593873e-01 3.45007002e-01 8.85962322e-02 2.71126717e-01
9.77311850e-01 -1.21196830e+00 -2.28042193e-02 4.13707167e-01
-8.91799778e-02 5.98235965e-01 1.30478001e+00 -5.21973789e-01
-8.87705624e-01 -9.96752679e-01 1.18987167e+00 -9.76704180e-01
1.05978131e+00 -4.25862402e-01 -7.11950719e-01 6.76193297e-01
-1.04566276e-01 -3.07796955e-01 1.49266315e+00 4.69896585e-01
-2.95529693e-01 -7.65045136e-02 -1.22123981e+00 4.15266246e-01
3.88158321e-01 -5.43805122e-01 -6.06347084e-01 4.49288845e-01
3.43375176e-01 -2.49648347e-01 -9.57779884e-01 3.15433621e-01
7.35485911e-01 -1.03756797e+00 2.48704597e-01 -4.59807307e-01
5.36599755e-01 -2.69569904e-01 -3.25897574e-01 -1.28603423e+00
3.29710960e-01 -4.00137544e-01 7.84168363e-01 1.39318883e+00
6.84766829e-01 -6.18872046e-01 9.27283406e-01 9.28386867e-01
-8.73301700e-02 -6.77267790e-01 -7.20896661e-01 -5.80668077e-02
2.44228363e-01 -1.01615584e+00 7.53095269e-01 1.41562414e+00
4.36001867e-01 2.02390775e-01 1.67726334e-02 -1.87440649e-01
-5.13306744e-02 8.42239410e-02 9.30727363e-01 -1.46142244e+00
2.73084909e-01 -4.67423528e-01 -2.49189869e-01 -2.31937870e-01
1.42020047e-01 -6.41878486e-01 -5.28677031e-02 -1.28916740e+00
-1.01675354e-01 -8.90994072e-01 -1.50039911e-01 6.01309776e-01
2.59144474e-02 6.08813345e-01 -1.53865749e-02 1.91364929e-01
-6.83693826e-01 3.10019925e-02 7.32154429e-01 1.09536685e-01
-1.42855689e-01 7.25319535e-02 -1.39045882e+00 1.12823379e+00
7.18351185e-01 -6.64477050e-01 -1.76068485e-01 -7.35213468e-03
1.41160691e+00 -5.35617948e-01 -3.80244921e-03 -5.01575649e-01
3.65467891e-02 -3.16658497e-01 1.68417692e-01 -3.94465983e-01
1.25814125e-01 -9.63740170e-01 1.46374911e-01 1.86301768e-02
-4.71101284e-01 -5.00440300e-02 6.69346601e-02 5.21278568e-02
-4.01228100e-01 -4.08000112e-01 5.98873675e-01 -1.70783177e-01
-3.53259265e-01 -1.95509464e-01 -5.80219150e-01 2.46132806e-01
8.96146834e-01 -1.77841708e-01 -4.34653729e-01 -4.26250398e-01
-7.19502449e-01 1.63096026e-01 7.16433525e-01 3.60415936e-01
-1.36468098e-01 -7.37658978e-01 -5.80134630e-01 -4.59540375e-02
6.29180908e-01 6.56844750e-02 -9.70981866e-02 7.34594882e-01
-4.46197778e-01 1.72140643e-01 2.11044282e-01 -2.76558608e-01
-1.18022871e+00 4.78284247e-02 5.64601496e-02 -3.06403071e-01
-3.25798988e-02 7.23553002e-01 -8.02580044e-02 -5.04412472e-01
-1.81881681e-01 -4.15509522e-01 -5.81208289e-01 9.16328609e-01
3.09320599e-01 -3.87404561e-02 5.64719856e-01 -1.16467392e+00
-3.66009116e-01 1.07834645e-01 -3.50620821e-02 -1.25777110e-01
1.31647658e+00 -4.27886546e-01 -3.08795363e-01 9.49517429e-01
1.03599834e+00 5.27789116e-01 -6.11828566e-01 1.23227634e-01
2.40525886e-01 -3.11138868e-01 -1.30171239e-01 -8.73317599e-01
-6.96214259e-01 3.70027155e-01 4.41277549e-02 1.02486169e+00
9.63464677e-01 -3.83649468e-02 5.16321361e-01 1.57389894e-01
3.41109186e-01 -1.54881036e+00 -2.98306733e-01 6.05488896e-01
5.60866714e-01 -1.47607076e+00 1.76374525e-01 -2.46426761e-01
-9.57166851e-01 9.79628980e-01 2.75362730e-01 -2.04371884e-02
8.59661520e-01 2.79230744e-01 6.07782960e-01 -5.64661384e-01
-6.66343570e-01 -1.26309648e-01 -7.60937035e-02 8.98273960e-02
7.21545219e-01 2.44163752e-01 -5.87042034e-01 1.18528855e+00
-1.10015285e+00 -2.40474135e-01 9.00190592e-01 1.13344491e+00
-3.22213709e-01 -9.58887696e-01 -3.79202515e-01 8.93092573e-01
-1.21794820e+00 5.17994724e-02 -6.26454830e-01 6.30334675e-01
4.14833367e-01 1.50289774e+00 9.15957168e-02 -3.45917612e-01
4.58665580e-01 1.22669764e-01 -2.60604262e-01 -7.27127492e-01
-1.28844106e+00 -1.27418652e-01 5.62128067e-01 -1.16441570e-01
-7.88730383e-01 -8.71026754e-01 -1.24628508e+00 -3.72094452e-01
-4.51019019e-01 5.59053600e-01 8.35154176e-01 1.25740254e+00
1.92461580e-01 1.64985895e-01 6.33027136e-01 -3.54448617e-01
-1.05791919e-01 -7.99829483e-01 -5.58708429e-01 9.72217381e-01
2.73726493e-01 -4.29891616e-01 -7.38331020e-01 3.36530983e-01] | [11.065223693847656, 6.914811611175537] |
f96e61b0-15d4-4d57-a26a-c842e906a2c3 | engineering-education-in-the-age-of | 2102.07900 | null | https://arxiv.org/abs/2102.07900v1 | https://arxiv.org/pdf/2102.07900v1.pdf | Engineering Education in the Age of Autonomous Machines | In the past few years, we have observed a huge supply-demand gap for autonomous driving engineers. The core problem is that autonomous driving is not one single technology but rather a complex system integrating many technologies, and no one single academic department can provide comprehensive education in this field. We advocate to create a cross-disciplinary program to expose students with technical background in computer science, computer engineering, electrical engineering, as well as mechanical engineering. On top of the cross-disciplinary technical foundation, a capstone project that provides students with hands-on experiences of working with a real autonomous vehicle is required to consolidate the technical foundation. | ['Hironori Kasahara', 'Jean-Luc Gaudiot', 'Shaoshan Liu'] | 2021-02-16 | null | null | null | null | ['electrical-engineering'] | ['miscellaneous'] | [-3.54985416e-01 2.72745788e-02 -1.68658540e-01 -2.14505777e-01
-4.44368482e-01 -6.03913903e-01 4.83464599e-02 -1.49173945e-01
-3.74230631e-02 5.27413309e-01 -4.65902746e-01 -9.91520464e-01
-2.85545141e-02 -7.59314656e-01 -7.24807620e-01 -2.48147205e-01
2.50843227e-01 1.00253582e-01 5.14606118e-01 -8.53371263e-01
4.13755000e-01 6.44298434e-01 -1.98956728e+00 -4.24361765e-01
1.15574324e+00 6.18697405e-01 1.81680962e-01 5.23152292e-01
-9.68171507e-02 5.38959146e-01 -4.12069350e-01 -1.45037457e-01
8.19613561e-02 -2.41305187e-01 -5.36209643e-01 -2.34552771e-01
2.36277625e-01 -7.11528286e-02 -3.84549797e-01 8.44461143e-01
-3.73372750e-04 -3.05639684e-01 4.43808325e-02 -1.96228886e+00
-1.37036577e-01 1.10398144e-01 -4.37568009e-01 5.07930852e-02
2.73050249e-01 4.40222800e-01 6.05292797e-01 -8.35465133e-01
4.08201337e-01 7.14103103e-01 5.76728225e-01 1.43837810e-01
-9.94143248e-01 -8.83507073e-01 1.99919119e-01 6.56555295e-02
-1.41959333e+00 -2.33721003e-01 7.38473117e-01 -8.32872748e-01
6.32546127e-01 -1.52692124e-01 8.56904447e-01 6.63776636e-01
6.41933322e-01 3.76125395e-01 7.47506678e-01 -2.08619341e-01
2.01561570e-01 8.44307244e-01 4.75983799e-01 5.85763872e-01
6.47872269e-01 3.69371086e-01 -1.20898359e-01 2.69854665e-01
4.58086520e-01 1.31014556e-01 1.30569294e-01 -6.54634893e-01
-1.18759525e+00 6.52858555e-01 -1.01996988e-01 4.14512992e-01
-6.43284544e-02 3.55391651e-01 2.97237009e-01 5.43943405e-01
-4.83964860e-01 3.65581959e-01 -3.22521538e-01 -5.92905045e-01
-5.50951362e-01 4.41700190e-01 1.06856537e+00 1.04927707e+00
8.90241742e-01 3.13845485e-01 8.39495659e-01 1.99267223e-01
3.78785372e-01 4.44076180e-01 -6.31338134e-02 -1.33089626e+00
2.39335354e-02 5.73054075e-01 1.96224719e-01 -9.55047011e-01
-1.33617902e-02 -2.35188335e-01 -2.04200268e-01 7.81094790e-01
1.46049261e-01 -6.00276649e-01 -3.21095496e-01 1.34995461e+00
8.86526257e-02 2.45478787e-02 7.55648734e-03 6.66632831e-01
6.64556086e-01 8.88307750e-01 1.09022051e-01 1.85936868e-01
1.45292175e+00 -7.05979705e-01 -6.12078309e-01 -5.55783868e-01
5.17245471e-01 -9.46658492e-01 7.99738050e-01 7.38456964e-01
-1.23520970e+00 -8.54395628e-01 -1.56669462e+00 2.46935919e-01
-6.09519064e-01 -1.57878906e-01 3.12348366e-01 8.70293558e-01
-1.04037642e+00 2.57442474e-01 -7.91951180e-01 -7.54946828e-01
-2.91630566e-01 2.77596146e-01 -2.20823556e-01 -1.98744640e-01
-1.11817122e+00 1.43061888e+00 4.67875935e-02 -3.88035595e-01
-6.94015563e-01 -9.38376844e-01 -7.42580950e-01 -1.90059423e-01
3.63969386e-01 -4.90140706e-01 1.32655251e+00 -1.27692223e-01
-1.19555879e+00 5.48669279e-01 -1.34521872e-01 8.91573876e-02
2.39008144e-01 -1.07220173e-01 -7.09607244e-01 -2.57790089e-01
1.76714107e-01 5.34720302e-01 2.44565710e-01 -1.39521992e+00
-1.12594318e+00 -1.92475572e-01 2.44234905e-01 1.25429705e-01
-3.82255167e-01 -3.76542136e-02 -7.88626000e-02 1.04546696e-02
7.27871945e-03 -1.07534575e+00 -2.43544519e-01 -6.71357140e-02
2.92685688e-01 -2.75247812e-01 1.45714414e+00 1.19155562e-02
9.73442137e-01 -2.29578757e+00 -1.77775338e-01 8.04943666e-02
4.79993708e-02 9.80610400e-02 3.54123145e-01 8.18336308e-01
8.01254064e-02 8.43639579e-03 3.63133699e-01 4.44050819e-01
3.65203023e-01 1.67567641e-01 -3.11756223e-01 2.12953895e-01
-1.53877605e-02 5.02913952e-01 -1.12933373e+00 -2.01753452e-01
5.20759642e-01 4.03725803e-01 -5.13298698e-02 -2.95587569e-01
2.29282603e-01 2.15321615e-01 -6.42947435e-01 3.92737627e-01
8.56580734e-01 -2.30004594e-01 2.56036520e-01 4.12771642e-01
-6.12929463e-01 2.26025864e-01 -1.30617678e+00 1.49378681e+00
-7.62907267e-01 8.83404016e-01 5.57519794e-01 -1.08731127e+00
1.02781463e+00 5.62878788e-01 6.62953317e-01 -3.28222573e-01
1.90913066e-01 7.24102020e-01 2.60491014e-01 -5.97139239e-01
9.83747125e-01 -1.55265108e-01 -3.09640050e-01 6.02565885e-01
-3.52246374e-01 -7.19066024e-01 4.54361439e-01 3.83763939e-01
1.12369108e+00 1.74170643e-01 -3.88771951e-01 -5.13128698e-01
4.08536166e-01 3.42506886e-01 4.33356285e-01 2.46782511e-01
-6.87780440e-01 -1.26374587e-01 2.90804207e-01 -3.65770340e-01
-1.11720788e+00 -9.59638119e-01 -1.90272823e-01 1.15004313e+00
5.16923845e-01 -3.55875701e-01 -4.62840647e-01 -7.54273161e-02
1.00443035e-01 6.04229450e-01 1.75936371e-01 -1.39893353e-01
-4.99159098e-01 -1.44408140e-02 1.59038231e-01 5.81646204e-01
2.93989182e-01 -8.62761676e-01 -8.78424585e-01 1.97210625e-01
1.28476858e-01 -1.51960433e+00 -2.76940763e-01 3.52718294e-01
-5.93820393e-01 -4.62741256e-01 -1.61378235e-02 -1.38080430e+00
3.58248204e-01 7.80568063e-01 1.04662097e+00 3.30682874e-01
-4.26120728e-01 5.27103603e-01 -1.90456249e-02 -6.79163516e-01
-6.02906227e-01 1.82175070e-01 4.74894047e-03 -7.09186375e-01
7.76939690e-01 -6.95265055e-01 -3.62181097e-01 5.13028204e-01
-6.58488512e-01 -1.06656730e-01 6.65066600e-01 1.53881520e-01
-1.92453742e-01 3.38278949e-01 8.39252889e-01 -2.10058197e-01
4.16782618e-01 -5.13298452e-01 -8.67450237e-01 -2.35534068e-02
-5.66433072e-01 -3.47961724e-01 4.25483078e-01 3.04918438e-02
-6.01090729e-01 1.35625526e-01 -1.50548846e-01 1.98270261e-01
-4.87892210e-01 4.94420052e-01 -2.76583970e-01 -2.49020427e-01
4.93271708e-01 2.00755119e-01 5.25763214e-01 3.25992644e-01
-8.98183063e-02 8.18407178e-01 6.60451531e-01 -6.72630608e-01
1.04474604e+00 2.27784649e-01 4.68519107e-02 -9.38623428e-01
-2.61126429e-01 -2.14434996e-01 -4.25265849e-01 -6.52204037e-01
7.96260059e-01 -9.60473239e-01 -1.00440979e+00 1.85656786e-01
-9.70447719e-01 -2.97386825e-01 -5.93691655e-02 7.60412753e-01
-2.67401814e-01 1.13808431e-01 -3.45236719e-01 -5.18638551e-01
3.98930699e-01 -1.65978801e+00 6.94476962e-01 5.48469126e-01
-4.22311395e-01 -8.23098242e-01 9.08146873e-02 5.92336714e-01
5.70046604e-01 1.34831041e-01 5.03885567e-01 1.63280845e-01
-1.10351038e+00 -6.57122612e-01 4.40000854e-02 2.66769052e-01
2.16947570e-02 4.92842883e-01 -6.64893568e-01 -3.03288728e-01
-1.63714036e-01 -2.35554963e-01 2.91542649e-01 1.19855255e-01
5.62696695e-01 5.38310587e-01 -8.50693524e-01 -1.21776402e-01
1.43974173e+00 6.52308643e-01 7.10688531e-01 4.08503920e-01
2.57626235e-01 7.02063859e-01 6.56528175e-01 7.58275986e-02
7.91807413e-01 4.56985772e-01 2.54498571e-01 -6.69663819e-03
2.23543569e-01 -1.76929921e-01 5.47149658e-01 8.67406189e-01
8.84070173e-02 2.50493586e-01 -1.15929520e+00 8.18849087e-01
-1.61684513e+00 -9.64027524e-01 -5.72385848e-01 2.14811158e+00
2.86095858e-01 5.96364439e-01 2.99405128e-01 3.08340132e-01
6.00276411e-01 -3.73369664e-01 -2.77633727e-01 -7.80925989e-01
6.77356660e-01 3.11502852e-02 5.49420118e-01 4.49936330e-01
-7.39704013e-01 7.08035469e-01 7.46063709e+00 3.85558426e-01
-1.22466147e+00 -2.24804386e-01 -6.71528056e-02 3.26314181e-01
-5.87158799e-01 4.34530646e-01 -8.93172741e-01 4.33290303e-01
1.22874701e+00 -6.57032788e-01 2.09127367e-01 1.07902658e+00
2.77406156e-01 -1.68771133e-01 -7.03824162e-01 4.60183471e-01
-4.81765509e-01 -1.19306421e+00 -7.24387944e-01 6.64007843e-01
7.99367011e-01 1.01810684e-02 1.76443189e-01 3.68831426e-01
6.82864130e-01 -1.06670904e+00 5.66712976e-01 6.22537285e-02
9.53958631e-02 -9.24359202e-01 6.90573633e-01 4.31635946e-01
-1.23308825e+00 -7.98734426e-02 -3.57179791e-02 -4.52528507e-01
2.05761403e-01 3.28266174e-01 -3.07182074e-01 4.80054498e-01
8.59267592e-01 3.05716604e-01 -7.14411661e-02 1.02185726e+00
1.23816393e-01 3.08392733e-01 -1.62942961e-01 -2.95518756e-01
7.78974891e-02 -3.06067884e-01 2.08018303e-01 8.17681134e-01
5.20080864e-01 3.36506993e-01 2.91914999e-01 6.21393681e-01
4.42958206e-01 -3.52390558e-01 -9.12923515e-01 -7.27274418e-02
7.64193177e-01 1.63831913e+00 -7.59016395e-01 -5.57815768e-02
-7.64962733e-01 2.09997848e-01 -5.25520265e-01 2.06507072e-01
-9.79675591e-01 -1.07574868e+00 9.43834007e-01 4.25714016e-01
5.07046938e-01 -7.55175471e-01 -7.31003046e-01 -3.48411828e-01
7.89953396e-02 -8.84556532e-01 -4.14298773e-01 -6.61746979e-01
-8.05830121e-01 4.96977568e-01 -1.52445853e-01 -1.36337733e+00
-1.81341320e-01 -6.37721419e-01 -7.86498427e-01 8.99316907e-01
-1.37297177e+00 -8.03595722e-01 -8.08085442e-01 1.21785691e-02
9.11939442e-02 -3.08954388e-01 4.88304287e-01 5.64205527e-01
-5.28163373e-01 2.51591921e-01 -1.11818891e-02 -2.87492663e-01
4.80060756e-01 -8.47397208e-01 4.89518046e-01 7.74986327e-01
-6.26638412e-01 8.51386368e-01 1.10910225e+00 -2.48485774e-01
-2.16127419e+00 -8.09350431e-01 1.03395891e+00 -6.72298789e-01
9.53000784e-01 -1.15043603e-01 -8.18799317e-01 7.63831854e-01
6.90096378e-01 -4.12724689e-02 5.21788120e-01 -8.76349509e-02
3.61842364e-02 -4.39429611e-01 -9.44126427e-01 4.12876636e-01
6.55637681e-01 -2.94090509e-01 -4.48256075e-01 1.52371347e-01
6.20458305e-01 -4.46359992e-01 -9.52727914e-01 3.72598708e-01
5.81519365e-01 -7.24732876e-01 5.57310283e-01 -7.52337426e-02
2.92516083e-01 -7.22444534e-01 -3.13711390e-02 -9.65618074e-01
-6.91017956e-02 -8.15354109e-01 4.81331825e-01 9.73789334e-01
3.54759395e-01 -8.79384577e-01 1.13321269e+00 6.21530473e-01
-7.21897781e-01 -5.19468725e-01 -4.00192231e-01 -1.13281560e+00
5.49564242e-01 -7.08881438e-01 4.39790547e-01 5.96060216e-01
6.27239704e-01 3.88585180e-01 1.03959620e-01 1.78584695e-01
4.38915581e-01 -1.71919651e-02 1.28144825e+00 -1.56314206e+00
-1.12177767e-02 -5.72438359e-01 -7.39923596e-01 -1.00989103e+00
1.20250508e-01 -4.23849255e-01 3.52180809e-01 -1.81437433e+00
-2.47260839e-01 -3.51696730e-01 1.16582289e-01 1.43575743e-01
3.23096961e-01 1.80750057e-01 -3.12649719e-02 -2.09451452e-01
-2.91271001e-01 1.88171305e-02 1.21923172e+00 2.59877920e-01
-1.43086627e-01 6.02054112e-02 -1.21732056e+00 6.91559911e-01
7.55843282e-01 -1.82438478e-01 -7.85925806e-01 -2.19865367e-01
1.73726991e-01 4.21622396e-03 4.08416212e-01 -1.42208481e+00
6.22141719e-01 -6.52973056e-01 -2.34241873e-01 -6.63571477e-01
4.64760929e-01 -1.15161872e+00 2.33709484e-01 6.74068272e-01
3.67265642e-01 2.21831501e-01 5.36099553e-01 -5.07191494e-02
-3.92502397e-01 -1.60531342e-01 7.24676430e-01 -1.64089680e-01
-8.66063178e-01 -1.11992553e-01 -9.14346039e-01 -3.27326171e-02
1.81325638e+00 -5.52218080e-01 -6.30253375e-01 -4.09554958e-01
-5.97854495e-01 5.34079671e-01 6.95006251e-01 5.15416443e-01
3.55153054e-01 -1.18563974e+00 -4.35545504e-01 1.12261586e-01
1.00132555e-01 -4.39361930e-01 2.76755154e-01 8.09296012e-01
-7.89274752e-01 7.72364140e-01 -5.83755732e-01 -6.05531931e-01
-1.03113592e+00 3.69563729e-01 7.14596063e-02 2.35573739e-01
-4.98966187e-01 5.05607009e-01 -3.04388821e-01 -7.06917703e-01
2.26479068e-01 1.51683286e-01 2.47363538e-01 -3.62329572e-01
4.67840046e-01 4.31212008e-01 2.06073627e-01 -3.03760737e-01
-4.31752801e-01 7.21136987e-01 2.14238137e-01 -3.97728682e-01
1.20211029e+00 -2.24755660e-01 9.61120054e-02 5.61778724e-01
1.06906104e+00 -1.27512366e-01 -7.22662151e-01 3.54439467e-01
-1.08775450e-02 -2.89155513e-01 1.86897427e-01 -4.23171937e-01
-8.76687765e-01 9.56988573e-01 5.40683031e-01 5.68396151e-01
7.22231030e-01 -2.37211153e-01 1.05559289e+00 1.22797437e-01
7.58110344e-01 -1.36751068e+00 5.18561415e-02 6.53744459e-01
5.39520025e-01 -9.82731581e-01 -1.92540493e-02 -5.62468171e-01
-5.78060091e-01 1.09307170e+00 1.00498855e+00 -1.08668059e-01
1.03430057e+00 6.57996535e-01 2.35936061e-01 -2.05778226e-01
-8.62813234e-01 -1.43266156e-01 -2.65210778e-01 6.08226597e-01
7.45585144e-01 -1.07820123e-01 -4.65839535e-01 3.46014053e-02
-4.97430831e-01 4.95967120e-01 1.11014068e+00 1.45232892e+00
-9.87125933e-01 -1.35032463e+00 -5.17233610e-01 1.07583307e-01
-1.26434550e-01 4.26665366e-01 -1.78329021e-01 9.88573909e-01
3.31293046e-01 1.18068326e+00 1.80588722e-01 -7.41526067e-01
4.34693933e-01 1.27825305e-01 2.60217935e-01 -4.72267866e-01
-3.32893372e-01 -4.80843872e-01 -7.03745708e-02 -8.15797374e-02
1.85436398e-01 -6.65559173e-01 -1.55038500e+00 -1.25061524e+00
-1.91362519e-02 4.90116745e-01 1.32388186e+00 7.56529689e-01
7.16366291e-01 7.27678180e-01 5.80915272e-01 -5.74361503e-01
-2.06304193e-01 -4.93397325e-01 -5.86988330e-01 -3.94413024e-01
1.66090786e-01 -6.85348868e-01 -1.45532027e-01 -1.63908854e-01] | [5.613361835479736, 1.0261492729187012] |
a2b2a3ca-70a5-4104-a2bf-d8ff9c1b5349 | proceedings-38th-international-conference-on | 2208.02685 | null | https://arxiv.org/abs/2208.02685v1 | https://arxiv.org/pdf/2208.02685v1.pdf | Proceedings 38th International Conference on Logic Programming | ICLP is the premier international event for presenting research in logic programming. Contributions to ICLP 2022 were sought in all areas of logic programming, including but not limited to: Foundations: Semantics, Formalisms, Nonmonotonic reasoning, Knowledge representation. Languages issues: Concurrency, Objects, Coordination, Mobility, Higher order, Types, Modes, Assertions, Modules, Meta-programming, Logic-based domain-specific languages, Programming techniques. Programming support: Program analysis, Transformation, Validation, Verification, Debugging, Profiling, Testing, Execution visualization. Implementation: Compilation, Virtual machines, Memory management, Parallel and Distributed execution, Constraint handling rules, Tabling, Foreign interfaces, User interfaces. Related Paradigms and Synergies: Inductive and coinductive logic programming, Constraint logic programming, Answer set programming, Interaction with SAT, SMT and CSP solvers, Theorem proving, Argumentation, Probabilistic programming, Machine learning. Applications: Databases, Big data, Data integration and federation, Software engineering, Natural language processing, Web and semantic web, Agents, Artificial intelligence, Computational life sciences, Cyber-security, Robotics, Education. | ['Tuncay Tekle', 'Martin Gebser', 'Veronica Dahl', 'Carmine Dodaro', 'Jose F. Morales', 'Yuliya Lierler'] | 2022-08-04 | null | null | null | null | ['data-integration', 'probabilistic-programming', 'automated-theorem-proving', 'automated-theorem-proving'] | ['knowledge-base', 'methodology', 'miscellaneous', 'reasoning'] | [-5.09687901e-01 3.78067344e-01 -2.08871275e-01 -4.69289422e-01
3.39967608e-01 -1.02106881e+00 5.19024670e-01 9.69981432e-01
2.09362566e-01 1.22281528e+00 -8.01486745e-02 -5.64293385e-01
-8.79004359e-01 -1.08875477e+00 -2.63309747e-01 -8.89859628e-03
-4.85328078e-01 1.27414417e+00 8.10610175e-01 -3.25239182e-01
4.36958700e-01 5.96846938e-01 -2.11756158e+00 4.39827442e-01
6.28767490e-01 1.05897868e+00 -6.95747197e-01 4.66312975e-01
-4.06036675e-01 1.45486331e+00 -8.49068537e-02 -2.55673409e-01
-1.31639481e-01 1.81538537e-01 -1.22825444e+00 -5.55426896e-01
-3.78266126e-01 2.91006535e-01 4.01514381e-01 1.43995452e+00
-1.11921817e-01 -1.23761714e-01 3.88640821e-01 -2.24724293e+00
-6.59135401e-01 9.81754482e-01 -3.91743213e-01 -9.60120633e-02
1.36868918e+00 -2.29231939e-01 7.10631967e-01 -3.94803792e-01
1.07069850e+00 1.74948800e+00 5.66555262e-01 3.33738059e-01
-9.84014690e-01 -4.48523521e-01 -8.80401507e-02 7.44298279e-01
-1.30915916e+00 2.97682315e-01 1.52089715e-01 -5.16100228e-01
1.42878461e+00 6.73716843e-01 5.02011478e-01 1.37947068e-01
2.00579539e-01 4.28258091e-01 1.24528360e+00 -8.16097379e-01
5.71709752e-01 1.27566493e+00 9.38562155e-01 8.34869564e-01
4.84079570e-01 1.12458728e-01 -5.79683185e-01 -6.16693795e-01
1.75470561e-01 -3.95515919e-01 2.24872917e-01 -4.18712705e-01
-9.61900592e-01 7.93013752e-01 -6.18101060e-01 4.87796605e-01
-3.34348649e-01 -2.00998262e-01 9.07073975e-01 6.76663756e-01
-5.79807580e-01 -1.19631728e-02 -1.11959350e+00 -1.57561362e-01
-2.75146306e-01 1.00385141e+00 1.67720044e+00 1.65163732e+00
6.51114106e-01 -2.83389658e-01 1.78836972e-01 1.66659191e-01
1.05910802e+00 4.20248866e-01 4.32612330e-01 -1.05242717e+00
1.21470056e-01 1.10602057e+00 2.53272355e-01 -8.80989671e-01
-6.15353525e-01 9.11666691e-01 -1.61824703e-01 2.29316622e-01
-2.62697190e-01 -1.55195624e-01 -4.37900513e-01 1.38394785e+00
9.17034268e-01 -4.29040641e-01 5.14648199e-01 5.46646655e-01
9.05650020e-01 4.99696523e-01 1.93037093e-01 -4.36286479e-01
1.68225253e+00 -4.05657887e-01 -7.68963993e-01 4.83500659e-01
2.48954639e-01 -6.50817573e-01 1.06990457e-01 1.03962076e+00
-1.41331792e+00 2.10702568e-01 -6.51085794e-01 2.78545171e-01
-1.40089500e+00 -4.63332027e-01 1.10176146e+00 6.49075508e-01
-9.07056570e-01 -6.21548891e-02 -6.74136460e-01 -5.55934966e-01
2.61349857e-01 6.46338642e-01 -4.55072552e-01 1.57461576e-02
-1.34803283e+00 1.36651587e+00 1.57392585e+00 -5.23179293e-01
-5.34221351e-01 -6.98634148e-01 -9.30697858e-01 -2.87371464e-02
4.73125666e-01 -1.30881533e-01 8.43658864e-01 -6.36359811e-01
-1.50987208e+00 9.31414068e-01 4.86157686e-02 -9.06298816e-01
2.33189762e-01 6.92541063e-01 -8.72556806e-01 -2.30096787e-01
2.36162677e-01 2.28080407e-01 -4.40824986e-01 -8.92223835e-01
-1.05704784e+00 -6.03966117e-01 3.46421659e-01 -4.45110172e-01
6.99132919e-01 7.06378222e-01 4.09882277e-01 3.75378430e-01
-3.00196446e-02 -5.22354901e-01 -1.16614796e-01 -4.85094219e-01
-3.03907663e-01 -7.18839526e-01 1.08454275e+00 -1.09118767e-01
8.18225265e-01 -1.85803747e+00 2.17881680e-01 1.03854752e+00
-2.40273893e-01 -1.48207486e-01 9.32474136e-01 1.01669300e+00
1.02274649e-01 3.05617392e-01 -1.02662466e-01 1.19910109e+00
8.62357199e-01 2.30077490e-01 -8.42289403e-02 9.97778028e-02
-1.08172901e-01 3.21298331e-01 -7.33594120e-01 -1.03156769e+00
7.11592138e-01 -1.99336886e-01 -6.26362324e-01 -3.13644886e-01
-1.15037775e+00 -2.81257659e-01 -6.49415374e-01 1.43591797e+00
7.79767931e-01 6.46319427e-03 7.85963476e-01 1.34417653e-01
-6.93696439e-01 -3.18635941e-01 -1.93172932e+00 1.40156984e+00
-5.80063879e-01 5.80091238e-01 1.64721847e-01 -9.96830881e-01
7.95894444e-01 8.75413895e-01 3.21935385e-01 -4.00590867e-01
-5.79159986e-03 5.86720705e-01 -2.22168729e-01 -1.15159130e+00
6.73437640e-02 3.49318564e-01 -9.38330814e-02 5.24332464e-01
3.15290242e-01 -2.48276845e-01 9.17216122e-01 1.06655367e-01
6.98020756e-01 4.47701633e-01 8.33142877e-01 -4.63718504e-01
1.05555642e+00 6.35055125e-01 5.31664848e-01 2.09422797e-01
-6.74179494e-02 -4.95388359e-01 9.57288384e-01 -7.43611276e-01
-6.99468374e-01 -1.05208588e+00 -4.52677637e-01 1.02691472e+00
2.30040938e-01 -5.03125429e-01 -4.12860274e-01 -2.38420084e-01
2.82000065e-01 8.87316227e-01 -1.87810659e-02 -8.97367205e-03
-2.76681423e-01 -4.38991427e-01 6.04147792e-01 4.71311994e-02
3.84582579e-01 -1.37763286e+00 -1.22069669e+00 4.47858810e-01
4.57514733e-01 -9.09307659e-01 9.45874631e-01 1.67402864e-01
-5.56860566e-01 -1.52383018e+00 8.59918952e-01 -6.46174967e-01
3.16836655e-01 -8.57880175e-01 1.27771616e+00 1.29593596e-01
-5.86277068e-01 2.95568347e-01 -1.41660646e-01 -1.12804878e+00
-4.00433719e-01 -9.10357535e-01 5.77921830e-02 -8.53132725e-01
8.92579854e-01 -5.79485059e-01 3.19902897e-01 1.91856146e-01
-8.69945049e-01 -3.13863605e-01 1.63295984e-01 7.47755229e-01
4.17914957e-01 6.17353797e-01 3.47486973e-01 -1.24026942e+00
9.44712400e-01 -6.05139434e-01 -1.33430266e+00 8.54827106e-01
-8.21663678e-01 -2.48808730e-02 6.22945189e-01 -1.90471694e-01
-9.63093460e-01 -3.84801865e-01 4.09910530e-01 7.98003469e-03
-5.68879724e-01 7.79621065e-01 -6.68416500e-01 6.50077537e-02
6.77587390e-01 1.51333913e-01 2.18508169e-02 1.32324234e-01
4.66285795e-01 5.81107795e-01 2.24478394e-01 -1.44810474e+00
4.06856805e-01 3.09773475e-01 2.47662187e-01 -3.53809237e-01
-1.53276131e-01 -2.29379386e-01 -3.06200176e-01 6.28384426e-02
6.76598132e-01 -4.04517591e-01 -1.63144040e+00 -1.82685345e-01
-1.36944306e+00 2.62414843e-01 -4.11187410e-01 4.87259448e-01
-5.87163866e-01 -3.17238450e-01 -2.96026677e-01 -1.18439674e+00
-4.09393102e-01 -8.83753002e-01 3.20362836e-01 5.36939085e-01
-2.57265210e-01 -1.19426763e+00 1.65444002e-01 1.23349607e-01
2.84966975e-01 8.22740614e-01 1.36852837e+00 -1.32846379e+00
-6.35593951e-01 -3.27225149e-01 -9.29949135e-02 9.87158641e-02
-2.81483591e-01 8.79280686e-01 -4.17857617e-01 5.80672443e-01
-6.95896387e-01 -6.84223950e-01 -4.54696238e-01 -2.50895936e-02
8.06194901e-01 -8.57277751e-01 -4.34348583e-01 -2.08060160e-01
1.69141829e+00 6.04975939e-01 6.13659203e-01 5.05534172e-01
-2.80339062e-01 7.74139822e-01 8.72333586e-01 6.44493699e-01
6.50164962e-01 3.05625439e-01 3.95015836e-01 1.10889006e+00
5.24776757e-01 2.81597853e-01 -7.58778006e-02 1.04623437e-01
-5.27701020e-01 5.20811200e-01 -1.51249933e+00 2.38037378e-01
-2.36569047e+00 -1.37022090e+00 -4.80023086e-01 2.28861070e+00
9.81199920e-01 3.89706552e-01 1.34109452e-01 5.90035081e-01
5.82642555e-01 -5.85690618e-01 9.05425102e-02 -1.06176555e+00
2.87565924e-02 6.00258335e-02 3.15660000e-01 7.96306133e-01
-7.83919394e-01 1.05425715e+00 5.12501764e+00 6.60739958e-01
-9.61799324e-01 5.24386406e-01 -2.73423344e-01 5.53700030e-01
-6.03316605e-01 3.30003202e-01 -8.87210846e-01 3.92122120e-01
1.15673494e+00 -8.39008510e-01 1.04941344e+00 1.20093596e+00
-8.61736313e-02 -3.37463289e-01 -1.23373175e+00 4.35835093e-01
-3.66260149e-02 -1.79775715e+00 -1.49486765e-01 -3.02226424e-01
8.11167955e-01 -3.56285907e-02 -7.62576878e-01 4.75341111e-01
6.57988846e-01 -8.56354415e-01 1.20723081e+00 5.99609375e-01
-2.85083234e-01 -1.13194895e+00 1.10510182e+00 6.41260818e-02
-6.79154098e-01 -2.89252251e-01 -8.21062103e-02 -7.18269348e-02
3.92056294e-02 5.03245533e-01 -6.50341213e-01 1.18431246e+00
8.22826147e-01 4.34657708e-02 -4.06918883e-01 9.25048292e-01
1.02738820e-01 -1.01432867e-01 -7.58914530e-01 -6.66575611e-01
-1.96883917e-01 -2.65550464e-01 3.44779253e-01 1.16901362e+00
-3.73636246e-01 1.61992818e-01 4.01924074e-01 1.12406850e+00
6.09707654e-01 -3.76472995e-02 -5.86310625e-01 8.67889524e-02
9.18264508e-01 1.04824269e+00 -7.42538810e-01 -4.65768039e-01
-3.18280518e-01 -6.16117835e-01 -4.07097936e-01 3.42827439e-01
-7.43903399e-01 -9.35209692e-01 2.92008996e-01 -1.32822916e-01
-4.73996967e-01 8.68527442e-02 -4.19761479e-01 -7.37683237e-01
1.19284820e-02 -1.31066489e+00 8.38713408e-01 -5.23560762e-01
-9.32812750e-01 2.43287191e-01 8.37740362e-01 -5.29347360e-01
-6.31657958e-01 -9.60990489e-01 -5.15839279e-01 4.59326029e-01
-1.02306354e+00 -1.39939845e+00 -1.18341424e-01 8.89576316e-01
-2.42448121e-01 -6.56379104e-01 1.19279540e+00 5.85033238e-01
-8.59121159e-02 -2.73134917e-01 -4.39509988e-01 -3.84649634e-01
3.65302935e-02 -1.42039442e+00 -8.98338258e-01 3.96566361e-01
-2.24850029e-01 4.96461302e-01 8.40862095e-01 -3.67658049e-01
-1.95620799e+00 -6.07285440e-01 9.28002715e-01 -1.38318434e-01
1.11175931e+00 -2.78167218e-01 -4.08447295e-01 1.04362190e+00
8.74819309e-02 4.13440615e-01 5.14260173e-01 2.37512842e-01
-3.36391598e-01 -3.47571701e-01 -2.05023360e+00 2.43731052e-01
4.29848880e-02 -4.98233765e-01 -7.44043648e-01 1.04590547e+00
6.48634315e-01 -6.61830604e-01 -1.28262269e+00 2.60123778e-02
6.86513066e-01 -1.05553865e+00 7.72161126e-01 -9.50611413e-01
1.31018341e-01 -7.24192858e-01 -5.50375640e-01 -1.00384548e-01
4.08800572e-01 -5.07826030e-01 -3.53893936e-01 1.48880649e+00
8.46367240e-01 -1.05633724e+00 4.30966705e-01 1.24224603e+00
3.53369047e-03 -3.28245074e-01 -1.03775132e+00 -5.40797949e-01
-3.72437537e-01 -6.08198762e-01 1.07024980e+00 1.26259005e+00
1.09823155e+00 -2.13438243e-01 4.53368038e-01 4.82558995e-01
4.78122354e-01 4.19034868e-01 5.24070859e-01 -1.60004699e+00
-2.98274219e-01 -5.50727010e-01 -1.28716934e+00 6.18394256e-01
3.35638553e-01 -6.61395252e-01 -4.37788904e-01 -1.89711070e+00
-3.93193990e-01 -6.12117946e-01 1.46390229e-01 7.17622995e-01
9.27982688e-01 -5.49456000e-01 -2.02023890e-02 -4.15717006e-01
-1.02375138e+00 -4.57384020e-01 7.26333737e-01 -1.89186886e-01
-4.18915302e-02 -1.61576886e-02 -1.17890872e-01 8.79845440e-01
5.09534597e-01 -5.30387580e-01 -3.80758792e-01 1.48403555e-01
1.02459526e+00 3.11136156e-01 3.11292559e-01 -7.24982619e-01
8.57399046e-01 -1.09589040e+00 -2.35694572e-01 -4.17540491e-01
-1.72984883e-01 -1.44731927e+00 6.55841231e-01 7.22774565e-01
-2.22812183e-02 1.06238741e-02 2.67805487e-01 -7.43044028e-03
-5.48722088e-01 -1.06518197e+00 5.00865698e-01 -4.69912261e-01
-7.03845263e-01 -6.33941367e-02 -5.19009590e-01 2.65598267e-01
2.00544477e+00 1.16286203e-01 -5.03496230e-01 3.80363971e-01
-1.07158196e+00 6.35764539e-01 6.73657730e-02 1.01461798e-01
1.66334599e-01 -1.09659958e+00 -2.10542202e-01 -1.74530834e-01
6.68569133e-02 4.97068793e-01 -2.48035014e-01 8.57284963e-01
-1.14531434e+00 7.97933996e-01 -5.50957382e-01 -1.79463610e-01
-1.11324441e+00 7.21604466e-01 2.86021888e-01 -2.34794050e-01
8.02987069e-02 5.46464860e-01 -9.46937382e-01 -1.04959106e+00
5.02900124e-01 -5.14439046e-01 -2.98665166e-01 -2.53716111e-01
5.51690578e-01 5.90793490e-01 5.07777594e-02 1.03472751e-02
-1.01633298e+00 3.77168119e-01 3.59233290e-01 -2.92363346e-01
1.22763097e+00 4.75803107e-01 -1.11414194e+00 8.46224129e-01
9.98431817e-02 -1.12116277e-01 2.68312335e-01 2.78181881e-01
1.03145492e+00 -1.38663456e-01 -3.91548753e-01 -1.47099352e+00
-2.38088235e-01 2.95893103e-01 3.34235013e-01 1.12691891e+00
6.04005337e-01 -2.29427591e-02 -4.19976979e-01 7.03165531e-01
1.06014431e+00 -1.59469569e+00 -8.09692264e-01 7.54077971e-01
9.88447666e-01 -1.03682292e+00 2.87699550e-01 -3.52322161e-01
-5.66942692e-01 1.51826990e+00 6.17981255e-01 9.04975533e-02
1.12429512e+00 1.00340188e+00 -3.29244494e-01 -6.68500006e-01
-1.22027087e+00 1.38020247e-01 -3.78278315e-01 1.16576195e+00
5.42201161e-01 2.98636705e-01 -6.44999504e-01 8.40519667e-01
-3.75507057e-01 7.10450470e-01 7.18826890e-01 1.43030035e+00
-2.15198368e-01 -1.24848402e+00 -8.92014563e-01 2.56936580e-01
-6.19665563e-01 1.25821412e-01 -2.34168127e-01 1.25609553e+00
4.97745454e-01 9.48775291e-01 -1.64864346e-01 2.01568469e-01
5.74274123e-01 8.47718045e-02 7.30456531e-01 -5.64990819e-01
-7.25245297e-01 -8.22347581e-01 6.25781238e-01 -2.51532942e-01
-4.92983967e-01 -5.98927319e-01 -1.88795412e+00 -4.33253020e-01
-3.64885122e-01 9.31357503e-01 1.30764127e+00 1.06776428e+00
6.02109693e-02 1.84274957e-01 -1.28934875e-01 -1.22103997e-01
-4.02726561e-01 -8.01753581e-01 -5.64942479e-01 -3.87972564e-01
-2.39624396e-01 -6.65728867e-01 -1.49191380e-01 2.01511160e-01] | [8.670280456542969, 6.73847770690918] |
d1377fd0-880a-47af-b24c-059e0a7c46da | video-transformer-for-deepfake-detection-with | 2108.05307 | null | https://arxiv.org/abs/2108.05307v1 | https://arxiv.org/pdf/2108.05307v1.pdf | Video Transformer for Deepfake Detection with Incremental Learning | Face forgery by deepfake is widely spread over the internet and this raises severe societal concerns. In this paper, we propose a novel video transformer with incremental learning for detecting deepfake videos. To better align the input face images, we use a 3D face reconstruction method to generate UV texture from a single input face image. The aligned face image can also provide pose, eyes blink and mouth movement information that cannot be perceived in the UV texture image, so we use both face images and their UV texture maps to extract the image features. We present an incremental learning strategy to fine-tune the proposed model on a smaller amount of data and achieve better deepfake detection performance. The comprehensive experiments on various public deepfake datasets demonstrate that the proposed video transformer model with incremental learning achieves state-of-the-art performance in the deepfake video detection task with enhanced feature learning from the sequenced data. | ['Hang Dai', 'Sohail A. Khan'] | 2021-08-11 | null | null | null | null | ['3d-face-reconstruction', 'face-reconstruction'] | ['computer-vision', 'computer-vision'] | [ 4.08769101e-02 -3.84192437e-01 -3.17905933e-01 -3.02887499e-01
-6.29247129e-01 -5.74578881e-01 3.34228337e-01 -9.62875128e-01
3.55356447e-02 2.56803215e-01 1.80733442e-01 -2.54721958e-02
6.31710812e-02 -6.10560119e-01 -7.78297842e-01 -8.81758511e-01
1.43273130e-01 -1.97238833e-01 2.94265356e-02 8.19839835e-02
2.83773124e-01 5.40444314e-01 -1.71892297e+00 7.65774310e-01
3.76285315e-01 1.46556830e+00 1.17950708e-01 5.72411954e-01
1.25575900e-01 7.48078883e-01 -6.04100406e-01 -7.33433783e-01
5.13303399e-01 -3.21765393e-01 -5.12863159e-01 4.91924644e-01
1.02661550e+00 -1.26772141e+00 -8.00197244e-01 1.02976131e+00
6.22932017e-01 -3.94794762e-01 4.67327118e-01 -1.49321592e+00
-5.15567064e-01 -5.96475378e-02 -9.57526267e-01 3.18572521e-01
3.63877863e-01 2.22287357e-01 2.49649867e-01 -1.40275228e+00
4.17648435e-01 1.65166414e+00 8.53340745e-01 6.98180795e-01
-6.47203147e-01 -1.29445112e+00 -6.99188560e-02 7.87237346e-01
-1.54335892e+00 -9.89832878e-01 8.99699628e-01 -2.11887956e-01
7.42864847e-01 1.93902701e-02 6.51749492e-01 1.29917109e+00
1.19110243e-02 6.75861716e-01 9.05564427e-01 -2.93969721e-01
-2.08998278e-01 2.73648906e-03 -3.12034607e-01 1.10900283e+00
1.67962223e-01 3.72767717e-01 -8.17201912e-01 -1.64476201e-01
7.00081229e-01 1.24550059e-01 -3.09257865e-01 -5.49693741e-02
-5.12197256e-01 7.45622516e-01 1.46305248e-01 -1.52929202e-01
7.20115080e-02 1.00239784e-01 3.48626465e-01 5.07017136e-01
2.74560452e-01 -4.24488425e-01 -2.54537404e-01 6.32180721e-02
-7.77512074e-01 -6.43932745e-02 3.58039618e-01 7.39383221e-01
8.69527996e-01 6.69089779e-02 -1.53591692e-01 9.41798449e-01
4.22692090e-01 8.22448373e-01 2.98641503e-01 -8.92726123e-01
5.99548519e-01 3.70408207e-01 -1.17759004e-01 -1.40980303e+00
-4.83766533e-02 1.28083766e-01 -6.80202186e-01 2.33744085e-01
4.73890811e-01 -5.72945178e-02 -8.56235027e-01 1.25420022e+00
5.34725547e-01 6.83332384e-01 -2.47509658e-01 8.00559282e-01
9.49620724e-01 6.45820200e-01 -3.99535388e-01 -3.12546611e-01
1.34634614e+00 -8.13035548e-01 -7.50465751e-01 -2.00365841e-01
4.82108444e-01 -8.96746755e-01 6.43211126e-01 7.04459667e-01
-5.72190821e-01 -5.81843317e-01 -1.05543780e+00 -5.21206222e-02
4.20041233e-02 7.02618182e-01 3.28232169e-01 1.35633910e+00
-9.76426482e-01 3.44519734e-01 -4.72154886e-01 -7.37231746e-02
1.06451929e+00 4.79101121e-01 -5.82278550e-01 -4.70699936e-01
-8.29666138e-01 5.41822970e-01 -3.34376171e-02 3.70500952e-01
-1.43110728e+00 -3.37873042e-01 -7.79032886e-01 -1.22505976e-02
6.26423955e-01 -1.32144332e-01 1.15379262e+00 -1.12087870e+00
-1.51603186e+00 9.79099751e-01 -4.88915652e-01 2.76234131e-02
6.24981225e-01 -8.51079971e-02 -6.51067078e-01 3.56747687e-01
-2.77964354e-01 4.05448496e-01 1.84233701e+00 -9.45765436e-01
-5.07033944e-01 -7.01089025e-01 -2.62058079e-01 -1.72114834e-01
-7.78713286e-01 3.04649770e-01 -5.60465097e-01 -6.34155512e-01
-1.92040354e-01 -7.70842671e-01 6.73960268e-01 4.65330303e-01
-1.77293271e-01 -4.06258494e-01 1.80211604e+00 -9.81864870e-01
1.12825346e+00 -2.04808974e+00 -1.47732198e-01 6.74889684e-02
3.59635562e-01 6.76959813e-01 -4.22632843e-01 -1.21341934e-02
-1.41028613e-01 -1.06212862e-01 1.88678697e-01 -4.48130131e-01
-4.72701073e-01 -7.28939772e-02 -2.51886517e-01 7.78048098e-01
1.71521962e-01 7.40841627e-01 -7.84129977e-01 -2.97613770e-01
2.68692106e-01 6.01053238e-01 -6.80819631e-01 2.48825684e-01
2.29617581e-01 8.65500271e-02 -2.19398662e-01 1.10039055e+00
1.51178730e+00 1.44533422e-02 2.79398918e-01 -5.06855071e-01
2.51393706e-01 -1.89585313e-01 -8.74317110e-01 1.35595453e+00
-3.09528857e-01 8.65504861e-01 1.77060336e-01 -8.26992393e-01
6.97846770e-01 2.12045193e-01 3.47269356e-01 -8.24984312e-01
3.66755486e-01 6.85352609e-02 -2.82322764e-01 -9.64397609e-01
8.28457326e-02 1.65798008e-01 1.87669113e-01 6.53605521e-01
2.66044587e-01 3.70323002e-01 -3.45046967e-01 -4.26812060e-02
7.35314667e-01 -3.99801619e-02 -1.30041242e-01 -1.45562794e-02
8.22877884e-01 -5.72825551e-01 5.99806190e-01 4.68331009e-01
-4.09866422e-01 3.40027303e-01 5.40593803e-01 -7.07875252e-01
-9.30360377e-01 -9.31710780e-01 -4.05952595e-02 1.03756320e+00
8.81534293e-02 -4.31297868e-01 -1.10055625e+00 -1.32038951e+00
1.39614180e-01 -5.30234054e-02 -6.18367612e-01 -3.69234920e-01
-5.21866381e-01 -7.01254427e-01 7.97019601e-01 2.18998373e-01
9.77067590e-01 -7.31611371e-01 -1.17449246e-01 -2.24018678e-01
-3.87904793e-01 -1.18357325e+00 -9.07413125e-01 -7.51750588e-01
-4.10331488e-01 -1.45884824e+00 -5.99288642e-01 -1.09169209e+00
6.91748679e-01 8.05655062e-01 5.33858836e-01 4.20722783e-01
-7.53944457e-01 2.75380582e-01 -2.65833884e-01 2.17881221e-02
-4.86535430e-01 -5.31586766e-01 3.31331402e-01 5.82346737e-01
5.74605346e-01 -4.56691757e-02 -8.29834819e-01 6.85437024e-01
-7.03194797e-01 -3.15063924e-01 3.59013468e-01 1.01502168e+00
-3.17361057e-02 3.24970514e-01 6.91794932e-01 -3.21940482e-01
2.86658555e-01 -2.63878524e-01 -6.66468561e-01 4.37149554e-01
-1.93150148e-01 -3.02094012e-01 2.30218232e-01 -4.93425727e-01
-1.30570722e+00 1.94601208e-01 -1.48284823e-01 -7.55759776e-01
1.23460941e-01 -2.08059102e-01 -4.65882450e-01 -7.68941104e-01
4.24342990e-01 4.95206863e-01 3.38206500e-01 -4.00269300e-01
1.12851143e-01 1.21367812e+00 5.53743184e-01 -1.49749339e-01
6.79944754e-01 6.34806335e-01 -1.19305193e-01 -1.05977833e+00
-4.49560553e-01 -3.07956666e-01 -4.74007487e-01 -5.71257412e-01
4.29743648e-01 -1.04534292e+00 -1.00984359e+00 1.37089777e+00
-1.23328233e+00 2.51286030e-02 7.09324479e-01 7.68762082e-02
-2.82704413e-01 9.64569747e-01 -6.25777781e-01 -8.99390757e-01
-3.26783836e-01 -1.27091229e+00 1.31128466e+00 6.07664250e-02
3.77996385e-01 -5.26842237e-01 -3.29660326e-01 5.51049531e-01
1.91104770e-01 -3.15139860e-01 4.65550303e-01 3.35133984e-03
-7.80208647e-01 -2.43309475e-02 -5.25949895e-01 6.43734753e-01
6.69715643e-01 -5.88693656e-02 -1.36125493e+00 -6.64696753e-01
1.38637543e-01 -5.59894085e-01 1.06019473e+00 3.27801630e-02
1.66655850e+00 -6.60890341e-01 -4.04362082e-01 8.43973577e-01
1.11922884e+00 1.52411073e-01 8.33349168e-01 -5.92898801e-02
9.19483304e-01 6.18127525e-01 5.03921747e-01 6.87871039e-01
1.45644501e-01 7.49642909e-01 4.21311408e-01 3.01407218e-01
-4.58984882e-01 -3.67058158e-01 8.49941432e-01 9.59920958e-02
1.49597645e-01 -2.79555142e-01 -5.43872535e-01 3.15286756e-01
-1.50834394e+00 -1.25231242e+00 2.19722703e-01 2.13438749e+00
3.65630209e-01 -4.27556425e-01 2.15911701e-01 1.44915432e-01
9.52715814e-01 9.24060196e-02 -4.73824978e-01 3.66375269e-03
-4.74566892e-02 -5.15900999e-02 4.69139427e-01 3.36641103e-01
-1.20562255e+00 1.09989238e+00 6.61773205e+00 1.11272013e+00
-1.47010028e+00 2.71468639e-01 9.69986796e-01 -2.77674407e-01
1.21448390e-01 -5.64640284e-01 -1.11399794e+00 7.90903091e-01
5.37678063e-01 6.25261366e-02 6.97756469e-01 7.46828794e-01
2.93015838e-01 1.95707202e-01 -9.68789041e-01 1.78443074e+00
6.11568570e-01 -1.51399064e+00 7.21204430e-02 2.19940305e-01
6.19845986e-01 -2.16343239e-01 4.03160930e-01 6.32938603e-03
-1.91249743e-01 -1.19493794e+00 4.35362786e-01 9.83770937e-02
1.19887185e+00 -8.18791568e-01 4.90521312e-01 6.08674623e-02
-1.25938714e+00 -6.24848068e-01 -3.71584088e-01 1.89682350e-01
-1.36205062e-01 2.03078389e-01 -9.97740388e-01 9.60382447e-02
1.10763526e+00 9.11724865e-01 -5.00530839e-01 9.83585179e-01
6.11000769e-02 3.73290777e-01 -2.57333636e-01 2.15214193e-01
-2.06474051e-01 1.04512945e-01 3.82090747e-01 9.42251205e-01
8.09099972e-01 -1.67768791e-01 -1.04670115e-01 4.27308977e-01
-4.17840362e-01 -1.26512155e-01 -6.16257548e-01 -5.24641424e-02
5.83344519e-01 1.24641109e+00 -2.56944627e-01 -2.25913614e-01
-6.33189499e-01 1.31503963e+00 1.85711324e-01 9.28741395e-02
-7.94688582e-01 -3.92988138e-02 1.08934927e+00 2.51889348e-01
6.33540630e-01 1.45084903e-01 4.31387216e-01 -1.22578752e+00
1.54756799e-01 -1.02498829e+00 5.16457438e-01 -6.20817423e-01
-1.32496905e+00 4.81468201e-01 -2.20655456e-01 -1.23189712e+00
-1.71031252e-01 -1.09337449e+00 -5.14140606e-01 4.49910700e-01
-1.60710001e+00 -1.20176947e+00 -6.20067954e-01 9.92139757e-01
6.70265555e-01 -5.86536109e-01 4.06217337e-01 6.05993807e-01
-7.22170591e-01 1.21590674e+00 1.07486151e-01 4.63616252e-01
9.10962284e-01 -2.89031625e-01 4.08304363e-01 7.08803833e-01
-1.80279329e-01 1.88187137e-01 9.80527848e-02 -7.70841956e-01
-1.87611091e+00 -1.17921674e+00 3.48515242e-01 -4.54195082e-01
3.48627388e-01 -5.35956979e-01 -6.93694055e-01 5.39862752e-01
-1.59260463e-02 3.49607170e-01 3.27788949e-01 -3.62336189e-01
-7.18507826e-01 -5.44512033e-01 -1.63386524e+00 1.64354920e-01
1.26939785e+00 -7.72904932e-01 -1.83395714e-01 2.51028687e-01
6.39649630e-02 -1.71999991e-01 -3.87993515e-01 3.99689227e-01
9.40341413e-01 -1.08608711e+00 1.08350670e+00 -2.99627870e-01
8.36406425e-02 -9.98537838e-02 -4.23537008e-02 -1.09340680e+00
-1.57346115e-01 -7.15737462e-01 -4.08095956e-01 1.24147832e+00
-2.41871208e-01 -6.30257070e-01 1.14596915e+00 -3.70539259e-03
3.07379097e-01 -3.80011946e-01 -1.16750669e+00 -6.83637500e-01
-1.55019209e-01 -1.70764267e-01 7.41415858e-01 6.99863553e-01
-1.10547192e-01 -1.74827963e-01 -1.01624095e+00 2.84922421e-01
9.68173146e-01 -1.04201563e-01 7.09910393e-01 -9.62502420e-01
-7.20055476e-02 -2.16842994e-01 -5.28388381e-01 -1.14225209e+00
2.11699352e-01 -5.33024848e-01 -2.56789953e-01 -6.52996421e-01
4.54498291e-01 -1.83616638e-01 -6.27612844e-02 4.56707507e-01
1.15715705e-01 7.29216218e-01 -7.82735366e-03 -9.80484635e-02
-3.23511779e-01 5.49764097e-01 1.16037679e+00 -3.72616053e-01
2.79876411e-01 -2.32171461e-01 -5.91396928e-01 5.87459087e-01
5.15030324e-01 -2.60358393e-01 -4.31275040e-01 -5.58649004e-01
-1.61082938e-01 1.30329385e-01 5.74105382e-01 -7.63729155e-01
-2.97198119e-03 9.15600918e-03 9.42289591e-01 -5.16363382e-01
4.94929701e-01 -7.44706273e-01 -1.78658932e-01 4.81062770e-01
2.44644284e-01 -1.46986365e-01 2.87265122e-01 6.63574159e-01
8.93134773e-02 -1.76326502e-02 8.44269693e-01 4.97843511e-03
-9.28759158e-01 7.03927457e-01 -3.31119359e-01 -4.46457893e-01
1.11731088e+00 -3.58063757e-01 -6.45930111e-01 -4.60378557e-01
-3.17660719e-01 1.76738203e-02 4.38768953e-01 9.00033414e-01
1.17428851e+00 -1.51879632e+00 -7.60377705e-01 8.80380273e-01
7.15582967e-02 -6.55528307e-01 7.55102158e-01 3.31614226e-01
-4.14722413e-01 2.34435916e-01 -4.04632658e-01 -6.73847139e-01
-2.02078700e+00 6.93041503e-01 6.54254615e-01 4.56064284e-01
-3.96185279e-01 1.07046711e+00 4.27354723e-01 -1.44615665e-01
2.24687248e-01 2.59329051e-01 -1.04379676e-01 1.09009571e-01
1.23153198e+00 4.73203629e-01 2.89352983e-01 -1.06055725e+00
-4.99925345e-01 7.89531112e-01 -3.45307291e-01 3.54706049e-01
9.92328823e-01 -3.73555422e-01 -1.37237823e-02 -5.52079260e-01
1.66906810e+00 1.26386300e-01 -1.65782785e+00 -2.15659320e-01
-4.87312526e-01 -1.31154060e+00 3.54920954e-01 -4.80807036e-01
-1.50603068e+00 8.32316101e-01 1.07736778e+00 -3.31896335e-01
1.30479074e+00 2.72518098e-02 9.50421572e-01 3.20345551e-01
5.11322081e-01 -1.00909650e+00 7.85483420e-01 5.92025071e-02
8.49616289e-01 -1.32162333e+00 -1.90818131e-01 -7.90095687e-01
-3.20509881e-01 1.36405563e+00 7.23723710e-01 1.59158796e-01
7.64393628e-01 2.45094255e-01 9.12771523e-02 4.09913436e-03
-5.25017440e-01 2.49920145e-01 1.34866253e-01 7.47511983e-01
-2.60146528e-01 -2.46355891e-01 3.35507125e-01 3.71215284e-01
-8.12643394e-03 1.45891249e-01 3.61301720e-01 5.34351528e-01
-3.91838491e-01 -1.11156225e+00 -5.94672859e-01 4.46920991e-01
-5.11773169e-01 2.14259356e-01 -5.55499852e-01 3.32278460e-01
2.06416488e-01 1.17925465e+00 4.24947888e-02 -6.77492440e-01
-1.79920942e-01 -1.80572599e-01 1.04618466e+00 -1.52952969e-01
1.08957489e-03 -1.61869377e-02 -1.09338015e-02 -7.59885311e-01
-5.26461229e-02 -6.72661006e-01 -5.79496622e-01 -7.05453992e-01
-3.13495338e-01 -3.18414927e-01 3.86757791e-01 9.63486850e-01
3.81245822e-01 -2.05570385e-01 1.19178414e+00 -1.04777515e+00
-4.90539640e-01 -8.60532522e-01 -4.81795311e-01 2.03532130e-01
8.32841456e-01 -8.72305930e-01 -5.13617158e-01 -2.65625492e-02] | [12.887310028076172, 1.0403811931610107] |
f9be96b6-c859-4e3b-bd4b-8f1866c03f56 | dependency-based-embeddings-for-sentence | null | null | https://aclanthology.org/N16-1175 | https://aclanthology.org/N16-1175.pdf | Dependency Based Embeddings for Sentence Classification Tasks | null | ['Man', 'ros', 'Suresh har', 'Alex Komninos'] | 2016-06-01 | null | null | null | naacl-2016-6 | ['learning-word-embeddings'] | ['methodology'] | [-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] | [-7.294086933135986, 3.6589674949645996] |
2c0d3568-556d-4974-84e3-7e1a5f8845c3 | are-these-birds-similar-learning-branched | null | null | https://ieeexplore.ieee.org/document/8960960 | http://artelab.dista.uninsubria.it/res/research/papers/2019/2019-IVCNZ-Nawaz-Birds.pdf | Are These Birds Similar: Learning Branched Networks for Fine-grained Representations | Fine-grained image classification is a challenging task due to the presence of hierarchical coarse-to-fine-grained distribution in the dataset. Generally, parts are used to discriminate various objects in fine-grained datasets, however, not all parts are beneficial and indispensable. In recent years, natural language descriptions are used to obtain information on discriminative parts of the object. This paper leverages on natural language description and proposes a strategy for learning the joint representation of natural language description and images using a two-branch network with multiple layers to improve the fine-grained classification task. Extensive experiments show that our approach gains significant improvements in accuracy for the fine-grained image classification task. Furthermore, our method achieves new state-of-the-art results on the CUB-200-2011 dataset. | ['Ignazio Gallo', 'Nicola Landro', 'Moreno Caraffini', 'Alessandro Calefati', 'Shah Nawaz'] | 2020-01-16 | null | null | null | null | ['multimodal-text-and-image-classification'] | ['methodology'] | [-1.97964236e-01 -4.08876300e-01 -6.71944678e-01 -7.67452776e-01
-9.09691513e-01 -6.67542517e-01 8.23252857e-01 9.43588391e-02
-2.35930458e-01 7.33765006e-01 6.39846683e-01 3.48469406e-01
-1.76420137e-01 -8.05207431e-01 -5.43922544e-01 -7.05201447e-01
1.76192775e-01 4.81761307e-01 8.15105662e-02 2.57452000e-02
1.68263972e-01 6.94608927e-01 -1.60013366e+00 8.46601129e-01
4.54951167e-01 1.47630823e+00 2.83767402e-01 3.84687036e-01
-3.95193487e-01 1.13525331e+00 -3.89614731e-01 -2.30087936e-01
1.90926567e-01 -3.36654708e-02 -9.64246571e-01 5.04922628e-01
9.09365594e-01 -4.96478796e-01 -3.36134791e-01 1.21463633e+00
1.10854805e-01 6.49615899e-02 1.07090557e+00 -1.16000569e+00
-9.91657197e-01 6.52050912e-01 -7.27019489e-01 1.34250477e-01
6.83851391e-02 -2.18529091e-03 1.31154704e+00 -9.32045162e-01
2.31840059e-01 1.53540778e+00 4.26098049e-01 4.46425825e-01
-1.28540885e+00 -9.08240259e-01 4.98254955e-01 2.45435879e-01
-1.93145692e+00 -2.50770658e-01 5.69731057e-01 -6.58879101e-01
9.86048400e-01 -2.02658642e-02 3.30984086e-01 1.03671849e+00
2.14929104e-01 6.97678328e-01 1.38942921e+00 -1.00148827e-01
-2.68193875e-02 -2.47952528e-02 4.27252322e-01 8.04372549e-01
2.09577277e-01 6.91286698e-02 -4.02920246e-01 8.95435959e-02
7.59529054e-01 3.66367728e-01 4.83471155e-02 -2.54176557e-01
-1.11482716e+00 9.65289474e-01 9.16443646e-01 3.60284626e-01
-5.01157761e-01 2.17182353e-01 5.02542913e-01 5.41702211e-02
6.71184480e-01 5.25531232e-01 -5.30193210e-01 1.55197224e-02
-9.37117755e-01 3.60997736e-01 6.56662464e-01 9.96591628e-01
1.02022028e+00 -2.49337748e-01 -7.04533756e-01 1.01686144e+00
8.90715197e-02 2.70621985e-01 4.53025967e-01 -7.77963459e-01
3.52345169e-01 7.94055581e-01 -1.02229662e-01 -7.82950342e-01
-3.38755935e-01 -6.48974419e-01 -1.50164771e+00 -1.12802520e-01
1.67032361e-01 5.83567202e-01 -1.12445402e+00 1.51039732e+00
-1.71723291e-01 -1.07649647e-01 -4.06050503e-01 8.88124824e-01
1.13663483e+00 4.97224867e-01 4.57803786e-01 3.00063103e-01
1.49038672e+00 -1.27860689e+00 -3.29309911e-01 -2.08326504e-01
2.10082114e-01 -2.88379520e-01 1.18252933e+00 1.60935149e-01
-6.83368444e-01 -9.89274383e-01 -8.22724700e-01 -2.48182133e-01
-7.07296491e-01 1.56748995e-01 6.95427656e-01 2.18156397e-01
-9.76285458e-01 3.10474306e-01 -3.05401981e-01 -1.04828022e-01
1.03275979e+00 5.95638826e-02 -6.58996105e-01 -4.50694770e-01
-9.64757860e-01 7.24887133e-01 8.39485645e-01 -3.06768686e-01
-1.07047522e+00 -8.22475851e-01 -1.00470722e+00 4.31761950e-01
1.08437382e-01 -8.75888288e-01 1.10683382e+00 -5.85328817e-01
-8.14017296e-01 1.24699152e+00 -5.09090126e-02 -4.13707346e-01
2.72255659e-01 5.65706342e-02 -8.67159888e-02 2.94992421e-02
6.59567297e-01 1.04440510e+00 8.72619331e-01 -1.18970811e+00
-9.99618709e-01 -2.75138706e-01 5.30120969e-01 -1.86068892e-01
-1.08126886e-01 -2.96850741e-01 -4.77812022e-01 -8.62344980e-01
-2.33243302e-01 -7.13496149e-01 -1.76313594e-01 -1.52176549e-03
-4.90537763e-01 -7.50121415e-01 4.20567125e-01 -1.23929679e-01
1.09102452e+00 -2.25717235e+00 -1.54606998e-01 -3.05014044e-01
5.78647435e-01 -1.89450115e-01 -3.01169783e-01 2.58881241e-01
1.33996573e-03 4.65855628e-01 1.44497246e-01 -2.79451877e-01
5.86290956e-01 2.34312832e-01 -6.33923650e-01 3.83267105e-01
3.72062415e-01 1.22695553e+00 -6.78019822e-01 -6.12410426e-01
3.28907609e-01 4.08911288e-01 -5.06170690e-01 4.34479803e-01
-1.87272325e-01 2.88783193e-01 -7.49092460e-01 8.12641919e-01
6.34392321e-01 -8.10675740e-01 -3.85962129e-01 -7.14998186e-01
2.62668030e-03 8.14642534e-02 -6.21781826e-01 1.68855226e+00
-6.42454743e-01 2.80130595e-01 -2.70890683e-01 -9.06596780e-01
5.67574799e-01 -7.25043118e-02 7.26319775e-02 -9.43551242e-01
5.53828180e-02 4.23468612e-02 -1.86954349e-01 -1.17820784e-01
5.59832096e-01 -4.76478994e-01 -7.89241314e-01 3.46526623e-01
2.60939479e-01 -3.57459188e-01 5.05539298e-01 1.47343457e-01
7.42183626e-01 6.15869761e-02 9.42665398e-01 -4.23557341e-01
4.25147444e-01 9.97162834e-02 3.41989070e-01 1.17298639e+00
-4.29375529e-01 6.93951368e-01 1.41600132e-01 -6.66233480e-01
-8.99475098e-01 -9.65294898e-01 -2.93788314e-01 1.56932676e+00
3.00450861e-01 -4.40202892e-01 -4.53489035e-01 -8.92751038e-01
2.27874339e-01 3.30244452e-01 -1.07646096e+00 -6.79074302e-02
-1.77349404e-01 -4.28454936e-01 5.35931349e-01 8.71652484e-01
9.64422047e-01 -1.06718719e+00 -1.84570611e-01 -2.80973036e-02
-2.85781473e-01 -1.40812790e+00 -6.26223564e-01 4.30848539e-01
-3.60573053e-01 -8.34936559e-01 -7.26784468e-01 -8.50777507e-01
4.91535693e-01 5.06167233e-01 1.69055176e+00 3.86593267e-02
-4.21205252e-01 1.11997947e-01 -4.34506655e-01 -4.66382839e-02
-5.27712964e-02 1.95830375e-01 -1.90825373e-01 -5.23471348e-02
4.85452116e-01 -1.79937452e-01 -4.90812331e-01 1.50542229e-01
-8.26471984e-01 1.48527831e-01 8.88995528e-01 1.00807524e+00
9.77185667e-01 5.06344199e-01 3.21358114e-01 -8.10116053e-01
4.61300224e-01 -3.07868063e-01 -4.14197475e-01 2.64383256e-01
-2.96858788e-01 2.88085401e-01 9.23127890e-01 -3.18271369e-01
-8.31801593e-01 -1.24383599e-01 -1.26968175e-01 -5.20354629e-01
-7.04527378e-01 3.78951401e-01 -1.63543075e-01 -8.04931000e-02
2.66952157e-01 3.76892775e-01 -7.92498350e-01 -6.23228371e-01
5.72718501e-01 7.07991064e-01 4.52159643e-01 -7.13324547e-01
7.19542861e-01 4.87736642e-01 -2.38300748e-02 -3.71905148e-01
-1.80979371e+00 -8.30015600e-01 -8.62804353e-01 2.91155398e-01
1.06782842e+00 -1.37659824e+00 -6.80957913e-01 4.01705891e-01
-9.57630754e-01 -2.38596827e-01 -2.37447932e-01 1.24528192e-01
-5.55180669e-01 -5.42363077e-02 -6.92633271e-01 -1.78784266e-01
-1.77764609e-01 -1.12881422e+00 1.72115600e+00 1.48107246e-01
-7.53745213e-02 -9.57727611e-01 -1.92566976e-01 2.99829572e-01
5.16178489e-01 -1.98695771e-02 1.02362025e+00 -3.66464347e-01
-6.38587832e-01 -5.54882325e-02 -1.05817676e+00 1.73555374e-01
4.11369473e-01 -4.15757179e-01 -1.07241619e+00 -3.24985921e-01
-3.07719439e-01 -9.11393046e-01 1.38229775e+00 3.96623403e-01
1.87019551e+00 -4.81873870e-01 -3.16607773e-01 7.02363849e-01
1.64836514e+00 -3.44647110e-01 2.68312722e-01 3.84854674e-01
9.91195381e-01 4.95840758e-01 7.43810952e-01 7.28068888e-01
7.27041364e-01 6.07339382e-01 4.36192930e-01 -9.45693105e-02
-5.82821906e-01 -3.13307434e-01 -2.31966704e-01 3.75075668e-01
2.20807776e-01 -3.22270274e-01 -6.89381540e-01 5.98867059e-01
-1.56184721e+00 -1.01725698e+00 3.85341257e-01 1.55593586e+00
8.90857279e-01 -1.33943530e-02 4.55333255e-02 -2.29709342e-01
6.66321754e-01 5.16679525e-01 -5.90369105e-01 -9.46668759e-02
-2.57461399e-01 1.25193402e-01 5.22266924e-01 2.28352770e-01
-1.70231831e+00 1.23544157e+00 6.35371780e+00 1.32500613e+00
-1.17791021e+00 3.02129667e-02 8.68400753e-01 1.79252252e-01
-5.18047772e-02 -5.32189786e-01 -1.03590417e+00 3.49488378e-01
4.10405338e-01 1.09174866e-02 3.53025585e-01 8.66009116e-01
-2.19963893e-01 1.99304327e-01 -1.22261167e+00 1.33544171e+00
1.93189904e-02 -1.57017803e+00 6.62021995e-01 -6.19260892e-02
9.08908606e-01 1.53479308e-01 2.49932483e-02 6.11076534e-01
4.71088678e-01 -1.52298009e+00 9.45151925e-01 4.86992836e-01
1.24401116e+00 -6.82043731e-01 7.15597153e-01 4.71390158e-01
-1.63893783e+00 -1.74504761e-02 -5.32620013e-01 -1.89810872e-01
-2.31609672e-01 5.22123814e-01 -5.71702302e-01 3.83678436e-01
1.00349355e+00 1.04960907e+00 -7.88433194e-01 6.23616636e-01
-2.99510986e-01 4.16877657e-01 2.15992227e-01 2.31019370e-02
6.56336844e-01 3.35048765e-01 -1.84567362e-01 1.52586055e+00
-9.65823159e-02 7.02476874e-02 8.33473086e-01 9.75308239e-01
-4.38110918e-01 -3.45189244e-01 -3.59024435e-01 -4.01095927e-01
2.10841492e-01 1.42001605e+00 -6.68942213e-01 -4.78129506e-01
-4.48579580e-01 8.35140109e-01 7.96004713e-01 2.69482136e-01
-6.66057646e-01 -2.01345921e-01 7.65679359e-01 2.31182147e-02
6.36153877e-01 -8.54295120e-02 -2.53735214e-01 -1.27015877e+00
-3.09093237e-01 -8.78312826e-01 5.68388641e-01 -7.74701595e-01
-2.13690400e+00 8.49419475e-01 -5.76674901e-02 -1.20771468e+00
-3.43846440e-01 -7.93511569e-01 7.22473040e-02 9.62561905e-01
-1.87782264e+00 -1.71995032e+00 -5.79181790e-01 7.08085418e-01
8.37255716e-01 -7.82430470e-02 1.06643474e+00 2.49537885e-01
-3.81225534e-02 5.88959873e-01 -2.26485282e-01 4.02910262e-01
5.98983586e-01 -1.42155004e+00 2.35935166e-01 3.91117901e-01
1.77292034e-01 5.62324166e-01 4.61572826e-01 -3.48593473e-01
-9.85225141e-01 -1.57811761e+00 8.21166337e-01 -3.07971448e-01
7.84458637e-01 -5.00877738e-01 -6.67504787e-01 5.53821445e-01
1.15981139e-01 5.95070422e-01 6.15995884e-01 1.67493850e-01
-1.13639998e+00 -1.98443115e-01 -1.13876963e+00 3.19363385e-01
8.51089120e-01 -1.15571129e+00 -6.46056831e-01 1.95101768e-01
6.88839495e-01 -1.49884418e-01 -9.21090841e-01 4.46613371e-01
7.48149455e-01 -9.13346469e-01 1.00224531e+00 -6.81235969e-01
6.44503295e-01 -4.00615454e-01 -7.30946541e-01 -1.36946833e+00
-1.12164330e+00 2.25450352e-01 9.85885412e-02 1.10407102e+00
-2.74972711e-02 -4.29034203e-01 6.06218517e-01 9.56877246e-02
7.56886825e-02 -7.31362879e-01 -6.16590977e-01 -8.03434968e-01
3.11964393e-01 -2.53772467e-01 7.83791065e-01 7.50918746e-01
-5.39309919e-01 6.04194105e-01 -3.07593435e-01 4.09584343e-02
7.77465701e-01 9.18818831e-01 5.76139331e-01 -1.26701498e+00
-1.97384447e-01 -7.77763724e-01 -6.66706979e-01 -1.38345599e+00
6.83544040e-01 -9.69434202e-01 3.36385936e-01 -1.76105392e+00
1.15468431e+00 -4.09464568e-01 -5.84360361e-01 7.19100773e-01
-1.77831382e-01 9.86360610e-01 2.80403793e-01 3.67716789e-01
-1.22000325e+00 5.14592707e-01 1.41561353e+00 -8.35732043e-01
5.61205566e-01 -3.37657362e-01 -1.07908392e+00 6.53447509e-01
4.63732392e-01 -4.02512580e-01 -2.10384741e-01 -3.47015560e-01
-5.75320087e-02 -2.62782246e-01 6.30717099e-01 -6.40175641e-01
7.05586234e-03 -3.42085183e-01 6.99632704e-01 -8.89447629e-01
3.38918626e-01 -7.38066137e-01 -1.34591997e-01 2.65158296e-01
-5.97411335e-01 -3.21959049e-01 2.11176440e-01 4.08168405e-01
-7.36080229e-01 2.21714586e-01 9.49393928e-01 -4.53423440e-01
-1.26838756e+00 9.40216243e-01 6.08751029e-02 2.74747610e-01
6.45304263e-01 6.91527203e-02 -5.29703140e-01 -1.80053130e-01
-5.09376228e-01 3.94976400e-02 5.31292319e-01 5.05697727e-01
4.16867286e-01 -1.66459095e+00 -9.26353335e-01 1.03696331e-01
8.32132518e-01 -1.15558222e-01 4.32502747e-01 4.58097190e-01
-9.56535488e-02 8.85840833e-01 -4.34625328e-01 -8.15374494e-01
-1.03459382e+00 7.17939794e-01 2.87298262e-01 -6.48785651e-01
-4.43922639e-01 1.04553378e+00 1.35014009e+00 -3.65330935e-01
9.68315601e-02 -4.48160857e-01 -3.58092278e-01 1.05822459e-01
7.45422542e-01 -7.38731325e-02 -1.49179980e-01 -1.04814255e+00
-5.87057829e-01 8.53408039e-01 -4.44785416e-01 4.76050198e-01
1.31418765e+00 -4.56498712e-01 -2.51174301e-01 5.02216697e-01
1.58606958e+00 -1.79023102e-01 -1.35755157e+00 -4.63762820e-01
-2.38986641e-01 -5.42356491e-01 1.40298083e-01 -1.08468020e+00
-9.17434156e-01 8.87723923e-01 2.74224043e-01 3.33213717e-01
1.21110117e+00 5.96529067e-01 4.44563031e-01 3.70524228e-01
7.31327832e-01 -4.93377119e-01 8.51433501e-02 8.05310547e-01
1.11596203e+00 -1.60110724e+00 -1.86217520e-02 -3.38255644e-01
-4.21873629e-01 1.00735021e+00 5.05316198e-01 -3.29440802e-01
7.55999327e-01 2.72259474e-01 -1.66666601e-02 -2.51942992e-01
-8.69205058e-01 -5.20183265e-01 8.82868826e-01 4.29139227e-01
5.30490100e-01 4.78457034e-01 2.67106235e-01 9.08156037e-01
-1.80668518e-01 -6.31430745e-02 6.29634014e-04 4.39211190e-01
-4.33717102e-01 -6.70664966e-01 -6.55186027e-02 6.93011403e-01
-6.69238269e-01 -3.84114265e-01 -4.18743223e-01 7.43921280e-01
4.54916626e-01 8.73821497e-01 2.04092309e-01 -2.54891723e-01
1.13592103e-01 -4.73241448e-01 7.59065688e-01 -8.32224727e-01
-5.24457097e-01 -7.40685910e-02 -1.18349157e-01 -8.60693276e-01
-5.08127213e-01 -1.11527115e-01 -1.05505979e+00 -2.50636250e-01
8.54084492e-02 1.51054123e-02 2.52894402e-01 1.22224534e+00
2.92880177e-01 7.08870769e-01 6.46393657e-01 -1.10742927e+00
-5.64344943e-01 -9.72609937e-01 -1.05285835e+00 5.49774349e-01
7.94295967e-01 -8.79433692e-01 -4.00824755e-01 -7.55442753e-02] | [9.632585525512695, 2.090695858001709] |
24ee5096-4aee-4066-87fa-8298c41d0613 | texturepose-supervising-human-mesh-estimation-1 | 1910.11322 | null | https://arxiv.org/abs/1910.11322v1 | https://arxiv.org/pdf/1910.11322v1.pdf | TexturePose: Supervising Human Mesh Estimation with Texture Consistency | This work addresses the problem of model-based human pose estimation. Recent approaches have made significant progress towards regressing the parameters of parametric human body models directly from images. Because of the absence of images with 3D shape ground truth, relevant approaches rely on 2D annotations or sophisticated architecture designs. In this work, we advocate that there are more cues we can leverage, which are available for free in natural images, i.e., without getting more annotations, or modifying the network architecture. We propose a natural form of supervision, that capitalizes on the appearance constancy of a person among different frames (or viewpoints). This seemingly insignificant and often overlooked cue goes a long way for model-based pose estimation. The parametric model we employ allows us to compute a texture map for each frame. Assuming that the texture of the person does not change dramatically between frames, we can apply a novel texture consistency loss, which enforces that each point in the texture map has the same texture value across all frames. Since the texture is transferred in this common texture map space, no camera motion computation is necessary, or even an assumption of smoothness among frames. This makes our proposed supervision applicable in a variety of settings, ranging from monocular video, to multi-view images. We benchmark our approach against strong baselines that require the same or even more annotations that we do and we consistently outperform them. Simultaneously, we achieve state-of-the-art results among model-based pose estimation approaches in different benchmarks. The project website with videos, results, and code can be found at https://seas.upenn.edu/~pavlakos/projects/texturepose. | ['Nikos Kolotouros', 'Georgios Pavlakos', 'Kostas Daniilidis'] | 2019-10-24 | texturepose-supervising-human-mesh-estimation | http://openaccess.thecvf.com/content_ICCV_2019/html/Pavlakos_TexturePose_Supervising_Human_Mesh_Estimation_With_Texture_Consistency_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Pavlakos_TexturePose_Supervising_Human_Mesh_Estimation_With_Texture_Consistency_ICCV_2019_paper.pdf | iccv-2019-10 | ['weakly-supervised-3d-human-pose-estimation'] | ['computer-vision'] | [-1.73563808e-02 8.97298604e-02 -2.07114458e-01 -4.12649184e-01
-4.86420214e-01 -4.94658172e-01 4.96540904e-01 -3.84969622e-01
-3.76178473e-01 5.20423770e-01 3.40141803e-01 2.86339670e-01
2.66541839e-01 -3.72676075e-01 -9.30949807e-01 -6.07180178e-01
1.23532951e-01 5.53880155e-01 2.95684576e-01 -2.95137197e-01
-1.48180485e-01 2.16480240e-01 -1.27055621e+00 4.24886085e-02
2.02609077e-01 1.01615858e+00 -5.89530962e-03 6.85611784e-01
4.48559433e-01 4.53285158e-01 -2.95430750e-01 -6.08508348e-01
6.05114222e-01 -3.10823560e-01 -8.65261197e-01 4.37797219e-01
1.01916373e+00 -5.59988260e-01 -5.06743848e-01 8.69440913e-01
4.45483178e-01 1.35540217e-01 5.04463434e-01 -1.14232683e+00
-2.83486009e-01 6.22885041e-02 -6.76053405e-01 -1.23378269e-01
4.00349498e-01 1.93813995e-01 8.56179416e-01 -8.49211395e-01
8.38602543e-01 1.31985104e+00 7.89625883e-01 6.89379513e-01
-1.18765891e+00 -2.85438091e-01 3.29009831e-01 9.65941790e-03
-1.29812419e+00 -6.03964925e-01 8.02212238e-01 -5.82803905e-01
4.80458945e-01 2.13065714e-01 8.44097853e-01 1.44019651e+00
1.83418810e-01 7.82272100e-01 9.04038787e-01 -4.62737262e-01
-1.17895231e-01 -1.34285361e-01 -1.18822314e-01 8.01155925e-01
3.02115709e-01 2.26227827e-02 -7.30770886e-01 1.00132249e-01
1.22687006e+00 8.92026871e-02 -3.73015493e-01 -8.99001777e-01
-1.25224698e+00 5.22676170e-01 1.81336462e-01 -1.13858700e-01
-1.43303812e-01 3.30125749e-01 3.16603929e-01 1.85027316e-01
4.73081797e-01 1.86270162e-01 -5.32873452e-01 -9.39090103e-02
-7.62864172e-01 5.14752984e-01 6.42653227e-01 9.97486413e-01
6.17889583e-01 -1.72926784e-01 8.70253425e-03 5.99250078e-01
2.55411297e-01 5.82734704e-01 1.95233285e-01 -1.25171340e+00
4.11724091e-01 2.05349684e-01 3.27241659e-01 -1.18999970e+00
-4.29761678e-01 -1.68516159e-01 -7.06865966e-01 1.48345932e-01
8.70985925e-01 -1.37485772e-01 -9.06586528e-01 1.93314826e+00
5.78172505e-01 6.05406761e-02 -4.04943556e-01 1.26581788e+00
4.91929859e-01 1.34686576e-02 -2.95843065e-01 1.83091342e-01
1.42274189e+00 -1.11285603e+00 -5.34088433e-01 -4.20893043e-01
3.49339217e-01 -6.76166832e-01 1.15783763e+00 4.69987273e-01
-1.14416778e+00 -5.27341664e-01 -1.01240253e+00 -2.69847125e-01
2.94972165e-03 1.13361895e-01 6.77282274e-01 5.28597414e-01
-9.30749059e-01 6.54894173e-01 -1.31165707e+00 -4.95809913e-01
1.30058482e-01 4.21834499e-01 -6.89571440e-01 -5.15054306e-03
-8.66550922e-01 8.32979441e-01 -3.25811058e-02 2.78721362e-01
-8.26178908e-01 -3.64203453e-01 -1.00621748e+00 -3.22291017e-01
7.77308285e-01 -1.04897237e+00 1.37787294e+00 -1.27486718e+00
-1.56030619e+00 9.77217257e-01 -1.73734069e-01 -2.72951275e-01
9.97232914e-01 -6.74146652e-01 1.35343075e-01 1.53037772e-01
-5.48505746e-02 6.98556960e-01 9.64112878e-01 -1.20031977e+00
-2.52810597e-01 -4.08481926e-01 2.93063015e-01 2.90333927e-01
-2.03793913e-01 -8.98051858e-02 -9.50579882e-01 -7.51750588e-01
8.48897174e-02 -1.47117519e+00 -1.03076719e-01 4.23606575e-01
-5.11435330e-01 6.65686876e-02 5.36792397e-01 -8.25200081e-01
9.18974996e-01 -1.98492050e+00 5.57698131e-01 1.37980729e-01
2.15674683e-01 -2.20750213e-01 1.35448173e-01 1.46737859e-01
3.87268737e-02 -8.46917853e-02 -1.27449006e-01 -7.39623725e-01
6.55213967e-02 2.29810953e-01 -9.85120013e-02 7.16349721e-01
3.18600126e-02 8.41521561e-01 -6.13004267e-01 -4.74143237e-01
2.26065651e-01 6.40799105e-01 -8.08088839e-01 1.03657454e-01
-2.06019640e-01 9.04345930e-01 -3.05070311e-01 4.85912859e-01
4.55929548e-01 -5.58850527e-01 3.54773879e-01 -2.68990576e-01
1.73451483e-01 2.57012367e-01 -1.32968330e+00 2.14476037e+00
-1.96464047e-01 4.78413850e-01 1.11977048e-01 -8.64543736e-01
4.67076927e-01 3.88493448e-01 6.14342451e-01 -2.95564711e-01
2.74583220e-01 -3.62364054e-02 -4.12349105e-02 -2.15789720e-01
3.80879819e-01 1.41533426e-04 3.20434906e-02 3.03790301e-01
1.13490015e-01 7.75073245e-02 6.47737309e-02 -1.30225435e-01
8.09456348e-01 7.04279542e-01 7.05927759e-02 -2.01175392e-01
2.47215882e-01 -1.66150227e-01 7.16667891e-01 5.10811865e-01
-1.97154358e-01 9.39835370e-01 2.91903526e-01 -7.06770658e-01
-1.30519128e+00 -1.00714016e+00 -5.77588156e-02 1.08270025e+00
2.26438224e-01 -4.40862566e-01 -7.89493322e-01 -4.82933819e-01
-2.18060371e-02 -1.86778679e-01 -8.83300245e-01 1.32804036e-01
-9.08269465e-01 -3.44873697e-01 3.81068677e-01 6.82996392e-01
4.36143547e-01 -6.04183435e-01 -7.12052882e-01 -6.44173995e-02
-4.11251456e-01 -1.21480644e+00 -7.70248711e-01 -9.38229933e-02
-8.25190783e-01 -9.96892273e-01 -9.51408625e-01 -4.93942112e-01
7.27079213e-01 1.99574202e-01 1.19491255e+00 9.59031582e-02
-1.67842031e-01 5.22449851e-01 -1.18518457e-01 -1.96315035e-01
1.36505961e-01 9.39562842e-02 3.38088900e-01 5.01208194e-02
2.21353341e-02 -5.43060958e-01 -8.23540688e-01 5.32900929e-01
-4.50579226e-01 2.53584206e-01 3.52951527e-01 8.09727490e-01
6.90698624e-01 -4.58260596e-01 5.45093976e-02 -7.39383221e-01
-3.58199328e-02 -6.21258393e-02 -5.34898043e-01 1.01856649e-01
-1.65320665e-01 4.62856591e-02 3.24953616e-01 -5.72525084e-01
-8.48973095e-01 2.85826594e-01 1.08607830e-02 -8.24165881e-01
-1.72145039e-01 2.47270212e-01 -1.99149236e-01 -6.89750612e-02
5.86558938e-01 -1.11410566e-01 1.29461929e-01 -6.30996168e-01
2.26857215e-01 8.40982199e-02 7.08445609e-01 -7.99990475e-01
8.11891019e-01 7.92039871e-01 2.43889913e-02 -7.08447635e-01
-9.75089729e-01 -3.88177842e-01 -8.94729316e-01 -2.21236914e-01
1.01349783e+00 -1.11030877e+00 -7.04512775e-01 5.15642762e-01
-1.01003659e+00 -5.52220285e-01 -2.92176493e-02 4.71710622e-01
-6.89597547e-01 4.89501178e-01 -7.01064110e-01 -5.97796261e-01
-5.13155498e-02 -1.03528953e+00 1.27246118e+00 4.17980030e-02
-4.75877106e-01 -9.42220867e-01 9.45585668e-02 5.15566051e-01
1.35395259e-01 4.75962758e-01 4.01300222e-01 -1.54647797e-01
-5.37534118e-01 -2.09749073e-01 2.15561986e-01 1.90105960e-01
1.75613984e-01 -1.72695458e-01 -9.48670149e-01 -5.05119860e-01
-1.37597710e-01 -4.23429877e-01 6.68902159e-01 6.00068152e-01
1.12513971e+00 -3.12207282e-01 -2.76056707e-01 7.64503658e-01
9.39850688e-01 -3.49481046e-01 5.08915424e-01 4.13842410e-01
1.09704661e+00 7.44798243e-01 4.55205381e-01 3.49156946e-01
5.82632661e-01 1.25059056e+00 2.24730313e-01 -1.65540785e-01
-1.54322699e-01 -3.37297320e-01 3.62625390e-01 5.61035275e-01
-6.48088992e-01 -4.38400358e-02 -8.38010550e-01 3.53083253e-01
-1.93397975e+00 -9.19894755e-01 9.57554355e-02 2.56164026e+00
7.92061388e-01 1.32426545e-01 4.66071308e-01 -1.70577541e-01
5.03721237e-01 1.19823791e-01 -5.31266391e-01 1.41033664e-01
6.41867332e-03 4.58268262e-02 5.61088443e-01 5.76012969e-01
-1.34613705e+00 8.10769796e-01 5.98917484e+00 2.96337664e-01
-1.26677930e+00 -2.11211294e-03 5.51415920e-01 -4.82268035e-01
2.45831218e-02 -1.07956171e-01 -7.14314938e-01 3.77760142e-01
5.85277975e-01 1.97349116e-01 3.53466392e-01 7.78168619e-01
2.36601293e-01 1.91945769e-02 -1.44104612e+00 1.09115648e+00
7.18009770e-02 -9.68958855e-01 -2.51064360e-01 2.85342127e-01
6.72914863e-01 -4.66867574e-02 1.56073108e-01 -4.32501025e-02
1.53847411e-01 -9.64535534e-01 9.89709437e-01 5.99774659e-01
7.42429614e-01 -4.88542169e-01 6.39905930e-01 2.00988248e-01
-1.19529140e+00 2.61305004e-01 -3.49009067e-01 -2.25688875e-01
2.16689840e-01 3.34832132e-01 -3.95676076e-01 6.13082290e-01
9.14979637e-01 7.59264171e-01 -5.56432664e-01 7.82763898e-01
-2.56134897e-01 3.08309674e-01 -4.86789316e-01 4.92851228e-01
-1.62758455e-01 -7.31746927e-02 3.56818408e-01 9.61840093e-01
1.21009015e-01 -3.73852551e-02 5.22481620e-01 4.76167023e-01
-3.99989448e-02 -1.43455192e-02 -4.58503395e-01 3.66719395e-01
2.41775766e-01 1.07956016e+00 -5.91679454e-01 -2.67899215e-01
-6.21372402e-01 1.14637172e+00 3.35272878e-01 2.83769816e-01
-8.50929737e-01 2.49052808e-01 7.63438880e-01 4.09685075e-01
3.75234783e-01 -3.99749309e-01 -2.00307161e-01 -1.50545359e+00
4.30498600e-01 -1.06890547e+00 2.71125197e-01 -5.57409167e-01
-1.25489330e+00 3.63753885e-01 2.53905863e-01 -1.26511598e+00
-4.96175587e-01 -6.89827919e-01 -3.64029795e-01 6.47170842e-01
-1.17060876e+00 -1.37006223e+00 -3.97719115e-01 7.05481887e-01
4.08706993e-01 2.40239590e-01 5.96953809e-01 3.49074870e-01
-5.04734576e-01 7.66614914e-01 -3.84142041e-01 4.24272388e-01
1.10053575e+00 -1.17873561e+00 4.47155625e-01 7.84411788e-01
2.28822961e-01 7.25947082e-01 8.36638927e-01 -6.01589382e-01
-1.52553856e+00 -7.37675786e-01 5.21094799e-01 -9.02816474e-01
4.15961981e-01 -6.21037900e-01 -7.82115400e-01 1.02183473e+00
1.46143781e-02 2.26085827e-01 4.27137077e-01 2.79407471e-01
-3.76611888e-01 9.33910906e-02 -6.73580408e-01 6.41252518e-01
1.24344540e+00 -3.57859015e-01 -3.33470285e-01 2.86097884e-01
4.65484440e-01 -9.19727802e-01 -8.54502738e-01 3.68567646e-01
1.02165616e+00 -9.48554933e-01 1.14247942e+00 -6.23331547e-01
4.98581082e-01 -3.98252904e-01 -2.09348395e-01 -9.71104145e-01
-2.34535620e-01 -6.00902259e-01 -1.77337766e-01 8.55398893e-01
2.37390071e-01 -3.94812584e-01 1.00186944e+00 9.76279676e-01
-3.35275866e-02 -8.08001935e-01 -8.46395850e-01 -8.01585495e-01
1.59036562e-01 -2.34123453e-01 2.79395819e-01 9.50398743e-01
-1.57862157e-01 2.58535206e-01 -9.37616289e-01 2.43867278e-01
6.68831646e-01 -6.51824549e-02 1.27422750e+00 -1.21347129e+00
-7.34579086e-01 -2.18399271e-01 -4.72756445e-01 -1.38868833e+00
7.25192279e-02 -4.21239436e-01 6.28614202e-02 -1.11504960e+00
2.90483534e-01 -2.37498865e-01 1.46667343e-02 6.81693673e-01
-1.21848449e-01 4.39676464e-01 2.84413606e-01 4.66474414e-01
-5.95209956e-01 4.27556545e-01 1.39735866e+00 5.34890480e-02
1.61380116e-02 -3.63094546e-02 -4.82102305e-01 1.07569766e+00
6.57100201e-01 -3.16881597e-01 -3.10262799e-01 -7.16361642e-01
1.41682193e-01 2.16146354e-02 7.56846905e-01 -9.15374100e-01
1.51073352e-01 -1.28310531e-01 6.53936327e-01 -2.12250188e-01
7.29062259e-01 -7.67856300e-01 3.81305307e-01 3.37054729e-01
-9.32371616e-02 2.21237227e-01 4.60215807e-02 5.06849647e-01
-3.36702578e-02 1.24893740e-01 7.04664946e-01 -2.86843687e-01
-5.28904498e-01 5.08146405e-01 1.62864983e-01 1.75821126e-01
6.18289411e-01 -3.52270693e-01 -1.14225492e-01 -6.20347261e-01
-8.64294827e-01 4.22407947e-02 9.58471060e-01 5.37242055e-01
2.47124210e-01 -1.40016162e+00 -5.60977459e-01 7.33496472e-02
7.55517259e-02 1.51979521e-01 1.93677023e-01 1.11273384e+00
-4.19706285e-01 2.51465321e-01 -3.13163906e-01 -8.79119039e-01
-1.28260338e+00 2.91047186e-01 5.46454072e-01 -1.93454117e-01
-8.42970788e-01 7.00561821e-01 6.19403005e-01 -3.56806517e-01
2.32404843e-01 -2.23131567e-01 1.83137566e-01 -3.13567251e-01
3.87245655e-01 1.54658750e-01 -9.65702534e-03 -8.59311819e-01
-3.76764387e-01 9.12483692e-01 -7.83342347e-02 -2.08191171e-01
1.25426304e+00 -2.10430861e-01 1.95702299e-01 5.54099381e-01
9.33269203e-01 1.11414954e-01 -1.77891624e+00 -1.60234660e-01
-3.10025603e-01 -6.58626974e-01 -2.03060865e-01 -5.67258298e-01
-1.15512490e+00 7.82423198e-01 5.16456783e-01 -2.22414315e-01
8.47458661e-01 3.55062783e-02 5.89034617e-01 3.91252577e-01
5.74968934e-01 -1.06703246e+00 2.32629761e-01 5.67174137e-01
8.67228389e-01 -1.39497292e+00 8.77257586e-02 -4.61873204e-01
-6.12234414e-01 8.90400648e-01 6.74181998e-01 -1.95173055e-01
4.16758567e-01 2.55830735e-01 1.19853407e-01 -1.82825983e-01
-6.21303916e-01 7.40706734e-03 5.00262439e-01 5.97129583e-01
6.82530403e-01 -4.72100079e-02 -1.17485048e-02 4.03343499e-01
-4.57756013e-01 -1.03243969e-01 3.24358195e-01 8.66864562e-01
-1.90754697e-01 -1.17614615e+00 -5.26172221e-01 1.24799408e-01
-5.28869689e-01 1.49224803e-01 -4.11441594e-01 9.44387615e-01
-4.53009177e-03 6.28823340e-01 -3.84210162e-02 -2.87930280e-01
3.91960740e-01 1.14202257e-02 8.60189676e-01 -5.08751392e-01
-1.41513616e-01 3.65432918e-01 9.39141363e-02 -7.63750017e-01
-5.89451134e-01 -8.04325163e-01 -9.52751696e-01 -5.37514091e-01
-1.90421656e-01 -2.24288046e-01 3.05634886e-01 9.53453422e-01
2.17298925e-01 2.73570061e-01 2.26700976e-01 -1.32294786e+00
-2.32520044e-01 -9.05402184e-01 -3.43891054e-01 5.48112273e-01
3.83530468e-01 -1.15612459e+00 -7.78581649e-02 4.16215390e-01] | [7.155539512634277, -1.1022436618804932] |
0f56fac5-f1e6-46a3-9c8e-0cb56a55ee29 | clothing-and-people-a-social-signal | 1704.02231 | null | http://arxiv.org/abs/1704.02231v1 | http://arxiv.org/pdf/1704.02231v1.pdf | Clothing and People - A Social Signal Processing Perspective | In our society and century, clothing is not anymore used only as a means for
body protection. Our paper builds upon the evidence, studied within the social
sciences, that clothing brings a clear communicative message in terms of social
signals, influencing the impression and behaviour of others towards a person.
In fact, clothing correlates with personality traits, both in terms of
self-assessment and assessments that unacquainted people give to an individual.
The consequences of these facts are important: the influence of clothing on the
decision making of individuals has been investigated in the literature, showing
that it represents a discriminative factor to differentiate among diverse
groups of people. Unfortunately, this has been observed after cumbersome and
expensive manual annotations, on very restricted populations, limiting the
scope of the resulting claims. With this position paper, we want to sketch the
main steps of the very first systematic analysis, driven by social signal
processing techniques, of the relationship between clothing and social signals,
both sent and perceived. Thanks to human parsing technologies, which exhibit
high robustness owing to deep learning architectures, we are now capable to
isolate visual patterns characterising a large types of garments. These
algorithms will be used to capture statistical relations on a large corpus of
evidence to confirm the sociological findings and to go beyond the state of the
art. | ['Petia Radeva', 'Federico Parezzan', 'Marco Cristani', 'Mariella Dimiccoli', 'Maedeh Aghaei'] | 2017-04-07 | null | null | null | null | ['human-parsing'] | ['computer-vision'] | [ 2.59288520e-01 2.85983860e-01 7.49485642e-02 -5.19111872e-01
-2.39871303e-03 -5.96550643e-01 8.35536599e-01 6.97777212e-01
-6.81832731e-01 4.42692727e-01 5.28924823e-01 1.35618910e-01
-3.39620590e-01 -8.12559068e-01 -5.45458734e-01 -6.41000628e-01
-1.81931213e-01 2.65675426e-01 1.90867379e-01 -4.51758802e-01
5.40191540e-03 3.94189090e-01 -1.67501163e+00 4.04768586e-01
2.07984224e-01 9.64516163e-01 -7.14041293e-02 4.93856877e-01
-6.09318074e-03 3.37619871e-01 -5.33582270e-01 -1.00462401e+00
-9.22529176e-02 -4.69256908e-01 -4.65649664e-01 3.01272646e-02
4.94647145e-01 -7.29937851e-02 9.60190669e-02 9.30947781e-01
5.91250956e-01 -5.93337119e-01 3.81862909e-01 -5.48944950e-01
-7.20646143e-01 8.77110124e-01 -1.93735331e-01 1.14499740e-01
7.11643040e-01 9.75030065e-02 1.05694830e+00 -2.10903898e-01
7.78591633e-01 1.11275077e+00 8.16368043e-01 4.16954547e-01
-1.34073901e+00 -3.45593214e-01 -8.22793096e-02 8.75277221e-02
-9.87229228e-01 -6.18485689e-01 9.41668212e-01 -7.13918805e-01
1.61368012e-01 6.55930936e-01 1.23098397e+00 1.64055586e+00
-8.64391625e-02 4.84688193e-01 1.58567607e+00 -4.95637566e-01
9.70611721e-02 5.79161406e-01 1.71229586e-01 4.36151147e-01
3.89499903e-01 7.25802034e-02 -5.37698627e-01 5.86593407e-04
3.54019970e-01 -3.47757310e-01 -8.90518501e-02 -6.24139886e-03
-1.22429121e+00 6.97343111e-01 3.22609872e-01 1.03855443e+00
-4.84276175e-01 -3.41549702e-02 6.30648553e-01 4.80350047e-01
7.00334013e-01 2.94373453e-01 -1.44434333e-01 -1.50512755e-01
-7.71201968e-01 1.12515867e-01 8.10722172e-01 1.68853670e-01
1.16059914e-01 -3.30430537e-01 -1.38199311e-02 8.12779725e-01
5.23550391e-01 5.24380982e-01 3.44006538e-01 -5.90528548e-01
3.69098112e-02 6.27927721e-01 -2.30703637e-01 -1.72446179e+00
-5.86004078e-01 -5.95510483e-01 -6.81544840e-01 1.85136408e-01
8.15310299e-01 -4.30271551e-02 -7.20593780e-02 1.77324951e+00
4.97667909e-01 -4.75195795e-01 -4.09058124e-01 1.14122927e+00
7.98286378e-01 -1.04849726e-01 3.78968924e-01 -2.36417815e-01
1.84103858e+00 1.83893070e-01 -5.15480876e-01 -1.23894818e-01
3.14438343e-02 -9.43470836e-01 1.03318441e+00 6.48219347e-01
-9.85766470e-01 -6.79695904e-01 -9.80646968e-01 2.56241888e-01
-5.10292351e-01 -8.83062556e-03 8.33125532e-01 1.06660140e+00
-8.45745862e-01 8.57936084e-01 -4.62654352e-01 -7.02497005e-01
4.66887742e-01 2.15590551e-01 -3.97177190e-01 3.36659044e-01
-1.29288709e+00 7.65483737e-01 -1.52825505e-01 4.76982772e-01
-2.64809519e-01 -3.36785167e-01 -5.52303970e-01 -4.33652699e-01
2.79299796e-01 -5.47674358e-01 6.02261126e-01 -1.37463653e+00
-1.40898895e+00 1.64910126e+00 3.37787300e-01 -4.46795344e-01
1.06861103e+00 -1.39632776e-01 -7.45747089e-01 2.25206196e-01
-9.27695557e-02 2.62071282e-01 9.97011602e-01 -1.25986481e+00
-1.09623782e-01 -6.83775425e-01 1.14219405e-01 -3.49605978e-01
-3.23447973e-01 4.10158753e-01 3.03288940e-02 -7.02070653e-01
-1.51178716e-02 -7.58584321e-01 7.00858608e-02 -8.19183812e-02
-4.42667782e-01 -8.86365026e-02 3.26400287e-02 -8.47780228e-01
1.03394020e+00 -2.32691503e+00 2.87827760e-01 2.46489808e-01
5.77086806e-01 1.04631804e-01 3.90718371e-01 6.09563053e-01
2.66740590e-01 1.23671710e-01 -1.34717792e-01 -3.91337067e-01
5.17822027e-01 6.84151649e-02 -7.78267756e-02 1.07624042e+00
-1.58027157e-01 8.57662201e-01 -7.33801425e-01 -6.78225756e-01
4.65744048e-01 5.69833517e-01 -1.79487765e-01 -1.04925066e-01
3.54696028e-02 5.99966824e-01 -3.48137856e-01 4.44557816e-01
2.95996726e-01 1.70131072e-01 5.01529098e-01 -2.89489686e-01
-2.44043022e-01 1.16718069e-01 -8.31013799e-01 1.33510673e+00
-1.23303771e-01 8.09409082e-01 2.76035547e-01 -9.14624631e-01
9.96855319e-01 8.57506394e-02 3.92321050e-01 -8.05187523e-01
6.93453133e-01 9.95444432e-02 2.66710997e-01 -7.30440438e-01
3.26485783e-01 -3.78888607e-01 -2.97453642e-01 2.08659634e-01
-1.29458800e-01 2.28133112e-01 1.84796631e-01 -2.19628051e-01
7.15410650e-01 4.64822054e-02 2.45882750e-01 -3.11774224e-01
5.09373426e-01 -4.00097370e-01 5.03594652e-02 5.39611578e-01
-1.85316324e-01 3.91970038e-01 5.57808161e-01 -3.22525561e-01
-9.18839753e-01 -8.99469614e-01 -4.34319496e-01 1.16608334e+00
-8.57759863e-02 -1.24082305e-01 -7.89756238e-01 -4.28015172e-01
2.02839538e-01 2.00764820e-01 -8.86416316e-01 -5.32668491e-04
-4.39808726e-01 -8.44650745e-01 4.59393233e-01 -8.42650887e-03
1.69720754e-01 -1.11896193e+00 -8.15147042e-01 1.11725844e-01
2.06359774e-01 -1.21227396e+00 2.40275458e-01 -2.21719757e-01
-4.97464657e-01 -1.06556988e+00 -6.92684174e-01 -2.42874786e-01
3.19561452e-01 -3.09721738e-01 1.14881170e+00 3.68775040e-01
-3.34258735e-01 5.59680820e-01 -4.60431159e-01 -4.55139935e-01
-7.48460710e-01 -6.09112680e-02 -1.18870130e-02 6.01278782e-01
4.48167682e-01 -8.49036753e-01 -6.12481594e-01 2.53953040e-01
-7.72689819e-01 -2.05298364e-01 7.67683506e-01 8.42416435e-02
1.43156638e-02 -3.60759825e-01 3.76005113e-01 -9.07185733e-01
6.27327859e-01 -5.43188691e-01 -1.23806611e-01 -8.07907879e-02
-2.88577884e-01 -4.45931137e-01 3.95262688e-01 -4.08850193e-01
-5.72601914e-01 -2.99745351e-01 -2.55131990e-01 2.92061716e-01
-6.61704242e-01 2.14388579e-01 8.37740302e-03 -8.72046351e-02
5.12216330e-01 1.40402056e-02 1.38837799e-01 -5.98473489e-01
3.42744678e-01 6.80082798e-01 4.09302056e-01 -4.31669265e-01
5.81957042e-01 7.63764620e-01 3.85125652e-02 -1.32866502e+00
-6.77509904e-01 -2.44677946e-01 -7.50034153e-01 -9.19734538e-01
1.19077301e+00 -5.29288113e-01 -1.19417429e+00 3.81021559e-01
-8.64894390e-01 -1.41202793e-01 -4.01885748e-01 5.10561049e-01
-8.71601924e-02 6.03458226e-01 -5.28495967e-01 -1.01427281e+00
-1.35433031e-02 -5.47132075e-01 9.42489505e-01 -1.94514990e-01
-7.93257296e-01 -9.57873166e-01 7.40041807e-02 6.32937849e-01
3.16540539e-01 7.22701907e-01 5.35273552e-01 -5.75120747e-01
-7.43943453e-02 -4.90724564e-01 2.25912742e-02 3.80235136e-01
-1.20795920e-01 -1.23897530e-02 -1.20645344e+00 1.07705399e-01
2.31670752e-01 -6.54345229e-02 6.60896301e-01 2.18626171e-01
7.03104496e-01 -8.37833807e-02 2.58684345e-03 5.34069166e-02
1.27823579e+00 -2.89353281e-01 6.04515493e-01 2.71588624e-01
4.11943346e-01 1.24345481e+00 3.06343306e-02 2.72026688e-01
2.51898557e-01 7.84352243e-01 3.58128548e-01 -1.27503663e-01
-2.77381510e-01 -2.12813690e-02 4.12418932e-01 7.80432343e-01
-5.79136252e-01 9.84127745e-02 -4.73039299e-01 4.10964102e-01
-1.41075540e+00 -9.55427051e-01 -6.50703728e-01 2.15240932e+00
5.28490961e-01 5.05296588e-01 5.91136813e-01 4.19753909e-01
5.53471148e-01 4.86506969e-01 1.33260399e-01 -4.84544486e-01
-3.11880738e-01 -4.45860066e-02 4.72651660e-01 2.73218453e-01
-9.55837011e-01 3.29614550e-01 6.14402914e+00 4.77506787e-01
-1.19537449e+00 6.76554441e-02 4.18727249e-01 -9.64085460e-02
-2.49960810e-01 -4.77791905e-01 -2.93036938e-01 8.23607206e-01
8.99359345e-01 3.35279167e-01 4.26264942e-01 4.45935965e-01
3.73547614e-01 -2.26473942e-01 -1.10443771e+00 7.37026095e-01
2.01825291e-01 -7.53586888e-01 -5.89607179e-01 2.74395615e-01
-1.05618574e-01 -1.54232502e-01 9.99415964e-02 -1.01662539e-01
-1.72794208e-01 -7.59380221e-01 1.12326908e+00 8.63480687e-01
3.76121014e-01 -1.49982244e-01 6.62926137e-01 1.88768674e-02
-6.28275752e-01 -3.51651683e-02 -8.96866992e-02 -3.89112294e-01
3.80374461e-01 8.84048283e-01 -4.97099191e-01 3.81319761e-01
5.63371718e-01 4.94507313e-01 -7.36140013e-01 5.61099827e-01
-2.17007414e-01 7.06284463e-01 -2.61930466e-01 -4.37627882e-01
-9.45298299e-02 -2.22103179e-01 7.67632246e-01 1.37583911e+00
-7.32457712e-02 -1.95604548e-01 -3.54443938e-01 9.20622170e-01
3.44484597e-01 4.16257709e-01 -6.18396997e-01 -2.97212332e-01
-1.22625805e-01 1.39981210e+00 -1.10942078e+00 -2.77343187e-02
-3.21063697e-01 8.11545789e-01 5.71379438e-02 -2.11211815e-01
-7.12247729e-01 1.98188201e-01 4.48416620e-01 5.30091584e-01
1.21244870e-01 -3.28052849e-01 -3.64401489e-01 -8.51852000e-01
2.87377626e-01 -7.62140632e-01 6.86087608e-02 -4.27035183e-01
-1.50306761e+00 4.60832089e-01 -1.28775969e-01 -7.97280133e-01
1.89313650e-01 -6.20027602e-01 -1.28885612e-01 4.54009742e-01
-1.01789331e+00 -1.28278637e+00 -2.06551060e-01 5.05114011e-02
-2.71256138e-02 7.90416449e-02 7.15694904e-01 5.46368003e-01
-3.13480943e-01 1.90511718e-01 -3.58551204e-01 1.48482904e-01
4.70404357e-01 -1.28242099e+00 1.11045957e-01 4.43403751e-01
2.87124783e-01 6.18512690e-01 1.11770284e+00 -5.04395008e-01
-1.37390900e+00 -2.51982003e-01 1.13533270e+00 -7.14710414e-01
9.33829844e-01 -7.04537511e-01 -5.16612053e-01 1.54089794e-01
2.16290206e-01 -3.43427241e-01 1.00973439e+00 4.62822080e-01
-1.62532926e-01 -2.39010155e-01 -1.14266467e+00 4.54874188e-01
1.29598236e+00 -6.43397331e-01 -7.30828583e-01 1.05648756e-01
4.05464292e-01 1.31902888e-01 -1.22540963e+00 1.51239872e-01
1.12784624e+00 -1.57773852e+00 1.12670028e+00 -2.68646866e-01
5.74161887e-01 1.08416915e-01 -2.07821652e-01 -8.22586775e-01
-2.28978425e-01 -3.99262697e-01 3.92507941e-01 1.58042920e+00
2.97012031e-01 -6.89606845e-01 6.42425179e-01 5.66884816e-01
9.33238268e-02 -4.67947870e-01 -1.02648771e+00 -3.05399150e-01
-2.73177773e-01 -9.02040184e-01 4.90730107e-01 1.15071201e+00
-2.00322829e-03 1.53374240e-01 -4.74259436e-01 -1.43736124e-01
5.65470755e-01 1.10577082e-03 8.18500876e-01 -1.50358689e+00
-6.21137500e-01 -6.65945649e-01 -7.91996241e-01 -2.83526629e-01
-8.88085142e-02 -7.24393368e-01 -4.65493053e-01 -1.09356689e+00
-2.16783416e-02 -2.62476087e-01 -1.88130066e-01 -8.64383876e-02
2.63452202e-01 7.96318471e-01 4.09625888e-01 -3.68770845e-02
-4.66519594e-01 -3.48629756e-03 1.21380520e+00 -8.37304220e-02
4.18214016e-02 4.45398502e-02 -9.51811671e-01 1.01057816e+00
5.60136378e-01 -1.75736725e-01 9.27980617e-02 -7.68448114e-02
8.63841295e-01 -2.09060937e-01 8.58051658e-01 -1.03712690e+00
-2.88274705e-01 2.01576948e-01 5.53008378e-01 -3.04515529e-02
3.80170971e-01 -1.25297964e+00 3.28211904e-01 5.32815099e-01
-4.32840884e-01 -4.04364169e-01 -8.97218809e-02 4.70134079e-01
7.57296830e-02 -1.14633098e-01 4.40314502e-01 -1.92803562e-01
-4.21774089e-01 -2.75904179e-01 -3.78238827e-01 -2.50710696e-01
8.45396340e-01 -2.96780080e-01 -1.43025935e-01 -3.37187290e-01
-1.06662250e+00 -3.20850343e-01 4.41293687e-01 4.19949204e-01
1.45916998e-01 -1.02119267e+00 -7.57413924e-01 3.01442435e-03
-3.91116776e-02 -7.58573830e-01 4.98464227e-01 1.22079980e+00
-3.21773916e-01 4.81706224e-02 -2.51613170e-01 -5.65525174e-01
-1.58302748e+00 6.48765266e-01 6.96089193e-02 1.55242622e-01
-5.88364303e-01 5.17410278e-01 -2.65706897e-01 -3.74523811e-02
5.91520481e-02 -4.87698764e-01 -5.21672010e-01 7.94110715e-01
3.08047652e-01 4.29446787e-01 -1.02017164e-01 -9.60851789e-01
-2.57873684e-01 6.82807207e-01 4.57006454e-01 -8.77129063e-02
1.39824021e+00 -3.21291238e-01 -3.48433137e-01 9.46451545e-01
1.03707397e+00 6.20958388e-01 -6.44053936e-01 1.25285760e-01
8.93681347e-02 -5.55684328e-01 -3.36180955e-01 -8.09861064e-01
-7.87792087e-01 6.29219592e-01 7.02104628e-01 1.25363207e+00
8.17535818e-01 3.75084221e-01 5.14143884e-01 -2.37957329e-01
4.43468332e-01 -1.16049707e+00 -3.56947154e-01 -3.35137099e-01
1.07229543e+00 -1.19385684e+00 1.83721483e-01 -5.06504774e-01
-4.53219980e-01 9.90549803e-01 -2.24522606e-01 -1.88850954e-01
5.24011731e-01 2.34663170e-02 2.06085201e-03 -5.75139284e-01
-1.62822649e-01 -5.29391468e-01 4.41758811e-01 5.54185808e-01
6.61136985e-01 4.83306795e-01 -9.94345844e-01 6.75564349e-01
-5.84224761e-01 -2.85000987e-02 1.48571596e-01 3.78918171e-01
-3.27245355e-01 -1.16497576e+00 -4.75820154e-01 1.45652011e-01
-9.28640604e-01 3.26703697e-01 -1.00886941e+00 9.85736787e-01
6.75498188e-01 9.79564428e-01 -1.12321317e-01 -4.66023803e-01
4.76038069e-01 4.32025529e-02 7.75914192e-01 -2.44229451e-01
-9.44993556e-01 -5.45757599e-02 6.56756461e-01 -4.62776542e-01
-1.03302610e+00 -9.22993660e-01 -4.83392656e-01 -3.56793165e-01
1.57897651e-01 -1.09616332e-01 9.40688252e-01 1.12267303e+00
-8.36454481e-02 5.83163917e-01 3.81280243e-01 -8.72414351e-01
-5.08204177e-02 -8.93950820e-01 -7.11720943e-01 6.67095840e-01
2.56767780e-01 -4.62383181e-01 -2.88350791e-01 -4.39721756e-02] | [11.478618621826172, 0.30778729915618896] |
66ce3863-cf65-4998-857e-0de880768e59 | large-language-models-struggle-to-learn-long | 2211.08411 | null | https://arxiv.org/abs/2211.08411v1 | https://arxiv.org/pdf/2211.08411v1.pdf | Large Language Models Struggle to Learn Long-Tail Knowledge | The internet contains a wealth of knowledge -- from the birthdays of historical figures to tutorials on how to code -- all of which may be learned by language models. However, there is a huge variability in the number of times a given piece of information appears on the web. In this paper, we study the relationship between the knowledge memorized by large language models and the information in their pre-training datasets. In particular, we show that a language model's ability to answer a fact-based question relates to how many documents associated with that question were seen during pre-training. We identify these relevant documents by entity linking pre-training datasets and counting documents that contain the same entities as a given question-answer pair. Our results demonstrate strong correlational and causal relationships between accuracy and relevant document count for numerous question answering datasets (e.g., TriviaQA), pre-training corpora (e.g., ROOTS), and model sizes (e.g., 176B parameters). Moreover, we find that while larger models are better at learning long-tail knowledge, we estimate that today's models must be scaled by many orders of magnitude to reach competitive QA performance on questions with little support in the pre-training data. Finally, we show that retrieval-augmentation can reduce the dependence on relevant document count, presenting a promising approach for capturing the long-tail. | ['Colin Raffel', 'Eric Wallace', 'Adam Roberts', 'Haikang Deng', 'Nikhil Kandpal'] | 2022-11-15 | null | null | null | null | ['triviaqa'] | ['miscellaneous'] | [-3.12167406e-01 -4.66843173e-02 -2.17996672e-01 -2.33246848e-01
-1.24892783e+00 -1.04835606e+00 6.74542665e-01 7.48186231e-01
-6.77427173e-01 7.77315795e-01 3.26332629e-01 -7.78286934e-01
-5.17289162e-01 -1.08539307e+00 -1.04375994e+00 -2.13613990e-03
5.14862649e-02 7.52222598e-01 5.91923296e-01 -5.95671773e-01
3.42425615e-01 1.01468913e-01 -1.22376001e+00 5.06420672e-01
1.08276927e+00 9.00517642e-01 2.32045829e-01 7.38444507e-01
-7.33168542e-01 9.45956171e-01 -7.35062659e-01 -8.44828546e-01
-2.24681213e-01 -2.15257108e-01 -1.23338485e+00 -3.90517741e-01
9.58796978e-01 -1.71699077e-01 -5.01742244e-01 6.39832139e-01
2.09129706e-01 1.79226696e-02 8.69071484e-01 -8.81556213e-01
-8.01583707e-01 6.59258008e-01 -3.86168033e-01 8.37827086e-01
4.61546898e-01 -1.99750345e-02 1.40350270e+00 -8.03971708e-01
6.30173147e-01 1.13624656e+00 5.87702334e-01 2.27885395e-01
-9.19959068e-01 -5.92755497e-01 -1.93536967e-01 4.04997587e-01
-1.07716894e+00 -2.18755081e-01 2.51745552e-01 -5.00178576e-01
9.57921624e-01 2.77313203e-01 2.95372277e-01 7.86466777e-01
3.74362133e-02 6.00963295e-01 8.11864793e-01 -6.98893666e-01
-1.20292313e-01 3.86767715e-01 6.81035817e-01 7.60482907e-01
4.99913722e-01 -3.93788576e-01 -6.62557065e-01 -3.05579871e-01
3.12645376e-01 -3.83444101e-01 -1.91219598e-01 -8.12810436e-02
-7.72483528e-01 1.00331199e+00 3.17492574e-01 5.54963589e-01
-2.16733426e-01 7.99989104e-02 3.26242357e-01 4.43228245e-01
2.53481060e-01 8.19494843e-01 -9.52872455e-01 -8.64582509e-02
-7.63849616e-01 4.03536618e-01 1.23970330e+00 8.21418464e-01
9.08243835e-01 -6.83580339e-01 -1.84585899e-01 9.58326340e-01
8.07057098e-02 5.39506316e-01 5.97971857e-01 -1.12588155e+00
9.08461869e-01 6.57567084e-01 1.44389272e-01 -1.03301919e+00
-2.72464156e-01 -6.63964272e-01 -3.24086070e-01 -6.69441521e-01
1.00139034e+00 -1.94337219e-01 -5.21056831e-01 1.82981956e+00
1.47562832e-01 -2.91033566e-01 2.35502142e-02 4.73829418e-01
8.59240472e-01 7.92824209e-01 2.67962605e-01 -1.62459128e-02
1.71716225e+00 -7.55089998e-01 -5.32905996e-01 -4.74254131e-01
8.62821877e-01 -8.16630602e-01 1.37365091e+00 1.11026324e-01
-1.12239420e+00 -4.70277071e-01 -5.94818652e-01 -3.63322735e-01
-6.87604010e-01 -3.12253181e-02 6.20912135e-01 6.51813388e-01
-8.12316775e-01 3.92063409e-01 -1.66517645e-01 -5.76757610e-01
1.38702273e-01 -1.39203295e-01 -1.09583378e-01 -3.49373847e-01
-1.50061369e+00 1.12834334e+00 3.73603225e-01 -5.33701301e-01
-5.03948212e-01 -9.20545340e-01 -4.73651737e-01 4.11787987e-01
4.69177574e-01 -8.92426372e-01 1.34280932e+00 -5.39036632e-01
-7.09803164e-01 8.58606875e-01 -3.36792141e-01 -4.87858951e-01
9.61701348e-02 -3.00156802e-01 -3.60874176e-01 3.25294286e-01
1.73864022e-01 3.63781631e-01 5.26131988e-01 -1.08748281e+00
-7.94849992e-01 -4.14966285e-01 5.55712521e-01 1.24548323e-01
-7.74010003e-01 5.17803542e-02 -6.96945608e-01 -3.42398345e-01
8.07325989e-02 -6.57046437e-01 2.33001009e-01 -2.79341042e-01
1.04946992e-03 -6.53798819e-01 3.19145888e-01 -9.68689382e-01
1.50400162e+00 -1.79679430e+00 -2.45928988e-01 2.74773449e-01
2.35496417e-01 1.14801377e-01 -4.22925860e-01 8.07184994e-01
1.92358688e-01 4.09080118e-01 9.10788849e-02 1.60478458e-01
8.35298598e-02 2.93655276e-01 -6.69651985e-01 3.08579635e-02
-1.11978315e-02 1.04926825e+00 -8.61386061e-01 -6.50538027e-01
-4.79049206e-01 1.80047691e-01 -5.58303535e-01 1.24933660e-01
-6.45545185e-01 -4.74186651e-02 -4.73447114e-01 3.35012585e-01
2.21543387e-01 -7.93125749e-01 3.28671709e-02 4.50027958e-02
3.51388872e-01 8.61255229e-01 -8.30586970e-01 1.40639853e+00
-7.08499074e-01 7.48920083e-01 -1.46085531e-01 -6.22489512e-01
5.59896708e-01 1.49321467e-01 6.47295192e-02 -9.54589069e-01
-1.61742732e-01 3.45607191e-01 1.63569674e-01 -8.02947640e-01
7.73678303e-01 -1.28458098e-01 1.98375508e-01 6.29913628e-01
2.23759472e-01 -9.87344533e-02 8.77256691e-01 7.25912631e-01
1.20821905e+00 -4.45192039e-01 -9.08197314e-02 -1.36455536e-01
5.35978377e-01 1.86175629e-01 -1.08592257e-01 9.99268889e-01
2.60637790e-01 1.14840053e-01 6.20974243e-01 -4.46990058e-02
-9.22628164e-01 -8.65239978e-01 -3.47884715e-01 1.56330121e+00
-1.34386331e-01 -6.39295101e-01 -5.88730812e-01 -7.40983427e-01
3.10523391e-01 9.96321976e-01 -4.88502771e-01 -1.70143366e-01
-6.23354256e-01 -4.78301048e-01 5.12070596e-01 4.30683702e-01
2.11597279e-01 -8.68862748e-01 -2.27547929e-01 5.30265085e-02
-5.48426151e-01 -1.09621334e+00 -2.88491577e-01 7.33588338e-02
-9.96158540e-01 -1.19794905e+00 -6.93127155e-01 -6.70325935e-01
5.33087194e-01 2.23097488e-01 1.89195335e+00 3.89661103e-01
5.91295809e-02 9.41139519e-01 -3.89117986e-01 -3.94895643e-01
-4.08870339e-01 5.94541550e-01 -4.45567608e-01 -5.63824594e-01
5.38246036e-01 -3.69720191e-01 -5.03035247e-01 2.17870072e-01
-1.07731938e+00 -4.67340827e-01 6.42767489e-01 6.16244853e-01
2.04973847e-01 3.22893523e-02 7.52860606e-01 -1.03646994e+00
9.51127470e-01 -9.04854655e-01 -4.59125131e-01 8.23458374e-01
-6.51920021e-01 2.76558876e-01 4.54743832e-01 -4.28861916e-01
-9.13633108e-01 -7.78537214e-01 -1.08770140e-01 2.37406120e-01
6.65385723e-02 1.04983997e+00 3.01174641e-01 1.33827180e-01
1.03893054e+00 7.20965192e-02 -3.65202665e-01 -6.14819705e-01
5.82321107e-01 4.39964354e-01 4.08010930e-01 -9.82519686e-01
9.22783196e-01 2.75450796e-02 -2.41583943e-01 -7.64482260e-01
-1.35142350e+00 -8.11910272e-01 -3.39085490e-01 7.56004453e-02
5.00580609e-01 -8.88498306e-01 -6.99004769e-01 3.82769629e-02
-1.15939987e+00 -1.73470810e-01 -2.74233222e-01 3.54091018e-01
-8.01048800e-02 3.63008946e-01 -7.52201796e-01 -6.25949562e-01
-3.28956127e-01 -4.16048288e-01 5.97455084e-01 3.36428255e-01
-3.12285095e-01 -1.17950892e+00 2.70079613e-01 8.38072240e-01
5.04349709e-01 -4.47742730e-01 1.70082891e+00 -9.41625357e-01
-7.21619189e-01 -2.66431272e-01 -3.24545205e-01 3.32389057e-01
-1.35309413e-01 -2.76489854e-01 -6.75751507e-01 -2.27784634e-01
-6.23515286e-02 -6.80865645e-01 9.84530807e-01 7.93395415e-02
1.04059100e+00 -4.47780818e-01 -1.43323556e-01 -1.75033882e-02
1.36130023e+00 -1.88462347e-01 5.02180576e-01 4.09150660e-01
3.29006732e-01 8.61843705e-01 3.54221284e-01 1.84954144e-02
7.40274012e-01 3.90874356e-01 -1.33368364e-02 3.44123691e-01
-2.41234869e-01 -5.03421664e-01 1.25192031e-01 1.20681894e+00
2.08584368e-01 -3.29183340e-01 -1.24212301e+00 7.94080615e-01
-1.36600375e+00 -8.59797776e-01 -4.15151566e-01 2.15017509e+00
1.36118841e+00 2.65486181e-01 6.28476068e-02 -1.86596245e-01
2.42437497e-01 -2.29594577e-03 -4.34788316e-01 -5.17152213e-02
-2.33814940e-01 4.46396649e-01 4.44989026e-01 6.74304068e-01
-6.17013335e-01 7.63179243e-01 6.47108173e+00 9.80513453e-01
-6.05040193e-01 2.05305398e-01 4.67110485e-01 2.05482587e-01
-6.95182562e-01 7.87624270e-02 -9.94321823e-01 4.43793744e-01
1.34227693e+00 -5.14020920e-01 2.34888747e-01 5.68805635e-01
-4.00279999e-01 -4.32869732e-01 -1.01919043e+00 6.01199627e-01
3.16208690e-01 -1.28707790e+00 3.11337799e-01 -6.51519373e-02
8.33512783e-01 -2.36284100e-02 8.68860856e-02 7.80764580e-01
3.84911299e-01 -8.85665655e-01 4.83122706e-01 5.31741917e-01
4.26397175e-01 -5.87552369e-01 6.64361060e-01 7.85225689e-01
-6.53026521e-01 -2.40373269e-01 -5.95478594e-01 1.58869222e-01
-1.03249811e-01 6.72174811e-01 -8.88622463e-01 3.73683542e-01
6.26165152e-01 5.23439981e-02 -1.20721936e+00 1.04085207e+00
-3.10773939e-01 9.73486781e-01 -3.49607825e-01 -3.31840098e-01
1.12052366e-01 -9.32304636e-02 4.56555262e-02 1.11664140e+00
1.84309050e-01 3.22231501e-01 -4.35004354e-01 6.82901263e-01
-5.80881774e-01 3.33855867e-01 -3.64288181e-01 -2.57767916e-01
4.93462026e-01 1.06245017e+00 -2.77009815e-01 -5.39106965e-01
-6.97122395e-01 4.77024615e-01 6.52870476e-01 4.91172194e-01
-6.67568505e-01 -4.39423770e-01 1.10163406e-01 4.96039927e-01
3.14677417e-01 -3.57220501e-01 -1.51847377e-01 -1.09911430e+00
2.50529408e-01 -1.09952843e+00 7.34239161e-01 -9.65544283e-01
-1.58288145e+00 2.27745965e-01 -5.15605211e-02 -4.77850586e-01
-4.40233678e-01 -5.66160738e-01 -3.09786022e-01 1.05271339e+00
-1.79018211e+00 -8.05821478e-01 -1.38732105e-01 5.75443864e-01
2.13499844e-01 1.03997774e-02 7.08024144e-01 5.71839750e-01
-1.06796078e-01 5.82685113e-01 3.96480680e-01 3.54922503e-01
8.90457034e-01 -1.30378175e+00 2.85899460e-01 4.28068250e-01
7.19113410e-01 1.06784546e+00 5.59770048e-01 -5.95484316e-01
-1.36317599e+00 -6.11184716e-01 1.43643856e+00 -1.17351484e+00
1.03801441e+00 -3.82006764e-02 -1.24677551e+00 7.49953151e-01
3.06424022e-01 -3.40540439e-01 7.54606903e-01 7.45182216e-01
-8.18179250e-01 -1.78657636e-01 -7.06840098e-01 3.87763262e-01
7.82296717e-01 -1.02424967e+00 -1.10505807e+00 5.73368132e-01
7.45240510e-01 -2.71193981e-01 -7.81822085e-01 1.80542737e-01
3.71680409e-01 -6.28453255e-01 9.71830785e-01 -1.08263743e+00
6.68696165e-01 8.81761014e-02 -1.07243605e-01 -1.16879356e+00
-1.48433417e-01 -1.25733718e-01 -4.48822707e-01 1.32467544e+00
8.98198247e-01 -4.30608630e-01 7.69384682e-01 7.66092658e-01
2.28585854e-01 -7.43530154e-01 -7.21356571e-01 -8.28887105e-01
6.37644053e-01 -2.55842745e-01 3.37637544e-01 1.00742269e+00
-1.63432732e-01 8.22950900e-01 3.71076405e-01 2.67251525e-02
3.21053147e-01 1.14515387e-01 7.65107989e-01 -1.26491344e+00
-3.11688542e-01 -2.58097678e-01 2.24278957e-01 -1.39383316e+00
1.35391146e-01 -1.09238493e+00 -4.09292907e-01 -1.74548066e+00
4.65738446e-01 -5.85131347e-01 -1.79733723e-01 2.61457324e-01
-4.99841601e-01 -2.70318478e-01 6.60825223e-02 2.51001507e-01
-8.21847320e-01 1.88368008e-01 1.20322514e+00 -1.00406650e-02
1.08043745e-01 -1.32314339e-01 -8.34797740e-01 6.51128411e-01
6.62953496e-01 -6.45650566e-01 -4.65767980e-01 -6.81326091e-01
9.11015928e-01 2.53986150e-01 2.61325300e-01 -8.75628293e-01
5.91471136e-01 2.68770196e-02 2.77069241e-01 -6.14464343e-01
9.26634893e-02 -8.16748500e-01 -4.81156468e-01 2.28839412e-01
-7.30803132e-01 3.14207852e-01 3.34972173e-01 6.22773051e-01
-3.06369752e-01 -7.80287325e-01 3.92102182e-01 -1.84720203e-01
-4.23769414e-01 5.40637784e-02 -1.62840262e-01 1.03387511e+00
2.72171289e-01 3.32579702e-01 -8.01240385e-01 -6.85529053e-01
-3.05033237e-01 3.78522366e-01 6.87260255e-02 4.31305021e-01
2.46685132e-01 -1.01899159e+00 -7.62055099e-01 -2.75252879e-01
1.76203772e-01 -2.78621733e-01 1.90798000e-01 6.03304565e-01
-3.40007722e-01 8.87907565e-01 2.00049847e-01 -2.52981484e-01
-1.01311791e+00 3.47666562e-01 1.12451635e-01 -7.43008673e-01
-4.08284627e-02 9.46111321e-01 -9.20534804e-02 -4.67187464e-01
3.31308633e-01 -3.39295864e-01 -2.82865882e-01 4.59846914e-01
5.15409946e-01 3.34256172e-01 2.63789058e-01 -1.78189412e-01
-4.04441133e-02 5.48000336e-01 -3.12293470e-01 -1.40589476e-01
1.19985521e+00 -1.04331732e-01 -2.99871147e-01 4.92601037e-01
1.32619059e+00 3.76846373e-01 -6.63133979e-01 -5.68247259e-01
4.78806764e-01 -3.01982969e-01 -2.80588269e-01 -1.23614442e+00
-7.03298450e-01 9.58022535e-01 2.46712729e-01 3.51840466e-01
6.67231500e-01 4.19417381e-01 1.00892258e+00 1.01913488e+00
3.40706527e-01 -1.06486940e+00 4.36870158e-01 9.11231875e-01
7.96796978e-01 -1.14687860e+00 9.26291123e-02 -2.10588902e-01
-2.56967843e-01 8.66963685e-01 5.73827744e-01 3.24405491e-01
4.95115995e-01 -2.57876396e-01 -9.89916269e-03 -4.95723248e-01
-9.38622057e-01 -2.49850988e-01 5.86294472e-01 1.69696599e-01
4.77203906e-01 -2.57583052e-01 -3.68874073e-01 6.46315336e-01
-5.53029358e-01 -2.01793000e-01 3.52947772e-01 6.82308674e-01
-7.76968598e-01 -9.26167250e-01 -3.46956104e-01 7.75212288e-01
-7.34169006e-01 -4.85545516e-01 -3.51749271e-01 8.56515348e-01
-1.09152690e-01 1.03384483e+00 -8.94175097e-03 -1.30956760e-02
4.56243426e-01 6.42737567e-01 5.94233632e-01 -7.43551314e-01
-7.83512056e-01 -5.49833059e-01 3.66527706e-01 -1.61189869e-01
-7.04065636e-02 -3.48714888e-01 -1.13066912e+00 -5.00837803e-01
-5.72144091e-01 5.86831629e-01 5.89321613e-01 9.28683341e-01
2.57069498e-01 2.63054103e-01 1.12861469e-01 3.53734553e-01
-9.52994525e-01 -1.13434136e+00 -4.91909713e-01 5.67189395e-01
1.69184223e-01 -3.04518849e-01 -6.12605333e-01 1.94761474e-02] | [11.021828651428223, 8.00523567199707] |
c095ce22-c931-46d4-8a04-b08bee76eaa1 | loss-attitude-aware-energy-management-for | 2301.07789 | null | https://arxiv.org/abs/2301.07789v1 | https://arxiv.org/pdf/2301.07789v1.pdf | Loss Attitude Aware Energy Management for Signal Detection | This work considers a Bayesian signal processing problem where increasing the power of the probing signal may cause risks or undesired consequences. We employ a market based approach to solve energy management problems for signal detection while balancing multiple objectives. In particular, the optimal amount of resource consumption is determined so as to maximize a profit-loss based expected utility function. Next, we study the human behavior of resource consumption while taking individuals' behavioral disparity into account. Unlike rational decision makers who consume the amount of resource to maximize the expected utility function, human decision makers act to maximize their subjective utilities. We employ prospect theory to model humans' loss aversion towards a risky event. The amount of resource consumption that maximizes the humans' subjective utility is derived to characterize the actual behavior of humans. It is shown that loss attitudes may lead the human to behave quite differently from a rational decision maker. | ['Pramod K. Varshney', 'Makan Fardad', 'Tianyun Zhang', 'Chen Quan', 'Baocheng Geng'] | 2023-01-18 | null | null | null | null | ['energy-management'] | ['time-series'] | [ 1.36615574e-01 3.27192783e-01 -2.40668803e-01 -4.06205565e-01
-4.74899650e-01 -3.03677082e-01 -1.30837172e-01 1.62421465e-01
-1.01698422e+00 7.22624779e-01 -8.72483552e-02 -2.93002367e-01
-8.49633962e-02 -9.71011519e-01 -9.58996490e-02 -6.74970567e-01
-9.60491076e-02 -6.68354258e-02 -1.98553175e-01 1.51207700e-01
2.57604241e-01 3.80272090e-01 -8.17375958e-01 -8.27163696e-01
6.50600433e-01 1.25274801e+00 5.02854623e-02 5.58234572e-01
5.76648653e-01 7.33515382e-01 -8.08785021e-01 -2.74648368e-01
4.88497585e-01 -5.46056628e-01 -2.86292583e-01 1.60992771e-01
-7.85817802e-01 -6.51355684e-01 2.49290094e-01 1.32150126e+00
7.29786813e-01 3.66932422e-01 7.85985351e-01 -1.40402901e+00
-2.36412540e-01 7.64304578e-01 -4.67892975e-01 2.85143554e-01
2.25731522e-01 3.32211643e-01 9.84367728e-01 1.18159287e-01
6.56365380e-02 1.16651237e+00 1.89350098e-02 5.13782322e-01
-1.38618171e+00 -7.17282593e-01 9.14077833e-02 5.39720282e-02
-1.30058789e+00 -4.45187986e-01 9.97861564e-01 -2.29679197e-01
5.34490049e-01 1.79052636e-01 8.18568289e-01 6.74962401e-01
6.50649190e-01 5.19455731e-01 9.77055848e-01 -5.66088676e-01
9.61899340e-01 3.82736415e-01 2.63038158e-01 2.02018872e-01
6.83482707e-01 5.17224133e-01 -5.14572918e-01 -5.80459297e-01
2.73732930e-01 -3.04523677e-01 -1.77728906e-01 -5.27957715e-02
-7.95719743e-01 8.47417414e-01 1.49592049e-02 1.91052184e-02
-1.21222186e+00 4.44568366e-01 -2.45451197e-01 2.76058167e-01
1.46698266e-01 4.60789263e-01 7.44025037e-02 -1.13932297e-01
-3.87422562e-01 3.48649234e-01 9.28639591e-01 4.81960118e-01
3.97516221e-01 3.72024745e-01 -2.63177574e-01 3.08514655e-01
7.07991302e-01 6.58816993e-01 3.04614007e-02 -1.42423713e+00
1.62909314e-01 9.65134725e-02 9.84162629e-01 -7.55275130e-01
-4.05956924e-01 -4.63102728e-01 -4.38743234e-01 3.48957896e-01
5.20734608e-01 -9.39230800e-01 -6.65102825e-02 2.06157708e+00
3.65093872e-02 -4.73217130e-01 4.95198257e-02 9.41542864e-01
-4.40605521e-01 6.40731573e-01 4.32771713e-01 -8.09417903e-01
1.55180132e+00 2.54156679e-01 -8.33881795e-01 -3.04767162e-01
8.48735869e-02 -3.78049642e-01 6.68586731e-01 5.16142130e-01
-1.40948641e+00 2.88075566e-01 -1.15620494e+00 8.43371212e-01
1.66961402e-01 -4.54196751e-01 2.12708235e-01 1.48490226e+00
-7.31496274e-01 3.95246625e-01 -6.53704107e-01 -2.16215849e-01
2.88528442e-01 1.93756506e-01 8.62267971e-01 4.55261976e-01
-1.26320767e+00 7.87616372e-01 1.23971209e-01 -2.28770614e-01
-9.96986389e-01 -6.85905695e-01 -5.44094861e-01 5.46769619e-01
6.12552404e-01 -7.42115557e-01 1.66926682e+00 -8.42382550e-01
-1.77421856e+00 4.20822501e-01 2.83923805e-01 -7.54624188e-01
6.64703369e-01 3.30249488e-01 -1.16567791e-01 1.24247827e-01
1.22620650e-01 3.58685344e-01 6.72257423e-01 -8.91247869e-01
-4.60084885e-01 -4.08217490e-01 5.43491775e-03 4.45149720e-01
-2.59105980e-01 1.74703553e-01 5.59474230e-01 -5.89939773e-01
-3.24408323e-01 -8.34883749e-01 -5.48285663e-01 -3.30692768e-01
-5.30270338e-01 -9.87413004e-02 1.54015515e-02 -1.35599136e-01
9.47795928e-01 -1.71868324e+00 -5.45547307e-01 5.22232711e-01
-3.32245119e-02 -7.77453780e-01 1.72776639e-01 -4.96127494e-02
5.71770370e-01 1.91800535e-01 5.54326847e-02 1.16566926e-01
6.59547567e-01 -2.70457029e-01 1.01859383e-01 5.29435515e-01
-1.77329108e-01 3.82102519e-01 -5.21517038e-01 -1.33061707e-01
-2.10125014e-01 -5.75071983e-02 -5.51023543e-01 3.82965207e-01
-4.69987728e-02 -2.71250427e-01 -1.12787318e+00 6.51191950e-01
4.49373424e-01 -2.19254512e-02 3.29990119e-01 2.04539925e-01
-1.78122655e-01 2.74338543e-01 -1.04336905e+00 5.18895924e-01
-3.60341460e-01 2.48355851e-01 5.66373885e-01 -8.74557316e-01
4.19738859e-01 3.20380747e-01 6.61588788e-01 -5.56118190e-01
7.47364640e-01 -7.64754936e-02 2.79186457e-01 -1.06735341e-01
5.28846264e-01 -9.34498608e-01 -5.89712799e-01 1.22709811e+00
-6.37701094e-01 -4.05711979e-02 -4.02061254e-01 2.71334797e-01
1.09847045e+00 -4.38185126e-01 8.52131844e-01 -5.18024504e-01
-1.50922403e-01 -1.68380007e-01 9.00375605e-01 8.72225106e-01
-7.57914722e-01 -4.30865169e-01 6.71192408e-01 2.75662780e-01
-5.38143873e-01 -1.06154931e+00 2.02663109e-01 1.01991880e+00
2.85642087e-01 4.86779332e-01 -7.97298849e-01 -1.87120363e-02
9.65505615e-02 1.57312226e+00 -1.07243568e-01 -3.63963604e-01
1.82395175e-01 -1.28665495e+00 1.87189057e-01 2.70763129e-01
7.99619615e-01 -9.01284873e-01 -1.77039146e+00 3.67707998e-01
-1.80862784e-01 -6.22814536e-01 -7.64585614e-01 1.18400350e-01
-3.42007250e-01 -4.00502950e-01 -6.48413479e-01 4.76115376e-01
5.51716805e-01 1.01301149e-01 1.00182712e+00 -3.47331017e-01
-3.48918736e-01 9.97634530e-01 7.32885823e-02 -1.00469220e+00
-1.14566229e-01 -5.42424738e-01 3.34003329e-01 2.57793553e-02
6.30526066e-01 -2.76285321e-01 -9.20598388e-01 3.77392083e-01
-5.40829837e-01 -5.97385168e-01 2.85845309e-01 1.77889600e-01
1.03681654e-01 7.27484882e-01 8.22431803e-01 -3.52921754e-01
1.26010180e+00 -6.63309038e-01 -8.93526495e-01 2.38169774e-01
-8.97251487e-01 -2.81752869e-02 2.37498820e-01 -5.67580760e-01
-1.65870881e+00 -1.37621209e-01 5.20024121e-01 3.10208052e-01
5.46277463e-02 8.30686241e-02 -7.64038086e-01 1.33703426e-01
3.85969460e-01 -1.08731225e-01 -3.33537668e-01 -1.62009776e-01
1.16872609e-01 5.91717184e-01 1.43319473e-01 -8.30247819e-01
5.01205206e-01 3.30871828e-02 1.66749939e-01 -8.53104413e-01
-3.71470809e-01 3.54883038e-02 1.79929674e-01 -5.86836278e-01
9.24086630e-01 -5.97528040e-01 -1.56246829e+00 2.89056748e-01
-8.58345389e-01 -2.09830433e-01 -4.16362673e-01 7.18157053e-01
-8.86075199e-01 2.17400655e-01 -1.19499542e-01 -1.81618059e+00
-7.90330619e-02 -7.17244864e-01 1.94575772e-01 5.14535129e-01
-5.22588074e-01 -6.52834058e-01 -3.32988381e-01 3.58218879e-01
3.63817096e-01 2.33409345e-01 7.44138122e-01 -3.17213595e-01
-9.83051896e-01 -1.86663732e-01 3.36148143e-01 -1.05091073e-01
1.28559107e-02 -5.07874191e-01 -6.09833658e-01 -1.32396132e-01
6.01961911e-01 -2.40143880e-01 1.65088102e-01 9.36072409e-01
4.68479425e-01 -7.79528439e-01 -1.94649979e-01 -2.17193905e-02
1.32022166e+00 8.88530314e-01 1.66405305e-01 2.72228330e-01
-4.11903113e-01 1.06756544e+00 5.75259209e-01 1.12286425e+00
3.76344889e-01 5.37574351e-01 3.13372910e-01 6.37495875e-01
9.49655592e-01 -1.63571090e-01 6.48292959e-01 -2.66255528e-01
1.52058914e-01 -7.16411173e-01 -4.81448859e-01 2.90298820e-01
-1.50589383e+00 -9.61237788e-01 5.76020777e-01 2.52109909e+00
4.68826771e-01 4.24835473e-01 6.89549088e-01 5.74315973e-02
7.89072812e-01 -1.32307127e-01 -9.88577545e-01 -4.81561571e-01
2.13574901e-01 -3.93667430e-01 1.14169955e+00 6.75093114e-01
-3.66776198e-01 2.25673303e-01 6.67898703e+00 5.58183968e-01
-4.12886858e-01 8.89611244e-02 1.04312086e+00 -5.36866546e-01
-5.71209729e-01 -9.54849347e-02 -6.98680639e-01 6.00704551e-01
1.28547311e+00 -1.17189288e+00 3.96914423e-01 5.79348564e-01
9.40177023e-01 -7.90234923e-01 -9.50306356e-01 5.91615915e-01
-5.20377696e-01 -6.04487777e-01 -6.59866095e-01 4.67047185e-01
3.22205666e-03 -7.48847425e-01 6.22261390e-02 -3.94702144e-02
1.12670112e+00 -5.47427356e-01 9.45507109e-01 7.64021873e-01
-5.69787696e-02 -1.03737271e+00 3.85717511e-01 6.97922945e-01
-8.31122458e-01 -3.13859731e-01 -3.19855303e-01 -5.09818852e-01
6.44679248e-01 8.77773225e-01 -8.72532785e-01 1.53917074e-02
3.36555868e-01 -3.91028285e-01 1.43119484e-01 7.64027596e-01
-1.87318757e-01 5.84150374e-01 -6.36287808e-01 -6.16073668e-01
-1.43376976e-01 -4.82518315e-01 7.34739423e-01 6.38924599e-01
6.74962580e-01 6.97063386e-01 4.81531993e-02 1.34041476e+00
-6.75561130e-02 2.69266292e-02 -3.78152192e-01 -2.52205998e-01
8.34386051e-01 9.40230072e-01 -9.40907776e-01 -1.63821161e-01
9.70816880e-04 6.21639729e-01 -5.69544077e-01 3.73376817e-01
-3.92660230e-01 -3.13763559e-01 5.92717528e-01 4.02372964e-02
-3.54599565e-01 3.21374297e-01 -5.55984318e-01 -6.47324383e-01
-2.33482733e-01 -2.16928303e-01 5.39045274e-01 -5.96963704e-01
-1.26028395e+00 -3.35475728e-02 2.91843861e-01 -8.51665914e-01
-4.97012377e-01 -2.26892233e-02 -1.05063856e+00 9.40384805e-01
-1.13308680e+00 -5.81293106e-02 3.18589985e-01 1.22630440e-01
2.58953124e-01 -3.68271880e-02 2.16694459e-01 -2.45243117e-01
-7.09336400e-01 4.31048185e-01 -2.32580468e-01 -2.78323680e-01
6.28252001e-03 -1.13565874e+00 -1.51603132e-01 8.28723848e-01
-8.07533622e-01 4.10893559e-01 1.10828185e+00 -7.86633253e-01
-1.20809686e+00 -5.22728622e-01 6.37047410e-01 1.00980766e-01
5.29999554e-01 2.30692560e-03 -2.26489156e-01 2.79096484e-01
3.05495948e-01 -6.91681862e-01 9.43796396e-01 -2.41073877e-01
2.69233882e-01 -1.84671469e-02 -1.94217861e+00 8.48013878e-01
6.19720697e-01 -9.89227295e-02 -5.11395037e-01 -1.44103557e-01
3.01349372e-01 6.40304625e-01 -6.06741309e-01 -2.07405403e-01
6.99640334e-01 -8.21072340e-01 8.45605731e-01 -2.90333442e-02
-4.26830471e-01 4.49907482e-02 -4.03249025e-01 -1.48589516e+00
-3.52839798e-01 -1.16667676e+00 4.11571950e-01 1.06313324e+00
3.54183406e-01 -8.93059134e-01 9.37232018e-01 1.58981931e+00
7.36840963e-01 -3.76504324e-02 -1.10046864e+00 -8.98267984e-01
1.10349588e-01 -4.23891366e-01 3.98912996e-01 3.25709134e-01
5.43449640e-01 2.29379103e-01 -4.16788071e-01 2.44635344e-01
1.58436120e+00 -1.36662751e-01 4.28284407e-02 -1.08196306e+00
-3.07060778e-01 -5.66419005e-01 1.05207913e-01 -8.66903365e-01
1.61046740e-02 -1.73396066e-01 2.61316121e-01 -1.01100612e+00
2.02101529e-01 -6.15093447e-02 -3.96520913e-01 1.55883348e-02
1.30143873e-02 -3.16721022e-01 5.24997652e-01 -1.26658708e-01
-3.38101178e-01 5.20675421e-01 5.75072587e-01 -1.84800133e-01
-3.85315686e-01 6.99450731e-01 -1.16427207e+00 8.32849026e-01
9.78169620e-01 -5.74143052e-01 -6.10133827e-01 4.71429348e-01
3.65742832e-01 8.09760273e-01 2.66696900e-01 -4.49740022e-01
1.38669342e-01 -9.55965281e-01 1.87389150e-01 -5.26754797e-01
3.45588684e-01 -1.02572882e+00 3.63574237e-01 1.00221479e+00
-6.09736681e-01 -2.76080459e-01 -4.46130574e-01 9.28254068e-01
5.19723237e-01 -5.51029801e-01 1.08661067e+00 -3.22285622e-01
-3.42367925e-02 -1.78432226e-01 -1.53064072e+00 -2.69295543e-01
1.30316913e+00 1.14954352e-01 -1.19931355e-01 -1.07538283e+00
-6.77036524e-01 6.76368177e-01 4.35023606e-01 -1.82577774e-01
4.03603941e-01 -1.13735235e+00 -3.70645821e-01 -4.29415584e-01
-3.80477369e-01 -9.58300650e-01 2.61916727e-01 3.79744768e-01
6.21992275e-02 2.74892926e-01 -3.05513084e-01 1.75739527e-01
-8.76491725e-01 3.60555381e-01 7.27573574e-01 -8.10718983e-02
1.73154939e-02 5.87506652e-01 1.19199669e-02 4.40553635e-01
2.38495007e-01 -1.04401901e-01 1.34559765e-01 3.33515048e-01
5.22558212e-01 1.08366346e+00 -4.72247422e-01 -9.09235179e-02
-4.04764056e-01 1.11025430e-01 4.36859041e-01 -9.09512877e-01
9.47248936e-01 -6.75758600e-01 4.26930994e-01 9.14953053e-02
4.88407761e-01 -2.33872179e-02 -1.37482464e+00 -9.53562185e-02
2.37387389e-01 -5.54394722e-01 4.86002535e-01 -8.95727217e-01
-6.80116177e-01 5.28834224e-01 5.49282312e-01 8.14155221e-01
1.29141891e+00 -4.73642141e-01 4.83083159e-01 3.87298614e-01
7.78391361e-01 -1.60543823e+00 8.33490640e-02 -4.83190358e-01
2.65504926e-01 -9.95922804e-01 5.57363629e-02 -3.05201739e-01
-8.30728233e-01 4.62562948e-01 2.53897995e-01 9.77079943e-02
9.82152104e-01 4.43083555e-01 -1.86198056e-01 2.42043123e-01
-8.56472433e-01 -2.05288216e-01 -3.33821446e-01 7.05358267e-01
8.62726495e-02 7.21217692e-01 -6.80848122e-01 1.03882015e+00
-6.24042787e-02 3.40318501e-01 1.07268882e+00 9.90875781e-01
-1.03945553e+00 -9.39853907e-01 -8.12470436e-01 4.82666194e-01
-8.33545685e-01 3.14420879e-01 -4.04311687e-01 1.78276509e-01
-1.91440776e-01 1.44627333e+00 1.86833277e-01 2.35011697e-01
4.29207116e-01 2.06388190e-01 2.32147262e-01 -4.36921805e-01
-1.03085428e-01 4.27462071e-01 2.01655492e-01 -2.84891039e-01
-3.98611546e-01 -7.29779601e-01 -1.01704061e+00 -3.62988621e-01
6.17249832e-02 3.36960047e-01 3.39891791e-01 6.52791560e-01
-1.52193934e-01 3.25912654e-01 9.84956980e-01 -4.88652080e-01
-1.16655362e+00 -5.53549111e-01 -1.26028121e+00 -1.14095956e-01
1.56929389e-01 -4.62811232e-01 -7.04956532e-01 -3.79081815e-01] | [4.306277751922607, 2.9803271293640137] |
07ba32c0-d3af-44c2-b863-94515de4444f | minimizing-safety-interference-for-safe-and | 2107.07316 | null | https://arxiv.org/abs/2107.07316v1 | https://arxiv.org/pdf/2107.07316v1.pdf | Minimizing Safety Interference for Safe and Comfortable Automated Driving with Distributional Reinforcement Learning | Despite recent advances in reinforcement learning (RL), its application in safety critical domains like autonomous vehicles is still challenging. Although punishing RL agents for risky situations can help to learn safe policies, it may also lead to highly conservative behavior. In this paper, we propose a distributional RL framework in order to learn adaptive policies that can tune their level of conservativity at run-time based on the desired comfort and utility. Using a proactive safety verification approach, the proposed framework can guarantee that actions generated from RL are fail-safe according to the worst-case assumptions. Concurrently, the policy is encouraged to minimize safety interference and generate more comfortable behavior. We trained and evaluated the proposed approach and baseline policies using a high level simulator with a variety of randomized scenarios including several corner cases which rarely happen in reality but are very crucial. In light of our experiments, the behavior of policies learned using distributional RL can be adaptive at run-time and robust to the environment uncertainty. Quantitatively, the learned distributional RL agent drives in average 8 seconds faster than the normal DQN policy and requires 83\% less safety interference compared to the rule-based policy with slightly increasing the average crossing time. We also study sensitivity of the learned policy in environments with higher perception noise and show that our algorithm learns policies that can still drive reliable when the perception noise is two times higher than the training configuration for automated merging and crossing at occluded intersections. | ['Christoph Stiller', 'Johannes Fischer', 'Marvin Busch', 'Tizian Engelgeh', 'Danial Kamran'] | 2021-07-15 | null | null | null | null | ['distributional-reinforcement-learning'] | ['methodology'] | [-6.28373027e-02 6.12780392e-01 -2.95342058e-01 -2.74817407e-01
-8.49327803e-01 -6.15787685e-01 4.77423817e-01 1.07106932e-01
-8.75445843e-01 1.25336421e+00 -2.51330763e-01 -6.97173893e-01
-2.29912296e-01 -9.49317217e-01 -9.12936568e-01 -1.07370043e+00
-3.63621682e-01 5.36091685e-01 5.24915397e-01 -5.20199418e-01
5.63310906e-02 7.73047924e-01 -1.88529134e+00 -2.76502907e-01
1.01613629e+00 8.44525218e-01 5.74695356e-02 7.64426708e-01
4.14586544e-01 8.49506140e-01 -8.66630375e-01 2.02285796e-01
5.14662743e-01 4.28796560e-02 -3.14215362e-01 -2.54484564e-01
1.80883259e-01 -5.70624173e-01 -1.07628711e-01 8.65599930e-01
5.78118205e-01 5.48072696e-01 6.06409907e-01 -1.69552636e+00
2.79785782e-01 5.09329498e-01 -5.12019694e-01 8.63757432e-02
1.00620598e-01 7.99182773e-01 3.41838926e-01 2.55213529e-01
3.34371418e-01 1.41414928e+00 1.70015350e-01 7.91276872e-01
-1.32745314e+00 -8.71927679e-01 5.29604614e-01 1.63607210e-01
-1.19930363e+00 -2.34623492e-01 5.12498260e-01 -2.73249187e-02
8.51958573e-01 -7.86997229e-02 4.02200907e-01 1.29350865e+00
4.86686051e-01 3.80867451e-01 1.21133602e+00 -9.03540254e-02
8.09759855e-01 2.86610126e-01 -4.47639108e-01 3.29315931e-01
5.16343355e-01 8.18451166e-01 1.93477899e-01 -2.50240177e-01
1.99708462e-01 -5.71229994e-01 6.39038384e-02 -4.75882947e-01
-7.52417147e-01 9.22085583e-01 1.21501558e-01 -1.62086785e-01
-5.14767408e-01 4.00168180e-01 5.14135778e-01 4.75200146e-01
-1.01783037e-01 4.71020162e-01 -4.79749203e-01 -4.35093939e-01
-3.50982726e-01 7.15630889e-01 8.30464125e-01 9.85852182e-01
5.62788963e-01 5.74391246e-01 -2.45557800e-01 4.88895774e-01
-7.14748027e-03 9.71487701e-01 -1.64097637e-01 -1.54730165e+00
5.07193744e-01 7.76495412e-03 7.65604854e-01 -7.98366368e-01
-5.53203583e-01 -2.32245743e-01 -4.23246890e-01 1.10955477e+00
4.45921987e-01 -7.60049880e-01 -5.86090863e-01 1.95983887e+00
4.69404191e-01 1.37061432e-01 3.70434850e-01 6.81076527e-01
-2.18933687e-01 7.16421366e-01 3.12523425e-01 -6.10154510e-01
6.85926378e-01 -2.29391247e-01 -7.66659975e-01 -6.36481941e-02
5.81581712e-01 -2.91634768e-01 1.09572208e+00 7.06520557e-01
-9.10674393e-01 -4.11341578e-01 -1.15224445e+00 8.46682847e-01
-1.64961398e-01 -3.49405587e-01 -3.01007018e-03 7.46861219e-01
-9.06431437e-01 3.45425785e-01 -5.37027657e-01 -4.14800972e-01
1.92102626e-01 3.94571304e-01 7.31557906e-02 3.07453662e-01
-1.30208254e+00 1.34460545e+00 4.65408981e-01 -1.83196560e-01
-1.46316743e+00 -5.86170733e-01 -7.75852621e-01 -2.61827618e-01
1.00184572e+00 -1.74595907e-01 1.48541629e+00 -6.24654412e-01
-2.01387787e+00 8.90260413e-02 2.85899371e-01 -9.23867822e-01
8.45504105e-01 -3.45090121e-01 -4.28953469e-01 2.86230296e-02
-5.02206534e-02 7.46042371e-01 6.62082553e-01 -1.73568463e+00
-1.07447815e+00 1.29500806e-01 5.44543087e-01 2.61387527e-01
1.00993931e-01 -4.53957349e-01 3.69948804e-01 -1.29393175e-01
-9.50436652e-01 -1.24683285e+00 -6.83002412e-01 -2.83230841e-01
5.07316180e-02 -2.02201903e-01 9.43793774e-01 1.32593930e-01
9.51556504e-01 -1.99146938e+00 -3.16091299e-01 5.21583855e-01
-4.80860978e-01 2.78857410e-01 -1.41799271e-01 4.78930861e-01
2.99299955e-01 -7.46017471e-02 -6.48166686e-02 2.29353771e-01
8.70205089e-02 8.21779251e-01 -7.29434431e-01 5.60820878e-01
2.76515745e-02 2.64933795e-01 -1.06487107e+00 -3.71492326e-01
5.29555798e-01 1.17508233e-01 -6.50646448e-01 6.05268061e-01
-4.96886671e-01 5.83327293e-01 -5.45645177e-01 2.20429659e-01
6.04192197e-01 8.86226475e-01 1.36289880e-01 3.39308709e-01
-4.41168547e-01 -2.94458687e-01 -1.30925369e+00 7.73218036e-01
-7.75573969e-01 4.01879340e-01 2.80557930e-01 -9.83184278e-01
9.10380900e-01 1.20434903e-01 5.61450243e-01 -1.07246482e+00
4.24018085e-01 -5.29965051e-02 1.22372076e-01 -5.79705775e-01
3.37194175e-01 -9.14427415e-02 -3.63026083e-01 2.26510778e-01
-4.71341193e-01 -5.20819604e-01 -3.65846767e-03 4.39890474e-02
1.06895459e+00 1.84780672e-01 2.46786669e-01 -4.23011035e-01
6.08515978e-01 6.50583506e-02 8.34930003e-01 1.00570107e+00
-6.19223952e-01 -4.06739473e-01 8.57975185e-01 -3.81653100e-01
-7.46017754e-01 -1.04092491e+00 1.08578242e-01 9.94090974e-01
5.13673306e-01 1.19682908e-01 -6.25769198e-01 -7.79396772e-01
2.14797243e-01 1.65925992e+00 -5.48519671e-01 -5.33465981e-01
-7.82884061e-01 -3.35867792e-01 6.52709186e-01 3.20630699e-01
3.39283288e-01 -9.82641757e-01 -1.21757936e+00 2.12206304e-01
2.46466294e-01 -1.24957728e+00 -1.32388070e-01 3.27371866e-01
-2.83633947e-01 -1.24800861e+00 1.60578806e-02 -2.84228235e-01
6.18403375e-01 4.35842089e-02 8.09661388e-01 -6.93098158e-02
1.25454009e-01 4.57951635e-01 -1.87450051e-01 -8.25164318e-01
-7.85435975e-01 -2.06057012e-01 5.11184335e-01 -2.97510445e-01
3.52299549e-02 -4.05003548e-01 -4.94144201e-01 5.89970708e-01
-8.53282869e-01 -4.40176517e-01 4.09811318e-01 6.23446941e-01
5.39864361e-01 4.53566968e-01 8.22143376e-01 -6.15456522e-01
9.68578815e-01 -4.95548099e-01 -1.30208313e+00 -8.17456841e-03
-7.40896285e-01 3.82968217e-01 1.21044374e+00 -6.22307420e-01
-1.06806505e+00 -1.10001013e-01 -1.06039792e-02 -2.54187793e-01
-5.08502960e-01 -3.96683425e-01 -2.08359420e-01 -1.68441162e-02
7.98208594e-01 -1.75092816e-01 1.70930266e-01 3.16571504e-01
4.85239029e-01 4.29098904e-01 1.61552757e-01 -1.17710710e+00
1.09835052e+00 2.91503578e-01 2.38823041e-01 -6.85857832e-01
-5.12409091e-01 9.30449888e-02 -6.98326230e-02 -6.00059569e-01
7.29489565e-01 -7.12825358e-01 -1.37639332e+00 1.03061497e-02
-7.59176433e-01 -8.87907982e-01 -4.91176903e-01 2.61009514e-01
-1.06512797e+00 6.52258024e-02 -9.44561213e-02 -1.36840689e+00
1.48048237e-01 -1.27548432e+00 8.20009053e-01 3.34790558e-01
4.99935560e-02 -7.16503143e-01 1.46057725e-01 -2.30676636e-01
4.84785587e-01 6.02184117e-01 7.80181885e-01 -3.33980739e-01
-2.49717906e-01 3.07657242e-01 4.29385513e-01 3.66270304e-01
-1.19981930e-01 1.41628966e-01 -7.44740367e-01 -5.40930152e-01
-2.70013154e-01 -6.22225523e-01 2.97943264e-01 2.82976300e-01
1.17332160e+00 -6.14010036e-01 -3.18691015e-01 1.03073455e-01
1.33399212e+00 7.84954846e-01 5.33096015e-01 5.89930654e-01
-5.97810894e-02 7.58739412e-01 1.41975808e+00 6.77462220e-01
4.60072309e-01 7.84275711e-01 8.64221454e-01 2.62318905e-02
5.28081536e-01 -3.11505258e-01 7.72398174e-01 -3.46677244e-01
6.09001555e-02 -1.94934294e-01 -6.79785788e-01 2.56242573e-01
-1.97845829e+00 -1.22788692e+00 2.56239533e-01 2.60342026e+00
7.51599252e-01 7.64420509e-01 5.16645968e-01 1.04117885e-01
3.35710734e-01 -1.17643803e-01 -7.74046063e-01 -1.14092302e+00
1.84835717e-01 -1.22244358e-01 1.12892807e+00 8.62256527e-01
-8.26470673e-01 9.58236575e-01 6.58402252e+00 7.64118075e-01
-1.08505988e+00 -1.76345706e-01 6.41887844e-01 -3.33653986e-01
-1.21524647e-01 -9.19299945e-02 -7.12544084e-01 5.03418088e-01
1.30487013e+00 -2.78360784e-01 4.81300950e-01 1.14118779e+00
1.01544976e+00 -6.22921586e-01 -9.46844339e-01 5.04298925e-01
-4.13676053e-01 -8.45059633e-01 -3.21104944e-01 -2.92674196e-03
6.15468979e-01 -1.91368580e-01 9.25987288e-02 8.43758762e-01
9.17038620e-01 -1.05738378e+00 7.90852666e-01 3.87731016e-01
4.54772562e-01 -1.67566347e+00 9.27744210e-01 7.38521159e-01
-8.47757459e-01 -5.96125484e-01 -2.65652388e-01 -1.77996662e-02
2.33520880e-01 1.64039537e-01 -9.57564235e-01 2.10979730e-01
6.51652694e-01 -7.58282840e-02 -9.91175473e-02 6.68216407e-01
-3.96297753e-01 6.60868645e-01 -4.61325258e-01 -2.92198509e-01
6.04554832e-01 -1.70208097e-01 6.58418775e-01 8.99332464e-01
7.42315277e-02 1.16428189e-01 6.20018065e-01 4.94268119e-01
6.26302421e-01 -2.59561211e-01 -1.10458708e+00 5.49867868e-01
6.68943167e-01 1.01574397e+00 -3.14835697e-01 -1.15369320e-01
1.37217775e-01 2.24034190e-01 2.52506375e-01 6.61415279e-01
-1.50982666e+00 -2.50356525e-01 1.14049327e+00 1.64642602e-01
1.07800886e-01 -3.64691079e-01 -7.41587430e-02 -1.54838741e-01
-2.28805378e-01 -8.17865908e-01 2.93914735e-01 -2.93423980e-01
-8.64193499e-01 5.60067654e-01 5.58992207e-01 -1.43963838e+00
-5.51703036e-01 -4.19209689e-01 -5.89840889e-01 3.13421905e-01
-1.60544014e+00 -4.21609938e-01 2.63814609e-02 5.35420060e-01
5.06463766e-01 -1.87716395e-01 5.12003362e-01 -3.03651728e-02
-5.70917308e-01 5.98130345e-01 -1.45640954e-01 -5.22466481e-01
7.86656916e-01 -1.09334135e+00 -2.80623555e-01 6.26092136e-01
-7.63709426e-01 1.28062531e-01 1.41965485e+00 -4.44499940e-01
-1.29519355e+00 -1.36786735e+00 -1.54165477e-01 -2.10117549e-01
5.61762691e-01 -2.49137446e-01 -6.41019881e-01 2.86555499e-01
3.56870890e-01 7.06860283e-03 1.01720937e-01 -2.68836647e-01
9.80053917e-02 -5.60396969e-01 -1.50288737e+00 1.00733793e+00
7.99487054e-01 1.19337842e-01 -2.23288894e-01 7.15790093e-02
8.42791855e-01 -3.30161065e-01 -5.21146476e-01 5.52037477e-01
2.89467305e-01 -9.24081981e-01 4.68781769e-01 -5.99781513e-01
-5.12210190e-01 -5.27069271e-01 -1.62991092e-01 -1.70875454e+00
7.14309886e-02 -9.02922809e-01 9.69787389e-02 8.62684429e-01
3.55666816e-01 -7.92863786e-01 5.30821502e-01 4.91267115e-01
-7.97299892e-02 -7.89564431e-01 -1.27439487e+00 -1.15550697e+00
3.57356727e-01 -7.17643499e-01 4.08024788e-01 1.96898803e-01
-1.97467074e-01 -6.94831386e-02 -5.16121447e-01 6.20179832e-01
8.08288157e-01 -3.53886813e-01 1.05107927e+00 -5.96139312e-01
-9.18174461e-02 -3.65692437e-01 -2.41747513e-01 -6.18169069e-01
7.32233167e-01 -1.73600078e-01 6.91820383e-01 -1.13391507e+00
-5.60119927e-01 -7.62474716e-01 -1.53995320e-01 4.32831377e-01
1.92344815e-01 -5.73042750e-01 7.01057585e-03 -5.55701554e-01
-7.62989938e-01 8.12711298e-01 9.47693408e-01 -1.86484113e-01
-3.55999857e-01 1.73405439e-01 -4.40510541e-01 5.75439215e-01
1.11881053e+00 -4.57378566e-01 -9.94197786e-01 2.66059265e-02
4.69149239e-02 2.68779278e-01 8.28238949e-03 -1.14357412e+00
-8.72291028e-02 -1.08529317e+00 -2.76047736e-01 -4.34814513e-01
9.86350924e-02 -1.30108726e+00 2.77188071e-03 7.37004042e-01
-5.09052038e-01 7.36344010e-02 6.46171331e-01 8.58358979e-01
1.25264511e-01 2.77595252e-01 1.15617788e+00 2.93679565e-01
-6.84399545e-01 8.74480009e-02 -9.17166293e-01 1.30280495e-01
1.91444147e+00 -1.19087949e-01 -2.23990545e-01 -6.83830321e-01
-9.51238349e-02 9.99869049e-01 4.10924315e-01 5.42730212e-01
5.56019127e-01 -1.09215784e+00 -5.15266061e-01 3.50539833e-02
4.03809026e-02 -1.82148412e-01 -1.54999383e-02 6.89589202e-01
-1.94238186e-01 3.00482124e-01 -3.50063592e-01 -5.80088794e-01
-9.20913815e-01 7.94622362e-01 6.31470799e-01 -1.09095663e-01
-4.39952314e-01 1.64198264e-01 5.25936186e-02 -3.16854656e-01
6.29892528e-01 -4.11856860e-01 -2.30985031e-01 -2.79739618e-01
5.07464111e-01 4.87044305e-01 -1.45018816e-01 -4.55106974e-01
-3.94495010e-01 4.27985221e-01 2.23994777e-01 -1.88284576e-01
9.17373538e-01 -8.54979530e-02 5.64995289e-01 2.77617782e-01
4.87244308e-01 6.26582354e-02 -1.89863360e+00 4.25522447e-01
4.15540673e-02 -3.65370899e-01 -1.37493357e-01 -7.60378003e-01
-9.36410904e-01 3.71413648e-01 8.16983223e-01 9.93183330e-02
9.97638106e-01 -3.44592988e-01 4.58626717e-01 5.41606605e-01
9.93148386e-01 -1.63747013e+00 1.89149491e-02 5.95507026e-01
8.22203696e-01 -1.27508843e+00 -3.04931164e-01 -7.41625950e-02
-1.09290123e+00 8.39648306e-01 1.28059125e+00 -3.48201841e-01
4.33566868e-01 7.91408539e-01 2.58626312e-01 3.70208055e-01
-1.13092995e+00 -2.66770452e-01 -4.04861689e-01 1.13332570e+00
-2.61621892e-01 2.47062117e-01 -2.98162520e-01 5.52573055e-02
-5.29936329e-02 -2.88921118e-01 8.27392697e-01 7.81768799e-01
-8.58373582e-01 -9.54963326e-01 -5.53787231e-01 7.41094053e-02
-2.50552315e-02 7.57274985e-01 1.44700974e-01 1.16070366e+00
2.60069728e-01 1.39759231e+00 -2.16653123e-02 -3.46171588e-01
7.43847251e-01 -4.21526790e-01 2.55107790e-01 -2.34407738e-01
-2.79888093e-01 -2.33812913e-01 5.03930509e-01 -1.11089313e+00
-2.02645048e-01 -4.64239359e-01 -1.82806206e+00 -3.14971477e-01
2.11972743e-01 2.75319964e-01 2.61487335e-01 8.47813487e-01
1.44867063e-01 6.31152630e-01 9.75618303e-01 -6.86818242e-01
-1.13177407e+00 -4.21291977e-01 -2.98929632e-01 6.52371868e-02
5.22401571e-01 -1.10777831e+00 -5.10765791e-01 -6.38425589e-01] | [4.871094703674316, 1.7690609693527222] |
12210fac-f8c2-4b61-baa1-00d5d99a5c4e | morphological-classification-of-astronomical | 2105.02958 | null | https://arxiv.org/abs/2105.02958v1 | https://arxiv.org/pdf/2105.02958v1.pdf | Morphological classification of astronomical images with limited labelling | The task of morphological classification is complex for simple parameterization, but important for research in the galaxy evolution field. Future galaxy surveys (e.g. EUCLID) will collect data about more than a $10^9$ galaxies. To obtain morphological information one needs to involve people to mark up galaxy images, which requires either a considerable amount of money or a huge number of volunteers. We propose an effective semi-supervised approach for galaxy morphology classification task, based on active learning of adversarial autoencoder (AAE) model. For a binary classification problem (top level question of Galaxy Zoo 2 decision tree) we achieved accuracy 93.1% on the test part with only 0.86 millions markup actions, this model can easily scale up on any number of images. Our best model with additional markup achieves accuracy of 95.5%. To the best of our knowledge it is a first time AAE semi-supervised learning model used in astronomy. | ['Sergey Gerasimov', 'Alex Meshcheryakov', 'Andrey Soroka'] | 2021-04-27 | null | null | null | null | ['morphology-classification'] | ['computer-vision'] | [-3.32681447e-01 3.92523319e-01 6.08922124e-01 -3.47992420e-01
-4.60880995e-01 -7.59112120e-01 5.59395611e-01 -3.56610212e-03
-7.66862988e-01 7.86275685e-01 -3.80697787e-01 -4.09040242e-01
1.16767786e-01 -1.30135083e+00 -8.30826342e-01 -9.67149079e-01
-2.29332209e-01 1.11055517e+00 6.04456007e-01 -1.51745498e-01
2.02099472e-01 6.38869166e-01 -1.19315648e+00 2.29020510e-02
7.29586303e-01 1.07627773e+00 4.57936488e-02 1.23088872e+00
-9.86758471e-02 1.17232633e+00 -8.50913048e-01 -1.07409203e+00
6.18922889e-01 -6.00095689e-02 -1.07380426e+00 3.79657954e-01
5.22904754e-01 -2.24205449e-01 -2.98104614e-01 1.11447716e+00
5.60837507e-01 3.41289401e-01 7.04046667e-01 -9.94601607e-01
-9.85336900e-01 4.81867373e-01 -1.56346455e-01 2.47995049e-01
-4.09834206e-01 4.37417626e-01 7.50525355e-01 -9.47415411e-01
4.63384986e-01 8.78687501e-01 8.27968419e-01 3.35957199e-01
-9.64920342e-01 -1.77531436e-01 -5.50838768e-01 6.35289401e-03
-1.27704096e+00 -3.08135718e-01 5.13626397e-01 -7.84491181e-01
1.00903904e+00 3.86815757e-01 9.70399380e-01 7.05969214e-01
7.83771574e-02 1.32706329e-01 1.34716499e+00 -2.73204714e-01
3.96532685e-01 5.05175710e-01 -1.93292424e-01 1.20485008e+00
4.07167792e-01 -3.66220996e-02 -1.21868543e-01 -3.25475842e-01
1.26412034e+00 -3.08024347e-01 2.58447677e-01 -2.85674870e-01
-9.47449684e-01 1.23000765e+00 5.01976192e-01 1.60565674e-01
2.10856065e-01 2.59213950e-02 1.56405374e-01 4.61157709e-01
6.64025605e-01 1.13751721e+00 -3.87619555e-01 -1.53856603e-02
-7.41800725e-01 1.06175371e-01 9.79076326e-01 6.25749350e-01
6.32934868e-01 4.37339753e-01 5.46049535e-01 9.12709117e-01
2.77976245e-01 2.52531022e-01 6.80690527e-01 -9.44657147e-01
-1.29078791e-01 8.88569176e-01 1.02795467e-01 -6.10244632e-01
-3.77137959e-02 -6.34081602e-01 -9.53918457e-01 8.44945192e-01
9.17693079e-01 -4.46713604e-02 -6.90079987e-01 9.54064250e-01
3.63732874e-01 6.86251745e-02 -1.80998236e-01 5.88905394e-01
1.22110808e+00 5.46781242e-01 -5.03201544e-01 2.11287990e-01
1.10070956e+00 -1.26323903e+00 6.84216321e-02 -1.14148930e-01
2.74890244e-01 -9.62619245e-01 1.08678651e+00 6.76991284e-01
-1.27402663e+00 -6.03684843e-01 -1.20010376e+00 -2.45640263e-01
-7.95678496e-01 2.51482353e-02 1.24190736e+00 9.89641488e-01
-1.10991228e+00 8.01415563e-01 -7.44528532e-01 -2.52606422e-01
5.51990330e-01 9.32061970e-01 -6.29199922e-01 5.98351300e-01
-5.49573779e-01 7.13657439e-01 3.54468584e-01 -4.00354922e-01
-1.13518882e+00 -3.84302169e-01 -5.47332585e-01 8.58472753e-03
1.61298424e-01 -6.59582198e-01 1.37513113e+00 -1.11611259e+00
-1.51900780e+00 1.49034917e+00 6.38618231e-01 -9.53624487e-01
4.42025632e-01 1.82836846e-01 -2.36796811e-01 1.33827165e-01
-3.79341632e-01 7.29376972e-01 9.36461449e-01 -9.04463291e-01
-3.74171972e-01 -2.35928521e-01 2.44957522e-01 -2.18940347e-01
-4.88671213e-01 2.72014737e-01 -2.74319351e-01 -7.82503247e-01
-2.31500998e-01 -1.10686970e+00 -2.56136358e-01 6.92888498e-02
8.36516172e-02 -2.59910077e-01 6.13030076e-01 -8.48604143e-01
4.73173171e-01 -1.90302920e+00 1.31173292e-02 -3.59472036e-01
5.08291245e-01 2.72677660e-01 3.43652844e-01 -7.18300864e-02
4.89420593e-02 1.69300467e-01 -4.06553000e-01 -4.64532822e-01
-1.33995444e-01 1.64879471e-01 -1.67045414e-01 5.72346210e-01
2.97694683e-01 1.07686138e+00 -6.00448191e-01 -4.25723612e-01
1.58351660e-01 -1.36964126e-02 -6.33127511e-01 7.14465261e-01
-1.05597489e-01 4.04658347e-01 1.96163043e-01 7.58012295e-01
6.17558300e-01 -4.23708826e-01 -4.92266119e-01 3.86676013e-01
-1.53056979e-01 -3.97527888e-02 -8.65245104e-01 1.49930024e+00
-3.30680549e-01 6.71788573e-01 1.87881365e-02 -1.10540533e+00
1.10448980e+00 6.14621900e-02 2.07847446e-01 6.52812719e-02
3.40857685e-01 2.49270931e-01 3.90579015e-01 -3.28604490e-01
5.43825269e-01 -9.87790227e-02 -1.70106173e-01 4.07658488e-01
7.46511519e-01 -4.93429154e-01 1.64555028e-01 1.75351068e-01
1.28562212e+00 -2.28984818e-01 3.77190530e-01 -2.14160994e-01
3.69356573e-01 2.02036709e-01 3.20353776e-01 9.47724879e-01
-4.08327520e-01 6.49763405e-01 3.29806417e-01 -6.16997182e-01
-1.65715837e+00 -9.35571551e-01 -2.39969231e-02 9.60551441e-01
-4.38513786e-01 8.28126371e-02 -7.62636065e-01 -8.77255976e-01
-1.95889488e-01 1.97046131e-01 -7.26986706e-01 -1.60873204e-01
-5.45018077e-01 -1.25835276e+00 7.94177532e-01 3.04286808e-01
8.91741097e-01 -1.11617088e+00 -9.03802440e-02 -7.08987191e-02
5.64209700e-01 -6.57881439e-01 9.64350812e-03 1.11140378e-01
-7.87830293e-01 -7.42635608e-01 -8.13304484e-01 -1.08731985e+00
5.67310810e-01 -1.30193442e-01 1.34566784e+00 1.64145097e-01
-4.93078887e-01 2.38427982e-01 -3.40126961e-01 -6.46372020e-01
-5.34695804e-01 -3.19377263e-03 2.43556015e-02 4.50480916e-02
1.70469645e-03 -7.81091630e-01 -6.61875367e-01 1.09450631e-01
-5.02967060e-01 -2.38555312e-01 6.56574190e-01 1.08229792e+00
4.35819119e-01 4.22312886e-01 2.55594850e-01 -1.10039639e+00
1.56394020e-01 -3.91569674e-01 -9.24950361e-01 1.34818316e-01
-5.51345646e-01 -8.90091509e-02 8.52611780e-01 -6.92459881e-01
-1.17171359e+00 2.05964163e-01 -4.41998065e-01 -1.38486400e-01
-2.50610113e-01 -6.85952231e-02 3.67235132e-02 -9.07004774e-01
8.42134595e-01 3.09332740e-02 -1.08693101e-01 -6.04538262e-01
1.77085385e-01 8.64992738e-01 8.65054071e-01 -1.59344181e-01
1.11940980e+00 6.48569465e-01 -6.50529098e-03 -6.07857168e-01
-7.51340568e-01 -5.65971196e-01 -7.12490678e-01 -1.80463478e-01
6.82326019e-01 -8.02705288e-01 -6.25127375e-01 7.34921694e-01
-7.14266956e-01 -3.63366425e-01 -5.35060525e-01 5.12490392e-01
-5.07072389e-01 4.41673577e-01 -6.94986284e-01 -5.73303580e-01
-5.63762069e-01 -9.16075230e-01 6.71300948e-01 3.32931429e-01
8.11920166e-02 -1.09114802e+00 3.18420172e-01 7.45078862e-01
4.03888851e-01 6.51006773e-03 5.64622760e-01 -9.32979167e-01
-6.52977526e-01 -3.95797342e-01 2.66716559e-03 4.10926342e-01
-1.27579749e-01 3.13516483e-02 -1.12977684e+00 -3.61564487e-01
4.96739417e-01 -8.35705340e-01 9.84109044e-01 -4.48363181e-03
1.59666073e+00 -3.54617208e-01 3.80822867e-01 7.84083843e-01
1.24726212e+00 3.39773834e-01 8.29418182e-01 3.31276596e-01
5.96166670e-01 3.66652310e-01 4.19117838e-01 2.49264777e-01
-2.72211611e-01 2.52518743e-01 4.74818081e-01 -3.15071523e-01
-2.19331354e-01 1.73765883e-01 1.01851232e-01 1.23589933e+00
-3.27562183e-01 -3.63087863e-01 -9.48047698e-01 3.34834278e-01
-1.42333865e+00 -8.85551274e-01 -8.72819349e-02 2.36159897e+00
7.66919672e-01 4.54361767e-01 2.00399712e-01 3.56856555e-01
3.92769337e-01 -2.62429804e-01 -4.80356753e-01 -4.88542169e-01
-1.59333050e-01 7.42489338e-01 5.51965594e-01 4.93545681e-01
-1.36769760e+00 8.11947703e-01 6.21728849e+00 1.18716526e+00
-1.07913506e+00 4.88404632e-01 8.98563147e-01 7.78554380e-02
2.31593877e-01 1.58262134e-01 -5.16629755e-01 5.05378425e-01
8.42862964e-01 2.30156019e-01 6.09712183e-01 1.24899733e+00
-3.81818652e-01 -1.74204007e-01 -5.77819288e-01 1.02287960e+00
1.35092795e-01 -1.35365129e+00 -1.86546013e-01 2.86600351e-01
7.10579455e-01 -2.23387498e-02 7.91042000e-02 5.09203255e-01
6.70598090e-01 -1.23653030e+00 3.92413318e-01 4.52425808e-01
7.80169427e-01 -8.85576546e-01 8.93303514e-01 4.98546511e-01
-7.70652056e-01 1.97318226e-01 -9.02000725e-01 -2.12857172e-01
-3.83924693e-01 5.05448639e-01 -1.47207856e+00 1.09303921e-01
7.75407314e-01 6.83154240e-02 -1.33978927e+00 1.31870937e+00
1.83316633e-01 1.17245722e+00 -3.14379126e-01 -3.27071041e-01
7.22929090e-02 -4.00334716e-01 3.79835010e-01 1.13903010e+00
3.80958676e-01 -2.32561026e-02 -8.30295235e-02 6.44373000e-01
-3.83598983e-01 8.83857086e-02 -5.17290056e-01 -4.58175510e-01
1.61722198e-01 1.59697330e+00 -1.13468051e+00 -6.24665320e-01
-5.14079511e-01 1.21855783e+00 9.16459084e-01 -5.01953363e-01
-7.73577690e-01 -5.29640913e-01 -1.19781651e-01 1.60253257e-01
2.78976172e-01 -4.52349305e-01 -3.86938006e-01 -1.18380821e+00
-5.01383424e-01 -7.19620705e-01 3.17583174e-01 -9.01766181e-01
-1.52832329e+00 4.60981637e-01 -6.74553216e-01 -1.16267455e+00
8.75728019e-03 -1.15211523e+00 -8.06016445e-01 2.97463119e-01
-3.46200705e-01 -1.15537262e+00 -4.41989988e-01 2.34136447e-01
4.92802531e-01 -1.14649391e+00 1.01354277e+00 2.54561841e-01
-5.81866264e-01 5.02884746e-01 4.98685539e-01 5.40247977e-01
7.12049365e-01 -1.89842594e+00 7.46460557e-01 8.42305958e-01
6.58773601e-01 4.30160761e-01 5.47843575e-01 -6.36319280e-01
-1.28124130e+00 -1.11243463e+00 6.85341954e-01 -5.83275676e-01
8.75643671e-01 -6.44305587e-01 -7.16394842e-01 6.20525837e-01
4.96737897e-01 1.70843944e-01 8.59944165e-01 -1.29271299e-01
-1.17355794e-01 3.38611417e-02 -1.42555690e+00 3.26747030e-01
9.65504646e-01 -4.79349345e-01 -6.94579542e-01 6.12238646e-01
6.83298528e-01 -1.50717020e-01 -1.17125058e+00 3.54365170e-01
1.62315831e-01 -1.11662531e+00 1.15275717e+00 -6.16248846e-01
4.05655831e-01 -1.19059764e-01 2.33257100e-01 -1.08852541e+00
-6.39300585e-01 -5.70642769e-01 -1.86137408e-01 1.23578501e+00
3.09916258e-01 -5.75364888e-01 1.16898918e+00 2.46016055e-01
-6.43038452e-02 -5.68531215e-01 -8.43438447e-01 -8.04808319e-01
3.06866556e-01 -9.06257629e-02 4.47867364e-01 9.45589364e-01
-5.44346154e-01 3.56896460e-01 -3.92429650e-01 4.96641323e-02
6.44561350e-01 1.83956698e-01 9.74109888e-01 -1.20672750e+00
-1.21387875e+00 -1.93560392e-01 -6.19579077e-01 -6.15253389e-01
-1.85675770e-01 -9.44779217e-01 -3.90822589e-01 -9.56023395e-01
2.32004240e-01 -4.76188362e-01 3.90876643e-02 1.84763387e-01
4.04430181e-02 8.28613222e-01 -4.42907065e-01 3.02292351e-02
-5.74702382e-01 2.74598300e-01 1.01866412e+00 -3.11912000e-01
3.18853885e-01 1.20126441e-01 -4.98364419e-01 1.06056213e+00
1.32919681e+00 -5.60578525e-01 -8.33145827e-02 -2.02898219e-01
3.41656357e-01 -2.26585895e-01 5.78661382e-01 -1.08738542e+00
2.03328893e-01 -1.91602513e-01 9.03409600e-01 -4.41557944e-01
6.96538627e-01 -2.46277049e-01 1.19972803e-01 4.04037714e-01
1.25580832e-01 -1.69033095e-01 -1.40173048e-01 3.25655699e-01
-2.05251798e-01 -9.91428614e-01 1.27117038e+00 -9.47029412e-01
-5.27995646e-01 3.90254200e-01 -5.85063756e-01 4.55587730e-02
8.81434321e-01 4.08447310e-02 -2.99784601e-01 -1.51973471e-01
-8.07901680e-01 -5.44721186e-01 7.51469493e-01 -2.00860932e-01
2.80654639e-01 -1.12981629e+00 -5.98157585e-01 2.66095493e-02
-1.31588057e-01 1.76161662e-01 -1.50040358e-01 3.65112841e-01
-1.24604023e+00 1.46607757e-01 -4.00927395e-01 -1.86442673e-01
-1.38488483e+00 6.49887025e-01 4.55478579e-01 -4.82905626e-01
-1.89562231e-01 1.66489768e+00 -7.08595291e-02 -6.44753337e-01
6.67909766e-03 2.67075390e-01 -1.98921248e-01 -1.29164919e-01
3.74255925e-01 6.32850409e-01 5.01583040e-01 -4.47354376e-01
2.35265419e-01 1.61923513e-01 -2.85431780e-02 -1.91484019e-01
1.36716068e+00 4.77434427e-01 -4.67842877e-01 3.83238584e-01
1.03723443e+00 5.29237986e-01 -9.84895647e-01 7.12929517e-02
-1.06119156e-01 -4.76345092e-01 4.69492488e-02 -6.64592803e-01
-8.92055690e-01 8.45511794e-01 7.54389703e-01 4.45098102e-01
7.31044829e-01 1.82840079e-01 6.04525566e-01 8.67503703e-01
4.37562466e-01 -8.37720156e-01 4.96209294e-01 3.56928259e-01
1.07729948e+00 -1.42358649e+00 2.10290536e-01 -2.72748470e-01
-3.23778749e-01 9.67661977e-01 8.37769985e-01 -6.57601357e-01
4.26363796e-01 5.14265001e-01 -1.78670242e-01 -2.21046686e-01
-6.62375331e-01 -2.54054546e-01 2.50894368e-01 5.19392073e-01
1.94970712e-01 2.25175112e-01 -1.50680110e-01 7.01562524e-01
-7.69866467e-01 -6.15158379e-01 5.30615687e-01 7.27158427e-01
-8.25810790e-01 -1.32759035e+00 -5.71102083e-01 7.50827849e-01
-5.67803264e-01 -1.25519231e-01 -1.13311636e+00 5.99228323e-01
5.43568730e-01 7.56400883e-01 3.28612745e-01 -2.95713425e-01
-2.83076614e-01 6.61822185e-02 8.49538088e-01 -7.21977711e-01
-1.01905787e+00 -6.84013963e-02 3.10470521e-01 3.51465434e-01
-1.81928441e-01 -4.69001651e-01 -8.87304664e-01 -7.08279490e-01
-4.05384362e-01 2.11489499e-01 3.53615731e-01 2.86933094e-01
-7.48030096e-02 9.02063027e-03 7.84911931e-01 -7.85953999e-01
-4.63175088e-01 -1.42908835e+00 -6.96701467e-01 4.92913783e-01
-3.46088260e-01 -6.62135363e-01 -7.14643598e-01 1.95789918e-01] | [7.945366859436035, 2.966765880584717] |
a74ddd6b-74ab-4bcc-bbc8-e5420fd66cba | rvt-robotic-view-transformer-for-3d-object | 2306.14896 | null | https://arxiv.org/abs/2306.14896v1 | https://arxiv.org/pdf/2306.14896v1.pdf | RVT: Robotic View Transformer for 3D Object Manipulation | For 3D object manipulation, methods that build an explicit 3D representation perform better than those relying only on camera images. But using explicit 3D representations like voxels comes at large computing cost, adversely affecting scalability. In this work, we propose RVT, a multi-view transformer for 3D manipulation that is both scalable and accurate. Some key features of RVT are an attention mechanism to aggregate information across views and re-rendering of the camera input from virtual views around the robot workspace. In simulations, we find that a single RVT model works well across 18 RLBench tasks with 249 task variations, achieving 26% higher relative success than the existing state-of-the-art method (PerAct). It also trains 36X faster than PerAct for achieving the same performance and achieves 2.3X the inference speed of PerAct. Further, RVT can perform a variety of manipulation tasks in the real world with just a few ($\sim$10) demonstrations per task. Visual results, code, and trained model are provided at https://robotic-view-transformer.github.io/. | ['Dieter Fox', 'Yu-Wei Chao', 'Valts Blukis', 'Yijie Guo', 'Jie Xu', 'Ankit Goyal'] | 2023-06-26 | null | null | null | null | ['robot-manipulation'] | ['robots'] | [-3.43432724e-01 -1.44483671e-01 2.03839620e-03 -2.67364025e-01
-7.57188380e-01 -7.32240736e-01 4.61694270e-01 -2.80813187e-01
-2.57225007e-01 3.11895460e-01 6.93066567e-02 -3.01021755e-01
1.49298921e-01 -4.14106220e-01 -1.24852729e+00 -2.55334705e-01
-5.82636148e-02 5.44672847e-01 3.08006853e-01 -2.01228321e-01
3.84901255e-01 6.17254734e-01 -1.54764283e+00 3.50404143e-01
3.83039981e-01 9.94517565e-01 7.67942488e-01 9.23769355e-01
2.77351081e-01 9.66696322e-01 -7.35771298e-01 3.01436558e-02
5.15089929e-01 2.20246911e-01 -6.47630095e-01 -9.38266963e-02
7.60657609e-01 -9.03953493e-01 -7.29749799e-01 6.04453146e-01
7.05810547e-01 3.59203249e-01 5.02505958e-01 -1.35424089e+00
-5.48176110e-01 5.30743062e-01 -7.77289212e-01 1.03236340e-01
6.12602949e-01 6.20365739e-01 7.12187767e-01 -1.09309244e+00
9.08621788e-01 1.57611215e+00 4.53989238e-01 5.41100204e-01
-1.03514850e+00 -6.83713853e-01 4.12681669e-01 -5.96624939e-03
-9.41252708e-01 -3.62158388e-01 2.07150847e-01 -4.65790600e-01
1.73222852e+00 4.15036418e-02 7.22355962e-01 1.24589264e+00
6.41356111e-01 8.59116852e-01 1.08074856e+00 1.86621144e-01
-9.43646282e-02 -4.65532035e-01 -7.90933445e-02 9.55980301e-01
6.55010790e-02 1.02820173e-01 -6.56393886e-01 5.43192811e-02
1.34951270e+00 2.22052023e-01 -2.79769838e-01 -9.14441943e-01
-1.66602397e+00 3.48028541e-01 7.90283263e-01 -2.26097867e-01
-3.37564617e-01 1.01102233e+00 6.14964843e-01 3.56588483e-01
3.06081653e-01 5.92906713e-01 -4.67139006e-01 -5.51850736e-01
-7.17051551e-02 6.01486921e-01 5.11820614e-01 1.63553298e+00
3.21191162e-01 1.05750687e-01 8.97140354e-02 5.21104634e-01
1.52918309e-01 6.09838486e-01 9.51171741e-02 -1.73377812e+00
8.44031632e-01 4.48327810e-01 2.14149013e-01 -6.32895470e-01
-4.99836951e-01 -1.40101492e-01 -3.28777969e-01 8.87675881e-01
4.83110756e-01 8.21488202e-02 -1.00408602e+00 1.41581428e+00
4.26605046e-01 -3.72764468e-01 -1.19814239e-01 1.01917279e+00
8.96538019e-01 5.64504862e-01 -2.46999577e-01 5.66311121e-01
1.19165492e+00 -1.32421732e+00 -5.96403778e-01 -2.88802564e-01
5.67539334e-01 -7.53097117e-01 1.27537405e+00 6.60276532e-01
-1.41996229e+00 -6.35369658e-01 -1.11069548e+00 -6.41036510e-01
-3.19582582e-01 1.76520705e-01 8.91258478e-01 -1.62191346e-01
-1.11724019e+00 7.22885191e-01 -1.25575078e+00 -3.16369385e-01
4.22135651e-01 3.21538627e-01 -6.42015040e-01 -3.34713548e-01
-3.09878975e-01 1.35325050e+00 1.31378444e-02 5.59854731e-02
-1.39374471e+00 -8.53926241e-01 -9.23517466e-01 -1.88597783e-01
5.57948172e-01 -8.83586645e-01 1.61215615e+00 -6.37719259e-02
-1.73472214e+00 7.52963483e-01 -9.69912186e-02 -1.71176106e-01
6.93508267e-01 -8.52747619e-01 3.86893183e-01 2.05245271e-01
2.01301172e-01 8.48030388e-01 7.12440372e-01 -1.40233505e+00
-3.00179273e-01 -5.19128859e-01 6.53461158e-01 4.97979492e-01
4.44914311e-01 -2.85637200e-01 -4.82984394e-01 -2.86972642e-01
2.62773663e-01 -1.05122530e+00 -2.57238239e-01 7.18090236e-01
-1.47402972e-01 -3.50914747e-01 8.93895388e-01 -5.46605945e-01
1.65228665e-01 -2.09364533e+00 6.47976041e-01 -3.96994054e-01
5.30577362e-01 -8.00175220e-02 9.52679850e-03 5.04199028e-01
2.03887179e-01 -1.63740084e-01 2.64308780e-01 -4.34766829e-01
2.04251066e-01 1.92870304e-01 -2.49727920e-01 6.35711908e-01
-6.86802566e-02 9.70027506e-01 -1.02631748e+00 -2.31634632e-01
6.83965981e-01 4.39045906e-01 -6.78965688e-01 3.25026155e-01
-2.95206785e-01 4.81927246e-01 -4.47288275e-01 4.92772430e-01
4.90561426e-01 -3.92246932e-01 5.96369803e-02 -1.15743779e-01
-7.73694292e-02 4.72196013e-01 -8.51519048e-01 2.46518207e+00
-8.07811916e-01 6.74376011e-01 4.22080755e-01 -6.85226202e-01
7.78938472e-01 1.02509387e-01 3.81425560e-01 -3.50603104e-01
2.26770297e-01 1.02197297e-01 -1.42482668e-01 -6.14208937e-01
6.42655015e-01 4.49474245e-01 -1.71276256e-01 5.28158605e-01
2.57593453e-01 -8.67509365e-01 1.23952897e-02 3.06950629e-01
1.36254191e+00 9.17003930e-01 1.28499418e-01 -2.44400967e-02
-3.04394662e-01 2.31798112e-01 1.17453657e-01 6.83978736e-01
-1.03473879e-01 5.32666802e-01 4.33844507e-01 -5.22333860e-01
-1.18257248e+00 -1.38511622e+00 2.66883731e-01 1.04488301e+00
4.99542296e-01 -4.54207838e-01 -3.51442277e-01 -5.20557106e-01
4.07591730e-01 7.08458960e-01 -4.07757461e-01 6.79856837e-02
-7.27087379e-01 9.81689692e-02 3.15681666e-01 9.42787588e-01
5.29391825e-01 -8.79160702e-01 -1.35705221e+00 -2.25468483e-02
-8.99500549e-02 -1.28216052e+00 -3.17391396e-01 2.39902183e-01
-1.07258332e+00 -1.17880344e+00 -5.48526764e-01 -5.36754191e-01
6.60065114e-01 8.04402828e-01 1.12960541e+00 -1.08799770e-01
-3.99988294e-01 6.15018249e-01 -4.02938664e-01 -3.88305992e-01
-2.69025236e-01 -1.21590495e-01 5.50332181e-02 -1.15143645e+00
-1.22554779e-01 -5.23092031e-01 -4.88291860e-01 2.17223763e-01
-2.94356436e-01 3.13841194e-01 6.11244977e-01 7.62640417e-01
4.91682857e-01 -6.35134220e-01 9.50266197e-02 -4.14344400e-01
2.88071811e-01 -1.49167016e-01 -7.63420045e-01 -1.83812961e-01
-1.81826010e-01 7.46266544e-03 4.00482237e-01 -5.72219908e-01
-9.03627694e-01 1.49344012e-01 4.66072820e-02 -1.02763784e+00
-1.82482693e-02 9.60049629e-02 2.43597314e-01 -3.78278084e-02
6.18027747e-01 -5.18799908e-02 2.62433439e-01 -5.11623263e-01
6.30136311e-01 2.37581328e-01 3.99693668e-01 -6.96859241e-01
6.50072515e-01 5.04517674e-01 -1.22584224e-01 -4.33614224e-01
-5.13127625e-01 -2.47120723e-01 -6.85101271e-01 -4.13954556e-01
8.84770095e-01 -1.13259232e+00 -1.24815094e+00 4.17480648e-01
-1.31780052e+00 -9.61394846e-01 -1.75255224e-01 8.18327725e-01
-1.06124198e+00 -4.50358279e-02 -7.90296376e-01 -5.67246854e-01
-8.69695842e-02 -1.63392472e+00 1.57917929e+00 -7.28351474e-02
-2.16180623e-01 -4.00336862e-01 -4.41691071e-01 4.33723986e-01
2.81405628e-01 4.18181092e-01 6.47939980e-01 -4.71539050e-02
-9.74681616e-01 -7.84540474e-02 -3.01415294e-01 -1.13185449e-02
1.77185779e-04 -2.09756017e-01 -8.98893654e-01 -4.34691191e-01
-2.07444593e-01 -7.52919078e-01 7.15384066e-01 4.24483001e-01
1.32397902e+00 1.27052203e-01 -4.68078941e-01 6.96800411e-01
1.21590722e+00 1.81972459e-01 3.91025305e-01 2.03227252e-01
7.63473868e-01 1.84193030e-01 9.11154211e-01 3.68812144e-01
5.27363181e-01 8.25885832e-01 9.52566326e-01 1.17596574e-01
-2.44604275e-01 -2.65750617e-01 5.51401138e-01 6.63018942e-01
-4.40350026e-01 -1.58334866e-01 -6.91173136e-01 3.35626036e-01
-1.77738702e+00 -9.21710074e-01 -9.17073265e-02 2.01935482e+00
4.16620195e-01 3.51025432e-01 -1.49722740e-01 -3.69924217e-01
2.94027507e-01 2.56941110e-01 -9.13537204e-01 -5.67650139e-01
4.57064807e-01 2.09907830e-01 7.04331756e-01 4.23679233e-01
-8.03011000e-01 9.70020711e-01 6.24055767e+00 3.63904476e-01
-9.18255508e-01 7.23368675e-02 5.66471508e-03 -5.19012988e-01
1.38365105e-01 -2.27272317e-01 -6.29062295e-01 -1.06700353e-01
3.12518626e-01 1.32371411e-01 7.68705249e-01 1.15490949e+00
1.60352439e-02 -3.08485597e-01 -1.47924697e+00 1.13906121e+00
2.35765189e-01 -1.17241585e+00 -2.09323987e-01 1.40241042e-01
4.69028831e-01 5.68209648e-01 2.01900303e-02 4.66867477e-01
7.68364906e-01 -9.24220204e-01 1.06719041e+00 2.70714194e-01
8.29445302e-01 -3.04470509e-01 3.46471816e-01 4.50561047e-01
-1.24140465e+00 -1.37225971e-01 -4.13145840e-01 -3.19767088e-01
1.89492688e-01 -7.40948366e-03 -8.63799810e-01 4.57976907e-01
1.09245408e+00 8.88868332e-01 -2.57467538e-01 7.54585683e-01
-2.93535709e-01 -2.55121380e-01 -4.01775301e-01 -1.21486366e-01
2.59375542e-01 1.62685305e-01 6.19291902e-01 7.25308359e-01
3.15592706e-01 1.13166943e-01 2.51509517e-01 8.76303673e-01
-3.30842733e-02 -6.53745353e-01 -1.13010716e+00 1.63001403e-01
5.16272247e-01 1.01194966e+00 -5.89741111e-01 -4.23998058e-01
-2.32684687e-01 1.11973846e+00 6.92217410e-01 2.81015754e-01
-1.22742772e+00 -4.11095917e-01 9.00482416e-01 -1.69409484e-01
4.94347602e-01 -9.80811179e-01 -4.38968465e-03 -9.32485938e-01
3.83066744e-01 -8.15617085e-01 -2.57298887e-01 -1.32749856e+00
-1.04278290e+00 4.39120203e-01 3.51551741e-01 -1.40845966e+00
-1.16024025e-01 -1.16318035e+00 -1.74669713e-01 5.88928878e-01
-1.16191530e+00 -1.02438462e+00 -7.38171279e-01 3.51014674e-01
1.05998421e+00 2.09912479e-01 8.24792624e-01 -4.60057221e-02
-7.72297606e-02 1.39588505e-01 -1.13198318e-01 -1.38745233e-01
8.56721878e-01 -1.37076879e+00 8.82481754e-01 2.11896688e-01
-2.82310486e-01 6.65997803e-01 4.61513430e-01 -5.70281744e-01
-2.26544118e+00 -7.49750435e-01 1.59473434e-01 -1.02135694e+00
4.74147588e-01 -6.51628077e-01 -4.46002305e-01 1.25668812e+00
3.39966565e-01 1.53342456e-01 -1.02733031e-01 2.85673924e-02
-4.88315612e-01 1.82265878e-01 -9.78436649e-01 6.45245731e-01
1.73824394e+00 -4.26197469e-01 -5.80897927e-01 5.94269812e-01
8.63184273e-01 -1.41276145e+00 -1.12118411e+00 1.74195066e-01
9.38096344e-01 -9.42794621e-01 1.25423729e+00 -3.40133190e-01
7.15967238e-01 -1.61207125e-01 -3.30241024e-01 -1.57353520e+00
-2.19065979e-01 -4.69595641e-01 -2.24121839e-01 4.43300515e-01
2.11876437e-01 -7.64470458e-01 3.65179449e-01 4.12893951e-01
-6.19965255e-01 -9.17803764e-01 -7.72568464e-01 -9.51252222e-01
5.02466820e-02 -2.83713371e-01 4.48603392e-01 5.32214999e-01
1.96100146e-01 2.83438504e-01 -6.91993013e-02 1.74827337e-01
5.31390071e-01 1.96868107e-01 1.37500453e+00 -8.09929907e-01
-3.23042989e-01 -3.19602072e-01 -3.21927577e-01 -1.62404180e+00
1.68118075e-01 -8.91128480e-01 2.24514812e-01 -1.88993275e+00
8.70660096e-02 -6.19077086e-01 4.21268344e-01 4.46596265e-01
2.02182263e-01 -1.15309022e-01 5.85512757e-01 -7.91396759e-03
-5.24479270e-01 6.55219138e-01 1.93711019e+00 -1.80721894e-01
2.80728228e-02 -2.46237695e-01 -2.37999827e-01 6.77632451e-01
8.40825737e-01 -1.71856567e-01 -3.56910586e-01 -1.26423466e+00
-9.38277617e-02 3.48364353e-01 6.71562314e-01 -1.08849943e+00
2.34436803e-02 -4.07366976e-02 6.28961086e-01 -7.43473470e-01
1.18202257e+00 -8.02661240e-01 8.66296515e-02 5.87174475e-01
-4.35023218e-01 7.19715059e-01 3.64789903e-01 5.85737526e-01
4.34051305e-01 -1.15052257e-02 4.04513001e-01 -6.85712814e-01
-6.71830416e-01 1.40621856e-01 -4.18203771e-01 2.17235391e-03
1.06076336e+00 -1.27765313e-01 -7.32498825e-01 -3.80763143e-01
-5.48677325e-01 4.30164754e-01 7.12620139e-01 6.56394184e-01
8.43964934e-01 -1.19771588e+00 -4.54949409e-01 -4.77525173e-03
-7.49303475e-02 5.11061966e-01 1.57850847e-01 7.14720249e-01
-8.47657919e-01 2.73229718e-01 -4.15545106e-01 -8.39574456e-01
-1.20823371e+00 6.19356871e-01 2.52158284e-01 8.75797793e-02
-1.28675997e+00 9.07449841e-01 2.54150301e-01 -7.76909113e-01
4.96916264e-01 -8.65961492e-01 4.06030625e-01 -6.30750835e-01
2.93700248e-01 5.06206274e-01 -1.82304546e-01 -2.39088938e-01
-1.56003177e-01 7.57483542e-01 8.34133774e-02 -1.25888065e-01
1.37740409e+00 1.44561768e-01 9.82193202e-02 6.54056013e-01
1.09392118e+00 -2.56978095e-01 -1.72526205e+00 8.77947435e-02
-6.68563128e-01 -7.03039348e-01 -8.08851495e-02 -6.81286991e-01
-9.33935821e-01 1.05514562e+00 3.30392241e-01 -2.48528004e-01
4.48615134e-01 1.95337951e-01 6.73729062e-01 8.79169405e-01
9.37394917e-01 -8.78193796e-01 4.32707787e-01 6.38759673e-01
1.41164410e+00 -1.22573709e+00 3.30474257e-01 -4.64392036e-01
-7.14704633e-01 1.03759050e+00 1.13349736e+00 -6.12851918e-01
3.54260921e-01 5.18537223e-01 -1.81325361e-01 -6.18853807e-01
-9.10010815e-01 1.38612539e-01 -2.27270797e-01 6.95349753e-01
2.88959712e-01 1.02041304e-01 4.79856491e-01 -6.13376535e-02
-3.35295230e-01 -1.22032218e-01 4.93316650e-01 1.37816215e+00
-2.88118899e-01 -4.89604533e-01 -4.69653644e-02 4.91030008e-01
-1.44878760e-01 2.73428202e-01 -3.06140006e-01 1.13570273e+00
-4.39928979e-01 7.72452056e-01 1.63981900e-01 -3.78694683e-01
6.80329025e-01 -1.90121278e-01 1.26739287e+00 -6.38889730e-01
-7.30072498e-01 -1.11340322e-01 2.06448704e-01 -1.36610889e+00
-3.27899694e-01 -5.79795420e-01 -1.33535171e+00 -5.19067228e-01
-3.33573788e-01 -4.68127042e-01 8.21666121e-01 4.67988521e-01
5.16610920e-01 7.98801959e-01 3.61174017e-01 -1.82797444e+00
-9.00732994e-01 -1.01268423e+00 -2.39239544e-01 1.15610413e-01
3.18886906e-01 -1.35360312e+00 -1.52773738e-01 -1.96380854e-01] | [4.770730018615723, 0.46780747175216675] |
0be72a91-d681-4169-9970-15f6ba558cce | the-statcan-dialogue-dataset-retrieving-data | 2304.01412 | null | https://arxiv.org/abs/2304.01412v2 | https://arxiv.org/pdf/2304.01412v2.pdf | The StatCan Dialogue Dataset: Retrieving Data Tables through Conversations with Genuine Intents | We introduce the StatCan Dialogue Dataset consisting of 19,379 conversation turns between agents working at Statistics Canada and online users looking for published data tables. The conversations stem from genuine intents, are held in English or French, and lead to agents retrieving one of over 5000 complex data tables. Based on this dataset, we propose two tasks: (1) automatic retrieval of relevant tables based on a on-going conversation, and (2) automatic generation of appropriate agent responses at each turn. We investigate the difficulty of each task by establishing strong baselines. Our experiments on a temporal data split reveal that all models struggle to generalize to future conversations, as we observe a significant drop in performance across both tasks when we move from the validation to the test set. In addition, we find that response generation models struggle to decide when to return a table. Considering that the tasks pose significant challenges to existing models, we encourage the community to develop models for our task, which can be directly used to help knowledge workers find relevant tables for live chat users. | ['Harm de Vries', 'Siva Reddy', 'Xing Han Lu'] | 2023-04-03 | null | null | null | null | ['dialogue-generation', 'table-retrieval', 'dialogue-generation'] | ['natural-language-processing', 'natural-language-processing', 'speech'] | [ 2.01562002e-01 4.03987259e-01 -2.40653940e-02 -3.75833571e-01
-1.50950372e+00 -1.09455061e+00 1.27156079e+00 2.92194188e-01
-5.13400257e-01 1.01469243e+00 9.27478850e-01 -4.72057790e-01
9.48534831e-02 -6.01102233e-01 -4.77427542e-01 -1.21157564e-01
1.67001486e-02 1.21663308e+00 2.70571679e-01 -4.74168986e-01
3.84420931e-01 2.89506279e-02 -1.17213500e+00 8.00476372e-01
5.58477223e-01 6.95066631e-01 -7.45534375e-02 9.87457275e-01
-1.40362307e-01 1.46011293e+00 -9.25380111e-01 -7.45509982e-01
1.06048003e-01 -4.84100133e-01 -1.53242016e+00 -1.18461482e-01
2.16453299e-01 -5.17479539e-01 -1.97537184e-01 6.62105143e-01
4.39941019e-01 3.32749277e-01 7.99796999e-01 -1.36461449e+00
-4.84019518e-01 9.81504679e-01 1.49448186e-01 4.42677349e-01
9.84988749e-01 2.86964536e-01 1.30112600e+00 -9.31691945e-01
1.02040720e+00 1.39713228e+00 5.17414629e-01 5.92135489e-01
-9.63044822e-01 -1.62808269e-01 1.90237135e-01 1.48917302e-01
-7.91270375e-01 -8.97648156e-01 3.41313094e-01 -5.04894733e-01
1.28865767e+00 5.61126888e-01 2.12650329e-01 1.29783356e+00
-1.20856255e-01 8.17306221e-01 7.63343096e-01 -2.15725332e-01
2.25707099e-01 3.00733238e-01 3.25287059e-02 4.63556677e-01
-1.27598181e-01 -4.31991130e-01 -9.27904665e-01 -5.29695332e-01
1.59786314e-01 -5.20145714e-01 -3.87806714e-01 1.94760919e-01
-1.43449807e+00 8.46064806e-01 8.93832147e-02 8.51570591e-02
-5.08324742e-01 -3.66851002e-01 4.62325066e-01 5.32235980e-01
4.22841549e-01 8.88013184e-01 -6.22268021e-01 -7.22315133e-01
-1.93688512e-01 8.32980335e-01 1.70521033e+00 1.13943839e+00
5.70008874e-01 -7.52904475e-01 -3.06500107e-01 9.81456816e-01
1.27852231e-01 3.84125412e-01 2.38286793e-01 -1.14360428e+00
1.08418918e+00 7.85582125e-01 5.19525051e-01 -9.07945395e-01
-5.41975498e-01 2.47517183e-01 -2.38486260e-01 -4.92410630e-01
9.81111407e-01 -4.57510442e-01 -3.59159768e-01 1.60906839e+00
1.25996187e-01 -6.55630231e-01 4.96375531e-01 6.94491863e-01
9.85721409e-01 6.86665773e-01 -1.27717406e-01 -2.43770957e-01
1.48434210e+00 -1.04871607e+00 -7.23015606e-01 -4.62121844e-01
7.35843837e-01 -1.02036798e+00 1.23997629e+00 2.87625402e-01
-1.33314574e+00 5.28083704e-02 -5.84699631e-01 -3.92962098e-01
-5.14472783e-01 -1.69235095e-01 6.08199477e-01 1.99694112e-01
-1.10586810e+00 1.57718584e-01 -5.48253357e-01 -7.04547346e-01
-1.41701341e-01 6.28082976e-02 -1.94651201e-01 1.28955409e-01
-1.28995502e+00 1.01381004e+00 -1.03070937e-01 5.29300235e-02
-5.58488488e-01 -4.68390435e-01 -6.11274600e-01 -2.06869200e-01
6.26747191e-01 -5.64696848e-01 1.92947650e+00 -4.63066608e-01
-1.58885920e+00 8.71770561e-01 -3.50957572e-01 -4.70741123e-01
8.65763128e-01 -1.12128012e-01 -1.09426193e-01 7.14502996e-03
5.45090973e-01 4.55811739e-01 6.39355630e-02 -8.95927191e-01
-9.54172790e-01 -2.31575593e-01 5.18031657e-01 4.33355033e-01
-7.63050616e-02 2.06718147e-01 -5.06524503e-01 -2.83619702e-01
-1.88934490e-01 -1.15479696e+00 1.33002698e-01 -5.49114466e-01
-7.13239670e-01 -7.69613862e-01 2.66236633e-01 -8.65880847e-01
1.24420142e+00 -1.51084054e+00 2.31851235e-01 -8.24602414e-03
1.08598145e-02 -2.92371869e-01 -4.50409018e-02 9.32764530e-01
3.93731922e-01 2.45021582e-01 -5.98388761e-02 -2.23456040e-01
1.66273013e-01 -1.02536052e-01 -5.05235851e-01 -1.71911255e-01
2.19932467e-01 9.00390923e-01 -8.72795820e-01 -2.47930095e-01
-4.59321946e-01 -4.32855897e-02 -6.31975710e-01 3.53922904e-01
-5.88529289e-01 2.35496700e-01 -5.96452475e-01 4.83628780e-01
9.08432901e-02 -5.34772635e-01 3.08441877e-01 1.81292802e-01
-4.63907011e-02 1.32121229e+00 -7.53129780e-01 1.44875515e+00
-3.78588736e-01 6.53410494e-01 1.68693185e-01 -3.15375447e-01
5.72272003e-01 3.00901175e-01 1.02597103e-01 -6.80265009e-01
-5.32178022e-02 2.65701413e-01 1.18191652e-01 -5.96228063e-01
7.46292353e-01 1.84857503e-01 -3.49128008e-01 1.04400575e+00
-3.60552639e-01 -1.04297131e-01 6.57123029e-01 6.39516652e-01
1.33508384e+00 -2.65331268e-01 3.77563536e-02 -1.14633046e-01
6.76439643e-01 5.36924899e-01 1.26077816e-01 9.99886096e-01
-5.69218211e-02 3.94936264e-01 9.61357236e-01 -6.77054405e-01
-7.78455198e-01 -6.54064953e-01 3.62790495e-01 1.44657946e+00
-1.69956740e-02 -6.39068186e-01 -7.87537158e-01 -8.22659016e-01
-9.47073698e-02 8.24629724e-01 -6.28825188e-01 1.15077987e-01
-7.88998425e-01 -5.64997137e-01 4.41802144e-01 4.58601892e-01
5.39556324e-01 -1.38016331e+00 -5.47666848e-01 3.97659451e-01
-9.87179101e-01 -1.25523710e+00 -7.87723601e-01 8.81161168e-02
-2.28503406e-01 -1.30951774e+00 -3.51023734e-01 -5.97895861e-01
2.83519059e-01 2.06331275e-02 1.53215969e+00 4.14782107e-01
2.01348096e-01 6.43254936e-01 -3.83227974e-01 -5.87780476e-01
-1.02016461e+00 6.07312083e-01 -5.11300042e-02 -2.87194371e-01
5.47753870e-01 -2.40347549e-01 -3.04510087e-01 4.42885399e-01
-4.11403388e-01 6.98910579e-02 1.88449129e-01 7.38293469e-01
-2.06220821e-01 -3.95297736e-01 6.69629931e-01 -1.12931979e+00
1.31563139e+00 -6.79374635e-01 -5.00309885e-01 4.13041979e-01
-4.60820228e-01 2.13994905e-01 5.99963427e-01 -3.14184427e-02
-1.04438126e+00 -2.93750346e-01 2.56912149e-02 4.29845721e-01
4.14270870e-02 5.89389384e-01 1.68919697e-01 6.56960309e-01
8.68750751e-01 1.01316907e-01 9.05273557e-02 -3.77154678e-01
1.35113657e-01 9.49870110e-01 4.00225997e-01 -7.48756170e-01
6.78153336e-01 1.03787594e-01 -6.89791083e-01 -5.41489005e-01
-8.25048745e-01 -3.44267458e-01 -3.52455288e-01 -1.52255043e-01
8.35781395e-01 -8.51936400e-01 -1.21516681e+00 2.37309605e-01
-1.39807618e+00 -8.64284575e-01 2.29512349e-01 9.68905613e-02
-6.57729387e-01 -7.42493421e-02 -7.56531656e-01 -8.35138738e-01
-3.27411175e-01 -1.16293085e+00 8.86522293e-01 1.67549968e-01
-7.85432518e-01 -1.00757110e+00 1.38659611e-01 7.91633666e-01
3.92824113e-01 -2.94629186e-01 1.05173743e+00 -1.28368390e+00
-7.44934976e-01 -8.40402842e-02 4.66608331e-02 -3.11920702e-01
2.58276939e-01 -6.63094893e-02 -9.78105962e-01 -1.83403064e-02
-3.40416312e-01 -9.23691690e-01 6.65832758e-01 -2.48171702e-01
4.96659338e-01 -8.18651855e-01 -3.81228894e-01 -1.85412034e-01
5.69333911e-01 4.88920242e-01 4.05899912e-01 5.92312574e-01
1.51401997e-01 1.01402557e+00 4.53740388e-01 3.37466091e-01
1.14064980e+00 7.12081313e-01 1.70419943e-02 3.52319360e-01
1.46563306e-01 -4.16978538e-01 4.05683815e-01 6.14173353e-01
6.56764880e-02 -6.32217824e-01 -1.08858573e+00 6.22596622e-01
-1.96487486e+00 -1.03853595e+00 2.05364645e-01 2.06546330e+00
1.21593165e+00 3.05029452e-01 4.94786859e-01 -3.33450705e-01
3.79675299e-01 7.46155158e-02 -5.10132313e-01 -3.44764739e-01
1.68257564e-01 -3.07119936e-01 1.11777149e-01 8.61587822e-01
-8.47017825e-01 8.14551651e-01 6.70527506e+00 8.91072303e-02
-8.56072545e-01 -2.01860785e-01 9.82579708e-01 -1.03011288e-01
-5.14501035e-01 -1.08493485e-01 -8.22048664e-01 4.36746448e-01
1.08292484e+00 -4.32652950e-01 8.41830909e-01 5.41878462e-01
3.99944447e-02 -1.79240420e-01 -1.55745351e+00 5.80126822e-01
5.22944406e-02 -1.36483920e+00 1.12624630e-01 -7.67543241e-02
4.34421659e-01 2.04944611e-02 -2.36329198e-01 6.34344280e-01
9.21831429e-01 -1.06699598e+00 5.61200023e-01 4.07245100e-01
3.04776371e-01 -4.82013851e-01 6.45411015e-01 5.30849695e-01
-8.14256430e-01 -1.51730636e-02 -7.69008994e-02 -2.11252272e-01
3.52855958e-02 -2.69773036e-01 -1.55524695e+00 2.91900128e-01
5.81993639e-01 4.27868843e-01 -6.23779714e-01 4.78711367e-01
-1.33988202e-01 3.38622749e-01 -3.35421532e-01 -4.79034632e-01
1.98402181e-01 -3.53090838e-02 4.28287685e-01 1.04554224e+00
-6.93640038e-02 2.58206695e-01 3.50146115e-01 8.47123086e-01
-4.83714581e-01 -8.40125829e-02 -6.42098129e-01 -4.02245730e-01
7.94667721e-01 1.10147226e+00 -7.24262953e-01 -4.27406669e-01
-4.41501349e-01 7.24798322e-01 5.75941801e-01 4.10009950e-01
-3.98097515e-01 -2.95336545e-01 5.45332313e-01 -8.53108801e-03
-3.68299941e-03 2.13870518e-02 -1.55767471e-01 -1.06298959e+00
3.71995687e-01 -1.39512157e+00 6.22762024e-01 -7.58374989e-01
-1.33789539e+00 8.69684994e-01 -8.57932121e-02 -7.73443580e-01
-1.24172509e+00 -2.16687307e-01 -5.00562191e-01 8.12653899e-01
-1.26989329e+00 -7.39176869e-01 -2.74198741e-01 4.06983137e-01
9.23979521e-01 -2.82830417e-01 8.63909304e-01 6.52386472e-02
-4.53888774e-01 4.47670430e-01 -3.34008157e-01 4.67742920e-01
9.15212452e-01 -1.32348323e+00 9.17922378e-01 5.27251542e-01
2.19714195e-01 8.92229736e-01 7.85273969e-01 -7.22577214e-01
-1.47115541e+00 -6.32922828e-01 1.54592085e+00 -1.31836343e+00
6.79372251e-01 -5.98032892e-01 -7.97862947e-01 1.02368319e+00
5.57368755e-01 -6.81711853e-01 5.56098223e-01 3.55185956e-01
-2.94845939e-01 2.86979795e-01 -7.87061155e-01 7.82912970e-01
1.01756907e+00 -6.82738543e-01 -7.29584455e-01 7.34986067e-01
6.91054225e-01 -6.81376040e-01 -5.49028397e-01 -8.27332810e-02
5.71005344e-01 -9.47841167e-01 6.29010141e-01 -7.87520170e-01
6.04253411e-01 5.24446405e-02 -7.02481493e-02 -1.22213531e+00
1.38469502e-01 -1.19920874e+00 3.20955843e-01 1.20518184e+00
1.16836631e+00 -6.08969867e-01 4.63308424e-01 1.21414173e+00
6.60104901e-02 -5.07612765e-01 -7.47304261e-01 -2.48415932e-01
5.14421538e-02 -1.52745187e-01 7.14103341e-01 8.17785442e-01
4.69085693e-01 9.20942903e-01 -7.47586563e-02 -5.37553728e-02
4.24917750e-02 8.66935924e-02 1.25042880e+00 -1.15557909e+00
-2.60526478e-01 -5.44667900e-01 4.65240687e-01 -1.31401265e+00
9.62239057e-02 -7.09526598e-01 3.38215411e-01 -1.80455351e+00
2.47174442e-01 -3.77231419e-01 4.09233391e-01 5.21119237e-01
-3.22760314e-01 -8.09278190e-02 1.98731452e-01 4.43430007e-01
-7.93282330e-01 3.37844729e-01 9.69222069e-01 -1.75729588e-01
-4.30556625e-01 2.46771678e-01 -1.07808077e+00 5.74825525e-01
5.43729901e-01 -2.95316428e-01 -5.17425954e-01 -4.54276890e-01
5.45623124e-01 4.53900725e-01 1.13432817e-01 -6.07418120e-01
6.09182775e-01 -2.54059792e-01 -1.18898069e-02 -3.71193379e-01
4.22653347e-01 -3.81934434e-01 -2.99615055e-01 9.00003314e-02
-1.00914955e+00 5.52302122e-01 1.58913687e-01 3.80023420e-01
-5.33102825e-02 -7.14606121e-02 1.62532777e-01 -2.99408913e-01
-1.73900574e-01 -1.94398478e-01 -7.21639276e-01 6.80784345e-01
5.22484243e-01 1.63588345e-01 -7.15210974e-01 -1.10429513e+00
-5.10734200e-01 5.77725053e-01 3.28437835e-01 6.88657463e-01
1.95788175e-01 -9.98282313e-01 -8.45368743e-01 -1.07222889e-02
2.64905840e-01 2.39532590e-02 -2.59000152e-01 5.72073102e-01
-3.83172244e-01 8.24143171e-01 1.55228972e-01 -3.40122998e-01
-1.03183234e+00 9.66676772e-02 2.61071473e-01 -5.09665251e-01
-4.24900323e-01 9.78897452e-01 -8.21999682e-04 -6.88547790e-01
4.91280615e-01 -4.33387816e-01 -3.27665776e-01 4.10163283e-01
8.62233758e-01 2.87647128e-01 2.55913883e-01 -2.82421768e-01
-4.21129286e-01 -1.28495917e-01 -5.86776316e-01 -5.93332946e-01
1.22033048e+00 -2.81324327e-01 -1.00867659e-01 5.79576790e-01
8.38562429e-01 2.83532679e-01 -1.05985904e+00 -3.95130336e-01
3.84347051e-01 -1.57581195e-01 -6.88635230e-01 -1.30573308e+00
-4.37252849e-01 3.22187781e-01 -3.03364515e-01 7.60360658e-01
5.25084376e-01 2.12503374e-01 8.37496996e-01 9.49181557e-01
4.64405388e-01 -1.01971424e+00 3.01925600e-01 1.01524091e+00
1.36683655e+00 -1.35351837e+00 -1.87432319e-01 -2.00091332e-01
-9.37940955e-01 1.18892229e+00 7.20049977e-01 4.34126705e-01
1.31738394e-01 1.93173379e-01 4.84699368e-01 -3.92358720e-01
-1.66905940e+00 5.71938120e-02 4.44126613e-02 4.61202234e-01
6.19692624e-01 -2.90826440e-01 -9.20944139e-02 4.80593771e-01
-7.19632566e-01 -1.95000678e-01 8.67709339e-01 9.20689464e-01
-1.63448602e-01 -1.04111373e+00 -1.07867725e-01 4.28349972e-01
-2.03604922e-01 -1.93917006e-01 -1.15996969e+00 7.87867785e-01
-7.84346104e-01 1.47316957e+00 1.42957285e-01 -2.72638470e-01
6.21203899e-01 4.98763978e-01 2.71323863e-02 -6.88073099e-01
-9.51562345e-01 -3.22298616e-01 9.26651657e-01 -4.34134930e-01
-1.80556193e-01 -8.47566366e-01 -1.17202532e+00 -5.14562726e-01
3.14231999e-02 5.87385356e-01 2.63255119e-01 9.84912694e-01
4.31414574e-01 1.97573677e-01 4.88014817e-01 -4.06501055e-01
-8.36586654e-01 -1.26460600e+00 7.60760978e-02 5.22594035e-01
3.42189461e-01 -4.30611879e-01 -4.41608399e-01 1.57293916e-01] | [12.423127174377441, 7.997835636138916] |
7b4d3e7d-2213-4f34-babf-0bbd83ea38f3 | vision-language-navigation-with-self | 1911.07883 | null | https://arxiv.org/abs/1911.07883v4 | https://arxiv.org/pdf/1911.07883v4.pdf | Vision-Language Navigation with Self-Supervised Auxiliary Reasoning Tasks | Vision-Language Navigation (VLN) is a task where agents learn to navigate following natural language instructions. The key to this task is to perceive both the visual scene and natural language sequentially. Conventional approaches exploit the vision and language features in cross-modal grounding. However, the VLN task remains challenging, since previous works have neglected the rich semantic information contained in the environment (such as implicit navigation graphs or sub-trajectory semantics). In this paper, we introduce Auxiliary Reasoning Navigation (AuxRN), a framework with four self-supervised auxiliary reasoning tasks to take advantage of the additional training signals derived from the semantic information. The auxiliary tasks have four reasoning objectives: explaining the previous actions, estimating the navigation progress, predicting the next orientation, and evaluating the trajectory consistency. As a result, these additional training signals help the agent to acquire knowledge of semantic representations in order to reason about its activity and build a thorough perception of the environment. Our experiments indicate that auxiliary reasoning tasks improve both the performance of the main task and the model generalizability by a large margin. Empirically, we demonstrate that an agent trained with self-supervised auxiliary reasoning tasks substantially outperforms the previous state-of-the-art method, being the best existing approach on the standard benchmark. | ['Xiaojun Chang', 'Fengda Zhu', 'Xiaodan Liang', 'Yi Zhu'] | 2019-11-18 | vision-language-navigation-with-self-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Zhu_Vision-Language_Navigation_With_Self-Supervised_Auxiliary_Reasoning_Tasks_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Zhu_Vision-Language_Navigation_With_Self-Supervised_Auxiliary_Reasoning_Tasks_CVPR_2020_paper.pdf | cvpr-2020-6 | ['vision-language-navigation'] | ['computer-vision'] | [ 3.84058766e-02 5.30189872e-02 -1.32822543e-01 -4.34713125e-01
-3.47174734e-01 -5.58504224e-01 1.04928434e+00 1.22109711e-01
-6.44726992e-01 5.17474055e-01 3.68861407e-01 -2.95125633e-01
3.91649380e-02 -6.59888923e-01 -8.93048823e-01 -6.52096331e-01
4.00307067e-02 6.01993263e-01 4.57291186e-01 -6.37761176e-01
4.21566039e-01 2.31022239e-01 -1.58103228e+00 2.09965304e-01
9.30845320e-01 9.31115568e-01 6.67153180e-01 4.94482279e-01
-1.32586494e-01 1.39690101e+00 -4.39483151e-02 3.34745161e-02
7.52377287e-02 -5.21671951e-01 -7.50230372e-01 -3.05643305e-02
2.69434959e-01 -3.54856819e-01 -4.48948354e-01 1.24098241e+00
-2.93752346e-02 6.53442740e-01 6.09409273e-01 -1.40436709e+00
-7.17646539e-01 3.50516379e-01 -2.18893319e-01 4.98413742e-02
5.73508978e-01 4.10759956e-01 1.00245261e+00 -6.73756778e-01
6.01311803e-01 1.38711393e+00 4.71345127e-01 7.29687631e-01
-1.06005239e+00 -2.64863014e-01 7.67230570e-01 5.64295471e-01
-8.12504649e-01 -3.32164615e-01 6.78352594e-01 -4.70117331e-01
9.55411315e-01 -1.89109445e-01 6.43758059e-01 1.36271334e+00
3.07838824e-02 9.77968633e-01 1.13529885e+00 -2.26542234e-01
4.94998306e-01 -9.64426324e-02 3.55274290e-01 1.27935708e+00
1.09933816e-01 3.40638220e-01 -6.69363260e-01 2.46067151e-01
6.67479634e-01 6.22556508e-02 -3.57198715e-01 -9.56012070e-01
-1.41334832e+00 7.63220370e-01 8.82718682e-01 7.77305514e-02
-5.47575951e-01 2.26599589e-01 3.85777652e-01 1.37752190e-01
-1.88030392e-01 2.87066817e-01 -4.02413815e-01 -3.41273062e-02
-2.76895463e-01 2.79146254e-01 6.85975194e-01 1.05408359e+00
1.07733071e+00 6.49250224e-02 -1.66065499e-01 4.20935363e-01
6.72237217e-01 7.27410018e-01 5.30498981e-01 -1.37583280e+00
6.26966298e-01 7.66676366e-01 1.03611477e-01 -7.59412169e-01
-6.48032129e-01 -1.73912585e-01 -6.10184371e-01 6.06675565e-01
7.07548022e-01 1.11199960e-01 -1.11862659e+00 2.09419799e+00
1.29404217e-01 -6.77343979e-02 5.91612697e-01 9.72504616e-01
8.89570057e-01 4.01920080e-01 2.16779321e-01 1.57263651e-01
1.46836889e+00 -1.70633316e+00 -5.77121198e-01 -8.72509837e-01
5.95468760e-01 -1.97130099e-01 1.26383829e+00 1.88710421e-01
-6.27724349e-01 -7.66961634e-01 -9.37269926e-01 -3.28584999e-01
-6.21465743e-01 1.98820353e-01 9.79615331e-01 1.25483677e-01
-1.08010602e+00 1.45088315e-01 -1.03460813e+00 -6.00963235e-01
3.83968621e-01 -4.82854843e-02 -5.54015994e-01 -2.09875360e-01
-1.02591896e+00 1.04565263e+00 4.39056993e-01 1.94969088e-01
-1.18583322e+00 -2.27385387e-01 -1.41576028e+00 -4.78440635e-02
6.28974199e-01 -9.66494620e-01 1.37099528e+00 -8.33662033e-01
-1.45203257e+00 7.81315923e-01 -3.26159000e-01 -6.46391571e-01
5.11048496e-01 -2.38491014e-01 7.40166456e-02 7.13080615e-02
5.68879366e-01 7.96259940e-01 5.47769308e-01 -1.39331079e+00
-9.02078629e-01 -6.96472824e-01 6.52965248e-01 5.02374828e-01
3.86988759e-01 -9.41149294e-01 -7.15644658e-01 -3.18879873e-01
3.75140399e-01 -1.01252866e+00 -4.39524591e-01 2.07844973e-02
-2.74658114e-01 -2.03414708e-01 5.98506153e-01 -6.73886836e-01
4.79689628e-01 -1.96307802e+00 4.05218184e-01 1.16283279e-02
1.28022954e-01 -7.92004094e-02 -2.68460006e-01 2.67011911e-01
3.30117375e-01 -7.46603012e-01 -1.80091187e-01 -3.88608426e-01
1.51830986e-01 5.53551495e-01 -4.70811546e-01 3.05449545e-01
-3.01600546e-01 1.20521784e+00 -1.21116447e+00 -2.20604509e-01
4.06251460e-01 2.56985873e-01 -5.51062405e-01 4.08092290e-01
-6.24751687e-01 9.10257459e-01 -5.65204024e-01 4.26638812e-01
2.04595849e-01 -3.14508498e-01 -3.40009853e-02 -1.24652788e-01
-4.58477996e-02 3.25884819e-01 -9.02368605e-01 2.37175369e+00
-5.72425604e-01 6.66903913e-01 -1.49968443e-02 -1.22560954e+00
6.65233135e-01 -1.54469430e-01 6.34284019e-02 -1.24697292e+00
-4.55244929e-02 -2.27843728e-02 -1.14729524e-01 -5.88003516e-01
2.03741238e-01 3.03767085e-01 -3.46167646e-02 3.40748906e-01
-1.96957309e-02 -1.89024466e-03 2.81730086e-01 3.19083780e-01
8.26410532e-01 6.28219187e-01 5.86525321e-01 -9.29161459e-02
7.83172190e-01 3.49522114e-01 2.36420497e-01 1.01815462e+00
-2.81346440e-01 6.09788261e-02 2.38578767e-01 -5.61194003e-01
-6.18804932e-01 -1.16846204e+00 5.60315669e-01 1.26437366e+00
6.55371130e-01 -3.94694328e-01 -6.68216705e-01 -7.52924085e-01
-1.72023505e-01 1.19824147e+00 -9.23533678e-01 -2.48541072e-01
-4.00288701e-01 -3.19311559e-01 1.74392134e-01 8.56893480e-01
8.37866545e-01 -1.23723888e+00 -1.04576790e+00 -2.07501978e-01
-4.31504697e-01 -1.43920207e+00 -2.85762548e-01 3.51111382e-01
-6.75049901e-01 -1.33071101e+00 -2.28827193e-01 -9.17732775e-01
8.62066865e-01 4.77608651e-01 1.08759522e+00 2.21723653e-02
-1.65228604e-03 8.66415739e-01 -2.99189806e-01 -8.31342191e-02
-3.58667612e-01 -2.31170714e-01 -1.04212433e-01 -2.77064323e-01
4.76777703e-01 -3.88854325e-01 -5.64348578e-01 -3.87119763e-02
-3.82636160e-01 4.74600941e-01 6.54232800e-01 9.22548473e-01
5.60711622e-01 -1.03339165e-01 8.56002644e-02 -7.49363244e-01
3.99250090e-01 -1.63969263e-01 -7.84682751e-01 2.86362112e-01
-6.04103208e-01 6.35338962e-01 5.97726822e-01 -1.31583542e-01
-1.38474000e+00 1.64015219e-01 -1.20723374e-01 -1.10639177e-01
-4.97779816e-01 5.03348827e-01 -2.52866834e-01 4.44379896e-02
5.88284194e-01 6.93694592e-01 1.33548483e-01 -8.50396454e-02
6.93423271e-01 -9.17307064e-02 8.07573915e-01 -5.88494718e-01
7.29169905e-01 7.87808001e-01 2.34845147e-01 -6.65671468e-01
-1.21919453e+00 -6.42085195e-01 -7.62171209e-01 -4.85905483e-02
1.02707982e+00 -1.11713469e+00 -1.07859206e+00 2.20880315e-01
-1.16332114e+00 -8.47508252e-01 -1.82882890e-01 4.97049391e-01
-1.05922604e+00 4.83241975e-01 -3.27348709e-01 -6.79076314e-01
1.63758233e-01 -1.35109186e+00 1.03794968e+00 8.24226663e-02
-7.54541233e-02 -1.16380870e+00 1.73953518e-01 4.98857170e-01
2.33988270e-01 -9.90373781e-04 8.36826146e-01 -6.05752826e-01
-8.68878603e-01 3.14789325e-01 -3.60100061e-01 -1.15639709e-01
2.36724168e-01 -7.17264414e-01 -8.29084039e-01 -8.72516856e-02
1.19059920e-01 -4.83915091e-01 1.24043190e+00 3.79231274e-01
9.01507735e-01 -4.96415831e-02 -3.34864736e-01 6.51876092e-01
1.27115917e+00 3.08215350e-01 3.96052957e-01 7.96115577e-01
8.85408640e-01 8.44377577e-01 7.36670315e-01 7.48317391e-02
9.82321858e-01 5.84205687e-01 8.07321072e-01 6.29899129e-02
-1.07289799e-01 -5.06612659e-01 4.98473108e-01 2.69169331e-01
-1.64280668e-01 3.68233249e-02 -9.88358557e-01 1.76600665e-01
-2.45892930e+00 -1.20497346e+00 1.12872392e-01 1.87789178e+00
3.44442964e-01 2.05290616e-01 -1.97875082e-01 -2.68825233e-01
2.36264005e-01 2.62757599e-01 -9.58572328e-01 -3.49789262e-02
2.53566131e-02 -4.95036930e-01 2.99482614e-01 7.99093127e-01
-1.20339918e+00 1.16298592e+00 5.57225323e+00 2.60014236e-01
-6.80160582e-01 -1.47202760e-01 9.75909904e-02 4.93532211e-01
-1.70069233e-01 -2.41474137e-02 -8.31570208e-01 -3.23946662e-02
4.55765188e-01 2.48497337e-01 7.51351893e-01 1.11072719e+00
1.36193499e-01 -3.91153991e-01 -1.53129387e+00 1.01597023e+00
4.08347100e-01 -1.12841558e+00 3.66667867e-01 -2.00281501e-01
5.11492789e-01 1.51920155e-01 4.73935008e-02 7.71094739e-01
6.30868733e-01 -9.39574957e-01 8.77216697e-01 7.76198149e-01
1.94562107e-01 -3.43109965e-01 6.00288749e-01 6.44364893e-01
-1.38357627e+00 -3.28194559e-01 -6.71190722e-03 -3.45492661e-01
2.50967503e-01 -1.86960891e-01 -4.81918573e-01 4.64409143e-01
6.08170331e-01 1.00978005e+00 -6.63686812e-01 5.48853815e-01
-7.55382895e-01 -5.49825430e-02 1.75690595e-02 -6.23221248e-02
7.30131984e-01 -5.45569479e-01 5.20608783e-01 8.45601439e-01
4.10573761e-04 -8.07390828e-03 5.16362786e-01 9.45296228e-01
3.38620335e-01 -3.50999802e-01 -6.72051728e-01 1.81276292e-01
2.31288932e-02 7.60347784e-01 -8.10633481e-01 -3.01296353e-01
-7.63350666e-01 1.14162040e+00 6.73779130e-01 7.79580355e-01
-8.34846735e-01 3.64078023e-02 7.84760416e-01 -3.69262069e-01
3.49226981e-01 -4.89548445e-01 -7.87493438e-02 -1.19602263e+00
4.64036614e-02 -6.78977251e-01 3.78504813e-01 -9.58422005e-01
-1.00770748e+00 6.27847791e-01 -2.25065306e-01 -1.02652037e+00
-7.13932276e-01 -1.07462072e+00 -3.45362246e-01 6.87118113e-01
-1.92676795e+00 -1.38752425e+00 -9.16475713e-01 8.00197542e-01
1.01120126e+00 -4.13260192e-01 9.02626157e-01 -2.59483606e-01
-1.79850727e-01 1.36692509e-01 -1.35358542e-01 2.79925555e-01
4.43748862e-01 -1.46793950e+00 2.90212572e-01 9.33064282e-01
3.55368882e-01 7.51191616e-01 7.60231316e-01 -3.73076618e-01
-1.51468050e+00 -9.45663214e-01 3.86453331e-01 -7.12096274e-01
6.58667147e-01 -2.43886873e-01 -6.81941152e-01 1.04102921e+00
6.70533404e-02 -6.35930225e-02 3.37928504e-01 1.21828780e-01
-7.27849782e-01 5.54180257e-02 -5.89262784e-01 1.10266697e+00
1.31381953e+00 -8.46394598e-01 -1.10029817e+00 1.90369144e-01
7.51860142e-01 -5.05567670e-01 1.10584646e-01 1.95383430e-01
5.75014234e-01 -1.21728230e+00 1.24939787e+00 -7.17219055e-01
2.98133075e-01 -5.31347096e-01 -4.53330904e-01 -1.25433207e+00
-3.21220964e-01 -4.62231040e-02 -1.51884109e-01 6.71521068e-01
2.47446090e-01 -6.14932835e-01 7.35494137e-01 5.02344012e-01
-1.30935460e-01 -3.28174293e-01 -6.41713738e-01 -6.37637496e-01
-4.36234087e-01 -5.12694120e-01 3.37972790e-01 5.65914333e-01
-1.90843921e-02 7.05985665e-01 -8.01449046e-02 4.18868750e-01
8.43320966e-01 4.87138897e-01 9.78734195e-01 -1.25633383e+00
-1.03749648e-01 -5.24316728e-01 -2.93253124e-01 -1.61010182e+00
8.07799518e-01 -8.37592185e-01 5.45065582e-01 -2.01213336e+00
1.62316889e-01 -4.48579304e-02 -2.35379159e-01 5.19288778e-01
-1.72298357e-01 -2.67315239e-01 1.50279418e-01 4.18872498e-02
-1.09590685e+00 8.37049961e-01 1.35892749e+00 -4.45146918e-01
-2.47526348e-01 9.11956429e-02 -6.58520460e-01 1.09578216e+00
6.32947922e-01 1.16215952e-01 -9.85903561e-01 -5.93300521e-01
1.66343719e-01 6.32970221e-03 8.39805365e-01 -9.78331745e-01
6.43556476e-01 -2.21923947e-01 2.59393483e-01 -6.94352925e-01
5.54936826e-01 -1.04470479e+00 -3.66781026e-01 6.90085888e-01
-5.31371593e-01 3.85407023e-02 3.09119254e-01 9.92023468e-01
-2.47093469e-01 -9.64475349e-02 4.64697957e-01 -4.40059811e-01
-1.66629100e+00 2.90209264e-01 -3.91897142e-01 2.20484078e-01
8.64878058e-01 -2.44516611e-01 -5.70907474e-01 -4.73547518e-01
-8.42414677e-01 5.90697050e-01 4.51089561e-01 5.50125062e-01
7.15449929e-01 -1.20081365e+00 -2.33143315e-01 2.79662967e-01
5.30435205e-01 5.69133200e-02 3.14827710e-01 7.92458177e-01
-4.06915367e-01 7.38520384e-01 -5.63937008e-01 -7.40586102e-01
-9.08550620e-01 9.01742101e-01 3.08642656e-01 -1.24439538e-01
-9.23032701e-01 6.72643304e-01 8.84205103e-01 -7.13766217e-01
5.48076510e-01 -4.62222725e-01 -6.46463394e-01 -1.55918777e-01
5.16550481e-01 1.25269294e-01 -3.01119924e-01 -7.48405337e-01
-3.48079711e-01 6.73724174e-01 1.36876032e-01 -2.14965150e-01
1.13909769e+00 -5.20986259e-01 -7.91392103e-02 6.58014178e-01
7.47178018e-01 -2.27992088e-01 -1.64079165e+00 -4.81384695e-01
9.46925357e-02 -1.84995934e-01 -1.35818601e-01 -7.06223011e-01
-6.93803847e-01 9.85631883e-01 3.14402133e-01 -1.47040576e-01
9.29456174e-01 4.37190309e-02 2.37691745e-01 1.02827501e+00
6.80395186e-01 -9.08421814e-01 4.73974675e-01 9.32066083e-01
9.10442352e-01 -1.72274637e+00 -2.09488124e-01 -1.97019726e-01
-9.59766448e-01 1.05073118e+00 9.84096825e-01 1.56902075e-01
3.97423029e-01 -2.24662602e-01 3.27689499e-01 -3.38447899e-01
-6.59649611e-01 -7.26664841e-01 3.57403666e-01 9.75750923e-01
-4.10105065e-02 -1.24604344e-01 4.19896126e-01 5.16175508e-01
-2.75691062e-01 -3.34292710e-01 1.37536958e-01 7.00831950e-01
-5.42712927e-01 -5.74278712e-01 -1.06959574e-01 7.48825520e-02
4.06711400e-01 -1.30112972e-02 -1.47740528e-01 7.91296422e-01
-2.36083716e-02 8.72111320e-01 -5.24253696e-02 4.04228009e-02
5.06960034e-01 -5.92598319e-02 4.21526372e-01 -5.34439504e-01
8.56795385e-02 -3.64472687e-01 2.02240180e-02 -1.17637718e+00
-7.67279804e-01 -5.98858356e-01 -1.61970770e+00 5.80990016e-02
3.70133519e-01 7.36614540e-02 3.97640973e-01 1.27602065e+00
5.01873996e-03 8.59312356e-01 -1.30123451e-01 -9.34580803e-01
-4.84667391e-01 -5.58959126e-01 -2.75460780e-01 5.51495135e-01
7.57095993e-01 -1.08197951e+00 -3.09358954e-01 1.07740305e-01] | [4.478273391723633, 0.492919921875] |
b839c710-aa65-406a-ab36-e37f7bfab555 | joint-topology-preserving-and-feature | null | null | http://openaccess.thecvf.com//content/ICCV2021/html/Cheng_Joint_Topology-Preserving_and_Feature-Refinement_Network_for_Curvilinear_Structure_Segmentation_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Cheng_Joint_Topology-Preserving_and_Feature-Refinement_Network_for_Curvilinear_Structure_Segmentation_ICCV_2021_paper.pdf | Joint Topology-Preserving and Feature-Refinement Network for Curvilinear Structure Segmentation | Curvilinear structure segmentation (CSS) is under semantic segmentation, whose applications include crack detection, aerial road extraction, and biomedical image segmentation. In general, geometric topology and pixel-wise features are two critical aspects of CSS. However, most semantic segmentation methods only focus on enhancing feature representations while existing CSS techniques emphasize preserving topology alone. In this paper, we present a Joint Topology-preserving and Feature-refinement Network (JTFN) that jointly models global topology and refined features based on an iterative feedback learning strategy. Specifically, we explore the structure of objects to help preserve corresponding topologies of predicted masks, thus design a reciprocative two-stream module for CSS and boundary detection. In addition, we introduce such topology-aware predictions as feedback guidance that refines attentive features by supplementing and enhancing saliencies. To the best of our knowledge, this is the first work that jointly addresses topology preserving and feature refinement for CSS. We evaluate JTFN on four datasets of diverse applications: Crack500, CrackTree200, Roads, and DRIVE. Results show that JTFN performs best in comparison with alternative methods. Code is available. | ['Jun Guo', 'Yajing Xu', 'Xuhong Guo', 'Kaili Zhao', 'Mingfei Cheng'] | 2021-01-01 | null | null | null | iccv-2021-1 | ['boundary-detection'] | ['computer-vision'] | [ 5.54696977e-01 1.51829943e-01 -2.59157866e-01 -3.67796898e-01
-3.34574342e-01 -5.51994503e-01 8.43414590e-02 4.40511107e-01
-5.16598113e-02 3.65722954e-01 -7.33640492e-02 -1.66572496e-01
-1.80935666e-01 -8.54448676e-01 -5.93018711e-01 -3.86209041e-01
-1.92433875e-02 4.33125049e-02 9.39346313e-01 -3.12947154e-01
7.63120770e-01 7.85917401e-01 -1.50895369e+00 1.36244506e-01
1.24529946e+00 1.00142813e+00 3.40743273e-01 4.90962654e-01
-1.71810269e-01 2.37287208e-01 -1.26895130e-01 -1.27126664e-01
2.60414869e-01 -7.28723709e-04 -1.09551597e+00 3.14116955e-01
2.41497427e-01 -1.47423133e-01 -1.09793082e-01 8.29330266e-01
3.97874475e-01 4.05248161e-03 7.88418233e-01 -1.09814656e+00
-4.04755592e-01 6.81734204e-01 -8.33370388e-01 6.92240968e-02
2.16729611e-01 3.17607448e-02 9.19546723e-01 -8.77926528e-01
6.90525353e-01 9.13464069e-01 8.04992497e-01 3.59320551e-01
-1.09392977e+00 -5.03123283e-01 4.02748555e-01 1.86813518e-01
-1.48084760e+00 -9.83636752e-02 1.35912633e+00 -3.49233568e-01
5.38530648e-01 3.72510046e-01 6.64348185e-01 3.83706897e-01
-3.57523398e-03 1.10481393e+00 7.52557039e-01 -2.66614795e-01
1.74810752e-01 -5.44488877e-02 5.20075485e-02 1.00464177e+00
4.65246849e-02 -1.97242975e-01 -5.90569377e-01 1.27143458e-01
1.08085990e+00 -8.28753710e-02 -3.32952499e-01 -8.26827407e-01
-1.01869512e+00 6.11016870e-01 6.69637024e-01 2.73375273e-01
-2.31159359e-01 1.92605898e-01 1.25696912e-01 -1.69514131e-03
4.40375835e-01 6.00912392e-01 -5.12143373e-01 2.55717903e-01
-1.05964410e+00 1.92489088e-01 4.86854076e-01 8.40682685e-01
1.05204403e+00 -1.65368959e-01 -4.51582074e-02 7.90142298e-01
3.27331215e-01 2.90960282e-01 -3.54999788e-02 -1.22966552e+00
2.70135164e-01 1.09775758e+00 -1.73936784e-01 -1.43091846e+00
-6.81381285e-01 -5.98359525e-01 -6.77720785e-01 1.11061603e-01
4.71775383e-02 -3.37740630e-02 -1.09473324e+00 1.33883107e+00
5.00291348e-01 3.45315754e-01 -3.82615060e-01 8.79634380e-01
9.04414594e-01 2.56002426e-01 -1.16338722e-01 1.65950149e-01
9.87349093e-01 -9.55038369e-01 -5.16838193e-01 -6.89126318e-03
6.01576746e-01 -6.49109423e-01 8.53980362e-01 4.21218663e-01
-1.11941004e+00 -4.02688056e-01 -1.10343981e+00 -2.92746518e-02
-4.13693100e-01 1.88795999e-01 5.29904425e-01 5.99934042e-01
-1.02624893e+00 7.48191476e-01 -9.10204649e-01 -2.19097152e-01
7.82403171e-01 3.90166759e-01 2.73779593e-02 1.31087840e-01
-9.94192183e-01 4.23166484e-01 1.04280494e-01 1.94250271e-02
-4.29327935e-01 -9.60087299e-01 -8.63291264e-01 -8.63628983e-02
5.52391112e-01 -5.47432125e-01 9.74871039e-01 -5.70672512e-01
-1.27909327e+00 6.85644150e-01 -8.52154046e-02 -2.87292928e-01
3.11120391e-01 -9.94223431e-02 -1.94475260e-02 5.61357975e-01
2.50627160e-01 1.11201978e+00 8.45317245e-01 -1.54889023e+00
-6.76923990e-01 -4.71294552e-01 9.34177637e-02 2.82938600e-01
-4.76044506e-01 -4.26411271e-01 -5.42369843e-01 -7.98230529e-01
6.72975242e-01 -3.39604825e-01 -5.14698565e-01 5.02150059e-01
-7.01167822e-01 -2.47560427e-01 1.29407322e+00 -5.71284175e-01
1.61073053e+00 -2.04037142e+00 9.19197276e-02 6.01082265e-01
2.79461533e-01 1.54724568e-02 -9.91480649e-02 3.13247681e-01
2.53721654e-01 4.62658286e-01 -8.58886063e-01 -2.33910680e-01
-3.86846662e-01 1.00021310e-01 6.19805902e-02 2.96323001e-01
5.33015251e-01 9.57385480e-01 -8.48657906e-01 -8.85254741e-01
3.10606390e-01 3.93290669e-01 -7.08424687e-01 -3.53948772e-01
-2.37624735e-01 4.05065775e-01 -8.10936034e-01 1.02407801e+00
7.69204080e-01 -3.23344529e-01 -2.24200636e-01 -5.71313858e-01
-3.85752827e-01 -2.15911046e-01 -1.32637823e+00 1.97199273e+00
-1.74593493e-01 3.40378195e-01 2.31493354e-01 -1.06798530e+00
1.11831450e+00 -1.90278795e-02 9.10255671e-01 -5.37921369e-01
1.28831223e-01 1.01578854e-01 -4.90191787e-01 -4.64756340e-01
6.23993456e-01 2.81824321e-01 5.25470190e-02 3.32800418e-01
-4.10672337e-01 -6.47973001e-01 1.43533424e-01 3.40304166e-01
8.29435647e-01 2.93097228e-01 -5.03682047e-02 -3.16968352e-01
5.16728997e-01 2.02608943e-01 5.94423950e-01 3.96683484e-01
-2.22030982e-01 9.10237074e-01 3.87885541e-01 -1.96439549e-01
-7.71630466e-01 -1.11374032e+00 -2.49029681e-01 7.46460378e-01
8.65450144e-01 -4.21329916e-01 -9.41502392e-01 -7.91748106e-01
3.95173468e-02 4.57437336e-01 -6.19520545e-01 -4.11266461e-02
-6.97570920e-01 -3.24100673e-01 3.67205203e-01 7.79375255e-01
6.52083457e-01 -9.39112306e-01 -8.46950293e-01 2.13892519e-01
-1.65765375e-01 -8.28925133e-01 -6.88894451e-01 7.98879638e-02
-1.07252002e+00 -1.43654704e+00 -5.67039251e-01 -1.05722666e+00
8.45281661e-01 4.03700143e-01 7.65672326e-01 4.79366273e-01
-6.58322453e-01 3.07090491e-01 -4.13038760e-01 -7.39478916e-02
1.42260179e-01 4.06542212e-01 -5.19835651e-01 -1.05174132e-01
-4.43239689e-01 -4.84757632e-01 -8.91812682e-01 6.21052742e-01
-9.39765096e-01 4.01707739e-01 5.62079132e-01 5.85394502e-01
9.20889556e-01 3.11387569e-01 5.07189989e-01 -1.01680470e+00
4.29871500e-01 -3.16014618e-01 -3.01459461e-01 3.29881907e-01
-7.16867208e-01 3.29639912e-02 2.31736615e-01 -2.31023766e-02
-1.11865175e+00 4.58005160e-01 -1.26279786e-01 -3.31013888e-01
-1.63702488e-01 3.33561331e-01 -7.24957436e-02 -2.43823722e-01
4.80079412e-01 1.87888205e-01 -9.54133943e-02 -3.43585610e-01
4.11197841e-01 4.11402851e-01 5.59890985e-01 -6.14568412e-01
7.28026271e-01 6.96064055e-01 6.63122535e-02 -7.91620255e-01
-6.16266549e-01 -6.41946495e-01 -9.77808237e-01 -4.13424820e-01
7.48519003e-01 -3.57667655e-01 -5.22343218e-01 6.92549765e-01
-8.60657930e-01 -4.03189629e-01 -3.59745830e-01 -4.57528979e-02
-6.35407984e-01 4.96315002e-01 -4.28100646e-01 -5.43214798e-01
-3.77186239e-01 -1.04782832e+00 1.25215065e+00 4.38906491e-01
2.19011642e-02 -8.70503068e-01 -2.68052965e-01 3.57176572e-01
3.81347179e-01 5.74044228e-01 8.37840021e-01 6.84006698e-03
-7.45815694e-01 1.66043296e-01 -3.52536529e-01 9.92398858e-02
4.04800415e-01 2.46407315e-01 -6.32154286e-01 -2.26832498e-02
-5.37629366e-01 -1.82250574e-01 1.05825937e+00 5.97862840e-01
1.57237816e+00 1.85387060e-01 -7.10687339e-01 6.41991019e-01
1.31529164e+00 8.73972923e-02 4.50698823e-01 1.76862881e-01
8.53447914e-01 8.35749686e-01 7.42164314e-01 4.88519549e-01
5.55005133e-01 3.84008855e-01 6.45785391e-01 -4.95603681e-01
-2.66396970e-01 -3.13564390e-01 -2.61414379e-01 4.08642828e-01
1.09136559e-01 -5.06688952e-02 -9.37190711e-01 6.93487465e-01
-1.58418810e+00 -5.33198893e-01 -4.56569940e-01 1.66835284e+00
6.98861778e-01 1.91988915e-01 -2.87743043e-02 3.85963351e-01
8.02385330e-01 7.97924623e-02 -9.57414687e-01 -8.48047808e-02
-9.29304585e-02 3.32077473e-01 4.60848004e-01 4.49222475e-01
-1.29532278e+00 1.16054440e+00 5.77117729e+00 9.37063277e-01
-9.17397738e-01 -1.66283652e-01 7.52701938e-01 3.42754662e-01
-7.30281889e-01 -4.96199615e-02 -5.02807140e-01 1.40631646e-01
-7.15955421e-02 2.50286967e-01 9.97558981e-02 5.87338209e-01
2.87268758e-01 -1.94294274e-01 -5.78847587e-01 6.98241174e-01
-1.23728484e-01 -1.53358245e+00 -2.00626627e-02 -2.22370118e-01
9.48295057e-01 -2.62105525e-01 1.28086656e-01 -3.00098091e-01
1.26325995e-01 -8.45139563e-01 7.91713178e-01 6.12154245e-01
8.43529582e-01 -9.13293242e-01 4.27584648e-01 8.35597888e-02
-1.65202057e+00 -1.39782175e-01 9.92797092e-02 4.51832354e-01
3.16067398e-01 6.67601168e-01 -6.54363692e-01 6.44178987e-01
7.80225337e-01 1.19107330e+00 -5.52535653e-01 1.18249500e+00
-2.04663306e-01 6.66163623e-01 -3.75392824e-01 1.89674526e-01
1.38119072e-01 3.80429476e-02 4.46734577e-01 1.06111729e+00
9.66490507e-02 2.72316724e-01 3.69049489e-01 1.08114779e+00
5.49218692e-02 2.66840994e-01 -2.90528238e-01 1.95929199e-01
6.20573997e-01 1.19647837e+00 -1.40649438e+00 3.37505335e-05
-2.13629887e-01 7.92001545e-01 6.30432293e-02 1.61430016e-01
-6.20307505e-01 -4.47019547e-01 4.42308992e-01 4.70216483e-01
4.91194546e-01 -1.85982287e-01 -1.03401411e+00 -6.80625677e-01
-1.31834209e-01 -1.85477227e-01 2.91658908e-01 -5.79840183e-01
-9.81235623e-01 2.46610314e-01 -9.09292400e-02 -9.80213881e-01
4.99808043e-01 -1.88269451e-01 -7.32618809e-01 3.97922218e-01
-1.76417351e+00 -1.30338621e+00 -4.08613950e-01 6.07467175e-01
8.57595921e-01 4.28650469e-01 3.01943302e-01 9.29795131e-02
-5.51006019e-01 4.05920476e-01 -2.56566882e-01 -5.33019640e-02
3.77367854e-01 -1.21961153e+00 3.14890891e-01 7.51334965e-01
-2.69081295e-01 3.45358282e-01 3.56686026e-01 -1.04252648e+00
-1.00787294e+00 -1.21247923e+00 3.36618453e-01 2.00033337e-01
3.40758741e-01 -1.18102536e-01 -1.03409421e+00 5.64631559e-02
-1.09798960e-01 1.00895576e-02 1.54709607e-01 -3.01878721e-01
6.90054670e-02 2.64425925e-03 -1.32192755e+00 4.89528865e-01
1.38698912e+00 -1.16784818e-01 -1.04925171e-01 -2.73905392e-03
7.92172968e-01 -4.05068278e-01 -1.00409901e+00 7.80879676e-01
5.74297190e-01 -8.65425467e-01 1.08645439e+00 2.18202956e-02
3.29604179e-01 -5.94603121e-01 1.04424439e-01 -1.04585493e+00
-3.88382703e-01 -4.28401023e-01 4.29100618e-02 1.10736811e+00
4.63038951e-01 -3.45078588e-01 1.27067983e+00 2.06528217e-01
-6.69087112e-01 -1.19281113e+00 -6.38497353e-01 -3.42481285e-01
-7.02992976e-02 -4.66696650e-01 7.71579921e-01 8.45367789e-01
-2.61465788e-01 -1.58800527e-01 7.49397129e-02 3.09745222e-01
6.65009081e-01 2.52275914e-01 3.48857850e-01 -1.43627560e+00
1.72710940e-01 -5.88647842e-01 -2.36364871e-01 -1.29919302e+00
-1.78253084e-01 -7.42879689e-01 1.39847070e-01 -1.74996269e+00
-1.05076944e-02 -9.62226570e-01 -1.86887577e-01 6.62947297e-01
-1.87589973e-01 3.34773362e-01 -1.59449875e-01 -9.39210355e-02
-4.82835859e-01 5.40603876e-01 1.71616721e+00 -2.24042669e-01
-4.58020806e-01 -2.29495112e-03 -8.13758731e-01 7.26138234e-01
1.24610651e+00 -1.47633389e-01 -6.89439356e-01 -3.21262926e-01
2.67325924e-03 -1.47972077e-01 3.82818729e-01 -8.87400568e-01
5.21398246e-01 -3.23939830e-01 2.77085721e-01 -1.10038733e+00
5.22549488e-02 -6.44311607e-01 -1.45512089e-01 5.84655464e-01
-2.96459883e-01 -2.82023698e-01 8.57596919e-02 6.22240961e-01
1.59928985e-02 -6.30533770e-02 7.69426465e-01 -4.87385727e-02
-9.90007579e-01 4.42683607e-01 -3.45214784e-01 2.77677234e-02
1.20583117e+00 -8.76279593e-01 -1.56465515e-01 -6.09226115e-02
-7.75227547e-01 6.08196199e-01 6.24763727e-01 3.00279677e-01
1.21406054e+00 -9.35266256e-01 -5.18522620e-01 4.79774773e-01
6.74807793e-03 5.50827920e-01 3.52548420e-01 9.01052117e-01
-5.75757504e-01 1.11196011e-01 -9.25900638e-02 -7.79393435e-01
-1.28040147e+00 1.41363874e-01 2.64236480e-01 -3.81296761e-02
-5.77504754e-01 9.30294991e-01 1.73930049e-01 -4.29997176e-01
1.23033412e-01 -5.57798445e-01 -4.66893911e-01 -1.19756535e-01
5.70905488e-03 6.65898860e-01 5.64298108e-02 -3.83930802e-01
-3.00081372e-01 1.02990222e+00 -1.31095544e-01 2.40082636e-01
1.40891171e+00 -4.70984161e-01 -1.95898060e-02 -7.78934732e-02
7.39266515e-01 -1.73040614e-01 -1.56449819e+00 -1.78261787e-01
1.44580919e-02 -3.43044400e-01 2.08915934e-01 -6.97706640e-01
-1.58328772e+00 7.56258845e-01 6.26895487e-01 1.12110339e-01
1.42823243e+00 2.49700323e-01 1.08142698e+00 -3.38923349e-03
3.76275092e-01 -1.29693115e+00 4.86601025e-01 1.84385821e-01
8.02825153e-01 -8.20895195e-01 1.50394281e-02 -1.20336998e+00
-3.63597691e-01 1.07645071e+00 8.99815559e-01 -1.50639445e-01
8.98162723e-01 3.83318424e-01 -2.85679966e-01 -5.45440316e-01
-2.82368273e-01 -2.27189124e-01 3.30341280e-01 6.04383290e-01
1.39846116e-01 -1.43490639e-02 -3.60747308e-01 3.56538564e-01
-1.30872250e-01 -1.47732794e-01 2.83507884e-01 1.18475616e+00
-8.65341067e-01 -1.16209674e+00 -3.41692924e-01 6.97131455e-01
-3.97970192e-02 5.09987492e-03 -4.51716930e-01 5.33139110e-01
4.28636789e-01 9.24163997e-01 9.25444365e-02 -5.41250408e-01
2.78025061e-01 -5.47250390e-01 3.08772117e-01 -6.56581342e-01
-4.45380688e-01 4.77448152e-03 -1.61815733e-01 -6.91885471e-01
-4.84537959e-01 -7.68272102e-01 -1.84239113e+00 1.44001663e-01
-6.00395381e-01 -8.54496062e-02 6.96171224e-01 7.37106502e-01
4.82373387e-01 6.83765173e-01 8.63240838e-01 -7.86601067e-01
4.09522764e-02 -4.95577574e-01 -5.74970841e-01 2.10214660e-01
6.71161637e-02 -9.15160060e-01 -8.25516284e-02 3.24647069e-01] | [9.357362747192383, -0.1574830412864685] |
9ae543c7-543b-4544-857b-b365d6ed7e03 | deep-contextual-recurrent-residual-networks | 1704.03594 | null | http://arxiv.org/abs/1704.03594v1 | http://arxiv.org/pdf/1704.03594v1.pdf | Deep Contextual Recurrent Residual Networks for Scene Labeling | Designed as extremely deep architectures, deep residual networks which
provide a rich visual representation and offer robust convergence behaviors
have recently achieved exceptional performance in numerous computer vision
problems. Being directly applied to a scene labeling problem, however, they
were limited to capture long-range contextual dependence, which is a critical
aspect. To address this issue, we propose a novel approach, Contextual
Recurrent Residual Networks (CRRN) which is able to simultaneously handle rich
visual representation learning and long-range context modeling within a fully
end-to-end deep network. Furthermore, our proposed end-to-end CRRN is
completely trained from scratch, without using any pre-trained models in
contrast to most existing methods usually fine-tuned from the state-of-the-art
pre-trained models, e.g. VGG-16, ResNet, etc. The experiments are conducted on
four challenging scene labeling datasets, i.e. SiftFlow, CamVid, Stanford
background and SUN datasets, and compared against various state-of-the-art
scene labeling methods. | ['Khoa Luu', 'T. Hoang Ngan Le', 'Marios Savvides', 'Dipan Pal', 'Chi Nhan Duong', 'Ligong Han'] | 2017-04-12 | null | null | null | null | ['scene-labeling'] | ['computer-vision'] | [ 2.14353025e-01 -1.82454228e-01 -1.78370133e-01 -6.35174394e-01
-3.72701824e-01 -2.41218284e-01 6.09936476e-01 -1.29292086e-01
-5.50902665e-01 5.50074577e-01 2.60777652e-01 -2.54675388e-01
1.65452272e-01 -6.37422383e-01 -6.87131226e-01 -4.66312200e-01
3.42401743e-01 1.31626114e-01 3.15706789e-01 -3.53834629e-01
1.63243040e-01 4.13661033e-01 -1.55527914e+00 1.12647630e-01
7.24248290e-01 9.33818042e-01 4.94208723e-01 4.59730178e-01
-2.21649632e-01 1.52347755e+00 -3.88040006e-01 -1.14523403e-01
8.77845362e-02 -2.91567236e-01 -7.91404605e-01 3.25827837e-01
6.38406277e-01 -2.67398655e-01 -7.52733052e-01 1.10853171e+00
3.81441087e-01 4.20917898e-01 3.50511789e-01 -8.30103576e-01
-8.92716408e-01 3.91229749e-01 -7.43282557e-01 2.53005922e-01
3.47625054e-02 1.92137703e-01 8.39721084e-01 -9.35245454e-01
8.56137097e-01 1.23316419e+00 5.48857510e-01 5.58086276e-01
-9.56573188e-01 -4.33549881e-01 5.97690940e-01 1.81735814e-01
-1.29971731e+00 -2.23990783e-01 1.06809723e+00 -3.80414277e-01
9.15361345e-01 -1.11979648e-01 3.20725054e-01 1.25172842e+00
1.23018874e-02 8.94839406e-01 1.12836289e+00 -2.30376124e-01
1.27215181e-02 -1.51791945e-01 2.25170106e-01 6.15322709e-01
-2.40903459e-02 3.66709568e-02 1.30329151e-02 2.98220754e-01
9.00305390e-01 4.87587124e-01 -2.20591024e-01 -3.82249147e-01
-1.10962105e+00 7.79242158e-01 1.02088034e+00 3.80938232e-01
-4.28434491e-01 4.68197763e-01 7.50624776e-01 9.21242386e-02
4.39718425e-01 7.31352419e-02 -5.14527261e-01 2.97953933e-01
-8.32510769e-01 1.30523652e-01 2.09845096e-01 9.97220218e-01
8.61806750e-01 5.02633095e-01 -2.92479306e-01 9.65657055e-01
3.36839646e-01 2.64759570e-01 5.68512261e-01 -4.33473349e-01
4.24214780e-01 7.50525117e-01 -1.56142727e-01 -9.65886474e-01
-6.33692265e-01 -9.16300774e-01 -1.17636967e+00 5.40993996e-02
6.02080412e-02 3.97019833e-02 -1.51788402e+00 1.81068552e+00
3.13598484e-01 6.16519451e-01 2.67284334e-01 1.18815362e+00
1.27817249e+00 8.29687595e-01 5.03445506e-01 1.51782304e-01
1.28245950e+00 -1.53421044e+00 -4.92597789e-01 -5.83676398e-01
5.14758825e-01 -7.49067187e-01 1.25305378e+00 1.40505850e-01
-7.90214658e-01 -1.13235056e+00 -9.83966053e-01 -4.54825819e-01
-5.61976910e-01 3.47711891e-01 7.95395911e-01 1.12002678e-01
-1.09813821e+00 3.40036362e-01 -5.97786307e-01 -3.80599737e-01
5.14783680e-01 -1.54700810e-02 -3.58537018e-01 -4.99511063e-01
-8.81491780e-01 7.59447157e-01 5.12856185e-01 4.95604903e-01
-1.37078416e+00 -3.94059300e-01 -1.09849703e+00 9.20613781e-02
5.64762473e-01 -6.77155375e-01 1.04867554e+00 -1.11619830e+00
-1.58204496e+00 1.06058156e+00 4.08551544e-02 -4.79496539e-01
4.76449251e-01 -2.98879445e-01 -3.66993368e-01 -6.70965314e-02
9.16032940e-02 6.14239991e-01 7.88479090e-01 -1.25562561e+00
-3.99223953e-01 -2.03552276e-01 2.86322057e-01 1.70612440e-01
1.23296864e-02 2.73632463e-02 -4.74500418e-01 -8.15411568e-01
-5.32084629e-02 -9.14312482e-01 -9.71365809e-01 -2.77842224e-01
-5.73918760e-01 -3.19587916e-01 9.94572163e-01 -4.00871128e-01
7.60962009e-01 -2.29999924e+00 1.65898561e-01 -2.06635833e-01
1.15149446e-01 6.98921204e-01 -4.27812189e-01 3.05255353e-01
-1.85635448e-01 -8.49104971e-02 -3.95673931e-01 -5.85288048e-01
-6.21010549e-02 1.89124748e-01 -2.88923770e-01 6.55319810e-01
1.08136185e-01 1.06057525e+00 -9.38130677e-01 -3.87439936e-01
7.70371854e-01 6.94626153e-01 -2.63425499e-01 5.04581809e-01
-5.80264747e-01 5.22089601e-01 -3.62257540e-01 6.01850569e-01
6.84036672e-01 -5.84693909e-01 2.49066278e-02 -1.24779284e-01
-9.85251218e-02 -1.18212746e-02 -9.07860637e-01 2.17140317e+00
-7.34447539e-01 6.63396418e-01 -1.16480216e-01 -1.19912004e+00
1.12569058e+00 -4.21044119e-02 9.16873366e-02 -8.86368275e-01
3.36307675e-01 -5.99108683e-03 -3.61000955e-01 -3.70612949e-01
8.09513927e-01 6.13495186e-02 -1.40102416e-01 -8.75839684e-03
2.63540000e-01 2.58019120e-01 1.28418475e-01 1.27209082e-01
7.98498690e-01 3.70330870e-01 3.17862093e-01 -2.44217023e-01
7.26109684e-01 2.85486821e-02 7.53531814e-01 5.47209620e-01
-1.63427815e-01 8.36525202e-01 1.22831598e-01 -6.95913017e-01
-9.79707241e-01 -6.25338554e-01 -1.78850833e-02 1.24109101e+00
5.17601848e-01 -3.53093892e-01 -4.70402122e-01 -6.73945129e-01
-3.62939835e-01 5.40372372e-01 -7.82903969e-01 -7.15074167e-02
-5.57516754e-01 -5.21735489e-01 3.87332350e-01 7.76507616e-01
1.04899633e+00 -1.38284564e+00 -7.32100308e-01 3.71338546e-01
5.02825826e-02 -1.46154428e+00 -2.75424719e-01 1.37997985e-01
-6.64993167e-01 -1.00155759e+00 -7.79717147e-01 -9.20862138e-01
5.67736804e-01 5.23263872e-01 1.20061707e+00 2.24458948e-01
-5.81986070e-01 9.36256871e-02 -4.52627897e-01 6.93299621e-03
7.12213963e-02 2.27707177e-01 -3.54851633e-01 1.67627364e-01
1.95874766e-01 -5.43312371e-01 -9.10304427e-01 3.20810556e-01
-1.16362548e+00 1.32070854e-01 7.35018909e-01 9.98035133e-01
8.53382349e-01 -7.63335526e-02 6.15974188e-01 -1.27033281e+00
7.70705342e-02 -4.63898748e-01 -6.21990085e-01 3.56857717e-01
-3.43424469e-01 -7.13299513e-02 8.08231592e-01 -3.30003798e-01
-1.36921406e+00 2.52153248e-01 -4.47035819e-01 -7.87492692e-01
-5.00473857e-01 7.20826924e-01 -1.27886534e-01 -1.07515931e-01
5.64922273e-01 2.59672701e-01 -5.37268341e-01 -5.13512492e-01
8.02636802e-01 3.67754966e-01 9.00872290e-01 -3.46040100e-01
6.79407001e-01 3.89910102e-01 -9.26007703e-02 -8.82515788e-01
-1.09468544e+00 -7.31623232e-01 -7.12454140e-01 4.34789918e-02
1.08137536e+00 -1.17136955e+00 -2.69474566e-01 5.39799213e-01
-1.06337607e+00 -6.79394364e-01 -2.76707530e-01 3.27171683e-01
-5.13247013e-01 2.92487323e-01 -6.54249728e-01 -5.15751719e-01
-4.67832893e-01 -1.13254404e+00 1.14163840e+00 4.64423567e-01
3.05788070e-01 -1.04999983e+00 2.00388461e-01 2.12572470e-01
6.45497620e-01 7.87504315e-01 6.33605361e-01 -3.02718431e-01
-5.89515269e-01 -1.47143722e-01 -6.59752011e-01 3.20061743e-01
-2.61771698e-02 -4.52797711e-02 -1.07770753e+00 -3.80788088e-01
-2.12269366e-01 -6.65585339e-01 1.31013691e+00 3.82174760e-01
1.36939800e+00 -4.03770693e-02 -2.82222569e-01 1.08590019e+00
1.86615252e+00 -9.54820365e-02 9.36545312e-01 3.15535963e-01
1.21194088e+00 4.14843500e-01 6.66344285e-01 2.27323458e-01
4.57695574e-01 5.33818066e-01 7.67170906e-01 -4.75502014e-01
-4.20302063e-01 -4.97775018e-01 -3.83766145e-02 5.78264296e-01
-4.37083561e-03 -3.51064563e-01 -7.66133428e-01 6.14045978e-01
-2.16746235e+00 -7.68740237e-01 -1.67337656e-02 1.77179873e+00
2.59526193e-01 1.41133308e-01 -1.38635650e-01 -2.02483758e-01
7.21704960e-01 8.24952304e-01 -7.62886047e-01 -1.22356758e-01
-2.48992741e-01 2.48832256e-02 6.80208027e-01 1.11224927e-01
-1.27703249e+00 1.61925459e+00 5.55160522e+00 8.24105263e-01
-1.36092842e+00 2.08149314e-01 7.29166865e-01 3.68400574e-01
-1.01195693e-01 4.99981157e-02 -6.00668609e-01 1.84781566e-01
6.12880945e-01 3.11853737e-01 1.82284847e-01 1.32674277e+00
-1.84354365e-01 1.88668236e-01 -8.11492503e-01 1.19601846e+00
1.21057890e-01 -1.42495060e+00 2.15220898e-02 -1.26836658e-01
9.82605577e-01 4.47041065e-01 1.07849598e-01 5.93968570e-01
4.76420641e-01 -1.16246176e+00 5.34310758e-01 3.20199758e-01
9.99265730e-01 -8.02504778e-01 8.74899089e-01 1.17754839e-01
-1.52468848e+00 -3.62837464e-02 -8.12689722e-01 1.71642639e-02
2.98065305e-01 6.15606964e-01 -3.85705471e-01 8.28941584e-01
6.96770549e-01 1.18223619e+00 -8.27708840e-01 9.39294338e-01
-3.89666587e-01 5.46971440e-01 1.93382744e-02 2.17214644e-01
8.15285444e-01 3.24796848e-02 1.53084174e-01 1.41517830e+00
-1.00135893e-01 4.38545458e-02 6.01877689e-01 8.20919096e-01
-3.28202307e-01 -3.87274772e-02 -8.09472501e-01 8.58101100e-02
5.74765913e-02 1.55571842e+00 -1.02461243e+00 -4.96754318e-01
-4.80700433e-01 8.97085965e-01 6.02171779e-01 6.39228165e-01
-8.56314898e-01 -3.54276538e-01 4.85293537e-01 -1.70070753e-01
3.36281240e-01 -2.77054459e-01 5.31269163e-02 -1.45065594e+00
-2.30648965e-01 -7.75225997e-01 3.78785461e-01 -7.52423108e-01
-1.38577902e+00 1.06519330e+00 -1.97199836e-01 -1.21218204e+00
-2.33420789e-01 -5.79151511e-01 -7.68561959e-01 4.80233908e-01
-1.98806119e+00 -1.55604649e+00 -6.40091538e-01 9.73371685e-01
8.81464839e-01 -2.44951807e-02 4.67495471e-01 3.67786199e-01
-5.94282150e-01 4.98542219e-01 -1.25714064e-01 3.28409076e-01
4.41068023e-01 -9.80302274e-01 6.36868775e-01 1.09789455e+00
2.82389551e-01 5.07784188e-01 4.30654556e-01 -3.70828658e-01
-1.39413774e+00 -1.71013343e+00 3.05123091e-01 -6.70462400e-02
2.74534643e-01 -4.11166668e-01 -9.10330594e-01 9.29345310e-01
4.41733241e-01 7.21285820e-01 1.99827105e-01 8.56593177e-02
-4.94700551e-01 -5.85177876e-02 -9.44217324e-01 5.62954724e-01
1.46070504e+00 -6.49155974e-01 -4.74312216e-01 3.40221524e-01
1.00659454e+00 -5.87662816e-01 -3.07589412e-01 7.47300029e-01
9.36528519e-02 -8.87057304e-01 1.08656502e+00 -5.38510621e-01
4.45403248e-01 -4.37120318e-01 -3.86409253e-01 -1.13720477e+00
-4.16241050e-01 -4.42874432e-01 1.84326679e-01 1.38265502e+00
1.16432859e-02 -4.87454981e-01 6.46718562e-01 2.20424652e-01
-4.70824808e-01 -6.77314758e-01 -5.29113054e-01 -5.77597737e-01
-2.10168660e-01 -2.87579834e-01 5.52920759e-01 1.15668607e+00
-6.78737283e-01 6.64249063e-01 -7.99515128e-01 8.36964399e-02
7.31604517e-01 4.15975958e-01 9.73281026e-01 -1.05649173e+00
-9.97511968e-02 -3.25355530e-01 -6.74082398e-01 -1.25937927e+00
5.14507711e-01 -8.26740205e-01 1.63671419e-01 -1.81227803e+00
1.54405609e-01 -5.96675515e-01 -5.72379827e-01 4.25088584e-01
-1.70023829e-01 4.49664414e-01 2.54993439e-01 1.91208385e-02
-9.63928699e-01 8.40160608e-01 1.12981832e+00 -3.19193363e-01
-1.96289241e-01 -3.14989328e-01 -5.87742984e-01 8.68808806e-01
7.16564655e-01 -3.14247429e-01 -6.47552371e-01 -5.22542894e-01
2.80409362e-02 -1.68588180e-02 4.62414354e-01 -1.04143465e+00
3.17845434e-01 -2.27552310e-01 4.48285431e-01 -8.75310719e-01
2.44646713e-01 -6.98671639e-01 3.43121104e-02 1.68100506e-01
-4.42632437e-01 8.50309953e-02 1.99665278e-01 8.09506178e-01
-4.89255488e-01 -4.71783690e-02 9.00195479e-01 -3.86772454e-01
-1.39580965e+00 6.42138004e-01 -4.52471226e-02 1.28971308e-01
1.02189445e+00 6.44780248e-02 -5.35120428e-01 -2.30302647e-01
-5.25304914e-01 2.85448104e-01 4.76146698e-01 5.56716502e-01
8.41222346e-01 -1.29094076e+00 -6.61240757e-01 -3.94750796e-02
3.35465699e-01 4.55731779e-01 7.07990229e-01 3.16300094e-01
-6.89990520e-01 3.90791476e-01 -2.83440024e-01 -7.96719849e-01
-9.81557429e-01 1.07309651e+00 2.58260489e-01 -4.48133469e-01
-1.04153919e+00 8.64786386e-01 6.07850134e-01 -5.87414622e-01
3.57568592e-01 -4.43277121e-01 -2.94994414e-01 -3.26338917e-01
5.28160214e-01 -5.64876385e-02 -1.75640091e-01 -7.87529230e-01
-3.87432128e-01 8.77359092e-01 -1.69405103e-01 3.11583042e-01
1.32057679e+00 -2.73959875e-01 6.61656633e-02 2.67102599e-01
1.36961234e+00 -4.23965752e-01 -1.49733043e+00 -5.99599719e-01
-1.16774783e-01 -4.79025364e-01 1.67935088e-01 -5.55072546e-01
-1.55397463e+00 1.07332861e+00 6.03968918e-01 -7.22563788e-02
1.14352274e+00 -1.44584805e-01 8.15111935e-01 4.82278347e-01
4.55951810e-01 -7.91161418e-01 3.37257683e-01 5.56493640e-01
7.07068622e-01 -1.43999970e+00 2.13529263e-02 -1.97854251e-01
-8.45057189e-01 9.40263629e-01 8.94001782e-01 -7.03240693e-01
5.10183156e-01 -1.10164426e-01 1.80038139e-01 -3.39501649e-01
-6.36476219e-01 -6.49857938e-01 1.51982725e-01 3.86033267e-01
3.69047552e-01 -8.25709850e-02 4.99331318e-02 1.85683489e-01
1.60877839e-01 7.16818571e-02 4.71087396e-01 8.89178455e-01
-2.72278398e-01 -7.70123780e-01 1.14537038e-01 1.39728785e-01
-5.46527863e-01 -6.34811744e-02 -1.02868207e-01 8.40269148e-01
-8.92733261e-02 8.55280459e-01 -1.44228429e-01 -3.51471990e-01
2.83709407e-01 -3.74329835e-01 1.67440996e-01 -6.24613941e-01
-5.42354047e-01 -4.42430675e-02 -4.48620804e-02 -7.42581904e-01
-7.04464734e-01 -1.76423743e-01 -1.13921797e+00 -9.94118601e-02
-2.46286735e-01 -9.58566591e-02 6.01432920e-01 8.57941508e-01
2.99144626e-01 9.53599811e-01 6.21389151e-01 -1.12421954e+00
-1.54925585e-01 -1.04302061e+00 -4.28645968e-01 4.94829327e-01
4.91605282e-01 -6.77891731e-01 -8.95412564e-02 -1.09962814e-01] | [9.569232940673828, 0.38557109236717224] |
f1c5e153-bee1-4014-92b6-fe87d8fbd9aa | variational-image-restoration-network | 2008.10796 | null | https://arxiv.org/abs/2008.10796v3 | https://arxiv.org/pdf/2008.10796v3.pdf | Deep Variational Network Toward Blind Image Restoration | Blind image restoration (IR) is a common yet challenging problem in computer vision. Classical model-based methods and recent deep learning (DL)-based methods represent two different methodologies for this problem, each with their own merits and drawbacks. In this paper, we propose a novel blind image restoration method, aiming to integrate both the advantages of them. Specifically, we construct a general Bayesian generative model for the blind IR, which explicitly depicts the degradation process. In this proposed model, a pixel-wise non-i.i.d. Gaussian distribution is employed to fit the image noise. It is with more flexibility than the simple i.i.d. Gaussian or Laplacian distributions as adopted in most of conventional methods, so as to handle more complicated noise types contained in the image degradation. To solve the model, we design a variational inference algorithm where all the expected posteriori distributions are parameterized as deep neural networks to increase their model capability. Notably, such an inference algorithm induces a unified framework to jointly deal with the tasks of degradation estimation and image restoration. Further, the degradation information estimated in the former task is utilized to guide the latter IR process. Experiments on two typical blind IR tasks, namely image denoising and super-resolution, demonstrate that the proposed method achieves superior performance over current state-of-the-arts. | ['Kwan-Yen K. Wong', 'Qian Zhao', 'Lei Zhang', 'Hongwei Yong', 'Zongsheng Yue', 'Deyu Meng'] | 2020-08-25 | null | null | null | null | ['image-deblocking'] | ['computer-vision'] | [ 2.66463488e-01 -4.41515446e-01 3.19679715e-02 -6.46541119e-02
-7.11206973e-01 -1.11985452e-01 6.57193303e-01 -5.56535661e-01
-2.15115607e-01 5.27084589e-01 2.90041149e-01 -4.45004180e-02
-2.79074520e-01 -4.25404459e-01 -4.79853839e-01 -1.20076108e+00
5.55996239e-01 2.49802228e-02 -9.70795974e-02 -6.77639320e-02
3.29634398e-01 2.82973379e-01 -1.56777537e+00 -1.82275340e-01
1.37085676e+00 1.06440139e+00 5.23312390e-01 3.22336733e-01
-6.12731576e-02 6.85223877e-01 -3.49120677e-01 -3.01940084e-01
1.23412088e-01 -4.06141400e-01 -3.12299371e-01 5.38865149e-01
1.07792057e-01 -4.10574406e-01 -6.30671084e-01 1.55517435e+00
6.23931706e-01 3.14844936e-01 9.35579538e-01 -8.10440660e-01
-1.01568973e+00 1.24575265e-01 -7.86709070e-01 2.28483975e-01
8.15331861e-02 1.33457616e-01 6.37196779e-01 -9.13528442e-01
1.03254087e-01 1.47856283e+00 3.77184212e-01 4.20500040e-01
-1.17792928e+00 -2.96157032e-01 2.08523616e-01 4.17433709e-01
-1.31170845e+00 -6.31170690e-01 1.03866279e+00 -6.26972377e-01
2.18301460e-01 1.53690614e-02 2.29952216e-01 1.17333949e+00
1.56539142e-01 7.74075925e-01 1.27370179e+00 -3.73787105e-01
3.36336881e-01 -3.06002218e-02 8.61330852e-02 2.68267095e-01
1.58913687e-01 1.41923174e-01 -3.33219618e-01 -4.33788300e-02
8.64030898e-01 1.18910849e-01 -6.13238752e-01 -3.77130747e-01
-9.00650918e-01 7.50108182e-01 4.32053804e-01 3.15480709e-01
-5.18252432e-01 -4.19200771e-02 1.01734675e-01 -1.21471390e-01
6.82779968e-01 -2.17242762e-01 2.16731802e-03 4.19475526e-01
-9.49252486e-01 1.10655613e-01 3.48483294e-01 5.93516350e-01
6.68248117e-01 3.28029871e-01 -5.04952073e-01 1.12830377e+00
8.34130228e-01 5.22332847e-01 4.25482601e-01 -1.04743886e+00
1.33811861e-01 1.19665548e-01 4.61053908e-01 -1.05585086e+00
-4.72828224e-02 -7.73471177e-01 -1.46152389e+00 2.85478950e-01
2.59034932e-01 1.80657044e-01 -1.13336551e+00 1.65168417e+00
2.75461257e-01 3.43882710e-01 9.65272188e-02 1.08023000e+00
6.66346014e-01 7.49844074e-01 6.23040134e-03 -6.54268503e-01
1.45976555e+00 -7.32839823e-01 -1.18922102e+00 -2.88033754e-01
-4.70205247e-01 -8.75007331e-01 8.52416158e-01 7.19105005e-01
-1.11744332e+00 -8.76328468e-01 -1.02185857e+00 -6.40333220e-02
9.75449309e-02 4.31146801e-01 2.37939924e-01 6.08570457e-01
-1.14676595e+00 4.27973270e-01 -6.75768316e-01 -1.41935378e-01
3.95407468e-01 -8.18410739e-02 1.10516720e-01 -3.42921525e-01
-1.05624104e+00 8.55918169e-01 1.22271568e-01 6.66162670e-01
-1.08286703e+00 -3.30717534e-01 -7.52423584e-01 -6.86877668e-02
2.18932003e-01 -1.06889713e+00 1.18565738e+00 -8.29666793e-01
-1.67483008e+00 6.60403430e-01 -5.16804218e-01 -2.06867203e-01
6.66340172e-01 -3.36210221e-01 -4.36392069e-01 1.72961846e-01
-5.95308654e-02 1.91387266e-01 1.62630641e+00 -1.85049069e+00
-2.39515424e-01 -4.41546559e-01 -1.27716765e-01 2.72249788e-01
-3.87818068e-01 4.34750952e-02 -8.09811354e-01 -8.69564533e-01
3.97485316e-01 -5.02901852e-01 -1.86425224e-01 -2.38412116e-02
-4.54590648e-01 -5.11719333e-03 4.98180330e-01 -1.12150800e+00
1.15454125e+00 -2.25815368e+00 4.85006392e-01 -6.71020374e-02
1.31525502e-01 4.08621669e-01 6.58251047e-02 2.33258992e-01
1.53674498e-01 -2.26752728e-01 -6.89318657e-01 -8.12893391e-01
1.50833532e-01 2.70497620e-01 -2.81433672e-01 7.34963059e-01
1.00380205e-01 5.04474103e-01 -6.67050838e-01 -4.25843477e-01
4.21065062e-01 1.00775754e+00 -2.25181818e-01 3.15666199e-01
-1.22950554e-01 9.84808922e-01 -3.79271924e-01 4.80916500e-01
1.06240487e+00 -6.07619286e-02 5.39425872e-02 -5.63135684e-01
-9.70941633e-02 -2.45947123e-01 -1.27063715e+00 1.68552744e+00
-4.51638609e-01 2.42183894e-01 5.29070854e-01 -1.35681987e+00
9.97500777e-01 4.41977203e-01 2.28501797e-01 -5.89968503e-01
1.53918460e-01 2.27624550e-01 -2.25844458e-01 -7.75004506e-01
2.31169462e-01 -3.88513625e-01 4.60082412e-01 -1.08564161e-02
-6.63910583e-02 -8.00598040e-03 4.13029268e-02 -1.31589055e-01
5.71717322e-01 2.54574478e-01 2.38614455e-01 -1.39600992e-01
9.50002968e-01 -4.74832714e-01 7.88303971e-01 8.31245542e-01
-3.11997563e-01 7.72830486e-01 1.28090540e-02 -8.32490325e-02
-8.90004992e-01 -1.09472430e+00 -3.48603994e-01 5.75988173e-01
4.30596054e-01 2.22825587e-01 -8.21778297e-01 -1.68117538e-01
-2.68262178e-01 5.71790755e-01 -2.87389755e-01 -1.11511290e-01
-3.46166551e-01 -1.13768673e+00 2.07345672e-02 1.64182544e-01
7.81076193e-01 -8.97730052e-01 -3.19944993e-02 1.95550248e-01
-4.79221821e-01 -1.11071837e+00 -3.35705221e-01 -2.90157229e-01
-7.84287333e-01 -8.28218400e-01 -1.12363231e+00 -8.43202770e-01
6.90059543e-01 4.91238445e-01 9.43142533e-01 -1.12842731e-01
-2.00071648e-01 4.18032885e-01 -2.73754239e-01 9.14351344e-02
-4.31434840e-01 -7.08384097e-01 1.95657656e-01 6.31071985e-01
1.62797496e-01 -7.31774628e-01 -8.50276589e-01 1.78192928e-01
-1.24660563e+00 -1.03365146e-01 7.23729670e-01 1.00994980e+00
7.29591966e-01 6.20380819e-01 5.46705067e-01 -3.35785955e-01
6.96902215e-01 -4.52991843e-01 -5.62445521e-01 1.87906101e-01
-5.41524053e-01 3.70354690e-02 4.92553771e-01 -4.23126578e-01
-1.67402899e+00 -7.56129473e-02 -1.57032490e-01 -5.68392634e-01
-2.67650783e-01 5.14296353e-01 -6.02875710e-01 -2.43196990e-02
2.81437159e-01 7.04994321e-01 7.75073990e-02 -9.58419740e-01
4.57906902e-01 8.19365621e-01 8.94398570e-01 -5.24810672e-01
1.01843357e+00 5.46536684e-01 -1.00716099e-01 -7.70949841e-01
-8.30104828e-01 -4.86901820e-01 -4.21838284e-01 -1.60346329e-01
9.75775063e-01 -1.10685730e+00 -7.12190807e-01 9.49167490e-01
-1.24643540e+00 -2.69177742e-02 1.44558534e-01 4.31416273e-01
-6.77500844e-01 8.55825365e-01 -6.41588807e-01 -1.23363340e+00
-3.55723917e-01 -1.44303751e+00 9.11380351e-01 4.22295272e-01
4.60949749e-01 -1.00218642e+00 5.14457487e-02 5.18960118e-01
4.69199687e-01 -8.11594948e-02 8.54003549e-01 -4.08380181e-02
-5.04773498e-01 5.00645442e-03 -5.55474162e-01 9.89940643e-01
2.52233297e-01 -2.29191169e-01 -1.07152951e+00 -3.61387312e-01
6.41836464e-01 1.16583057e-01 1.10918629e+00 7.97066450e-01
1.16918337e+00 -2.94766039e-01 -3.58682983e-02 6.66851223e-01
1.65368140e+00 2.20311612e-01 9.02079880e-01 3.15226704e-01
6.43051565e-01 5.64049482e-01 5.27449429e-01 5.43951213e-01
2.83626556e-01 6.87412918e-01 6.49105310e-01 -1.60016958e-02
-2.62984276e-01 -1.57813374e-02 4.72347319e-01 8.34721208e-01
-3.23065631e-02 -2.85347730e-01 -5.41846156e-01 4.48451310e-01
-1.93121088e+00 -8.12621415e-01 -1.18212499e-01 2.25291348e+00
8.08325112e-01 -1.52814150e-01 -3.04604977e-01 1.40307039e-01
9.66972351e-01 2.98379421e-01 -6.43684447e-01 1.83086798e-01
-2.73569286e-01 2.43299585e-02 1.21020533e-01 6.09949648e-01
-1.16313171e+00 5.76414645e-01 5.57520008e+00 1.00831032e+00
-6.76440597e-01 1.41328752e-01 5.87799489e-01 3.92947972e-01
-2.99804419e-01 -9.10860151e-02 -5.54974854e-01 6.31208718e-01
4.22009319e-01 1.94824830e-01 7.30882287e-01 2.67783284e-01
7.06741512e-01 -1.32651225e-01 -6.58977747e-01 1.21422923e+00
1.60713628e-01 -8.13044012e-01 1.75351426e-01 -2.64032427e-02
6.81377769e-01 -3.72204483e-01 3.27221096e-01 -5.10310754e-02
3.50840725e-02 -9.12036002e-01 7.09693432e-01 1.11888254e+00
5.77625334e-01 -6.16273820e-01 7.62975037e-01 4.86227095e-01
-8.12974989e-01 -2.23491058e-01 -4.78410244e-01 9.49645340e-02
3.91039342e-01 1.04159451e+00 1.04961753e-01 9.99748468e-01
8.16537678e-01 9.67100561e-01 -2.93858141e-01 1.27532673e+00
-4.52003747e-01 4.80727971e-01 1.15980133e-01 6.40119672e-01
-9.53575745e-02 -7.17577875e-01 8.88076901e-01 9.33609545e-01
5.34138560e-01 -1.82012826e-01 4.40881103e-02 1.11620486e+00
6.90457374e-02 -1.36938244e-01 -1.24955364e-01 2.41682559e-01
2.50623196e-01 1.25308740e+00 -3.20277423e-01 -3.10967118e-01
-3.59646708e-01 1.19754791e+00 -3.83150019e-02 8.07561100e-01
-7.17901289e-01 -6.38129637e-02 7.21086562e-01 -2.81030029e-01
4.29873675e-01 -2.28356734e-01 -2.39858717e-01 -1.41704547e+00
1.73739985e-01 -9.90055561e-01 5.18391617e-02 -9.54738677e-01
-1.70898628e+00 4.64062780e-01 -9.83204395e-02 -1.39214575e+00
-3.72425164e-03 -6.22933149e-01 -6.98812246e-01 1.25928640e+00
-2.01971674e+00 -1.08750439e+00 -4.30960149e-01 7.84867287e-01
5.13950586e-01 1.89677428e-03 4.70194846e-01 4.50309277e-01
-1.00283909e+00 1.74155787e-01 6.23026669e-01 -2.02751458e-01
7.85009503e-01 -1.10622132e+00 -4.07175012e-02 1.40665805e+00
-1.51170969e-01 6.66990817e-01 1.01737905e+00 -5.32339513e-01
-1.39348769e+00 -8.37641001e-01 3.43492031e-01 -8.21672566e-03
6.20931089e-01 5.24904393e-02 -1.07123411e+00 2.93360800e-01
1.42044306e-01 -4.96408120e-02 1.79134339e-01 -4.11467701e-01
-3.27294499e-01 -2.50862122e-01 -1.18124795e+00 5.20635664e-01
6.99821651e-01 -5.01411200e-01 -5.53935409e-01 2.88922787e-01
5.27733862e-01 -7.69954175e-02 -7.14308202e-01 4.95007843e-01
1.98175281e-01 -1.08206630e+00 1.38986182e+00 -9.74635035e-03
3.83309335e-01 -7.58114100e-01 -3.47906977e-01 -1.31271982e+00
-4.09748077e-01 -7.47293055e-01 -5.02464592e-01 1.50414991e+00
-2.87604600e-01 -7.04216480e-01 1.25074387e-01 2.14038208e-01
-3.28931332e-01 -2.35350847e-01 -8.85181904e-01 -6.85101271e-01
-2.96057183e-02 -3.64675462e-01 2.34564409e-01 6.02496803e-01
-5.53916574e-01 1.35552406e-01 -7.75554419e-01 5.71818531e-01
1.30090129e+00 -8.07218552e-02 4.51526523e-01 -1.23974645e+00
-4.15189683e-01 -4.67887819e-01 -3.87454480e-02 -1.27411509e+00
1.82334810e-01 -4.80479479e-01 3.84390175e-01 -1.85861981e+00
3.24719638e-01 -1.34868324e-01 -3.43913078e-01 5.61305089e-03
-4.95660484e-01 2.23015159e-01 -8.18069503e-02 4.96605963e-01
-2.39430040e-01 8.86934459e-01 1.29742730e+00 -2.86029816e-01
-1.19058304e-01 2.13128686e-01 -7.72379756e-01 6.57816827e-01
6.49470568e-01 -1.17209934e-01 -3.79172653e-01 -5.91072857e-01
-9.08527970e-02 7.03041479e-02 7.42522061e-01 -8.55953574e-01
2.81563818e-01 -2.99543384e-02 3.01163524e-01 -4.77927357e-01
5.40944636e-01 -7.52343535e-01 1.47318989e-01 1.89051032e-01
3.80252637e-02 -6.12755299e-01 -9.13075432e-02 9.79021013e-01
-4.89209622e-01 -2.68488526e-01 1.13099277e+00 -8.54654312e-02
-6.59150779e-01 3.60101938e-01 -3.32521021e-01 -3.04772854e-01
3.99689078e-01 -5.92391938e-02 -2.38763198e-01 -4.42521960e-01
-8.51301312e-01 -5.17891459e-02 3.21276188e-01 2.27100551e-01
6.95885718e-01 -1.19936538e+00 -8.59778762e-01 5.12065254e-02
1.12257944e-02 -8.29701051e-02 7.74880707e-01 1.13242614e+00
-9.78943557e-02 7.84966946e-02 1.40638143e-01 -6.30599558e-01
-8.18144143e-01 8.98557842e-01 4.01749372e-01 -1.88718110e-01
-6.85432911e-01 6.57941818e-01 4.56881136e-01 -1.15216263e-01
3.35306734e-01 3.84433828e-02 -5.13590217e-01 1.17621208e-02
7.68201649e-01 4.51319993e-01 -1.05370693e-01 -8.63546431e-01
-7.44005144e-02 7.28578150e-01 1.48376584e-01 -1.34408355e-01
1.24334252e+00 -7.63418376e-01 -4.76616442e-01 3.08027208e-01
9.73363936e-01 -1.93491817e-01 -1.57401109e+00 -6.07070684e-01
-1.62021473e-01 -4.71029758e-01 5.74904561e-01 -7.11488724e-01
-1.17961109e+00 9.87305522e-01 8.22010100e-01 1.98700100e-01
1.51207614e+00 -2.34181181e-01 5.70337117e-01 -1.99648798e-01
2.36773431e-01 -8.49825144e-01 -6.18935935e-02 1.69104919e-01
1.01110375e+00 -1.27919745e+00 3.88527289e-02 -3.01931888e-01
-3.91571969e-01 1.00998926e+00 2.48605415e-01 -1.16619281e-01
5.60630441e-01 -2.51223564e-01 5.41094206e-02 1.12992615e-01
-3.70226353e-02 -3.96901906e-01 4.87804413e-01 7.60268569e-01
9.03030038e-02 -1.12112306e-01 -3.90123338e-01 6.33867323e-01
3.58909845e-01 1.05319113e-01 3.50374967e-01 5.36880493e-01
-5.24618447e-01 -9.27493095e-01 -8.76219749e-01 3.57170263e-03
-4.15704131e-01 -1.80541143e-01 2.71986187e-01 1.86943933e-01
1.11583568e-01 1.47310519e+00 -2.81938940e-01 -9.77722183e-02
1.26704946e-01 -1.38624012e-01 3.37369591e-01 -2.04955742e-01
2.00068615e-02 5.14092207e-01 -3.98225963e-01 -3.28822702e-01
-8.37877095e-01 -5.31813741e-01 -7.47281075e-01 -1.78682227e-02
-2.29544520e-01 -1.33499429e-01 7.34014094e-01 1.03124511e+00
8.58734846e-02 6.76890552e-01 7.41782188e-01 -1.24024582e+00
-8.17000866e-01 -1.12875473e+00 -8.78449619e-01 2.98066884e-01
6.54056609e-01 -8.32124591e-01 -8.23160887e-01 2.36032143e-01] | [11.521479606628418, -2.4591922760009766] |
166fb127-e370-4676-80d2-2f5c03bd3ba0 | capsule-networks-for-brain-tumor | 1811.00597 | null | http://arxiv.org/abs/1811.00597v1 | http://arxiv.org/pdf/1811.00597v1.pdf | Capsule Networks for Brain Tumor Classification based on MRI Images and Course Tumor Boundaries | According to official statistics, cancer is considered as the second leading
cause of human fatalities. Among different types of cancer, brain tumor is seen
as one of the deadliest forms due to its aggressive nature, heterogeneous
characteristics, and low relative survival rate. Determining the type of brain
tumor has significant impact on the treatment choice and patient's survival.
Human-centered diagnosis is typically error-prone and unreliable resulting in a
recent surge of interest to automatize this process using convolutional neural
networks (CNNs). CNNs, however, fail to fully utilize spatial relations, which
is particularly harmful for tumor classification, as the relation between the
tumor and its surrounding tissue is a critical indicator of the tumor's type.
In our recent work, we have incorporated newly developed CapsNets to overcome
this shortcoming. CapsNets are, however, highly sensitive to the miscellaneous
image background. The paper addresses this gap. The main contribution is to
equip CapsNet with access to the tumor surrounding tissues, without distracting
it from the main target. A modified CapsNet architecture is, therefore,
proposed for brain tumor classification, which takes the tumor coarse
boundaries as extra inputs within its pipeline to increase the CapsNet's focus.
The proposed approach noticeably outperforms its counterparts. | ['Parnian Afshar', 'Konstantinos N. Plataniotis', 'Arash Mohammadi'] | 2018-11-01 | null | null | null | null | ['miscellaneous'] | ['miscellaneous'] | [-6.69151098e-02 9.29299593e-02 -1.77872583e-01 -5.43603525e-02
-3.42335373e-01 -2.83105731e-01 6.56862855e-01 2.57819921e-01
-6.04888916e-01 6.54647171e-01 7.20032379e-02 -3.26166868e-01
8.92806277e-02 -7.28232086e-01 -2.39231095e-01 -9.44573045e-01
2.93256715e-02 3.46552402e-01 4.29806441e-01 -1.39001131e-01
6.20149411e-02 7.83570647e-01 -1.14346659e+00 2.20135286e-01
8.77306938e-01 1.32528865e+00 4.10788625e-01 2.35908091e-01
-3.52919847e-01 6.72554255e-01 -5.14647722e-01 -4.86778244e-02
1.79961585e-02 -3.48319441e-01 -6.35354817e-01 -3.04346889e-01
1.07157446e-01 -6.73081279e-02 -3.04188877e-01 1.12632847e+00
3.07112932e-01 -3.19905072e-01 6.99681520e-01 -1.13823390e+00
2.38592084e-02 3.99620622e-01 -4.55753535e-01 3.40921611e-01
-1.95720300e-01 1.60390779e-01 6.82525218e-01 -7.53382683e-01
4.54670131e-01 6.03621721e-01 6.65648937e-01 6.08753741e-01
-7.46396184e-01 -6.80461586e-01 1.77851662e-01 3.99512321e-01
-1.43330252e+00 -2.68920660e-01 7.19012976e-01 -4.24599737e-01
5.51193416e-01 3.97951245e-01 9.70227063e-01 1.28875160e+00
5.41257083e-01 7.50274718e-01 1.20044219e+00 -1.83746606e-01
3.58416677e-01 1.18033826e-01 5.43324463e-02 5.14614344e-01
2.92483687e-01 -9.86408815e-02 -1.61991939e-01 1.36395529e-01
5.86065054e-01 3.15770656e-01 -5.52639723e-01 -3.51130098e-01
-1.14609468e+00 5.40357769e-01 9.53964055e-01 7.06461668e-01
-3.68404686e-01 8.38773027e-02 4.39249396e-01 -2.92667955e-01
3.98515105e-01 1.24512911e-01 -2.01068595e-01 -7.64879212e-02
-1.04454374e+00 4.77842055e-02 6.00484550e-01 5.12813210e-01
1.42616322e-02 -1.63997591e-01 -1.51887298e-01 7.13977635e-01
1.66617855e-01 1.34016633e-01 8.45742166e-01 1.07384302e-01
7.03261569e-02 8.82022500e-01 -2.37061501e-01 -9.73455667e-01
-8.42273414e-01 -1.23250282e+00 -1.33405876e+00 1.72408730e-01
6.27070546e-01 1.28816068e-01 -1.17903268e+00 1.53189635e+00
1.69326469e-01 1.34610727e-01 -1.40986845e-01 1.02072752e+00
1.04021883e+00 2.62822062e-01 1.94711417e-01 9.35060978e-02
1.47773457e+00 -8.88685822e-01 -5.23572981e-01 -2.21722856e-01
7.02265739e-01 -5.18280149e-01 5.99673510e-01 2.21854657e-01
-7.29334116e-01 -1.06900312e-01 -8.92487526e-01 2.01319799e-01
-6.71752095e-01 2.10420132e-01 6.46338224e-01 6.05460703e-01
-1.12058973e+00 1.41861677e-01 -1.02449942e+00 -6.81983948e-01
7.56151974e-01 3.70094508e-01 -5.53882837e-01 -9.53290835e-02
-1.06307280e+00 1.23368442e+00 4.59317267e-01 2.48521626e-01
-7.95606315e-01 -8.02622616e-01 -4.98267561e-01 1.50593773e-01
5.37234068e-01 -4.35625494e-01 9.00447011e-01 -1.06234598e+00
-1.12432325e+00 5.69901526e-01 -7.22998288e-03 -6.49570465e-01
7.66370893e-01 3.67499053e-01 -3.83785307e-01 7.06225932e-02
-1.14981644e-01 6.57564938e-01 6.13769770e-01 -1.00692761e+00
-8.03883791e-01 -3.59005511e-01 -1.59135118e-01 -2.71297581e-02
-4.38363284e-01 -1.94583505e-01 -5.41707456e-01 -6.83139324e-01
1.35354072e-01 -8.68957281e-01 -3.98189038e-01 1.78329661e-01
-4.24578249e-01 -1.03313223e-01 1.06895196e+00 -5.20988464e-01
1.01523292e+00 -2.25055528e+00 -2.82296687e-02 1.45152852e-01
6.13780677e-01 4.65191007e-01 1.71952367e-01 4.56987023e-02
-7.65499547e-02 8.45884159e-02 -4.22028661e-01 -1.87415913e-01
-3.40619147e-01 6.00399151e-02 1.55927718e-01 6.63162589e-01
3.01048756e-01 8.74947727e-01 -7.78361738e-01 -5.55486202e-01
2.70180434e-01 7.48309076e-01 -1.82559147e-01 2.45847553e-02
-1.59233883e-02 5.24552763e-01 -3.69380087e-01 8.29902709e-01
6.52471542e-01 -1.43324018e-01 -1.98513821e-01 -2.41654113e-01
-1.14336617e-01 -8.88513923e-02 -6.48221970e-01 1.29791915e+00
-3.04554552e-01 7.87290037e-01 3.19846839e-01 -9.60932910e-01
8.26953232e-01 4.29733127e-01 8.29066813e-01 -6.79025352e-01
3.50955784e-01 3.85571182e-01 5.72708011e-01 -3.92945647e-01
1.33116156e-01 -7.46574551e-02 2.91746020e-01 -1.39952853e-01
-3.52468938e-01 2.01827049e-01 -2.19259150e-02 1.79347619e-01
1.33837008e+00 -2.17733890e-01 5.65305114e-01 -3.73322368e-01
6.33343041e-01 2.35631213e-01 7.82293677e-01 2.63340682e-01
-5.76186955e-01 7.80524492e-01 7.65023947e-01 -5.97563505e-01
-5.40908396e-01 -9.32747304e-01 -4.30149078e-01 4.07946825e-01
6.78541139e-02 -6.47118140e-04 -7.12138772e-01 -8.64002049e-01
-1.84103191e-01 3.51861298e-01 -8.38340044e-01 -2.29399264e-01
-4.93401557e-01 -8.68466854e-01 6.14044845e-01 5.64748824e-01
8.15911949e-01 -8.65459979e-01 -8.52663636e-01 2.05666572e-01
-1.06763199e-01 -1.08877695e+00 -2.81737950e-02 4.88606989e-01
-5.72035253e-01 -1.31646550e+00 -1.17973471e+00 -6.49391353e-01
9.44196522e-01 1.61866531e-01 8.14670622e-01 2.13441014e-01
-5.22796333e-01 -1.31578788e-01 -4.21773940e-01 -5.41327953e-01
-3.50566417e-01 4.83581185e-01 -2.41310060e-01 1.17216498e-01
3.59975010e-01 -3.46793234e-01 -7.73972452e-01 2.54132718e-01
-8.80673945e-01 3.02899927e-01 8.34798932e-01 9.76438403e-01
3.09246182e-01 1.72148719e-01 4.58789825e-01 -7.49818444e-01
4.24729496e-01 -5.69202244e-01 -4.66234446e-01 -5.15046008e-02
-1.82269573e-01 -3.01834583e-01 9.39740896e-01 -2.41479099e-01
-8.02347064e-01 2.95409441e-01 -2.71459281e-01 -1.75118953e-01
-4.57258105e-01 6.33879066e-01 -2.47428030e-01 -2.08955124e-01
4.05600846e-01 1.28612384e-01 2.51030207e-01 -1.82832226e-01
-3.11992705e-01 5.30489564e-01 5.82961500e-01 -7.78117627e-02
4.93910909e-01 5.51593125e-01 3.12384069e-01 -9.39160287e-01
-6.38156474e-01 -6.44590735e-01 -6.32739842e-01 -4.36731726e-01
9.59214151e-01 -5.82144916e-01 -4.81931269e-01 5.20448565e-01
-1.07340181e+00 -6.81749284e-02 2.03300267e-01 4.01273608e-01
-9.21616927e-02 4.73451279e-02 -3.78469914e-01 -3.93965453e-01
-2.60818243e-01 -1.49908447e+00 6.20481968e-01 2.29835391e-01
-7.67420679e-02 -9.70124841e-01 -2.46397749e-01 1.46396071e-01
9.04562235e-01 4.47956443e-01 9.18019295e-01 -8.88208270e-01
-5.36947906e-01 -4.36057776e-01 -5.44894934e-01 1.46718785e-01
1.45986483e-01 5.29913418e-02 -1.17646670e+00 -2.18032569e-01
-1.38213173e-01 1.99345604e-01 9.81722116e-01 4.19004977e-01
1.12611806e+00 1.53619468e-01 -7.85011113e-01 7.05809474e-01
1.37520468e+00 3.01871717e-01 5.95755577e-01 5.16842246e-01
6.47189975e-01 5.07996857e-01 3.70121211e-01 2.01906979e-01
1.50811672e-01 5.90266466e-01 1.12651229e+00 -3.85926604e-01
-2.85922974e-01 2.48403922e-01 1.63977265e-01 5.84530532e-01
-6.19396269e-02 -3.75704765e-01 -1.31214547e+00 6.16649330e-01
-1.53655696e+00 -7.16171443e-01 -2.19623879e-01 2.05426073e+00
4.10429269e-01 1.52037099e-01 -1.55782536e-01 1.80809051e-01
5.37682354e-01 -8.01251549e-03 -4.94040221e-01 4.59138788e-02
-2.43934076e-02 6.88580722e-02 7.44840741e-01 -2.67072804e-02
-1.20440412e+00 6.48709178e-01 5.23495531e+00 7.59522796e-01
-1.66436505e+00 7.80816302e-02 7.58550584e-01 -3.65295302e-04
2.12573603e-01 -3.47102195e-01 -6.11857235e-01 6.14184260e-01
5.59116960e-01 1.71806365e-01 -1.74983010e-01 7.52020121e-01
3.34304214e-01 -3.14963847e-01 -1.06202471e+00 9.22710240e-01
6.08672388e-02 -1.30297124e+00 -2.16975287e-01 2.72017956e-01
2.61254013e-01 8.05615112e-02 4.02997732e-02 1.64794266e-01
-7.63137937e-02 -1.36269367e+00 6.28380656e-01 3.24855298e-01
6.62697077e-01 -7.84758151e-01 1.30992365e+00 4.38465983e-01
-1.09931612e+00 -3.41862030e-02 -1.85696229e-01 2.00677276e-01
1.51155321e-02 6.54671788e-01 -1.20122623e+00 3.39131027e-01
6.58556640e-01 6.36940718e-01 -7.07152724e-01 1.46322048e+00
1.12498038e-01 6.71104968e-01 -3.27958882e-01 -1.01625942e-01
4.45380777e-01 1.85768545e-01 5.50706565e-01 1.22161090e+00
4.76209164e-01 -1.62854826e-03 2.07331866e-01 8.86542022e-01
1.20602615e-01 1.28716961e-01 -4.88560200e-01 7.02781826e-02
1.63969696e-01 1.47543967e+00 -1.25092614e+00 -1.20886259e-01
-4.72350746e-01 7.23440230e-01 2.78172255e-01 1.17428675e-01
-7.81059325e-01 -2.22023979e-01 6.18742049e-01 2.45948628e-01
5.25386892e-02 9.69032571e-03 -4.73206520e-01 -9.64299560e-01
-4.59026359e-02 -5.87241411e-01 2.27825209e-01 -3.49956810e-01
-9.88969982e-01 8.49775612e-01 -2.33500227e-01 -1.24913466e+00
1.90641940e-01 -8.11338782e-01 -7.23527730e-01 8.84120941e-01
-1.60719728e+00 -1.21604860e+00 -6.83378100e-01 4.59503204e-01
5.78766704e-01 -9.91662592e-02 7.58060217e-01 2.84095556e-01
-8.41923833e-01 5.61996877e-01 -1.59255907e-01 2.27324903e-01
6.21666968e-01 -1.12895238e+00 -1.16664179e-01 6.93692088e-01
-4.74576235e-01 3.45068842e-01 6.30101621e-01 -4.70084965e-01
-1.08289826e+00 -1.20382798e+00 5.94259739e-01 6.90078139e-02
7.33349025e-01 -3.46259505e-01 -9.66639996e-01 2.54967183e-01
1.57757923e-01 4.47797716e-01 3.91692400e-01 -3.49948376e-01
-8.29555094e-02 -2.69415915e-01 -1.09802222e+00 8.06427777e-01
5.42548895e-01 -2.43556157e-01 -1.33967638e-01 1.95950046e-01
3.05975586e-01 -4.41804111e-01 -7.18413830e-01 5.14853060e-01
3.50058138e-01 -1.19052565e+00 6.18502438e-01 -1.92751214e-02
4.42566037e-01 -3.07800561e-01 2.15494514e-01 -1.53929734e+00
-3.27040553e-01 7.82962516e-02 3.13902199e-01 9.57400680e-01
4.56596315e-01 -8.36882889e-01 1.08846986e+00 4.98773962e-01
-4.19056445e-01 -1.14220357e+00 -9.55298424e-01 -6.79108143e-01
1.27465412e-01 -2.61893004e-01 5.41980743e-01 9.10423279e-01
-2.58244257e-02 -6.38807714e-02 8.86897221e-02 1.92423418e-01
2.89564580e-01 -3.88104022e-01 5.22467911e-01 -1.23805797e+00
1.58326030e-01 -9.21796024e-01 -8.85038972e-01 -5.58411717e-01
-1.16259955e-01 -9.02197957e-01 2.28507407e-02 -1.60327399e+00
2.04781085e-01 -5.70758045e-01 -5.15890539e-01 4.78759766e-01
1.10655949e-01 3.65484089e-01 1.99463636e-01 3.74944136e-02
-1.40054688e-01 2.84905106e-01 1.31430972e+00 -3.40730667e-01
2.95768175e-02 5.37346862e-02 -3.88945758e-01 7.76414335e-01
1.03594041e+00 -2.03854039e-01 -2.98390538e-01 -2.05839545e-01
-4.91570421e-02 -1.17164940e-01 4.70061868e-01 -1.27137566e+00
5.65418065e-01 -1.65993333e-01 5.18066823e-01 -6.96855068e-01
3.39832187e-01 -1.25398195e+00 1.54971525e-01 6.89518869e-01
-2.29344338e-01 -1.27591789e-02 2.40636662e-01 3.37973535e-01
-4.83193219e-01 -1.59009293e-01 9.55241859e-01 -2.41991252e-01
-5.59280455e-01 4.58716661e-01 -6.52867913e-01 -2.80651063e-01
1.43435109e+00 -4.49465394e-01 -2.84894228e-01 -3.45949898e-03
-5.54139853e-01 1.01787783e-01 4.53309804e-01 1.86479062e-01
5.61345637e-01 -9.50631440e-01 -5.28959394e-01 1.32867381e-01
2.14900449e-01 3.64798844e-01 2.04781666e-01 1.46264005e+00
-7.65160084e-01 6.37282431e-01 -2.28247136e-01 -6.39376700e-01
-1.33161855e+00 3.96869153e-01 6.49473846e-01 -3.58180374e-01
-8.57225537e-01 7.05164075e-01 3.88450831e-01 4.33063470e-02
5.99600732e-01 -6.60843909e-01 -5.43998957e-01 -5.86308539e-02
5.61304986e-01 1.38021186e-01 3.53780061e-01 -6.95101857e-01
-4.07946050e-01 1.16141275e-01 -5.29216826e-01 2.73507625e-01
1.16907060e+00 2.92847157e-01 -3.83163095e-01 1.52510434e-01
1.07697201e+00 -1.57939062e-01 -9.59491193e-01 -8.88031721e-02
5.48431538e-02 -2.86907881e-01 4.63503599e-01 -9.98231053e-01
-1.56265652e+00 8.52917135e-01 6.23144984e-01 1.72728300e-01
1.11471236e+00 -1.51864558e-01 7.47904837e-01 -1.58614680e-01
3.01801562e-01 -7.69719362e-01 -3.55174333e-01 4.55549031e-01
7.29851961e-01 -1.29948151e+00 -1.76429212e-01 -5.51057577e-01
-2.93521285e-01 1.28712249e+00 7.49377787e-01 9.41314846e-02
8.11190605e-01 4.62775260e-01 2.42924660e-01 -2.18642682e-01
-6.10412121e-01 -1.71639919e-01 1.08872972e-01 5.37730277e-01
5.18910587e-01 2.29719400e-01 -2.77207106e-01 5.77848136e-01
-1.49175987e-01 -2.34079827e-02 3.08772594e-01 9.25530553e-01
-4.21514601e-01 -9.27779794e-01 -5.13885736e-01 6.40086353e-01
-5.71062267e-01 -1.44563049e-01 -5.44072986e-01 1.01517856e+00
2.83197850e-01 5.32500565e-01 1.23655662e-01 -6.24425113e-02
9.10439119e-02 -1.74738705e-01 9.27201957e-02 -3.31634372e-01
-7.43019879e-01 -9.53850970e-02 -3.02416444e-01 -4.02675003e-01
-2.36929223e-01 -4.84275967e-01 -1.34217644e+00 -7.38874003e-02
-2.09760934e-01 -9.28200483e-02 9.08118784e-01 1.02739787e+00
1.81353971e-01 7.68144906e-01 2.79956222e-01 -6.66381478e-01
-3.35044831e-01 -9.02134895e-01 -4.17470783e-01 8.55768695e-02
3.53453398e-01 -7.55423129e-01 -1.24591336e-01 -2.70858467e-01] | [14.800202369689941, -2.5410823822021484] |
dc9e6c3b-44d0-4e5b-a241-975dce03c068 | audio-spectral-enhancement-leveraging | 2108.03703 | null | https://arxiv.org/abs/2108.03703v1 | https://arxiv.org/pdf/2108.03703v1.pdf | Audio Spectral Enhancement: Leveraging Autoencoders for Low Latency Reconstruction of Long, Lossy Audio Sequences | With active research in audio compression techniques yielding substantial breakthroughs, spectral reconstruction of low-quality audio waves remains a less indulged topic. In this paper, we propose a novel approach for reconstructing higher frequencies from considerably longer sequences of low-quality MP3 audio waves. Our technique involves inpainting audio spectrograms with residually stacked autoencoder blocks by manipulating individual amplitude and phase values in relation to perceptual differences. Our architecture presents several bottlenecks while preserving the spectral structure of the audio wave via skip-connections. We also compare several task metrics and demonstrate our visual guide to loss selection. Moreover, we show how to leverage differential quantization techniques to reduce the initial model size by more than half while simultaneously reducing inference time, which is crucial in real-world applications. | ['Harshavardhan Abichandani', 'Darshan Deshpande'] | 2021-08-08 | null | null | null | null | ['spectral-reconstruction'] | ['computer-vision'] | [ 4.85365748e-01 -1.17345728e-01 -2.45833565e-02 -2.95157917e-02
-1.21576333e+00 -5.19115984e-01 -6.38412759e-02 6.55289069e-02
-9.36913565e-02 6.88287556e-01 6.25272453e-01 -6.58790767e-02
-1.33701682e-01 -5.95487416e-01 -8.07237864e-01 -5.36032557e-01
-3.26356560e-01 -3.00280869e-01 1.26171425e-01 -1.63662121e-01
1.54171079e-01 3.28168154e-01 -1.79951334e+00 6.67433083e-01
4.62038368e-01 1.11348796e+00 8.44935849e-02 1.22993863e+00
3.36816818e-01 1.10940313e+00 -7.65131354e-01 -4.15417910e-01
3.30359578e-01 -6.37213767e-01 -6.96771443e-01 -8.49127099e-02
6.53789997e-01 -6.25105858e-01 -5.42747557e-01 8.15449119e-01
8.20598364e-01 3.27537626e-01 4.22951370e-01 -9.39175189e-01
-7.33327344e-02 7.43082762e-01 -4.06628639e-01 2.71632433e-01
3.82839918e-01 1.31314531e-01 1.38983023e+00 -5.46763003e-01
3.20864528e-01 1.00580847e+00 1.00802708e+00 1.66476250e-01
-1.59968698e+00 -6.63637757e-01 -4.23735201e-01 7.30530560e-01
-1.50912344e+00 -9.32174563e-01 1.11944914e+00 -2.19215930e-01
1.10035801e+00 4.03799891e-01 7.55186439e-01 7.65440643e-01
2.04587411e-02 4.90972579e-01 5.42790711e-01 -5.58403909e-01
1.20427087e-01 -3.63807142e-01 -4.56002057e-01 4.05594260e-01
-2.70923495e-01 -2.91698296e-02 -1.22398293e+00 -1.49464160e-01
8.36127400e-01 -3.68988037e-01 -5.54004908e-01 6.96546361e-02
-7.87377059e-01 5.65225005e-01 1.56673752e-02 7.87406713e-02
-3.68345737e-01 6.16215229e-01 3.76706213e-01 6.57119453e-01
4.63210464e-01 4.34057295e-01 -3.35005850e-01 -4.81096387e-01
-1.49495006e+00 5.27645290e-01 5.43731391e-01 6.76529348e-01
5.62005937e-01 6.45262182e-01 7.43317679e-02 9.56085801e-01
9.77633055e-04 1.49111211e-01 3.85939002e-01 -1.62013018e+00
4.45011020e-01 -2.62906104e-01 1.14091821e-01 -1.08727217e+00
-1.65424347e-01 -3.78731191e-01 -7.50094771e-01 6.45984933e-02
1.85937509e-01 -2.32891858e-01 -4.69541371e-01 1.69295287e+00
7.82462656e-02 4.22088712e-01 -1.95958376e-01 7.07845092e-01
3.06195647e-01 1.03500915e+00 -1.11230671e-01 -2.09536701e-01
1.14636660e+00 -6.27910852e-01 -8.56057703e-01 1.16418995e-01
1.66580856e-01 -1.02119088e+00 9.79920983e-01 8.58478904e-01
-1.68420446e+00 -6.13522887e-01 -1.31788993e+00 -5.00618041e-01
2.86103010e-01 -1.22458220e-01 2.30086267e-01 6.00841880e-01
-1.20033956e+00 1.21330523e+00 -7.71272004e-01 3.73444051e-01
3.56030107e-01 3.41905385e-01 -1.09590381e-01 3.09793562e-01
-1.17572677e+00 4.30400521e-01 2.25682110e-01 -4.19144303e-01
-7.24654675e-01 -1.21531260e+00 -7.01788962e-01 4.44127262e-01
1.37618721e-01 -4.56060708e-01 1.59230220e+00 -7.17746198e-01
-1.64449918e+00 3.15902412e-01 -1.40350506e-01 -9.06539917e-01
1.97206780e-01 -4.14958835e-01 -6.52171552e-01 8.01684737e-01
-3.18237484e-01 7.45997012e-01 1.42881978e+00 -8.57651234e-01
-7.91580021e-01 7.28399260e-03 -9.24808234e-02 1.67499498e-01
-5.12941718e-01 -1.97708160e-01 -1.80493474e-01 -1.25908566e+00
1.17432214e-01 -4.46914345e-01 1.18686020e-01 1.84802815e-01
6.78320229e-02 2.72454530e-01 6.52333260e-01 -1.14800644e+00
1.48956609e+00 -2.46829629e+00 2.63774306e-01 6.57403097e-03
2.12012276e-01 -1.57649722e-02 -1.33412644e-01 5.03099203e-01
-1.14337146e-01 6.87058568e-02 -3.09336483e-01 -4.87105250e-01
6.97257742e-02 -1.19762704e-01 -7.94492126e-01 2.47544408e-01
3.28918725e-01 3.77354711e-01 -5.90963304e-01 -5.39725840e-01
5.11801355e-02 8.66467893e-01 -1.14132977e+00 2.18869478e-01
-9.05622095e-02 2.14260221e-01 3.08244050e-01 4.56186086e-01
4.55136836e-01 2.15591818e-01 7.61472285e-02 -5.42957306e-01
4.65660729e-02 9.13921952e-01 -1.24132752e+00 1.98941755e+00
-7.31706321e-01 9.96911883e-01 5.20848870e-01 -6.43042564e-01
7.18227863e-01 6.07614100e-01 7.51329660e-01 -6.26139224e-01
-3.24864574e-02 1.23270541e-01 -2.05432013e-01 -3.49706501e-01
7.11947739e-01 -5.34914613e-01 3.20705682e-01 4.10336524e-01
3.05815250e-01 -5.19679189e-01 9.16542634e-02 -1.34000897e-01
1.13786697e+00 4.24305089e-02 9.55338776e-02 1.30022421e-01
1.47293717e-01 -4.14013475e-01 4.64222401e-01 4.30376768e-01
-1.69459835e-01 1.06165683e+00 3.36335689e-01 -1.49601907e-01
-1.23260581e+00 -1.29172599e+00 -1.12635039e-01 1.18884146e+00
-3.27877164e-01 -8.50828886e-01 -8.11533809e-01 6.66964650e-02
-1.25693783e-01 6.71624959e-01 -4.83837277e-02 -1.27329126e-01
-8.04449797e-01 -2.40907341e-01 1.04080451e+00 3.24727565e-01
2.56504476e-01 -8.64536524e-01 -9.23758745e-01 5.03647804e-01
-5.52549481e-01 -8.74637783e-01 -5.32166719e-01 3.97675365e-01
-1.04367983e+00 -7.11422384e-01 -6.74285591e-01 -6.21035516e-01
-2.68105090e-01 2.10183144e-01 1.18827319e+00 -1.65317833e-01
-2.90791452e-01 9.64029282e-02 -3.95620078e-01 -2.80976325e-01
-2.57800460e-01 4.05456722e-02 5.09288982e-02 -1.33979172e-01
-3.29676345e-02 -1.38961935e+00 -6.72353864e-01 -1.46379501e-01
-1.09289205e+00 -7.56686106e-02 2.99488425e-01 4.78511781e-01
7.66992152e-01 4.88571465e-01 4.85104591e-01 -2.06280455e-01
7.00852215e-01 -2.67565966e-01 -3.85195583e-01 -4.40619290e-01
-2.66719580e-01 -1.51912039e-02 8.18542778e-01 -4.66927022e-01
-7.80346453e-01 -2.45943695e-01 -5.56422949e-01 -6.61572278e-01
1.08339176e-01 1.00732356e-01 -2.60943975e-02 7.54323453e-02
6.14365220e-01 9.25363898e-02 -1.88962594e-01 -7.21754134e-01
3.19345772e-01 7.42185831e-01 8.25198412e-01 -5.55213213e-01
5.44369578e-01 4.97907698e-01 3.50676514e-02 -9.31960940e-01
-5.31738698e-01 -2.07925394e-01 -3.03382456e-01 -4.43636179e-02
3.79786134e-01 -1.09214115e+00 -6.11115158e-01 1.25290498e-01
-1.11008143e+00 -2.94976026e-01 -8.19353640e-01 5.55361688e-01
-8.71421456e-01 2.97691882e-01 -1.09233069e+00 -6.34502709e-01
-4.16684449e-01 -9.61044610e-01 1.04753506e+00 -2.90173590e-01
-4.82220978e-01 -4.48617548e-01 1.56595051e-01 8.26470777e-02
5.40396631e-01 -1.26346529e-01 9.66999590e-01 -6.49559638e-03
-4.73396242e-01 1.66646615e-01 1.28098622e-01 5.29110730e-01
6.85768435e-03 -2.37801477e-01 -1.24753547e+00 -1.47719905e-01
3.06576431e-01 -2.59314269e-01 8.93239498e-01 5.09788275e-01
1.50663042e+00 -5.22955358e-01 3.41433734e-01 8.68900537e-01
1.17735434e+00 1.83552265e-01 8.74775469e-01 -1.21971481e-02
4.40156877e-01 5.88958144e-01 1.83774263e-01 7.97867715e-01
-6.02640770e-02 6.74272478e-01 3.16250995e-02 1.48970857e-01
-7.38673806e-01 -4.50539500e-01 4.80569541e-01 1.31764615e+00
4.18257155e-02 -9.36995000e-02 -5.08447528e-01 6.47240043e-01
-1.22027552e+00 -1.32941210e+00 2.41033182e-01 2.03667641e+00
1.42048895e+00 1.45002142e-01 1.28229767e-01 9.31460261e-01
4.58263993e-01 4.99220937e-01 -3.47990304e-01 -4.74845648e-01
-1.07794695e-01 9.10856426e-01 4.02343154e-01 7.12826848e-01
-9.54840183e-01 4.75238979e-01 7.19106579e+00 1.12397838e+00
-1.06811404e+00 2.22342491e-01 4.63645160e-01 -7.68937051e-01
-4.40646023e-01 -9.50061753e-02 -3.89225036e-01 2.50054777e-01
1.33588481e+00 -8.51462334e-02 8.32919538e-01 4.40209657e-01
4.03269976e-01 1.75437823e-01 -9.94701624e-01 1.24225056e+00
-7.28823021e-02 -1.35521328e+00 1.05781883e-01 -1.74869239e-01
5.05215406e-01 -2.14763492e-01 2.95311809e-01 -3.05278338e-02
-1.21523872e-01 -1.04345465e+00 1.08120453e+00 2.75322974e-01
1.08738685e+00 -1.12285006e+00 2.27090895e-01 3.97289805e-02
-1.40983188e+00 -9.63223428e-02 -3.05340618e-01 -3.26305002e-01
2.80515283e-01 7.94780493e-01 -7.23174930e-01 2.21104354e-01
9.00678515e-01 5.91268182e-01 -1.49380192e-01 1.04141581e+00
-7.77702853e-02 1.07543576e+00 -4.52256382e-01 6.05113268e-01
-9.93897617e-02 2.40610517e-03 5.77290237e-01 1.29475451e+00
5.26050448e-01 3.56759503e-02 -3.47446382e-01 6.36312664e-01
-3.55456740e-01 -4.49974686e-02 -4.28614318e-01 5.14857769e-02
6.25326335e-01 7.59375393e-01 -1.48584664e-01 -4.90848236e-02
-2.12130114e-01 8.70932698e-01 -1.10785440e-01 3.25406551e-01
-7.45601177e-01 -7.28180885e-01 1.00489664e+00 3.48441780e-01
6.57394230e-01 -2.82493114e-01 -2.93738335e-01 -8.70167136e-01
1.41416252e-01 -1.10055387e+00 1.04582772e-01 -8.29475462e-01
-8.04631352e-01 3.70920390e-01 -1.98020980e-01 -1.40223491e+00
-5.54822326e-01 -1.09830782e-01 -1.30536214e-01 6.77050233e-01
-1.51818812e+00 -5.53024292e-01 1.84898943e-01 5.61800599e-01
6.90881431e-01 9.14194360e-02 8.45918477e-01 7.47347057e-01
-2.77492963e-03 5.71112990e-01 4.29004654e-02 -1.60491437e-01
7.28551865e-01 -9.46823001e-01 3.44837010e-01 7.17606008e-01
4.39109564e-01 4.25428391e-01 1.03370249e+00 -2.86478817e-01
-1.28158164e+00 -9.37688947e-01 7.66134381e-01 2.70125359e-01
6.39443934e-01 -3.08159351e-01 -9.46723402e-01 4.87236023e-01
4.38463748e-01 -2.99271017e-01 9.18495715e-01 -1.00952625e-01
-6.11046493e-01 -5.20045221e-01 -1.02647686e+00 5.91608047e-01
8.24868560e-01 -1.10648358e+00 -4.65800822e-01 4.11738306e-02
1.03191733e+00 -3.40039968e-01 -9.97712135e-01 2.35710531e-01
6.20328069e-01 -1.23793697e+00 1.26618266e+00 -2.04056665e-01
7.08430707e-01 -3.21271420e-01 -5.37614644e-01 -1.22000146e+00
-3.11914623e-01 -1.19064271e+00 -5.99951446e-01 1.16458285e+00
4.85940017e-02 5.35253845e-02 8.44385147e-01 -2.07544807e-02
-1.93756834e-01 -5.18747568e-01 -1.06982636e+00 -4.93540347e-01
-1.83576103e-02 -8.40830505e-01 5.82025707e-01 5.18154919e-01
1.00253358e-01 1.35864347e-01 -7.53290832e-01 1.53012350e-01
5.50602734e-01 -1.05414711e-01 4.11569685e-01 -8.42345595e-01
-7.61863053e-01 -5.14465988e-01 -2.92654991e-01 -1.18005908e+00
-7.60039091e-02 -5.52826643e-01 9.37620625e-02 -9.26898718e-01
-2.82281905e-01 6.20000698e-02 -3.62677097e-01 2.13413596e-01
2.97596425e-01 7.11043656e-01 3.79065931e-01 3.02386340e-02
-2.32134581e-01 6.91675246e-01 8.36654186e-01 -3.24333251e-01
-2.54810512e-01 -9.25806016e-02 -5.02382934e-01 7.48767734e-01
8.56105447e-01 -5.71932971e-01 -5.97541094e-01 -5.98207235e-01
4.00586009e-01 4.36565608e-01 4.08148706e-01 -1.56361890e+00
1.17238455e-01 1.98085368e-01 3.31032366e-01 -6.83401227e-01
7.58576274e-01 -7.59047866e-01 2.74948418e-01 1.33415833e-01
-6.39263570e-01 7.83173591e-02 4.81039792e-01 4.64430511e-01
-6.54386401e-01 -2.89581448e-01 8.81659806e-01 8.65386575e-02
-3.47590744e-01 7.77414627e-03 -5.17199039e-01 1.34609252e-01
3.02789181e-01 2.00043060e-02 2.81834602e-01 -8.60866010e-01
-7.82420933e-01 -3.82376045e-01 2.15909675e-01 -2.60572191e-02
7.51666427e-01 -1.23344600e+00 -7.46675014e-01 2.64859438e-01
-4.33334053e-01 -1.46655872e-01 4.41753566e-01 3.86012346e-01
-7.85303891e-01 2.42750525e-01 -2.60704368e-01 -2.63540328e-01
-1.26102984e+00 1.62030429e-01 2.50285745e-01 -6.95513785e-02
-7.76101351e-01 9.72049892e-01 -3.04480374e-01 3.74716580e-01
6.22526109e-01 -2.73757160e-01 1.02596402e-01 1.18233770e-01
8.17449510e-01 6.87575042e-01 3.71187866e-01 -4.62083936e-01
1.64553002e-02 4.73334789e-01 1.48897722e-01 -6.20845079e-01
1.47073627e+00 -3.43603492e-01 9.41229016e-02 5.44813097e-01
1.55737352e+00 4.01388526e-01 -1.49288130e+00 4.27475572e-03
-2.95786083e-01 -5.07931888e-01 3.62891853e-01 -4.20834869e-01
-1.18827271e+00 1.15146208e+00 5.71481526e-01 2.86887109e-01
1.72928536e+00 -3.20038170e-01 1.29766095e+00 2.57706225e-01
2.60337442e-01 -1.17681587e+00 1.50172383e-01 1.89352274e-01
8.43167424e-01 -4.73107994e-01 1.99105278e-01 -2.75017411e-01
-2.95377702e-01 1.07107234e+00 -1.97552741e-01 -2.95090407e-01
6.55676961e-01 6.33513212e-01 -8.25272575e-02 2.57617056e-01
-8.35984707e-01 4.10102420e-02 1.64020270e-01 6.67005777e-01
5.77878416e-01 -1.13535061e-01 -7.07734004e-03 5.36491513e-01
-7.90839612e-01 -1.19894505e-01 3.54404569e-01 6.90313756e-01
-6.52821600e-01 -1.14370573e+00 -5.77733457e-01 3.44796687e-01
-8.02945018e-01 -4.18230087e-01 -4.99824658e-02 2.53167570e-01
2.07328558e-01 1.10391104e+00 2.50414193e-01 -5.96842647e-01
2.56452441e-01 2.26728141e-01 6.27571821e-01 -1.80050343e-01
-6.39218032e-01 5.65239072e-01 -6.47080243e-02 -6.89278781e-01
-3.35861564e-01 -5.81418216e-01 -1.15970576e+00 -3.99020404e-01
1.50381718e-02 -3.25461663e-02 4.94876742e-01 4.62799221e-01
5.08469284e-01 8.03379297e-01 5.73264718e-01 -1.08857715e+00
-4.85502988e-01 -8.31109583e-01 -7.56165743e-01 3.42117190e-01
8.46894383e-01 -1.86895654e-01 -3.41283530e-01 6.01268709e-01] | [15.534723281860352, 5.804286003112793] |
dcb0c426-bd85-40d6-b68c-85dfd89b1f4b | investigating-and-exploiting-image-resolution | 2006.14715 | null | https://arxiv.org/abs/2006.14715v1 | https://arxiv.org/pdf/2006.14715v1.pdf | Investigating and Exploiting Image Resolution for Transfer Learning-based Skin Lesion Classification | Skin cancer is among the most common cancer types. Dermoscopic image analysis improves the diagnostic accuracy for detection of malignant melanoma and other pigmented skin lesions when compared to unaided visual inspection. Hence, computer-based methods to support medical experts in the diagnostic procedure are of great interest. Fine-tuning pre-trained convolutional neural networks (CNNs) has been shown to work well for skin lesion classification. Pre-trained CNNs are usually trained with natural images of a fixed image size which is typically significantly smaller than captured skin lesion images and consequently dermoscopic images are downsampled for fine-tuning. However, useful medical information may be lost during this transformation. In this paper, we explore the effect of input image size on skin lesion classification performance of fine-tuned CNNs. For this, we resize dermoscopic images to different resolutions, ranging from 64x64 to 768x768 pixels and investigate the resulting classification performance of three well-established CNNs, namely DenseNet-121, ResNet-18, and ResNet-50. Our results show that using very small images (of size 64x64 pixels) degrades the classification performance, while images of size 128x128 pixels and above support good performance with larger image sizes leading to slightly improved classification. We further propose a novel fusion approach based on a three-level ensemble strategy that exploits multiple fine-tuned networks trained with dermoscopic images at various sizes. When applied on the ISIC 2017 skin lesion classification challenge, our fusion approach yields an area under the receiver operating characteristic curve of 89.2% and 96.6% for melanoma classification and seborrheic keratosis classification, respectively, outperforming state-of-the-art algorithms. | ['Georg Dorffner', 'Isabella Ellinger', 'Rupert Ecker', 'Gerald Schaefer', 'Chunliang Wang', 'Amirreza Mahbod'] | 2020-06-25 | null | null | null | null | ['skin-lesion-classification'] | ['medical'] | [ 6.21388912e-01 6.67896634e-03 -1.94831118e-01 -4.94139455e-03
-4.27292496e-01 -3.53258729e-01 4.35159355e-01 1.52358413e-01
-7.61621058e-01 6.58882201e-01 -1.26943126e-01 -4.07932192e-01
-1.37231067e-01 -8.34817648e-01 -4.18122649e-01 -7.99337626e-01
3.52920443e-01 -2.80398399e-01 3.30768973e-01 -8.62599760e-02
1.42930880e-01 5.70139229e-01 -1.48654437e+00 4.53457803e-01
1.25047457e+00 1.04035175e+00 -8.29485729e-02 1.05892003e+00
1.04577951e-01 5.48084199e-01 -7.29753017e-01 -4.94216532e-01
4.04349238e-01 -7.08228629e-03 -8.14500451e-01 2.92027920e-01
9.00238454e-01 -2.16792375e-01 -1.98748335e-01 1.21670783e+00
4.92517233e-01 -1.74099624e-01 6.42219722e-01 -5.02928913e-01
-6.48674548e-01 1.64039969e-01 -6.96503639e-01 3.03705603e-01
-9.97213796e-02 2.33792007e-01 3.69316190e-01 -2.18192130e-01
6.76292360e-01 7.69894838e-01 9.32987809e-01 7.74730444e-01
-1.08489025e+00 -4.61372465e-01 -2.08376855e-01 1.42496675e-01
-1.34591460e+00 -9.13659558e-02 7.84597099e-02 -3.69747341e-01
8.73195708e-01 4.83585626e-01 6.70717955e-01 1.01352191e+00
4.66165304e-01 1.13184877e-01 1.58395207e+00 -5.31301796e-01
4.80923289e-03 3.78464401e-01 -1.87648088e-01 7.64923275e-01
4.43520546e-01 1.08132407e-01 -2.71333531e-02 1.65462513e-02
1.02340245e+00 8.98963809e-02 -3.56118292e-01 2.90347278e-01
-8.49126697e-01 7.95718312e-01 8.43741834e-01 3.76044482e-01
-4.95475858e-01 -5.78749599e-03 4.21008855e-01 3.30718368e-01
5.12510478e-01 5.98967731e-01 -1.02735922e-01 2.64025956e-01
-8.30072880e-01 -1.86819836e-01 5.14974236e-01 -2.36752443e-02
4.16817933e-01 -1.27086267e-01 -3.32448125e-01 1.24124324e+00
-2.29542896e-01 1.57745168e-01 6.80081546e-01 -4.02138203e-01
1.10490032e-01 9.05898273e-01 -1.95067525e-01 -6.43525302e-01
-5.24751842e-01 -4.42413718e-01 -1.35202968e+00 5.04204035e-01
6.13264084e-01 -1.94425166e-01 -1.59887409e+00 1.22698891e+00
3.50708127e-01 1.28714755e-01 2.02790320e-01 1.00263393e+00
8.07819664e-01 1.60079688e-01 4.73789662e-01 1.03898071e-01
1.50052166e+00 -1.00195861e+00 -2.31559858e-01 -4.18566056e-02
5.15792072e-01 -7.72492349e-01 8.89589310e-01 4.10706669e-01
-7.99341857e-01 -6.76921368e-01 -1.10974431e+00 1.48469403e-01
-4.15659577e-01 4.87676203e-01 5.33748567e-01 9.79395747e-01
-1.14966404e+00 7.52507091e-01 -6.68823421e-01 -7.35371292e-01
6.42376721e-01 5.41380584e-01 -6.83760464e-01 -2.63242662e-01
-9.88898873e-01 1.03020060e+00 3.40951204e-01 6.34554923e-02
-5.76837599e-01 -8.83803606e-01 -5.78759909e-01 -1.31761700e-01
9.93117318e-02 -7.79382288e-01 8.64587963e-01 -1.26929200e+00
-1.37229216e+00 1.07950711e+00 1.98930427e-01 -7.36100018e-01
5.77661216e-01 8.85945484e-02 -7.63466835e-01 4.59774494e-01
-3.96202058e-01 6.13140047e-01 1.08920693e+00 -8.71229649e-01
-8.20834696e-01 -3.70947689e-01 3.10459942e-01 1.81601465e-01
-5.59038103e-01 -9.34632868e-02 -3.71850640e-01 -4.66372609e-01
-4.19730932e-01 -1.07671535e+00 -6.06623530e-01 8.74507055e-02
-6.62372291e-01 1.60114467e-01 5.45757532e-01 -5.77363610e-01
1.16057718e+00 -1.93937874e+00 -2.17652917e-01 4.58627820e-01
5.56509256e-01 9.87713695e-01 -3.80645394e-01 5.34503311e-02
-2.41149858e-01 3.48808497e-01 -1.20221991e-02 9.26211774e-02
-6.21498048e-01 -2.10307598e-01 4.20819342e-01 4.88067001e-01
4.05947596e-01 8.40369046e-01 -4.95318919e-01 -3.21786433e-01
6.90398991e-01 8.08838129e-01 8.43717065e-03 -1.52225226e-01
1.56787768e-01 8.46883878e-02 -2.34116167e-01 7.18353331e-01
7.95301795e-01 -6.47438586e-01 1.28632516e-01 -5.61423004e-01
1.43387318e-01 -2.64434099e-01 -7.73640513e-01 1.23320723e+00
-7.26360857e-01 7.25979924e-01 6.07442968e-02 -5.06634355e-01
5.93058705e-01 2.41926253e-01 2.21637905e-01 -6.71868086e-01
1.56230778e-01 7.28260428e-02 3.01255792e-01 -5.94105721e-01
3.51184547e-01 -2.46411100e-01 3.96430850e-01 9.63640884e-02
-1.49583906e-01 7.47572035e-02 3.35403621e-01 -9.03510377e-02
1.23412240e+00 -4.84088153e-01 6.33940756e-01 -8.89592022e-02
6.27999783e-01 6.18694201e-02 -1.35145159e-02 6.05034292e-01
-2.94803083e-01 5.61773717e-01 2.23202109e-01 -5.07668614e-01
-1.00687265e+00 -8.85471225e-01 -5.14524937e-01 7.08743811e-01
3.42128128e-02 -1.53583035e-01 -1.06846654e+00 -8.13260138e-01
8.13683402e-03 3.28292809e-02 -1.18441820e+00 -9.73355100e-02
-2.04094455e-01 -1.21201277e+00 7.46920884e-01 4.80152965e-01
8.22978318e-01 -9.51063156e-01 -5.80043316e-01 -4.94247079e-02
3.18293244e-01 -8.96289170e-01 -1.48520157e-01 3.04779671e-02
-6.05261028e-01 -1.53912735e+00 -1.14394665e+00 -5.80218971e-01
1.02637231e+00 2.52331614e-01 7.51119673e-01 2.84765542e-01
-9.97900248e-01 7.59828463e-02 -3.45397830e-01 -3.55873346e-01
-6.84144318e-01 3.27055752e-01 -2.36138493e-01 1.05356976e-01
2.34250009e-01 -7.26839602e-02 -9.02179778e-01 1.05588444e-01
-1.22133577e+00 4.61836345e-02 1.20363474e+00 1.03507769e+00
4.80471373e-01 2.02825859e-01 3.67426544e-01 -1.32369709e+00
8.97329390e-01 -1.39772356e-01 -2.94340134e-01 4.91457194e-01
-5.31770408e-01 -1.47100314e-01 7.22353458e-01 -6.22486174e-01
-1.00696075e+00 1.78762001e-03 -1.51394755e-01 -4.68473762e-01
-4.11308587e-01 4.15123731e-01 6.12596452e-01 -7.90611088e-01
1.25816536e+00 -6.40323386e-02 3.05224180e-01 -3.44866514e-02
-2.05883216e-02 8.29072118e-01 4.26321357e-01 2.20316544e-01
6.47585034e-01 4.70378757e-01 1.64271295e-01 -9.59290445e-01
-6.87354267e-01 -5.10566175e-01 -2.96214283e-01 -1.69395611e-01
8.15789044e-01 -8.56773615e-01 -6.07249558e-01 9.37965333e-01
-5.60584486e-01 -3.88148785e-01 -1.84357956e-01 2.24430919e-01
1.23490959e-01 3.99485022e-01 -7.92722881e-01 -4.36987638e-01
-7.02753663e-01 -9.84047890e-01 9.81261134e-01 7.39554644e-01
-1.53881967e-01 -1.39933383e+00 -9.45925787e-02 4.63330895e-01
7.46099710e-01 5.66263556e-01 7.72243083e-01 -3.29628229e-01
-7.43009970e-02 -3.89710784e-01 -6.50384724e-01 5.78136981e-01
3.89410406e-01 2.18090057e-01 -1.13310289e+00 -4.91837919e-01
-5.24841547e-01 -2.87587166e-01 1.31555855e+00 5.97418070e-01
1.25605583e+00 -1.64131805e-01 -3.62981915e-01 6.92555666e-01
1.75740635e+00 -2.50697080e-02 8.60640943e-01 3.51245999e-01
6.50848627e-01 5.93326390e-01 3.35350901e-01 2.02434361e-01
-3.55214290e-02 1.64327577e-01 5.92913449e-01 -7.53668249e-01
-5.64505100e-01 1.62867561e-01 -2.13040367e-01 -3.22845541e-02
-5.46792209e-01 -1.91377953e-01 -6.58242285e-01 5.13233244e-01
-1.19223166e+00 -5.72967947e-01 -8.06404203e-02 2.24860191e+00
7.49747813e-01 -3.37923467e-02 -7.86461774e-03 1.84413448e-01
9.54642177e-01 8.02012458e-02 -7.59639323e-01 -4.58144605e-01
-1.12422787e-01 7.85515368e-01 1.01064801e+00 2.88297117e-01
-1.27881718e+00 7.90593624e-01 5.66441679e+00 1.12451088e+00
-1.80732143e+00 -8.98113176e-02 8.84719849e-01 -5.61186485e-02
1.01925567e-01 -5.65401495e-01 -5.58391154e-01 2.70873278e-01
9.51006830e-01 1.65672451e-01 3.01228493e-01 5.10058105e-01
-4.17069942e-02 -4.89571899e-01 -5.08246124e-01 9.31243896e-01
1.59461908e-02 -1.59934151e+00 1.42137423e-01 2.49503389e-01
1.10479140e+00 -3.44272368e-02 5.27250648e-01 -1.16376758e-01
2.24164054e-02 -1.62610376e+00 -1.38183460e-01 4.49242264e-01
1.40440214e+00 -6.36657834e-01 1.00079060e+00 -7.29545578e-02
-8.40567946e-01 -1.07472099e-01 -4.93035972e-01 4.13205296e-01
-3.13222319e-01 7.34003723e-01 -1.20656335e+00 3.83387595e-01
6.94528222e-01 5.16506791e-01 -1.06159234e+00 1.11616182e+00
9.03751627e-02 5.26867330e-01 -1.62333384e-01 -1.42850548e-01
4.10142392e-01 2.33741954e-01 1.02982529e-01 1.19555187e+00
2.47953132e-01 -2.56474704e-01 -4.07147765e-01 5.41720688e-01
-8.83311555e-02 1.96127862e-01 -2.83830583e-01 -1.57647189e-02
1.29917532e-01 1.64177489e+00 -6.42814696e-01 -2.21202806e-01
-2.12539241e-01 9.00776207e-01 1.15643248e-01 3.55952531e-01
-5.61971486e-01 -5.36989868e-01 5.96828401e-01 3.03881556e-01
2.18231708e-01 5.00129342e-01 -1.64377272e-01 -8.97331059e-01
-2.43956640e-01 -9.65224862e-01 3.79773378e-01 -4.44010466e-01
-1.27743065e+00 8.20878327e-01 -5.15758336e-01 -1.11361349e+00
-9.83749703e-02 -9.57996786e-01 -7.70101607e-01 1.04117918e+00
-1.69818747e+00 -1.37717378e+00 -9.38413262e-01 6.69067860e-01
2.33801395e-01 -7.92787373e-02 9.54072893e-01 -2.40827464e-02
-5.83186090e-01 7.95587122e-01 1.29202172e-01 8.68678614e-02
8.77163410e-01 -1.31348634e+00 2.75325954e-01 6.27508759e-01
-3.23530465e-01 5.15177011e-01 2.03773499e-01 -5.54480255e-01
-8.77634525e-01 -1.47394240e+00 2.72927523e-01 -5.51711991e-02
4.21792358e-01 2.20208228e-01 -7.61513174e-01 -4.92271446e-02
2.23072886e-01 2.73154855e-01 8.63042533e-01 -1.10344645e-02
-4.17034715e-01 -2.14859888e-01 -1.55399752e+00 5.87688565e-01
4.39094007e-01 -4.73993689e-01 1.29297987e-01 3.62553924e-01
6.67786524e-02 -5.78051090e-01 -1.29287541e+00 5.48713505e-01
6.81071818e-01 -1.17614865e+00 9.22743440e-01 -4.22257692e-01
5.89524686e-01 5.18693700e-02 2.78481781e-01 -1.50106215e+00
-2.00880736e-01 -2.03518897e-01 2.68167913e-01 6.15365326e-01
4.59098458e-01 -7.89934695e-01 1.10395455e+00 2.37297997e-01
2.00954109e-01 -1.13595998e+00 -6.39380515e-01 -3.15441608e-01
1.35858357e-01 8.87881964e-02 1.91089898e-01 6.92608416e-01
-3.94840539e-01 4.21354771e-02 -1.13455519e-01 1.09524585e-01
5.11068821e-01 -1.91357926e-01 4.90106195e-01 -1.07307386e+00
-3.43038775e-02 -4.99810278e-01 -6.30698264e-01 -1.95912436e-01
-1.85186982e-01 -6.22139335e-01 -4.04423952e-01 -1.44185388e+00
3.61047477e-01 -3.90850991e-01 -5.11615336e-01 5.48963010e-01
-4.22201872e-01 8.81111741e-01 1.15121007e-01 -4.99391276e-03
-2.16270268e-01 -3.17294598e-01 1.71967494e+00 -2.96088636e-01
-1.60615876e-01 1.06806658e-01 -8.09797704e-01 6.18571818e-01
9.51970339e-01 2.45216325e-01 -2.07787171e-01 -5.32551892e-02
-1.36296839e-01 -1.32699549e-01 6.02360308e-01 -1.30685866e+00
2.76051521e-01 -6.03617728e-02 8.09845388e-01 -9.21946540e-02
3.60004514e-01 -3.72629285e-01 1.89265445e-01 7.71827877e-01
-3.97348940e-01 -5.71816981e-01 4.68573332e-01 4.35832322e-01
-2.88387716e-01 -2.09379241e-01 1.22806799e+00 -2.02832982e-01
-8.05844247e-01 3.07706833e-01 -2.86610544e-01 -4.48171765e-01
1.27315784e+00 -6.04685307e-01 -7.51377523e-01 -8.13540891e-02
-6.80629551e-01 -3.18850845e-01 5.62396348e-01 3.08864862e-01
6.04501009e-01 -8.58762920e-01 -7.45710254e-01 1.49024531e-01
2.91150600e-01 -2.14793727e-01 8.99085641e-01 1.02260387e+00
-8.79335344e-01 4.57896650e-01 -5.83145499e-01 -5.01274467e-01
-1.73409629e+00 1.15139291e-01 6.74592495e-01 -5.02653003e-01
-2.09217727e-01 1.04617989e+00 7.36110434e-02 -1.40175357e-01
-7.13378936e-02 -5.37404180e-01 -3.79608065e-01 -2.11610511e-01
5.74621022e-01 3.78903717e-01 3.74344140e-01 -3.90562475e-01
3.64716328e-03 7.39270687e-01 -6.37717068e-01 4.59324807e-01
1.02083123e+00 1.63593888e-01 7.94011727e-03 -3.42062891e-01
1.03913629e+00 -7.80156925e-02 -1.13608897e+00 -3.64858657e-02
-5.53113401e-01 -5.16396284e-01 2.06854105e-01 -1.22907126e+00
-1.20747328e+00 8.16792309e-01 1.09312856e+00 3.34622443e-01
1.47101188e+00 -3.00734580e-01 6.31108642e-01 1.32045820e-01
2.37085745e-02 -9.52065945e-01 -3.22995037e-02 9.22737923e-03
5.30365705e-01 -1.40243006e+00 1.13215679e-02 -6.40045762e-01
-6.86086833e-01 1.19859815e+00 9.03340816e-01 -3.14902127e-01
2.95319885e-01 1.56054229e-01 4.11999404e-01 -7.95735270e-02
-5.70456147e-01 -3.55870008e-01 6.30358458e-01 7.35217512e-01
2.69058406e-01 2.25740239e-01 -8.40234831e-02 8.38482529e-02
-6.65333644e-02 5.29028475e-02 5.82979679e-01 5.53994775e-01
-3.08531255e-01 -8.88260484e-01 -3.40570897e-01 1.15630770e+00
-8.57418180e-01 -1.81790546e-01 -5.28433442e-01 1.02177811e+00
3.47363234e-01 8.54920506e-01 3.82929109e-02 -4.29124236e-01
7.96254277e-02 -4.52109516e-01 6.28406763e-01 -5.94294250e-01
-1.00792265e+00 -1.45837262e-01 1.48165390e-01 -4.88903522e-01
-5.79996765e-01 -1.46009386e-01 -6.46011055e-01 -3.81724924e-01
-4.41170454e-01 -2.27490321e-01 8.23574305e-01 7.20565736e-01
8.25170055e-02 8.18165600e-01 3.77202183e-01 -4.89908606e-01
-5.84698617e-01 -1.21200943e+00 -7.18972981e-01 2.98075676e-01
4.40483093e-01 -3.03772748e-01 -1.77947104e-01 1.36527028e-02] | [15.67648983001709, -2.9951796531677246] |
2a897b09-923c-45ca-9887-ae7bcdbdee5f | deliberated-domain-bridging-for-domain | 2209.07695 | null | https://arxiv.org/abs/2209.07695v3 | https://arxiv.org/pdf/2209.07695v3.pdf | Deliberated Domain Bridging for Domain Adaptive Semantic Segmentation | In unsupervised domain adaptation (UDA), directly adapting from the source to the target domain usually suffers significant discrepancies and leads to insufficient alignment. Thus, many UDA works attempt to vanish the domain gap gradually and softly via various intermediate spaces, dubbed domain bridging (DB). However, for dense prediction tasks such as domain adaptive semantic segmentation (DASS), existing solutions have mostly relied on rough style transfer and how to elegantly bridge domains is still under-explored. In this work, we resort to data mixing to establish a deliberated domain bridging (DDB) for DASS, through which the joint distributions of source and target domains are aligned and interacted with each in the intermediate space. At the heart of DDB lies a dual-path domain bridging step for generating two intermediate domains using the coarse-wise and the fine-wise data mixing techniques, alongside a cross-path knowledge distillation step for taking two complementary models trained on generated intermediate samples as 'teachers' to develop a superior 'student' in a multi-teacher distillation manner. These two optimization steps work in an alternating way and reinforce each other to give rise to DDB with strong adaptation power. Extensive experiments on adaptive segmentation tasks with different settings demonstrate that our DDB significantly outperforms state-of-the-art methods. Code is available at https://github.com/xiaoachen98/DDB.git. | ['Yi Jin', 'Kai Chen', 'Miao Zheng', 'Huaian Chen', 'Xin Jin', 'Zhixiang Wei', 'Lin Chen'] | 2022-09-16 | null | null | null | null | ['synthetic-to-real-translation'] | ['computer-vision'] | [ 3.18816125e-01 3.07638794e-01 -1.57726452e-01 -5.16652822e-01
-8.13855410e-01 -5.52600741e-01 6.14710093e-01 8.07151422e-02
-3.67864907e-01 8.01546335e-01 1.25856632e-02 -1.99079573e-01
-2.89526419e-03 -7.64677048e-01 -7.84870982e-01 -7.71300256e-01
7.18188763e-01 8.56509089e-01 5.19604921e-01 -2.92387456e-01
-1.19904384e-01 -1.82710290e-02 -1.11518228e+00 6.81162551e-02
1.51113379e+00 6.03944540e-01 3.36623818e-01 2.62960494e-01
-5.78083158e-01 3.55783999e-01 -4.00683761e-01 -6.39973402e-01
3.01257432e-01 -7.18142509e-01 -9.75765228e-01 2.82322526e-01
1.50266007e-01 -2.12281942e-01 4.40169983e-02 1.08453667e+00
4.95272428e-01 2.33073875e-01 5.33340156e-01 -1.10190678e+00
-7.47494102e-01 7.88614988e-01 -7.46570051e-01 -1.86116800e-01
5.78546524e-02 2.02872664e-01 8.62456560e-01 -8.29753637e-01
6.02986038e-01 1.16834080e+00 7.00014830e-01 8.27995420e-01
-1.54687762e+00 -7.41719723e-01 3.07457298e-01 2.20970213e-02
-1.03449929e+00 -2.10313439e-01 1.07694197e+00 -4.20113504e-01
4.00898784e-01 -1.93741247e-01 7.28954077e-01 1.31109858e+00
-5.50794363e-01 1.05838645e+00 1.18794000e+00 -4.92128044e-01
2.61264354e-01 3.64692956e-01 5.02680726e-02 4.16763693e-01
2.31783018e-02 -2.36032486e-01 -3.85963857e-01 1.46613106e-01
7.30655432e-01 -1.97716922e-01 -5.63573427e-02 -5.38647532e-01
-9.76255834e-01 7.73441195e-01 3.65899533e-01 3.13748330e-01
-4.11328048e-01 -4.93585914e-01 2.61826158e-01 2.89600641e-01
6.30244732e-01 4.67108190e-01 -6.04617059e-01 -9.15480107e-02
-9.77033854e-01 4.04427230e-01 6.68865442e-01 8.92952621e-01
8.55100751e-01 -1.39406785e-01 -1.37596637e-01 1.41409004e+00
3.60620350e-01 1.16812430e-01 5.99972010e-01 -8.93074572e-01
6.17471933e-01 7.57532418e-01 -5.03064506e-02 -4.61603165e-01
-3.15731913e-02 -4.72512573e-01 -6.72941923e-01 1.83513090e-01
8.92067373e-01 -2.80688524e-01 -1.27109778e+00 2.02618289e+00
9.30142105e-01 4.39707160e-01 1.59445360e-01 8.82762074e-01
5.94868541e-01 5.83355665e-01 3.43600243e-01 1.61020622e-01
1.19313133e+00 -1.21775711e+00 -3.36568564e-01 -4.38933879e-01
4.93421674e-01 -7.47381806e-01 1.36330605e+00 3.91396374e-01
-1.33020508e+00 -7.13821352e-01 -8.76655161e-01 -3.19275588e-01
-2.87246168e-01 -8.17729719e-03 2.06187665e-01 2.87386417e-01
-7.77016401e-01 5.55103958e-01 -8.50883365e-01 -2.54694462e-01
6.95964694e-01 9.36447084e-02 -1.64618656e-01 -1.48761436e-01
-1.20595419e+00 7.90040731e-01 6.32414699e-01 -1.55007929e-01
-8.21518600e-01 -1.09587884e+00 -7.48834014e-01 -2.14326084e-01
4.71297383e-01 -8.93674135e-01 1.43916833e+00 -1.10196733e+00
-1.98195028e+00 1.08144593e+00 8.86835083e-02 -5.00575602e-01
9.08823848e-01 -3.39319021e-01 -9.45763886e-02 -1.52495146e-01
2.36269742e-01 1.13735056e+00 8.49747539e-01 -1.34803116e+00
-7.09904969e-01 -3.47992063e-01 -7.06383362e-02 4.75436568e-01
-3.14425588e-01 -2.94282585e-01 -6.43155634e-01 -8.45805883e-01
1.54241964e-01 -8.51782680e-01 -3.90444636e-01 -1.44711733e-01
-4.11767453e-01 -4.36903685e-01 7.57273257e-01 -6.82801902e-01
1.33311772e+00 -2.06760120e+00 5.60221553e-01 5.28438501e-02
1.97923422e-01 6.90117419e-01 -1.76777840e-01 1.47024125e-01
-9.73521397e-02 -1.61632672e-01 -8.00274730e-01 -7.73094714e-01
9.35832784e-02 3.64692062e-01 -3.21268588e-01 1.08172461e-01
4.17645425e-01 6.69599354e-01 -1.08022845e+00 -6.02208018e-01
1.20922670e-01 4.82928872e-01 -7.86455214e-01 4.33630437e-01
-6.26291394e-01 9.33710098e-01 -5.04703581e-01 4.47404027e-01
7.74083257e-01 -2.51214266e-01 1.58741608e-01 1.05939239e-01
4.53479029e-02 4.90938336e-01 -1.15069401e+00 1.98738337e+00
-3.18298131e-01 1.71771735e-01 2.70215452e-01 -1.24882805e+00
1.12814391e+00 1.37699306e-01 3.09045941e-01 -6.94294930e-01
1.36496961e-01 4.41427201e-01 -7.19488114e-02 -2.01969638e-01
3.56664449e-01 -3.94244552e-01 -5.80802895e-02 2.34097749e-01
3.32567543e-01 -3.69325578e-01 1.36645734e-01 1.50333375e-01
7.40564048e-01 6.72929823e-01 7.97259659e-02 -6.34239465e-02
5.98320305e-01 1.92678675e-01 9.17182505e-01 3.87005925e-01
-3.88442844e-01 8.90347600e-01 5.42213440e-01 -9.14518386e-02
-1.16398287e+00 -1.26421571e+00 -8.09388459e-02 1.11615050e+00
3.14167082e-01 -8.90108347e-02 -1.18008876e+00 -9.33977544e-01
-4.61377166e-02 8.16735268e-01 -5.25019526e-01 -1.16643086e-01
-6.46742582e-01 -6.08341455e-01 4.39886689e-01 4.72291470e-01
8.07383776e-01 -9.82139349e-01 -1.88250408e-01 3.89842182e-01
-2.06398666e-01 -9.89504695e-01 -4.83175963e-01 2.59643286e-01
-8.43593478e-01 -7.40413845e-01 -1.22076368e+00 -8.52789640e-01
6.49214089e-01 8.00226554e-02 1.02496326e+00 -1.62696064e-01
1.68127000e-01 -1.01315871e-01 -3.46298367e-01 -3.11130852e-01
-6.60787702e-01 3.88154387e-01 -1.56386852e-01 -5.73314838e-02
2.93733358e-01 -8.79824936e-01 -6.80496097e-01 3.88265193e-01
-8.88242245e-01 3.92015249e-01 5.69172919e-01 8.72074664e-01
7.83942223e-01 -1.52217284e-01 6.53546333e-01 -1.29128015e+00
6.12088084e-01 -7.25389898e-01 -5.31553686e-01 2.16659233e-01
-6.58938289e-01 1.19408555e-01 7.44351625e-01 -7.79012799e-01
-1.49238336e+00 1.99926030e-02 -4.23913598e-01 -6.33494377e-01
-5.52632570e-01 2.04935849e-01 -4.48889256e-01 3.50581646e-01
7.11408138e-01 2.32416213e-01 8.96816403e-02 -7.08647192e-01
6.20498061e-01 5.40114820e-01 5.69494545e-01 -9.54021573e-01
7.82419562e-01 1.91956207e-01 -6.53452754e-01 -4.46486264e-01
-9.44768488e-01 -2.66910821e-01 -7.90860891e-01 1.49551451e-01
8.89149129e-01 -9.08162534e-01 1.47822827e-01 8.30257297e-01
-9.24754262e-01 -9.56676245e-01 -5.47563732e-01 2.71263719e-01
-3.90717924e-01 1.77791461e-01 -4.59477007e-01 -2.65702724e-01
-8.03702027e-02 -1.22321177e+00 8.01371336e-01 5.96432209e-01
-2.51395762e-01 -1.09648883e+00 3.50294948e-01 7.18235314e-01
3.23430598e-01 9.20136794e-02 7.90135801e-01 -8.27884495e-01
-1.94488257e-01 2.54915655e-01 -2.13364914e-01 4.96369034e-01
2.36648262e-01 -1.30793840e-01 -8.35518122e-01 -1.71641186e-01
-2.25104541e-01 -3.61896902e-01 7.41630554e-01 2.51208752e-01
1.10689235e+00 -9.07330215e-02 -2.60773987e-01 6.50761306e-01
1.05356658e+00 1.51932061e-01 5.22828519e-01 5.02955973e-01
8.41816783e-01 6.96741879e-01 7.30368853e-01 2.10970238e-01
6.51065886e-01 5.83013654e-01 2.35463046e-02 -2.54410893e-01
-5.20915329e-01 -4.50982213e-01 2.61826336e-01 9.20421720e-01
1.66245759e-01 -1.05883278e-01 -1.08655512e+00 9.15403247e-01
-1.74889326e+00 -5.40047109e-01 3.61362398e-02 1.99641311e+00
1.51853764e+00 3.21133763e-01 4.24496055e-01 -1.68920964e-01
7.75518835e-01 2.69768108e-02 -8.19586933e-01 -3.17798436e-01
9.82227623e-02 3.34557801e-01 1.69961795e-01 5.92309654e-01
-1.00551260e+00 1.25210524e+00 4.56132221e+00 1.12714303e+00
-1.19267654e+00 3.28642368e-01 6.95527792e-01 1.37555093e-01
-5.03985167e-01 1.62376061e-01 -8.46617043e-01 7.32147515e-01
7.04968452e-01 7.12447688e-02 2.68606931e-01 8.34704101e-01
2.34846715e-02 5.65305948e-02 -7.79662013e-01 5.38736880e-01
-4.70881581e-01 -9.95024979e-01 1.38386954e-02 -6.40545860e-02
9.67258751e-01 -6.82381094e-02 -3.31891328e-02 5.53985059e-01
7.94888973e-01 -5.03295183e-01 7.49126673e-01 1.10579006e-01
6.19696259e-01 -5.55311203e-01 3.44830632e-01 4.45706248e-01
-1.04521060e+00 9.85578746e-02 -1.11732550e-01 2.68080026e-01
2.18789399e-01 6.38120532e-01 -9.95034337e-01 5.65274954e-01
6.74482763e-01 6.25972927e-01 -2.70978779e-01 6.90401614e-01
-5.48080683e-01 8.14592123e-01 -2.71755815e-01 4.55253601e-01
4.17293847e-01 -5.80128133e-01 6.67295873e-01 1.12211967e+00
1.91903979e-01 1.34003341e-01 1.28689528e-01 1.23778808e+00
-1.68615162e-01 -9.84424353e-02 -2.43290693e-01 3.27879041e-02
7.17142045e-01 1.05024445e+00 -6.73902392e-01 -4.41919357e-01
-3.37278843e-01 1.08200216e+00 5.04890323e-01 3.71459961e-01
-1.04477358e+00 -3.38609934e-01 7.88842916e-01 2.71170467e-01
3.23889524e-01 -1.30373895e-01 -5.74743032e-01 -1.07647502e+00
-6.05735518e-02 -1.00381708e+00 4.71479625e-01 -4.75580424e-01
-1.47814834e+00 4.56135750e-01 1.09813407e-01 -1.24496841e+00
2.35899910e-03 -7.12946504e-02 -7.26770937e-01 1.02972734e+00
-1.57182014e+00 -1.23518395e+00 -3.50169033e-01 6.48616076e-01
8.30268323e-01 1.85077060e-02 5.18884838e-01 4.95664001e-01
-8.12837601e-01 7.66356647e-01 2.07832586e-02 6.36488050e-02
9.75762367e-01 -1.52484441e+00 4.66392994e-01 8.34753871e-01
-1.16828948e-01 4.55420822e-01 6.20627403e-01 -7.36228228e-01
-6.33699000e-01 -1.10586548e+00 5.94358683e-01 -2.37592056e-01
5.93753099e-01 -2.84737974e-01 -1.46665931e+00 5.86519718e-01
2.90072441e-01 -1.42727822e-01 5.64285576e-01 1.49853034e-02
-3.00923198e-01 -1.68624297e-01 -1.14841819e+00 5.89149177e-01
9.55082774e-01 -2.67242372e-01 -8.15470695e-01 2.28730962e-02
8.64417195e-01 -5.74820220e-01 -9.19983864e-01 3.74692589e-01
2.16873348e-01 -9.03135717e-01 1.02294934e+00 -4.87258106e-01
6.60881042e-01 -3.13430518e-01 1.81700498e-01 -1.53876221e+00
-9.78820473e-02 -5.16500592e-01 9.26019698e-02 1.74684668e+00
4.39267278e-01 -7.46684313e-01 8.58690262e-01 5.57111919e-01
-3.15189511e-01 -9.30979192e-01 -6.99851573e-01 -6.46213770e-01
5.47959566e-01 -2.16103241e-01 7.26563811e-01 1.35691750e+00
-3.00844848e-01 3.83643270e-01 4.28988934e-02 7.12818354e-02
5.42959929e-01 9.67557430e-02 8.67150724e-01 -1.13306975e+00
-4.94743675e-01 -7.85549819e-01 3.33509296e-01 -1.38109124e+00
2.42683794e-02 -8.25793684e-01 1.66356623e-01 -1.47883880e+00
4.94172164e-05 -1.06107438e+00 -2.10562408e-01 6.07573390e-01
-4.97488916e-01 -6.33068308e-02 1.51504606e-01 3.06779981e-01
-4.14831936e-01 6.75146639e-01 1.47009909e+00 -5.58047229e-03
-6.55292094e-01 8.44904408e-02 -8.92930686e-01 7.26482689e-01
8.17269981e-01 -5.11754513e-01 -6.84818029e-01 -6.09217227e-01
-2.81442642e-01 4.83437069e-03 1.50442466e-01 -8.46852839e-01
7.89072216e-02 -2.09895492e-01 1.69802532e-01 -2.32970878e-01
1.59731492e-01 -5.36121666e-01 7.81338010e-03 6.64156526e-02
-3.39744717e-01 -4.89381313e-01 2.30464295e-01 4.01730686e-01
-4.15304780e-01 -1.82057798e-01 1.25983727e+00 -1.80229306e-01
-7.12818384e-01 2.26890191e-01 2.22214520e-01 3.88282895e-01
1.10898018e+00 -4.47865903e-01 -6.58350810e-02 5.21479584e-02
-8.97076905e-01 5.27685285e-01 5.02855539e-01 4.07891631e-01
2.23775253e-01 -1.08619416e+00 -7.50386119e-01 2.62285501e-01
-2.31761619e-01 9.04791594e-01 4.26815033e-01 9.13575888e-01
-2.73524016e-01 -1.51419058e-01 -2.47389108e-01 -6.86929703e-01
-9.76746440e-01 1.63413793e-01 3.36944312e-01 -6.19001627e-01
-4.87918019e-01 1.29383850e+00 3.75296801e-01 -8.56330633e-01
2.72663593e-01 -1.72589719e-01 -4.78822961e-02 2.94787884e-01
1.54438332e-01 1.73063815e-01 -1.20415397e-01 -3.76662374e-01
-1.75411478e-01 5.64067543e-01 -3.60805839e-01 -7.20345527e-02
1.42389345e+00 -2.81056225e-01 2.05079392e-01 1.51155770e-01
8.64709616e-01 -2.15924561e-01 -1.88344109e+00 -5.41901469e-01
3.49266501e-03 -3.19486707e-01 -2.18802616e-01 -9.48471785e-01
-1.00249970e+00 1.03774786e+00 3.78313899e-01 1.30472839e-01
1.38598716e+00 1.39614746e-01 1.17069173e+00 -3.51852328e-01
-2.41981521e-02 -1.22603357e+00 2.90351927e-01 4.56379265e-01
5.66941261e-01 -1.33293593e+00 -3.74456167e-01 -3.74302983e-01
-9.58748281e-01 7.40147591e-01 9.89428401e-01 -2.38634385e-02
3.24541301e-01 1.17190875e-01 9.45389941e-02 2.14924887e-01
-4.96787220e-01 -2.90083289e-01 2.09122941e-01 5.17483890e-01
2.73213029e-01 5.48790768e-02 -2.76673257e-01 9.34264183e-01
-1.57867506e-01 3.32681984e-02 5.77720851e-02 8.65750968e-01
-3.26796561e-01 -1.67948580e+00 -3.20486307e-01 -6.13385532e-03
-2.67841578e-01 7.14862794e-02 -3.02079946e-01 8.09772730e-01
4.98078227e-01 5.61373830e-01 8.93967878e-03 -2.14708567e-01
4.47281718e-01 3.02801102e-01 4.05522525e-01 -6.92435265e-01
-5.57156265e-01 2.12353945e-01 -1.44755065e-01 -3.26690376e-01
-2.61271119e-01 -7.60073423e-01 -1.49862087e+00 -1.52710885e-01
-8.59455764e-02 1.47491455e-01 4.56442684e-01 1.00435925e+00
4.00359660e-01 6.11850500e-01 3.66885275e-01 -7.12421358e-01
-5.14524162e-01 -9.25344706e-01 -1.86532646e-01 6.49768114e-01
2.87962317e-01 -7.40219116e-01 -8.24264362e-02 1.55293971e-01] | [9.632830619812012, 1.3774003982543945] |
746728e7-7921-4426-9dfa-c3c192bf1162 | face-sketch-synthesis-style-similaritya-new | 1804.02975 | null | http://arxiv.org/abs/1804.02975v1 | http://arxiv.org/pdf/1804.02975v1.pdf | Face Sketch Synthesis Style Similarity:A New Structure Co-occurrence Texture Measure | Existing face sketch synthesis (FSS) similarity measures are sensitive to
slight image degradation (e.g., noise, blur). However, human perception of the
similarity of two sketches will consider both structure and texture as
essential factors and is not sensitive to slight ("pixel-level") mismatches.
Consequently, the use of existing similarity measures can lead to better
algorithms receiving a lower score than worse algorithms. This unreliable
evaluation has significantly hindered the development of the FSS field. To
solve this problem, we propose a novel and robust style similarity measure
called Scoot-measure (Structure CO-Occurrence Texture Measure), which
simultaneously evaluates "block-level" spatial structure and co-occurrence
texture statistics. In addition, we further propose 4 new meta-measures and
create 2 new datasets to perform a comprehensive evaluation of several
widely-used FSS measures on two large databases. Experimental results
demonstrate that our measure not only provides a reliable evaluation but also
achieves significantly improved performance. Specifically, the study indicated
a higher degree (78.8%) of correlation between our measure and human judgment
than the best prior measure (58.6%). Our code will be made available. | ['Paul L. Rosin', 'Ming-Ming Cheng', 'Yu-Huan Wu', 'Shengchuan Zhang', 'Deng-Ping Fan', 'Bo Ren', 'Rongrong Ji'] | 2018-04-09 | null | null | null | null | ['face-sketch-synthesis'] | ['computer-vision'] | [ 2.11000815e-01 -6.05691075e-01 -5.60211614e-02 -4.07484263e-01
-4.98724669e-01 -2.95135617e-01 6.94607019e-01 -9.50770900e-02
1.79605708e-02 5.61437964e-01 1.23375610e-01 1.28107101e-01
-9.12964866e-02 -8.32175136e-01 -1.64486974e-01 -5.04394710e-01
3.18517983e-01 -1.80031255e-01 4.47856992e-01 -3.09868753e-01
7.61272252e-01 7.69926190e-01 -1.77179158e+00 4.08798575e-01
8.89585793e-01 1.02021527e+00 2.04690218e-01 2.57857680e-01
-2.75360137e-01 2.50435799e-01 -8.60159278e-01 -6.82730138e-01
2.26278901e-01 -5.22952259e-01 -3.65229249e-01 -4.05529402e-02
8.14374387e-01 -2.91726530e-01 -1.41830474e-01 1.14864731e+00
6.70554519e-01 -4.12322022e-03 7.16237128e-01 -1.21960139e+00
-8.68251562e-01 -4.11353074e-02 -7.35492527e-01 8.43599737e-02
5.97388029e-01 1.10608600e-01 8.06298971e-01 -1.10702407e+00
5.09633780e-01 1.46033275e+00 7.47558773e-01 3.01489174e-01
-1.08062387e+00 -9.06379104e-01 -1.87227249e-01 2.31478304e-01
-1.54896379e+00 -5.07191062e-01 1.07599068e+00 -1.51141703e-01
5.29911757e-01 5.75870395e-01 4.18043107e-01 9.72365797e-01
1.51235819e-01 5.80655396e-01 1.56588233e+00 -4.86951768e-01
7.25471899e-02 2.07819358e-01 -2.90978819e-01 8.14773560e-01
3.26529413e-01 8.05881768e-02 -6.40040100e-01 -1.85944602e-01
1.07761073e+00 -4.64978740e-02 -9.44122374e-02 -2.89629698e-01
-1.09953499e+00 4.43787545e-01 1.57369420e-01 6.13045812e-01
-1.20541357e-01 -1.97497442e-01 2.06443220e-01 4.47742492e-01
3.92932504e-01 2.70778179e-01 1.97679251e-02 -2.06608161e-01
-9.65316176e-01 1.85014352e-01 6.21678591e-01 7.83324897e-01
5.96155226e-01 -1.85213834e-02 -4.47343588e-01 1.25788689e+00
1.88796490e-01 8.08878362e-01 3.05640399e-01 -9.54730809e-01
2.26970956e-01 5.07067025e-01 5.35096154e-02 -1.86333561e+00
-7.85329342e-02 -2.14287162e-01 -1.05679262e+00 4.30288941e-01
3.93126220e-01 4.99777734e-01 -5.96602559e-01 1.66400337e+00
1.35585919e-01 1.08995005e-01 -3.80980432e-01 9.53869104e-01
9.43716109e-01 3.60972434e-01 -1.18473165e-01 -2.20687836e-01
1.35293269e+00 -7.40856111e-01 -9.02678132e-01 1.28568979e-02
-1.29523978e-01 -1.42597055e+00 1.41474009e+00 2.69985229e-01
-1.22674394e+00 -9.81082201e-01 -1.38196778e+00 2.86268085e-01
-3.35158139e-01 2.60473907e-01 5.82959294e-01 8.66221607e-01
-1.12391722e+00 7.79946208e-01 -3.19661379e-01 -4.72718626e-01
3.08834374e-01 -2.63037309e-02 -3.78830224e-01 -1.11774705e-01
-9.40880001e-01 9.72219586e-01 -1.71450511e-01 -1.35661229e-01
-2.96210945e-01 -5.18670261e-01 -4.12965149e-01 -4.97895665e-03
2.21369371e-01 -4.54455376e-01 7.50977397e-01 -9.00225580e-01
-1.51824665e+00 9.81070101e-01 -3.07673544e-01 4.36068654e-01
4.46951866e-01 1.80174962e-01 -6.67553067e-01 3.02696526e-01
5.77492267e-02 4.87027407e-01 1.09376466e+00 -1.47976494e+00
-3.35436612e-01 -3.69030803e-01 -1.55108988e-01 9.47464556e-02
-4.51696455e-01 1.85131043e-01 -5.11413991e-01 -1.11052942e+00
3.53971303e-01 -6.36449635e-01 3.35284382e-01 5.64307034e-01
1.49298400e-01 -2.47538447e-01 7.19076693e-01 -6.32437646e-01
1.44849169e+00 -2.22613096e+00 -2.73319602e-01 4.03101176e-01
1.01719208e-01 5.60660839e-01 -3.83918226e-01 4.61754233e-01
9.74987745e-02 1.21290490e-01 -3.35258603e-01 -1.98006835e-02
-2.42686700e-02 -3.04698329e-02 -4.82438579e-02 2.75130570e-01
1.32375896e-01 6.78164840e-01 -8.46658766e-01 -8.17283034e-01
2.13233024e-01 4.50430721e-01 -3.04563433e-01 2.72916049e-01
2.99663991e-01 7.51660839e-02 -2.10537404e-01 1.03065908e+00
1.07100213e+00 4.06094305e-02 5.91688044e-02 -5.39293051e-01
-7.69544067e-03 -2.28448004e-01 -1.33588839e+00 1.43378198e+00
-3.03487450e-01 6.69538379e-01 -1.23576857e-01 -6.71693206e-01
1.50580943e+00 1.07698679e-01 4.28258598e-01 -1.12289762e+00
-6.80321455e-02 2.56502301e-01 -1.70215353e-01 -4.48597997e-01
4.14890587e-01 -7.57699311e-02 3.92197132e-01 3.56767267e-01
-2.38544881e-01 -1.81002602e-01 7.95267746e-02 3.58402096e-02
8.76613736e-01 -1.01563744e-01 3.16552222e-01 -4.86140281e-01
8.42188895e-01 -5.70877135e-01 4.66347694e-01 6.06904507e-01
-5.98305464e-01 8.20325315e-01 4.81571138e-01 -3.48700255e-01
-1.06540287e+00 -1.30473006e+00 -1.49290994e-01 6.65797293e-01
4.06015784e-01 -6.04813337e-01 -7.55170822e-01 -4.88939166e-01
2.40420878e-01 1.79775462e-01 -5.27557611e-01 -1.82315782e-01
-2.72778690e-01 -4.48821157e-01 6.10654354e-01 4.91665274e-01
9.81499374e-01 -9.98893023e-01 -2.27066413e-01 -2.09305197e-01
-2.00898275e-01 -8.87246132e-01 -7.59915709e-01 -9.00649011e-01
-7.22878516e-01 -1.04463267e+00 -1.02483308e+00 -8.39949131e-01
6.40091598e-01 7.25059509e-01 1.06241167e+00 5.40895581e-01
-3.51836890e-01 3.69000047e-01 -3.03297341e-01 -8.01084042e-02
-2.47696787e-01 -3.06727976e-01 7.95553327e-02 1.79540753e-01
2.52204716e-01 -6.21532023e-01 -8.69157493e-01 8.70237291e-01
-8.72099698e-01 -3.21201272e-02 6.65096462e-01 7.87718058e-01
3.79991025e-01 -4.91612032e-02 6.43273056e-01 -2.74841815e-01
1.04898560e+00 5.23560159e-02 -3.33482385e-01 5.93614399e-01
-1.02215111e+00 -4.95701730e-02 3.88006717e-01 -4.40674543e-01
-1.27650619e+00 -4.90023136e-01 2.27319766e-02 -1.76258221e-01
-7.37888142e-02 1.46470398e-01 -1.51460782e-01 -4.50697392e-01
4.72871125e-01 2.11489752e-01 3.02146733e-01 -4.72570270e-01
-9.53194946e-02 8.03822815e-01 5.61013997e-01 -7.82997787e-01
7.80053854e-01 4.04796183e-01 1.29746392e-01 -1.01105857e+00
-2.42242828e-01 -2.40810484e-01 -4.66297746e-01 -4.99025136e-01
2.90156841e-01 -5.41326940e-01 -8.66723120e-01 7.45193124e-01
-1.05866718e+00 2.07642719e-01 1.90579042e-01 3.60456407e-01
-2.68905669e-01 8.91910195e-01 -4.52848107e-01 -8.01427126e-01
-2.72897065e-01 -9.78294075e-01 9.84902143e-01 3.72900128e-01
-2.59664923e-01 -7.47788250e-01 -2.82202065e-02 4.29219991e-01
7.88692296e-01 2.10192785e-01 7.11393774e-01 1.16174906e-01
-3.10345739e-01 -8.44148770e-02 -7.89027452e-01 4.46673781e-01
4.52448428e-01 4.45025116e-01 -9.56139445e-01 -2.29341269e-01
-6.97623044e-02 5.86718619e-02 6.06673956e-01 7.45908171e-02
1.16622746e+00 -4.66864295e-02 -7.07447603e-02 3.82089168e-01
1.30410898e+00 4.13960159e-01 9.95396733e-01 2.20261291e-01
4.15508121e-01 7.28585243e-01 8.29205155e-01 5.21182835e-01
9.71411318e-02 1.05425525e+00 -1.21401966e-01 -1.86257079e-01
-5.98174810e-01 -3.98976415e-01 2.79271215e-01 9.08408999e-01
-3.85165751e-01 -1.88475624e-02 -6.45824373e-01 1.99878439e-01
-1.55887806e+00 -1.01340234e+00 -1.10115901e-01 2.27398610e+00
8.25210571e-01 -2.98117176e-02 1.30873874e-01 3.36006492e-01
9.24592197e-01 2.81772375e-01 -2.39015907e-01 -4.52473283e-01
-3.70439649e-01 3.72336775e-01 7.24858195e-02 3.99965078e-01
-7.91931987e-01 9.04476941e-01 6.78704691e+00 1.29266429e+00
-1.05442142e+00 -2.52607107e-01 6.01871550e-01 2.24157125e-01
-4.73942310e-01 -1.66567862e-01 -3.19886893e-01 8.52488399e-01
3.57476026e-01 -1.24421500e-01 3.91988933e-01 5.82564116e-01
3.19713689e-02 -2.64244616e-01 -7.74312496e-01 1.59423840e+00
2.78505415e-01 -1.03171730e+00 1.57629669e-01 -1.11334443e-01
7.81809568e-01 -8.40756893e-01 2.76947707e-01 -1.03190713e-01
-3.89802247e-01 -9.82091069e-01 5.43289840e-01 8.28818381e-01
1.08385003e+00 -6.03203118e-01 6.63231671e-01 -2.92017221e-01
-1.42990959e+00 3.88762265e-01 -4.60388303e-01 2.92379670e-02
-1.61771223e-01 5.83564699e-01 -4.91718024e-01 3.86966735e-01
6.02270961e-01 5.66346705e-01 -7.80160606e-01 9.27252769e-01
-9.27637592e-02 2.19671130e-01 -1.21468060e-01 -2.41908640e-01
-1.42888948e-01 -3.56588095e-01 3.25273305e-01 1.14850366e+00
3.88607115e-01 4.66471240e-02 -1.72802195e-01 8.44960451e-01
1.33832023e-01 4.01837587e-01 -6.19314909e-01 -4.65160161e-02
7.00217903e-01 1.12620378e+00 -8.49396110e-01 -3.94294620e-01
-5.53310096e-01 1.21912622e+00 1.57449804e-02 2.72605360e-01
-7.66036212e-01 -5.87535322e-01 7.37079859e-01 1.05486445e-01
-1.42170712e-02 -2.26866364e-01 -6.40015483e-01 -8.70929480e-01
2.88983196e-01 -9.47157800e-01 1.60178974e-01 -8.52841139e-01
-1.39931953e+00 4.32545632e-01 -1.35942087e-01 -1.31659412e+00
2.13487208e-01 -5.86682856e-01 -6.91532254e-01 8.49468052e-01
-1.34480739e+00 -9.79433119e-01 -6.93962634e-01 5.56450725e-01
3.41631919e-01 -4.53257531e-01 7.07027078e-01 5.89141667e-01
-4.39853430e-01 1.06352901e+00 -9.44075882e-02 -1.02966391e-01
1.12490177e+00 -9.06357765e-01 4.23730701e-01 5.47607780e-01
-3.27425636e-02 6.71312690e-01 4.83220667e-01 -7.10929930e-01
-1.25739598e+00 -3.48641336e-01 8.80728066e-01 -2.22221106e-01
3.50844234e-01 1.03192506e-02 -9.43634033e-01 -1.86075851e-01
-4.78067361e-02 -1.15766130e-01 4.81125265e-01 -8.63655359e-02
-5.92742264e-01 -5.41270733e-01 -1.49511170e+00 7.62820780e-01
1.32839704e+00 -7.16873527e-01 -4.86673057e-01 -3.07755530e-01
1.51876863e-02 -7.64967799e-02 -1.03214490e+00 6.77902758e-01
1.24849284e+00 -1.43119943e+00 1.19785666e+00 -1.38568103e-01
3.61399740e-01 -3.29724491e-01 -3.20553541e-01 -9.31952894e-01
-5.86351156e-01 -3.60973597e-01 5.60490265e-02 1.40107691e+00
1.29438087e-01 -6.33566916e-01 5.85371912e-01 4.80534196e-01
1.55587971e-01 -8.38258862e-01 -7.20651984e-01 -9.82505083e-01
-3.46045732e-01 -6.32191375e-02 8.19057584e-01 1.06667113e+00
2.85268612e-02 1.79619212e-02 -3.92821878e-01 -1.99563563e-01
6.27878129e-01 1.28913477e-01 7.71421313e-01 -1.15590191e+00
2.19498739e-01 -9.00082886e-01 -6.43209636e-01 -7.82332599e-01
-1.17165700e-01 -5.65276444e-01 -2.97401160e-01 -1.26933873e+00
4.43717778e-01 -5.02013087e-01 -2.91207761e-01 1.94664344e-01
-3.86541456e-01 5.47396600e-01 1.50360331e-01 1.75772935e-01
-3.23203713e-01 6.07687473e-01 1.64072680e+00 -1.76561981e-01
1.58369586e-01 -8.81544501e-02 -6.01268530e-01 5.71619093e-01
9.12884772e-01 5.67583032e-02 -3.30873191e-01 -1.51455060e-01
-5.21779023e-02 -8.80064219e-02 2.65401810e-01 -1.17322004e+00
9.62347239e-02 -4.27174926e-01 5.62936246e-01 -3.94643188e-01
3.62270325e-01 -7.02417433e-01 1.57801777e-01 4.63278115e-01
-7.63630494e-02 2.35055879e-01 9.12145376e-02 2.25119516e-01
-3.01489085e-01 -1.03471026e-01 9.62759972e-01 3.80682312e-02
-7.86677361e-01 1.16756856e-01 3.80131900e-02 -3.39100033e-01
7.99466789e-01 -6.95022345e-01 -3.23328406e-01 -4.43033367e-01
-8.47112611e-02 -3.58935148e-01 7.08943963e-01 6.55878425e-01
9.91316736e-01 -1.72555804e+00 -7.28070855e-01 4.77126479e-01
1.73602253e-01 -7.75302172e-01 2.76483864e-01 5.59328735e-01
-6.40787184e-01 3.01385254e-01 -5.80075085e-01 -4.71339494e-01
-1.60493195e+00 3.32674801e-01 -1.31646730e-03 2.32683271e-01
-2.18906105e-01 6.66025758e-01 5.09907715e-02 -5.31884693e-02
2.50927776e-01 -8.87908265e-02 -3.06234881e-02 -4.19640578e-02
6.71442926e-01 8.16587687e-01 -2.33753631e-03 -6.03537440e-01
-4.68595386e-01 9.91684437e-01 3.94738019e-02 -6.79102316e-02
9.11600053e-01 -2.29818281e-02 -2.00604156e-01 1.64500296e-01
1.19896340e+00 2.13516206e-01 -9.70458090e-01 -1.21065378e-01
-6.38093054e-02 -1.13422871e+00 -1.75424382e-01 -9.05636311e-01
-1.12822640e+00 6.89156175e-01 9.63707626e-01 1.09056167e-01
1.25201094e+00 -2.95142412e-01 8.70164990e-01 7.52789974e-02
3.42170388e-01 -1.33993304e+00 5.30893505e-01 1.57424167e-01
1.20540249e+00 -1.28349757e+00 1.75324470e-01 -5.81674576e-01
-5.50039113e-01 1.11913824e+00 7.84419417e-01 6.09222315e-02
6.15573943e-01 2.64265358e-01 6.46474138e-02 -1.52265623e-01
-3.64012152e-01 -6.38771653e-02 6.46557331e-01 6.57882929e-01
5.91647983e-01 5.73659688e-02 -7.22285032e-01 3.92424196e-01
-3.53822373e-02 -1.08523257e-02 -4.45385464e-02 7.79473722e-01
-5.03983140e-01 -1.30580378e+00 -5.92135489e-01 4.03000146e-01
-1.65657297e-01 1.39735088e-01 -6.16578579e-01 5.83700776e-01
8.86541903e-02 8.95431161e-01 -2.42752749e-02 -6.91942871e-01
5.42895675e-01 -1.73858970e-01 8.13300610e-01 -2.85125580e-02
-1.72653273e-01 -1.15463808e-01 -1.40165269e-01 -6.52495623e-01
-6.03596807e-01 -4.43759918e-01 -6.67016268e-01 -9.68091309e-01
-1.35353699e-01 -1.02564991e-01 6.56708121e-01 5.40638328e-01
3.14768225e-01 1.51978835e-01 8.37718904e-01 -5.80353022e-01
-3.84082258e-01 -9.68495190e-01 -6.05224490e-01 7.22557127e-01
-1.74334615e-01 -9.36358929e-01 -2.56885827e-01 -2.29972005e-01] | [12.759565353393555, 0.21592658758163452] |
2f0b1fac-99aa-4e2b-a5a7-f874c196f7ca | putting-people-in-their-place-monocular | 2112.08274 | null | https://arxiv.org/abs/2112.08274v3 | https://arxiv.org/pdf/2112.08274v3.pdf | Putting People in their Place: Monocular Regression of 3D People in Depth | Given an image with multiple people, our goal is to directly regress the pose and shape of all the people as well as their relative depth. Inferring the depth of a person in an image, however, is fundamentally ambiguous without knowing their height. This is particularly problematic when the scene contains people of very different sizes, e.g. from infants to adults. To solve this, we need several things. First, we develop a novel method to infer the poses and depth of multiple people in a single image. While previous work that estimates multiple people does so by reasoning in the image plane, our method, called BEV, adds an additional imaginary Bird's-Eye-View representation to explicitly reason about depth. BEV reasons simultaneously about body centers in the image and in depth and, by combing these, estimates 3D body position. Unlike prior work, BEV is a single-shot method that is end-to-end differentiable. Second, height varies with age, making it impossible to resolve depth without also estimating the age of people in the image. To do so, we exploit a 3D body model space that lets BEV infer shapes from infants to adults. Third, to train BEV, we need a new dataset. Specifically, we create a "Relative Human" (RH) dataset that includes age labels and relative depth relationships between the people in the images. Extensive experiments on RH and AGORA demonstrate the effectiveness of the model and training scheme. BEV outperforms existing methods on depth reasoning, child shape estimation, and robustness to occlusion. The code and dataset are released for research purposes. | ['Michael J. Black', 'Tao Mei', 'Yili Fu', 'Qian Bao', 'Wu Liu', 'Yu Sun'] | 2021-12-15 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Sun_Putting_People_in_Their_Place_Monocular_Regression_of_3D_People_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Sun_Putting_People_in_Their_Place_Monocular_Regression_of_3D_People_CVPR_2022_paper.pdf | cvpr-2022-1 | ['3d-depth-estimation'] | ['computer-vision'] | [-1.86918914e-01 2.66043186e-01 -4.18068133e-02 -4.79466826e-01
-1.56569660e-01 -6.62217379e-01 1.90128744e-01 3.56346332e-02
-1.90397039e-01 2.45185331e-01 1.72069445e-01 1.36984408e-01
3.23273152e-01 -8.82861316e-01 -7.82547116e-01 -4.17167604e-01
4.11787242e-01 9.32348430e-01 1.91796198e-01 -9.48539376e-02
6.57501668e-02 3.57566208e-01 -1.62695277e+00 -1.53594837e-01
7.36472368e-01 8.43149424e-01 -1.88342661e-01 7.71656573e-01
1.84825010e-04 4.66950119e-01 -4.84778970e-01 -7.25164413e-01
4.27002221e-01 -2.35369965e-01 -5.98965287e-01 4.64422524e-01
1.01269996e+00 -9.43658471e-01 -1.37744755e-01 6.56242907e-01
5.14730692e-01 -3.30838412e-02 8.12496364e-01 -1.24193823e+00
-3.65335643e-01 3.59754592e-01 -1.20615482e+00 -2.20096901e-01
7.82353997e-01 -1.73055768e-01 5.20860672e-01 -7.61527777e-01
4.32590663e-01 1.51154923e+00 8.29310119e-01 8.71492624e-01
-1.09239507e+00 -7.40717590e-01 5.00905216e-01 -2.94111669e-01
-1.32453978e+00 -3.44035208e-01 8.15271914e-01 -8.27470899e-01
4.32231486e-01 8.55235159e-02 1.15290916e+00 8.76658440e-01
-3.76228034e-01 7.37739801e-01 7.43076980e-01 -3.87672603e-01
4.67473194e-02 -3.07167083e-01 -4.01057452e-02 9.73366797e-01
4.09792393e-01 -1.36371583e-01 -6.12739801e-01 1.29301637e-01
1.14769852e+00 5.73290661e-02 -8.85802074e-05 -7.65894949e-01
-9.49776709e-01 7.21858919e-01 2.49181688e-01 -2.37922415e-01
4.47942019e-02 1.47331595e-01 -9.91251320e-02 -1.29334733e-01
5.08216739e-01 -1.71684232e-02 -5.11115849e-01 -4.56604585e-02
-7.85400510e-01 5.46836495e-01 8.73895824e-01 9.93304133e-01
8.14418018e-01 -3.10883790e-01 1.03097297e-01 6.95308268e-01
4.39122915e-01 6.73770308e-01 1.06555507e-01 -1.26699412e+00
5.40302455e-01 7.61824012e-01 4.69147712e-02 -9.98437881e-01
-4.98674959e-01 9.41662118e-02 -6.10274911e-01 3.75502110e-01
9.41006124e-01 -2.82218069e-01 -1.09383452e+00 1.97277665e+00
9.78497684e-01 2.24099848e-02 -2.17974365e-01 9.09376204e-01
1.06083941e+00 1.28190458e-01 -3.23397934e-01 -2.87041487e-03
1.46004188e+00 -9.87100184e-01 -3.69063675e-01 -6.09316528e-01
1.72851995e-01 -3.88414592e-01 6.76358759e-01 4.86949474e-01
-1.55792403e+00 -3.67917001e-01 -8.21566641e-01 -3.79955947e-01
-1.64850950e-01 -3.60662714e-02 8.73441458e-01 7.59945035e-01
-9.66437519e-01 3.46131444e-01 -9.91577625e-01 -4.59305406e-01
2.39059925e-01 5.13832271e-01 -4.11950171e-01 1.09353602e-01
-7.14509368e-01 7.92790055e-01 3.57063636e-02 -8.40634182e-02
-6.01728797e-01 -7.06450164e-01 -1.27455485e+00 -2.20517397e-01
4.05202270e-01 -1.03029370e+00 1.41055155e+00 -8.92871022e-01
-1.26408124e+00 1.31233966e+00 -3.10590595e-01 -2.12148540e-02
9.62021410e-01 -3.45911264e-01 2.74039716e-01 2.90088624e-01
1.35606989e-01 9.90963399e-01 8.87156725e-01 -1.19074786e+00
-4.68886346e-01 -9.99261916e-01 3.46294403e-01 5.28996706e-01
-3.26123238e-02 -6.77756593e-02 -9.26683247e-01 -3.41152430e-01
6.75065696e-01 -8.00116420e-01 1.98917519e-02 7.51474261e-01
-2.19022691e-01 -1.29692405e-01 4.20638621e-01 -7.81439662e-01
7.57359982e-01 -2.00475144e+00 3.07156175e-01 2.22495478e-02
5.80266953e-01 -1.39671251e-01 2.34125316e-01 -1.49666136e-02
1.88029930e-01 -1.35680288e-01 -5.02661429e-02 -6.59148157e-01
-9.46278125e-02 3.04472387e-01 1.88043475e-01 6.09920442e-01
-1.94698170e-01 6.95485055e-01 -8.47006083e-01 -8.75278533e-01
1.76107466e-01 6.52572989e-01 -8.51560056e-01 3.18286091e-01
-2.09661964e-02 6.36424839e-01 -3.45428646e-01 8.01791966e-01
8.88203859e-01 -2.20943347e-01 4.87336069e-02 -5.55394171e-03
-6.91020638e-02 -1.00133926e-01 -1.22584021e+00 1.69320691e+00
-2.62393415e-01 1.65153220e-01 2.33299702e-01 -7.13063717e-01
6.98091269e-01 2.70394146e-01 6.24009848e-01 -3.94508153e-01
2.87431747e-01 -4.54655476e-02 -2.07495019e-01 -4.25302356e-01
6.41619321e-03 -1.15395553e-01 8.18071216e-02 2.34805956e-01
-1.60572439e-01 -4.10661668e-01 3.16221148e-01 5.23660630e-02
5.67124963e-01 3.54476631e-01 3.91620010e-01 1.96000129e-01
2.72899866e-01 -4.50666875e-01 8.14088047e-01 3.94041419e-01
-1.22143835e-01 1.05162406e+00 4.69850332e-01 -5.88864565e-01
-8.61680806e-01 -1.36075068e+00 -1.91358049e-04 1.01913166e+00
3.58512044e-01 -2.38820896e-01 -9.68212247e-01 -4.93127406e-01
2.54050672e-01 2.51322597e-01 -8.05175900e-01 2.92777956e-01
-6.00079656e-01 -2.18828171e-01 2.75306731e-01 7.73654997e-01
4.18114871e-01 -6.36516929e-01 -9.88093555e-01 -1.48809269e-01
-3.60219061e-01 -1.21524012e+00 -7.36772835e-01 -4.15149152e-01
-7.29374826e-01 -1.22741365e+00 -8.82804990e-01 -7.07902908e-01
1.05257189e+00 3.51479687e-02 1.17847025e+00 1.86141282e-01
-3.47802609e-01 7.43116438e-01 -2.54687160e-01 -7.17017114e-01
6.00115955e-02 -2.61300415e-01 1.26850471e-01 -2.17150807e-01
2.05514804e-01 -7.63748705e-01 -8.91132712e-01 2.52135843e-01
-4.33630943e-01 5.28848708e-01 1.42281771e-01 3.73900712e-01
4.69725996e-01 -1.91329151e-01 -1.88929722e-01 -6.85481966e-01
-2.09888145e-01 -1.62684262e-01 -6.94393635e-01 2.22699583e-01
-1.50035143e-01 -1.60359535e-02 1.31916463e-01 -6.87135637e-01
-9.16991055e-01 3.36710900e-01 -2.22108439e-01 -4.23199564e-01
-2.05667675e-01 -1.25236765e-01 -2.84165472e-01 1.32707551e-01
3.42022896e-01 -9.06978101e-02 1.08296342e-01 -4.71248031e-01
1.79457903e-01 3.04041862e-01 9.06983912e-01 -7.57328749e-01
8.14234257e-01 7.33069062e-01 1.30763695e-01 -8.14918280e-01
-1.14289939e+00 -3.14220339e-01 -1.07176089e+00 -3.36884856e-01
1.06001616e+00 -1.13409925e+00 -1.14408696e+00 8.13222468e-01
-1.19241667e+00 -3.47622722e-01 3.84311378e-02 4.16301638e-01
-5.10122895e-01 2.88873911e-01 -5.21601737e-01 -9.26424384e-01
-3.41017395e-01 -8.80467832e-01 1.30673659e+00 4.15195018e-01
-3.26540679e-01 -1.07708323e+00 -1.59293607e-01 7.11923540e-01
-2.50044554e-01 5.20065844e-01 5.69486499e-01 -6.44438788e-02
-3.32666636e-01 2.21741106e-02 -4.32361998e-02 -3.57744023e-02
4.73255999e-02 1.29773423e-01 -8.20090353e-01 -3.24197382e-01
-2.10611671e-01 -2.78687119e-01 5.02349675e-01 6.59660578e-01
1.05344224e+00 -2.16372982e-01 -3.20187181e-01 9.45782483e-01
1.12576151e+00 -5.08864112e-02 2.36418545e-01 -1.92794539e-02
1.08252883e+00 9.61786747e-01 4.71049160e-01 7.27949142e-01
1.11360562e+00 6.30940676e-01 4.82032806e-01 -1.35838792e-01
-2.19833672e-01 -6.74401581e-01 9.99779329e-02 5.74839592e-01
-3.88075620e-01 2.44994760e-02 -1.01225495e+00 3.65146190e-01
-1.55779982e+00 -6.67918861e-01 -1.16906054e-01 2.33886957e+00
7.93861270e-01 -6.31673634e-02 4.59745258e-01 6.92459196e-02
5.24911821e-01 -8.52066278e-02 -7.80809700e-01 -3.51578712e-01
2.41649032e-01 -1.18572459e-01 4.38702911e-01 5.08815289e-01
-8.50247324e-01 6.88270569e-01 6.36211681e+00 6.12542778e-02
-9.42304492e-01 -1.45227119e-01 5.61407030e-01 -1.77356973e-01
-2.89777398e-01 -1.50339007e-02 -9.92601871e-01 3.03109616e-01
1.47467069e-02 8.09887797e-02 5.19676208e-01 7.18655586e-01
-1.42917395e-01 -3.84779483e-01 -1.42262137e+00 1.36742079e+00
5.19838870e-01 -4.98916715e-01 -3.56474280e-01 -9.60691571e-02
6.13527477e-01 -5.05297363e-01 -8.93903673e-02 1.57959759e-01
2.99117684e-01 -9.25789416e-01 1.13433456e+00 3.45391244e-01
9.29858983e-01 -5.99410415e-01 1.87815994e-01 8.33651960e-01
-1.25436163e+00 1.11856699e-01 -2.50237435e-01 -5.53065538e-01
2.37358939e-02 3.75482142e-01 -5.95034719e-01 1.19085602e-01
8.25636148e-01 5.44178963e-01 -5.78608513e-01 6.99651659e-01
-4.51981395e-01 -2.03710701e-02 -4.94756609e-01 3.28343093e-01
-4.39351618e-01 -2.52074361e-01 1.44850507e-01 6.98651493e-01
3.79479051e-01 5.11701286e-01 2.39303783e-01 7.98417807e-01
5.07387370e-02 -7.15394542e-02 -4.81959313e-01 2.88260490e-01
3.36816549e-01 1.04874206e+00 -7.39354253e-01 -1.97761059e-01
-5.89956522e-01 1.06243634e+00 4.47034150e-01 1.52709872e-01
-8.23795795e-01 2.52067119e-01 6.95381641e-01 4.90251094e-01
3.50285828e-01 -1.66494653e-01 -3.35991949e-01 -1.29963565e+00
1.15880482e-01 -6.48139000e-01 3.38977873e-01 -8.50679636e-01
-9.92077708e-01 7.19530284e-02 3.31523120e-01 -1.01101339e+00
-3.94143820e-01 -4.34799910e-01 -3.78451854e-01 6.55605495e-01
-1.06492054e+00 -1.43612051e+00 -5.70491433e-01 4.44104761e-01
6.54529512e-01 4.07843053e-01 4.57291722e-01 1.04398675e-01
-4.32095945e-01 7.27809191e-01 -7.01800048e-01 4.25418735e-01
5.69798768e-01 -1.38963890e+00 5.05981266e-01 6.16116583e-01
-9.21517015e-02 5.43810070e-01 7.53824651e-01 -7.54251599e-01
-1.39535367e+00 -5.26377261e-01 8.10342669e-01 -8.19969416e-01
1.82932571e-01 -5.59485734e-01 -5.98693669e-01 9.19078946e-01
-2.95958936e-01 1.06033809e-01 3.11905622e-01 2.65760779e-01
-5.28681099e-01 -4.35870923e-02 -1.28413045e+00 5.89364707e-01
1.40199852e+00 -1.22509763e-01 -6.55573845e-01 -6.07939512e-02
5.73563278e-01 -1.15434396e+00 -6.82627320e-01 5.49451351e-01
1.03014624e+00 -1.37335718e+00 1.33428228e+00 -1.59598753e-01
4.98296827e-01 -1.28426328e-01 1.18694179e-01 -1.04217339e+00
6.70066699e-02 -2.04569370e-01 -4.01452780e-01 1.34653139e+00
-2.59475708e-02 -5.00675321e-01 1.11970913e+00 1.00606501e+00
3.49951267e-01 -8.89922440e-01 -5.44876218e-01 -4.75517482e-01
2.20033526e-01 -3.10764104e-01 7.38990128e-01 6.54036522e-01
-2.92757481e-01 2.43398726e-01 -4.43440318e-01 3.98880243e-01
7.35547423e-01 1.79412469e-01 1.09588468e+00 -1.51163065e+00
-3.94256026e-01 -2.72147626e-01 -4.02794152e-01 -1.34484053e+00
3.61428000e-02 -3.44102234e-01 5.15003279e-02 -1.69830847e+00
4.20381248e-01 -1.88194931e-01 5.28640330e-01 5.63232541e-01
-3.23593974e-01 1.18083611e-01 1.07906729e-01 -1.99700788e-01
-3.30766439e-01 1.02561690e-01 1.56811523e+00 8.27696100e-02
2.09565536e-04 2.66495317e-01 -6.43493772e-01 1.36455083e+00
4.25171375e-01 -1.59367442e-01 -5.19474328e-01 -8.10207248e-01
4.58526194e-01 4.96895671e-01 4.19784874e-01 -8.68949234e-01
1.28716484e-01 -2.21605644e-01 8.76010060e-01 -1.00575840e+00
6.88420951e-01 -6.07675314e-01 9.39000100e-02 4.44164932e-01
1.19563013e-01 7.02577531e-02 -1.72404617e-01 7.86262974e-02
2.18302101e-01 -2.02612817e-01 7.73307264e-01 -3.37360054e-01
-4.74835962e-01 7.31193662e-01 2.32960343e-01 4.01013464e-01
9.25245762e-01 -4.65971857e-01 8.06959579e-04 -5.24173141e-01
-5.73906898e-01 6.15068674e-01 9.63261545e-01 3.17721367e-01
6.69043541e-01 -1.06617475e+00 -6.37711644e-01 4.75263119e-01
2.31651831e-02 6.85778975e-01 3.24000448e-01 7.86038697e-01
-7.45267034e-01 -1.82355508e-01 2.35887505e-02 -6.83008075e-01
-1.76298451e+00 6.04443192e-01 3.63236457e-01 2.27228269e-01
-6.22050941e-01 1.10927773e+00 6.70830548e-01 -6.09985054e-01
4.63047534e-01 -3.34104508e-01 -2.56281555e-01 1.76338598e-01
6.34639919e-01 4.26533401e-01 -5.09908676e-01 -8.38830948e-01
-2.47982413e-01 1.26294982e+00 1.74990788e-01 -2.48710841e-01
1.06271768e+00 -3.89768124e-01 -1.46716774e-01 5.81049323e-01
8.89994442e-01 3.42245698e-01 -1.58291614e+00 -8.68392065e-02
-6.07797086e-01 -7.25913644e-01 -4.83280927e-01 -4.29324061e-01
-1.28212237e+00 1.12431884e+00 3.41939896e-01 -1.06648028e-01
1.02217638e+00 3.03891569e-01 6.72505796e-01 -1.33892924e-01
5.18896759e-01 -9.48937416e-01 2.14292035e-01 3.94164652e-01
6.95552349e-01 -1.44617736e+00 2.43208528e-01 -7.53150702e-01
-5.27633965e-01 1.02503657e+00 1.09890342e+00 3.07225913e-01
4.26562697e-01 3.86502206e-01 1.73308641e-01 -1.83354616e-01
-3.84333342e-01 -2.19964311e-01 4.43124056e-01 7.00923026e-01
3.59098464e-01 1.84220403e-01 -1.75538197e-01 3.80860180e-01
-7.24070549e-01 -2.05976650e-01 3.80569547e-01 8.05470824e-01
-2.67779082e-01 -9.79867578e-01 -6.78918064e-01 2.15086266e-01
-4.04256672e-01 3.19438308e-01 -3.53527516e-01 6.92241609e-01
6.29245520e-01 8.78235817e-01 2.53629863e-01 -7.58684352e-02
3.42721134e-01 -2.48911202e-01 1.00483656e+00 -8.11958849e-01
5.08450624e-03 3.88110541e-02 -2.63443053e-01 -4.18864369e-01
-3.00020784e-01 -8.32553327e-01 -1.51920974e+00 -3.84521872e-01
-8.03088993e-02 -4.02030110e-01 6.51848435e-01 9.00291026e-01
-3.86860251e-01 1.34801120e-01 4.83631700e-01 -9.85296726e-01
-1.94072098e-01 -5.74102581e-01 -6.79133475e-01 5.88920176e-01
5.39229512e-01 -8.80056441e-01 -3.65682989e-01 3.13871540e-02] | [7.086113929748535, -1.1091279983520508] |
5cd38144-7e27-497a-b535-0d03fb256520 | learning-nonautonomous-systems-via-dynamic | 2306.15618 | null | https://arxiv.org/abs/2306.15618v1 | https://arxiv.org/pdf/2306.15618v1.pdf | Learning Nonautonomous Systems via Dynamic Mode Decomposition | We present a data-driven learning approach for unknown nonautonomous dynamical systems with time-dependent inputs based on dynamic mode decomposition (DMD). To circumvent the difficulty of approximating the time-dependent Koopman operators for nonautonomous systems, a modified system derived from local parameterization of the external time-dependent inputs is employed as an approximation to the original nonautonomous system. The modified system comprises a sequence of local parametric systems, which can be well approximated by a parametric surrogate model using our previously proposed framework for dimension reduction and interpolation in parameter space (DRIPS). The offline step of DRIPS relies on DMD to build a linear surrogate model, endowed with reduced-order bases (ROBs), for the observables mapped from training data. Then the offline step constructs a sequence of iterative parametric surrogate models from interpolations on suitable manifolds, where the target/test parameter points are specified by the local parameterization of the test external time-dependent inputs. We present a number of numerical examples to demonstrate the robustness of our method and compare its performance with deep neural networks in the same settings. | ['Daniel M. Tartakovsky', 'Hannah Lu'] | 2023-06-27 | null | null | null | null | ['dimensionality-reduction'] | ['methodology'] | [-2.86426932e-01 8.87930542e-02 -1.56514591e-03 1.47725970e-01
-6.57375574e-01 -4.22149807e-01 7.30221331e-01 -4.11210477e-01
-2.57974744e-01 1.04489601e+00 -1.79121718e-01 -2.14180395e-01
-5.79023778e-01 -3.74277562e-01 -9.26918089e-01 -1.15405881e+00
-2.97738701e-01 6.66056871e-01 -3.44026893e-01 -3.87238920e-01
-2.04140902e-01 7.91685939e-01 -1.01042449e+00 -5.24402320e-01
1.00845957e+00 7.62965500e-01 -2.85409063e-01 7.59648442e-01
6.13999009e-01 3.50071967e-01 -9.83314663e-02 1.50736734e-01
7.58625805e-01 -4.25430238e-01 -6.87503040e-01 1.54989153e-01
2.94188768e-01 -2.15510726e-01 -7.50687420e-01 9.78859603e-01
2.23905802e-01 5.89566112e-01 1.03444874e+00 -1.01165676e+00
-7.53712416e-01 1.63249195e-01 4.32568928e-03 -1.51652589e-01
-2.39624053e-01 1.98154494e-01 6.28887892e-01 -1.17077649e+00
7.19750166e-01 1.26068735e+00 8.68044913e-01 4.57534403e-01
-1.86636674e+00 -7.34434351e-02 -3.91290277e-01 -9.68301743e-02
-1.54371965e+00 -3.49614859e-01 1.01978886e+00 -9.35356021e-01
3.00908208e-01 2.22543776e-01 6.04600668e-01 9.22749102e-01
4.40164179e-01 9.93582606e-02 8.86144817e-01 -3.13568115e-01
3.24069500e-01 2.26236612e-01 2.90977031e-01 8.41194332e-01
-5.34339324e-02 4.89562511e-01 2.16544971e-01 -4.53848273e-01
1.29923153e+00 -1.43357679e-01 -3.33278209e-01 -9.28352952e-01
-1.35465884e+00 9.55078781e-01 3.60058606e-01 8.01451653e-02
-5.36668479e-01 -2.59677339e-02 3.16173702e-01 6.20657802e-01
4.63535756e-01 5.55616438e-01 -3.53094608e-01 1.96211666e-01
-6.50111854e-01 2.79932141e-01 8.74260545e-01 8.77705872e-01
1.01172066e+00 4.53634977e-01 -3.78308445e-02 5.05447745e-01
1.10584207e-01 3.83955002e-01 2.44426683e-01 -1.12318730e+00
2.95307487e-01 4.04777020e-01 5.95268667e-01 -6.29085720e-01
-4.71699417e-01 -4.69183564e-01 -1.20746636e+00 2.16521099e-01
7.49415755e-01 -3.78829002e-01 -6.27709806e-01 1.77584326e+00
5.81490815e-01 1.81393668e-01 8.50961581e-02 1.03773749e+00
-5.09142913e-02 9.20296252e-01 -3.37905675e-01 -4.10872757e-01
7.33031392e-01 -5.89911044e-01 -4.96176213e-01 6.69026196e-01
7.83715725e-01 -2.93194950e-01 1.10614717e+00 2.89829135e-01
-1.13064981e+00 -7.64749467e-01 -1.11498857e+00 -4.06131484e-02
-2.88444191e-01 5.48850536e-01 -1.80991665e-01 5.04555181e-03
-1.08450055e+00 1.23992038e+00 -1.01639009e+00 -1.33844167e-02
-2.45082155e-01 5.05469322e-01 -3.98307413e-01 6.96490884e-01
-1.47053111e+00 1.02356815e+00 3.40132922e-01 5.38108289e-01
-1.28683877e+00 -9.35855329e-01 -6.57530367e-01 -3.16376269e-01
5.06297983e-02 -8.17033887e-01 8.67075384e-01 -8.32459331e-01
-1.81473637e+00 4.64689940e-01 1.08430512e-01 -5.59410036e-01
7.49568999e-01 2.34860554e-02 -3.37807268e-01 2.26168588e-01
-1.49625137e-01 1.83718771e-01 1.49568462e+00 -1.10732770e+00
1.37593180e-01 3.44100073e-02 1.89022928e-01 1.47398887e-02
-1.32310390e-01 -4.28749651e-01 1.92600295e-01 -5.25097251e-01
5.30952439e-02 -1.28791165e+00 -2.52689302e-01 -7.76732191e-02
-4.44490910e-01 7.06504881e-02 9.30558681e-01 -9.31164861e-01
1.03200960e+00 -2.07284093e+00 9.95836377e-01 2.25734979e-01
3.12615216e-01 2.32456937e-01 -5.19430302e-02 7.88710594e-01
-5.35834968e-01 -8.13854393e-03 -5.44559538e-01 -2.24743679e-01
-7.72865787e-02 -2.48313937e-02 -5.44794559e-01 8.69826674e-01
3.16352665e-01 6.88633382e-01 -7.53710270e-01 -2.21730679e-01
3.71127278e-01 5.63591957e-01 -2.56099284e-01 1.62254870e-01
-2.44269609e-01 1.17755949e+00 -4.89077955e-01 1.52381673e-01
4.49941784e-01 7.15946630e-02 -3.48897368e-01 -3.49837661e-01
-3.68350059e-01 -1.08754762e-01 -1.23576832e+00 1.10105669e+00
-7.84639776e-01 4.62396383e-01 5.17599642e-01 -1.44603682e+00
9.12967801e-01 4.89671767e-01 7.37138033e-01 1.54327163e-02
1.31371737e-01 4.70563889e-01 5.63815720e-02 -5.87618887e-01
1.53245732e-01 -3.52553517e-01 -6.90555125e-02 3.61694694e-02
2.33288944e-01 -1.41838104e-01 -2.37770882e-02 -6.24049734e-03
7.12504208e-01 4.31233317e-01 2.80415803e-01 -7.20286727e-01
1.11266363e+00 1.79762080e-01 2.93469310e-01 2.09166601e-01
-1.29868239e-01 2.70711362e-01 6.27307951e-01 -7.63788044e-01
-1.68761706e+00 -1.02430594e+00 -6.25793874e-01 4.60973501e-01
-1.83963314e-01 6.04384504e-02 -9.08284783e-01 -1.47772431e-01
1.22619390e-01 4.25458521e-01 -7.80072570e-01 -4.02990699e-01
-8.85827303e-01 -6.25426590e-01 2.53434330e-01 2.40133688e-01
4.43798959e-01 -5.78326523e-01 -3.09444815e-02 3.04304332e-01
2.38039374e-01 -7.91959643e-01 -5.03005207e-01 1.04541540e-01
-1.08500528e+00 -8.53806436e-01 -8.97350967e-01 -6.61935210e-01
9.17995632e-01 -3.50245774e-01 4.43432152e-01 -4.33476090e-01
3.31022404e-02 5.00253975e-01 3.92736286e-01 2.00467825e-01
-8.47604573e-01 -9.75423381e-02 7.97779977e-01 5.23387849e-01
-4.35445547e-01 -6.95245087e-01 -4.07045275e-01 5.31962335e-01
-8.89436841e-01 1.26507487e-02 1.17844000e-01 1.29250240e+00
7.43742585e-01 -2.29661584e-01 7.68724144e-01 -4.17612314e-01
6.82479382e-01 -6.74521685e-01 -1.21090078e+00 -1.30788505e-01
-3.97980869e-01 4.58463401e-01 1.46503580e+00 -8.35808635e-01
-8.02202821e-01 3.72797310e-01 2.80868381e-01 -8.65243435e-01
1.45215020e-01 4.93993193e-01 2.51192674e-02 -4.75949109e-01
9.79801178e-01 1.06467180e-01 3.70448500e-01 -6.39518619e-01
3.47325325e-01 6.02406979e-01 6.04463756e-01 -8.81011844e-01
1.10206068e+00 4.44347441e-01 6.46418989e-01 -1.06850612e+00
-4.14051622e-01 -2.76975393e-01 -1.19257212e+00 -2.14897364e-01
5.29082894e-01 -7.62999654e-01 -6.71800017e-01 5.37321091e-01
-9.91088390e-01 -5.84377468e-01 -7.91004419e-01 6.62640393e-01
-1.07916880e+00 1.88469425e-01 -7.61912048e-01 -8.95828068e-01
-5.36429416e-03 -1.18150580e+00 7.96242476e-01 -2.21892223e-01
2.04177760e-02 -1.47438705e+00 3.82074922e-01 -2.55194694e-01
3.04547787e-01 5.89979768e-01 1.07924521e+00 -4.56581146e-01
-4.92318422e-01 -5.78480363e-01 4.34577048e-01 7.59211838e-01
-1.72377825e-01 8.66393838e-03 -6.01360679e-01 -4.37546909e-01
6.30205154e-01 9.81525108e-02 5.15806079e-01 4.36443210e-01
7.99897730e-01 -5.32555699e-01 -2.44546071e-01 9.09704447e-01
1.43411493e+00 7.43463188e-02 1.63174450e-01 9.89651680e-03
1.08815396e+00 5.18087387e-01 3.11811805e-01 1.16523027e-01
5.11268303e-02 6.67222202e-01 2.47209474e-01 1.64755791e-01
3.88242990e-01 -1.96939409e-01 5.77532530e-01 1.02760053e+00
-4.70834345e-01 5.17345846e-01 -7.51650691e-01 5.71352839e-01
-1.87832248e+00 -5.75065374e-01 -2.92818666e-01 2.59322977e+00
6.90742612e-01 -3.64870608e-01 5.52573681e-01 -1.93363819e-02
9.46817756e-01 -2.42020026e-01 -1.00508356e+00 -4.68974531e-01
4.25441600e-02 -1.19105570e-01 4.34474736e-01 7.90133476e-01
-1.11753142e+00 4.44016963e-01 5.90248775e+00 5.63754678e-01
-1.21636462e+00 8.55968222e-02 3.49359542e-01 2.33697236e-01
1.55978158e-01 2.05506518e-01 -6.79277301e-01 5.05168736e-01
1.46997094e+00 -5.44063985e-01 7.01358140e-01 8.09143066e-01
6.96199656e-01 4.20084894e-01 -1.20415556e+00 6.65602505e-01
-5.81448853e-01 -1.19148684e+00 -2.06899315e-01 3.16389829e-01
1.08632815e+00 -1.40086725e-01 1.35419786e-01 2.58681774e-01
-9.56352204e-02 -5.57904661e-01 5.46438873e-01 1.07775784e+00
7.75891542e-01 -5.52511811e-01 3.75680506e-01 6.85796916e-01
-1.00118601e+00 -1.77443370e-01 -4.23103154e-01 -1.05999112e-01
1.15121275e-01 4.83945936e-01 -6.45207345e-01 4.68646526e-01
-8.04926604e-02 8.22460830e-01 -4.33419049e-02 8.56403947e-01
2.76945144e-01 5.41442573e-01 -3.89263928e-01 9.19059813e-02
2.30398104e-01 -1.18859851e+00 1.16599143e+00 5.69469333e-01
4.31936026e-01 2.11042371e-02 -1.21321499e-01 1.22177279e+00
9.55219120e-02 4.68399115e-02 -7.38904595e-01 -1.60400897e-01
3.25013958e-02 1.36697102e+00 -1.54224513e-02 -3.17802668e-01
-4.64747427e-03 5.40307999e-01 2.68983275e-01 8.43285859e-01
-6.84907615e-01 -4.47564363e-01 5.51365018e-01 1.61541745e-01
1.61004975e-01 -8.47402632e-01 6.24249689e-02 -1.38920844e+00
2.28604823e-01 -6.36852324e-01 1.02218010e-01 -4.11455929e-01
-1.18868935e+00 4.28414911e-01 3.48692060e-01 -1.78781784e+00
-6.78164899e-01 -7.05545366e-01 -6.62831664e-01 1.20901752e+00
-8.20566595e-01 -9.79576707e-01 2.26358265e-01 5.59168875e-01
2.49889493e-02 -1.66877776e-01 7.72079289e-01 2.39911750e-02
-6.97759748e-01 1.42827287e-01 8.70866358e-01 -1.64780915e-01
3.23789805e-01 -1.41402590e+00 2.89114892e-01 8.04606140e-01
-4.73621190e-01 8.26042712e-01 7.94616997e-01 -4.95104909e-01
-1.58808756e+00 -1.26080894e+00 2.57046551e-01 -4.92041051e-01
1.15584910e+00 -3.87436122e-01 -1.20261407e+00 9.06391799e-01
-3.47518116e-01 2.70384967e-01 -1.55562669e-01 -4.72145915e-01
1.43180937e-01 -2.01712951e-01 -1.14644861e+00 8.31793666e-01
6.40520275e-01 -5.96155763e-01 -4.62736726e-01 5.14132380e-01
6.25655293e-01 -3.94959092e-01 -1.35245597e+00 2.64922976e-01
3.70419890e-01 -2.81326622e-01 1.04811418e+00 -8.86803746e-01
1.23686165e-01 -4.51728314e-01 1.60506994e-01 -1.69141936e+00
-2.87470698e-01 -1.45998657e+00 -5.15630007e-01 7.28359103e-01
1.64589986e-01 -1.00964391e+00 2.65575528e-01 6.85402095e-01
-2.86093473e-01 -9.26813424e-01 -1.10136282e+00 -1.20581055e+00
7.23502696e-01 1.76097192e-02 3.30670416e-01 8.29671860e-01
1.84931472e-01 1.39753908e-01 -5.13185382e-01 3.50732476e-01
7.13002920e-01 -1.92878768e-01 7.09636629e-01 -1.17420304e+00
-2.31113806e-01 -7.03926161e-02 -2.36878887e-01 -8.45856547e-01
4.63877499e-01 -8.74128342e-01 5.84331155e-02 -9.12452817e-01
-5.01509488e-01 -3.04626852e-01 -2.19195858e-02 -1.52343214e-01
2.98774660e-01 -1.97325602e-01 -9.08776373e-02 5.99718034e-01
1.93370864e-01 8.69300306e-01 1.52682316e+00 1.23062022e-01
-6.97419167e-01 3.16969186e-01 3.88270132e-02 7.78868616e-01
6.48015559e-01 -1.45287469e-01 -3.74405861e-01 2.41563782e-01
3.39524895e-02 6.00426316e-01 7.83993065e-01 -1.04351878e+00
1.37421444e-01 -1.26128003e-01 6.27121255e-02 -3.46042693e-01
5.02874732e-01 -6.15055859e-01 4.04735386e-01 4.42793965e-01
-4.43062425e-01 -2.24705674e-02 6.38183653e-02 6.51541948e-01
-1.48746982e-01 -2.30300039e-01 9.55569685e-01 2.06161425e-01
-6.94671646e-02 4.61832315e-01 -3.82897794e-01 -1.07632279e-01
8.41474056e-01 -6.41535670e-02 -3.15378830e-02 -3.85269463e-01
-1.33202577e+00 -3.18035134e-03 4.29021001e-01 -1.41956240e-01
2.60759145e-01 -1.58692205e+00 -3.55929554e-01 4.71662641e-01
-2.39694774e-01 -7.08897859e-02 2.95048803e-01 1.30275202e+00
-4.70839292e-01 6.32378280e-01 -2.83847570e-01 -5.41072130e-01
-3.96223783e-01 7.91896522e-01 9.38501775e-01 3.45793031e-02
-6.39078379e-01 1.07549854e-01 1.51234448e-01 -5.85409164e-01
-1.17533319e-01 -6.25695586e-01 1.13876075e-01 -1.91110253e-01
-1.75420176e-02 8.51854563e-01 -1.83055010e-02 -9.69231904e-01
1.33286268e-01 6.49306178e-01 3.99214536e-01 -4.09961611e-01
1.20304227e+00 -2.01444194e-01 -2.27044061e-01 8.32541049e-01
1.67103922e+00 -2.03860164e-01 -1.75542259e+00 -1.48608103e-01
-2.60193199e-01 1.58232331e-01 -5.50346263e-02 -1.51690975e-01
-6.91831946e-01 7.50288785e-01 3.76643062e-01 5.24610341e-01
8.67938101e-01 -3.41641247e-01 6.22637510e-01 7.07658231e-01
1.85446605e-01 -1.01959383e+00 -4.31386232e-01 6.88410938e-01
1.31203210e+00 -8.17810178e-01 -4.83147711e-01 -8.88491794e-02
-2.22716391e-01 1.46611142e+00 2.40635112e-01 -9.89630699e-01
8.48704815e-01 -1.75942078e-01 -2.39554048e-01 1.08373210e-01
-5.53582311e-01 2.27344811e-01 6.62843704e-01 4.18605149e-01
-7.46138021e-03 9.15297419e-02 -4.61584926e-01 2.17333034e-01
-8.11437070e-02 -1.70943618e-01 7.94773221e-01 2.78115988e-01
7.50493184e-02 -9.28573430e-01 -6.06193960e-01 2.23608777e-01
7.42359161e-02 3.25169951e-01 -8.21613967e-02 1.21992660e+00
-2.21095800e-01 2.64939576e-01 -1.23578824e-01 -3.15896302e-01
2.68130243e-01 3.59371603e-01 3.04512948e-01 -3.29422414e-01
-1.36087030e-01 -9.15330253e-04 -5.73731400e-02 -3.10349077e-01
-9.80167314e-02 -8.00647736e-01 -1.11914015e+00 -7.69762471e-02
-5.40615022e-02 2.26927206e-01 6.62009001e-01 1.05503774e+00
4.44266230e-01 2.04215318e-01 9.44752216e-01 -1.32306302e+00
-1.30180585e+00 -1.01746452e+00 -7.91968763e-01 3.01355094e-01
9.96975183e-01 -8.55045080e-01 -8.63501906e-01 2.50922710e-01] | [6.48972225189209, 3.447875738143921] |
d9991ec9-c0b8-4324-aa3f-efcf94e4cd59 | entred-benchmarking-relation-extraction-with | 2305.13551 | null | https://arxiv.org/abs/2305.13551v2 | https://arxiv.org/pdf/2305.13551v2.pdf | How Fragile is Relation Extraction under Entity Replacements? | Relation extraction (RE) aims to extract the relations between entity names from the textual context. In principle, textual context determines the ground-truth relation and the RE models should be able to correctly identify the relations reflected by the textual context. However, existing work has found that the RE models memorize the entity name patterns to make RE predictions while ignoring the textual context. This motivates us to raise the question: ``are RE models robust to the entity replacements?'' In this work, we operate the random and type-constrained entity replacements over the RE instances in TACRED and evaluate the state-of-the-art RE models under the entity replacements. We observe the 30\% - 50\% F1 score drops on the state-of-the-art RE models under entity replacements. These results suggest that we need more efforts to develop effective RE models robust to entity replacements. We release the source code at https://github.com/wangywUST/RobustRE. | ['Muhao Chen', 'Manjuan Duan', 'Jing Tang', 'Wenxuan Zhou', 'Yuxuan Liang', 'Yujun Cai', 'Fei Wang', 'Bryan Hooi', 'Yiwei Wang'] | 2023-05-22 | null | null | null | null | ['causal-inference', 'causal-inference', 'relation-extraction'] | ['knowledge-base', 'miscellaneous', 'natural-language-processing'] | [-1.06753506e-01 5.36714256e-01 -4.51388389e-01 -4.20071304e-01
-4.72088635e-01 -5.79060972e-01 6.09568357e-01 1.26243696e-01
-4.37042236e-01 7.18932509e-01 2.81523973e-01 -3.78998458e-01
-7.27285147e-02 -8.04821491e-01 -7.10725605e-01 -3.28146219e-02
-8.82150978e-02 4.82763052e-01 3.66011411e-01 -3.92734259e-01
7.91026279e-02 1.57902628e-01 -1.48397124e+00 3.89102668e-01
1.09703434e+00 7.66228139e-01 -4.78437990e-02 4.37835932e-01
-3.03649604e-01 1.06251085e+00 -6.18182242e-01 -9.00795758e-01
1.10502914e-01 -3.08236003e-01 -1.20399904e+00 -3.72997671e-01
3.59891146e-01 4.67776693e-02 -5.07154703e-01 9.70912397e-01
2.65098870e-01 7.09731430e-02 6.58758640e-01 -1.27851176e+00
-1.01532650e+00 1.25296175e+00 -4.06852841e-01 5.31768024e-01
5.96598268e-01 -3.20326447e-01 1.48645341e+00 -1.16605520e+00
8.98732245e-01 1.03217244e+00 8.24277341e-01 4.51984882e-01
-1.20634985e+00 -8.37423503e-01 4.55714464e-01 4.33003217e-01
-1.84091485e+00 -6.17415786e-01 3.71538132e-01 -3.37835819e-01
1.64325964e+00 6.99934959e-01 4.65025194e-02 1.13600504e+00
-2.63794269e-02 6.15960121e-01 8.33009779e-01 -6.35741174e-01
-2.88094521e-01 1.61052704e-01 4.83192056e-01 3.79956186e-01
6.76777482e-01 5.42431995e-02 -5.14198780e-01 -1.43496081e-01
3.24795187e-01 -2.31479570e-01 -2.35621274e-01 3.35045934e-01
-9.53124702e-01 3.68091822e-01 2.42336199e-01 6.20514512e-01
-2.71101147e-01 -2.00667396e-01 1.65481493e-01 2.26627156e-01
6.91434801e-01 1.05370951e+00 -1.12103868e+00 9.26726311e-02
-7.42062330e-01 4.58631903e-01 1.08003032e+00 1.52106273e+00
6.12956166e-01 -3.77860546e-01 -2.49197796e-01 9.94833708e-01
1.94878653e-01 4.43040460e-01 3.95788908e-01 -6.37181044e-01
7.28294849e-01 8.11157584e-01 1.52152613e-01 -1.08919895e+00
-4.41020817e-01 -6.02066338e-01 -5.62209547e-01 -7.62102187e-01
2.10500404e-01 -2.02462405e-01 -8.08037758e-01 1.69096541e+00
3.28834683e-01 3.02644283e-01 1.26854122e-01 3.02702487e-01
1.37807345e+00 2.63849556e-01 3.71314436e-01 -2.37355813e-01
1.50515389e+00 -8.36059690e-01 -9.94095325e-01 -6.85021162e-01
9.73394036e-01 -8.73747647e-01 8.11048090e-01 -2.46887654e-01
-7.50368297e-01 -5.22830129e-01 -8.93766999e-01 -7.10747689e-02
-5.61567128e-01 1.63787663e-01 5.85764229e-01 4.51770484e-01
-4.73660439e-01 7.80613959e-01 -6.99576378e-01 -5.02831340e-01
-4.39189859e-02 3.72630835e-01 -6.03763938e-01 5.41981822e-03
-1.77108669e+00 1.38012111e+00 6.84255004e-01 3.29255015e-01
-1.17389120e-01 -7.20441759e-01 -9.50107098e-01 6.63177371e-02
9.71393764e-01 -5.56054115e-01 1.43524635e+00 -4.01679516e-01
-7.56785989e-01 9.19918537e-01 -6.11763060e-01 -2.83499002e-01
1.53032929e-01 -6.60257936e-01 -8.40905786e-01 -4.74376231e-01
2.48195410e-01 7.82026425e-02 6.06143028e-02 -1.31554997e+00
-9.01545942e-01 -2.51678452e-02 -3.02430017e-05 -8.74024071e-03
2.70381812e-02 3.33781093e-01 -3.86750132e-01 -8.14264655e-01
2.98527062e-01 -1.14739430e+00 4.19446118e-02 -1.04443073e+00
-8.67821693e-01 -6.43275559e-01 2.96180099e-01 -7.29936481e-01
2.07148695e+00 -1.84042144e+00 -6.98361993e-02 8.44333768e-02
2.33169779e-01 2.93605864e-01 -4.74047698e-02 7.21999764e-01
-5.35860658e-01 7.83883452e-01 1.06664367e-01 -1.60930946e-01
3.77385393e-02 1.34879306e-01 -4.18952376e-01 -4.01216671e-02
2.57710755e-01 1.16486192e+00 -7.40400076e-01 -5.59737265e-01
-3.04504186e-01 1.30344570e-01 -3.30664456e-01 3.32337976e-01
-2.17912063e-01 -3.47998999e-02 -4.96211052e-01 6.89268112e-01
4.75955904e-01 -4.41140622e-01 5.66903532e-01 -3.91669989e-01
1.25051111e-01 8.67949963e-01 -1.21761930e+00 9.44812953e-01
-6.15706407e-02 2.74449557e-01 -5.49090028e-01 -4.94916588e-01
7.99690127e-01 4.88036722e-01 3.58841121e-01 -6.60621703e-01
-5.47175389e-03 3.98649871e-01 1.41115040e-01 -7.11790085e-01
7.78614402e-01 -9.45665538e-02 -1.77588627e-01 1.21219434e-01
-1.08518392e-01 7.71934167e-02 3.10540617e-01 2.77687281e-01
1.20449710e+00 2.64311671e-01 7.58554518e-01 -8.12133309e-03
4.22096372e-01 1.35953814e-01 1.00594246e+00 7.00663209e-01
2.34892935e-01 3.08156878e-01 2.06594855e-01 -3.92962188e-01
-6.89175487e-01 -6.67339027e-01 -2.39721909e-01 1.19870734e+00
1.88805819e-01 -1.01473165e+00 -4.43084270e-01 -1.08352864e+00
1.84171990e-01 1.18448400e+00 -6.77771389e-01 -1.25253901e-01
-8.33120704e-01 -8.44183028e-01 5.90244412e-01 6.87819004e-01
4.21016634e-01 -1.04411697e+00 1.02539912e-01 3.37003350e-01
-6.75407112e-01 -1.44629943e+00 -3.95509571e-01 2.40047440e-01
-5.71459413e-01 -1.20271873e+00 -9.06303972e-02 -6.44581497e-01
4.23926771e-01 3.28767346e-04 1.55466294e+00 3.27365458e-01
2.15040281e-01 -2.40886193e-02 -7.77798355e-01 -3.55166107e-01
-4.04176027e-01 6.17057920e-01 5.55741461e-03 -5.52577257e-01
1.05285096e+00 -6.01277113e-01 -2.61839002e-01 5.60536981e-01
-5.94727993e-01 -8.47703218e-02 5.37944973e-01 6.21215165e-01
5.74862123e-01 1.52963936e-01 6.07693136e-01 -1.64321518e+00
4.93882269e-01 -7.05604851e-01 -1.05440766e-01 6.77671731e-01
-1.09878957e+00 2.74880499e-01 3.33852470e-01 -5.01316071e-01
-1.04312348e+00 -1.98736548e-01 -8.64759758e-02 1.55310899e-01
-1.09929800e-01 9.22810972e-01 -2.26053298e-01 4.96669114e-01
8.53352904e-01 -2.00633943e-01 -9.75925922e-01 -7.43353367e-01
4.03323501e-01 6.95806980e-01 4.86292571e-01 -8.99782002e-01
8.84159923e-01 -1.58892125e-01 -2.94421077e-01 -3.62448752e-01
-1.39053273e+00 -5.08622706e-01 -6.74272895e-01 2.61378080e-01
5.78321636e-01 -8.77009809e-01 -2.39993513e-01 1.26970187e-01
-1.23710191e+00 -1.24115333e-01 -1.43377692e-01 2.76839674e-01
-3.19859870e-02 2.65389353e-01 -6.46464109e-01 -5.41824579e-01
-3.20134699e-01 -7.20688105e-01 6.91567004e-01 2.40940660e-01
-9.31305170e-01 -7.81082809e-01 1.45115465e-01 3.89252871e-01
1.46444589e-01 2.37933710e-01 9.66957390e-01 -1.23301172e+00
-3.32628667e-01 -2.85586417e-01 -4.73827757e-02 -1.99550483e-02
5.31363845e-01 2.06366256e-01 -1.07216525e+00 5.07545657e-02
-2.56624222e-01 1.65436313e-01 6.61883175e-01 -9.31157321e-02
7.93024480e-01 -6.24895275e-01 -8.36139262e-01 4.15112019e-01
1.05468678e+00 1.87138885e-01 8.25743735e-01 4.70833540e-01
8.37674677e-01 6.63865387e-01 8.69590163e-01 7.19863996e-02
7.65813887e-01 6.81858540e-01 -1.31433621e-01 1.87490761e-01
-1.16370939e-01 -6.31252944e-01 4.48642895e-02 8.26874375e-01
-2.04093307e-01 -4.88461822e-01 -1.27152085e+00 6.31993055e-01
-1.89057350e+00 -8.68026257e-01 -3.41270417e-01 1.84083974e+00
1.47233284e+00 2.73680836e-01 -1.62330911e-01 -4.59161494e-03
7.81093776e-01 1.49196116e-02 -2.32823670e-01 -1.23368718e-01
-1.02424316e-01 2.08384439e-01 5.08819163e-01 5.08121669e-01
-1.22990322e+00 1.29170239e+00 6.13646746e+00 7.30881691e-01
-8.40292573e-01 -1.28727749e-01 4.13267881e-01 2.32664317e-01
-4.44603086e-01 4.40926135e-01 -1.38157213e+00 3.80927712e-01
1.26185834e+00 -3.98138821e-01 6.62880316e-02 7.76745319e-01
-2.42573023e-01 -4.56234179e-02 -1.39945161e+00 4.90958333e-01
-1.08869009e-01 -1.06008625e+00 -1.09635629e-01 7.71003077e-03
7.26678669e-01 2.06940938e-02 -3.46357942e-01 6.49504304e-01
6.47720397e-01 -1.29963672e+00 5.04694700e-01 5.14304638e-01
8.58913600e-01 -4.59567457e-01 1.04239011e+00 3.61557573e-01
-1.42156863e+00 1.55775711e-01 -6.65865242e-02 6.72312975e-02
1.58171856e-03 5.59209883e-01 -1.19496095e+00 8.57562006e-01
7.21591294e-01 6.18698180e-01 -1.01025593e+00 6.29306614e-01
-5.52600503e-01 5.90945542e-01 -2.03062296e-01 1.61797896e-01
-4.55449760e-01 2.98933685e-01 4.96266961e-01 1.39101982e+00
9.85109210e-02 3.65526259e-01 -6.53771078e-03 7.06092298e-01
-2.90141582e-01 1.80122465e-01 -5.11837721e-01 -2.89323807e-01
1.02688587e+00 9.03282881e-01 -5.60817599e-01 -2.04528868e-01
-5.61676979e-01 6.64794564e-01 5.71708381e-01 5.14279485e-01
-6.85389817e-01 -5.32361686e-01 5.65988600e-01 3.98366630e-01
3.42463970e-01 -6.30165339e-02 -3.91674846e-01 -1.32463121e+00
2.25820586e-01 -9.72315490e-01 6.99716747e-01 -5.95046639e-01
-1.54548764e+00 8.40611696e-01 3.34537357e-01 -9.74824011e-01
-4.09147650e-01 -3.58544230e-01 -3.11881691e-01 7.10398316e-01
-1.42269957e+00 -1.17057717e+00 -5.15802950e-02 1.35582825e-02
2.65384734e-01 1.00026295e-01 1.00921905e+00 3.51527780e-01
-7.80880868e-01 1.03469038e+00 -3.80024105e-01 6.77253902e-01
1.00374877e+00 -1.31239808e+00 8.52432907e-01 1.07348108e+00
4.50630933e-01 1.33215332e+00 9.72201645e-01 -1.16650820e+00
-8.60413969e-01 -8.61577451e-01 2.01933169e+00 -9.69135463e-01
7.39495814e-01 -5.38510233e-02 -1.15726686e+00 1.36796677e+00
1.38773233e-01 6.33529574e-02 9.09912586e-01 7.75191009e-01
-8.92343521e-01 2.82881837e-02 -1.04770887e+00 6.56239808e-01
1.45052314e+00 -5.05425990e-01 -1.13963282e+00 2.80891005e-02
7.81616807e-01 -5.63479364e-01 -1.16434407e+00 8.89341354e-01
3.71001661e-01 -4.51918840e-01 9.04117882e-01 -9.18834329e-01
4.02320534e-01 -3.54754746e-01 -2.76881963e-01 -1.09539938e+00
-4.80289400e-01 -4.85232681e-01 -4.86947775e-01 1.75430429e+00
1.17162108e+00 -6.57249272e-01 4.13215011e-01 1.29537666e+00
1.38450190e-01 -6.18759811e-01 -7.12779045e-01 -8.86069655e-01
8.54713470e-02 -3.90279919e-01 8.50640893e-01 1.38297415e+00
2.65278816e-01 5.94839692e-01 -6.95425570e-02 4.68342811e-01
-1.69071164e-02 8.04645568e-02 5.12136757e-01 -1.16768503e+00
-2.71246791e-01 -5.55650853e-02 -1.25244483e-01 -8.74304831e-01
4.52440888e-01 -9.63586032e-01 6.34966092e-03 -1.57453191e+00
2.91637897e-01 -7.08222151e-01 -3.44024122e-01 9.16501522e-01
-7.48018265e-01 -1.33044928e-01 1.16151527e-01 3.67981315e-01
-6.20479584e-01 2.64625788e-01 9.06080484e-01 6.34253919e-02
-1.23690151e-01 7.51192272e-02 -1.21660614e+00 6.04979336e-01
7.83218443e-01 -9.16332245e-01 -3.05100143e-01 -3.41887206e-01
6.94867730e-01 -3.49793345e-01 -9.32388678e-02 -4.55044150e-01
3.41914028e-01 -5.29543698e-01 1.75067559e-01 -4.36079174e-01
-9.99262035e-02 -6.67479992e-01 3.46587151e-01 -6.67318404e-02
-3.92846018e-01 3.17755759e-01 2.57788211e-01 1.35188609e-01
-2.39214063e-01 -2.76670516e-01 1.94857359e-01 -2.12350890e-01
-7.45417893e-01 -4.97676693e-02 5.62514830e-03 6.64615512e-01
6.52032733e-01 1.91751756e-02 -7.28311419e-01 -1.09896816e-01
-7.06091702e-01 3.26102763e-01 2.69562900e-01 7.10432827e-01
3.13407898e-01 -1.25923479e+00 -8.89226258e-01 -7.48359710e-02
4.34092224e-01 2.54416317e-01 -6.64256513e-02 5.48990786e-01
-1.62392825e-01 5.27071595e-01 4.43564981e-01 -3.85227017e-02
-1.48048437e+00 4.82365817e-01 4.54396248e-01 -4.56287265e-01
-4.14958864e-01 1.19901121e+00 -6.40269592e-02 -5.97869813e-01
1.82297323e-02 -4.21345830e-01 -3.96853685e-01 2.36953539e-03
4.39386517e-01 2.50626326e-01 2.38868400e-01 -6.59156561e-01
-7.29034662e-01 3.66523355e-01 -4.86435801e-01 1.56039372e-01
1.35271478e+00 -2.04013050e-01 -2.11523086e-01 3.75181794e-01
8.50000024e-01 4.92848247e-01 -5.83233595e-01 -6.97175384e-01
9.17330325e-01 -3.86126935e-01 -3.02431554e-01 -1.09442389e+00
-8.56880367e-01 4.59827222e-02 -1.05527230e-01 3.39360595e-01
7.23833263e-01 4.05952662e-01 6.44277692e-01 4.10910994e-01
3.93320441e-01 -9.42696810e-01 -4.95449275e-01 8.86664629e-01
9.87804294e-01 -1.11605823e+00 3.77707988e-01 -1.03899276e+00
-8.42238903e-01 6.44905686e-01 9.97426093e-01 1.77615091e-01
7.98971236e-01 4.45860296e-01 1.73083723e-01 -2.81285405e-01
-1.00453615e+00 -4.42301810e-01 7.04676032e-01 3.95467311e-01
9.51504827e-01 1.22825496e-01 -4.20479059e-01 1.13193607e+00
-6.02166653e-01 -2.28034109e-01 2.69513577e-01 8.73890519e-01
-5.94903715e-02 -1.50035441e+00 4.04060222e-02 5.10065615e-01
-7.17381775e-01 -5.19637644e-01 -8.93176317e-01 8.78717959e-01
3.29144180e-01 1.13255799e+00 -3.88318777e-01 -7.99547970e-01
8.43115985e-01 5.34047365e-01 1.86244771e-01 -1.10299158e+00
-7.98251450e-01 -3.21045727e-01 8.37869644e-01 -3.82879138e-01
-2.54074901e-01 -6.79207563e-01 -1.34510720e+00 -4.11346108e-01
-7.18867004e-01 2.99443364e-01 8.82546306e-02 1.08230245e+00
5.11976779e-01 4.59143847e-01 3.20757419e-01 6.22946359e-02
-2.65339017e-01 -1.30458641e+00 -2.50153273e-01 5.50189495e-01
7.67844319e-02 -7.67291546e-01 -5.00114560e-01 -9.09260511e-02] | [9.442658424377441, 8.678647994995117] |
b4355538-667f-4080-b553-941d5c44d5fa | fice-text-conditioned-fashion-image-editing | 2301.02110 | null | https://arxiv.org/abs/2301.02110v1 | https://arxiv.org/pdf/2301.02110v1.pdf | FICE: Text-Conditioned Fashion Image Editing With Guided GAN Inversion | Fashion-image editing represents a challenging computer vision task, where the goal is to incorporate selected apparel into a given input image. Most existing techniques, known as Virtual Try-On methods, deal with this task by first selecting an example image of the desired apparel and then transferring the clothing onto the target person. Conversely, in this paper, we consider editing fashion images with text descriptions. Such an approach has several advantages over example-based virtual try-on techniques, e.g.: (i) it does not require an image of the target fashion item, and (ii) it allows the expression of a wide variety of visual concepts through the use of natural language. Existing image-editing methods that work with language inputs are heavily constrained by their requirement for training sets with rich attribute annotations or they are only able to handle simple text descriptions. We address these constraints by proposing a novel text-conditioned editing model, called FICE (Fashion Image CLIP Editing), capable of handling a wide variety of diverse text descriptions to guide the editing procedure. Specifically with FICE, we augment the common GAN inversion process by including semantic, pose-related, and image-level constraints when generating images. We leverage the capabilities of the CLIP model to enforce the semantics, due to its impressive image-text association capabilities. We furthermore propose a latent-code regularization technique that provides the means to better control the fidelity of the synthesized images. We validate FICE through rigorous experiments on a combination of VITON images and Fashion-Gen text descriptions and in comparison with several state-of-the-art text-conditioned image editing approaches. Experimental results demonstrate FICE generates highly realistic fashion images and leads to stronger editing performance than existing competing approaches. | ['Simon Dobrišek', 'Vitomir Štruc', 'Clinton Fookes', 'Martin Pernuš'] | 2023-01-05 | null | null | null | null | ['virtual-try-on'] | ['computer-vision'] | [ 7.19375432e-01 -8.19316581e-02 -1.45230204e-01 -4.36772853e-01
-5.67586958e-01 -6.99337244e-01 8.47285986e-01 -3.16739470e-01
-2.23440498e-01 5.93872845e-01 9.83874053e-02 9.25031006e-02
5.51731139e-02 -7.19512224e-01 -1.05649269e+00 -4.42080528e-01
7.11381435e-01 5.21152437e-01 -1.92586228e-01 -4.64481175e-01
-8.85622203e-02 2.11355418e-01 -1.36165833e+00 1.65055469e-01
9.49513435e-01 9.93014753e-01 4.86337394e-01 3.47443432e-01
8.58372599e-02 7.57542849e-01 -3.81127775e-01 -7.79224396e-01
4.40576375e-01 -7.49399781e-01 -3.49980235e-01 6.46866679e-01
5.80527008e-01 -2.41222814e-01 -3.42074364e-01 9.18451607e-01
1.27008691e-01 2.76544720e-01 6.27010882e-01 -1.23197985e+00
-1.17284405e+00 3.86967897e-01 -6.40291393e-01 -4.66662437e-01
4.97106165e-01 1.77021131e-01 9.08086956e-01 -9.65936124e-01
8.89971733e-01 1.06674469e+00 6.04850292e-01 5.29219389e-01
-1.54343104e+00 -5.03902674e-01 2.88680375e-01 -5.16590774e-02
-1.41596735e+00 -4.73159760e-01 1.15578914e+00 -3.56843054e-01
3.93729717e-01 4.30666178e-01 7.48886585e-01 1.36911190e+00
-5.32750003e-02 8.67938817e-01 1.27213991e+00 -6.70749903e-01
1.36915118e-01 3.06885123e-01 -3.92470837e-01 7.49919415e-01
-1.39287874e-01 -2.05993075e-02 -3.50538701e-01 2.38351077e-01
1.07196605e+00 1.26510829e-01 -2.16999620e-01 -5.55586159e-01
-1.31495893e+00 8.54025900e-01 3.50545704e-01 1.33771691e-02
-4.89749581e-01 2.52240211e-01 3.54112089e-01 2.62633741e-01
4.40498143e-01 4.49023604e-01 -6.89435154e-02 2.69141614e-01
-1.09795189e+00 3.59867215e-01 5.94566166e-01 1.38914096e+00
6.52489126e-01 2.61506557e-01 -3.47017437e-01 1.00912273e+00
8.92266259e-02 4.35118675e-01 3.82261947e-02 -9.12154078e-01
3.72681111e-01 2.57609665e-01 5.38043156e-02 -8.95782232e-01
2.42038161e-01 -3.66373152e-01 -9.49844301e-01 1.55417651e-01
1.96181357e-01 1.29950225e-01 -1.24319839e+00 1.89382875e+00
4.09175046e-02 -1.19855747e-01 -6.76816255e-02 1.07241464e+00
5.60491920e-01 6.83717847e-01 9.41735283e-02 -2.07583562e-01
1.36536765e+00 -1.18383646e+00 -8.91865671e-01 -2.96783477e-01
-1.29675522e-01 -9.29751873e-01 1.49789608e+00 2.49316871e-01
-1.16388559e+00 -7.93032587e-01 -8.65508437e-01 -2.49194190e-01
-3.04190129e-01 3.40880722e-01 6.44550264e-01 5.00910759e-01
-9.35141027e-01 9.38801914e-02 -5.06913841e-01 -4.00348485e-01
2.15956539e-01 1.33955836e-01 -2.93585420e-01 -3.57242823e-01
-9.90883052e-01 6.99876487e-01 2.91551173e-01 1.98881865e-01
-9.26394224e-01 -6.08710408e-01 -1.24057102e+00 -3.95664871e-02
7.01533258e-01 -1.00825691e+00 9.64302063e-01 -1.52443290e+00
-1.64703953e+00 9.16380405e-01 -2.26747394e-02 -1.03477754e-01
8.36466074e-01 -2.64371425e-01 -2.51174510e-01 1.08131871e-01
2.05455154e-01 8.60095561e-01 1.21528280e+00 -1.69224989e+00
-7.99952745e-02 7.64262825e-02 3.47457528e-01 2.30480313e-01
-1.68246344e-01 -9.52769592e-02 -1.21360552e+00 -1.27535093e+00
-1.42970145e-01 -1.22575414e+00 -2.94690013e-01 2.17302352e-01
-7.58153498e-01 3.45997214e-01 7.91899502e-01 -6.63024545e-01
7.70070732e-01 -2.20589662e+00 4.84755278e-01 2.96530128e-01
7.03345314e-02 8.29259902e-02 -2.86760330e-01 4.96358544e-01
1.85599342e-01 -9.47431698e-02 -4.34080243e-01 -7.55929708e-01
2.25966439e-01 4.98107374e-01 -4.09445286e-01 1.79542392e-01
2.84322768e-01 1.10079134e+00 -7.27606833e-01 -5.36309063e-01
6.22962594e-01 7.05064952e-01 -7.95084119e-01 2.31987223e-01
-5.54061174e-01 9.18864906e-01 -3.56098235e-01 4.96083468e-01
4.74168241e-01 -1.77244410e-01 2.88759917e-01 -6.00951552e-01
3.72857414e-03 -4.02465463e-01 -1.04916155e+00 1.93692410e+00
-7.83949614e-01 3.85262400e-01 3.10071949e-02 -7.96141684e-01
8.75909388e-01 5.66332154e-02 2.25426316e-01 -6.31798208e-01
1.52375787e-01 -7.14387223e-02 -4.06165034e-01 -3.21036875e-01
7.16778934e-01 -2.81948209e-01 -3.95701915e-01 2.05052331e-01
8.20752159e-02 -4.74275321e-01 2.55713552e-01 2.75643200e-01
5.53927839e-01 6.92680180e-01 1.41787350e-01 -2.60543991e-02
4.12192136e-01 -4.72354615e-04 5.93474269e-01 7.20368981e-01
2.89292216e-01 8.95206094e-01 1.38626233e-01 -1.37186497e-01
-1.41106367e+00 -1.11958849e+00 4.64368276e-02 8.86457205e-01
3.33628327e-01 -3.27706099e-01 -6.96592391e-01 -4.84960198e-01
-1.49914935e-01 9.87422466e-01 -6.99390054e-01 1.75120626e-02
-5.69993079e-01 -2.59635806e-01 2.91199535e-01 7.15096474e-01
6.70270264e-01 -8.87367964e-01 -2.09285721e-01 1.27062529e-01
-3.02621484e-01 -1.50380385e+00 -1.04757190e+00 -1.94995061e-01
-3.44826460e-01 -7.02423871e-01 -8.53131652e-01 -1.09229660e+00
1.08418643e+00 1.83069333e-02 1.01996338e+00 -3.42369340e-02
-3.17385554e-01 7.57516801e-01 -4.41517502e-01 -1.41514614e-01
-6.17752016e-01 -2.64282048e-01 -2.84181554e-02 4.29630607e-01
-1.42222688e-01 -4.95028585e-01 -5.04497170e-01 3.66578192e-01
-1.16842473e+00 5.94625354e-01 7.26138890e-01 1.14617777e+00
9.41251099e-01 4.84363548e-02 4.82011437e-01 -1.09447181e+00
4.19316530e-01 -4.58436199e-02 -4.23042923e-01 5.01138926e-01
-4.07491416e-01 -4.30377945e-02 9.98550832e-01 -7.15241373e-01
-1.26564538e+00 3.02790940e-01 -7.89852440e-02 -8.44105482e-01
-7.71938935e-02 2.84054726e-01 -2.98200756e-01 -4.99358065e-02
2.47633532e-01 5.02312720e-01 -2.19561327e-02 -4.01910752e-01
8.57615352e-01 2.75237113e-01 8.27302873e-01 -7.30982184e-01
9.85944510e-01 4.38446015e-01 -1.49582088e-01 -8.93277407e-01
-7.31939733e-01 -7.63528347e-02 -6.52709782e-01 -3.46920699e-01
1.04669738e+00 -9.72568393e-01 -4.37019497e-01 4.04017210e-01
-9.23894346e-01 -3.69099557e-01 -3.83640289e-01 3.45009774e-01
-8.88247430e-01 3.78861487e-01 -6.84382141e-01 -5.37529409e-01
-1.06418006e-01 -1.33308947e+00 1.14301789e+00 -1.23900644e-01
-2.78956503e-01 -9.83048558e-01 -4.10455704e-01 6.68828547e-01
4.19282734e-01 5.90260565e-01 9.83032048e-01 4.57079411e-02
-6.55930161e-01 -7.64940225e-04 -2.14334384e-01 5.04021883e-01
2.16859981e-01 -1.61045447e-01 -7.24600196e-01 -2.36237392e-01
-2.29811758e-01 -3.73903960e-01 7.44500220e-01 3.88872296e-01
1.18457413e+00 -2.96365231e-01 -8.04130137e-02 7.57076561e-01
1.63798904e+00 5.44760041e-02 6.36391699e-01 9.65584069e-02
1.02616334e+00 5.31461537e-01 6.92155957e-01 2.83440292e-01
3.25832903e-01 1.07794607e+00 2.86691397e-01 -5.44952214e-01
-2.94465840e-01 -6.58463001e-01 2.04027638e-01 8.87705863e-01
-2.64277160e-01 -3.52965921e-01 -3.66347164e-01 4.62935299e-01
-1.90064812e+00 -7.82314360e-01 1.09632380e-01 1.92223418e+00
8.73538136e-01 -8.61469731e-02 6.98035304e-03 -1.31239697e-01
7.10691273e-01 1.48224920e-01 -6.16573870e-01 -2.97881246e-01
-1.07861474e-01 1.26675785e-01 3.93975109e-01 2.75365591e-01
-1.03828859e+00 1.01326668e+00 6.01804018e+00 8.19762766e-01
-9.82685566e-01 2.29230762e-01 5.40659249e-01 1.66454241e-02
-3.93763125e-01 -6.68297485e-02 -3.97223443e-01 4.21742648e-01
1.72594100e-01 4.79945652e-02 7.80701041e-01 6.84768021e-01
2.46610433e-01 1.61321461e-01 -1.22068357e+00 1.10317707e+00
4.56783950e-01 -1.25793540e+00 3.77916902e-01 -1.52802631e-01
8.56798947e-01 -7.73669481e-01 3.78138036e-01 2.96915531e-01
2.03133270e-01 -8.87867510e-01 1.30981195e+00 5.85720122e-01
1.40469909e+00 -6.58116937e-01 3.37056845e-01 7.41116377e-03
-1.24470031e+00 8.96110237e-02 -1.31267518e-01 2.07361892e-01
4.66053158e-01 3.26581299e-01 -2.79917747e-01 7.88856030e-01
3.54629904e-01 6.62119687e-01 -3.22570354e-01 4.70736533e-01
-3.86540651e-01 3.78792197e-01 -1.31632775e-01 3.79650414e-01
2.34745562e-01 -5.38149059e-01 4.65966731e-01 1.10058403e+00
2.09721893e-01 -5.95170259e-02 5.75495720e-01 1.31452930e+00
-1.82799742e-01 1.55206174e-01 -6.16742074e-01 4.99959625e-02
2.58105308e-01 1.11964452e+00 -7.18549848e-01 -2.93444842e-01
-5.41517556e-01 1.62747908e+00 7.14462698e-02 5.53912044e-01
-1.11255836e+00 -1.41368940e-01 3.00268799e-01 1.17779352e-01
5.05851984e-01 -2.94557512e-01 -1.39233574e-01 -1.13797116e+00
1.42546594e-01 -9.81995940e-01 -6.41703904e-02 -1.01714921e+00
-1.35066664e+00 5.96637487e-01 8.68374854e-02 -1.27391219e+00
-8.58163387e-02 -3.93522382e-01 -4.89621788e-01 6.68653846e-01
-1.34531021e+00 -1.98413086e+00 -3.61140460e-01 7.71295011e-01
8.97244036e-01 -4.80000600e-02 6.61163390e-01 4.66853678e-01
-4.07912284e-01 6.79942310e-01 2.33926363e-02 1.07654139e-01
6.57281876e-01 -1.07632899e+00 3.15583080e-01 7.93190897e-01
1.13325968e-01 7.04553366e-01 7.48101234e-01 -6.30876184e-01
-1.69893539e+00 -1.41464376e+00 3.60642433e-01 -3.62233013e-01
3.10989469e-01 -6.96239114e-01 -4.28641677e-01 1.04281902e+00
2.23936632e-01 -1.46089926e-01 5.23076296e-01 -1.06115051e-01
-1.76123172e-01 -7.08050802e-02 -1.00036609e+00 7.53451228e-01
1.18208706e+00 -5.87131083e-01 -4.58025634e-01 3.32625449e-01
6.50015950e-01 -4.98708338e-01 -8.10067236e-01 2.41873592e-01
5.65783322e-01 -6.23500168e-01 1.08948529e+00 -2.72460550e-01
6.40700340e-01 -4.52708572e-01 -3.48431468e-01 -1.32428670e+00
-3.78284991e-01 -5.86338341e-01 1.93547979e-01 1.54508686e+00
2.18638480e-01 -2.63446212e-01 5.21542311e-01 5.84336460e-01
-1.61173075e-01 -5.87481678e-01 -4.65150237e-01 -7.74798930e-01
-3.18982631e-01 -3.53175879e-01 5.07546365e-01 1.15966761e+00
-5.01502693e-01 2.98911542e-01 -1.07673633e+00 -6.57153968e-03
7.50285447e-01 2.24888206e-01 9.77924466e-01 -6.50566280e-01
-5.91699183e-01 -7.12322965e-02 -3.47368926e-01 -1.10527313e+00
1.99933231e-01 -8.89962912e-01 1.92766681e-01 -1.55426347e+00
2.13134199e-01 -5.64521372e-01 -6.83798939e-02 4.84127790e-01
-9.10539031e-02 7.35814273e-01 5.32287002e-01 2.34803960e-01
-5.45478642e-01 7.34905541e-01 1.51667297e+00 -2.92936772e-01
-1.52288556e-01 -2.89041311e-01 -7.47111976e-01 6.60980761e-01
4.22600120e-01 -1.26734644e-01 -6.12726986e-01 -4.63958442e-01
-1.06145795e-02 4.78899181e-02 7.39409626e-01 -8.32460940e-01
2.03608815e-03 -2.20036313e-01 4.81053144e-01 -2.67176747e-01
6.34353817e-01 -9.56501126e-01 6.63236499e-01 1.47343218e-01
-4.59836096e-01 -2.76749507e-02 5.38417473e-02 7.57892370e-01
-2.39774823e-01 5.84904179e-02 7.96604812e-01 -1.82164922e-01
-9.69792187e-01 2.58910179e-01 -2.62039155e-01 -1.61013559e-01
1.21122050e+00 -3.12768370e-01 2.71929000e-02 -6.44789875e-01
-7.51350582e-01 4.10294682e-02 8.89456987e-01 5.88720739e-01
6.64431810e-01 -1.49371970e+00 -6.50501549e-01 4.21919972e-01
2.92546511e-01 -1.96401238e-01 3.18201452e-01 7.87547767e-01
-3.85772526e-01 7.15206191e-02 -2.45412648e-01 -6.05084777e-01
-1.22057366e+00 8.61443698e-01 -5.07777557e-02 -1.57072440e-01
-9.37047780e-01 4.65330899e-01 7.66169906e-01 -4.55169111e-01
1.50708901e-02 -1.70949757e-01 1.20631546e-01 -3.21750343e-01
1.49617940e-01 -1.91642389e-01 -2.74509668e-01 -8.74230444e-01
-2.23751236e-02 8.17152917e-01 -1.80509478e-01 -2.06192151e-01
1.24566853e+00 -2.37026140e-01 6.75465465e-02 2.30001137e-01
8.73073101e-01 2.16015667e-01 -1.41217148e+00 -3.91210824e-01
-4.79377717e-01 -6.74824834e-01 2.90331077e-02 -9.12872076e-01
-1.19705546e+00 4.24274325e-01 4.03393358e-01 -2.57826388e-01
1.38458252e+00 -1.04332842e-01 8.47253978e-01 5.44908049e-04
4.72974896e-01 -1.04717016e+00 2.09864289e-01 8.34892914e-02
1.17355871e+00 -1.08482325e+00 5.70497327e-02 -7.86944568e-01
-1.04571903e+00 8.91408265e-01 4.70796794e-01 -1.19800687e-01
2.21168235e-01 1.33388206e-01 7.40728006e-02 -1.87321395e-01
-3.72843713e-01 -1.61669508e-01 3.80783737e-01 6.04546905e-01
1.58312470e-01 1.32495850e-01 -2.64200866e-01 4.44145352e-01
-1.10449776e-01 1.12894043e-01 3.27798992e-01 9.14013326e-01
2.82936364e-01 -1.23360527e+00 -4.19786185e-01 3.16673815e-01
-2.45845035e-01 -1.22809440e-01 -2.57981986e-01 8.61083031e-01
2.72683203e-01 8.18882763e-01 -6.01006672e-02 -2.24063441e-01
4.89167452e-01 -1.06216244e-01 7.55845308e-01 -5.64428687e-01
-4.79275763e-01 2.35930473e-01 1.20453693e-01 -3.34256619e-01
-6.39677942e-01 -4.71448243e-01 -8.00221443e-01 -2.31613263e-01
-3.79771024e-01 -9.08314139e-02 6.47622943e-01 8.38979065e-01
2.61895478e-01 7.30005980e-01 5.27514458e-01 -1.08991969e+00
-1.65333897e-01 -5.68049610e-01 -6.82125032e-01 9.28796411e-01
1.04800247e-01 -8.02275717e-01 4.50503379e-02 6.23511672e-01] | [11.664528846740723, -0.5637024641036987] |
a16a647b-b116-4b36-991b-57b81365f14f | fine-grained-post-training-for-improving | null | null | https://aclanthology.org/2021.naacl-main.122 | https://aclanthology.org/2021.naacl-main.122.pdf | Fine-grained Post-training for Improving Retrieval-based Dialogue Systems | Retrieval-based dialogue systems display an outstanding performance when pre-trained language models are used, which includes bidirectional encoder representations from transformers (BERT). During the multi-turn response selection, BERT focuses on training the relationship between the context with multiple utterances and the response. However, this method of training is insufficient when considering the relations between each utterance in the context. This leads to a problem of not completely understanding the context flow that is required to select a response. To address this issue, we propose a new fine-grained post-training method that reflects the characteristics of the multi-turn dialogue. Specifically, the model learns the utterance level interactions by training every short context-response pair in a dialogue session. Furthermore, by using a new training objective, the utterance relevance classification, the model understands the semantic relevance and coherence between the dialogue utterances. Experimental results show that our model achieves new state-of-the-art with significant margins on three benchmark datasets. This suggests that the fine-grained post-training method is highly effective for the response selection task. | ['Jungyun Seo', 'Youngjoong Ko', 'Byoungjae Kim', 'Taesuk Hong', 'Janghoon Han'] | 2021-05-24 | null | null | null | naacl-2021-4 | ['conversational-response-selection'] | ['natural-language-processing'] | [ 3.59499335e-01 1.69205844e-01 -1.69081792e-01 -8.75183940e-01
-9.44661438e-01 -3.05079222e-01 8.37863803e-01 1.83119386e-01
-2.88086563e-01 6.32772446e-01 7.06607282e-01 -2.39975184e-01
-2.28443053e-02 -6.60653770e-01 -2.55328119e-01 -3.93575549e-01
3.36784095e-01 6.06626451e-01 2.07429528e-01 -9.23685312e-01
5.36655962e-01 -1.44842893e-01 -1.13498318e+00 9.29266751e-01
8.73649657e-01 1.03302741e+00 5.80472469e-01 7.42099226e-01
-4.94642377e-01 1.14686191e+00 -7.56204844e-01 -3.37902844e-01
-3.93925130e-01 -9.67385888e-01 -1.57597065e+00 -1.29754180e-02
-5.14613092e-02 -6.52887821e-01 -1.11134380e-01 5.93592465e-01
3.66314501e-01 5.12318730e-01 4.99461919e-01 -7.12865353e-01
-4.68728691e-01 9.84233320e-01 -2.87433006e-02 7.52924979e-02
8.92695665e-01 -1.69987902e-01 1.51759017e+00 -1.05903459e+00
4.13483679e-01 1.65925157e+00 1.07066855e-01 7.86465228e-01
-8.08239400e-01 -2.62715459e-01 4.62462246e-01 5.16181111e-01
-8.77311707e-01 -4.97417241e-01 7.49567807e-01 -1.56461999e-01
1.14340615e+00 5.39054632e-01 2.81108409e-01 1.04025352e+00
1.42384529e-01 1.02437103e+00 8.97997677e-01 -6.50623798e-01
-4.41290550e-02 2.75459558e-01 5.42593062e-01 2.26698011e-01
-9.01546478e-01 -2.68778563e-01 -7.71096408e-01 -2.44255915e-01
3.09475005e-01 -2.91314244e-01 -3.59286994e-01 2.53412753e-01
-8.79476011e-01 1.04478204e+00 3.15478355e-01 3.95282626e-01
-2.08548322e-01 -3.29821140e-01 7.36944258e-01 8.17395389e-01
5.41146576e-01 6.62883222e-01 -4.81091172e-01 -3.13063800e-01
-4.46562439e-01 1.43643752e-01 1.15624344e+00 9.35992658e-01
7.73513675e-01 -4.14337158e-01 -6.19144022e-01 1.19186187e+00
3.76971930e-01 1.66257843e-02 4.03468549e-01 -7.94663131e-01
6.09690249e-01 8.32795978e-01 -3.89130414e-02 -8.77809763e-01
-3.45442027e-01 -5.86908385e-02 -7.47354269e-01 -5.71868777e-01
2.12685600e-01 -5.97634837e-02 -3.23407710e-01 1.66308618e+00
4.20509905e-01 -3.06577414e-01 3.25137585e-01 1.03030312e+00
1.14142239e+00 8.57966602e-01 1.33876771e-01 -5.51652074e-01
1.43828881e+00 -1.31963217e+00 -1.11027896e+00 -2.48063937e-01
6.60319984e-01 -8.55414450e-01 1.39231491e+00 7.08681494e-02
-1.00759077e+00 -6.45100892e-01 -7.74184525e-01 -2.98520535e-01
-5.81734814e-02 8.88030604e-02 5.20751595e-01 4.72934693e-02
-7.74474680e-01 2.60119677e-01 -1.81785777e-01 -3.20234358e-01
-4.98783052e-01 2.53305823e-01 -2.79919710e-02 -4.22739722e-02
-1.86723542e+00 1.01577592e+00 6.43732375e-04 3.05996716e-01
-7.81026185e-01 -2.32057571e-01 -6.85467482e-01 2.79344767e-01
4.82243091e-01 -5.43307841e-01 1.69474173e+00 -1.00963700e+00
-2.15533948e+00 5.53650260e-01 -3.91351849e-01 -1.42520353e-01
2.75224537e-01 -3.12534273e-01 -2.41961479e-01 1.72675416e-01
-1.89198721e-02 2.67219007e-01 6.71512604e-01 -1.05193615e+00
-8.20947587e-01 -2.37134621e-01 7.31544137e-01 7.62255907e-01
-2.35568255e-01 3.63042057e-01 -5.49671829e-01 -1.30529255e-01
-4.46525142e-02 -7.76432931e-01 -5.46247438e-02 -7.59921610e-01
-3.96425933e-01 -9.26877439e-01 6.34974658e-01 -4.71664608e-01
1.52648652e+00 -2.02192807e+00 2.94290900e-01 -1.08387098e-01
-3.51534411e-02 2.40146834e-02 -1.84025869e-01 1.04262710e+00
3.36023927e-01 1.02179579e-01 7.46615902e-02 -3.97412062e-01
6.36610910e-02 6.82980381e-03 -5.82368493e-01 -6.76443949e-02
2.12400272e-01 6.99473619e-01 -1.04561245e+00 -3.63126546e-01
4.19842787e-02 1.42626733e-01 -5.76103330e-01 1.13505650e+00
-3.31153601e-01 7.95798242e-01 -8.24147940e-01 2.16440663e-01
2.09266454e-01 -3.61828804e-01 5.97292721e-01 -2.47330323e-01
1.51443973e-01 1.05397260e+00 -6.04655027e-01 1.72937369e+00
-8.07160974e-01 4.56266969e-01 1.06593035e-01 -9.16740835e-01
1.00521219e+00 5.64388037e-01 1.38960332e-01 -9.85169291e-01
4.47818339e-02 1.96734115e-01 8.57496336e-02 -7.57473707e-01
9.12965715e-01 -3.74498777e-02 -3.93800467e-01 7.39954233e-01
6.76733926e-02 -2.25016668e-01 4.46555428e-02 4.21924323e-01
7.78146207e-01 -3.50212492e-02 3.26750755e-01 -9.35797989e-02
9.41958368e-01 -2.16384992e-01 3.63865733e-01 7.14793146e-01
-3.62850726e-02 3.24392170e-01 6.06461108e-01 -2.75722712e-01
-4.82717097e-01 -2.75677413e-01 3.48974764e-02 1.85428059e+00
4.50147897e-01 -6.52581453e-01 -7.22410679e-01 -8.31506431e-01
-3.88533831e-01 8.25756907e-01 -6.08242929e-01 -3.35643053e-01
-8.11065197e-01 -4.39894050e-01 2.13173330e-01 2.78954774e-01
4.85704154e-01 -1.16919696e+00 -3.48329544e-01 3.49918365e-01
-8.26546252e-01 -1.21974659e+00 -7.09276676e-01 1.49943277e-01
-6.29596829e-01 -1.01779711e+00 -2.18969434e-01 -7.78342843e-01
4.40261781e-01 3.07812244e-01 1.26534712e+00 3.57860059e-01
4.81349349e-01 3.62857252e-01 -8.94041955e-01 9.63841751e-02
-6.78160608e-01 3.33146542e-01 -2.49225602e-01 1.36048377e-01
4.21568185e-01 -8.00875798e-02 -5.49983144e-01 6.95016384e-01
-6.37328327e-01 2.54792094e-01 2.62586236e-01 1.30274355e+00
2.01476738e-01 -2.81765491e-01 1.10566580e+00 -9.99688864e-01
1.36694503e+00 -4.36579198e-01 1.19701080e-01 8.31276357e-01
-5.34213960e-01 4.44409102e-02 6.88194692e-01 -2.49722347e-01
-1.48492372e+00 -3.95101130e-01 -3.96057963e-01 3.77504230e-01
6.57281578e-02 7.05015242e-01 -1.36698127e-01 4.01282638e-01
4.23536181e-01 1.32426098e-01 -1.41713455e-01 -4.20475841e-01
3.63202065e-01 1.12161446e+00 5.84973544e-02 -8.09168398e-01
-6.54926822e-02 -1.36266336e-01 -6.03463650e-01 -6.67148352e-01
-8.97293091e-01 -6.59625292e-01 -4.91229445e-01 -3.26935709e-01
7.58445382e-01 -7.24490821e-01 -8.55055749e-01 1.55600443e-01
-1.37878418e+00 -3.99488181e-01 9.86651480e-02 3.83455276e-01
-5.90139985e-01 3.05977196e-01 -9.12464023e-01 -9.90190387e-01
-5.65881014e-01 -1.41109896e+00 9.23692226e-01 2.58339643e-01
-4.64787811e-01 -9.91312802e-01 -1.63476199e-01 5.52031398e-01
4.23185498e-01 -4.89655495e-01 1.12480783e+00 -9.60842252e-01
-3.88238311e-01 1.04286829e-02 -7.58860633e-02 2.46310338e-01
3.95051748e-01 -1.37093380e-01 -1.02446282e+00 -1.92073226e-01
4.02372003e-01 -9.59136128e-01 5.90896130e-01 3.45467078e-03
1.02938902e+00 -4.26978230e-01 -1.20809644e-01 -4.48487513e-02
7.91276157e-01 4.01686430e-01 3.97796899e-01 1.10156760e-01
3.72886896e-01 1.20215154e+00 1.14850628e+00 4.19138104e-01
8.86507690e-01 9.71396267e-01 2.68619657e-01 6.13436140e-02
1.28226683e-01 -3.63582343e-01 3.05396765e-01 1.18629241e+00
3.22280437e-01 -4.38753158e-01 -5.90124249e-01 3.41675639e-01
-2.14380312e+00 -6.95708156e-01 -5.52254431e-02 1.99871767e+00
1.45679247e+00 -4.64412980e-02 -1.28809035e-01 -2.49444708e-01
5.86798251e-01 2.29044065e-01 -4.06671971e-01 -7.91295767e-01
1.25044763e-01 -1.10470474e-01 -3.18127960e-01 8.96242619e-01
-7.49027252e-01 1.09599972e+00 6.18048334e+00 7.26127505e-01
-1.04587984e+00 -2.52365787e-02 7.52246141e-01 4.05317366e-01
-5.24732113e-01 8.32097605e-02 -9.65995848e-01 1.85417563e-01
1.02661049e+00 -2.10964322e-01 3.76218200e-01 4.96545762e-01
3.22192997e-01 -2.31367752e-01 -1.46877468e+00 4.54227179e-01
1.12792604e-01 -1.09020686e+00 9.12551433e-02 -3.97201896e-01
3.00509900e-01 -4.68702435e-01 -2.12967217e-01 7.49075592e-01
1.65932089e-01 -9.23914075e-01 2.94468552e-01 3.91219109e-01
7.11160898e-01 -5.60980201e-01 8.34961951e-01 6.17150605e-01
-1.02142560e+00 -1.22067153e-01 -3.36215168e-01 -8.26462805e-02
1.64320976e-01 1.93954125e-01 -1.38563514e+00 6.72142804e-01
3.32394719e-01 4.41631258e-01 -2.65950441e-01 2.30246544e-01
-4.05443341e-01 5.48334420e-01 5.20517193e-02 -3.64532292e-01
2.41475090e-01 -2.67686583e-02 2.91220784e-01 1.36732054e+00
-1.36875927e-01 3.98498595e-01 3.77448261e-01 5.67371666e-01
-9.15533677e-02 3.91848236e-01 -3.23733687e-01 -8.81278946e-04
6.67197466e-01 1.13223720e+00 -1.12958536e-01 -3.77936751e-01
-3.70085955e-01 9.12618279e-01 4.46981043e-01 3.89496237e-01
-4.52037305e-01 -2.39864364e-01 4.81540650e-01 -3.06120098e-01
-2.07547262e-01 8.61582756e-02 -4.52732854e-02 -9.52713847e-01
-9.91101339e-02 -1.21944058e+00 3.66862059e-01 -4.37248617e-01
-1.19989693e+00 9.13033068e-01 -1.20349778e-02 -9.97912228e-01
-8.03267241e-01 -2.07423121e-01 -5.87052405e-01 9.76504982e-01
-1.65633297e+00 -1.16017389e+00 -7.40960613e-02 4.84196395e-01
1.23057032e+00 -9.03983712e-02 1.25258696e+00 9.07185003e-02
-4.71588552e-01 7.11728036e-01 -3.25239509e-01 5.65273799e-02
1.08519089e+00 -1.26615465e+00 8.10860544e-02 3.09099585e-01
-3.40915114e-01 8.39665174e-01 6.01425588e-01 -3.69540066e-01
-1.43156636e+00 -5.41483045e-01 1.43720841e+00 -1.83881968e-01
4.59787875e-01 -2.80602574e-01 -1.08023453e+00 4.23139900e-01
6.94627285e-01 -5.44754446e-01 1.04180133e+00 5.34320533e-01
-1.16739206e-01 -1.07956462e-01 -8.02300930e-01 4.75820303e-01
6.41761482e-01 -1.04804254e+00 -9.66750443e-01 3.73932809e-01
1.11888707e+00 -6.42552257e-01 -8.27788591e-01 4.52118993e-01
5.49840212e-01 -8.01064789e-01 6.99581563e-01 -7.28454530e-01
7.50542879e-01 1.47466347e-01 -3.30613375e-01 -1.50612891e+00
-1.22364432e-01 -7.01647937e-01 -1.06707767e-01 1.30966043e+00
4.44591582e-01 -3.06464314e-01 1.22136876e-01 9.24032152e-01
-2.88639158e-01 -9.85163152e-01 -6.60500288e-01 -8.23975578e-02
1.89557243e-02 -5.03761098e-02 7.46978879e-01 9.38103914e-01
6.50358140e-01 1.07298446e+00 -5.72764158e-01 -2.55882025e-01
-1.25147447e-01 4.65915918e-01 7.99531043e-01 -8.79382551e-01
-2.78949022e-01 -3.86275679e-01 3.75584841e-01 -1.86783648e+00
4.25175428e-01 -5.34381628e-01 3.97873640e-01 -1.52714074e+00
7.29622245e-02 -4.80636001e-01 -2.26672381e-01 1.17540307e-01
-7.14483738e-01 -5.31763613e-01 1.08652391e-01 3.10060829e-01
-8.70466590e-01 8.78286362e-01 1.46615732e+00 -2.71400481e-01
-3.65482211e-01 3.06601524e-01 -7.68208981e-01 3.01099062e-01
6.65188670e-01 -1.73153058e-01 -6.81830347e-01 -5.28796256e-01
1.13218226e-01 6.86141908e-01 -2.45679870e-01 -1.50491148e-01
5.34951925e-01 -4.42873210e-01 -2.38511503e-01 -5.93316317e-01
4.96513695e-01 -5.75942576e-01 -4.75404114e-01 6.98362535e-04
-1.31260073e+00 -8.19754079e-02 -1.96475387e-01 4.34018910e-01
-5.91979086e-01 -3.77961487e-01 4.65995640e-01 -1.73770592e-01
-7.15597808e-01 -2.69889403e-02 -3.66566837e-01 1.07096486e-01
4.15441245e-01 1.96201056e-01 -2.89767295e-01 -8.11608791e-01
-5.75886011e-01 6.85676575e-01 -4.90019470e-02 6.87866211e-01
8.32026064e-01 -1.16253769e+00 -8.13484192e-01 1.47680864e-01
2.95055419e-01 7.97140673e-02 3.23130667e-01 6.84216022e-01
2.04388514e-01 5.84235847e-01 1.17543474e-01 -6.50749207e-01
-1.47376812e+00 2.78841946e-02 4.08984721e-01 -6.40495837e-01
-3.42137516e-01 1.16497624e+00 5.50745249e-01 -6.65101647e-01
4.24328923e-01 -2.11320624e-01 -8.77151608e-01 2.47975707e-01
7.41470218e-01 -7.93709084e-02 7.90205076e-02 -5.70083857e-01
-1.36972040e-01 1.55693546e-01 -4.14111406e-01 -1.35870740e-01
9.95892584e-01 -8.05871427e-01 -3.85888845e-01 7.46013999e-01
1.26452124e+00 -2.30897337e-01 -9.91368055e-01 -7.23895133e-01
1.86626360e-01 -4.19130117e-01 -1.65192053e-01 -8.52410376e-01
-5.66011667e-01 9.13672268e-01 5.74190654e-02 5.98680794e-01
1.09216332e+00 -7.04512298e-02 7.53826976e-01 7.89789617e-01
2.73748547e-01 -1.45890427e+00 4.58352357e-01 1.25145018e+00
1.36137605e+00 -1.37688911e+00 -1.25699237e-01 -4.44769979e-01
-1.05430245e+00 1.34821069e+00 9.59426641e-01 3.31315637e-01
4.52268094e-01 -1.89843148e-01 2.56744891e-01 -2.02456832e-01
-1.33840156e+00 -2.94734631e-02 2.95401752e-01 9.97719690e-02
1.04444611e+00 9.01545584e-02 -6.02982819e-01 5.93798518e-01
-2.05416605e-01 -2.90455997e-01 3.79806668e-01 7.51920044e-01
-4.29041117e-01 -1.45746613e+00 9.93503481e-02 1.92179427e-01
-3.42964143e-01 -1.48720756e-01 -8.21959615e-01 2.34885097e-01
-6.31467581e-01 1.58783627e+00 -2.18723223e-01 -6.90345645e-01
4.36702281e-01 2.33393535e-01 8.01673755e-02 -9.38253164e-01
-1.12247062e+00 1.50676891e-01 6.68770552e-01 -5.29527783e-01
-4.83598560e-01 -1.21931486e-01 -1.32284069e+00 -1.67515263e-01
-7.57351637e-01 6.51814222e-01 4.95731711e-01 1.26716387e+00
1.93709105e-01 6.83786809e-01 1.41057158e+00 -4.73010391e-01
-1.05188191e+00 -1.38534009e+00 -1.70671806e-01 5.05139172e-01
4.45092082e-01 -4.51595753e-01 -2.62756735e-01 -2.43704885e-01] | [12.571048736572266, 7.830224514007568] |
48d87f47-b17d-46e1-8bfe-4e19810d458b | dipnet-efficiency-distillation-and-iterative | 2304.07018 | null | https://arxiv.org/abs/2304.07018v1 | https://arxiv.org/pdf/2304.07018v1.pdf | DIPNet: Efficiency Distillation and Iterative Pruning for Image Super-Resolution | Efficient deep learning-based approaches have achieved remarkable performance in single image super-resolution. However, recent studies on efficient super-resolution have mainly focused on reducing the number of parameters and floating-point operations through various network designs. Although these methods can decrease the number of parameters and floating-point operations, they may not necessarily reduce actual running time. To address this issue, we propose a novel multi-stage lightweight network boosting method, which can enable lightweight networks to achieve outstanding performance. Specifically, we leverage enhanced high-resolution output as additional supervision to improve the learning ability of lightweight student networks. Upon convergence of the student network, we further simplify our network structure to a more lightweight level using reparameterization techniques and iterative network pruning. Meanwhile, we adopt an effective lightweight network training strategy that combines multi-anchor distillation and progressive learning, enabling the lightweight network to achieve outstanding performance. Ultimately, our proposed method achieves the fastest inference time among all participants in the NTIRE 2023 efficient super-resolution challenge while maintaining competitive super-resolution performance. Additionally, extensive experiments are conducted to demonstrate the effectiveness of the proposed components. The results show that our approach achieves comparable performance in representative dataset DIV2K, both qualitatively and quantitatively, with faster inference and fewer number of network parameters. | ['Shuaicheng Liu', 'Haoqiang Fan', 'Qi Wu', 'Ting Jiang', 'Youwei Li', 'Xinpeng Li', 'Lei Yu'] | 2023-04-14 | null | null | null | null | ['image-super-resolution'] | ['computer-vision'] | [ 1.04067937e-01 -2.47780636e-01 -4.61152613e-01 -4.28859800e-01
-8.53876472e-01 -1.38841018e-01 1.61064208e-01 -3.15965325e-01
-3.67602110e-01 7.98200607e-01 2.14570150e-01 -1.63063675e-01
-3.04538250e-01 -8.76835406e-01 -7.48746753e-01 -6.27957344e-01
5.77718914e-02 -4.81398143e-02 3.61417562e-01 -1.39395088e-01
9.41559859e-03 5.15404224e-01 -1.50912046e+00 3.17031503e-01
1.20893276e+00 1.03943372e+00 1.04765527e-01 6.16987348e-01
1.16079882e-01 8.80800366e-01 -4.26773012e-01 -2.36712724e-01
3.45292449e-01 1.38345107e-01 -5.66834211e-01 -4.51480180e-01
9.87498820e-01 -9.04795706e-01 -3.79355341e-01 1.17368102e+00
8.16963971e-01 3.18012029e-01 2.20595952e-02 -7.45603919e-01
-5.10769725e-01 9.83890295e-01 -9.58175659e-01 6.03913724e-01
-2.58777123e-02 -1.25112832e-01 1.00640035e+00 -9.74344313e-01
3.53642195e-01 1.31234360e+00 9.25592661e-01 3.98646683e-01
-1.26262414e+00 -1.30995631e+00 2.54955769e-01 3.79216999e-01
-1.53482449e+00 -4.92333204e-01 6.26826823e-01 1.70819506e-01
6.19012475e-01 1.49815809e-03 5.00811815e-01 8.36219490e-01
-3.67112070e-01 5.87221622e-01 9.21806157e-01 -7.81842768e-02
2.89705712e-02 -1.03067376e-01 8.71408060e-02 9.01784122e-01
1.67268842e-01 -2.56377812e-02 -8.27859700e-01 1.76778920e-02
1.28009737e+00 2.48535089e-02 -1.90164849e-01 -9.63511318e-03
-9.79118288e-01 7.94188976e-01 8.53423536e-01 1.05300561e-01
-3.00458640e-01 2.35851377e-01 4.37789440e-01 1.06389578e-02
6.32790267e-01 -5.85942306e-02 -4.00498450e-01 8.61137360e-02
-1.25612533e+00 2.92532504e-01 5.48508167e-01 7.71606565e-01
5.79966247e-01 3.07606727e-01 -1.23206809e-01 9.87019777e-01
-1.47692546e-01 2.13215441e-01 2.16245368e-01 -1.55365014e+00
6.15537763e-01 4.31457907e-01 -2.39466578e-01 -1.18332291e+00
-3.43774736e-01 -8.10277402e-01 -1.50286770e+00 1.65288612e-01
2.73345798e-01 -1.81608647e-01 -7.38977611e-01 1.75346482e+00
5.33776224e-01 6.56823695e-01 3.39729479e-03 1.06058919e+00
9.44698036e-01 6.58960044e-01 -2.61355434e-02 -9.64422897e-02
1.38266301e+00 -1.18804014e+00 -5.20810246e-01 1.97793986e-03
2.45489627e-01 -4.57211465e-01 1.06637621e+00 4.66542691e-01
-1.45020914e+00 -8.64409864e-01 -1.28545809e+00 -4.46457833e-01
1.32733285e-01 4.27492196e-03 8.74513566e-01 2.67065883e-01
-9.91701245e-01 8.32140923e-01 -9.18695271e-01 2.85257399e-01
8.39328587e-01 5.24362624e-01 -5.98276667e-02 -1.37727857e-01
-1.27166486e+00 4.08482879e-01 4.30995136e-01 6.04415499e-03
-5.43287873e-01 -1.29655552e+00 -6.33410811e-01 3.68596345e-01
4.79005814e-01 -7.99607754e-01 1.14628577e+00 -6.25606000e-01
-1.40361166e+00 3.25596094e-01 -2.29112357e-01 -7.95419633e-01
3.43864888e-01 -3.94395947e-01 -3.59888494e-01 5.41612089e-01
-1.89459845e-01 8.09691906e-01 8.95704627e-01 -9.86199677e-01
-9.57378209e-01 -3.42887044e-01 1.71145618e-01 3.44670266e-01
-8.99393439e-01 -1.24358833e-01 -4.82951820e-01 -7.52844512e-01
2.67771721e-01 -6.97120965e-01 -3.00885230e-01 2.16731146e-01
-1.54078409e-01 -2.45832354e-01 8.46011579e-01 -6.77289248e-01
1.31176722e+00 -1.99562609e+00 9.43370759e-02 1.93579182e-01
8.92719924e-01 4.57236201e-01 2.37888098e-02 -5.30625463e-01
4.64067422e-02 1.31519645e-01 2.03129519e-02 -3.39391440e-01
-3.96914214e-01 1.50085121e-01 -2.25004777e-01 1.82549641e-01
-7.43125677e-02 7.67408133e-01 -6.45011187e-01 -7.13315845e-01
5.75697795e-02 1.10474539e+00 -8.76522541e-01 -1.20478958e-01
9.00816247e-02 3.83750856e-01 -5.64722121e-01 5.74164450e-01
8.14228833e-01 -6.49494171e-01 1.25810847e-01 -7.37787783e-01
-1.32151634e-01 3.93835127e-01 -1.25382578e+00 1.73734212e+00
-6.81518316e-01 3.64840209e-01 4.73412514e-01 -9.62241352e-01
7.79902041e-01 1.47039413e-01 3.74987215e-01 -9.71667647e-01
-2.13890344e-01 1.40222147e-01 -2.60222852e-01 -9.82521027e-02
7.57603168e-01 1.62232175e-01 3.01374674e-01 3.14802706e-01
-1.90355986e-01 3.80286098e-01 1.76457629e-01 3.23795468e-01
8.28573167e-01 1.38211146e-01 -9.23484098e-03 -8.71587992e-02
4.57656801e-01 -3.12307119e-01 9.10678148e-01 6.69634104e-01
-1.22829229e-01 3.94297212e-01 3.13769579e-01 -5.23313761e-01
-1.15198946e+00 -1.02064478e+00 -9.25725773e-02 1.53245866e+00
7.36483037e-02 -3.50374788e-01 -7.96965718e-01 -2.97091693e-01
-2.01029912e-01 1.67880625e-01 -3.22256923e-01 1.30434811e-01
-1.09936571e+00 -8.23275447e-01 6.14492238e-01 7.17089772e-01
1.10959244e+00 -6.11222684e-01 -3.31021905e-01 -1.17638651e-02
-3.21205020e-01 -1.43108511e+00 -4.27071273e-01 -1.54398307e-01
-1.26275110e+00 -6.20490432e-01 -6.27504647e-01 -8.51134896e-01
5.53122520e-01 6.11342132e-01 1.03307021e+00 3.13958734e-01
-1.75201118e-01 -2.64622897e-01 7.20464960e-02 1.02338441e-01
4.49414225e-03 6.60966992e-01 1.41430959e-01 -4.26471412e-01
9.94444713e-02 -1.08776081e+00 -9.94939148e-01 2.66556069e-02
-6.59689844e-01 3.01107556e-01 6.90786600e-01 6.24409854e-01
8.36339653e-01 3.16111952e-01 5.95313489e-01 -8.04247141e-01
4.85519737e-01 -2.84453154e-01 -5.56130350e-01 5.57779260e-02
-7.33504057e-01 6.16342276e-02 9.05656457e-01 -5.38691461e-01
-1.24307883e+00 -2.13324308e-01 -9.44428965e-02 -5.30201972e-01
3.54743153e-01 5.23158908e-01 1.49528369e-01 -3.95931989e-01
6.70353711e-01 1.54149100e-01 -3.49792913e-02 -5.67871928e-01
3.95013332e-01 4.11123037e-01 9.52781200e-01 -6.56346977e-01
1.15967989e+00 5.90706587e-01 8.92512798e-02 -4.92654741e-01
-1.50468957e+00 -2.27566093e-01 -4.62943912e-01 9.27528590e-02
6.28009439e-01 -1.54665864e+00 -9.12859857e-01 2.15224907e-01
-8.18221688e-01 -1.92206115e-01 9.91902426e-02 2.88767964e-01
-1.27940446e-01 2.02793598e-01 -1.02237844e+00 -4.02575284e-01
-9.61419046e-01 -9.55488265e-01 7.38902092e-01 6.07304275e-01
1.22285426e-01 -9.48419094e-01 -3.34033072e-01 6.87501431e-01
8.09063613e-01 1.35458097e-01 7.02536106e-01 -2.01484427e-01
-8.30748558e-01 1.51552945e-01 -9.41380084e-01 3.71699393e-01
-6.84222281e-02 -1.24010973e-01 -7.28304565e-01 -4.12266880e-01
-2.02812433e-01 -3.62273782e-01 9.89842474e-01 3.40954602e-01
1.72589505e+00 -3.42016876e-01 -7.25327283e-02 1.27775025e+00
1.41475105e+00 -4.62210596e-01 3.67528886e-01 3.93634170e-01
1.04223299e+00 2.52276093e-01 3.62202525e-01 4.71215367e-01
5.93851388e-01 6.56122565e-01 3.80521208e-01 -2.10063010e-01
-3.91678095e-01 -1.84998244e-01 9.67447236e-02 1.04405403e+00
-5.16658902e-01 4.97633576e-01 -6.07074320e-01 3.31367850e-01
-1.72688949e+00 -1.06635869e+00 1.30573168e-01 1.93852806e+00
1.17123199e+00 3.55552137e-01 1.73221096e-01 -1.21902593e-03
6.58543885e-01 4.41704571e-01 -7.88401783e-01 -1.76718738e-02
-7.67010376e-02 5.30821681e-01 4.96052533e-01 4.14556116e-01
-9.50959027e-01 9.31861043e-01 6.37740660e+00 1.13518715e+00
-1.14228547e+00 2.52930939e-01 6.48443401e-01 -5.14042437e-01
-1.98999122e-01 -3.90334815e-01 -1.21938419e+00 2.90541828e-01
9.89660203e-01 -2.05299854e-01 6.44366980e-01 9.96335983e-01
2.37250090e-01 2.45125994e-01 -6.57188952e-01 8.88805449e-01
-1.18968308e-01 -1.83072543e+00 1.30548134e-01 -1.45200178e-01
8.65513504e-01 9.32626203e-02 2.35560477e-01 4.48256969e-01
2.35134661e-01 -1.09013104e+00 1.62698150e-01 2.38255322e-01
9.02847230e-01 -1.22166979e+00 4.74836648e-01 2.13278517e-01
-1.56631947e+00 -2.25904837e-01 -5.52296042e-01 -1.50507405e-01
-3.86739615e-03 7.77593732e-01 -3.63191336e-01 4.25233662e-01
1.10857332e+00 5.86467564e-01 -4.41411227e-01 6.67652249e-01
-1.06652059e-01 5.90974391e-01 -4.71420825e-01 3.16861182e-01
7.84798265e-02 -6.74229860e-02 3.34984362e-01 1.05105317e+00
2.20990032e-01 1.99717030e-01 2.08886310e-01 8.02570999e-01
-4.96208847e-01 -2.28734374e-01 3.06742731e-02 4.05101508e-01
9.98142242e-01 1.67693174e+00 -4.47640985e-01 -4.46933359e-01
-4.37477946e-01 6.44497991e-01 7.40329266e-01 2.50282854e-01
-1.01469123e+00 -2.98256636e-01 8.37688386e-01 1.75255507e-01
3.99798185e-01 -3.03055763e-01 -4.93802756e-01 -1.10033572e+00
1.23125590e-01 -9.80031669e-01 3.74163836e-01 -5.86917520e-01
-9.36218083e-01 3.55880141e-01 -5.82004823e-02 -8.45332146e-01
2.01339982e-02 -1.70762017e-01 -5.29568076e-01 7.36236215e-01
-1.95090544e+00 -1.13284588e+00 -6.48058534e-01 6.96100950e-01
4.47796017e-01 -3.76477651e-02 4.49405581e-01 6.69077933e-01
-9.06594753e-01 9.74679828e-01 -7.11424202e-02 9.36410278e-02
7.07452953e-01 -1.01955700e+00 3.32732797e-01 8.87683511e-01
-2.16672227e-01 8.56879890e-01 3.99078250e-01 -3.47169697e-01
-1.18703210e+00 -1.24382746e+00 3.01886082e-01 1.39132857e-01
6.72388136e-01 -1.62279144e-01 -1.25927496e+00 5.09136558e-01
-2.11588189e-01 4.03346747e-01 4.09543484e-01 2.99374908e-01
-6.45884514e-01 -5.09720445e-01 -1.08753717e+00 6.76319659e-01
1.13489807e+00 -5.50699055e-01 -2.00135872e-01 2.72827446e-01
1.22684276e+00 -7.74812937e-01 -1.31063044e+00 6.94345057e-01
7.07456708e-01 -9.98399615e-01 1.54147696e+00 -2.27169976e-01
8.96213055e-01 -1.58862799e-01 1.70482080e-02 -8.17416549e-01
-5.55974662e-01 -6.22824371e-01 -9.08650517e-01 1.08994555e+00
1.91982478e-01 -3.98667663e-01 1.23629737e+00 4.34700429e-01
8.60424489e-02 -1.10957575e+00 -8.01860332e-01 -4.85808671e-01
-6.25953497e-03 -1.85818672e-01 6.14596844e-01 8.91024351e-01
-5.57376623e-01 3.65920991e-01 -5.26190698e-01 4.05022442e-01
1.12416947e+00 1.61321566e-01 6.95377886e-01 -1.18127155e+00
-2.55979121e-01 -4.11005944e-01 1.38146445e-01 -1.30691886e+00
1.13143019e-01 -7.15590775e-01 -3.47861230e-01 -1.28738368e+00
4.11107451e-01 -6.71725035e-01 -3.83659393e-01 4.36482787e-01
-4.77896124e-01 6.30799770e-01 6.53397292e-02 3.06772053e-01
-7.95701742e-01 2.89571881e-01 1.48184216e+00 3.27442229e-01
-2.86289901e-01 -6.68819398e-02 -9.80720103e-01 9.36221242e-01
8.50269496e-01 -2.53974408e-01 -4.12500262e-01 -6.73941791e-01
3.59608561e-01 6.22132309e-02 4.48962659e-01 -1.05343783e+00
4.90864754e-01 -6.55746181e-03 5.41803658e-01 -6.24966383e-01
2.77261436e-01 -4.95279759e-01 -2.08330870e-01 3.52890253e-01
-3.88792038e-01 -8.93555433e-02 1.41291097e-01 4.15194243e-01
-8.77066478e-02 8.70783255e-02 1.05800521e+00 2.85768099e-02
-6.62573397e-01 7.81626761e-01 3.59437346e-01 1.85667444e-02
6.06406510e-01 -1.98669508e-01 -4.78867561e-01 -2.45530471e-01
-5.84335566e-01 3.52535516e-01 1.84331879e-01 2.12929338e-01
5.58290720e-01 -1.39448059e+00 -7.84708083e-01 1.75208244e-02
-6.31310761e-01 6.54840469e-01 5.86477697e-01 8.78429651e-01
-5.63303292e-01 3.99892218e-02 -1.81664675e-01 -4.81108606e-01
-1.51009774e+00 2.60512948e-01 1.69495523e-01 -6.67390108e-01
-9.59964693e-01 9.59332764e-01 -5.27056344e-02 -3.77596080e-01
4.74482089e-01 1.46327810e-02 -2.37023607e-01 2.37317421e-02
1.02383792e+00 7.84154475e-01 -1.44818068e-01 -1.66275799e-01
-6.65475726e-02 6.40712857e-01 -6.00945592e-01 2.37409875e-01
1.58680189e+00 -2.44789630e-01 -8.61771405e-02 -1.20742723e-01
1.13838816e+00 -5.72265945e-02 -1.59030581e+00 -7.00412929e-01
-3.63977700e-01 -3.14763635e-01 5.92717052e-01 -4.64439988e-01
-1.57034516e+00 6.68294787e-01 5.69957137e-01 -1.14766523e-01
1.42792368e+00 -3.87945354e-01 1.28918219e+00 4.90742892e-01
2.27041826e-01 -1.16191757e+00 2.52013147e-01 4.34822053e-01
4.43257481e-01 -1.27321613e+00 5.87657034e-01 -6.25478148e-01
-1.06933191e-01 1.05838633e+00 8.86872113e-01 -3.36402774e-01
4.96098250e-01 5.90090871e-01 -2.84142464e-01 4.17314582e-02
-8.04778337e-01 1.50684208e-01 2.10990846e-01 3.87512863e-01
4.09640640e-01 -1.42402828e-01 -1.17118470e-02 5.43355644e-01
-5.00586092e-01 2.46841520e-01 2.56354004e-01 5.32229185e-01
-5.79540253e-01 -7.10157633e-01 -1.87618703e-01 4.58129585e-01
-8.22791636e-01 -3.48637611e-01 3.90478492e-01 7.07366586e-01
-4.31514420e-02 4.89061207e-01 1.47470072e-01 -3.12921524e-01
1.75856099e-01 -4.03571218e-01 5.13201773e-01 -3.09060782e-01
-6.04470968e-01 -1.35144249e-01 -4.43403758e-02 -7.77130008e-01
-6.04172468e-01 -1.70302555e-01 -1.42716765e+00 -1.01742601e+00
-1.90528303e-01 -1.76857620e-01 6.02059603e-01 8.71603012e-01
5.69063425e-01 8.03476751e-01 5.23981154e-01 -1.04204237e+00
-6.97501242e-01 -6.69338226e-01 -1.76073909e-01 -7.50090480e-02
3.74740571e-01 -3.55972439e-01 -3.64247233e-01 -9.36979428e-02] | [10.935439109802246, -1.7189263105392456] |
b6cb9d33-4a5c-43c0-ba3c-963924b2b890 | text-ranking-and-classification-using-data | 2109.11577 | null | https://arxiv.org/abs/2109.11577v2 | https://arxiv.org/pdf/2109.11577v2.pdf | Text Ranking and Classification using Data Compression | A well-known but rarely used approach to text categorization uses conditional entropy estimates computed using data compression tools. Text affinity scores derived from compressed sizes can be used for classification and ranking tasks, but their success depends on the compression tools used. We use the Zstandard compressor and strengthen these ideas in several ways, calling the resulting language-agnostic technique Zest. In applications, this approach simplifies configuration, avoiding careful feature extraction and large ML models. Our ablation studies confirm the value of individual enhancements we introduce. We show that Zest complements and can compete with language-specific multidimensional content embeddings in production, but cannot outperform other counting methods on public datasets. | ['Igor L. Markov', 'Nitya Kasturi'] | 2021-09-23 | null | https://openreview.net/forum?id=mCyM2CWFZX5 | https://openreview.net/pdf?id=mCyM2CWFZX5 | neurips-workshop-icbinb-2021-12 | ['text-categorization'] | ['natural-language-processing'] | [ 1.20220803e-01 -2.99020201e-01 -6.19689584e-01 -2.64538199e-01
-8.38352561e-01 -8.20748389e-01 9.15617704e-01 7.28841007e-01
-1.11527312e+00 5.07418036e-01 3.55986089e-01 -5.12933612e-01
-3.58144045e-01 -7.64588416e-01 -6.47878945e-02 -4.55439925e-01
6.74599633e-02 1.01838815e+00 4.20434058e-01 -1.53976217e-01
7.71585763e-01 3.10555816e-01 -1.86907399e+00 2.70496994e-01
5.20495474e-01 8.58371973e-01 1.87940806e-01 1.12111068e+00
-7.52621949e-01 6.51346684e-01 -6.14194214e-01 -9.39048529e-01
2.21481681e-01 5.53644598e-02 -9.40291345e-01 -5.43786526e-01
4.28849667e-01 -4.66319561e-01 -3.99713218e-01 6.89722836e-01
4.20916110e-01 -6.06339099e-03 1.11866915e+00 -1.02760708e+00
-7.41192818e-01 1.11805892e+00 -4.27764595e-01 3.47493857e-01
2.11484179e-01 -2.79759854e-01 1.57704282e+00 -7.60684550e-01
4.72077638e-01 1.23907387e+00 8.64186227e-01 2.49009475e-01
-1.31533587e+00 -5.98434389e-01 -4.44890797e-01 -2.85658222e-02
-1.35847676e+00 -4.69378203e-01 2.41782829e-01 -2.68239260e-01
1.36243534e+00 5.63115001e-01 3.21272701e-01 9.18131530e-01
-3.16371359e-02 7.79070914e-01 1.07341850e+00 -7.60536671e-01
9.53813270e-02 4.93688196e-01 4.84989166e-01 8.29180837e-01
8.73650074e-01 -3.38602394e-01 -4.98080492e-01 -5.27757466e-01
1.00498326e-01 -1.11577027e-01 4.35188413e-03 -3.97577226e-01
-1.02472484e+00 1.38823950e+00 -2.14301139e-01 5.37210345e-01
3.49552661e-01 3.95633429e-01 8.72865200e-01 5.03477991e-01
6.97429180e-01 8.78625453e-01 -8.20772171e-01 -6.29774094e-01
-1.19726431e+00 3.32801044e-01 1.03977466e+00 9.06975806e-01
3.64817679e-01 -3.92921895e-01 -6.71466738e-02 1.02482474e+00
-5.41107506e-02 2.62783676e-01 1.20355833e+00 -9.39582348e-01
4.86090213e-01 5.65169632e-01 -2.50246555e-01 -8.04523587e-01
-5.01451731e-01 -2.49476597e-01 -4.91577834e-01 -2.06488967e-01
6.31665707e-01 2.64491469e-01 -5.58447659e-01 1.37757277e+00
-1.14287280e-01 -5.50943017e-01 -6.20956346e-02 1.98892817e-01
3.33403617e-01 2.60350406e-01 1.20672643e-01 1.05555952e-01
1.41268134e+00 -6.45547330e-01 -4.28851753e-01 -2.07641274e-01
1.48560059e+00 -9.05052722e-01 1.35991144e+00 4.36611682e-01
-8.33377719e-01 -1.22824781e-01 -1.07455158e+00 -6.07495546e-01
-8.47454607e-01 1.83164388e-01 1.00245011e+00 1.34048796e+00
-9.36351240e-01 9.41510618e-01 -6.41501069e-01 -4.14943844e-01
4.44128573e-01 4.32471812e-01 -3.29922885e-01 8.03114846e-02
-1.13408697e+00 1.12435174e+00 6.20868564e-01 -8.87816072e-01
2.35995963e-01 -5.36912918e-01 -7.07156479e-01 2.23231539e-01
5.27168112e-03 -3.79004151e-01 1.26214552e+00 -3.83260906e-01
-1.23148775e+00 1.03211629e+00 -2.06518620e-02 -7.49386728e-01
4.23752576e-01 -8.50010663e-02 -4.58741449e-02 1.77378163e-01
-1.85142517e-01 5.74508369e-01 6.40399098e-01 -4.98988837e-01
-5.91592312e-01 -1.61102667e-01 -1.88967481e-01 1.50588676e-01
-1.36781001e+00 1.09021939e-01 -2.80020624e-01 -5.83368301e-01
2.67217448e-03 -7.38370657e-01 2.14272931e-01 2.16661409e-01
-1.47791013e-01 -3.81052256e-01 5.52971244e-01 -5.53060949e-01
1.56750202e+00 -2.02776027e+00 -5.90019301e-02 8.70586857e-02
4.24610585e-01 1.56389832e-01 -1.18150666e-01 5.29546678e-01
2.98361272e-01 7.76883841e-01 -9.01021361e-02 -3.55795681e-01
4.17507231e-01 1.63052425e-01 -3.25590700e-01 3.52831364e-01
-1.23673357e-01 7.99035549e-01 -7.48048484e-01 -9.14288044e-01
9.29942206e-02 1.78648978e-01 -8.03695321e-01 -3.02180350e-01
-5.33111654e-02 -9.63447094e-01 -2.35382870e-01 3.59875381e-01
4.20789689e-01 -2.00058207e-01 2.63823956e-01 9.84181687e-02
-5.64247631e-02 6.70394361e-01 -9.23145294e-01 1.49594700e+00
-7.91109562e-01 9.26139355e-01 -4.97312546e-01 -7.90870667e-01
5.81847131e-01 -1.09622069e-01 2.43462339e-01 -3.20515394e-01
2.24352762e-01 3.31163675e-01 8.33746642e-02 -1.60010368e-01
1.06372905e+00 -3.57944481e-02 -3.14695328e-01 8.65625918e-01
2.93846518e-01 -4.31189358e-01 3.92266035e-01 5.51564872e-01
1.19675040e+00 -3.77387732e-01 5.41471004e-01 -5.07099807e-01
1.93678126e-01 5.47250248e-02 -1.68569610e-01 9.33017790e-01
-7.34185651e-02 5.07638454e-01 7.37789631e-01 -1.85805663e-01
-1.40504360e+00 -8.25132906e-01 -6.62677526e-01 1.66412163e+00
-1.49576440e-01 -1.02637434e+00 -5.13794184e-01 -9.87847507e-01
4.95001912e-01 8.99864554e-01 -7.06070840e-01 -2.42007852e-01
-3.48938257e-01 -9.65406060e-01 9.81763184e-01 5.74468911e-01
-1.34761017e-02 -3.89663458e-01 -5.90510070e-01 3.49917077e-02
-3.05698756e-02 -8.76949787e-01 -4.15994197e-01 7.17179537e-01
-1.01338470e+00 -8.24003577e-01 -6.83940649e-01 -5.15543580e-01
3.74401540e-01 3.19313914e-01 1.27458465e+00 3.35457772e-01
-4.30485308e-01 3.78346771e-01 -5.49328804e-01 -4.39468652e-01
-5.50554037e-01 7.72143602e-01 2.29147941e-01 -7.72690952e-01
9.47705090e-01 -3.38316798e-01 -2.39987150e-01 2.52987593e-02
-8.50415945e-01 -2.78870672e-01 6.75933301e-01 1.09588242e+00
1.55824810e-01 3.19479406e-02 2.37065658e-01 -1.11576331e+00
1.07350624e+00 -2.97093928e-01 -5.24282694e-01 2.45684683e-01
-1.08663106e+00 8.32031369e-01 7.23106146e-01 -5.22635043e-01
-4.22402501e-01 -2.11426109e-01 -2.04833403e-01 -3.60870101e-02
2.22897217e-01 2.79559404e-01 3.97142172e-01 -4.47031995e-03
8.62001181e-01 2.42367834e-01 -2.91466974e-02 -5.29924929e-01
6.00161076e-01 1.11440480e+00 1.70066953e-04 -6.27959728e-01
6.46473467e-01 2.01479629e-01 -1.37213767e-01 -7.05551505e-01
-6.81226254e-01 -5.26537597e-01 -6.63066745e-01 4.84666198e-01
5.32757819e-01 -5.35507679e-01 -5.57022870e-01 -3.79228927e-02
-9.89557147e-01 -6.88617080e-02 -4.79555368e-01 4.17310447e-01
-4.15857315e-01 5.75745642e-01 -6.88088596e-01 -7.26802647e-01
-4.61159319e-01 -5.89573920e-01 1.04541469e+00 -3.68375957e-01
-5.14319301e-01 -1.03301215e+00 1.38970822e-01 9.23960656e-02
5.99096298e-01 -5.74636698e-01 1.29103756e+00 -1.26314807e+00
-3.55665907e-02 -6.33254111e-01 -4.57504719e-01 4.22927022e-01
-1.97363913e-01 1.48283869e-01 -9.56063211e-01 -1.59371123e-01
-3.58220309e-01 -6.48118317e-01 1.29145277e+00 5.46520725e-02
1.55185926e+00 -2.26950213e-01 -2.69037634e-01 5.79445064e-01
1.38915682e+00 -3.43343437e-01 3.65040660e-01 4.90380853e-01
3.90512705e-01 5.65026224e-01 8.19461942e-02 4.55331624e-01
1.86127692e-01 4.42991495e-01 -1.96767226e-01 5.31072855e-01
-1.48582354e-01 -3.32151383e-01 1.97353303e-01 1.19752049e+00
3.57965797e-01 -4.02066022e-01 -7.72565484e-01 4.55989778e-01
-1.23531187e+00 -1.06142211e+00 -4.56780121e-02 2.24295235e+00
9.88309085e-01 5.25908947e-01 6.66958168e-02 7.01801658e-01
2.20841438e-01 4.89624962e-02 -2.21677244e-01 -8.67315650e-01
-1.86272413e-01 5.78978956e-01 1.11610341e+00 4.04760957e-01
-1.00581157e+00 9.03366208e-01 7.56778145e+00 1.23744941e+00
-6.20853186e-01 2.83150405e-01 4.78805870e-01 -1.83580652e-01
-4.03625816e-01 -1.39386937e-01 -9.50239837e-01 4.43609416e-01
1.20966399e+00 -6.98503137e-01 4.18941617e-01 1.01992571e+00
-6.98723197e-01 -2.49360308e-01 -1.15038252e+00 1.06451750e+00
2.57469445e-01 -1.20363820e+00 5.05017601e-02 8.57834667e-02
2.33074069e-01 1.71945885e-01 2.45643735e-01 5.25842786e-01
6.80783331e-01 -8.06524038e-01 5.27065217e-01 -1.23458415e-01
1.08883071e+00 -7.89267063e-01 7.41461813e-01 2.05564335e-01
-9.73584235e-01 -2.68119007e-01 -7.54805624e-01 -5.44578768e-03
-2.57987618e-01 6.54285014e-01 -8.91588211e-01 1.15017414e-01
3.88461798e-01 1.72553286e-01 -1.07284260e+00 6.49195611e-01
1.91265047e-01 6.53141916e-01 -5.71813762e-01 -6.43540800e-01
-1.98002700e-02 1.29396588e-01 1.48439944e-01 1.68158293e+00
4.69816804e-01 -4.08736616e-01 -4.19788450e-01 4.01522607e-01
-2.65264928e-01 4.09651756e-01 -8.07412028e-01 -3.80796611e-01
5.95811605e-01 1.28673208e+00 -9.03841734e-01 -6.13074660e-01
-3.93518865e-01 1.04449093e+00 5.41240513e-01 -3.42741400e-01
-5.50420105e-01 -8.91732037e-01 6.15339518e-01 5.50245978e-02
4.08293366e-01 -3.50777179e-01 -5.98553598e-01 -1.37794018e+00
-2.44116366e-01 -7.30626523e-01 6.35146856e-01 -2.24850848e-01
-1.41895485e+00 5.54129124e-01 2.18998000e-01 -1.01986372e+00
-4.80494678e-01 -9.93950784e-01 6.06445409e-02 6.85706556e-01
-1.30426276e+00 -5.11538982e-01 1.27156198e-01 7.36049786e-02
4.44646925e-01 -3.34193170e-01 1.21278834e+00 2.43391380e-01
-2.25983277e-01 1.06964338e+00 6.02478504e-01 1.17696464e-01
7.70257056e-01 -1.56374383e+00 5.63563943e-01 4.02093023e-01
5.77181995e-01 7.29114771e-01 6.44580007e-01 -2.90282398e-01
-1.33869302e+00 -7.14524448e-01 1.35716093e+00 -8.54701817e-01
9.29597497e-01 -7.72482157e-01 -6.54029489e-01 3.41479272e-01
-7.03643560e-02 -1.57132745e-01 8.97785783e-01 5.16028464e-01
-9.44046795e-01 1.63291708e-01 -1.25967216e+00 6.66138291e-01
1.16930187e+00 -7.70543575e-01 -7.02767789e-01 4.01994973e-01
7.04569876e-01 -1.51775880e-02 -6.15660608e-01 -1.07140191e-01
7.80521512e-01 -7.74635315e-01 9.29430008e-01 -6.40339911e-01
6.26910567e-01 3.71849418e-01 -4.10821199e-01 -1.22106481e+00
-3.54878247e-01 -2.01187149e-01 -2.57178545e-01 9.76578236e-01
4.31066751e-01 -7.84946918e-01 7.97394097e-01 5.74113727e-01
4.29862589e-01 -4.87839073e-01 -8.56576681e-01 -1.07766294e+00
6.96874976e-01 -7.10399747e-01 6.55681551e-01 9.52283740e-01
3.77304256e-01 2.87847549e-01 5.20984735e-03 -6.78900003e-01
4.22441930e-01 -4.67677973e-02 5.11339068e-01 -1.60605109e+00
-4.64310616e-01 -9.86904383e-01 -5.27118206e-01 -1.06290603e+00
1.96947590e-01 -1.37214327e+00 -3.75972033e-01 -1.15455329e+00
4.76641089e-01 -6.79291606e-01 -2.18005553e-01 2.25183338e-01
9.65420306e-02 4.88613814e-01 1.42365322e-01 1.61625966e-01
-4.42934275e-01 2.22707808e-01 5.23811996e-01 -2.20511124e-01
1.28190547e-01 -4.86738324e-01 -9.87559378e-01 6.94854856e-01
9.14908648e-01 -8.47156644e-01 -4.03557628e-01 -5.36749661e-01
5.65681338e-01 -5.46971262e-01 -1.21480383e-01 -9.96537149e-01
1.57043442e-01 8.82297829e-02 4.14471805e-01 -3.81207645e-01
1.53016090e-01 -6.38582170e-01 -5.89699924e-01 4.29089040e-01
-8.97937655e-01 5.09222567e-01 5.91959953e-02 4.65085089e-01
9.37872529e-02 -7.90394306e-01 5.90529501e-01 -3.33149917e-02
-1.38520554e-01 8.17800537e-02 -4.58480567e-01 3.14515382e-01
5.93382955e-01 -1.43420652e-01 -4.35536891e-01 -2.81692982e-01
7.80981481e-02 -2.66671240e-01 5.88629127e-01 3.15044016e-01
3.81329358e-01 -1.12309158e+00 -6.23400509e-01 1.46324441e-01
3.41858298e-01 -6.88781679e-01 -4.38432336e-01 5.44016898e-01
-7.73840308e-01 6.80643499e-01 6.50672466e-02 -2.67705530e-01
-1.30836594e+00 6.17202401e-01 -7.63294399e-02 -5.32085180e-01
-6.20821357e-01 8.47133338e-01 -3.86777818e-01 -5.75619876e-01
1.81468770e-01 -4.10889208e-01 -9.01954770e-02 4.66337204e-01
6.58423126e-01 4.80444580e-01 4.18773979e-01 -2.12073416e-01
-2.23953888e-01 2.35882476e-01 -3.52337122e-01 -5.57053506e-01
1.06642473e+00 -1.40451808e-02 -3.40423286e-02 5.52649736e-01
1.71686876e+00 2.30124503e-01 -3.51149023e-01 -1.10845692e-01
3.21538359e-01 -6.95603311e-01 3.75650167e-01 -4.93129611e-01
-6.35431468e-01 1.03716600e+00 5.44061899e-01 4.71907228e-01
7.65181005e-01 -1.80397168e-01 9.21033561e-01 8.44602764e-01
3.73883933e-01 -1.39334238e+00 -1.91820398e-01 6.11258566e-01
2.69779593e-01 -9.72866178e-01 3.94214511e-01 -2.75871586e-02
-4.62829351e-01 1.35731578e+00 8.28353316e-02 1.24787934e-01
6.21944785e-01 5.21555483e-01 -2.66397864e-01 1.00719750e-01
-9.81871009e-01 -2.22717568e-01 1.72118664e-01 4.57864672e-01
7.45331526e-01 5.40364161e-03 -7.21028030e-01 5.77902496e-01
-6.31754220e-01 -2.98171550e-01 4.59242523e-01 9.32971418e-01
-5.12518644e-01 -1.46695840e+00 -4.83314833e-03 1.34812617e+00
-5.99002481e-01 -6.39580071e-01 -5.38221061e-01 8.33426535e-01
-7.67070204e-02 6.11190319e-01 2.21380398e-01 -6.82868838e-01
-2.46350572e-01 6.07884288e-01 6.28061950e-01 -4.29958791e-01
-4.55826432e-01 -5.28650045e-01 2.53265470e-01 -1.40818417e-01
1.48449197e-01 -8.03910971e-01 -9.38297808e-01 -7.48693347e-01
-6.01315439e-01 1.09303400e-01 8.62992823e-01 6.27895057e-01
2.66234517e-01 -6.06410578e-02 4.30607170e-01 -5.42244852e-01
-1.05885696e+00 -1.01253402e+00 -5.34835279e-01 4.38002020e-01
7.26180226e-02 -6.13106191e-01 -7.85142541e-01 -2.52887666e-01] | [10.596752166748047, 8.427888870239258] |
6f08b18e-5ca0-46f6-83bd-32e57ab5d27c | federated-learning-for-semantic-parsing-task | 2305.17221 | null | https://arxiv.org/abs/2305.17221v1 | https://arxiv.org/pdf/2305.17221v1.pdf | Federated Learning for Semantic Parsing: Task Formulation, Evaluation Setup, New Algorithms | This paper studies a new task of federated learning (FL) for semantic parsing, where multiple clients collaboratively train one global model without sharing their semantic parsing data. By leveraging data from multiple clients, the FL paradigm can be especially beneficial for clients that have little training data to develop a data-hungry neural semantic parser on their own. We propose an evaluation setup to study this task, where we re-purpose widely-used single-domain text-to-SQL datasets as clients to form a realistic heterogeneous FL setting and collaboratively train a global model. As standard FL algorithms suffer from the high client heterogeneity in our realistic setup, we further propose a novel LOss Reduction Adjusted Re-weighting (Lorar) mechanism to mitigate the performance degradation, which adjusts each client's contribution to the global model update based on its training loss reduction during each round. Our intuition is that the larger the loss reduction, the further away the current global model is from the client's local optimum, and the larger weight the client should get. By applying Lorar to three widely adopted FL algorithms (FedAvg, FedOPT and FedProx), we observe that their performance can be improved substantially on average (4%-20% absolute gain under MacroAvg) and that clients with smaller datasets enjoy larger performance gains. In addition, the global model converges faster for almost all the clients. | ['Huan Sun', 'Yu Su', 'Wei-Han Lee', 'Changchang Liu', 'Tianshu Zhang'] | 2023-05-26 | null | null | null | null | ['text-to-sql', 'semantic-parsing'] | ['computer-code', 'natural-language-processing'] | [-7.39302412e-02 3.50089282e-01 -3.47233385e-01 -8.42460275e-01
-1.26875722e+00 -5.00608861e-01 -2.15313374e-03 -2.07979679e-01
-4.81195867e-01 5.38876414e-01 2.60052472e-01 -3.13892335e-01
1.56432409e-02 -6.65438771e-01 -8.72992218e-01 -6.06423080e-01
2.73740321e-01 8.65242660e-01 4.22194391e-01 8.64419565e-02
-2.42149413e-01 2.27944389e-01 -1.19894445e+00 5.84634900e-01
7.88411379e-01 1.04358602e+00 5.82951963e-01 6.37948930e-01
-6.72705114e-01 1.17048228e+00 -4.33592826e-01 -8.74260783e-01
4.00857985e-01 -2.02072740e-01 -1.08309460e+00 -4.45367619e-02
3.61220926e-01 -5.39968848e-01 -1.35511354e-01 9.00278807e-01
4.64557081e-01 1.52069449e-01 -5.63307926e-02 -1.23244905e+00
-4.50889856e-01 1.08202279e+00 -4.40345794e-01 -1.62652820e-01
-1.26055613e-01 1.58863440e-01 1.32877862e+00 -5.34461737e-01
6.29743040e-01 1.58896756e+00 7.01874077e-01 8.44188750e-01
-1.19774508e+00 -6.72623158e-01 7.23223209e-01 -2.63908859e-02
-7.01655746e-01 -6.73516810e-01 5.95751107e-01 2.53637224e-01
9.40663874e-01 1.48598492e-01 -3.38017732e-01 9.05410171e-01
-5.02713442e-01 1.15318048e+00 7.65560091e-01 -3.08248311e-01
4.02848184e-01 2.20753625e-01 2.30361581e-01 5.90611756e-01
3.56910080e-02 -3.79130244e-01 -6.76963031e-01 -3.57337981e-01
3.18393856e-01 4.64091413e-02 -2.28961278e-02 -4.95048255e-01
-4.47006673e-01 1.01917934e+00 4.40788805e-01 2.22631302e-02
-4.18490589e-01 3.09622973e-01 5.45895398e-01 5.80206573e-01
4.72813755e-01 -8.61228704e-02 -1.08335900e+00 -3.91326249e-01
-4.90062684e-01 3.00105095e-01 1.11810946e+00 9.44409311e-01
9.14905548e-01 -3.86263907e-01 5.04213870e-02 1.31337523e+00
3.28733802e-01 2.37646610e-01 5.38321078e-01 -1.47590792e+00
1.18064797e+00 7.78213382e-01 -1.21345408e-01 -2.34578907e-01
-1.69740200e-01 -2.79717326e-01 -1.40923068e-01 9.90142524e-02
6.22349262e-01 -4.37927783e-01 -6.14342868e-01 2.26262045e+00
3.93719763e-01 -4.94763479e-02 3.37435305e-01 7.43317783e-01
4.69214648e-01 3.64729166e-01 4.02227700e-01 6.10573143e-02
9.80121374e-01 -1.35458016e+00 -1.47711098e-01 -5.75309038e-01
1.11196744e+00 -4.07216251e-01 1.28930318e+00 1.87560573e-01
-1.14702499e+00 5.53791178e-03 -5.48948824e-01 -1.12145588e-01
4.98444475e-02 -2.31420428e-01 7.20418215e-01 5.63917875e-01
-1.02588844e+00 6.67545319e-01 -1.05779314e+00 -4.25457239e-01
8.94558728e-01 3.62030476e-01 -2.45040625e-01 -4.17845070e-01
-6.10631943e-01 4.10941064e-01 1.04955800e-01 -5.10396779e-01
-7.19390452e-01 -8.17930102e-01 -3.60465586e-01 4.16683525e-01
6.54506683e-01 -5.83771825e-01 1.89928675e+00 -1.14505768e+00
-1.52436316e+00 8.03389013e-01 -3.06797475e-01 -4.14602846e-01
8.39228392e-01 -2.46082768e-01 1.29307359e-01 -1.84084892e-01
1.18206041e-02 4.83483106e-01 3.76147479e-01 -1.36348796e+00
-8.73468399e-01 -7.63563812e-01 2.08192438e-01 3.57611924e-01
-4.62567747e-01 2.27744833e-01 -8.25057983e-01 -3.12318802e-01
3.01398903e-01 -7.05037653e-01 -1.43048495e-01 2.48490736e-01
-2.73368079e-02 -3.98273975e-01 9.27706003e-01 -4.45421070e-01
8.47956359e-01 -2.17605734e+00 -9.07183066e-02 -1.61562003e-02
2.77337104e-01 2.35074714e-01 -5.76503813e-01 3.35463405e-01
4.18590963e-01 -2.84218229e-03 -1.52828395e-01 -8.96772861e-01
4.87335846e-02 6.51015997e-01 -1.00084767e-01 -2.78701540e-02
-1.61160618e-01 8.06420505e-01 -6.75697327e-01 -3.77070546e-01
-3.61314029e-01 1.78169549e-01 -1.09815931e+00 4.87533033e-01
-6.20508850e-01 2.03487977e-01 -8.18109691e-01 6.07621610e-01
8.88293028e-01 -5.82313240e-01 7.32076526e-01 1.13312617e-01
3.91744792e-01 4.57429171e-01 -1.09265745e+00 1.78556716e+00
-9.65149343e-01 1.97177649e-01 7.72941172e-01 -1.11585534e+00
8.91966939e-01 6.52593598e-02 4.17016983e-01 -9.82226014e-01
-9.56077222e-03 2.02724963e-01 -3.50734144e-01 -2.80048132e-01
5.45027517e-02 -2.06971932e-02 1.01758428e-01 9.12150621e-01
6.02210239e-02 6.11651123e-01 -1.15726501e-01 4.44119364e-01
1.53825688e+00 -5.39358647e-04 -3.75783324e-01 6.89625889e-02
2.55875885e-01 -1.44576728e-01 8.84781778e-01 9.24849033e-01
-3.06306064e-01 3.72590244e-01 7.69765794e-01 -3.90917510e-01
-7.74864554e-01 -9.81666207e-01 3.13254535e-01 2.19978189e+00
1.10856451e-01 -7.17679560e-02 -9.22911108e-01 -1.23314905e+00
3.92544389e-01 6.28294945e-01 -9.33084711e-02 -4.17758934e-02
-8.46394122e-01 -7.95635343e-01 5.59847891e-01 7.47052670e-01
6.27681136e-01 -1.11783051e+00 -3.64153206e-01 4.06149119e-01
-2.87856311e-01 -1.16258407e+00 -3.97040099e-01 3.51152658e-01
-1.07082546e+00 -8.72389317e-01 -3.76881063e-01 -6.61610365e-01
3.74654472e-01 3.81318331e-01 1.26713264e+00 1.84639812e-01
8.04570913e-02 9.67419371e-02 -4.83107120e-01 -9.81677026e-02
-6.60034776e-01 2.80198157e-01 -4.43057626e-01 -5.64973913e-02
4.27159727e-01 -6.73563361e-01 -5.02651215e-01 4.41932231e-01
-6.66813970e-01 -2.81369630e-02 3.50254208e-01 7.54316628e-01
2.12860599e-01 -3.20569962e-01 8.01573157e-01 -1.43390548e+00
5.72746038e-01 -7.04171121e-01 -5.66799760e-01 4.32025403e-01
-8.92823160e-01 3.25134277e-01 7.93187022e-01 -3.75074506e-01
-1.61580682e+00 -2.28687212e-01 -2.67948568e-01 -2.75936216e-01
1.31184518e-01 1.50860325e-01 -4.91374552e-01 -4.22188863e-02
4.87823695e-01 -1.82327196e-01 1.07408769e-01 -1.13741803e+00
5.60213029e-01 8.34817529e-01 4.72658426e-01 -1.01359606e+00
3.92455995e-01 3.74127984e-01 -5.17415643e-01 1.20852940e-01
-8.31053257e-01 -3.65412176e-01 -3.97068821e-02 1.51097119e-01
4.90849763e-01 -9.04292881e-01 -9.94950950e-01 5.97198188e-01
-1.11626935e+00 -9.03834224e-01 -2.48948961e-01 1.44421786e-01
-5.23062050e-01 2.81784594e-01 -8.83015752e-01 -7.02560127e-01
-6.13755167e-01 -1.12273443e+00 8.25151026e-01 2.42402956e-01
1.59393519e-01 -9.34477270e-01 -2.97201369e-02 8.24508131e-01
7.57501245e-01 -3.59871686e-01 1.12905908e+00 -1.04913461e+00
-6.25239491e-01 8.48709121e-02 -6.63208187e-01 5.46350062e-01
-9.06283632e-02 -5.53117454e-01 -1.19111884e+00 -3.03225845e-01
-9.58776101e-03 -7.54588187e-01 8.49339366e-01 3.98550853e-02
1.31021678e+00 -5.27595460e-01 -3.18869025e-01 7.41489470e-01
1.62958241e+00 1.59115627e-01 1.16758712e-01 4.23243642e-01
6.95927978e-01 6.10736191e-01 2.81558663e-01 5.29916763e-01
7.29675591e-01 6.79572105e-01 6.71168447e-01 1.83955312e-01
-2.36005560e-01 -3.44351172e-01 5.02085626e-01 5.67135632e-01
4.09088075e-01 -5.65829933e-01 -6.92938685e-01 4.25854772e-01
-2.11613011e+00 -4.59099084e-01 3.01959515e-01 2.14919639e+00
8.26150894e-01 8.63557458e-02 5.78353070e-02 -3.40151459e-01
6.15750253e-01 1.31403610e-01 -1.12846470e+00 -5.36538184e-01
-6.34160712e-02 2.71677554e-01 7.05755889e-01 4.58909422e-01
-7.17435539e-01 1.06378734e+00 5.46517515e+00 7.14136004e-01
-9.59683657e-01 6.02005422e-01 8.94648731e-01 -4.50757235e-01
-2.45129958e-01 8.77414197e-02 -9.13717628e-01 5.65986931e-01
1.18187344e+00 -8.32714289e-02 7.67174006e-01 1.37082839e+00
-2.55025737e-02 1.10545963e-01 -1.13116181e+00 6.29910707e-01
-4.56450552e-01 -1.25221550e+00 -2.55996227e-01 -4.19279262e-02
5.29556990e-01 7.96023726e-01 -2.01884687e-01 5.37961602e-01
1.14970815e+00 -5.50193846e-01 5.40563703e-01 -4.09712121e-02
4.47543085e-01 -6.21388674e-01 5.87618411e-01 4.85269248e-01
-9.39329088e-01 -5.46840131e-01 -4.00916994e-01 2.29714736e-01
2.60012478e-01 3.57437938e-01 -6.48989618e-01 3.38643342e-01
8.80361915e-01 1.22127086e-01 -3.68749470e-01 5.17186940e-01
1.17755622e-01 8.19531977e-01 -5.96040308e-01 2.75782257e-01
2.98785537e-01 4.70180362e-02 1.52980223e-01 1.02073848e+00
3.16834621e-06 -1.83231652e-01 1.29230887e-01 7.18482077e-01
-7.53844082e-01 2.10798845e-01 1.10027634e-01 2.42289171e-01
8.85582387e-01 1.12616026e+00 -3.48200709e-01 -2.86907107e-01
-7.68220246e-01 9.45761621e-01 9.41408634e-01 3.03848803e-01
-6.65092587e-01 -8.97662491e-02 1.10267472e+00 -4.69502397e-02
4.53816473e-01 4.06947136e-01 -3.56372565e-01 -1.02119648e+00
4.21026796e-01 -8.48876536e-01 7.84005582e-01 -4.50929284e-01
-1.51668775e+00 5.78551471e-01 -4.05669570e-01 -3.39879692e-01
-2.64960438e-01 -3.85709047e-01 -8.06957722e-01 7.93953478e-01
-1.72431374e+00 -1.11211717e+00 -2.73272246e-01 5.43734908e-01
6.84484899e-01 -2.44916394e-01 6.18827522e-01 6.18106067e-01
-7.08618879e-01 1.33769178e+00 4.34173316e-01 2.58346107e-02
6.40399277e-01 -9.47286844e-01 4.71252769e-01 5.53060710e-01
-1.22035630e-01 2.89088219e-01 2.54685789e-01 -3.82060021e-01
-1.53237379e+00 -1.25363243e+00 7.43205070e-01 -3.24611485e-01
5.03585398e-01 -3.10353667e-01 -1.19183624e+00 1.00950432e+00
-3.82047385e-01 2.54882276e-01 3.17817539e-01 3.00478220e-01
-7.70761788e-01 -4.19217646e-01 -1.76270115e+00 2.79207915e-01
1.43118107e+00 -3.17703992e-01 -5.45182936e-02 4.27079767e-01
1.14753580e+00 -1.93938881e-01 -8.29609513e-01 1.50053456e-01
4.55440104e-01 -1.12087798e+00 8.18555117e-01 -8.19580972e-01
5.01020811e-02 5.02774596e-01 -5.30507386e-01 -9.35560286e-01
-8.85876194e-02 -6.11374855e-01 -1.96531549e-01 1.60284984e+00
4.15254474e-01 -1.10026050e+00 1.52276182e+00 1.11462188e+00
-3.53786126e-02 -7.92940915e-01 -1.07506669e+00 -7.73974538e-01
3.21032912e-01 -5.94038963e-01 1.00780249e+00 6.07645035e-01
-2.56377041e-01 5.57885095e-02 5.28938510e-02 -2.05067359e-02
6.98354781e-01 1.46191716e-02 8.52973580e-01 -1.12487626e+00
-6.19197309e-01 -3.67127806e-01 1.35342836e-01 -1.41551960e+00
2.95819432e-01 -9.99266088e-01 -1.11021921e-02 -1.29083252e+00
4.48883474e-01 -1.17075109e+00 -5.00356555e-01 8.70346844e-01
-1.57396570e-01 -3.09704840e-01 5.02807617e-01 4.45078909e-01
-8.79653513e-01 3.06370348e-01 8.07140827e-01 3.15680295e-01
-3.12281787e-01 1.28649786e-01 -1.08731163e+00 5.93041956e-01
6.31377637e-01 -7.65508473e-01 -5.02244532e-01 -9.57374930e-01
-6.68237954e-02 1.49668872e-01 8.02365020e-02 -4.86168921e-01
2.48568997e-01 -1.33843139e-01 -2.20328376e-01 -7.04857633e-02
1.00334004e-01 -7.82244384e-01 -8.64519328e-02 1.67633116e-01
-4.71106142e-01 -7.29286671e-02 -1.09025806e-01 4.90190864e-01
5.80810308e-02 -3.11424881e-01 9.36959803e-01 -2.96742916e-01
-5.30360579e-01 3.24304789e-01 6.41369224e-02 4.33272779e-01
6.76146150e-01 1.34751648e-01 -5.63988447e-01 -4.43203419e-01
-5.21713912e-01 4.11575973e-01 5.50321996e-01 3.55628222e-01
4.69433405e-02 -1.00642622e+00 -5.80336690e-01 1.54050872e-01
3.88783924e-02 2.12657273e-01 3.06939453e-01 5.35053849e-01
-3.99001956e-01 2.55963057e-01 2.34544069e-01 -3.48187894e-01
-1.09885597e+00 1.96249813e-01 3.31962973e-01 -6.29866660e-01
-7.57830381e-01 1.36043692e+00 3.83670628e-01 -8.30855250e-01
7.22485125e-01 -1.46362290e-01 4.25502479e-01 -2.34447941e-01
4.19159472e-01 6.15278900e-01 3.51699650e-01 -2.73567945e-01
-2.82177985e-01 1.36645079e-01 -4.43023533e-01 3.18346806e-02
1.73696840e+00 -2.59965569e-01 1.17271259e-01 -3.02655190e-01
1.50296760e+00 -2.48474985e-01 -1.72598290e+00 -8.19924295e-01
1.26073956e-01 -4.53649491e-01 2.82820128e-02 -1.05322564e+00
-1.67543840e+00 5.64493537e-01 5.61887503e-01 -7.05625191e-02
1.18707728e+00 4.48054522e-01 1.15761828e+00 5.05312383e-01
5.13503492e-01 -1.03367627e+00 6.72987252e-02 2.74986744e-01
3.56724650e-01 -1.19562304e+00 -4.91260827e-01 -2.43195280e-01
-8.68224561e-01 7.81070709e-01 8.10091853e-01 1.16688289e-01
2.58846611e-01 4.26628053e-01 2.38108993e-01 -4.25076298e-03
-1.22856522e+00 1.43599078e-01 -6.59170151e-01 5.87641954e-01
5.11532165e-02 2.81369817e-02 1.34682789e-01 1.07432258e+00
4.07924093e-02 5.45810051e-02 8.94069374e-02 9.82594609e-01
-5.02939820e-01 -1.52802622e+00 -1.56214252e-01 5.05232096e-01
-6.97546661e-01 1.51204139e-01 -4.70007285e-02 3.90504271e-01
-1.47105411e-01 1.00384462e+00 9.57931876e-02 -1.10882618e-01
3.98992836e-01 3.47358108e-01 2.15127975e-01 -4.88152891e-01
-6.38914764e-01 -1.18286297e-01 2.68644333e-01 -1.08646345e+00
-9.55216412e-04 -6.01401210e-01 -1.64309084e+00 -6.32853627e-01
-2.62201607e-01 2.78092045e-02 9.04239714e-01 8.39552402e-01
6.25874758e-01 1.28236651e-01 8.30949366e-01 -3.71714532e-01
-1.07046723e+00 -7.42015243e-01 -2.98606873e-01 5.36086202e-01
-4.60655093e-02 -5.15553415e-01 -5.08492887e-01 -2.17042848e-01] | [5.814409255981445, 6.302374839782715] |
0b63885b-eca6-478e-9ddb-fc55a8526c63 | object-centric-anomaly-detection-by-attribute | null | null | http://openaccess.thecvf.com/content_cvpr_2013/html/Saleh_Object-Centric_Anomaly_Detection_2013_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2013/papers/Saleh_Object-Centric_Anomaly_Detection_2013_CVPR_paper.pdf | Object-Centric Anomaly Detection by Attribute-Based Reasoning | When describing images, humans tend not to talk about the obvious, but rather mention what they find interesting. We argue that abnormalities and deviations from typicalities are among the most important components that form what is worth mentioning. In this paper we introduce the abnormality detection as a recognition problem and show how to model typicalities and, consequently, meaningful deviations from prototypical properties of categories. Our model can recognize abnormalities and report the main reasons of any recognized abnormality. We also show that abnormality predictions can help image categorization. We introduce the abnormality detection dataset and show interesting results on how to reason about abnormalities. | ['Ahmed Elgammal', 'Ali Farhadi', 'Babak Saleh'] | 2013-06-01 | null | null | null | cvpr-2013-6 | ['image-categorization'] | ['computer-vision'] | [ 3.38377655e-01 1.23952843e-01 2.30589025e-02 -7.52193809e-01
-2.52574116e-01 -4.49296504e-01 5.89396179e-01 8.22937727e-01
9.32392403e-02 2.72460580e-01 2.12588772e-01 -1.98740825e-01
-3.14829528e-01 -5.93699038e-01 -6.31523967e-01 -4.63944167e-01
-1.11520410e-01 4.50970769e-01 3.23887080e-01 7.01475714e-04
6.94118023e-01 6.96392953e-01 -2.02540135e+00 6.49831116e-01
6.74824119e-01 1.00395954e+00 -4.48758572e-01 7.55751193e-01
-4.41234946e-01 7.84663796e-01 -9.02519107e-01 -8.88877809e-02
-1.42552987e-01 -5.79577088e-01 -9.98965263e-01 9.75834250e-01
6.14536524e-01 -8.10442716e-02 3.87527287e-01 1.16705441e+00
-5.05146198e-02 4.10976410e-02 1.14151835e+00 -1.59308088e+00
-9.53086913e-01 5.16827643e-01 -5.63152730e-01 6.03705406e-01
5.99875510e-01 6.35565445e-02 1.11580825e+00 -6.03591740e-01
5.56933045e-01 1.19547772e+00 3.80394727e-01 7.60954499e-01
-1.15670800e+00 -3.97383980e-03 5.57210147e-01 5.89509130e-01
-1.25746369e+00 -1.44920588e-01 4.99984920e-01 -7.53084600e-01
1.12152171e+00 8.28139007e-01 5.78812420e-01 7.66942441e-01
1.61686793e-01 8.26585174e-01 8.27668190e-01 -5.98741531e-01
1.45617113e-01 2.67156869e-01 8.10051322e-01 7.40006983e-01
7.06804752e-01 -2.80852824e-01 -9.34330001e-02 -6.28954312e-03
3.06262076e-01 1.75120771e-01 -2.45972067e-01 -2.47074440e-01
-1.27863276e+00 4.18717951e-01 2.68903613e-01 6.10105038e-01
-3.04588139e-01 1.07529335e-01 3.98811519e-01 2.25646883e-01
1.31363213e-01 6.25444293e-01 -2.85517037e-01 6.77028596e-02
-4.41783607e-01 -2.53451313e-03 4.31563824e-01 9.38333392e-01
6.14198267e-01 -1.22823149e-01 -1.58557341e-01 7.32546091e-01
-2.77615301e-02 1.95684731e-01 7.16617405e-01 -7.87268579e-01
-3.78165007e-01 1.10653913e+00 -1.37061983e-01 -1.28446186e+00
-4.72914517e-01 -2.73893386e-01 -7.52686739e-01 6.11348115e-02
2.34884515e-01 5.26346087e-01 -1.07181966e+00 1.37458813e+00
-2.88692087e-01 -1.08716726e-01 -9.02696922e-02 5.84573686e-01
9.00023758e-01 3.05095524e-01 5.77190965e-02 -2.06321418e-01
1.31138909e+00 -7.03626215e-01 -7.29473174e-01 -1.78944126e-01
8.70769858e-01 -6.22999847e-01 1.08015931e+00 5.87130368e-01
-9.39465761e-01 -6.26917839e-01 -6.99106991e-01 2.61758775e-01
-7.38601089e-01 5.72111271e-02 4.90001082e-01 4.29173082e-01
-9.82314229e-01 4.37013954e-01 -6.49522126e-01 -8.46306503e-01
1.09944023e-01 1.37272254e-01 -4.36621428e-01 2.50522554e-01
-4.35629845e-01 9.14085686e-01 7.11358607e-01 -1.44456401e-01
-2.38533229e-01 -3.31481695e-01 -7.99459815e-01 1.67488847e-02
3.55966240e-01 -4.60825086e-01 9.91575241e-01 -1.13878810e+00
-2.38838077e-01 1.36761975e+00 -5.41944027e-01 -5.23577571e-01
2.34498382e-01 1.52212292e-01 -1.02364182e+00 1.36575550e-01
3.14454399e-02 8.09126496e-01 7.38607407e-01 -1.32881725e+00
-7.09117651e-01 -7.19654560e-02 -1.33128926e-01 -2.51904130e-01
-4.26606424e-02 3.46081778e-02 -2.34391779e-01 -6.08185291e-01
7.80331552e-01 -6.51625991e-01 -3.39629734e-03 7.09974021e-02
-1.06534767e+00 -4.75648671e-01 5.56470215e-01 -4.17730778e-01
1.22486639e+00 -2.40848303e+00 -2.73699611e-01 6.92311883e-01
4.68063354e-01 -2.25660428e-01 1.25487626e-01 1.18979208e-01
-4.70866740e-01 6.86166644e-01 -2.17953354e-01 2.31357127e-01
1.49064749e-01 6.70199335e-01 -3.69748890e-01 3.79121631e-01
4.71333474e-01 6.98603988e-01 -6.59699678e-01 -5.45896411e-01
3.63557190e-01 -2.90107578e-02 -4.57900524e-01 4.10342142e-02
-1.68011129e-01 4.61103380e-01 -2.71180779e-01 8.44040871e-01
6.13860428e-01 -4.19313818e-01 -1.73949331e-01 -3.45632225e-01
1.16886079e-01 4.08111289e-02 -1.04684389e+00 4.99855429e-01
1.14309438e-01 6.37188971e-01 -7.57005334e-01 -1.31535161e+00
8.99302304e-01 -2.34004725e-02 1.44899637e-01 -4.61329073e-01
4.87598917e-03 3.49778950e-01 4.27911997e-01 -9.29373682e-01
2.56258339e-01 -4.58066240e-02 -1.37738315e-02 4.69925553e-01
-3.40821207e-01 -4.55178022e-02 4.27914768e-01 1.91244870e-01
1.07832539e+00 -6.06604815e-01 9.05632257e-01 -2.24165738e-01
9.56385732e-01 2.97652096e-01 4.40714896e-01 9.66946840e-01
-1.71846643e-01 8.48279238e-01 8.46154451e-01 -7.55417228e-01
-7.73462057e-01 -1.16281426e+00 -1.17582098e-01 7.65465736e-01
2.87746370e-01 -4.67439055e-01 -3.68129522e-01 -8.68796051e-01
2.53179055e-02 9.88320231e-01 -9.36535835e-01 -3.41270030e-01
-3.26647609e-01 -5.59661686e-01 1.54607818e-01 7.08642066e-01
2.44190812e-01 -1.21176076e+00 -6.86381459e-01 -5.39019942e-01
-1.21310838e-01 -9.12637889e-01 -9.40483734e-02 5.53073026e-02
-7.37627506e-01 -1.32859743e+00 -1.72871515e-01 -6.86992526e-01
1.47926450e+00 8.91206414e-02 1.14302933e+00 9.13404822e-01
-6.56142950e-01 7.01840699e-01 -5.11535406e-01 -4.96629983e-01
-6.43788934e-01 -6.26402259e-01 -4.83458638e-02 8.88539478e-02
6.50771797e-01 -2.68731147e-01 -2.46058375e-01 3.78729671e-01
-1.05286098e+00 -1.24969371e-01 6.12731934e-01 6.01758599e-01
8.38166475e-01 3.80034775e-01 -1.04124919e-01 -1.12370741e+00
3.45956355e-01 -2.18120262e-01 -5.54249808e-02 5.05092323e-01
-6.62166774e-01 2.79825002e-01 3.41509312e-01 -4.11011040e-01
-6.92411959e-01 -5.91869652e-02 -1.02610085e-02 -5.24123199e-02
-1.06529474e+00 1.17506847e-01 -9.60658267e-02 1.75870121e-01
6.29543841e-01 5.54788232e-01 -3.63880813e-01 -3.17318171e-01
3.06005269e-01 4.21395093e-01 5.26814461e-01 -4.59362894e-01
7.09180057e-01 5.91331780e-01 1.18902614e-02 -9.90479231e-01
-7.39266932e-01 -6.54881001e-01 -7.68938959e-01 -1.87018976e-01
6.75801218e-01 -1.91652596e-01 -4.36861366e-01 1.15954150e-02
-1.08679569e+00 2.71200627e-01 -6.16949618e-01 2.00149775e-01
-7.39823878e-01 4.77556199e-01 -2.93306410e-01 -9.50161338e-01
2.03250095e-01 -9.60658669e-01 8.80784214e-01 1.10855117e-01
-7.40723431e-01 -9.07386601e-01 -4.23572570e-01 -2.74336394e-02
3.41284543e-01 3.21529299e-01 1.41005945e+00 -1.18551993e+00
-3.78435880e-01 -2.69804180e-01 -3.63306463e-01 1.76487729e-01
4.15350854e-01 3.40414733e-01 -7.18725026e-01 1.12310864e-01
-1.07887775e-01 1.61026493e-01 9.48170066e-01 3.01563293e-01
1.91887438e+00 -4.34739769e-01 -4.88872766e-01 4.43745583e-01
1.22199547e+00 3.36608231e-01 5.26294112e-01 3.14061552e-01
3.45574141e-01 9.01620507e-01 5.43134511e-01 2.81451613e-01
-6.76818267e-02 4.90377069e-01 4.05639499e-01 -9.25614089e-02
-1.04657561e-01 6.29675984e-02 5.70838302e-02 5.46061754e-01
9.69576985e-02 -1.01158604e-01 -1.08011830e+00 7.06856847e-01
-1.67783201e+00 -9.50148165e-01 -6.80522323e-01 1.82612646e+00
3.12525809e-01 3.08129758e-01 1.21187098e-01 4.16693896e-01
1.03829765e+00 -3.48849088e-01 -4.41766798e-01 -8.37516308e-01
-3.26172024e-01 -3.47364873e-01 1.30266063e-02 2.91000754e-01
-1.09687817e+00 5.08352518e-01 7.45753002e+00 4.18416053e-01
-8.80032122e-01 -2.17929617e-01 8.00711334e-01 3.45277935e-01
-1.93303853e-01 -2.48537868e-01 -5.00785768e-01 2.68474519e-01
5.20925224e-01 -3.10381889e-01 -1.79859668e-01 8.37416470e-01
-1.21200182e-01 -3.15903664e-01 -1.65484869e+00 9.80572581e-01
4.87536192e-01 -9.76369798e-01 6.18662000e-01 -1.37490734e-01
4.28996980e-01 -7.36596346e-01 2.75451273e-01 -1.48568332e-01
-1.35699838e-01 -9.72419858e-01 6.97756410e-01 7.23441005e-01
3.41065139e-01 -3.02387565e-01 8.40488017e-01 2.31289029e-01
-9.35326874e-01 -1.85197309e-01 -5.34062207e-01 -1.85345769e-01
-1.81577474e-01 6.18539453e-01 -1.03221416e+00 1.38364062e-01
7.18191326e-01 8.73600364e-01 -1.32487190e+00 1.41790950e+00
-4.23110902e-01 3.80042791e-01 -7.26530105e-02 2.13702247e-02
-5.84018417e-02 1.49244368e-01 5.93033969e-01 1.37186098e+00
6.11521423e-01 -5.70900030e-02 -1.65394638e-02 1.09698713e+00
4.60579932e-01 1.96099207e-01 -6.56195164e-01 -4.43107449e-02
2.05662716e-02 9.20839787e-01 -1.33594334e+00 -6.46125197e-01
-4.60603654e-01 9.44398463e-01 -1.69334665e-01 6.06834292e-02
-5.33089697e-01 -2.43576393e-01 7.12747514e-01 1.60499841e-01
8.07449371e-02 3.70878428e-01 -4.09867465e-01 -1.13293540e+00
2.03731850e-01 -7.52669036e-01 7.51126647e-01 -9.55522656e-01
-1.65076578e+00 6.03030324e-01 1.61786318e-01 -1.54986179e+00
-1.51568547e-01 -9.24185216e-01 -8.67871165e-01 3.90082300e-01
-1.19386756e+00 -8.51936400e-01 -4.39286917e-01 4.59326744e-01
7.11315989e-01 1.00189261e-01 8.28058422e-01 -2.66710937e-01
-6.18228614e-01 3.38602692e-01 -3.81676018e-01 3.97367388e-01
5.17191589e-01 -1.51977313e+00 5.15686989e-01 1.06138349e+00
3.58488470e-01 8.27064455e-01 1.12251139e+00 -6.15806699e-01
-5.46059966e-01 -1.00463569e+00 1.08716047e+00 -4.89529341e-01
6.16857231e-01 1.02493711e-01 -1.28287232e+00 8.46611977e-01
-1.10931382e-01 1.74399629e-01 6.70229077e-01 5.33491932e-02
-4.92378533e-01 2.01781809e-01 -1.14937389e+00 5.93829393e-01
1.10620594e+00 -2.87984163e-01 -1.11135554e+00 4.78162110e-01
3.47653061e-01 -4.20336984e-02 -1.88356102e-01 3.11511457e-01
3.65520000e-01 -1.39025426e+00 1.00154173e+00 -1.03175175e+00
2.64323890e-01 -4.61800754e-01 -1.29107714e-01 -1.32940900e+00
-4.12029952e-01 8.99374411e-02 -3.04921493e-02 1.10508513e+00
6.79702103e-01 -7.56734073e-01 5.95702291e-01 6.22742712e-01
-5.96395358e-02 -2.89006621e-01 -4.64107662e-01 -1.13209593e+00
-1.68512046e-01 -7.04683185e-01 7.50767529e-01 8.27942550e-01
4.98020023e-01 -3.27899218e-01 7.41489530e-02 3.73692334e-01
4.87292469e-01 3.19420695e-01 4.03966248e-01 -1.38789570e+00
-1.46013439e-01 -8.30056906e-01 -1.18429315e+00 -4.91411567e-01
-1.89953372e-02 -7.30596960e-01 1.60228282e-01 -1.51850653e+00
4.14281785e-01 -9.28207412e-02 -4.90702838e-01 5.08138597e-01
-1.66611716e-01 3.64110112e-01 -3.44373770e-02 1.47780031e-01
-9.76481199e-01 -3.62359703e-01 7.57817328e-01 -4.50035334e-01
1.30710393e-01 1.28139351e-02 -9.22510207e-01 1.23754430e+00
9.78389621e-01 -2.01644808e-01 -1.78523362e-01 -2.41245896e-01
3.96999955e-01 -5.46051741e-01 4.98135567e-01 -1.12100291e+00
1.17757916e-01 -4.33606684e-01 6.13052845e-01 -6.57959998e-01
1.88255068e-02 -1.02736962e+00 2.68441159e-02 9.10603225e-01
-8.05613875e-01 3.57660353e-01 6.99679716e-04 4.44180727e-01
-3.04548234e-01 -4.53230560e-01 6.90411270e-01 -1.97610870e-01
-1.27345681e+00 -2.55457610e-01 -5.40092945e-01 -1.05473101e-01
1.19467890e+00 -4.51700956e-01 -6.07457936e-01 -4.56538826e-01
-1.36909997e+00 6.42575398e-02 6.27609670e-01 4.24909979e-01
1.00407231e+00 -1.08867681e+00 -4.48050499e-01 5.34221470e-01
8.08201134e-01 -6.17971659e-01 2.46098384e-01 1.00972629e+00
-5.57500839e-01 4.22211736e-01 -3.62238407e-01 -9.63535249e-01
-1.62391424e+00 8.33246708e-01 4.21177953e-01 3.31930518e-01
-7.86710262e-01 7.92365968e-01 4.67911124e-01 -1.00455821e-01
3.87652427e-01 -8.41790438e-01 -5.22292793e-01 -2.61653960e-01
9.23881590e-01 3.01866502e-01 1.21590599e-01 -7.96359599e-01
-7.48452485e-01 9.17027295e-01 -8.63376930e-02 3.87922227e-01
9.68235612e-01 -2.26627827e-01 -5.45547247e-01 7.44415343e-01
8.39855492e-01 -1.76481634e-01 -3.17254841e-01 -1.04006603e-01
5.38590193e-01 -6.69129729e-01 -5.46841085e-01 -1.06250000e+00
-8.70596349e-01 9.36832130e-01 6.56592131e-01 7.58245230e-01
1.43382275e+00 6.87279165e-01 2.51047909e-01 7.12213576e-01
1.45430386e-01 -7.27931321e-01 -2.78395489e-02 1.87945694e-01
1.03989768e+00 -1.19551170e+00 -3.48016396e-02 -9.41960514e-01
-7.17110097e-01 1.33857787e+00 8.92995358e-01 -1.17170706e-01
4.82747257e-01 3.86236608e-02 4.35777605e-02 -4.09295827e-01
-9.14669693e-01 -5.75885236e-01 6.47451460e-01 9.54352975e-01
3.56961340e-01 1.69415012e-01 -3.28826070e-01 4.01505798e-01
-1.60264984e-01 -6.21488690e-01 8.96180451e-01 7.53743112e-01
-7.94970810e-01 -9.14186478e-01 -7.06228495e-01 1.07130849e+00
-4.91688788e-01 2.79198200e-01 -9.25315142e-01 6.74444854e-01
6.39742434e-01 9.87364709e-01 4.51673716e-01 -4.50560898e-01
4.97287184e-01 2.54662603e-01 3.56606007e-01 -7.34885991e-01
-2.57856995e-01 -1.83809951e-01 -1.20468184e-01 -6.07358277e-01
-3.34508330e-01 -6.61615431e-01 -1.38903630e+00 -1.67275533e-01
-1.08473763e-01 2.51174480e-01 1.75920576e-01 1.13942325e+00
2.31614634e-02 5.43892324e-01 1.65301293e-01 -1.21789239e-01
-2.91729942e-02 -7.07377672e-01 -8.05978954e-01 1.07268953e+00
6.57659590e-01 -6.50725842e-01 -7.28746891e-01 3.52103114e-01] | [10.051623344421387, 1.4001904726028442] |
38ae41fd-fe8f-4f52-9006-17f050697e37 | web-image-search-engine-based-on-lsh-index | 2108.13301 | null | https://arxiv.org/abs/2108.13301v1 | https://arxiv.org/pdf/2108.13301v1.pdf | Web image search engine based on LSH index and CNN Resnet50 | To implement a good Content Based Image Retrieval (CBIR) system, it is essential to adopt efficient search methods. One way to achieve this results is by exploiting approximate search techniques. In fact, when we deal with very large collections of data, using an exact search method makes the system very slow. In this project, we adopt the Locality Sensitive Hashing (LSH) index to implement a CBIR system that allows us to perform fast similarity search on deep features. Specifically, we exploit transfer learning techniques to extract deep features from images; this phase is done using two famous Convolutional Neural Networks (CNNs) as features extractors: Resnet50 and Resnet50v2, both pre-trained on ImageNet. Then we try out several fully connected deep neural networks, built on top of both of the previously mentioned CNNs in order to fine-tuned them on our dataset. In both of previous cases, we index the features within our LSH index implementation and within a sequential scan, to better understand how much the introduction of the index affects the results. Finally, we carry out a performance analysis: we evaluate the relevance of the result set, computing the mAP (mean Average Precision) value obtained during the different experiments with respect to the number of done comparison and varying the hyper-parameter values of the LSH index. | ['Stefano Poleggi', 'Alice Nannini', 'Marco Parola'] | 2021-08-20 | null | null | null | null | ['content-based-image-retrieval'] | ['computer-vision'] | [-2.85389066e-01 -5.25324166e-01 2.72947829e-02 -2.56416708e-01
-7.20872164e-01 -3.71549815e-01 7.63855994e-01 5.95664859e-01
-1.05573487e+00 3.33615154e-01 1.16372310e-01 -4.64336425e-02
-3.57744873e-01 -1.19689262e+00 -7.29071975e-01 -5.45127988e-01
-3.16356957e-01 5.00866115e-01 7.02495873e-01 -4.76681888e-01
6.73951328e-01 7.66508520e-01 -1.89460504e+00 4.17170525e-01
1.98651791e-01 1.44282484e+00 2.09482595e-01 6.06855094e-01
-1.80183411e-01 8.33326578e-01 -5.35686433e-01 -2.79213279e-01
4.45110887e-01 5.02944402e-02 -1.05341601e+00 -8.09526920e-01
2.64992446e-01 -5.41042507e-01 -3.42795581e-01 9.32071090e-01
5.52135229e-01 3.50867897e-01 6.10906720e-01 -9.12204325e-01
-4.73425448e-01 5.76920748e-01 -2.56494731e-01 5.84501803e-01
4.65382934e-01 1.12969242e-01 9.90876377e-01 -7.77278841e-01
6.24349475e-01 1.01084173e+00 6.95385575e-01 -7.03782449e-03
-7.83474624e-01 -5.87917984e-01 -6.64640963e-01 7.03481138e-01
-1.85552192e+00 -2.66004980e-01 5.55033565e-01 -7.60898143e-02
1.02829492e+00 8.03144798e-02 6.34259939e-01 4.80908632e-01
2.83546776e-01 5.27722299e-01 7.77875781e-01 -5.45572996e-01
4.19374496e-01 1.46533340e-01 2.94925749e-01 5.25164604e-01
3.13548632e-02 8.09698701e-02 -3.19194168e-01 -2.38339052e-01
4.43059593e-01 2.77415216e-01 -2.01735925e-02 -1.69282168e-01
-8.79749835e-01 1.07023132e+00 1.27393198e+00 9.21691656e-01
-4.57613558e-01 4.65511113e-01 6.24871314e-01 5.81870258e-01
5.32439277e-02 5.21739125e-01 -2.06332713e-01 2.90197104e-01
-1.09898877e+00 2.52684981e-01 8.29835176e-01 3.61209452e-01
1.15629351e+00 -8.22313607e-01 -3.93116951e-01 7.26024628e-01
-1.54504487e-02 7.38043264e-02 9.67589080e-01 -4.71749306e-01
-1.69936847e-02 6.64409697e-01 -1.25766441e-01 -1.28058398e+00
-3.66218060e-01 -6.01768643e-02 -8.54917586e-01 1.15487784e-01
2.54202515e-01 4.41223294e-01 -7.40051687e-01 1.30108166e+00
1.23546205e-01 -5.44857532e-02 -1.82516575e-01 8.63830507e-01
6.33444250e-01 6.73459709e-01 7.68283159e-02 3.06838453e-01
1.46613264e+00 -8.84822488e-01 -2.32483357e-01 2.55308419e-01
8.02080810e-01 -9.54338729e-01 1.08891499e+00 2.03420326e-01
-7.79918492e-01 -7.70893514e-01 -1.25468969e+00 -2.87669182e-01
-1.09629667e+00 -7.30649084e-02 3.02751064e-01 1.69022456e-01
-1.43873870e+00 9.53520894e-01 -5.49121320e-01 -3.35382372e-01
1.19572513e-01 5.32183528e-01 -4.75719690e-01 -1.44688517e-01
-1.49441373e+00 9.41703439e-01 7.39686012e-01 -7.32157826e-02
-7.34206498e-01 -3.25507730e-01 -5.60440123e-01 5.57269216e-01
1.92599539e-02 -3.01112115e-01 9.25707102e-01 -8.25309634e-01
-1.09750140e+00 9.64087844e-01 2.61177987e-01 -7.39169657e-01
2.28753090e-01 -2.17991218e-01 6.88533578e-03 2.60790706e-01
-1.48020580e-01 7.54110515e-01 7.00615466e-01 -6.53312743e-01
-4.17449623e-01 -3.21376115e-01 1.36220068e-01 -7.93231875e-02
-6.33197010e-01 1.66230917e-01 -4.58138257e-01 -2.55040467e-01
-1.88617930e-01 -8.25236619e-01 -1.38649896e-01 1.20427899e-01
-9.59363952e-02 -4.85241383e-01 7.03520477e-01 -3.73358935e-01
9.65563774e-01 -2.37905979e+00 -1.39018834e-01 5.34478486e-01
4.22543474e-03 5.80566943e-01 -3.08285475e-01 6.16780162e-01
6.37810081e-02 -8.27387255e-03 9.89450067e-02 -2.10845143e-01
-5.59922047e-02 -6.98486269e-02 -1.49015710e-01 3.87045741e-01
-1.51092991e-01 8.75805080e-01 -5.49031019e-01 -6.85930789e-01
2.53297716e-01 6.33396089e-01 -5.62082946e-01 2.69797206e-01
-9.87733975e-02 -1.52695552e-01 -4.46529269e-01 2.35581666e-01
7.58699775e-01 -3.31743509e-01 -7.65438080e-02 -3.79577309e-01
-1.72244832e-01 3.30513239e-01 -1.05164123e+00 1.71319377e+00
-7.19527721e-01 6.22752190e-01 -4.85541731e-01 -9.66015041e-01
9.30477798e-01 1.20498858e-01 3.68724614e-01 -1.23181033e+00
3.22553098e-01 4.50839937e-01 -3.50658268e-01 -1.37777328e-01
6.99008763e-01 3.55689079e-01 5.60722500e-02 5.90770125e-01
1.01924263e-01 3.68743353e-02 1.39109641e-01 7.70744458e-02
1.18054330e+00 -4.96980131e-01 3.78011853e-01 -3.03760022e-01
8.49677026e-01 6.79375529e-02 -3.78419578e-01 9.07408774e-01
-9.94975567e-02 5.27327597e-01 1.45480931e-01 -1.18784523e+00
-1.12250209e+00 -7.01539338e-01 -1.78125441e-01 1.23732376e+00
1.96011111e-01 -3.01613033e-01 -6.89396977e-01 -4.38424110e-01
-1.02912620e-01 1.98643476e-01 -7.31612325e-01 -4.77179527e-01
-6.30115151e-01 -6.72694087e-01 5.72934031e-01 2.36492679e-01
9.52519476e-01 -1.42707717e+00 -1.02741885e+00 5.71936853e-02
1.15254477e-01 -8.31973493e-01 -2.13418558e-01 3.43848050e-01
-4.87801731e-01 -1.00459468e+00 -7.27826238e-01 -8.54326487e-01
2.29773596e-02 2.71649718e-01 1.10982656e+00 6.40831530e-01
-5.89922249e-01 8.80696326e-02 -7.21300066e-01 -3.04093654e-03
-2.79254228e-01 6.83298171e-01 -2.58612692e-01 -2.33137816e-01
6.63775682e-01 -4.09414679e-01 -9.52477574e-01 1.69125259e-01
-1.32604647e+00 -6.25740707e-01 6.97886288e-01 6.65871680e-01
4.76003855e-01 7.16768503e-02 -1.75560322e-02 -3.58857155e-01
7.44115114e-01 -2.96913058e-01 -8.16310108e-01 2.76987821e-01
-6.73299789e-01 4.15815830e-01 7.08411038e-01 -3.15303862e-01
-2.05769360e-01 2.27554683e-02 -4.57706273e-01 -4.66397911e-01
-4.46523055e-02 4.80052561e-01 5.73022544e-01 -4.77429986e-01
8.72749984e-01 5.00493884e-01 -6.74718022e-02 -3.60882670e-01
2.85268068e-01 8.21861625e-01 1.42551884e-01 -1.55957431e-01
5.74787438e-01 3.13631982e-01 -3.21232975e-02 -4.91365910e-01
-5.76159179e-01 -6.16157174e-01 -5.35899043e-01 2.55132943e-01
9.12390053e-01 -6.40787363e-01 -8.29512715e-01 3.30020308e-01
-1.13963771e+00 -2.77718723e-01 -1.20899968e-01 4.06949788e-01
-4.02621299e-01 1.83464244e-01 -8.03247511e-01 -2.75347859e-01
-8.98810446e-01 -1.34715199e+00 1.05269432e+00 3.37256752e-02
1.67921185e-01 -6.66717887e-01 5.89141369e-01 -1.83316275e-01
1.04526353e+00 -3.83273154e-01 8.88368428e-01 -1.09123409e+00
-6.47308528e-01 -4.61336553e-01 -6.68296337e-01 2.33194545e-01
-1.95216015e-01 -1.52733862e-01 -1.02701652e+00 -3.90946239e-01
-1.21514104e-01 -4.73883003e-01 1.12657487e+00 1.68039143e-01
1.44573653e+00 -3.63278568e-01 -2.26488814e-01 6.95222080e-01
1.87350082e+00 8.13723262e-03 9.97733295e-01 7.48020351e-01
2.56489813e-01 2.93964058e-01 4.76263165e-01 3.93770128e-01
1.75644204e-01 1.07018185e+00 5.68265915e-01 3.12802456e-02
-1.12570368e-01 -4.20726724e-02 -1.51804388e-02 5.06881714e-01
1.34252205e-01 2.04210854e-04 -7.98765719e-01 5.09456217e-01
-1.42917490e+00 -8.02598655e-01 4.74428624e-01 2.27324533e+00
7.07771301e-01 7.21369684e-02 9.04735550e-02 2.35553831e-01
4.36456054e-01 1.71225309e-01 4.80226651e-02 -5.66248536e-01
3.42931867e-01 5.34531891e-01 5.88353992e-01 3.20356458e-01
-1.14570844e+00 8.53358448e-01 5.90942669e+00 1.08738589e+00
-1.62477219e+00 1.54785156e-01 5.48657894e-01 5.54639585e-02
1.40694261e-01 -2.06845924e-01 -7.19422877e-01 5.27276933e-01
1.08959830e+00 2.09779710e-01 6.81111217e-01 1.06802821e+00
-4.76654291e-01 -1.73482984e-01 -1.02411270e+00 1.04413307e+00
-8.10952261e-02 -1.45039868e+00 1.82237878e-01 -8.70113373e-02
2.75902182e-01 4.87222344e-01 5.10498099e-02 3.16928744e-01
3.90195139e-02 -1.07029593e+00 4.19998497e-01 4.15303171e-01
5.83478391e-01 -8.10798526e-01 1.17873728e+00 2.29850113e-01
-1.27097034e+00 -2.26273373e-01 -8.00108731e-01 1.50377885e-01
-4.06808347e-01 5.30185282e-01 -7.59746850e-01 1.73400104e-01
1.17274559e+00 9.15047824e-02 -7.17959046e-01 1.27726483e+00
2.22125337e-01 -4.57244217e-02 -5.58197141e-01 -2.34984756e-01
5.02928793e-01 2.57993877e-01 6.45121858e-02 1.26532018e+00
3.92000377e-01 -3.46176475e-01 -1.74790889e-01 7.29139686e-01
-2.21675798e-01 4.10627395e-01 -7.85724163e-01 3.11569721e-01
4.21881706e-01 1.40749872e+00 -7.69581676e-01 -3.33450437e-01
-1.25528455e-01 1.03715599e+00 5.04011989e-01 -1.93492264e-01
-6.02376580e-01 -7.26464033e-01 4.01575744e-01 1.18944280e-01
5.90498149e-01 -3.99966948e-02 3.68784547e-01 -8.49153221e-01
-1.53285831e-01 -7.21523881e-01 5.39361179e-01 -6.82090878e-01
-9.77324843e-01 9.26452160e-01 -5.11579998e-02 -9.03104424e-01
-4.67328668e-01 -5.27824879e-01 -5.14645994e-01 7.75624394e-01
-1.69498825e+00 -8.19344699e-01 -4.73333269e-01 8.06834698e-01
2.19247475e-01 1.09781772e-02 8.56948495e-01 6.49825573e-01
-9.49394330e-02 8.29651177e-01 7.84072429e-02 3.54384363e-01
7.09306479e-01 -7.71138489e-01 3.59683871e-01 2.96740919e-01
3.14598322e-01 7.95123756e-01 3.57064039e-01 -8.63973871e-02
-1.35745418e+00 -8.31576645e-01 8.54544938e-01 -1.08821399e-03
5.36954522e-01 -1.21283300e-01 -9.63657916e-01 4.07987744e-01
1.36454031e-01 4.35965568e-01 1.79470658e-01 -1.71375602e-01
-6.59115493e-01 -4.64458704e-01 -1.34576666e+00 2.05074191e-01
4.60287362e-01 -7.50890613e-01 -4.37229335e-01 3.74738783e-01
7.54985154e-01 -7.61512816e-02 -9.10468817e-01 2.90536523e-01
7.64302492e-01 -1.24198592e+00 1.18242311e+00 -1.38802901e-01
3.82151544e-01 -1.36276767e-01 -4.19895947e-01 -1.03433585e+00
-2.94533789e-01 7.31789917e-02 2.87105948e-01 6.99874222e-01
1.18950374e-01 -5.93886077e-01 6.53718054e-01 1.64613500e-01
3.30280066e-01 -6.77276134e-01 -8.87361228e-01 -6.36465430e-01
-5.74853197e-02 4.42356383e-03 9.54817533e-01 6.60726666e-01
-2.03926831e-01 5.96184060e-02 -3.25997025e-02 -9.81222987e-02
2.48502612e-01 2.26022899e-01 6.10413969e-01 -1.12289810e+00
-2.50427812e-01 -5.15855610e-01 -8.53541732e-01 -5.42055666e-01
-1.67998061e-01 -7.62516379e-01 4.76549193e-02 -9.87386227e-01
3.60240638e-01 -7.09140122e-01 -7.25291669e-01 5.72739005e-01
2.03968555e-01 7.16670275e-01 2.55567133e-01 4.78189826e-01
-7.02832460e-01 2.66181052e-01 7.45459318e-01 -1.99134320e-01
-1.11255847e-01 -2.07310870e-01 -1.93316326e-01 3.11291307e-01
6.50702298e-01 -6.92401826e-01 -9.80410501e-02 -4.71565068e-01
5.54875076e-01 -2.48826385e-01 5.41077733e-01 -1.44521093e+00
6.38871491e-01 3.54817390e-01 4.23819065e-01 -6.02350235e-01
2.42640316e-01 -9.52708125e-01 6.22411352e-03 7.49874234e-01
-4.71820503e-01 3.27802747e-01 5.98052442e-02 3.16621959e-02
-5.74906290e-01 -7.37819612e-01 8.48717391e-01 -3.56243193e-01
-8.63641620e-01 2.17689171e-01 2.06669979e-02 -4.53076661e-01
7.83205330e-01 1.08506560e-01 -2.51765013e-01 -2.21613303e-01
-3.47664356e-01 -1.69751033e-01 5.74889362e-01 -2.54278257e-02
5.77397883e-01 -1.16098595e+00 -2.99538612e-01 2.38172799e-01
3.39953929e-01 -2.38137975e-01 4.83421870e-02 6.17744803e-01
-1.06601584e+00 5.97877860e-01 -5.24949491e-01 -4.43856657e-01
-9.64374602e-01 1.04113507e+00 4.57273811e-01 -5.97702563e-01
-4.27547008e-01 6.55698180e-01 -3.10607105e-01 -3.90132159e-01
2.99629092e-01 -1.89679667e-01 -3.83062035e-01 1.75008088e-01
8.95470202e-01 2.86777943e-01 5.99718034e-01 -2.61218965e-01
-4.52919334e-01 8.52433264e-01 -2.94117361e-01 1.09053943e-02
1.31126165e+00 1.34076536e-01 -3.93724978e-01 -1.07997365e-01
1.81994426e+00 -3.48337620e-01 -4.22715724e-01 -3.27242553e-01
7.30428426e-03 -2.64980614e-01 2.02246025e-01 -2.97260791e-01
-1.15894949e+00 8.56957376e-01 1.13067400e+00 4.24528062e-01
1.24106979e+00 -1.24804311e-01 9.57103193e-01 8.98918390e-01
4.86365348e-01 -8.32311988e-01 1.30609691e-01 5.47464788e-01
7.09856033e-01 -1.32113540e+00 7.82824680e-02 1.82138532e-01
-1.85845885e-02 1.22396529e+00 1.30535603e-01 -6.27457261e-01
9.27762330e-01 7.91809782e-02 -1.14295028e-01 -4.43925530e-01
-5.96909225e-01 -3.00070465e-01 2.73329735e-01 1.06504008e-01
1.58718199e-01 -1.85143828e-01 -4.50550914e-01 -1.31936431e-01
-2.11594746e-01 3.72829109e-01 -1.71298832e-02 7.86390901e-01
-5.70139349e-01 -1.07954240e+00 -2.34468207e-01 2.59909868e-01
-6.21190906e-01 -2.82442242e-01 -2.10897386e-01 7.64966965e-01
-4.15153280e-02 4.96686578e-01 3.16767693e-01 -6.01983130e-01
-1.04162376e-02 -1.16647646e-01 2.90710121e-01 -2.05659494e-01
-9.59181368e-01 -5.57004035e-01 -4.76764083e-01 -8.36549520e-01
-2.54718006e-01 -1.54664670e-03 -9.75075781e-01 -4.51952904e-01
-2.63001323e-01 8.86554644e-02 1.06815231e+00 9.35727000e-01
2.51262963e-01 4.60821316e-02 7.74234056e-01 -9.21803951e-01
-7.35805392e-01 -9.95946288e-01 -3.49478185e-01 4.90697503e-01
2.34871939e-01 -4.38349485e-01 -4.18975204e-01 -5.14877081e-01] | [10.553587913513184, 0.4427817761898041] |
f45d94d2-9869-401e-9a94-ee9d76f71dbe | old-swedish-part-of-speech-tagging-between | null | null | https://aclanthology.org/W16-2104 | https://aclanthology.org/W16-2104.pdf | Old Swedish Part-of-Speech Tagging between Variation and External Knowledge | null | ['Gerlof Bouma', 'Yvonne Adesam'] | 2016-08-01 | null | null | null | ws-2016-8 | ['morphological-tagging'] | ['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] | [-7.253198623657227, 3.805964708328247] |
4b63d9ad-991c-4198-8b1c-b63c0d777245 | 3d-concept-grounding-on-neural-fields | 2207.06403 | null | https://arxiv.org/abs/2207.06403v1 | https://arxiv.org/pdf/2207.06403v1.pdf | 3D Concept Grounding on Neural Fields | In this paper, we address the challenging problem of 3D concept grounding (i.e. segmenting and learning visual concepts) by looking at RGBD images and reasoning about paired questions and answers. Existing visual reasoning approaches typically utilize supervised methods to extract 2D segmentation masks on which concepts are grounded. In contrast, humans are capable of grounding concepts on the underlying 3D representation of images. However, traditionally inferred 3D representations (e.g., point clouds, voxelgrids, and meshes) cannot capture continuous 3D features flexibly, thus making it challenging to ground concepts to 3D regions based on the language description of the object being referred to. To address both issues, we propose to leverage the continuous, differentiable nature of neural fields to segment and learn concepts. Specifically, each 3D coordinate in a scene is represented as a high-dimensional descriptor. Concept grounding can then be performed by computing the similarity between the descriptor vector of a 3D coordinate and the vector embedding of a language concept, which enables segmentations and concept learning to be jointly learned on neural fields in a differentiable fashion. As a result, both 3D semantic and instance segmentations can emerge directly from question answering supervision using a set of defined neural operators on top of neural fields (e.g., filtering and counting). Experimental results show that our proposed framework outperforms unsupervised/language-mediated segmentation models on semantic and instance segmentation tasks, as well as outperforms existing models on the challenging 3D aware visual reasoning tasks. Furthermore, our framework can generalize well to unseen shape categories and real scans. | ['Chuang Gan', 'Joshua B. Tenenbaum', 'Chunru Lin', 'Yilun Du', 'Yining Hong'] | 2022-07-13 | null | null | null | null | ['visual-reasoning', 'visual-reasoning'] | ['computer-vision', 'reasoning'] | [ 2.28274092e-01 4.62626964e-01 -2.70925164e-01 -6.41303599e-01
-5.78724384e-01 -9.37736094e-01 7.68688500e-01 6.32283866e-01
-1.73035562e-01 8.20398480e-02 -1.12107269e-01 -4.15464789e-01
-1.71988040e-01 -1.08365679e+00 -9.60016370e-01 -4.17629868e-01
1.31581038e-01 6.65866137e-01 1.12978458e-01 2.56563425e-02
4.25236344e-01 6.97534502e-01 -1.75575376e+00 2.02619016e-01
7.78277516e-01 1.27146256e+00 1.07060000e-01 2.92912364e-01
-9.19213712e-01 4.00939196e-01 -3.85836095e-01 -1.39212489e-01
1.92619264e-01 -2.28860185e-01 -1.07685220e+00 6.65073276e-01
7.51691520e-01 -1.57443002e-01 1.67083010e-01 1.19147813e+00
-2.61716396e-01 3.69726747e-01 7.89110839e-01 -1.30655527e+00
-1.06148648e+00 8.25053006e-02 -4.70371872e-01 -2.18978509e-01
6.68489993e-01 1.84008852e-02 1.21115792e+00 -1.16252947e+00
6.71184480e-01 1.46600819e+00 3.23791593e-01 4.25606012e-01
-1.31962872e+00 -2.19645157e-01 6.69899583e-01 -1.56963930e-01
-1.35191226e+00 1.10412292e-01 1.03756070e+00 -7.11941838e-01
7.24345505e-01 1.30837724e-01 9.49706256e-01 7.92028964e-01
-2.99908131e-01 9.55764771e-01 1.08508885e+00 -2.25255653e-01
7.34766424e-01 1.23401456e-01 2.14316607e-01 9.83395576e-01
1.47463739e-01 -2.94043660e-01 -5.40926933e-01 -1.29809976e-02
1.13725483e+00 3.27388793e-01 2.07416974e-02 -8.13099146e-01
-1.38687968e+00 8.17346871e-01 9.02720094e-01 8.93899128e-02
-4.29886013e-01 2.56701827e-01 3.29666771e-02 -7.88261965e-02
4.86645013e-01 3.60966176e-01 -3.50792110e-01 4.22426879e-01
-9.42743897e-01 1.68599427e-01 7.03697443e-01 1.26286483e+00
1.41119027e+00 -2.75262862e-01 -2.47522667e-01 3.46058011e-01
5.70659757e-01 5.43659687e-01 -1.64155047e-02 -1.08717918e+00
4.04202729e-01 1.15043020e+00 -2.40023085e-03 -1.06565261e+00
-1.80583447e-01 -1.48545757e-01 -5.42120039e-01 2.33874142e-01
4.32904214e-01 3.94629300e-01 -1.31614470e+00 1.58449316e+00
6.73998475e-01 2.12874800e-01 1.02458056e-02 1.20886016e+00
1.00857961e+00 3.75994891e-01 1.44849911e-01 2.66104221e-01
1.51267719e+00 -4.77850974e-01 -3.62169832e-01 -3.47818524e-01
3.44211340e-01 -5.11903651e-02 1.27782691e+00 1.91180892e-02
-9.06894684e-01 -6.54233694e-01 -9.82568860e-01 -5.23337305e-01
-7.78700709e-01 -2.75278181e-01 6.88409448e-01 3.79900157e-01
-9.29511428e-01 2.25859180e-01 -9.95457053e-01 -5.79276443e-01
7.84245193e-01 1.69993773e-01 -3.20629090e-01 -2.41927952e-01
-8.70985508e-01 4.96088356e-01 6.13092244e-01 1.07796825e-01
-8.96623492e-01 -6.61798298e-01 -1.07714689e+00 -3.64695378e-02
6.26732886e-01 -8.83231044e-01 7.98483551e-01 -8.25345814e-01
-1.03292859e+00 1.44522381e+00 -9.66483504e-02 -3.44755858e-01
2.41659984e-01 -1.72595561e-01 1.15791783e-01 6.37693048e-01
2.77474195e-01 1.11828279e+00 9.85934436e-01 -1.67451882e+00
-5.03418088e-01 -7.15477705e-01 7.09436953e-01 3.76357138e-01
-4.44533071e-03 -7.59420991e-01 -5.48317850e-01 -1.78955153e-01
8.51134479e-01 -5.86135089e-01 -1.76607922e-01 6.23991072e-01
-7.08613992e-01 -3.43158394e-01 8.11958075e-01 -3.84528220e-01
3.55656147e-01 -2.09744596e+00 2.20927179e-01 5.08776248e-01
4.32558894e-01 -2.66042680e-01 1.56922191e-01 5.95403509e-03
2.30934918e-01 3.21515709e-01 -5.78818202e-01 -9.98937935e-02
3.05947572e-01 6.63590431e-01 -4.52813864e-01 5.34781575e-01
6.83545709e-01 1.11731923e+00 -1.14215958e+00 -7.03974843e-01
4.25423771e-01 5.02100945e-01 -7.11408734e-01 3.31838369e-01
-8.42886627e-01 6.81560636e-01 -8.79801095e-01 8.65177155e-01
5.12775660e-01 -5.16889691e-01 -8.19961503e-02 -1.57864407e-01
9.63008404e-02 1.95208833e-01 -1.14710629e+00 2.25856972e+00
-3.99142921e-01 2.87979126e-01 -5.05749956e-02 -1.42069352e+00
1.23936820e+00 1.15502886e-01 3.85698348e-01 -5.42267442e-01
-1.08284773e-02 4.20564674e-02 -6.69039726e-01 -5.26139557e-01
3.16658467e-01 -2.09706083e-01 -3.22352856e-01 3.59095663e-01
1.43412188e-01 -7.55456269e-01 -1.83955714e-01 2.92850196e-01
5.61882377e-01 4.61618274e-01 1.16631761e-01 -2.14369759e-01
4.61486191e-01 3.11914235e-01 1.88243657e-01 7.47976184e-01
7.79170841e-02 5.37344337e-01 4.70155805e-01 -2.46702582e-01
-6.92227542e-01 -1.61291432e+00 -1.44750252e-01 8.34015846e-01
6.81861699e-01 -1.63820162e-01 -7.57482409e-01 -7.04304695e-01
2.98957974e-01 7.07995296e-01 -7.31054962e-01 1.25070065e-02
-2.93285161e-01 7.58260489e-02 1.58584833e-01 5.73230982e-01
4.08664972e-01 -9.33693051e-01 -9.68821764e-01 -4.82714251e-02
9.48698074e-03 -1.34874010e+00 -7.81604275e-02 2.49127120e-01
-1.04113293e+00 -1.16982841e+00 -3.70075226e-01 -8.70805264e-01
1.10113966e+00 2.93886662e-01 1.14190018e+00 3.01927716e-01
-2.41350785e-01 1.19763672e+00 -2.66278744e-01 -2.23341048e-01
5.65205626e-02 -1.48839980e-01 -3.18319678e-01 2.09572881e-01
5.81767499e-01 -5.88823676e-01 -6.51587367e-01 -6.87061772e-02
-1.07256663e+00 1.37930825e-01 4.69912410e-01 5.14018297e-01
1.29677343e+00 -3.03400040e-01 2.33667076e-01 -8.20584118e-01
4.00956899e-01 -5.84005117e-01 -6.30231082e-01 4.83229458e-01
-2.47552961e-01 3.39757770e-01 3.54121000e-01 -3.19157630e-01
-9.28392291e-01 1.33369207e-01 2.59937912e-01 -7.09518075e-01
-7.97961593e-01 5.80314279e-01 -1.53451085e-01 3.02631944e-01
5.03500879e-01 1.40392676e-01 -9.64107811e-02 -4.62334394e-01
1.08875263e+00 3.18435758e-01 6.29517436e-01 -1.02639508e+00
9.66557860e-01 9.46590662e-01 2.06663430e-01 -7.93803394e-01
-1.14409494e+00 -6.98389471e-01 -1.00629854e+00 -1.51670158e-01
1.54985285e+00 -8.08359861e-01 -5.75862885e-01 -1.34115711e-01
-1.20774424e+00 -3.48603427e-02 -4.83492225e-01 2.18777061e-01
-8.77162278e-01 1.39658689e-01 -1.02306746e-01 -8.49079013e-01
5.27949557e-02 -8.99557948e-01 1.64071620e+00 2.59431362e-01
-4.40263227e-02 -1.10169804e+00 -4.51309383e-01 4.06509012e-01
-1.46587610e-01 6.58881366e-01 1.35381639e+00 -4.90649104e-01
-9.39615548e-01 -2.17092447e-02 -4.45878834e-01 1.67932898e-01
8.12235251e-02 -3.12840670e-01 -9.95278656e-01 1.51512874e-02
1.37446653e-02 -5.52292049e-01 6.49185061e-01 1.87580943e-01
1.54619491e+00 -1.25327229e-01 -3.32378656e-01 5.44421434e-01
1.43859410e+00 -5.94507195e-02 1.59085199e-01 -1.04545571e-01
9.14686799e-01 8.98925483e-01 5.63794613e-01 2.00581372e-01
5.39433002e-01 2.46365562e-01 8.38319600e-01 -3.11980337e-01
3.17337201e-03 -6.54182136e-01 -2.75976241e-01 4.52403963e-01
1.16717830e-01 1.51070356e-01 -1.12481201e+00 5.93040824e-01
-1.64583874e+00 -6.15108967e-01 -2.26897299e-02 2.03223562e+00
6.94739163e-01 1.65737391e-01 -2.14073747e-01 -2.98991110e-02
6.46398604e-01 1.22431181e-02 -9.57110763e-01 -2.12972417e-01
1.39950991e-01 3.32718134e-01 1.02415808e-01 4.16268289e-01
-9.45553780e-01 9.78701472e-01 4.81424093e+00 2.66666234e-01
-9.46419477e-01 -4.86810021e-02 5.04518688e-01 2.13129297e-01
-7.13914216e-01 2.57585764e-01 -5.02008259e-01 -1.12784654e-01
2.91725010e-01 6.77977204e-02 3.35969955e-01 7.83943057e-01
-1.38082549e-01 -2.46732198e-02 -1.64274764e+00 1.07765019e+00
5.74591495e-02 -1.32333374e+00 4.24215615e-01 7.64799565e-02
6.11025393e-01 -4.48606014e-01 6.73949346e-02 2.54364908e-01
2.11764559e-01 -1.27401638e+00 1.04525363e+00 7.11594284e-01
7.94873714e-01 -4.13075507e-01 9.43941027e-02 4.10755575e-01
-1.26250553e+00 7.02523142e-02 -2.85719097e-01 5.75650483e-02
5.88755943e-02 4.34829205e-01 -8.23325038e-01 6.53175950e-01
7.37514734e-01 7.81287849e-01 -4.57178056e-01 6.92321658e-01
-4.02864814e-01 3.81569862e-01 -3.23977977e-01 1.65894609e-02
5.53460538e-01 -3.94744217e-01 2.15911940e-01 8.42408359e-01
1.77291378e-01 2.90653855e-01 4.93018597e-01 1.66865325e+00
-1.68927817e-03 -1.16969578e-01 -5.63246608e-01 -8.79536942e-02
3.90726388e-01 1.06829858e+00 -1.14430141e+00 -3.48684311e-01
-5.17997026e-01 8.50001454e-01 4.90670294e-01 7.74454415e-01
-6.21284366e-01 -1.95009723e-01 7.58428991e-01 7.86155835e-02
4.47183222e-01 -5.63598633e-01 -4.00061756e-01 -1.05758977e+00
5.41258678e-02 -2.97240347e-01 3.16247135e-01 -1.01683187e+00
-1.42075646e+00 3.09120923e-01 2.58639395e-01 -1.16889238e+00
-4.99756150e-02 -8.69119585e-01 -3.22330803e-01 8.43574882e-01
-1.55019319e+00 -1.22860956e+00 -5.18455684e-01 7.95040905e-01
3.68349373e-01 4.47680652e-01 7.71039903e-01 -2.86311239e-01
7.74764493e-02 -6.10636845e-02 -4.87805396e-01 2.83989280e-01
-1.28513835e-02 -1.46344197e+00 2.89080709e-01 4.27688092e-01
6.96046650e-01 8.26909423e-01 2.61712819e-01 -4.04988825e-01
-1.54663861e+00 -1.09835839e+00 4.01961178e-01 -6.39158070e-01
5.16090333e-01 -8.82281482e-01 -1.19481492e+00 5.56366265e-01
-2.64584690e-01 3.03174734e-01 5.36420226e-01 9.51990113e-02
-7.19499946e-01 1.74184427e-01 -1.19638121e+00 5.06175518e-01
1.37751341e+00 -1.04562986e+00 -9.48552191e-01 3.99588078e-01
9.55196977e-01 -5.26947379e-01 -9.32087779e-01 7.19682276e-02
1.37693375e-01 -7.64782965e-01 1.31106007e+00 -9.18245196e-01
2.97572315e-01 -4.60537076e-01 -5.29678464e-01 -8.08973014e-01
1.84811041e-01 -4.72362041e-02 -1.36626795e-01 9.64435458e-01
2.00050592e-01 -3.16832960e-01 7.71739721e-01 6.64813519e-01
-5.01699448e-02 -8.01083207e-01 -9.94779527e-01 -5.62720120e-01
2.04515025e-01 -6.97873116e-01 7.69438446e-01 9.98027444e-01
-3.48640442e-01 3.26728255e-01 4.60925192e-01 5.32360971e-01
8.54885697e-01 7.07124829e-01 4.99188542e-01 -1.53302538e+00
-2.00917218e-02 -3.24315667e-01 -7.06155241e-01 -1.45440221e+00
4.20461297e-01 -1.28408325e+00 1.54419273e-01 -1.96778214e+00
-1.77136764e-01 -5.49325049e-01 -1.18140630e-01 5.08307397e-01
1.04825776e-02 -3.26991384e-03 1.27217382e-01 1.47623479e-01
-6.48589134e-01 5.80599487e-01 1.50124013e+00 -4.45268422e-01
-9.53825489e-02 -4.03154850e-01 -6.39634669e-01 6.63164496e-01
6.38773263e-01 -2.88318485e-01 -6.82991326e-01 -6.68681264e-01
3.14440042e-01 1.85062259e-01 1.08155322e+00 -6.89461589e-01
2.39590704e-01 -3.74815613e-01 4.41697508e-01 -6.89750731e-01
4.86558229e-01 -9.98485386e-01 -3.96244168e-01 -1.25048263e-02
-4.20599699e-01 -3.47660959e-01 1.48004564e-02 8.67362559e-01
-2.23583713e-01 -6.12684675e-02 2.44817317e-01 -4.85670388e-01
-9.82865810e-01 4.89691705e-01 9.91828516e-02 3.52215111e-01
8.76895130e-01 -5.12567580e-01 1.24574630e-02 -1.12719141e-01
-8.55401814e-01 3.55480939e-01 4.72792774e-01 5.38057804e-01
9.44919646e-01 -1.35779226e+00 -1.54911637e-01 3.24724853e-01
4.39099997e-01 8.34779561e-01 3.91448252e-02 5.05262315e-01
-2.52595037e-01 3.57123047e-01 -9.70367715e-02 -1.23301780e+00
-5.48954189e-01 7.07249999e-01 3.73310804e-01 4.55331028e-01
-7.76215613e-01 8.17538261e-01 5.99023581e-01 -6.98748827e-01
3.73015493e-01 -8.58096838e-01 -1.70936838e-01 -5.09060221e-03
7.43304342e-02 -2.38882557e-01 -1.11467898e-01 -7.10367084e-01
-4.34200972e-01 9.08785224e-01 2.91213959e-01 -1.30025357e-01
1.03555608e+00 -8.62089023e-02 -1.32694796e-01 7.28154898e-01
1.34496593e+00 -4.63455111e-01 -1.44388437e+00 -4.17115599e-01
2.76903242e-01 -4.26667303e-01 -2.81561036e-02 -4.59092796e-01
-9.49990332e-01 1.20513642e+00 3.22777748e-01 3.77196789e-01
8.93870115e-01 7.59595573e-01 4.28515524e-01 5.10250390e-01
4.78714734e-01 -8.97650480e-01 4.92559075e-01 2.28087619e-01
8.31602216e-01 -1.22599030e+00 -1.79298550e-01 -6.44228935e-01
-4.29681003e-01 1.10315204e+00 6.25317216e-01 -1.68914989e-01
7.16170490e-01 -3.88919443e-01 2.03872770e-02 -8.84011209e-01
-2.69489020e-01 -4.15389717e-01 6.22353017e-01 7.86359251e-01
1.21198930e-01 1.77210733e-01 3.59591663e-01 4.01028931e-01
-1.38323843e-01 -3.53992760e-01 9.47525203e-02 1.03613353e+00
-5.62093496e-01 -6.05646908e-01 -5.05446792e-01 4.09514517e-01
2.29860857e-01 1.24270387e-01 -4.75808859e-01 7.69969225e-01
3.26963156e-01 8.60679865e-01 5.27368486e-01 -4.25846968e-03
3.28898549e-01 1.25059590e-01 5.61927676e-01 -1.08318484e+00
-7.20641837e-02 -3.28633368e-01 -4.78967041e-01 -6.69255674e-01
-8.06035757e-01 -5.45811713e-01 -1.95788836e+00 3.41691166e-01
-1.11410366e-02 5.57933599e-02 7.44582355e-01 1.15933526e+00
1.30008668e-01 2.60526329e-01 2.61706710e-01 -7.29494393e-01
-4.19255048e-01 -3.16699117e-01 -6.03320599e-01 8.60161841e-01
3.82159889e-01 -9.65475917e-01 -3.73527110e-01 2.81126410e-01] | [8.04708480834961, -3.2346935272216797] |
d8f8baa5-42e2-4606-8266-22162f1294d4 | hierarchical-spatial-aware-siamese-network | 1711.09539 | null | http://arxiv.org/abs/1711.09539v2 | http://arxiv.org/pdf/1711.09539v2.pdf | Hierarchical Spatial-aware Siamese Network for Thermal Infrared Object Tracking | Most thermal infrared (TIR) tracking methods are discriminative, treating the
tracking problem as a classification task. However, the objective of the
classifier (label prediction) is not coupled to the objective of the tracker
(location estimation). The classification task focuses on the between-class
difference of the arbitrary objects, while the tracking task mainly deals with
the within-class difference of the same objects. In this paper, we cast the TIR
tracking problem as a similarity verification task, which is coupled well to
the objective of the tracking task. We propose a TIR tracker via a Hierarchical
Spatial-aware Siamese Convolutional Neural Network (CNN), named HSSNet. To
obtain both spatial and semantic features of the TIR object, we design a
Siamese CNN that coalesces the multiple hierarchical convolutional layers.
Then, we propose a spatial-aware network to enhance the discriminative ability
of the coalesced hierarchical feature. Subsequently, we train this network end
to end on a large visible video detection dataset to learn the similarity
between paired objects before we transfer the network into the TIR domain.
Next, this pre-trained Siamese network is used to evaluate the similarity
between the target template and target candidates. Finally, we locate the
candidate that is most similar to the tracked target. Extensive experimental
results on the benchmarks VOT-TIR 2015 and VOT-TIR 2016 show that our proposed
method achieves favourable performance compared to the state-of-the-art
methods. | ['Nana Fan', 'Hongzhi Wang', 'Qiao Liu', 'Zhenyu He', 'Xin Li'] | 2017-11-27 | null | null | null | null | ['thermal-infrared-object-tracking'] | ['computer-vision'] | [ 9.35713351e-02 -4.84240770e-01 -9.72992256e-02 -2.41264924e-01
-7.68855691e-01 -5.54451585e-01 5.21914423e-01 -4.92865503e-01
-4.89962488e-01 1.02500573e-01 -1.83391467e-01 5.11680404e-03
5.19756973e-02 -2.97418743e-01 -6.04609191e-01 -1.13423777e+00
2.83239394e-01 2.04780445e-01 7.17532396e-01 2.34268889e-01
4.10263166e-02 4.74215686e-01 -1.42135811e+00 8.17616284e-02
6.93050981e-01 1.64013553e+00 3.21033865e-01 2.65192777e-01
2.29823157e-01 7.41846263e-01 -4.54229921e-01 2.96521261e-02
4.24811929e-01 -1.72426984e-01 -3.56064796e-01 -4.21294495e-02
1.01544750e+00 -1.48460835e-01 -3.90911281e-01 1.01806962e+00
2.79857755e-01 3.86734158e-01 6.01642072e-01 -1.40084600e+00
-3.32651079e-01 5.28748482e-02 -6.48305714e-01 3.71959120e-01
-1.39495909e-01 1.27153456e-01 7.58921564e-01 -1.02071071e+00
2.62916028e-01 1.18866622e+00 7.57889450e-01 3.83876532e-01
-8.21539044e-01 -1.00840878e+00 2.79326528e-01 2.88526565e-01
-1.44703102e+00 -3.62416029e-01 8.44165623e-01 -6.72127783e-01
1.76240981e-01 1.04833297e-01 3.55363637e-01 8.54559660e-01
4.44347151e-02 9.18045461e-01 4.91720974e-01 1.06468327e-01
3.34280878e-02 -1.29287522e-02 8.88170600e-02 5.93224943e-01
4.41795513e-02 4.33549196e-01 -1.37899727e-01 6.68897331e-02
2.61119783e-01 2.16612652e-01 -1.94056302e-01 -5.57047367e-01
-1.20585644e+00 5.40599346e-01 7.65760601e-01 3.97173703e-01
-1.15977891e-01 3.73833090e-01 3.75601113e-01 -1.77778304e-01
5.27615070e-01 -1.96120013e-02 -3.40249240e-01 2.75946051e-01
-1.00802433e+00 4.48965915e-02 1.46999463e-01 8.74948978e-01
6.72272742e-01 -3.59719694e-02 -6.39516592e-01 5.18088996e-01
5.98550677e-01 7.47953296e-01 8.01791847e-02 -5.68720162e-01
4.34585780e-01 5.42033732e-01 2.42547542e-01 -1.03852999e+00
-2.06868663e-01 -8.94575119e-01 -6.28760874e-01 2.58639097e-01
5.78405023e-01 -4.20313962e-02 -8.71440411e-01 1.63717580e+00
7.91818321e-01 4.68693674e-01 -5.09467348e-02 1.32424581e+00
8.12489212e-01 6.70112371e-01 1.44912526e-01 -7.94980600e-02
1.44467080e+00 -1.20702720e+00 -4.43436056e-01 -3.92185599e-01
7.36144483e-01 -8.85108769e-01 3.60493511e-01 -4.79723923e-02
-5.54809153e-01 -9.66159821e-01 -7.87912309e-01 1.40658125e-01
-2.52190709e-01 8.67822826e-01 6.41698912e-02 2.96921253e-01
-6.80358529e-01 2.71538377e-01 -5.55382073e-01 -3.04617494e-01
4.84965891e-01 3.21491331e-01 -2.17559159e-01 -1.20498367e-01
-1.03385568e+00 7.64715254e-01 7.17539847e-01 5.81774056e-01
-8.65200102e-01 -5.97441673e-01 -8.76197040e-01 -9.80836675e-02
6.82999372e-01 -5.14856815e-01 1.05301011e+00 -1.08036041e+00
-1.17344677e+00 5.88931859e-01 -1.29368722e-01 -1.85243636e-01
4.64267373e-01 -8.19275603e-02 -4.89567369e-01 -1.19160049e-01
3.99724901e-01 5.13751268e-01 1.08883774e+00 -1.35442734e+00
-1.22456515e+00 -2.42009789e-01 -1.54799625e-01 8.90005380e-02
-1.38103589e-01 2.19612345e-01 -6.77802980e-01 -4.68679368e-01
3.77312079e-02 -1.15940845e+00 3.09560388e-01 2.71298498e-01
-4.22100544e-01 -7.90210128e-01 1.61814952e+00 -5.08318484e-01
9.67898726e-01 -2.55407548e+00 -7.25758821e-02 -9.29124430e-02
1.70481950e-01 4.76654798e-01 -2.14265257e-01 -2.26792041e-02
-1.41066015e-01 -4.56781775e-01 1.13134876e-01 -5.11381388e-01
4.93550673e-02 -2.84651101e-01 -2.10954830e-01 8.47319841e-01
5.83706833e-02 8.53865981e-01 -1.04231930e+00 -7.37749338e-01
3.55178833e-01 2.88241625e-01 -1.83853861e-02 3.87900531e-01
-1.82848394e-01 6.85662091e-01 -7.94212937e-01 6.10798419e-01
9.10552561e-01 -1.29912242e-01 -1.95572034e-01 -8.74844253e-01
-6.08532965e-01 -1.40829578e-01 -8.60513270e-01 1.32957506e+00
-2.49257579e-01 6.20509624e-01 -1.85056269e-01 -8.65323961e-01
1.06689501e+00 1.37397496e-03 6.96607590e-01 -6.03451848e-01
2.21806273e-01 1.52436152e-01 9.70311761e-02 -4.20722008e-01
4.03678089e-01 1.14208691e-01 6.43089339e-02 1.54784903e-01
-1.95312873e-01 4.10832614e-01 -2.15428062e-02 -2.02382635e-02
8.23404431e-01 4.20633048e-01 -3.44417959e-01 -5.55079766e-02
7.87479520e-01 1.47304848e-01 9.36422169e-01 6.14684939e-01
-3.76221925e-01 4.54317540e-01 -6.50642663e-02 -5.72163999e-01
-8.95844758e-01 -1.04687142e+00 2.27890350e-02 1.14600527e+00
6.70835137e-01 8.35343078e-02 -3.46096039e-01 -1.11880767e+00
1.05836410e-02 3.18413526e-01 -7.66723871e-01 -2.76982188e-01
-6.03623986e-01 -3.11165065e-01 3.62864524e-01 7.22271502e-01
7.51375556e-01 -8.88383448e-01 -7.06230640e-01 -7.06250742e-02
-2.31201097e-01 -1.43189549e+00 -9.87697482e-01 2.21237559e-02
-4.86789882e-01 -1.02859318e+00 -6.33049190e-01 -1.02209473e+00
3.49627316e-01 6.65223956e-01 5.15015602e-01 6.42884872e-04
-1.54270917e-01 1.49667084e-01 -2.25183442e-01 -1.47913337e-01
2.07133815e-01 5.61136417e-02 9.76423770e-02 5.80397785e-01
3.22499931e-01 7.59532750e-02 -8.35524917e-01 6.65317655e-01
-6.44853175e-01 -1.12804241e-01 7.67788112e-01 6.62893236e-01
4.71104324e-01 1.46943867e-01 1.65125817e-01 -1.16379932e-01
-2.28423640e-01 -3.02290678e-01 -8.73664856e-01 3.77117693e-01
5.48056792e-03 -5.84659316e-02 5.84782720e-01 -5.74235559e-01
-1.01934004e+00 7.02739298e-01 3.48545283e-01 -1.08914590e+00
-1.37949780e-01 2.94711441e-01 -2.24505037e-01 -3.09668899e-01
1.63797766e-01 4.77095962e-01 -2.09635496e-01 -4.57029104e-01
2.35764757e-01 5.88452697e-01 7.59262323e-01 -2.67358869e-01
1.49483895e+00 6.46985054e-01 7.97312036e-02 -7.76821971e-01
-1.31290483e+00 -8.78824413e-01 -7.22720563e-01 -5.77427447e-01
1.34101343e+00 -9.22296286e-01 -9.41102147e-01 3.57662022e-01
-1.20468247e+00 -1.35167256e-01 -4.66634845e-03 6.81344509e-01
-2.87931472e-01 4.16349888e-01 -5.85466400e-02 -9.38640356e-01
-3.58664125e-01 -1.10357964e+00 1.70727217e+00 5.23330271e-01
3.54717493e-01 -8.66118312e-01 2.35172778e-01 5.36009133e-01
1.89586610e-01 2.67638266e-01 4.13736671e-01 -4.81586576e-01
-8.22098851e-01 -5.08659959e-01 -6.14710689e-01 2.63484299e-01
3.46146263e-02 9.67113078e-02 -1.03488386e+00 -4.86944556e-01
-2.30521679e-01 -2.74069041e-01 1.03026080e+00 4.29603100e-01
1.11819279e+00 6.04889728e-02 -8.46134901e-01 7.84287691e-01
1.25671518e+00 2.96393275e-01 4.26488519e-01 2.88675040e-01
1.16932130e+00 5.83198726e-01 1.34543061e+00 4.44866717e-02
2.05089360e-01 1.26653862e+00 6.08752847e-01 -1.66298896e-01
-1.52211459e-02 -1.01859309e-01 5.27725637e-01 4.19778347e-01
1.32240206e-01 -1.25110865e-01 -7.41701484e-01 5.19325435e-01
-2.17091155e+00 -1.06833410e+00 -2.40026027e-01 2.19735026e+00
1.94901481e-01 -1.92412660e-02 2.93488353e-01 -3.87895167e-01
1.05957949e+00 2.28907466e-01 -4.85800147e-01 3.31659138e-01
1.69553190e-01 -4.11900371e-01 5.85973144e-01 -1.81631523e-03
-1.62345529e+00 1.10526001e+00 4.57807970e+00 1.25275469e+00
-1.27877450e+00 1.23508640e-01 1.73047647e-01 6.75669611e-02
4.17062521e-01 1.37775894e-02 -1.04981017e+00 6.42007172e-01
5.03576338e-01 1.48153290e-01 4.23805900e-02 8.94293070e-01
3.38524640e-01 2.10214406e-03 -1.17040741e+00 8.91341448e-01
6.18511401e-02 -9.37017679e-01 -4.32700157e-01 3.38875083e-03
5.44743121e-01 1.60034761e-01 3.37865382e-01 3.39417458e-01
-7.84660280e-02 -7.71090090e-01 1.00039995e+00 7.07608402e-01
6.82047963e-01 -6.45118535e-01 7.69664109e-01 3.58338416e-01
-2.04974031e+00 -1.90555036e-01 -5.16946733e-01 5.88019550e-01
-7.50991106e-02 3.35552603e-01 -4.46137875e-01 8.73104095e-01
9.57882345e-01 1.09156239e+00 -6.22481287e-01 1.49028242e+00
-7.55451471e-02 2.55907863e-01 -1.55711174e-01 1.84726760e-01
4.47979003e-01 -1.86468959e-01 5.68566084e-01 9.89349902e-01
3.63986880e-01 -2.23373294e-01 6.88574374e-01 9.48122799e-01
1.54858515e-01 -2.29676381e-01 -5.50007343e-01 1.18503414e-01
5.05454361e-01 1.67380333e+00 -5.01771569e-01 -2.92828441e-01
-3.05080235e-01 7.71731317e-01 2.86768377e-01 3.46446216e-01
-1.23478723e+00 -4.20116723e-01 6.59587622e-01 -2.05825746e-01
7.99839735e-01 -1.85836419e-01 2.26480842e-01 -8.95640790e-01
9.09710452e-02 -5.78042686e-01 4.17882323e-01 -1.00364327e+00
-1.28835762e+00 6.17243171e-01 -1.95741504e-01 -1.89075387e+00
1.55164853e-01 -5.62997043e-01 -9.79516089e-01 9.30562258e-01
-1.47602844e+00 -1.57558179e+00 -7.68423676e-01 6.01980805e-01
3.14373404e-01 8.81345421e-02 1.32419253e-02 4.64657605e-01
-8.54850531e-01 6.27655327e-01 1.91924095e-01 5.04129231e-01
7.35192060e-01 -8.29468012e-01 7.06829503e-02 9.76948261e-01
-2.31776372e-01 2.26990566e-01 3.50515991e-01 -7.93250263e-01
-1.31091881e+00 -1.80189419e+00 7.34256148e-01 -5.46231687e-01
8.20093989e-01 -4.25169498e-01 -8.39993536e-01 4.65145081e-01
-1.47381142e-01 4.98616248e-01 7.32630212e-03 -3.62285793e-01
-3.07386428e-01 -3.43339324e-01 -7.46039271e-01 3.54260921e-01
1.02440464e+00 -6.06286347e-01 -4.24145520e-01 4.67827171e-01
8.48183990e-01 -4.11415726e-01 -8.09517384e-01 5.27862549e-01
6.86365068e-01 -6.04981005e-01 1.14058185e+00 -2.68660665e-01
-4.16179821e-02 -9.59819973e-01 9.85414609e-02 -9.85327542e-01
-4.22698885e-01 -1.14181876e-01 3.30271432e-03 1.25770211e+00
-6.67829737e-02 -4.01268393e-01 8.57373655e-01 1.39902532e-01
-3.90493393e-01 -4.77715969e-01 -9.18215215e-01 -1.26260054e+00
-2.97791064e-01 -1.83319360e-01 2.33335540e-01 7.56092906e-01
-5.28251469e-01 3.38628560e-01 -4.80900198e-01 5.65145969e-01
1.03367043e+00 5.05322039e-01 7.13697135e-01 -1.31895673e+00
-3.87334861e-02 -2.30471313e-01 -3.97409827e-01 -1.20589137e+00
3.37885618e-01 -8.98329437e-01 6.89029574e-01 -1.25297797e+00
2.87831575e-01 -5.67528546e-01 -3.32759053e-01 3.36388052e-01
-3.53825748e-01 3.45206320e-01 3.43591481e-01 3.52541357e-01
-1.08947122e+00 8.86813402e-01 1.17788744e+00 -5.34718037e-01
-2.93842833e-02 2.08184674e-01 -1.31221399e-01 4.30830300e-01
5.25134146e-01 -7.39175558e-01 -1.23030499e-01 -3.93915176e-01
-3.07528406e-01 -8.54360014e-02 8.20824265e-01 -1.34950519e+00
6.85258567e-01 -1.89077422e-01 4.39552397e-01 -1.25865901e+00
3.10031772e-01 -1.16401958e+00 1.49268284e-01 5.99543452e-01
-2.15199307e-01 -2.43955433e-01 1.18571846e-02 7.49424815e-01
-1.61187917e-01 -1.22396506e-01 1.00089979e+00 4.28606898e-01
-8.95088136e-01 8.47004771e-01 -3.62896919e-02 1.11041823e-02
1.25065660e+00 -2.67308950e-01 -6.24714077e-01 7.47668669e-02
-3.81983370e-01 4.93585646e-01 3.26583475e-01 5.85697174e-01
4.49088097e-01 -1.51835072e+00 -4.62959021e-01 -2.31713295e-01
5.22602856e-01 6.48014843e-02 3.10026497e-01 1.35612726e+00
5.72161712e-02 4.86850858e-01 6.07170500e-02 -1.10011721e+00
-1.53863311e+00 8.56993258e-01 6.71615481e-01 3.58910486e-02
-5.09586334e-01 7.29743063e-01 7.55886197e-01 -1.77312180e-01
4.79758590e-01 -2.54410803e-01 -4.54951733e-01 -2.27645282e-02
3.75243276e-01 2.52281696e-01 -3.38051975e-01 -1.05652535e+00
-8.04879129e-01 1.09423292e+00 1.13707587e-01 4.36700046e-01
9.47997212e-01 -1.07283972e-01 3.69527601e-02 2.01633036e-01
1.39553404e+00 -4.15070236e-01 -1.46203053e+00 -6.03605807e-01
6.45091161e-02 -4.59829003e-01 2.30100483e-01 -4.46768343e-01
-1.40872085e+00 7.74022758e-01 1.04247463e+00 -5.68400212e-02
1.03567398e+00 1.10582530e-01 8.94502223e-01 2.91106641e-01
9.03699100e-02 -9.76026535e-01 3.63690525e-01 6.18403435e-01
4.31884080e-01 -1.31074655e+00 -2.84971982e-01 -3.96593034e-01
-5.60035765e-01 1.01819909e+00 9.00985420e-01 -5.36539890e-02
4.45868492e-01 -7.83925131e-02 1.02795653e-01 -2.43653223e-01
-4.97797519e-01 -6.50799811e-01 8.35153937e-01 5.85790515e-01
1.48333415e-01 -2.47869432e-01 1.37096673e-01 1.19400717e-01
3.46117079e-01 -4.42110747e-03 -2.49537334e-01 6.68464065e-01
-6.47375166e-01 -6.21211171e-01 -5.48038840e-01 1.43667921e-01
-6.36329949e-02 5.02408966e-02 -4.02373374e-01 6.26074612e-01
5.52419722e-01 9.64886725e-01 2.36351714e-01 -8.29082251e-01
2.77419448e-01 -2.52683848e-01 1.44966692e-01 -4.39178616e-01
-5.85823834e-01 1.87049955e-01 -9.38874409e-02 -6.64470255e-01
-5.50329804e-01 -7.63432324e-01 -1.00127387e+00 -1.50187746e-01
-5.86097717e-01 1.82270184e-01 5.26659846e-01 1.11147606e+00
8.89824405e-02 6.49991274e-01 9.84310687e-01 -9.58490133e-01
-5.26613474e-01 -6.97014809e-01 -4.20565784e-01 2.76535928e-01
4.22745824e-01 -9.84391034e-01 -2.93762207e-01 -2.29129225e-01] | [6.33281135559082, -2.192270040512085] |
4d2db566-1bbe-4b41-97cd-835608dde280 | utility-based-evaluation-metrics-for-models | null | null | https://aclanthology.org/W15-1108 | https://aclanthology.org/W15-1108.pdf | Utility-based evaluation metrics for models of language acquisition: A look at speech segmentation | null | ['Lawrence Phillips', 'Lisa Pearl'] | 2015-06-01 | null | null | null | ws-2015-6 | ['scene-labeling'] | ['computer-vision'] | [-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] | [-7.341408729553223, 3.7981784343719482] |
7a35483d-b750-4611-b3ef-56c33bba3e3f | a-job-assignment-heuristic-for-lifelong-multi | 2003.07108 | null | http://arxiv.org/abs/2003.07108v2 | http://arxiv.org/pdf/2003.07108v2.pdf | A Job-Assignment Heuristic for Lifelong Multi-Agent Path Finding Problem with Multiple Delivery Locations | In this paper we proposed multiple job-assignment heuristics to generate
low-total-cost solutions and determine the best performing method amongst them. | ['Polat Faruk', 'Semiz Fatih'] | 2020-04-25 | null | null | null | null | ['multi-agent-path-finding'] | ['playing-games'] | [-1.09989643e-01 9.16506350e-02 -2.52648443e-01 -3.81923735e-01
-3.80747199e-01 -3.93217683e-01 -1.11827835e-01 4.21671383e-02
-3.28008771e-01 1.61871719e+00 -3.94792259e-01 -2.45096564e-01
-8.61123502e-01 -7.26511359e-01 -1.14123404e-01 -6.77130997e-01
-1.02752440e-01 1.65083647e+00 2.17419818e-01 -2.07201466e-01
7.70622194e-01 9.22522426e-01 -1.43251657e+00 8.33221823e-02
9.82975841e-01 2.65615106e-01 1.09143841e+00 6.83757722e-01
-5.02307892e-01 4.41055268e-01 -9.53848600e-01 -1.49888828e-01
5.63509226e-01 -3.82626891e-01 -1.33578730e+00 4.95516360e-01
-8.24619472e-01 4.36336517e-01 4.31187242e-01 4.85919088e-01
6.04868293e-01 4.58135307e-01 4.65548456e-01 -1.67303157e+00
-1.20230369e-01 5.63556850e-01 -9.14358377e-01 5.98704100e-01
6.97106838e-01 -2.61436224e-01 6.35008454e-01 -3.49843770e-01
9.80413377e-01 9.92718816e-01 2.29784161e-01 2.52191871e-01
-1.55302334e+00 -3.93973917e-01 -3.50270152e-01 1.99426219e-01
-1.55774355e+00 -3.31662223e-02 5.22797287e-01 3.57269831e-02
1.73900914e+00 7.71924436e-01 5.58269083e-01 8.34434181e-02
6.09459281e-01 3.46541762e-01 1.49138463e+00 -8.08202922e-01
-2.25646254e-02 -1.30828656e-02 -1.94008127e-01 3.88345152e-01
4.50895637e-01 -2.75865436e-01 6.19201735e-02 -5.53727329e-01
2.72651970e-01 -1.91005185e-01 1.87376544e-01 -6.24781735e-02
-1.05232072e+00 6.04843497e-01 1.46327808e-01 4.13716108e-01
-8.29529643e-01 9.94620994e-02 2.99962401e-01 5.95407635e-02
1.33112922e-01 1.22378564e+00 -6.99221253e-01 -4.44180630e-02
-8.28851223e-01 7.66571164e-01 6.43707812e-01 1.28690863e+00
6.05637729e-01 -2.08875835e-01 -3.51010025e-01 8.03494513e-01
-8.59335512e-02 2.40530685e-01 -1.08901709e-01 -9.91121709e-01
4.52294320e-01 3.46547723e-01 6.59371436e-01 -8.75817895e-01
-9.76891220e-01 -5.80399930e-01 -3.09473306e-01 -1.95642054e-01
-3.58021185e-02 -1.08260266e-01 -7.79165745e-01 7.92597055e-01
2.06803426e-01 -4.94155645e-01 -1.27195805e-01 1.13349783e+00
2.01112643e-01 9.67783988e-01 1.42337665e-01 -9.18034375e-01
8.50251079e-01 -1.35470796e+00 -8.73929322e-01 -1.96531981e-01
3.16543847e-01 -1.32592893e+00 1.44446626e-01 2.92197049e-01
-1.46812224e+00 -1.84357613e-01 -6.11831248e-01 4.15257037e-01
-3.75514627e-01 -3.09584826e-01 8.25566173e-01 4.13107306e-01
-1.19918728e+00 4.07350600e-01 -2.70450205e-01 -1.39148995e-01
-6.16731703e-01 9.10042465e-01 -1.72193721e-01 -3.17492515e-01
-9.00364637e-01 1.36617899e+00 7.66340613e-01 1.30246863e-01
-5.65420270e-01 8.54449570e-02 -1.38761804e-01 -6.03990555e-02
7.41962016e-01 -7.62157977e-01 1.27384341e+00 -4.72435892e-01
-1.04774189e+00 5.92841208e-01 -3.14665705e-01 -2.17219517e-02
1.38665855e-01 9.10954416e-01 -2.53548324e-01 2.54991632e-02
2.47727886e-01 5.04480720e-01 3.21998924e-01 -1.69024003e+00
-1.02626336e+00 1.92397088e-01 1.24765068e-01 2.56494761e-01
1.29388705e-01 5.23516178e-01 -1.71292692e-01 -1.30113021e-01
1.03672795e-01 -1.04982543e+00 -9.11120474e-01 -1.55079174e+00
-5.52855730e-01 -3.76297683e-01 -3.48196179e-02 -3.55300248e-01
1.42304718e+00 -1.40439606e+00 4.93862212e-01 6.17206156e-01
-1.84402078e-01 -1.39657021e-01 -2.31959954e-01 1.13044500e+00
1.56491578e-01 1.50585026e-01 3.29850823e-01 -3.00989568e-01
8.77915323e-02 9.32915390e-01 3.26645494e-01 2.97747344e-01
-8.91648680e-02 4.67071325e-01 -9.34397519e-01 -9.67667401e-01
8.66206288e-02 -1.11573301e-02 2.94424258e-02 4.01641339e-01
-6.56210631e-02 1.08064301e-01 -4.91730332e-01 1.10493827e+00
1.02491319e+00 3.30493391e-01 3.07022959e-01 2.78214335e-01
-5.10191798e-01 -1.65207982e-01 -1.20725882e+00 1.12021041e+00
-6.04339838e-01 3.03173102e-02 5.51727235e-01 -9.06824827e-01
1.11124158e+00 2.74840951e-01 1.05609632e+00 -8.59868944e-01
3.06806058e-01 4.62807745e-01 1.23670593e-01 -5.32996118e-01
1.10247600e+00 -1.57768697e-01 -3.48291576e-01 3.36875796e-01
-2.59897441e-01 2.68293340e-02 9.69364822e-01 -1.56125119e-02
1.12755704e+00 -1.93985596e-01 2.79495239e-01 -8.35584104e-01
5.29713809e-01 7.10749030e-01 7.51967967e-01 8.33668470e-01
-2.32337594e-01 3.46399307e-01 3.30698282e-01 -6.02259278e-01
-1.00750566e+00 -4.90678608e-01 -1.33814380e-01 1.05432856e+00
3.22550207e-01 4.08534780e-02 -5.58490157e-01 -1.45942852e-01
1.16187960e-01 5.63413858e-01 2.82494426e-02 6.38402462e-01
-6.68226063e-01 -1.01746297e+00 -1.14759266e-01 -4.12926190e-02
-3.35438699e-01 -1.15050960e+00 -7.34781325e-01 9.17322755e-01
-3.93961132e-01 -8.71133387e-01 -2.91955858e-01 7.70475030e-01
-4.77425307e-01 -8.93362522e-01 -7.92220592e-01 -1.05643749e+00
1.29534435e+00 6.10107243e-01 1.28461635e+00 5.03960073e-01
-7.38493800e-01 -3.94858152e-01 -4.99929249e-01 -3.88461322e-01
-1.22412570e-01 5.50663710e-01 -1.44754320e-01 -6.72262371e-01
-1.57999977e-01 -8.79731849e-02 -1.14928233e-03 5.51594913e-01
-5.68637848e-01 -1.06685132e-01 7.44682789e-01 6.40749991e-01
7.94663072e-01 1.06701767e+00 4.97946560e-01 -9.20296252e-01
1.27343881e+00 -3.92723322e-01 -6.75489306e-01 4.60146070e-01
-9.04429793e-01 -9.04304385e-02 1.08678555e+00 1.31119907e-01
-8.75302672e-01 4.56851363e-01 1.21294402e-01 1.50112268e-02
-1.75636202e-01 4.81754839e-01 2.28129923e-01 -5.46817183e-01
1.22173958e-01 1.56105325e-01 -4.98763949e-01 -4.07605082e-01
-1.81832924e-01 3.32682371e-01 4.28318053e-01 -1.00268984e+00
6.55429244e-01 -1.94875121e-01 6.78816319e-01 -2.31995478e-01
-2.89733380e-01 -7.80053318e-01 -4.23341632e-01 -4.79034007e-01
5.32700717e-01 -1.63964301e-01 -8.43878627e-01 -1.82148740e-01
-1.32928860e+00 8.82798284e-02 4.84820664e-01 1.64353475e-01
-5.99292696e-01 -1.03049381e-02 -3.48015994e-01 -1.22071445e+00
-1.53939158e-01 -1.52457249e+00 5.21372080e-01 2.66115427e-01
-2.85529137e-01 -5.65142512e-01 9.64080021e-02 5.72374284e-01
5.52434981e-01 4.27390277e-01 6.95932031e-01 -8.34096074e-02
-6.87670290e-01 -3.42070088e-02 -2.54411817e-01 -7.10513890e-01
-1.15863115e-01 2.73752004e-01 -7.76088014e-02 -5.08393049e-01
-6.89047873e-01 2.63203591e-01 1.63313329e-01 6.26616597e-01
1.13891447e+00 -3.83448154e-01 -8.79614294e-01 1.84045851e-01
1.95322490e+00 9.36062276e-01 4.52767462e-01 1.02552009e+00
-9.96880699e-03 7.79723644e-01 1.29197001e+00 4.95619655e-01
4.23848718e-01 9.44115877e-01 4.39153463e-01 -3.48819613e-01
3.77274007e-01 6.18768811e-01 -3.77633750e-01 5.89559019e-01
-5.76078773e-01 -7.17221677e-01 -1.25332665e+00 1.02793097e+00
-2.04867268e+00 -1.13972294e+00 -5.67396700e-01 1.45576608e+00
6.35380208e-01 4.69651788e-01 6.79259002e-01 4.62506384e-01
8.29956055e-01 -4.58143443e-01 5.72357416e-01 -1.46492994e+00
1.64614290e-01 1.40480861e-01 9.90601778e-01 7.27723718e-01
-7.94027984e-01 4.01325941e-01 7.53811646e+00 6.80428088e-01
-5.17995596e-01 -8.00086036e-02 1.34851128e-01 -2.20215619e-01
-3.07185799e-01 1.79198429e-01 -7.04122722e-01 6.69815063e-01
1.08417833e+00 -6.50966048e-01 8.55973840e-01 7.65883267e-01
3.37315768e-01 -5.34437656e-01 -3.06731790e-01 3.65043819e-01
-2.09570214e-01 -1.25447237e+00 -6.37911379e-01 2.07610518e-01
1.11397326e+00 -6.72312915e-01 -6.94045663e-01 -1.02955043e-01
1.97055131e-01 -7.83762753e-01 5.83693624e-01 5.85539281e-01
-7.61697739e-02 -1.57421899e+00 1.30295050e+00 3.69793445e-01
-1.14785564e+00 -2.52611071e-01 -4.51939613e-01 -4.26722556e-01
7.26618707e-01 7.48059154e-01 -1.37044704e+00 1.01554954e+00
5.06796479e-01 -5.54286540e-01 -1.53057367e-01 1.75095069e+00
3.58988643e-01 -2.45140851e-01 -2.73794010e-02 -3.73790771e-01
3.78418297e-01 -1.47153556e-01 4.46119666e-01 1.58358049e+00
1.85571223e-01 1.26526982e-01 1.11086249e+00 2.35690624e-01
6.36742413e-01 3.11016440e-01 -2.88538873e-01 9.13948789e-02
7.34673083e-01 1.45113611e+00 -1.67815745e+00 1.98877767e-01
2.80468632e-02 6.92483962e-01 1.05465658e-01 -1.13984868e-01
-8.76574814e-01 -8.04816008e-01 2.64898121e-01 3.20070051e-02
3.62594947e-02 -4.50045347e-01 -4.03303236e-01 -1.44947901e-01
-1.88709006e-01 -2.26166710e-01 4.49094415e-01 -8.14370513e-01
-1.10071480e+00 8.75785172e-01 6.93881273e-01 -7.83250749e-01
-4.10922825e-01 -2.55469620e-01 -4.92077827e-01 1.15459442e+00
-1.58132005e+00 -8.36939216e-01 1.15999855e-01 9.47041959e-02
5.53790808e-01 -2.49825612e-01 7.84609854e-01 6.69484556e-01
-6.88212812e-01 1.34581327e-01 2.86184717e-02 -8.30485880e-01
1.90233976e-01 -1.32653952e+00 -1.64556891e-01 8.27146053e-01
-7.40744531e-01 5.07543921e-01 1.48331082e+00 -6.49396122e-01
-1.54603708e+00 -3.79595846e-01 1.60417461e+00 3.75106990e-01
3.91382962e-01 6.99477047e-02 -3.21128279e-01 8.84426981e-02
6.49763227e-01 -5.74492335e-01 2.75590777e-01 1.16628565e-01
8.45607281e-01 -1.05319075e-01 -1.47810197e+00 1.69369563e-01
6.32201374e-01 4.62904572e-01 -2.49674931e-01 6.12585962e-01
5.41468441e-01 -7.71946192e-01 -7.18385100e-01 1.50062710e-01
4.79081972e-03 -5.69992781e-01 9.97260153e-01 -1.78792953e-01
2.24087104e-01 -4.10103142e-01 2.31778935e-01 -1.64793730e+00
-1.20465684e+00 -7.07241237e-01 4.72731829e-01 1.30195475e+00
7.85315812e-01 -5.40443242e-01 8.83349299e-01 7.86881804e-01
-6.51600003e-01 -1.10804856e+00 -9.72493947e-01 -1.11655283e+00
-5.55136323e-01 3.68636757e-01 9.06315625e-01 9.99188602e-01
-1.01960883e-01 2.48479359e-02 -6.16396070e-01 -3.83192822e-02
7.99270391e-01 7.19231367e-01 2.69001156e-01 -1.10184884e+00
-1.56660259e-01 -4.26971763e-01 -2.24109963e-01 3.54078412e-01
4.31204021e-01 -9.89245296e-01 4.40449029e-01 -2.16358232e+00
4.06177938e-01 -8.83930683e-01 -1.94861799e-01 8.21386755e-01
-5.23456298e-02 8.59037712e-02 4.20621485e-01 -1.53736901e-02
-7.40130126e-01 -3.79336715e-01 1.24825370e+00 8.31069425e-03
-3.24134260e-01 4.31511402e-02 -5.05210161e-01 -2.25174800e-01
1.21339047e+00 -1.12692368e+00 -3.17066252e-01 -4.08074707e-01
2.72810429e-01 6.56666875e-01 -4.02004957e-01 -7.24819660e-01
2.71942407e-01 -1.40364122e+00 1.26727879e-01 -7.44961262e-01
9.28000584e-02 -1.30477440e+00 8.52800965e-01 6.39468431e-01
-7.11504519e-02 7.95936048e-01 5.19772843e-02 -1.59876095e-03
-2.31233798e-02 -1.03057754e+00 3.46450835e-01 -6.32314861e-01
-9.91149247e-01 -1.52773261e-01 -6.38330519e-01 -4.82624650e-01
1.61979496e+00 -1.01342469e-01 -3.53516936e-01 2.43659332e-01
-5.62091589e-01 8.05658579e-01 3.42674375e-01 2.37122312e-01
2.77345926e-01 -9.72599626e-01 -6.28398061e-01 -5.92826188e-01
-2.17397854e-01 -4.56356287e-01 1.83085874e-01 6.59299493e-01
-1.22839046e+00 9.45520043e-01 -1.02966106e+00 -1.73844606e-01
-1.44766843e+00 9.94465947e-01 -1.83619991e-01 -5.67623079e-01
-1.09572314e-01 6.44403636e-01 -1.19900513e+00 -3.89680356e-01
-4.88227187e-03 4.78853881e-01 -1.30274326e-01 -1.51968434e-01
-1.71024221e-04 9.36908603e-01 3.06290716e-01 -5.45824647e-01
-9.36571300e-01 4.31646943e-01 3.58341932e-01 -3.36151458e-02
1.45507467e+00 -2.31944531e-01 -9.94437397e-01 -2.45106265e-01
9.03544605e-01 -1.54689789e-01 -2.79534250e-01 4.65267599e-01
4.77938086e-01 -1.17428362e+00 -1.71086583e-02 -8.50747645e-01
-9.21726584e-01 -1.69317141e-01 -4.42496687e-02 9.55971479e-01
1.73508239e+00 -2.83083081e-01 6.85782850e-01 2.14770839e-01
1.00260723e+00 -1.62676132e+00 -6.52103364e-01 2.11859688e-01
7.52587080e-01 -8.08295667e-01 6.06494784e-01 -8.27308238e-01
-7.30044961e-01 1.33437121e+00 1.00128162e+00 7.46766478e-02
8.03780630e-02 6.66327894e-01 -2.01764688e-01 -2.60994494e-01
-1.09224594e+00 -5.89436948e-01 -2.40805522e-01 9.14443195e-01
4.43367898e-01 4.69358772e-01 -1.62088025e+00 4.45804596e-02
-8.59581381e-02 -2.67973598e-02 1.00451422e+00 1.25911701e+00
-5.65300226e-01 -1.66907847e+00 -8.15241337e-01 4.71554935e-01
-4.76213485e-01 7.35324919e-01 -4.17192250e-01 6.09533250e-01
3.40760648e-01 1.36983347e+00 -1.95626885e-01 -4.54955995e-01
4.82323617e-01 -2.32783064e-01 4.74830210e-01 -5.11700451e-01
-1.00361919e+00 3.06004614e-01 8.40082586e-01 -2.98130155e-01
-5.06814063e-01 -2.97017217e-01 -1.39417410e+00 -6.90179288e-01
-3.50487411e-01 6.37239516e-01 8.94518256e-01 7.07500458e-01
1.75453231e-01 8.22005153e-01 1.20162737e+00 -1.23955834e+00
-7.65517801e-02 -4.24021810e-01 -6.76479995e-01 2.00841114e-01
-2.31688306e-01 -8.64774108e-01 1.10149056e-01 -5.77334344e-01] | [5.015581130981445, 2.120194911956787] |
c0c6a6fb-77f5-4ce8-b926-f4f829a164cb | cross-modality-fusion-transformer-for | 2111.00273 | null | https://arxiv.org/abs/2111.00273v4 | https://arxiv.org/pdf/2111.00273v4.pdf | Cross-Modality Fusion Transformer for Multispectral Object Detection | Multispectral image pairs can provide the combined information, making object detection applications more reliable and robust in the open world. To fully exploit the different modalities, we present a simple yet effective cross-modality feature fusion approach, named Cross-Modality Fusion Transformer (CFT) in this paper. Unlike prior CNNs-based works, guided by the transformer scheme, our network learns long-range dependencies and integrates global contextual information in the feature extraction stage. More importantly, by leveraging the self attention of the transformer, the network can naturally carry out simultaneous intra-modality and inter-modality fusion, and robustly capture the latent interactions between RGB and Thermal domains, thereby significantly improving the performance of multispectral object detection. Extensive experiments and ablation studies on multiple datasets demonstrate that our approach is effective and achieves state-of-the-art detection performance. Our code and models are available at https://github.com/DocF/multispectral-object-detection. | ['Wang Zhaokui', 'Han Dapeng', 'Fang Qingyun'] | 2021-10-30 | null | null | null | null | ['multispectral-object-detection'] | ['computer-vision'] | [ 2.32887924e-01 -6.36353672e-01 -8.19399208e-02 -2.76878119e-01
-1.10680759e+00 -6.57336771e-01 6.93243206e-01 -2.09208414e-01
-2.32338816e-01 2.12258235e-01 6.65930510e-02 -1.40551987e-04
-1.09113298e-01 -7.85440207e-01 -5.77094436e-01 -9.01071489e-01
3.65225405e-01 -3.98983240e-01 2.75526494e-01 -7.30400681e-02
-4.06572260e-02 4.67659652e-01 -1.73702061e+00 3.63347858e-01
9.23689306e-01 1.19193065e+00 3.41823786e-01 6.02840662e-01
1.32621273e-01 5.46705484e-01 6.28288463e-02 -2.79742509e-01
5.11732399e-01 3.64051852e-03 -6.74252689e-01 2.31495097e-01
6.02139652e-01 -4.58272368e-01 -5.00750482e-01 1.21006334e+00
5.58701694e-01 -4.18798141e-02 3.18292171e-01 -1.26118493e+00
-9.52232182e-01 3.17475617e-01 -9.95556056e-01 1.81910172e-01
2.10416943e-01 5.75749397e-01 9.88567352e-01 -9.74440694e-01
1.19135909e-01 1.11985838e+00 7.55600214e-01 1.67607203e-01
-1.21906245e+00 -8.31882715e-01 1.81773469e-01 2.16450602e-01
-1.46632135e+00 -4.24380898e-01 7.49584973e-01 -1.91762820e-01
8.25559378e-01 2.04373837e-01 7.65958667e-01 1.04327655e+00
-1.03859797e-01 9.74960148e-01 1.31799901e+00 -4.36013162e-01
-4.57863510e-01 -1.75613999e-01 -7.26821795e-02 8.22705209e-01
1.87601075e-01 4.79185760e-01 -6.31827056e-01 -8.72287303e-02
7.92487144e-01 4.89806652e-01 -2.66112208e-01 -1.45041049e-01
-1.37446153e+00 5.14976919e-01 1.03581095e+00 3.49676490e-01
-4.04180914e-01 2.26005271e-01 1.12790748e-01 3.44407074e-02
5.43969095e-01 3.59037481e-02 -3.24915618e-01 2.96584904e-01
-6.99293435e-01 -4.39617895e-02 1.13054439e-01 7.12716818e-01
9.46412861e-01 -2.01137006e-01 -2.21541196e-01 8.90004098e-01
4.98616278e-01 8.50809038e-01 1.11440629e-01 -8.93797278e-01
2.94028580e-01 9.01436567e-01 5.61162829e-02 -6.02168798e-01
-4.25167561e-01 -5.90777040e-01 -8.02134037e-01 9.64336097e-02
1.82367280e-01 -1.18256949e-01 -1.10652661e+00 1.68679583e+00
5.25569022e-01 4.03357506e-01 -8.88132975e-02 1.05409718e+00
8.60804498e-01 3.79196525e-01 2.52850384e-01 1.10106803e-01
1.59352958e+00 -1.07458901e+00 -3.31375748e-01 -3.94327551e-01
1.74619645e-01 -1.03735507e+00 7.44238198e-01 2.29162462e-02
-7.37415373e-01 -6.65618420e-01 -8.40087414e-01 -2.54371494e-01
-4.78121370e-01 3.86781365e-01 8.02049994e-01 4.02467608e-01
-9.77959931e-01 1.75916508e-01 -9.26697850e-01 -6.13016725e-01
6.51526272e-01 8.51589739e-02 -4.99268770e-01 -4.68531549e-01
-8.77549827e-01 7.67556310e-01 5.26851535e-01 3.71588767e-01
-8.47326577e-01 -6.95528805e-01 -7.22609162e-01 -9.86859351e-02
4.61467028e-01 -8.76056492e-01 1.20970440e+00 -7.47483552e-01
-1.14366162e+00 7.15215683e-01 -3.28911692e-01 -1.41813964e-01
3.89656216e-01 -3.36660326e-01 -3.44947070e-01 1.45600215e-01
1.73578292e-01 8.07885826e-01 9.02473092e-01 -1.33762169e+00
-9.24782872e-01 -5.06414175e-01 2.10494950e-01 2.65554160e-01
-4.71897215e-01 1.45840555e-01 -6.81401908e-01 -3.40382844e-01
2.86973745e-01 -7.25430250e-01 -4.00699712e-02 1.63435653e-01
-5.63885510e-01 -1.61106125e-01 1.02788889e+00 -5.30394495e-01
7.62752414e-01 -2.30903244e+00 6.74439296e-02 -4.26700041e-02
2.77736813e-01 3.62277508e-01 -4.79117781e-01 3.10410947e-01
2.35602763e-02 -5.32839671e-02 -2.62302846e-01 -4.89910066e-01
-5.33074327e-02 4.97944057e-02 -2.70073622e-01 5.51976502e-01
4.34822619e-01 1.20710516e+00 -8.84128690e-01 -4.27490771e-01
5.33885300e-01 8.06446612e-01 -6.17833883e-02 1.64854363e-01
8.46030936e-02 4.64229107e-01 -5.34819722e-01 1.36852980e+00
8.71603191e-01 -3.35942537e-01 -1.43381044e-01 -7.29956329e-01
-1.64572805e-01 -1.02333613e-01 -9.17702556e-01 1.93160868e+00
-5.21519482e-01 5.51748276e-01 1.63276091e-01 -5.56431234e-01
6.50999486e-01 8.79854485e-02 4.62169796e-01 -7.95477986e-01
2.65244842e-01 2.05904156e-01 -3.44735384e-01 -3.70315939e-01
5.17973483e-01 6.59764232e-03 1.56888068e-01 2.72291631e-01
3.01902860e-01 1.91081557e-02 4.70911115e-02 6.09026812e-02
8.00368547e-01 3.41707408e-01 2.24863932e-01 1.36155412e-01
3.97736102e-01 -2.16589093e-01 4.72123593e-01 8.74042034e-01
-3.93388391e-01 6.38145030e-01 -8.90178606e-02 -2.93963701e-01
-7.09931135e-01 -1.12019777e+00 -2.27516264e-01 1.10147238e+00
2.92326003e-01 -2.01626927e-01 -3.86282980e-01 -5.70316553e-01
1.21217556e-01 1.23552606e-01 -7.45083570e-01 -1.25030369e-01
-1.94254182e-02 -1.00657952e+00 6.04131460e-01 6.56616509e-01
1.00998116e+00 -7.40359545e-01 -4.80410129e-01 1.04164500e-02
-5.96596479e-01 -1.30178511e+00 -2.68215328e-01 6.62080273e-02
-6.97570980e-01 -1.23645163e+00 -5.62082529e-01 -2.89185882e-01
3.17951769e-01 9.97148931e-01 7.41175294e-01 1.37629703e-01
-6.20168328e-01 5.74182868e-01 -4.14638847e-01 -2.90288866e-01
1.45913213e-01 6.45556599e-02 -2.23912910e-01 2.92319089e-01
4.87325490e-01 -4.70226437e-01 -7.53331304e-01 9.08779427e-02
-9.97372806e-01 2.12291896e-01 8.23838413e-01 8.75271976e-01
5.16379237e-01 4.28651907e-02 2.49877855e-01 -2.60751009e-01
1.39959157e-01 -3.92855883e-01 -6.19551241e-01 5.24811327e-01
-3.40596557e-01 -2.76674688e-01 3.48809324e-02 -2.62391150e-01
-1.38603067e+00 3.12015951e-01 1.11500584e-01 -4.88665432e-01
-4.00218040e-01 3.40169728e-01 -1.21477291e-01 -5.38745761e-01
6.15729451e-01 3.63309056e-01 -2.64772743e-01 -5.88025749e-01
6.57030761e-01 7.01232731e-01 6.71206892e-01 -4.06847268e-01
1.07445943e+00 9.26732421e-01 -1.96876258e-01 -4.81253833e-01
-1.08192039e+00 -8.05092812e-01 -6.41925097e-01 -3.37092876e-01
8.38600755e-01 -1.36517370e+00 -5.66843688e-01 1.00276005e+00
-1.00144529e+00 -1.51124328e-01 -3.21918428e-02 2.97431678e-01
-1.17014058e-01 2.69789249e-01 -5.57777941e-01 -9.21895385e-01
-3.41604233e-01 -1.07228684e+00 1.56255674e+00 6.34947181e-01
6.35909200e-01 -8.26010942e-01 -2.24227086e-01 5.39733946e-01
5.98360717e-01 1.94283202e-01 2.67978311e-01 -6.40099198e-02
-9.50026929e-01 -2.39690244e-01 -1.00920117e+00 3.01779747e-01
3.54595989e-01 9.00687203e-02 -1.33930290e+00 -2.20894054e-01
-4.45075810e-01 -3.83756697e-01 1.48121321e+00 2.52225757e-01
9.39953327e-01 2.48574927e-01 -4.34706748e-01 7.79192567e-01
1.62762225e+00 -4.17198718e-01 4.98860329e-01 3.39790463e-01
1.12421060e+00 4.58162814e-01 5.89656055e-01 3.50752324e-01
7.36701012e-01 5.30064404e-01 6.79901183e-01 -5.21239281e-01
-3.20768297e-01 -6.10767975e-02 3.11506838e-01 2.08993867e-01
-1.87457338e-01 -7.42089823e-02 -9.56849277e-01 5.75143516e-01
-2.02239847e+00 -9.50539231e-01 -5.55403642e-02 1.89467192e+00
6.09527230e-01 -3.02500099e-01 -2.20481995e-02 -2.13109270e-01
7.58660614e-01 3.04585516e-01 -6.37885153e-01 4.13419068e-01
-4.91764605e-01 -4.39918526e-02 6.96893036e-01 3.07037771e-01
-1.48390269e+00 8.84868860e-01 6.15633917e+00 9.46603715e-01
-1.16887820e+00 4.24718648e-01 3.60906214e-01 -1.57294601e-01
-7.15132058e-02 -2.26942748e-02 -5.60441315e-01 1.62709698e-01
5.13341010e-01 3.13837230e-01 4.73259866e-01 4.44327027e-01
-2.23143958e-02 -3.02238941e-01 -5.66537976e-01 8.71516824e-01
-6.19220696e-02 -1.18903744e+00 -1.62295073e-01 1.29139036e-01
6.89082742e-01 6.64825261e-01 1.21761151e-01 1.03729181e-01
3.70574147e-01 -6.17985964e-01 7.24783719e-01 6.73827529e-01
7.47783244e-01 -4.96042550e-01 6.61297977e-01 5.30289039e-02
-1.73945177e+00 -4.49684113e-01 -1.45537585e-01 1.77106485e-01
1.43741267e-02 7.60731578e-01 -2.13219583e-01 1.06571400e+00
1.06963730e+00 1.17254484e+00 -1.01901329e+00 1.13141370e+00
-2.84145206e-01 3.47852528e-01 -5.05219579e-01 5.65746963e-01
1.35129541e-01 2.96552703e-02 4.82483208e-01 1.19165111e+00
3.55243742e-01 -7.71323070e-02 4.38493133e-01 9.49965596e-01
-1.97824202e-02 -3.95038784e-01 -4.93050486e-01 -4.99738269e-02
5.14719069e-01 1.74018538e+00 -5.65955460e-01 -1.33335933e-01
-6.23907328e-01 8.95669699e-01 3.08319449e-01 5.67078173e-01
-9.00284946e-01 -1.44178465e-01 7.60042310e-01 -4.32551980e-01
4.68412668e-01 -1.84122965e-01 -2.47231513e-01 -1.43801188e+00
1.70764208e-01 -4.96762782e-01 4.59339499e-01 -1.03635073e+00
-1.58481908e+00 4.23727274e-01 -1.23262644e-01 -1.34575987e+00
2.89029866e-01 -7.63237357e-01 -4.67257202e-01 9.24603641e-01
-2.10064769e+00 -2.19233346e+00 -7.22860873e-01 8.09873343e-01
1.03427343e-01 9.61864665e-02 6.77245617e-01 3.67524207e-01
-7.49333084e-01 2.54233837e-01 -5.38509525e-02 1.46363705e-01
7.85671890e-01 -1.03640890e+00 1.68380156e-01 1.27257359e+00
9.42501575e-02 5.05979419e-01 1.26735538e-01 -3.59056652e-01
-1.50273693e+00 -1.21046412e+00 2.12598264e-01 -4.43545282e-01
8.92342865e-01 -1.14065781e-01 -8.31732512e-01 5.31123340e-01
3.15999895e-01 4.35157120e-01 6.53666914e-01 2.00307310e-01
-9.99417663e-01 -2.85819590e-01 -8.54932010e-01 2.00906470e-01
1.11423028e+00 -1.00023687e+00 -2.80403823e-01 3.44437033e-01
6.73859775e-01 -3.90807331e-01 -8.58714938e-01 6.78305686e-01
6.62306011e-01 -1.11413741e+00 1.29169953e+00 -1.33830771e-01
2.75715798e-01 -6.06747985e-01 -3.95218372e-01 -1.08996725e+00
-5.20073354e-01 -2.52101690e-01 -2.39879668e-01 1.40442348e+00
2.07463488e-01 -8.08209002e-01 8.70074779e-02 2.46381581e-01
1.17078412e-03 -4.38053489e-01 -9.06934083e-01 -6.25843763e-01
-2.05229655e-01 -5.80664337e-01 7.36135423e-01 8.65673959e-01
-4.68879431e-01 -4.78245690e-02 -3.61616313e-01 6.62735999e-01
9.31123435e-01 5.22166133e-01 4.96744573e-01 -1.14351356e+00
-2.93454409e-01 -5.27926862e-01 -2.56693840e-01 -6.78864002e-01
-3.59234139e-02 -7.45358467e-01 4.95328084e-02 -1.63170445e+00
6.01505399e-01 -3.87338191e-01 -7.22805738e-01 1.00193357e+00
-4.07938302e-01 9.69422221e-01 3.95699114e-01 3.38800937e-01
-8.02791238e-01 5.50208390e-01 1.13610351e+00 -2.09882006e-01
7.09848404e-02 -4.39586669e-01 -9.77300167e-01 5.69670498e-01
7.07322240e-01 -2.80415893e-01 5.40597513e-02 -7.09839642e-01
-9.17584673e-02 -2.21756771e-01 1.07149243e+00 -9.99399006e-01
3.88766915e-01 -2.38202825e-01 6.03807807e-01 -5.34202933e-01
5.63964248e-01 -9.01331067e-01 -7.75426701e-02 1.89127117e-01
4.51270342e-02 -4.20891047e-01 3.75838757e-01 7.02101529e-01
-2.68614084e-01 3.63140404e-01 6.66131735e-01 -3.28886934e-04
-1.05018103e+00 5.30157149e-01 1.45932548e-02 -4.70131665e-01
8.80545497e-01 -5.21529242e-02 -8.22170675e-01 2.58970261e-03
-4.39522207e-01 3.51814926e-01 4.83694255e-01 6.56715035e-01
4.68598664e-01 -1.42589366e+00 -7.02473998e-01 1.94968134e-01
6.28705800e-01 -5.97440861e-02 6.84747934e-01 1.23370624e+00
1.99927818e-02 1.37498692e-01 -2.48121217e-01 -9.42134976e-01
-1.05260086e+00 2.36371920e-01 6.75109029e-01 4.37414832e-02
-4.01786327e-01 8.60711336e-01 1.96166992e-01 -5.62254846e-01
-8.19715336e-02 -1.35876849e-01 1.00044094e-01 8.25287700e-02
7.09676802e-01 1.40922412e-01 2.08080169e-02 -7.44449854e-01
-4.30893809e-01 6.70374691e-01 -1.29909571e-02 6.00179918e-02
1.33827198e+00 -4.04709548e-01 -3.09572786e-01 1.90208972e-01
1.04706240e+00 -2.53630251e-01 -1.55532634e+00 -8.23785007e-01
-4.29003745e-01 -7.58063912e-01 4.97690320e-01 -1.04732406e+00
-1.31452715e+00 8.35404575e-01 9.13216591e-01 2.87435204e-01
1.65756023e+00 1.01577371e-01 5.84205985e-01 2.90555954e-01
2.38626257e-01 -7.17431068e-01 6.95942417e-02 3.83958578e-01
6.60481930e-01 -1.72946465e+00 -1.36006959e-02 -5.87516606e-01
-3.56968462e-01 1.04578102e+00 7.50778258e-01 1.11271292e-01
5.18165827e-01 2.10895717e-01 2.19608247e-01 -1.43910304e-01
-4.72992867e-01 -8.62180829e-01 4.97422189e-01 6.42537951e-01
3.10935974e-01 2.40916476e-01 4.35900450e-01 2.26129875e-01
3.84259909e-01 -5.00443485e-03 5.68916164e-02 9.57103372e-01
-3.77341032e-01 -1.06192386e+00 -5.53381264e-01 3.36917400e-01
-1.41304478e-01 -2.47156054e-01 -4.00439531e-01 6.75094068e-01
4.01132017e-01 1.16698074e+00 -1.90359414e-01 -5.85907340e-01
1.64531231e-01 -2.67723314e-02 5.32473564e-01 -3.14092636e-01
-5.12415826e-01 2.86870509e-01 -2.30321854e-01 -9.09239471e-01
-1.02499187e+00 -7.37467349e-01 -8.32357943e-01 -2.58196175e-01
-7.11337805e-01 -5.18685997e-01 6.15567803e-01 1.02107692e+00
6.21325254e-01 5.41703522e-01 5.33732533e-01 -1.21119082e+00
-2.42291793e-01 -1.05753040e+00 -4.11595702e-01 2.31079161e-01
6.49390459e-01 -8.23519945e-01 -2.34590501e-01 -1.30306765e-01] | [10.103190422058105, -1.6004008054733276] |
5f934725-9969-4b86-ade1-cee4ce7c3f43 | semidefinite-programming-based | 1310.2273 | null | http://arxiv.org/abs/1310.2273v2 | http://arxiv.org/pdf/1310.2273v2.pdf | Semidefinite Programming Based Preconditioning for More Robust Near-Separable Nonnegative Matrix Factorization | Nonnegative matrix factorization (NMF) under the separability assumption can
provably be solved efficiently, even in the presence of noise, and has been
shown to be a powerful technique in document classification and hyperspectral
unmixing. This problem is referred to as near-separable NMF and requires that
there exists a cone spanned by a small subset of the columns of the input
nonnegative matrix approximately containing all columns. In this paper, we
propose a preconditioning based on semidefinite programming making the input
matrix well-conditioned. This in turn can improve significantly the performance
of near-separable NMF algorithms which is illustrated on the popular successive
projection algorithm (SPA). The new preconditioned SPA is provably more robust
to noise, and outperforms SPA on several synthetic data sets. We also show how
an active-set method allow us to apply the preconditioning on large-scale
real-world hyperspectral images. | ['Nicolas Gillis', 'Stephen A. Vavasis'] | 2013-10-08 | null | null | null | null | ['hyperspectral-unmixing'] | ['computer-vision'] | [ 8.00402761e-01 -2.85867989e-01 -1.78636581e-01 -3.69175188e-02
-7.67809451e-01 -8.05180013e-01 2.60628641e-01 -3.25049728e-01
-2.01475978e-01 6.00162387e-01 2.59415150e-01 -3.57474029e-01
-5.38625419e-01 -5.33067822e-01 -5.15503168e-01 -1.26134753e+00
4.39351536e-02 5.63880444e-01 -7.12212861e-01 -9.80165526e-02
-1.58783317e-01 5.66432357e-01 -1.37125158e+00 3.39210033e-01
1.27800083e+00 5.92767060e-01 3.21965665e-01 4.87506598e-01
1.48489222e-01 4.11527276e-01 -1.28035441e-01 1.42403007e-01
8.30742598e-01 -2.22354650e-01 -6.37547612e-01 8.23844135e-01
6.10796571e-01 -1.41593115e-02 -2.65172601e-01 1.37908697e+00
1.68547794e-01 4.13840443e-01 4.43798453e-01 -1.33788300e+00
-4.99274164e-01 4.19757187e-01 -1.01670003e+00 -1.24601714e-01
1.83616340e-01 -3.66846293e-01 9.99901950e-01 -1.05297613e+00
5.90125680e-01 1.05679870e+00 4.25050735e-01 1.51298329e-01
-1.58272004e+00 -5.83797336e-01 -1.52875381e-02 -1.19518051e-02
-1.28144896e+00 -4.79136527e-01 5.90025961e-01 -4.95419830e-01
6.24517441e-01 9.27896261e-01 7.67811537e-01 4.57942396e-01
-3.35919082e-01 8.54008973e-01 1.36808825e+00 -7.07840979e-01
2.19302982e-01 -1.61550656e-01 4.67710704e-01 3.71244997e-01
5.82338333e-01 -4.36198637e-02 -5.00684559e-01 -6.74328566e-01
2.72867024e-01 2.08608389e-01 -8.99234176e-01 -7.09695697e-01
-1.36133337e+00 8.95248175e-01 1.12860836e-01 3.65351997e-02
-4.96904910e-01 -3.22404385e-01 3.12635675e-02 3.44696671e-01
5.23799300e-01 4.24660474e-01 -5.50227500e-02 3.99548441e-01
-1.17974329e+00 4.12755348e-02 7.51300693e-01 6.52247727e-01
8.55944216e-01 1.94190934e-01 -1.26093738e-02 9.50333059e-01
2.55907267e-01 1.12481987e+00 -5.11071533e-02 -9.67158556e-01
7.28412688e-01 3.87290627e-01 4.48306888e-01 -1.04285693e+00
-4.11555976e-01 -4.87529993e-01 -1.13557327e+00 1.23548999e-01
2.88727850e-01 -6.94168806e-02 -8.33760202e-01 1.66506636e+00
2.39499807e-01 2.38662124e-01 3.02177310e-01 1.01298547e+00
3.79289269e-01 1.06742120e+00 -5.00745177e-01 -9.88594413e-01
8.30270767e-01 -8.85660946e-01 -7.63512909e-01 -4.17283714e-01
3.09109539e-01 -8.04571390e-01 7.15351939e-01 7.04131126e-01
-7.74516046e-01 5.62698394e-03 -1.13020551e+00 1.50914475e-01
8.86780843e-02 2.53700584e-01 8.28001440e-01 8.81878197e-01
-9.30927753e-01 2.20717922e-01 -8.75304520e-01 -7.39155337e-02
1.61422610e-01 5.87629795e-01 -8.97573471e-01 -6.08121336e-01
-7.37986386e-01 4.04457062e-01 4.35586214e-01 5.31207323e-01
-5.31004190e-01 -5.39079249e-01 -8.14929605e-01 5.38290031e-02
4.24602628e-01 -6.79108739e-01 6.85473204e-01 -1.28508031e+00
-1.20980740e+00 8.11075807e-01 -5.25471747e-01 -2.44366452e-01
3.15138549e-01 -2.02958420e-01 -4.30924952e-01 3.60039979e-01
1.23731479e-01 1.29901573e-01 1.33003676e+00 -1.33070397e+00
-2.47126117e-01 -7.17919767e-01 -6.46198466e-02 4.21960145e-01
-7.38879442e-01 5.13442010e-02 -3.55877697e-01 -6.45695210e-01
8.44373643e-01 -1.47872734e+00 -6.31250203e-01 -3.75551343e-01
-5.45691192e-01 2.88264602e-01 8.35298121e-01 -7.87748039e-01
1.10099721e+00 -2.31813383e+00 7.85657585e-01 7.06004202e-01
2.02743530e-01 1.47014201e-01 -4.36736554e-01 5.89501143e-01
-7.96942770e-01 -3.29277277e-01 -7.19129801e-01 -3.55638653e-01
-2.06130937e-01 3.12429845e-01 -6.14329696e-01 9.77079213e-01
-2.90508240e-01 4.35824811e-01 -8.89637291e-01 3.03224772e-01
1.06314324e-01 3.07349324e-01 -3.83534819e-01 -1.33216992e-01
1.10322088e-01 3.49142104e-01 8.99908170e-02 4.31574821e-01
1.17929983e+00 -1.94132745e-01 3.83046985e-01 -2.25617006e-01
-1.78840384e-01 -3.75264823e-01 -1.76754260e+00 1.71456957e+00
-1.05394796e-01 4.80407953e-01 6.52669132e-01 -1.04395068e+00
4.49327052e-01 3.96369904e-01 6.68526947e-01 -2.32986823e-01
-1.62229985e-01 3.24750811e-01 1.69098768e-02 -2.00962558e-01
4.43646461e-01 -3.15648377e-01 4.71929312e-01 7.53447056e-01
-2.57035285e-01 -4.60568517e-02 7.16621697e-01 5.68539262e-01
6.70858920e-01 -5.57892561e-01 2.24962980e-01 -7.56227255e-01
7.04986691e-01 2.29171477e-02 6.48003757e-01 5.80733061e-01
1.92424789e-01 6.80830717e-01 3.04635137e-01 -1.57084521e-02
-7.81606376e-01 -7.43224502e-01 -2.99262285e-01 7.25295782e-01
-3.08316685e-02 -5.55439591e-01 -4.87803340e-01 -2.72106469e-01
2.43555717e-02 3.96382719e-01 -5.55673480e-01 1.70455888e-01
-7.64774755e-02 -1.59384191e+00 -6.36484846e-02 2.91943133e-01
1.93334773e-01 -1.71447262e-01 1.58222109e-01 -2.16712300e-02
-3.57931137e-01 -1.02480865e+00 -3.04333478e-01 4.50947881e-01
-9.57495093e-01 -9.18707728e-01 -8.72323573e-01 -5.22440910e-01
1.06486404e+00 1.10202098e+00 5.52020550e-01 -3.44016403e-01
7.17609897e-02 3.91462713e-01 -2.50021785e-01 -9.58285555e-02
-2.80375153e-01 -3.00308406e-01 4.83490407e-01 5.76790392e-01
-1.21700227e-01 -5.69226444e-01 -1.67832911e-01 1.65072829e-01
-1.34848082e+00 2.95907438e-01 1.66730762e-01 9.93139803e-01
8.73035312e-01 3.77587587e-01 4.10748832e-02 -1.13985074e+00
5.49571514e-01 -2.98150241e-01 -7.06256092e-01 4.37932551e-01
-4.34604526e-01 -6.77131414e-02 5.90176642e-01 -1.74887523e-01
-7.61514425e-01 4.53825384e-01 4.27588433e-01 -5.59229791e-01
4.33323801e-01 8.49206924e-01 -5.55045068e-01 -3.54354352e-01
7.54570484e-01 3.32797617e-01 -9.33104195e-03 -3.81005645e-01
5.36146462e-01 7.28372991e-01 5.48195243e-01 -5.60408056e-01
1.32036328e+00 1.15094674e+00 4.19743568e-01 -1.08777738e+00
-1.01719403e+00 -1.01463306e+00 -6.16915286e-01 -8.95803515e-03
3.90900671e-01 -1.21440458e+00 -1.67112067e-01 3.27991962e-01
-6.77050173e-01 -8.13758820e-02 -1.55178845e-01 6.39514029e-01
-4.14243847e-01 7.99599171e-01 -4.39234942e-01 -8.77040863e-01
-2.98619866e-01 -9.22963202e-01 6.88269734e-01 -1.56519994e-01
2.36488253e-01 -7.71004796e-01 2.33199313e-01 6.71255708e-01
1.13690468e-02 1.37592271e-01 8.86913478e-01 -2.46361136e-01
-2.83205450e-01 -2.86160201e-01 -3.19741488e-01 6.90997720e-01
1.68924063e-01 3.62446043e-03 -1.01550448e+00 -8.04243386e-01
3.12343121e-01 1.04321577e-01 1.19325614e+00 4.30639356e-01
7.85116613e-01 -2.30012447e-01 -2.37848103e-01 9.19725358e-01
1.63915575e+00 -5.38942479e-02 4.83359993e-01 8.15281495e-02
9.80202854e-01 5.32461107e-01 4.93557006e-01 4.87460136e-01
-3.46965998e-01 2.75043637e-01 2.24759817e-01 -3.20314914e-01
3.93513471e-01 3.74194235e-01 3.42310369e-01 9.81694162e-01
-3.06179613e-01 -2.26579905e-01 -9.64350939e-01 2.16825575e-01
-2.11043525e+00 -1.14685953e+00 -7.25716114e-01 2.40414739e+00
4.45117563e-01 -5.40687799e-01 -7.60693196e-03 6.94419920e-01
6.70503736e-01 2.63565361e-01 -3.20515215e-01 -4.18526828e-02
-7.99137712e-01 6.51621670e-02 8.08294654e-01 8.43813479e-01
-1.25254154e+00 4.43235815e-01 6.39239216e+00 6.32484794e-01
-1.00284803e+00 1.92590535e-01 3.26990932e-01 -8.93364921e-02
-5.12441278e-01 1.44671813e-01 -2.73529321e-01 7.73909129e-03
4.78885531e-01 -2.30666593e-01 8.36478412e-01 4.92991328e-01
3.51972133e-01 -1.64282739e-01 -7.68948197e-01 1.41040456e+00
3.35247785e-01 -1.10795057e+00 -2.24287598e-03 2.17702791e-01
1.36616504e+00 -6.70741349e-02 2.40784615e-01 -3.78287464e-01
-2.80148052e-02 -8.64166856e-01 4.52735931e-01 1.17003120e-01
7.90243268e-01 -8.10615778e-01 2.68533766e-01 5.20633996e-01
-9.09720600e-01 -3.25842798e-01 -5.68808258e-01 -9.08821076e-02
1.47753909e-01 1.15288496e+00 -3.77390563e-01 8.82416248e-01
2.92455912e-01 9.16508138e-01 -2.04717904e-01 1.07684040e+00
-2.69195646e-01 5.86635888e-01 -6.78390026e-01 6.82049096e-01
2.33495265e-01 -1.15048909e+00 9.00388122e-01 8.83507133e-01
3.08453083e-01 3.14566374e-01 3.24514866e-01 3.99683654e-01
6.39499202e-02 4.61436480e-01 -4.84774381e-01 -4.90180224e-01
-1.47439301e-01 1.39687061e+00 -8.35237980e-01 -2.52333879e-01
-5.59965432e-01 1.25739312e+00 5.93108833e-02 8.40600431e-01
-4.10120249e-01 1.84484646e-01 5.00213325e-01 -1.60734296e-01
1.22357413e-01 -4.24935609e-01 -2.10736021e-01 -1.71777105e+00
9.68830884e-02 -1.16333508e+00 5.57615817e-01 -6.52534544e-01
-1.17056143e+00 3.77023488e-01 -2.55459279e-01 -1.26961076e+00
3.40347141e-01 -7.28687167e-01 -2.41914973e-01 1.07101190e+00
-1.33902943e+00 -9.35946882e-01 -2.96404302e-01 9.18399274e-01
-1.34370714e-01 -5.45896813e-02 9.78713334e-01 4.01327223e-01
-1.01343310e+00 -5.77193871e-02 9.66934800e-01 -1.93347126e-01
6.80428743e-01 -1.20611823e+00 -3.10709327e-01 1.46859980e+00
5.15908539e-01 8.37999046e-01 7.44849861e-01 -5.08766115e-01
-1.81571245e+00 -9.53100562e-01 4.11864728e-01 4.20402959e-02
7.91698456e-01 -1.83338329e-01 -6.02569938e-01 7.03871429e-01
9.79196653e-02 6.60821274e-02 1.31672144e+00 2.07556620e-01
-5.42396009e-01 -1.90999344e-01 -8.19485486e-01 3.11109245e-01
8.66635144e-01 -6.08999431e-01 -2.60171235e-01 1.19211364e+00
1.68363124e-01 -3.17567348e-01 -4.81901556e-01 4.11740899e-01
2.73807764e-01 -8.23669434e-01 9.34829116e-01 -8.03792357e-01
5.22279702e-02 -7.13676572e-01 -5.33145070e-01 -1.29819655e+00
-6.07168555e-01 -9.75969255e-01 -1.39508754e-01 6.97838783e-01
4.09409612e-01 -8.24234068e-01 7.59664118e-01 5.78931451e-01
-2.28106119e-02 -2.50444353e-01 -7.23054051e-01 -1.07889736e+00
-2.41832316e-01 -4.37322974e-01 2.15143532e-01 1.26982689e+00
1.37074828e-01 2.81485021e-01 -6.50052845e-01 6.59322619e-01
8.49410772e-01 7.16018617e-01 5.57550073e-01 -1.28316844e+00
-5.35857797e-01 1.36014953e-01 -1.41339883e-01 -9.17459071e-01
4.15170163e-01 -1.23556745e+00 -1.63586572e-01 -1.38941336e+00
6.26116335e-01 -4.03333038e-01 -1.41215354e-01 4.83215541e-01
-2.19831139e-01 4.81097728e-01 3.69473070e-01 5.56799293e-01
-1.33442447e-01 3.55318218e-01 1.22301257e+00 -4.37110513e-01
-3.01429003e-01 -1.39609836e-02 -7.87649095e-01 4.77373362e-01
4.65752155e-01 -3.70035917e-01 -4.69723701e-01 -4.36012238e-01
6.77731872e-01 1.65492699e-01 -1.41307801e-01 -7.82731354e-01
7.94505477e-02 -3.75112623e-01 1.15831114e-01 -4.72630888e-01
4.19981211e-01 -1.01632071e+00 6.07931197e-01 3.99136782e-01
4.63801064e-02 -1.48538008e-01 1.02677383e-01 7.21047342e-01
-3.52762640e-01 -2.72363901e-01 8.07032287e-01 2.22092077e-01
-4.21876103e-01 4.25688297e-01 -2.19438463e-01 -3.06401849e-01
9.82522011e-01 -2.30120584e-01 -4.81285378e-02 -3.79211128e-01
-1.02939844e+00 6.08546324e-02 5.06824553e-01 -1.37272686e-01
2.88475096e-01 -1.23003173e+00 -8.51296842e-01 3.57918471e-01
2.62672156e-02 -2.10654289e-01 4.33045954e-01 1.28474653e+00
-5.85812330e-01 4.32899684e-01 1.34843707e-01 -6.40394151e-01
-1.60221040e+00 7.43791282e-01 1.42197788e-01 -1.27894461e-01
-3.42129618e-01 9.92232680e-01 2.42672205e-01 -2.57886171e-01
-1.52270570e-01 -1.22679591e-01 1.57540619e-01 3.05476785e-01
9.06769633e-01 5.59917748e-01 2.32460901e-01 -9.26357210e-01
-1.69225439e-01 4.11562085e-01 1.90818667e-01 -2.85711884e-01
1.49336886e+00 -3.03334054e-02 -9.80985582e-01 1.54970527e-01
1.16797483e+00 6.03485167e-01 -8.49836290e-01 -2.15094641e-01
-4.22494531e-01 -7.20815659e-01 2.49667361e-01 -3.39014679e-01
-1.24340129e+00 6.88815117e-01 3.66636932e-01 2.95387328e-01
1.43971443e+00 -6.05692506e-01 4.65997815e-01 6.89058542e-01
3.29691350e-01 -7.68550575e-01 -5.80811858e-01 4.40472305e-01
8.71648371e-01 -1.17812812e+00 3.54560882e-01 -1.15794885e+00
-2.77741015e-01 1.29691195e+00 -3.71885970e-02 7.93644935e-02
6.22907817e-01 2.77458876e-01 -2.00875793e-02 5.89058064e-02
-2.03930721e-01 -2.12821290e-01 3.39185804e-01 3.84445369e-01
-7.00841006e-03 4.50086057e-01 -2.60013938e-01 2.14663759e-01
1.60433308e-04 -3.50087196e-01 5.96025348e-01 8.36556435e-01
-3.27397525e-01 -1.23309755e+00 -1.05300820e+00 5.71213007e-01
-4.44664396e-02 -3.34184170e-01 -4.66848373e-01 9.27892849e-02
-1.90579534e-01 1.02305090e+00 -4.09705490e-01 -4.00497317e-02
-1.76800802e-01 1.08664567e-02 6.69206917e-01 -9.42468643e-01
-1.18873604e-01 3.43685150e-01 -2.88488623e-02 -3.24239701e-01
-8.04962516e-01 -9.77929831e-01 -1.04791355e+00 -1.95629314e-01
-6.50240421e-01 4.12206948e-01 2.94961751e-01 8.55718136e-01
1.62833482e-01 -1.92055255e-01 7.31426775e-01 -7.00973511e-01
-5.05163193e-01 -6.46784842e-01 -1.15325320e+00 4.41279531e-01
2.75907665e-01 -2.84917176e-01 -7.06188798e-01 9.71144587e-02] | [10.04980182647705, -1.9789663553237915] |
00e1c27a-2930-44c0-93d5-e5fb911237d5 | learning-to-disentangle-interleaved | null | null | https://aclanthology.org/N18-1164 | https://aclanthology.org/N18-1164.pdf | Learning to Disentangle Interleaved Conversational Threads with a Siamese Hierarchical Network and Similarity Ranking | An enormous amount of conversation occurs online every day, such as on chat platforms where multiple conversations may take place concurrently. Interleaved conversations lead to difficulties in not only following discussions but also retrieving relevant information from simultaneous messages. Conversation disentanglement aims to separate intermingled messages into detached conversations. In this paper, we propose to leverage representation learning for conversation disentanglement. A Siamese hierarchical convolutional neural network (SHCNN), which integrates local and more global representations of a message, is first presented to estimate the conversation-level similarity between closely posted messages. With the estimated similarity scores, our algorithm for conversation identification by similarity ranking (CISIR) then derives conversations based on high-confidence message pairs and pairwise redundancy. Experiments were conducted with four publicly available datasets of conversations from Reddit and IRC channels. The experimental results show that our approach significantly outperforms comparative baselines in both pairwise similarity estimation and conversation disentanglement. | ['Yan-Ying Chen', 'Jyun-Yu Jiang', 'Francine Chen', 'Wei Wang'] | 2018-06-01 | null | null | null | naacl-2018-6 | ['conversation-disentanglement'] | ['natural-language-processing'] | [ 3.55673224e-01 -7.70867104e-03 8.49145055e-02 -6.36144876e-01
-1.19872582e+00 -6.82678461e-01 1.14343131e+00 1.59845814e-01
-1.93225041e-01 7.45109379e-01 1.12656510e+00 -1.03148520e-01
-3.77523601e-02 -4.52552885e-01 -1.06808744e-01 -5.41290224e-01
-2.44895324e-01 7.10038185e-01 -3.34019333e-01 -4.28028017e-01
4.09925461e-01 2.99636513e-01 -1.50044191e+00 1.00693178e+00
4.83807653e-01 7.80304193e-01 8.50393176e-02 1.10695779e+00
-5.28416216e-01 1.04115474e+00 -8.25777829e-01 -5.94763339e-01
6.75194561e-02 -6.23475730e-01 -1.13799167e+00 2.95170337e-01
3.41795772e-01 -4.52244997e-01 -5.42217374e-01 4.58874732e-01
4.93827909e-01 4.23508555e-01 5.70059419e-01 -1.35468447e+00
-3.07108253e-01 9.87852097e-01 -2.56120950e-01 4.70815569e-01
8.08073819e-01 -1.40120417e-01 1.55850542e+00 -1.03433740e+00
4.36719328e-01 1.56969488e+00 5.27105749e-01 3.05763215e-01
-1.27796209e+00 -7.97127426e-01 1.63879588e-01 2.21589804e-01
-1.17830789e+00 -6.10592127e-01 6.43324316e-01 -3.14812571e-01
1.06877935e+00 6.84930801e-01 2.74641633e-01 1.39832699e+00
-2.52421379e-01 9.57838356e-01 7.97218084e-01 -1.02100514e-01
-2.01361582e-01 1.10553943e-01 3.78736168e-01 1.06760912e-01
-1.97274312e-01 -3.32981825e-01 -9.77337420e-01 -5.51166117e-01
3.12451571e-01 -2.72432361e-02 -3.26181114e-01 1.61098033e-01
-1.51018536e+00 9.29773867e-01 3.34282547e-01 3.76095325e-01
-1.98428199e-01 -3.94172341e-01 7.63406575e-01 7.38764465e-01
4.76899385e-01 5.33136904e-01 -4.92951311e-02 -6.25235736e-01
-4.61886972e-01 2.88267165e-01 1.30881453e+00 9.44825470e-01
8.02370310e-01 -6.18238628e-01 -4.44742203e-01 1.16218913e+00
-2.13446766e-02 -4.81017940e-02 5.63482225e-01 -1.06190336e+00
9.58907604e-01 7.46208489e-01 1.21029496e-01 -1.20335734e+00
-3.67850095e-01 7.15841725e-02 -1.15225351e+00 -4.39285904e-01
4.64630872e-01 -3.67607445e-01 -2.16286760e-02 1.47331953e+00
9.25731361e-02 4.08411503e-01 1.94082394e-01 7.40297973e-01
1.17365611e+00 6.02959812e-01 -4.70717043e-01 -2.34858096e-01
1.37882674e+00 -1.10867465e+00 -8.10033798e-01 -2.80190930e-02
5.69092095e-01 -1.00493205e+00 7.69173801e-01 -7.10385293e-03
-1.04851377e+00 -3.68447632e-01 -8.49715412e-01 -3.41116279e-01
-1.87511727e-01 -2.31970862e-01 5.28189242e-01 3.05718422e-01
-1.06817496e+00 4.19522256e-01 -1.41797379e-01 -3.28515917e-01
1.04642987e-01 2.82889396e-01 -5.58191121e-01 1.68027401e-01
-1.49029636e+00 8.80815923e-01 -4.76328284e-02 2.99575835e-01
-4.31524754e-01 -4.61309195e-01 -8.48407924e-01 3.51838231e-01
-3.98585834e-02 -3.76881152e-01 1.48963106e+00 -9.35904741e-01
-1.64955580e+00 8.95545661e-01 -4.45707470e-01 -5.19821703e-01
5.78656018e-01 -1.76222786e-01 -3.72165114e-01 1.08457273e-02
3.60613875e-02 3.56208891e-01 4.43289757e-01 -1.10927522e+00
-5.33597469e-01 4.58475016e-02 1.77714750e-01 5.85797787e-01
-9.24547315e-02 2.14283988e-01 -3.23816806e-01 -3.09814483e-01
6.58913478e-02 -1.03903198e+00 1.77381635e-01 -2.74824888e-01
-6.60659969e-01 -5.29346645e-01 6.34933710e-01 -6.15062773e-01
9.59542751e-01 -1.94762862e+00 1.97025582e-01 1.82226717e-01
4.47343737e-01 1.29815014e-02 -4.20666963e-01 1.08193672e+00
7.56696016e-02 -9.00511444e-02 -6.58021495e-03 -8.61861408e-01
1.16370983e-01 3.84691097e-02 -3.49650145e-01 3.25336605e-01
-1.45470845e-02 7.24583864e-01 -8.78987253e-01 -2.54807085e-01
4.00950499e-02 5.55546284e-01 -3.34964663e-01 6.42889261e-01
1.99243560e-01 6.24785721e-01 -2.98330728e-02 5.92698753e-02
7.86770821e-01 -6.04781568e-01 7.63375223e-01 1.16972886e-01
1.67886019e-01 1.11111128e+00 -7.66750336e-01 1.58581734e+00
-8.54294002e-01 1.27212107e+00 1.46438971e-01 -7.33880222e-01
1.00027132e+00 5.02665818e-01 2.57360458e-01 -3.89013052e-01
2.15691790e-01 -2.72389144e-01 9.08792987e-02 -3.54154944e-01
7.71750748e-01 3.19222599e-01 -3.34474564e-01 1.08441257e+00
-2.32132003e-01 2.39187703e-01 7.37494379e-02 7.28263438e-01
1.06758916e+00 -6.82252169e-01 4.47379917e-01 -2.47348547e-02
7.52162635e-01 -6.09465599e-01 3.27314854e-01 7.27239430e-01
-3.05822581e-01 8.23835194e-01 8.01267862e-01 -3.26986700e-01
-6.66635215e-01 -1.23516810e+00 1.43902823e-01 1.28306019e+00
1.85231790e-01 -5.60078502e-01 -4.03166652e-01 -6.92602456e-01
1.52245536e-01 4.80388254e-01 -6.04546964e-01 1.00305103e-01
-7.02022612e-01 -3.56348425e-01 5.53872883e-01 3.81154776e-01
4.02375579e-01 -9.32751954e-01 4.49161291e-01 8.03495273e-02
-9.25522625e-01 -1.23425376e+00 -9.70723629e-01 -4.65633631e-01
-3.52133989e-01 -1.12413025e+00 -4.98248965e-01 -7.81309068e-01
4.16170448e-01 1.03795207e+00 1.28277957e+00 1.90804079e-01
4.37846221e-02 1.99013185e-02 -3.69886965e-01 2.92402059e-01
-6.37741387e-01 2.78243065e-01 -4.15621437e-02 4.07806128e-01
6.73990428e-01 -8.54829848e-01 -5.57924211e-01 6.34924948e-01
-4.41919416e-01 8.76764879e-02 3.93691212e-01 9.07843590e-01
-3.85741830e-01 -5.44039965e-01 9.74118769e-01 -8.81225884e-01
1.23081434e+00 -7.99108684e-01 1.39318379e-02 1.97778150e-01
-6.40745983e-02 -9.41377878e-02 3.42284828e-01 -2.63052493e-01
-1.35237646e+00 -5.00173092e-01 1.87792536e-02 1.23020031e-01
-1.50572971e-01 2.64917284e-01 -8.06017146e-02 5.68578720e-01
4.45957333e-01 3.18821281e-01 2.69523025e-01 -4.16203827e-01
3.72206777e-01 1.33196032e+00 2.71379203e-01 -3.88442874e-01
4.29679155e-01 4.05842900e-01 -8.90942574e-01 -8.10040891e-01
-6.94202900e-01 -9.39646482e-01 -4.85586524e-01 -2.63220280e-01
5.77791750e-01 -1.12931311e+00 -1.55769634e+00 3.46203804e-01
-1.51102686e+00 4.68276255e-02 5.28093874e-01 4.93405491e-01
-2.71014482e-01 6.13627851e-01 -1.02161896e+00 -9.46953475e-01
-4.90598917e-01 -9.86685932e-01 1.03741932e+00 1.22602715e-03
-9.84273732e-01 -9.58156288e-01 2.20752046e-01 1.07278907e+00
4.13254410e-01 -1.75285578e-01 4.92414117e-01 -1.19181108e+00
-6.31562173e-01 -3.14779162e-01 -5.14683187e-01 2.01322079e-01
5.27864635e-01 -2.18513891e-01 -1.22185218e+00 -3.11391979e-01
-4.33885545e-01 -4.11004335e-01 8.14414978e-01 -2.36281946e-01
8.63735974e-01 -6.21051252e-01 -2.41098091e-01 2.93134321e-02
3.98060441e-01 -1.67800918e-01 5.53698480e-01 -4.36295308e-02
5.60342014e-01 1.05549812e+00 2.47605994e-01 6.00744903e-01
6.64591372e-01 6.86102569e-01 1.56848475e-01 4.07603145e-01
6.29191473e-02 -3.72185521e-02 3.77837986e-01 1.37081587e+00
2.15655208e-01 -4.46945280e-01 -5.64997375e-01 5.30817270e-01
-2.03248119e+00 -1.20340312e+00 -4.41178381e-02 2.04604840e+00
9.67989504e-01 3.43503058e-02 1.95675656e-01 -2.27813348e-01
1.21282101e+00 3.82145643e-01 -2.75893241e-01 -5.26342452e-01
-6.33246005e-02 -4.89902258e-01 -1.57017663e-01 9.27057207e-01
-8.71323407e-01 2.80863464e-01 6.13300562e+00 4.41788584e-01
-5.48510730e-01 -7.64089823e-02 6.19370461e-01 -3.79894763e-01
-4.53579307e-01 -2.10483953e-01 -7.51510561e-01 7.01762378e-01
8.96707237e-01 -3.98910493e-01 6.01521432e-01 5.14586270e-01
1.65518105e-01 1.69259503e-01 -1.46006393e+00 1.15516567e+00
2.61804700e-01 -1.56135213e+00 8.20993260e-02 -1.73126474e-01
6.94709778e-01 1.31738499e-01 -2.32579365e-01 4.08875972e-01
6.64076447e-01 -9.60770905e-01 -5.22244088e-02 4.68441367e-01
2.03358367e-01 -9.04238045e-01 8.92868817e-01 3.39335829e-01
-1.12757337e+00 1.77402020e-01 3.40120457e-02 -4.17317301e-01
1.99349880e-01 4.49101686e-01 -1.36858463e+00 3.18077564e-01
4.02853817e-01 8.61822665e-01 -3.23137715e-02 5.56273758e-01
2.49492452e-02 1.90411106e-01 -5.37740588e-02 -3.77788454e-01
1.08813725e-01 -3.11955869e-01 5.69180131e-01 1.52326250e+00
-6.72941282e-03 -1.60984114e-01 -3.48225236e-04 6.62511289e-01
-7.10082352e-01 -1.25732394e-02 -5.78105688e-01 -5.92988543e-02
1.05979204e+00 1.41128695e+00 -2.47902036e-01 -5.77591121e-01
-4.96935457e-01 1.08557367e+00 6.93936646e-01 2.93855667e-01
-4.38622326e-01 -4.61613208e-01 1.51057589e+00 -5.64563751e-01
-1.58153940e-02 -7.31832609e-02 -1.22781895e-01 -1.27214134e+00
1.51657030e-01 -9.23401654e-01 4.49408054e-01 -4.99224663e-01
-1.95853388e+00 6.96957588e-01 -2.89241642e-01 -1.25340486e+00
-6.16119206e-01 -4.20437641e-02 -1.10069466e+00 1.04185593e+00
-1.22666395e+00 -7.59940445e-01 -4.32887107e-01 4.49474424e-01
9.72971737e-01 -1.91374645e-01 9.72971141e-01 2.07124114e-01
-5.63479006e-01 9.63635921e-01 3.04317147e-01 6.06106281e-01
9.35626686e-01 -1.10592425e+00 8.67695093e-01 1.25014558e-01
2.20817868e-02 9.68226612e-01 5.74408710e-01 -2.81771690e-01
-9.99298811e-01 -7.47031868e-01 1.63096333e+00 -6.63014352e-01
7.97265828e-01 -9.30446088e-01 -1.10569978e+00 6.63870752e-01
6.94832504e-01 -6.07781112e-01 1.25744498e+00 6.21663153e-01
-7.44879186e-01 -5.47993816e-02 -8.19633901e-01 8.70591402e-01
9.26685631e-01 -1.27720487e+00 -6.74211085e-01 5.32782435e-01
8.45639706e-01 -2.27250621e-01 -7.67757237e-01 -2.88903028e-01
8.06844413e-01 -1.14578259e+00 9.66067314e-01 -6.45431757e-01
4.27100778e-01 2.40881264e-01 -4.63042110e-02 -1.26153743e+00
-3.19731086e-02 -1.17512226e+00 -2.22204882e-03 1.28839540e+00
5.25530159e-01 -6.72929525e-01 6.95527554e-01 6.69048965e-01
1.73073694e-01 -3.24847937e-01 -8.23666871e-01 -4.82162148e-01
-1.61924049e-01 -1.03224576e-01 7.75057316e-01 1.26739562e+00
8.25264215e-01 8.42148125e-01 -5.65203786e-01 6.23308308e-02
2.14593470e-01 5.05877733e-01 1.08964932e+00 -1.29293489e+00
-3.51563305e-01 -5.08101761e-01 -2.74579734e-01 -1.42424738e+00
5.40189743e-01 -8.76001835e-01 2.90764570e-01 -1.38938367e+00
4.68491942e-01 -1.12565078e-01 -7.07888678e-02 -1.79457381e-01
-3.41212273e-01 -3.85876559e-02 8.28086212e-02 3.44949841e-01
-8.71409237e-01 5.68441451e-01 9.67118561e-01 -3.12162101e-01
-1.21940173e-01 1.99568272e-01 -7.99868405e-01 4.45758104e-01
9.34998691e-01 -1.37336597e-01 -4.15703535e-01 -1.88272044e-01
1.98708132e-01 3.38679552e-01 3.03535555e-02 -3.54237050e-01
3.08218867e-01 1.44442916e-01 -1.92531645e-01 -6.00581408e-01
8.06619942e-01 -4.47126418e-01 -4.57732350e-01 -9.01042521e-02
-1.23813570e+00 1.46287128e-01 -1.73539281e-01 9.09700394e-01
-4.31706518e-01 2.48188123e-01 3.39128971e-01 8.58385637e-02
-1.51076093e-01 -1.01906471e-01 -8.87992620e-01 1.71924487e-01
5.65288782e-01 1.04602925e-01 -6.42210007e-01 -1.26862729e+00
-7.62996197e-01 4.07718271e-01 -7.52393678e-02 9.67185318e-01
6.14648759e-01 -1.49162900e+00 -1.06565368e+00 3.37410569e-02
2.94740111e-01 -2.64331728e-01 4.80646223e-01 7.83259094e-01
-7.93636963e-02 3.76189440e-01 7.40225539e-02 -4.74443644e-01
-1.84809446e+00 7.67181888e-02 1.13891959e-01 -2.69893259e-01
-4.34901744e-01 1.03265333e+00 2.71151960e-01 -9.19368982e-01
5.60267508e-01 -2.58731455e-01 -2.20682800e-01 4.03516531e-01
9.90297139e-01 4.59014922e-01 1.09605767e-01 -7.36188889e-01
-3.49366397e-01 -1.36973009e-01 -6.78677559e-01 -9.78307724e-02
8.97774756e-01 -6.10057294e-01 -2.55955428e-01 5.40898561e-01
1.98997247e+00 -9.93649587e-02 -1.08942389e+00 -7.69649446e-01
5.80444634e-02 -6.71394587e-01 -4.38001543e-01 -4.95212942e-01
-5.77190876e-01 7.94464648e-01 -1.30520612e-01 5.79546332e-01
5.07894814e-01 1.82659939e-01 1.07995927e+00 9.20767844e-01
-1.81822732e-01 -9.06336784e-01 4.00837749e-01 7.91673958e-01
1.16976023e+00 -1.63057578e+00 -2.61913210e-01 -4.11928892e-01
-9.88252401e-01 9.77784336e-01 4.52036291e-01 1.04643935e-02
5.65234721e-01 1.36673227e-01 2.36276627e-01 -2.70839203e-02
-1.22341073e+00 -6.95174411e-02 2.31922030e-01 3.63273382e-01
7.90174782e-01 -2.81348880e-02 3.81541587e-02 2.75136173e-01
-2.84834951e-01 -6.80208862e-01 5.56479752e-01 6.26841605e-01
-3.29531461e-01 -1.00661504e+00 -1.34941846e-01 3.45977694e-01
-1.33293912e-01 -3.45627457e-01 -1.03759599e+00 6.05277240e-01
-6.51611507e-01 1.54672909e+00 6.63380146e-01 -5.45360863e-01
-9.38436091e-02 8.31498578e-02 -1.50425836e-01 -5.08660436e-01
-9.57164288e-01 -3.40210408e-01 9.61260498e-01 -4.41295773e-01
-4.30993557e-01 -7.21729815e-01 -1.09765673e+00 -8.37754786e-01
-3.14258128e-01 5.18786252e-01 4.34821725e-01 1.07628238e+00
6.66670322e-01 3.71391147e-01 1.15765393e+00 -9.46156442e-01
-4.74753231e-01 -1.30223083e+00 -2.90041059e-01 6.86964095e-01
6.65711880e-01 -3.64477873e-01 -7.04316616e-01 -1.94470897e-01] | [12.638164520263672, 7.7619500160217285] |
13c9ae6e-9de5-4dcd-8c76-b5baa18004d4 | artificial-intelligence-solution-for | 2203.05563 | null | https://arxiv.org/abs/2203.05563v1 | https://arxiv.org/pdf/2203.05563v1.pdf | Artificial Intelligence Solution for Effective Treatment Planning for Glioblastoma Patients | Glioblastomas are the most common malignant brain tumors in adults. Approximately 200000 people die each year from Glioblastoma in the world. Glioblastoma patients have a median survival of 12 months with optimal therapy and about 4 months without treatment. Glioblastomas appear as heterogeneous necrotic masses with irregular peripheral enhancement, surrounded by vasogenic edema. The current standard of care includes surgical resection, radiotherapy and chemotherapy, which require accurate segmentation of brain tumor subregions. For effective treatment planning, it is vital to identify the methylation status of the promoter of Methylguanine Methyltransferase (MGMT), a positive prognostic factor for chemotherapy. However, current methods for brain tumor segmentation are tedious, subjective and not scalable, and current techniques to determine the methylation status of MGMT promoter involve surgically invasive procedures, which are expensive and time consuming. Hence there is a pressing need to develop automated tools to segment brain tumors and non-invasive methods to predict methylation status of MGMT promoter, to facilitate better treatment planning and improve survival rate. I created an integrated diagnostics solution powered by Artificial Intelligence to automatically segment brain tumor subregions and predict MGMT promoter methylation status, using brain MRI scans. My AI solution is proven on large datasets with performance exceeding current standards and field tested with data from teaching files of local neuroradiologists. With my solution, physicians can submit brain MRI images, and get segmentation and methylation predictions in minutes, and guide brain tumor patients with effective treatment planning and ultimately improve survival time. | ['Vikram Goddla'] | 2022-03-09 | null | null | null | null | ['brain-tumor-segmentation'] | ['medical'] | [ 4.03203100e-01 2.48068303e-01 -3.63687962e-01 -1.63964540e-01
-6.61008239e-01 -3.95543486e-01 2.26959571e-01 2.40913495e-01
-6.84357285e-01 8.93964589e-01 4.43469226e-01 -6.53921008e-01
2.13693708e-01 -5.87824583e-01 1.27350107e-01 -1.19166839e+00
1.84378382e-02 7.82670557e-01 5.02907149e-02 2.97527254e-01
2.95704216e-01 5.17552555e-01 -1.10716963e+00 -5.57423048e-02
1.50249958e+00 6.36538386e-01 5.89535356e-01 6.98933542e-01
-2.60554254e-01 4.51243997e-01 -4.57502276e-01 2.20640942e-01
2.34697666e-02 -4.21957403e-01 -7.04114914e-01 7.44957998e-02
-3.11791509e-01 -2.86407351e-01 -1.75733462e-01 1.22261333e+00
4.13316518e-01 -3.33507955e-01 1.05072868e+00 -1.02031898e+00
2.07091242e-01 3.31122607e-01 -5.20216644e-01 2.63853818e-01
6.69141263e-02 3.46535474e-01 3.68318588e-01 -4.83985782e-01
5.75311303e-01 1.28396273e-01 3.32128316e-01 8.10568273e-01
-6.96987033e-01 -6.23322070e-01 -2.02710435e-01 4.03178424e-01
-1.13068414e+00 -2.68593043e-01 -1.25378564e-01 -7.96521842e-01
1.11822200e+00 3.94739032e-01 1.32272482e+00 7.45322168e-01
9.99606729e-01 6.52317703e-01 1.08040059e+00 -8.88617039e-02
6.31952703e-01 -3.37210298e-01 3.23695868e-01 1.00386178e+00
2.39869997e-01 -1.84556432e-02 -4.07714576e-01 -4.85993214e-02
4.99073088e-01 3.79233569e-01 -7.46797919e-01 9.44469124e-03
-1.27861989e+00 6.97620213e-01 1.99196264e-01 4.63172168e-01
-2.49798551e-01 -9.90176499e-02 3.28386456e-01 -1.74018919e-01
2.69232154e-01 -1.46460146e-01 -1.81266099e-01 -2.08254024e-01
-1.00543773e+00 -9.29459855e-02 5.38787961e-01 5.67980170e-01
2.78680086e-01 -4.23061222e-01 6.68707043e-02 7.02727616e-01
5.07009208e-01 5.57199061e-01 1.29876769e+00 -5.44897020e-01
-3.47232670e-02 7.41420329e-01 -1.81135669e-01 -1.91120148e-01
-1.04309106e+00 -2.97236443e-01 -1.05678594e+00 4.24614429e-01
6.12048626e-01 -3.97833943e-01 -1.41377103e+00 1.26127207e+00
-9.52653587e-03 5.23598380e-02 6.24987576e-03 9.07379985e-01
6.20598912e-01 2.54306674e-01 1.39027983e-01 -3.63389015e-01
1.63492715e+00 -8.23886693e-01 -6.12909615e-01 -4.84700978e-01
1.36280167e+00 -2.57962972e-01 5.17910063e-01 4.01470065e-01
-7.47134089e-01 6.52979314e-01 -1.04444432e+00 2.85548210e-01
-2.92979509e-01 -3.79462987e-01 8.10024023e-01 1.05597675e+00
-1.13348520e+00 9.24473107e-02 -1.36000276e+00 -4.94111478e-01
8.06347549e-01 5.01262486e-01 -6.26546860e-01 -2.52662927e-01
-8.10865402e-01 1.29134667e+00 5.29237926e-01 -2.84471393e-01
-7.05464721e-01 -8.46393228e-01 -6.21595502e-01 -5.70685044e-02
-1.17075503e-01 -9.17179883e-01 1.26943839e+00 -6.98967814e-01
-1.49191451e+00 8.22244525e-01 -6.69285595e-01 -4.73508954e-01
3.46623302e-01 6.16906881e-01 1.36604141e-02 2.35995650e-01
4.24375683e-02 7.84750819e-01 4.57945198e-01 -3.26990098e-01
-1.08054972e+00 -8.20302010e-01 -9.82352674e-01 3.58247697e-01
5.69340289e-02 -6.96514174e-02 -1.28794713e-02 -1.79804429e-01
4.00396943e-01 -7.34942675e-01 -6.79467440e-01 -3.52454573e-01
-8.29465315e-02 5.53483330e-03 6.37858391e-01 -1.11119664e+00
6.15614951e-01 -1.78074503e+00 -4.95432951e-02 2.86513805e-01
5.87052047e-01 5.55587076e-02 3.11770886e-01 -5.33088624e-01
-3.56602222e-02 2.82211661e-01 -4.18632388e-01 3.54933739e-01
-9.57302824e-02 -1.49999812e-01 4.86229777e-01 6.53478384e-01
-3.49891335e-01 1.04704809e+00 -9.21237051e-01 -2.83933222e-01
1.87007457e-01 4.03660864e-01 -1.86704218e-01 -7.33693093e-02
-7.70305395e-02 5.00173986e-01 -3.05998117e-01 1.05885530e+00
3.25396925e-01 -1.61965445e-01 2.47968473e-02 5.72625041e-01
1.22580022e-01 -7.14681149e-02 -4.01048124e-01 1.49358487e+00
1.73497975e-01 7.57768631e-01 8.63156542e-02 -8.62454116e-01
7.05793083e-01 5.60531378e-01 7.91019619e-01 -6.63689196e-01
5.14101386e-01 4.30536240e-01 2.99545318e-01 -5.82919419e-01
-2.77073443e-01 -5.23275197e-01 4.25045222e-01 4.59960192e-01
-4.10359591e-01 -4.64595020e-01 2.81456202e-01 -3.71105745e-02
1.93306220e+00 -4.05499369e-01 6.45252347e-01 -2.85177201e-01
3.00220728e-01 1.74224004e-01 6.85248673e-01 2.79385507e-01
-6.76950514e-01 5.94066978e-01 5.28895855e-01 -2.43151501e-01
-8.06372464e-01 -1.03133178e+00 -4.21359658e-01 3.00949425e-01
-2.76778340e-01 1.63243935e-01 -1.22355342e+00 -4.51964378e-01
-4.01932955e-01 8.42175961e-01 -5.44944763e-01 9.46771130e-02
-6.16725236e-02 -1.38367343e+00 3.69566888e-01 3.51250112e-01
6.47422671e-01 -6.51109159e-01 -1.00791597e+00 6.08264267e-01
-4.13531423e-01 -5.62731087e-01 -2.98782557e-01 5.85537672e-01
-1.04042435e+00 -1.23176014e+00 -1.21609747e+00 -8.34733129e-01
1.03425992e+00 -4.39288206e-02 3.55397135e-01 1.68762445e-01
-9.02375221e-01 -1.37951434e-01 -1.95514429e-02 -7.23511398e-01
-1.25146165e-01 9.25319344e-02 -1.24882072e-01 -5.26376247e-01
6.13414526e-01 -7.16857553e-01 -6.08056366e-01 2.09289432e-01
-5.49453974e-01 6.19223177e-01 8.18460643e-01 7.99804032e-01
4.65888441e-01 -1.75640315e-01 5.01210988e-01 -5.04131019e-01
4.30941373e-01 -5.41882873e-01 -5.56484580e-01 -4.05388810e-02
-4.88853574e-01 -1.96951568e-01 2.77775586e-01 1.54172301e-01
-8.44082773e-01 2.48596326e-01 -1.44617453e-01 4.33444172e-01
-5.57205498e-01 6.85007691e-01 -2.31129453e-01 6.99089095e-02
5.59568107e-01 3.80692542e-01 3.24082822e-01 3.95778716e-01
-2.51228541e-01 8.82428110e-01 6.72847509e-01 1.13622770e-01
7.46728033e-02 4.92345691e-01 6.60379454e-02 -9.49653864e-01
-4.27417815e-01 -6.43549502e-01 -4.99382973e-01 -4.08452839e-01
1.10991299e+00 -5.34217000e-01 -5.06048620e-01 6.43759012e-01
-8.62756968e-01 -5.60975492e-01 3.41836423e-01 6.36482596e-01
-6.54656708e-01 2.22623274e-01 -6.26359105e-01 -4.00510311e-01
-7.38426268e-01 -1.46672356e+00 5.52619815e-01 5.17192781e-01
-4.92642134e-01 -9.14035559e-01 4.53511905e-03 6.58619881e-01
4.42714334e-01 2.36497328e-01 1.15564847e+00 -6.01323903e-01
-6.91038370e-01 -4.24897999e-01 -1.54101089e-01 -2.73488045e-01
3.75114918e-01 -1.31024823e-01 -7.52942979e-01 -6.12899959e-02
-1.18384495e-01 1.74124986e-01 6.88438654e-01 8.26223850e-01
8.29971731e-01 7.43129849e-02 -1.03339267e+00 7.32906282e-01
1.07201862e+00 8.23027492e-01 7.68954456e-01 5.12062490e-01
2.09822118e-01 4.32250857e-01 2.63663948e-01 1.91486314e-01
2.50701517e-01 -9.12874565e-02 5.63419878e-01 2.00849801e-01
1.36160895e-01 6.15824819e-01 2.23159000e-01 4.92748111e-01
-2.17676774e-01 -2.20674023e-01 -1.55148375e+00 5.47280312e-01
-1.53206885e+00 -9.32112992e-01 -2.48428926e-01 2.13759255e+00
8.25251460e-01 4.59684320e-02 -2.44167119e-01 1.68941751e-01
4.89942163e-01 -6.35971725e-01 -7.05518425e-01 -2.09940583e-01
-6.65182620e-02 4.23056334e-02 7.21597970e-01 6.83436155e-01
-6.96105301e-01 6.02070093e-01 6.81115055e+00 3.24561805e-01
-1.31367993e+00 2.30963141e-01 9.86170113e-01 -3.13601285e-01
-2.43276414e-02 -2.58361906e-01 -4.44574684e-01 5.40966272e-01
1.04439974e+00 -6.49815679e-01 2.28706568e-01 4.47281003e-01
4.38908458e-01 -8.13181639e-01 -8.78376186e-01 1.07588172e+00
5.76973893e-02 -1.53327262e+00 -7.29711533e-01 4.96612489e-01
5.73878765e-01 4.21749741e-01 -2.92838931e-01 -1.22660678e-02
3.54481488e-01 -1.32663596e+00 1.14098094e-01 8.32507551e-01
4.08080369e-01 -9.26090419e-01 1.16123581e+00 6.17070675e-01
-5.85057378e-01 4.23813649e-02 -4.48593050e-02 6.86124265e-02
7.63121322e-02 7.94625878e-01 -1.49058354e+00 -2.55928957e-03
4.75996137e-01 3.31669092e-01 -3.17666680e-01 1.55534804e+00
-3.52182776e-01 4.26027358e-01 -4.17894453e-01 -3.76335740e-01
7.33609200e-02 -2.46143609e-01 4.06121582e-01 8.17906857e-01
5.39690197e-01 4.78991359e-01 -5.90863936e-02 5.34324884e-01
4.03075904e-01 2.41976067e-01 -3.09064984e-01 -1.15595646e-01
2.06580326e-01 1.29846466e+00 -1.35407877e+00 -3.18564832e-01
-3.81066203e-01 8.89237285e-01 1.50602207e-01 2.14515835e-01
-4.69642848e-01 -1.93058595e-01 7.08256364e-01 5.40320985e-02
-3.32276970e-01 -2.36414820e-02 -9.49598670e-01 -8.86835515e-01
-3.38590473e-01 -6.63168669e-01 2.68078476e-01 -6.25023425e-01
-6.33841872e-01 2.17649728e-01 -4.70750213e-01 -7.12631643e-01
-4.46514904e-01 -6.94067121e-01 -8.79774749e-01 9.58651006e-01
-1.21087945e+00 -6.74711168e-01 -5.15720963e-01 2.60315984e-01
2.98660219e-01 -1.35106504e-01 1.08617699e+00 -3.20068479e-01
-6.62537873e-01 1.41907394e-01 -5.59999645e-02 3.04621141e-02
4.07368332e-01 -1.22791398e+00 -2.41907254e-01 6.73676014e-01
-7.14044333e-01 2.24448010e-01 6.14411116e-01 -8.00820887e-01
-1.07437742e+00 -1.07254410e+00 8.34940791e-01 -4.65141842e-03
6.74208045e-01 5.52540086e-02 -6.24195755e-01 6.88245773e-01
2.67552465e-01 -3.82540554e-01 1.14528489e+00 -4.63619024e-01
3.30775857e-01 3.95313531e-01 -1.38832235e+00 9.35936272e-01
5.90663970e-01 -5.16578741e-02 -5.46637595e-01 7.47531712e-01
4.92809713e-02 -4.11842555e-01 -8.25439453e-01 4.09335434e-01
2.96535552e-01 -8.51143301e-01 2.91202873e-01 -9.17132273e-02
7.11207986e-02 -3.26297492e-01 2.66667694e-01 -1.64189720e+00
-3.14263016e-01 -2.37975180e-01 3.54827285e-01 2.72138655e-01
8.07386994e-01 -9.35810566e-01 1.26005113e+00 1.30140376e+00
-4.00021404e-01 -9.00071442e-01 -1.08063400e+00 -4.56918806e-01
2.28835851e-01 -3.52081180e-01 4.94192302e-01 6.29943609e-01
1.04471374e+00 -4.40277085e-02 6.40080333e-01 2.93398295e-02
6.89525247e-01 -4.24767941e-01 2.16383085e-01 -1.09345782e+00
3.46471220e-01 -1.14826548e+00 -8.94068480e-01 -1.95734188e-01
1.58173084e-01 -1.18341494e+00 -1.79524941e-03 -2.09764194e+00
5.62695205e-01 1.13379853e-02 -2.63937302e-02 8.40140879e-01
-7.95272291e-02 6.79849982e-02 -4.06092674e-01 -8.34927484e-02
1.30579770e-01 1.69093356e-01 1.04196596e+00 -5.05081475e-01
-6.61158934e-02 -2.26666883e-01 -5.50650537e-01 9.86082017e-01
1.17951357e+00 -3.12873960e-01 -3.59361976e-01 1.37007004e-02
-1.75814420e-01 2.92608380e-01 1.33253466e-02 -1.18189704e+00
5.27749240e-01 -5.46954036e-01 6.31287456e-01 -7.92294502e-01
-1.76701266e-02 -6.18826687e-01 1.44103110e-01 1.00773561e+00
1.60848305e-01 -1.99099913e-01 -2.56911851e-02 6.96137995e-02
9.45364162e-02 -4.46990699e-01 9.30147350e-01 -1.98848471e-01
-6.13515675e-01 5.80337465e-01 -1.59126949e+00 -3.34536910e-01
1.61617851e+00 -4.72479194e-01 -4.02481884e-01 -2.46209651e-01
-1.05459821e+00 2.73648798e-01 6.35518134e-01 -3.80698554e-02
6.58360541e-01 -9.11377907e-01 -7.67229915e-01 1.00095108e-01
8.29679146e-03 2.11410314e-01 3.16936433e-01 1.44686747e+00
-9.63079870e-01 4.68852580e-01 -3.29835862e-01 -7.30507195e-01
-1.38089502e+00 2.47512653e-01 7.20599890e-01 -1.14098981e-01
-8.61573756e-01 8.79451931e-01 1.58814900e-02 -1.40615448e-01
2.59183884e-01 -5.01057804e-01 -2.41944671e-01 -3.78719941e-02
1.11785948e+00 4.44592983e-02 3.97613257e-01 -1.42429858e-01
-2.62417674e-01 -3.36987927e-04 -2.97931522e-01 -2.50380486e-01
1.05771148e+00 1.06108271e-01 -5.72977483e-01 -4.58245948e-02
7.98309624e-01 -5.17098606e-01 -8.94002855e-01 2.19044089e-01
1.73634961e-01 -2.40834400e-01 4.99892503e-01 -1.05213654e+00
-1.18391860e+00 4.95204031e-01 7.80039966e-01 -2.84896433e-01
1.11951602e+00 -1.67451397e-01 6.89161122e-01 2.52707809e-01
5.21414280e-01 -1.00067818e+00 -4.76467341e-01 5.93470931e-01
4.51671034e-01 -8.82746756e-01 -3.07757765e-01 -2.64471501e-01
-2.55698621e-01 1.12965095e+00 5.79599917e-01 2.68055171e-01
6.59235775e-01 9.03149843e-01 4.82756257e-01 -3.35993953e-02
-8.19375992e-01 3.86307091e-02 -2.61454135e-01 8.54383051e-01
6.31236970e-01 5.60250580e-01 -5.24817884e-01 6.03318870e-01
-6.13044024e-01 1.72843784e-01 7.08177745e-01 1.13055849e+00
-1.04989636e+00 -1.00401354e+00 -5.65391779e-01 1.02172267e+00
-2.28393078e-01 -1.74222723e-01 -3.67474228e-01 3.36050302e-01
-7.65592009e-02 7.81635940e-01 1.09765641e-01 1.00432575e-01
-2.89277077e-01 6.63859189e-01 5.77395141e-01 -3.89105082e-01
-2.17441738e-01 1.38472706e-01 -3.16282199e-03 -1.83865219e-01
-2.46520713e-02 -1.07682848e+00 -2.00450611e+00 -1.11107215e-01
-1.25626862e-01 -3.85599248e-02 1.05046129e+00 1.36572933e+00
4.03320119e-02 6.36021316e-01 4.71816212e-02 -7.29418278e-01
9.68108103e-02 -9.55259323e-01 -8.53316009e-01 -4.01806712e-01
1.33817568e-01 -4.10789937e-01 -2.39946187e-01 1.52984299e-02] | [14.784815788269043, -2.5181639194488525] |
2c1c4910-2d87-4448-90e0-7ff3c3e5c9aa | deep-clustering-with-incomplete-noisy | 2305.19391 | null | https://arxiv.org/abs/2305.19391v1 | https://arxiv.org/pdf/2305.19391v1.pdf | Deep Clustering with Incomplete Noisy Pairwise Annotations: A Geometric Regularization Approach | The recent integration of deep learning and pairwise similarity annotation-based constrained clustering -- i.e., $\textit{deep constrained clustering}$ (DCC) -- has proven effective for incorporating weak supervision into massive data clustering: Less than 1% of pair similarity annotations can often substantially enhance the clustering accuracy. However, beyond empirical successes, there is a lack of understanding of DCC. In addition, many DCC paradigms are sensitive to annotation noise, but performance-guaranteed noisy DCC methods have been largely elusive. This work first takes a deep look into a recently emerged logistic loss function of DCC, and characterizes its theoretical properties. Our result shows that the logistic DCC loss ensures the identifiability of data membership under reasonable conditions, which may shed light on its effectiveness in practice. Building upon this understanding, a new loss function based on geometric factor analysis is proposed to fend against noisy annotations. It is shown that even under $\textit{unknown}$ annotation confusions, the data membership can still be $\textit{provably}$ identified under our proposed learning criterion. The proposed approach is tested over multiple datasets to validate our claims. | ['Xiao Fu', 'Shahana Ibrahim', 'Tri Nguyen'] | 2023-05-30 | null | null | null | null | ['deep-clustering', 'deep-clustering'] | ['miscellaneous', 'natural-language-processing'] | [ 1.49426848e-01 1.44513041e-01 9.07143131e-02 -6.16240084e-01
-1.25528550e+00 -7.59284914e-01 1.12262361e-01 3.39643598e-01
-4.75557894e-01 6.46988034e-01 -1.78056553e-01 -1.70207441e-01
-6.01231039e-01 -2.30156034e-01 -9.03599501e-01 -1.08429027e+00
-3.28426361e-01 4.86160159e-01 -2.65116766e-02 2.20974609e-01
1.93255633e-01 1.48387268e-01 -1.45843434e+00 -2.13068984e-02
9.48044717e-01 9.66967702e-01 1.68614626e-01 3.20175678e-01
2.85120010e-01 4.62687761e-01 -5.76510727e-01 -6.52673483e-01
4.03005034e-01 -3.03832233e-01 -8.47549438e-01 -1.12309389e-01
4.59741145e-01 -7.39245617e-04 3.08644809e-02 1.24885941e+00
5.52470326e-01 2.05360815e-01 6.92999899e-01 -1.41567612e+00
-8.02764237e-01 7.24505842e-01 -4.02311742e-01 -2.42301315e-01
5.98746054e-02 3.71012203e-02 1.40941238e+00 -9.65323269e-01
4.18094456e-01 1.00749457e+00 1.24517000e+00 3.28193396e-01
-1.23592401e+00 -5.78798532e-01 1.20365076e-01 -1.38291031e-01
-1.85168052e+00 -1.38109475e-01 6.38762414e-01 -3.48673642e-01
5.10535836e-01 2.40450770e-01 1.97541386e-01 8.29872906e-01
-4.33920354e-01 7.41659939e-01 9.09691215e-01 -5.09481013e-01
3.71421605e-01 -7.45544583e-02 1.53075278e-01 6.61583126e-01
4.70670789e-01 -9.73389223e-02 -4.41495478e-01 -1.66638210e-01
3.02895963e-01 -2.30912209e-01 -2.51358598e-01 -4.75834906e-01
-1.03394926e+00 8.58900428e-01 4.24529821e-01 1.56588420e-01
3.84762794e-01 3.80766690e-01 4.94630903e-01 5.14420532e-02
5.24137914e-01 5.20383775e-01 -4.04205769e-01 -1.14681080e-01
-9.93433654e-01 2.55412489e-01 7.10056961e-01 1.26778901e+00
7.65807152e-01 -2.14371458e-01 8.55671614e-02 8.53552699e-01
2.26563856e-01 4.87700254e-01 -1.40617117e-01 -1.38346076e+00
3.98339212e-01 5.42551875e-01 1.85390338e-01 -1.36361897e+00
-5.79150379e-01 -4.80310827e-01 -1.02896965e+00 -1.01311415e-01
5.97480595e-01 -1.81420043e-01 -3.97049010e-01 2.17290521e+00
2.02403158e-01 1.68896705e-01 -4.04744744e-01 1.05431974e+00
2.39323050e-01 8.30332283e-03 -1.52635857e-01 -2.24601582e-01
9.31531131e-01 -4.82564598e-01 -7.24826455e-01 1.61263555e-01
9.35086429e-01 -5.71450830e-01 1.18963242e+00 3.63218158e-01
-9.81425345e-01 -3.54384750e-01 -1.16020370e+00 -2.14690492e-01
-3.01621288e-01 -8.88086669e-03 6.84157431e-01 7.97108054e-01
-1.15951836e+00 4.71707910e-01 -9.42396045e-01 -2.63347685e-01
6.77822351e-01 5.86968005e-01 -2.58666664e-01 -3.38525862e-01
-1.06768417e+00 4.69985276e-01 1.99384853e-01 4.28300351e-01
-6.66503787e-01 -6.14839971e-01 -7.20904589e-01 4.61111590e-02
6.58095241e-01 -3.91634434e-01 8.09013128e-01 -6.15756035e-01
-9.02619243e-01 8.36945832e-01 -8.57795402e-02 -4.07928526e-01
6.01914227e-01 -1.63422242e-01 -1.25999063e-01 2.89060593e-01
3.48444253e-01 6.36848688e-01 3.79786879e-01 -1.57599199e+00
-3.61873299e-01 -4.20583665e-01 3.08914352e-02 8.88326839e-02
-5.13301253e-01 -2.56164242e-02 -7.76114881e-01 -6.73782408e-01
2.26741239e-01 -1.03882527e+00 -3.32972616e-01 1.73222363e-01
-6.14112079e-01 -3.36645246e-01 5.91821730e-01 -1.16962686e-01
1.26125085e+00 -2.13589430e+00 5.72879519e-03 5.98245621e-01
3.50514889e-01 1.50873885e-01 -6.30065054e-02 2.61591792e-01
4.96343262e-02 5.85359693e-01 -7.05669522e-01 -7.61045456e-01
4.13122624e-01 2.22414553e-01 9.09653828e-02 9.06078577e-01
1.86505049e-01 7.71955609e-01 -8.32874835e-01 -4.85833466e-01
4.56136540e-02 4.89980876e-01 -7.92304456e-01 -1.39327064e-01
-2.09731072e-01 1.37142122e-01 -1.39041677e-01 7.70688295e-01
9.59710062e-01 -5.25221288e-01 5.21385431e-01 -2.16885060e-01
1.52829990e-01 -3.17808330e-01 -1.22534132e+00 1.89344811e+00
1.36675805e-01 5.68155825e-01 4.81669694e-01 -1.56273115e+00
6.94978952e-01 2.02269722e-02 7.18250811e-01 -2.52392113e-01
1.55891687e-01 3.61265123e-01 -1.45421311e-01 -2.97549307e-01
4.14476335e-01 -1.14917560e-02 -3.38130116e-01 3.89561385e-01
2.24807853e-04 -2.30579674e-02 8.52582827e-02 4.27859604e-01
1.18550766e+00 -5.81804924e-02 -1.55412987e-01 -6.40650570e-01
2.41435334e-01 -3.23717631e-02 8.79498065e-01 1.01295745e+00
-5.34993351e-01 9.08805490e-01 5.47095358e-01 -4.43125479e-02
-1.08566523e+00 -8.77544761e-01 -3.71313244e-01 9.99202609e-01
4.61177707e-01 -5.43470144e-01 -1.10018718e+00 -7.41471887e-01
7.65888542e-02 1.14167169e-01 -6.10983551e-01 -1.90094694e-01
-3.40099394e-01 -1.09624612e+00 9.97460544e-01 6.21871054e-01
2.74289876e-01 -4.53899473e-01 -8.51974115e-02 -1.82734177e-01
-2.56695777e-01 -1.09639442e+00 -6.35676444e-01 3.78729999e-01
-3.41005892e-01 -1.34537566e+00 -5.31825602e-01 -9.72205520e-01
7.93848872e-01 2.80027777e-01 9.65569496e-01 4.53122914e-01
-2.03483343e-01 5.46368837e-01 -5.62940955e-01 -9.17002372e-03
-1.17415853e-01 1.67504027e-01 4.29658115e-01 -1.33511826e-01
6.92503929e-01 -5.79916000e-01 -7.21535265e-01 6.38484359e-01
-1.00944638e+00 -5.55432081e-01 2.91136682e-01 8.35154653e-01
7.04194605e-01 2.93412268e-01 7.08502531e-01 -8.31910431e-01
4.24548537e-01 -3.92645985e-01 -5.70490479e-01 3.36258560e-01
-6.34202003e-01 -4.44494709e-02 6.63641632e-01 -2.24179104e-01
-6.61759079e-01 1.58084393e-01 5.92143275e-02 -6.28367662e-01
-4.51553427e-02 5.88665962e-01 -4.30292815e-01 2.20902320e-02
6.22951329e-01 -2.02215001e-01 1.35380588e-02 -4.53788370e-01
4.91362184e-01 5.84955990e-01 8.42213452e-01 -1.05156565e+00
6.38329268e-01 7.75922775e-01 6.82771504e-02 -4.96802002e-01
-1.07722926e+00 -5.21705747e-01 -8.77380610e-01 -1.05223276e-01
1.01562500e+00 -9.17441308e-01 -1.24469829e+00 3.19404781e-01
-8.61536801e-01 -3.54076982e-01 3.63343693e-02 2.69745231e-01
-5.35887003e-01 7.64677882e-01 -5.15221000e-01 -9.74292636e-01
7.59978518e-02 -1.20542955e+00 1.10723770e+00 -2.19667375e-01
-5.93928099e-02 -1.12863421e+00 -3.45595837e-01 6.44218743e-01
1.90304771e-01 3.23482990e-01 7.76532352e-01 -9.87401128e-01
-4.94190782e-01 -4.31278408e-01 -3.85323673e-01 5.62925100e-01
-4.33432013e-02 3.55347805e-02 -9.92227554e-01 -5.81682980e-01
-9.09255594e-02 -6.74826801e-01 7.83519924e-01 3.71432185e-01
1.63942528e+00 1.60920341e-02 -3.40106100e-01 6.71446443e-01
1.42331707e+00 -1.48682445e-01 4.50144291e-01 -1.51099358e-02
6.90198600e-01 5.18260598e-01 5.50048113e-01 5.18530190e-01
4.70354140e-01 5.75023651e-01 6.44910574e-01 6.29326003e-03
9.59445462e-02 -3.61628681e-02 1.82992872e-02 7.92938113e-01
-9.51189175e-03 -3.88377190e-01 -1.09122109e+00 4.13646847e-01
-2.00317407e+00 -9.08674061e-01 -3.35235149e-01 2.16801572e+00
8.31411600e-01 1.67940706e-01 -1.12114206e-01 1.94463208e-01
1.03220439e+00 -2.06792548e-01 -7.00029612e-01 1.54371291e-01
-6.62895441e-01 7.59021193e-02 4.11370218e-01 5.70517421e-01
-1.39639115e+00 7.90277243e-01 6.35118628e+00 1.08733952e+00
-6.27378821e-01 6.30455166e-02 8.04284394e-01 -1.33417323e-01
-2.67154723e-01 -4.00976799e-02 -6.47560656e-01 6.82279229e-01
6.33711219e-01 2.27583274e-01 2.88055956e-01 7.22386539e-01
3.40862542e-01 -1.31490424e-01 -1.28760505e+00 9.33664203e-01
2.31668159e-01 -1.38966429e+00 -2.04833597e-01 2.64034480e-01
8.15836847e-01 -1.16197541e-01 3.06624562e-01 2.12929741e-01
6.26779735e-01 -1.11318541e+00 7.08446205e-01 3.10949534e-01
1.05021930e+00 -9.87735569e-01 8.69407773e-01 2.50761867e-01
-1.15875447e+00 -2.56245285e-01 -5.10401905e-01 -1.18159233e-02
5.47441691e-02 8.89839292e-01 -3.98304552e-01 6.98375821e-01
9.35607314e-01 6.51821375e-01 -5.88899553e-01 7.48661041e-01
1.32885635e-01 6.16810739e-01 -5.44993401e-01 1.90584004e-01
1.77129179e-01 -2.43386716e-01 3.21909636e-01 1.23208976e+00
2.50458539e-01 1.43129602e-01 4.28222537e-01 9.30129111e-01
-4.33510512e-01 -2.21100301e-01 -3.84513080e-01 2.21773282e-01
9.57289934e-01 9.90281641e-01 -9.48216975e-01 3.35178524e-02
-1.77698985e-01 9.68634069e-01 7.13759005e-01 3.15922171e-01
-9.50634539e-01 -4.33288008e-01 6.13969564e-01 -1.32923990e-01
3.23728174e-01 -3.62218827e-01 -6.33913696e-01 -1.03154314e+00
2.39389598e-01 -6.82682931e-01 4.59113270e-01 -4.93109822e-01
-1.78553545e+00 1.61307335e-01 -1.58876091e-01 -1.02159953e+00
2.82242507e-01 -6.60206378e-01 -2.84500986e-01 4.82874215e-01
-9.55223203e-01 -1.00128400e+00 -1.59111276e-01 5.36449492e-01
8.40432271e-02 3.51818576e-02 6.53979003e-01 5.84639370e-01
-8.06019187e-01 1.13803756e+00 6.37012899e-01 3.07849437e-01
8.80988359e-01 -1.41666460e+00 -9.31596532e-02 9.30751443e-01
-1.09742209e-01 8.19008887e-01 5.45149744e-01 -5.75205207e-01
-1.31126237e+00 -1.22842526e+00 5.93049586e-01 -8.35601449e-01
5.65434635e-01 -8.68068099e-01 -9.96625662e-01 4.06088680e-01
-7.97800347e-02 3.21729034e-01 1.02080142e+00 9.35133323e-02
-5.44820845e-01 -6.12112172e-02 -1.31526005e+00 3.92068446e-01
1.42535937e+00 -6.57583117e-01 -1.06012665e-01 4.57781941e-01
1.05872405e+00 -2.57433224e-02 -1.06677198e+00 5.42099535e-01
3.24984431e-01 -1.08466768e+00 9.59434628e-01 -5.48664331e-01
2.39781871e-01 -4.01659399e-01 -7.84865677e-01 -8.21368515e-01
-1.27689138e-01 -7.54874647e-01 8.02222416e-02 1.43384910e+00
4.62096095e-01 -1.19127162e-01 8.80951583e-01 1.02950370e+00
-4.83529061e-01 -6.14399850e-01 -1.12657237e+00 -9.62096095e-01
5.46234190e-01 -8.61291349e-01 2.70622581e-01 1.36166608e+00
1.53779998e-01 -3.41570042e-02 -2.37206995e-01 4.80728477e-01
8.39524090e-01 -3.57508987e-01 5.10811269e-01 -1.30335891e+00
1.66463796e-02 -4.27366912e-01 -2.62478948e-01 -9.15356815e-01
4.58564371e-01 -1.07678401e+00 1.83839515e-01 -1.12277043e+00
2.67376989e-01 -8.92571688e-01 -2.14016467e-01 4.47948337e-01
-3.06046575e-01 5.06489873e-01 1.20695405e-01 3.05298746e-01
-1.21603620e+00 5.13928831e-01 8.00644577e-01 4.59059281e-03
2.33681872e-01 -2.10761949e-01 -7.96422422e-01 6.21135712e-01
6.40925229e-01 -3.50152791e-01 -3.02787036e-01 -4.71667737e-01
5.41615129e-01 -3.97667438e-01 3.68095875e-01 -8.58422697e-01
4.79232967e-01 1.25394359e-01 4.66533229e-02 -5.64977527e-01
2.37084568e-01 -9.09372330e-01 -1.71882346e-01 1.43491328e-01
-5.58258414e-01 -2.05588087e-01 -1.42489344e-01 1.05289793e+00
-2.31062882e-02 -1.20822497e-01 7.75069952e-01 9.74516049e-02
-2.38425776e-01 2.30089635e-01 -1.02049142e-01 4.69477654e-01
8.14536333e-01 -1.81785867e-01 -2.49677852e-01 -2.48592094e-01
-6.47422969e-01 5.25631964e-01 6.79018438e-01 -1.54076889e-01
3.17584991e-01 -1.33517957e+00 -5.89880884e-01 7.14328792e-03
1.93013564e-01 5.79727232e-01 2.12183893e-01 1.01915705e+00
-4.21242297e-01 2.53616601e-01 5.97042143e-01 -9.04889047e-01
-9.46408689e-01 7.10549951e-01 3.55548590e-01 1.07875094e-01
-3.13930631e-01 1.02326727e+00 -5.25449626e-02 -6.61405802e-01
6.28248453e-01 -2.27199107e-01 5.55509865e-01 -7.03959540e-02
1.69864938e-01 4.32606667e-01 1.27137274e-01 -5.40053785e-01
-4.59320366e-01 5.27765453e-01 7.86621571e-02 -3.53124887e-02
1.14732051e+00 -4.13311630e-01 -2.14737222e-01 1.94448501e-01
1.42259240e+00 -1.16079971e-01 -1.48947787e+00 -2.23979250e-01
3.69411528e-01 -3.01392674e-01 -3.38944137e-01 -8.50524783e-01
-9.86532569e-01 8.92777562e-01 3.37441832e-01 3.22038949e-01
9.01157379e-01 1.59324348e-01 4.59820807e-01 6.32848799e-01
3.19644094e-01 -1.39585030e+00 1.96699008e-01 4.38040197e-01
4.54940945e-01 -1.46997058e+00 -7.51850381e-02 -3.40454102e-01
-4.66653228e-01 6.50163949e-01 4.85939175e-01 2.09968779e-02
8.53191972e-01 3.81514668e-01 -1.32678345e-01 -3.02368224e-01
-4.25768018e-01 -1.80001065e-01 -1.77436639e-02 8.04467797e-01
3.53544354e-01 -1.27767816e-01 -2.21769646e-01 1.04188871e+00
-9.80495438e-02 -4.36966002e-01 3.71583819e-01 6.33639991e-01
-3.15622330e-01 -7.59924650e-01 -3.20829272e-01 2.44695500e-01
-5.33901989e-01 -2.18784809e-02 -5.70986331e-01 8.49927545e-01
4.26607072e-01 1.22630584e+00 8.94882381e-02 -3.46227318e-01
-1.24137439e-01 5.56803867e-02 1.37306154e-01 -3.12591285e-01
-3.41807157e-01 1.61375776e-01 -2.14069948e-01 -3.87978256e-01
-7.78365314e-01 -6.84427917e-01 -1.39481175e+00 -3.96189868e-01
-5.92874885e-01 2.85233021e-01 4.71554875e-01 8.62897873e-01
3.85515779e-01 1.33881094e-02 6.46106362e-01 -6.95651233e-01
-6.04848623e-01 -6.97213411e-01 -7.25932181e-01 7.17431605e-01
1.51010200e-01 -7.56796062e-01 -8.22990298e-01 8.21229145e-02] | [9.145593643188477, 3.2785425186157227] |
ec7755d9-eab2-45e0-96d7-4c999de9e448 | learning-the-night-sky-with-deep-generative | 2302.02030 | null | https://arxiv.org/abs/2302.02030v1 | https://arxiv.org/pdf/2302.02030v1.pdf | Learning the Night Sky with Deep Generative Priors | Recovering sharper images from blurred observations, referred to as deconvolution, is an ill-posed problem where classical approaches often produce unsatisfactory results. In ground-based astronomy, combining multiple exposures to achieve images with higher signal-to-noise ratios is complicated by the variation of point-spread functions across exposures due to atmospheric effects. We develop an unsupervised multi-frame method for denoising, deblurring, and coadding images inspired by deep generative priors. We use a carefully chosen convolutional neural network architecture that combines information from multiple observations, regularizes the joint likelihood over these observations, and allows us to impose desired constraints, such as non-negativity of pixel values in the sharp, restored image. With an eye towards the Rubin Observatory, we analyze 4K by 4K Hyper Suprime-Cam exposures and obtain preliminary results which yield promising restored images and extracted source lists. | ['Yashil Sukurdeep', 'Tamas Budavari', 'Daniel Hall', 'Fausto Navarro'] | 2023-02-03 | null | null | null | null | ['deblurring', 'astronomy'] | ['computer-vision', 'miscellaneous'] | [ 3.81127536e-01 -4.74082261e-01 6.83386803e-01 -3.38524282e-01
-9.49885070e-01 -6.77678823e-01 5.89272082e-01 -4.08910930e-01
-3.62520486e-01 8.13605607e-01 4.64696854e-01 4.66758311e-02
-6.67035818e-01 -5.65423608e-01 -8.13040853e-01 -1.02180016e+00
1.92644417e-01 1.38081685e-01 5.52840196e-02 -9.01933014e-02
3.41930121e-01 7.75326550e-01 -1.21783662e+00 -2.08801832e-02
7.09419191e-01 7.18905151e-01 2.31941223e-01 7.90300012e-01
4.54338282e-01 7.12789059e-01 -5.54187238e-01 -3.06381851e-01
3.95732850e-01 -6.49149716e-01 -6.29185319e-01 5.80223680e-01
6.78306639e-01 -4.60561424e-01 -7.19138324e-01 1.21715903e+00
2.71450013e-01 4.35604066e-01 7.15572536e-01 -5.08973181e-01
-1.02562439e+00 8.90703201e-02 -6.72439396e-01 3.80891085e-01
-5.92360981e-02 2.73475736e-01 5.05331755e-01 -8.19378495e-01
2.50632077e-01 1.03197730e+00 8.03919137e-01 -8.49290192e-02
-1.47499216e+00 2.54092403e-02 -4.16604906e-01 1.38452023e-01
-1.22527623e+00 -6.53852701e-01 7.58076131e-01 -4.08932269e-01
5.49445927e-01 3.96246344e-01 1.79172814e-01 8.45159590e-01
2.10806146e-01 2.84412922e-03 1.19172394e+00 -4.94596094e-01
-3.34665515e-02 -2.40546763e-01 -4.39673066e-02 2.49064282e-01
3.79415542e-01 2.11482897e-01 -5.46993911e-01 -2.36765906e-01
1.08947599e+00 7.72052854e-02 -1.00910521e+00 7.67783001e-02
-1.15764487e+00 7.17186213e-01 5.01190126e-01 2.06628457e-01
-7.53299117e-01 2.14731708e-01 -3.62521499e-01 1.53966665e-01
6.36717796e-01 6.95950210e-01 -2.94045627e-01 3.05646569e-01
-1.04419529e+00 3.48744154e-01 4.31782544e-01 3.72635573e-01
9.69901860e-01 2.36056328e-01 -1.41138345e-01 7.96022177e-01
2.50552922e-01 7.79404700e-01 1.49184659e-01 -1.54748976e+00
-2.48007685e-01 -2.42007613e-01 5.34811139e-01 -7.07044780e-01
-1.50198877e-01 -5.68467736e-01 -9.98644531e-01 4.04558480e-01
5.28581738e-01 -3.78483742e-01 -1.04366505e+00 1.55409920e+00
1.40640419e-02 2.07719043e-01 3.02875698e-01 1.18319845e+00
5.66870451e-01 5.90841949e-01 -4.87056732e-01 -3.31984788e-01
1.26513743e+00 -4.95097727e-01 -7.37143636e-01 -5.77742279e-01
-3.02132726e-01 -1.07230210e+00 5.18568635e-01 4.18305576e-01
-1.24741912e+00 -4.66682851e-01 -7.89844155e-01 -1.85326129e-01
1.89023882e-01 9.29945558e-02 4.00294423e-01 3.03334683e-01
-1.20970154e+00 8.56874943e-01 -6.52328432e-01 -5.71736656e-02
3.16029102e-01 -3.33789401e-02 -2.82522678e-01 -1.36890918e-01
-7.60372579e-01 9.84811068e-01 7.16588497e-02 2.57454842e-01
-9.52567399e-01 -8.59421313e-01 -7.07181871e-01 2.96099603e-01
1.86696112e-01 -7.41555810e-01 1.16079235e+00 -7.86726892e-01
-1.32048881e+00 6.55282140e-01 -3.25069696e-01 -5.07431686e-01
2.12434426e-01 -4.10202146e-01 -3.26720595e-01 4.80546683e-01
-1.19944997e-01 2.58117884e-01 1.22599387e+00 -1.27437103e+00
-7.47346133e-02 -8.24213177e-02 -1.66793436e-01 1.01517960e-01
2.79652447e-01 1.64968193e-01 -2.42870599e-01 -7.20951498e-01
3.36922586e-01 -6.98444963e-01 -3.49444240e-01 -1.62338287e-01
-1.72688678e-01 5.19401968e-01 5.04329681e-01 -9.30168152e-01
3.93130779e-01 -2.42340899e+00 2.89154172e-01 -2.23782599e-01
3.07413012e-01 -8.15174878e-02 -7.23072961e-02 2.83270687e-01
-2.47604534e-01 -2.83371985e-01 -6.85571730e-01 -2.50926137e-01
-3.53324592e-01 4.09130529e-02 -4.09201086e-01 8.90739501e-01
3.28963369e-01 5.27905762e-01 -5.94599664e-01 1.40058562e-01
4.43260610e-01 6.42305851e-01 -3.02213162e-01 3.59558135e-01
-1.65061817e-01 8.31520081e-01 2.18410417e-02 2.96351075e-01
1.17512965e+00 -3.42686892e-01 -2.67890811e-01 -3.75006646e-01
-4.64428961e-01 -1.59582451e-01 -8.73349845e-01 1.49870753e+00
-3.00620347e-01 1.02808547e+00 4.86868620e-01 -7.23058999e-01
8.46490324e-01 2.93495119e-01 3.11308414e-01 -3.51710469e-01
5.12576103e-02 8.78098458e-02 -1.89011008e-01 -7.39268482e-01
7.15909898e-01 -6.13269806e-01 4.04458314e-01 1.98520795e-01
1.88939899e-01 -5.74314058e-01 -1.94908231e-01 1.81710929e-01
9.41683650e-01 2.39544865e-02 -4.56880685e-03 -3.33739907e-01
1.81651682e-01 -1.32550150e-01 3.01675797e-01 1.07282984e+00
2.42001295e-01 1.50194514e+00 2.64514297e-01 -2.95269310e-01
-1.29246449e+00 -1.03977990e+00 -2.64786482e-01 4.40515995e-01
1.40795922e-02 2.42570579e-01 -5.37071586e-01 8.04932881e-03
-2.09284320e-01 7.67657161e-01 -3.96231502e-01 -8.85175839e-02
-3.50873023e-01 -1.26519763e+00 2.46237770e-01 2.67932545e-02
6.68197870e-01 -7.18237817e-01 -4.03401941e-01 9.03895497e-02
-4.12966996e-01 -1.17359531e+00 -2.96815276e-01 2.42788121e-01
-6.27692401e-01 -1.00293648e+00 -9.82715368e-01 -3.89433920e-01
6.08736277e-01 8.18371415e-01 1.13511217e+00 -2.31334537e-01
-2.24069461e-01 3.24754328e-01 -1.67340279e-01 -1.64797127e-01
-3.80119830e-01 -7.37097144e-01 -7.74063841e-02 3.60123366e-01
-2.25526601e-01 -8.15169454e-01 -5.60611784e-01 5.58689162e-02
-1.35829318e+00 -2.39025980e-01 6.12476707e-01 7.79278636e-01
3.52567047e-01 2.73251951e-01 1.95001557e-01 -3.70918959e-01
4.29005146e-01 -3.06480944e-01 -9.28139925e-01 1.95389509e-01
-2.04279214e-01 2.36693129e-01 2.96632588e-01 -3.34540546e-01
-1.64964616e+00 -1.05698228e-01 5.29993623e-02 -6.14079475e-01
-3.31530660e-01 3.96631539e-01 1.69039890e-01 -4.52559143e-01
1.06037128e+00 3.26096714e-01 1.00541040e-01 -6.33348405e-01
4.53052223e-01 4.85267043e-01 1.16714585e+00 -3.76481801e-01
9.07112479e-01 7.98738599e-01 2.10232794e-01 -1.22528183e+00
-1.23376799e+00 -3.93529713e-01 -4.12800163e-01 -2.20700964e-01
9.22281742e-01 -1.17818642e+00 -4.83038366e-01 7.35087812e-01
-1.24438989e+00 -2.79355168e-01 -2.44807899e-01 7.79756308e-01
-2.35438839e-01 6.39532864e-01 -5.62447488e-01 -7.42622197e-01
-1.30439341e-01 -1.04812169e+00 9.05900002e-01 7.19253004e-01
1.31044894e-01 -9.91384566e-01 -1.29848961e-02 2.90410191e-01
7.20344782e-01 1.20763697e-01 5.50862968e-01 3.23435850e-02
-8.69346440e-01 -1.19492047e-01 -4.58774656e-01 6.52925253e-01
2.34415218e-01 -1.04809828e-01 -1.21209466e+00 -1.31326661e-01
8.33598852e-01 -7.18550906e-02 1.21623445e+00 1.04279888e+00
1.10394287e+00 -1.55330166e-01 2.48089686e-01 9.37307239e-01
1.57473564e+00 -3.09654057e-01 9.58130062e-01 2.46780902e-01
5.06285191e-01 4.99918401e-01 -8.16212222e-03 4.96212184e-01
-1.70201689e-01 2.60020912e-01 6.09539092e-01 -2.06239849e-01
-3.57117385e-01 5.58951855e-01 5.80615997e-02 1.51439026e-01
-2.76012152e-01 -4.64578658e-01 -3.88223529e-01 7.20323265e-01
-1.48013783e+00 -1.11638975e+00 -7.06648767e-01 2.27656841e+00
6.11215353e-01 -1.46585554e-01 -5.07262528e-01 -4.24893528e-01
6.66580141e-01 4.43054408e-01 -2.65423357e-01 2.05952436e-01
-5.85567653e-01 5.23808062e-01 8.05413067e-01 8.87401998e-01
-8.90012801e-01 4.54341114e-01 6.70511723e+00 5.54295242e-01
-9.48178113e-01 1.11465834e-01 5.65727890e-01 -5.01863547e-02
-5.01357257e-01 1.82734147e-01 -3.46146107e-01 3.08997571e-01
7.78848171e-01 5.43212611e-03 8.27441514e-01 1.85093313e-01
3.92763615e-01 -5.56713939e-01 -6.54467940e-01 1.08046615e+00
1.54771745e-01 -1.48386621e+00 -3.50874782e-01 -2.85658464e-02
9.71885145e-01 2.83191532e-01 5.56028262e-02 -3.32819700e-01
4.06391919e-01 -8.33942831e-01 6.49799228e-01 1.29702175e+00
4.67975885e-01 -5.42271435e-01 6.33501887e-01 1.65015876e-01
-3.39762062e-01 1.92935452e-01 -4.82316941e-01 -8.99104103e-02
4.91634339e-01 1.28400218e+00 -3.12799424e-01 8.49615812e-01
8.53543937e-01 4.97395158e-01 -4.85647321e-01 1.31014335e+00
-3.28528136e-01 5.49302518e-01 -2.62064070e-01 6.72909558e-01
1.00706533e-01 -8.19172978e-01 8.17822456e-01 8.18818510e-01
7.29396224e-01 4.10272479e-01 -3.30754936e-01 1.15723133e+00
-2.03480065e-01 -7.96270013e-01 -3.99882048e-01 2.35196054e-01
3.47982440e-03 1.42418218e+00 -5.90544999e-01 -3.27895999e-01
-4.04142261e-01 1.04423475e+00 -1.94054842e-01 8.71938407e-01
-6.83780372e-01 -2.28364050e-01 7.23412991e-01 7.30381310e-02
4.55484778e-01 -4.76623565e-01 -3.33159417e-01 -1.23106456e+00
-1.13252282e-01 -7.29557097e-01 2.95310263e-02 -1.53496122e+00
-1.32077658e+00 4.91520733e-01 1.35177106e-01 -8.51107955e-01
1.44584209e-01 -4.33015198e-01 -9.53623772e-01 1.44343603e+00
-1.37074995e+00 -8.03890407e-01 -6.01012230e-01 2.39874989e-01
4.59371895e-01 1.49897277e-01 3.64863157e-01 5.91001622e-02
-3.51462096e-01 -3.86553347e-01 5.63978851e-01 -2.93351054e-01
9.49135065e-01 -1.17320216e+00 2.53576696e-01 1.39267516e+00
-7.86229372e-02 4.05866295e-01 1.30771685e+00 -5.56898952e-01
-1.11871338e+00 -9.36428249e-01 4.80677724e-01 -2.95569539e-01
6.40574217e-01 2.99670905e-01 -1.32880569e+00 7.80427694e-01
6.79164827e-01 1.19821079e-01 1.04285985e-01 -2.72431731e-01
-1.57939941e-02 -9.14085284e-02 -1.06189144e+00 3.26864064e-01
3.49256337e-01 -4.95970666e-01 -6.46376371e-01 5.86782515e-01
5.72879434e-01 -4.52358246e-01 -5.16298831e-01 3.46210033e-01
4.72172536e-02 -1.21730947e+00 1.22137809e+00 -2.41780624e-01
5.02131402e-01 -5.08932292e-01 -2.13736996e-01 -1.57879245e+00
-7.28163242e-01 -7.64682114e-01 2.32618332e-01 1.11802971e+00
1.96779966e-01 -5.97372651e-01 3.42144012e-01 5.03295660e-01
-4.18599963e-01 -6.06294721e-03 -6.03205502e-01 -6.77524090e-01
-1.85452223e-01 -5.80053069e-02 2.20913783e-01 8.71393859e-01
-8.31826866e-01 2.44174868e-01 -7.69291878e-01 8.82867873e-01
1.13733351e+00 1.80551335e-01 4.75359499e-01 -1.12388062e+00
-4.78424370e-01 -2.11790547e-01 2.78050840e-01 -1.06153631e+00
-1.44792885e-01 -3.58655900e-01 4.06747848e-01 -1.34288216e+00
1.82743281e-01 2.25502104e-01 2.05226287e-01 9.76870880e-02
-1.85012177e-01 4.46692407e-01 -1.76111981e-01 2.32814878e-01
-1.49057716e-01 5.24487078e-01 1.14568329e+00 1.06507026e-01
5.78252524e-02 -8.02310854e-02 -6.86371386e-01 8.22249353e-01
6.42072916e-01 -5.19817829e-01 5.30626662e-02 -8.98059249e-01
2.21388325e-01 2.99663842e-01 8.88482213e-01 -1.05108321e+00
1.51108950e-01 -3.04470509e-01 7.04896986e-01 -4.86618847e-01
5.95202506e-01 -3.42400789e-01 6.19252086e-01 -6.76849559e-02
-1.77823469e-01 -3.98302764e-01 3.30619477e-02 5.21897972e-01
-2.96491265e-01 -7.69805908e-01 1.27416766e+00 -4.61220801e-01
-2.30135933e-01 -2.99469624e-02 -4.99376267e-01 -1.58361658e-01
4.59006160e-01 3.53191048e-01 -5.05169690e-01 -6.41622126e-01
-5.39352059e-01 -1.86511025e-01 5.70966482e-01 -2.03161687e-01
4.49914545e-01 -8.80343437e-01 -1.02528429e+00 1.17198350e-02
-2.83452153e-01 1.55825913e-01 7.76669502e-01 1.02953768e+00
-8.90793622e-01 -3.92274447e-02 7.85755832e-03 -4.06466156e-01
-1.07241666e+00 4.09349442e-01 7.01538026e-01 2.07021847e-01
-6.33968115e-01 1.11864126e+00 3.10617417e-01 8.36586673e-03
-2.82402009e-01 -1.21907108e-01 1.18059285e-01 -1.69961125e-01
6.08996153e-01 4.36560512e-01 1.43149987e-01 -4.86285388e-01
4.81579453e-02 4.20297503e-01 1.06081530e-01 -2.47306362e-01
1.57944965e+00 -3.94450516e-01 -4.85914201e-01 9.13371146e-03
9.36810732e-01 1.89237088e-01 -1.76352584e+00 -4.11622703e-01
-4.32056010e-01 -6.53416753e-01 6.00215495e-01 -5.65957546e-01
-9.90389407e-01 6.79367006e-01 4.77872640e-01 4.43113506e-01
1.23011482e+00 1.34861171e-01 5.04773259e-01 2.76245564e-01
-3.64593506e-01 -5.76361358e-01 1.28703654e-01 6.05364203e-01
1.11909914e+00 -1.26770782e+00 3.54365587e-01 -9.98515040e-02
-3.79453182e-01 1.34262633e+00 5.11506498e-02 -2.58568853e-01
3.92322958e-01 -5.46751656e-02 4.71863486e-02 -3.99340183e-01
-2.93087184e-01 -2.50446200e-01 3.24893415e-01 3.45113784e-01
-3.99964228e-02 -2.77167201e-01 1.03636272e-01 1.85111582e-01
1.92055013e-02 -2.27599993e-01 1.12157917e+00 4.94121373e-01
-6.95343256e-01 -7.09802568e-01 -1.28421009e+00 1.50331363e-01
-6.14716828e-01 -3.16913754e-01 -5.87291196e-02 3.27017695e-01
-1.66736729e-02 1.02992558e+00 1.47022963e-01 3.10753703e-01
3.50502133e-02 -5.27670644e-02 6.51749015e-01 -2.93495893e-01
-1.91338032e-01 6.49230719e-01 -4.76913042e-02 3.36926710e-03
-6.80571496e-01 -7.80197501e-01 -7.51200855e-01 -2.71885842e-01
-2.61001170e-01 1.30875334e-01 5.47066033e-01 1.08384490e+00
1.67007446e-01 6.84169769e-01 6.63527787e-01 -1.36554205e+00
-3.80639821e-01 -1.33568442e+00 -8.01846623e-01 2.31956437e-01
1.03236139e+00 -2.57297933e-01 -9.51808155e-01 4.44778651e-01] | [11.499863624572754, -2.6683311462402344] |
9886a50f-405e-40fc-bba5-5bdca2e2e2d0 | m2rnet-multi-modal-and-multi-scale-refined | 2109.07922 | null | https://arxiv.org/abs/2109.07922v1 | https://arxiv.org/pdf/2109.07922v1.pdf | M2RNet: Multi-modal and Multi-scale Refined Network for RGB-D Salient Object Detection | Salient object detection is a fundamental topic in computer vision. Previous methods based on RGB-D often suffer from the incompatibility of multi-modal feature fusion and the insufficiency of multi-scale feature aggregation. To tackle these two dilemmas, we propose a novel multi-modal and multi-scale refined network (M2RNet). Three essential components are presented in this network. The nested dual attention module (NDAM) explicitly exploits the combined features of RGB and depth flows. The adjacent interactive aggregation module (AIAM) gradually integrates the neighbor features of high, middle and low levels. The joint hybrid optimization loss (JHOL) makes the predictions have a prominent outline. Extensive experiments demonstrate that our method outperforms other state-of-the-art approaches. | ['Hongpeng Wang', 'Xiuli Shao', 'Ruixun Zhang', 'Jinchao Zhu', 'Xian Fang'] | 2021-09-16 | null | null | null | null | ['rgb-d-salient-object-detection'] | ['computer-vision'] | [-1.26656681e-01 8.13027695e-02 -9.10604279e-03 -2.77511656e-01
-7.48921156e-01 4.96179648e-02 4.19820160e-01 -4.17379141e-02
-4.44089830e-01 5.77949584e-01 2.44993538e-01 1.85795635e-01
-1.99615926e-01 -7.05464125e-01 -4.54339564e-01 -7.97130644e-01
1.54581144e-01 -1.61600009e-01 9.46603358e-01 -2.60583609e-01
2.55619049e-01 7.46515870e-01 -1.74933481e+00 3.14816713e-01
9.75776374e-01 1.47156024e+00 2.94313729e-01 3.05109471e-01
-1.97702691e-01 7.51427412e-01 -7.12897703e-02 -1.92756772e-01
4.80634809e-01 -3.00270349e-01 -5.13424337e-01 7.41387680e-02
3.62059116e-01 -5.49547195e-01 -3.34063530e-01 1.15785873e+00
6.87663555e-01 1.35894120e-01 3.93393666e-01 -1.58521140e+00
-5.71695745e-01 3.53074968e-01 -1.14107466e+00 5.40267766e-01
6.66435808e-02 1.50684685e-01 1.09592819e+00 -1.05533576e+00
3.57743412e-01 1.41732383e+00 5.72971284e-01 8.90697166e-02
-8.66285264e-01 -4.71590161e-01 5.42741954e-01 3.65891486e-01
-1.34316921e+00 3.91569324e-02 1.25260305e+00 -1.90663517e-01
6.94848776e-01 -1.55561483e-02 8.84021878e-01 5.06412089e-01
1.14732876e-01 1.08628321e+00 1.22648454e+00 -2.53174037e-01
7.20077194e-03 1.11907497e-01 1.14254490e-01 1.03228974e+00
1.65573478e-01 1.16498664e-01 -6.60027027e-01 1.40762981e-02
1.15745544e+00 3.34229946e-01 -2.48997673e-01 -5.81240654e-01
-1.07779193e+00 9.43939030e-01 9.87334430e-01 2.82601237e-01
-6.00351274e-01 4.14114026e-03 7.69987702e-02 -9.44894105e-02
4.14675951e-01 2.95343157e-02 -3.56400639e-01 3.49350214e-01
-8.28430295e-01 5.16612194e-02 1.48305297e-01 6.02608323e-01
1.06327367e+00 -1.22706823e-01 -1.99605986e-01 6.37782395e-01
5.62296808e-01 2.13495851e-01 5.10552168e-01 -9.21091437e-01
2.27980033e-01 1.09121823e+00 3.53237987e-02 -1.09735620e+00
-6.42639160e-01 -3.85054916e-01 -8.43129337e-01 5.34711540e-01
2.26843938e-01 6.71749283e-03 -9.92365539e-01 1.46067023e+00
6.95117414e-01 2.23566875e-01 -1.05952971e-01 1.25192046e+00
1.18597710e+00 4.59084243e-01 2.39869952e-01 -6.75645918e-02
1.11812246e+00 -1.29557204e+00 -5.80211520e-01 -1.60516903e-01
-3.35117057e-02 -5.21996200e-01 8.79095078e-01 7.36533524e-03
-1.28601611e+00 -7.36261070e-01 -1.00361919e+00 -4.70639616e-01
-5.66211820e-01 -1.63070001e-02 8.52552235e-01 2.64649212e-01
-1.11449301e+00 4.77570266e-01 -6.67921424e-01 -1.41516909e-01
4.69564468e-01 3.94636542e-01 -3.11916262e-01 8.26458335e-02
-1.12017751e+00 7.81300485e-01 2.78136402e-01 4.64884609e-01
-5.54071546e-01 -5.60000598e-01 -8.90691757e-01 -1.10434815e-01
3.22349608e-01 -7.42785752e-01 9.34197366e-01 -9.68114078e-01
-1.35017610e+00 7.20302701e-01 -6.43210784e-02 -1.71384290e-01
7.02252924e-01 -3.89772207e-01 -8.19587335e-02 4.66594160e-01
1.86649665e-01 9.02129531e-01 7.34825015e-01 -1.35507941e+00
-1.19781601e+00 -5.20081043e-01 1.39304027e-01 5.05458474e-01
-3.56410712e-01 -1.60283267e-01 -6.19907200e-01 -5.81229270e-01
5.11283755e-01 -3.79745632e-01 -4.24683273e-01 2.62407482e-01
-5.12666762e-01 -4.11804408e-01 7.26883352e-01 -4.03265655e-01
1.05633223e+00 -2.05687523e+00 4.38631982e-01 1.39899686e-01
3.82589132e-01 1.11516133e-01 5.84517568e-02 -1.09203599e-01
1.02374099e-01 -6.72723055e-02 -5.68279400e-02 -4.74558234e-01
-3.86803225e-02 -4.81934957e-02 -8.21430013e-02 4.55301911e-01
3.22245091e-01 1.03878474e+00 -9.54320133e-01 -1.04189384e+00
5.24733067e-01 6.34044528e-01 -3.39918196e-01 2.06372216e-01
9.95668843e-02 2.35545024e-01 -6.83950484e-01 1.04323375e+00
6.99902833e-01 -3.43854547e-01 -3.32334161e-01 -7.01893508e-01
-4.97949481e-01 -1.69155702e-01 -1.32413530e+00 1.75166059e+00
2.04684818e-03 1.86680466e-01 4.81760055e-02 -6.49035275e-01
7.44085908e-01 -5.42078838e-02 6.16369367e-01 -8.59126627e-01
2.73707390e-01 2.56283790e-01 -1.70202225e-01 -4.60263848e-01
4.95601416e-01 -6.05400605e-03 7.12361783e-02 1.26135513e-01
5.53929284e-02 -1.91987585e-02 3.05499695e-02 1.43271759e-01
7.61675596e-01 2.46348292e-01 3.71222466e-01 -8.66839662e-02
6.70403719e-01 -4.14860159e-01 8.88721704e-01 7.35217988e-01
-7.64397621e-01 7.61314929e-01 5.65874875e-01 -6.01340175e-01
-8.17102432e-01 -1.09726977e+00 -5.95033616e-02 1.03645408e+00
6.53836310e-01 -2.11653076e-02 -3.39069873e-01 -9.16412652e-01
2.35968456e-01 1.39704481e-01 -1.03288412e+00 -9.79012847e-02
-4.58391845e-01 -6.22594535e-01 2.09280048e-02 8.02750885e-01
8.19515109e-01 -1.24582779e+00 -9.94728804e-01 2.27748975e-01
-9.25449207e-02 -1.03877640e+00 -2.47454658e-01 1.97674051e-01
-8.71061146e-01 -9.46354866e-01 -8.37720692e-01 -6.74087822e-01
5.60967386e-01 4.28427547e-01 9.71432388e-01 1.37354836e-01
-4.34747577e-01 2.55259335e-01 -4.22405303e-01 -2.85733551e-01
2.17997119e-01 3.53996046e-02 -1.18106283e-01 2.23421022e-01
2.75212884e-01 -7.35438347e-01 -9.11491513e-01 1.79716110e-01
-8.80950809e-01 2.88118392e-01 1.07264912e+00 6.83840454e-01
8.80554020e-01 -1.09175287e-01 5.66379189e-01 -4.45246458e-01
1.06240161e-01 -2.75759965e-01 -5.50190747e-01 4.53139663e-01
-3.93151283e-01 -5.18928319e-02 1.79012179e-01 -2.38089964e-01
-1.07869136e+00 2.82747716e-01 -1.98958784e-01 -5.97083390e-01
-1.92515954e-01 1.16586335e-01 -3.44349056e-01 -2.78018355e-01
-5.19130193e-03 1.40586212e-01 -2.43976593e-01 -3.39556903e-01
4.21154767e-01 3.48434240e-01 5.27748346e-01 -2.19604343e-01
7.57425189e-01 6.75875247e-01 -1.44342538e-02 -6.49178624e-01
-1.09603667e+00 -4.46840793e-01 -8.95136178e-01 -4.55028445e-01
1.03035498e+00 -8.67300034e-01 -6.77394927e-01 7.20813811e-01
-1.05880177e+00 -1.10173129e-01 -4.99864370e-01 3.53244960e-01
-4.32891339e-01 2.55837142e-01 -5.90659380e-01 -9.44486916e-01
-3.33315849e-01 -1.15398848e+00 1.26860857e+00 8.00562203e-01
4.33230758e-01 -7.53143132e-01 -1.80778980e-01 2.42989227e-01
3.79373699e-01 4.80675042e-01 6.06659055e-01 -3.06126177e-01
-8.83593678e-01 9.46803987e-02 -8.42209935e-01 1.27787724e-01
1.05798982e-01 -1.68578066e-02 -1.03494072e+00 2.11611595e-02
-2.04109848e-01 -5.14734209e-01 1.16399837e+00 5.88831007e-01
1.02416158e+00 1.21058933e-01 -2.66210288e-01 6.92299604e-01
1.52246153e+00 -2.38347147e-02 4.98179585e-01 5.23551047e-01
8.37779164e-01 4.34549302e-01 7.49569058e-01 4.51884568e-01
6.19414985e-01 4.36574757e-01 8.22734296e-01 -5.19812047e-01
-6.08186349e-02 -1.04793971e-02 6.61883205e-02 4.97745574e-01
-1.51409253e-01 1.51047602e-01 -7.03012347e-01 7.06386924e-01
-2.02200484e+00 -7.53980935e-01 6.44887388e-02 1.81610978e+00
6.14589155e-01 3.85387063e-01 2.81324834e-01 1.16945477e-02
7.23962367e-01 1.98035970e-01 -8.03830624e-01 4.96975556e-02
-4.58479941e-01 -2.21112341e-01 3.69674981e-01 3.20722550e-01
-1.39901590e+00 7.33362377e-01 6.05389881e+00 7.62815773e-01
-8.62728059e-01 1.52790740e-01 8.78459573e-01 -8.18068609e-02
-2.35229194e-01 -2.70389259e-01 -8.85897458e-01 4.63376343e-01
1.90937161e-01 2.64039874e-01 -8.78272653e-02 8.20318460e-01
-7.68201128e-02 -4.94232029e-01 -7.64606535e-01 8.91138077e-01
1.31435201e-01 -1.08701193e+00 8.24114159e-02 -1.11057922e-01
7.06038475e-01 2.55861580e-01 3.56064066e-02 1.39766291e-01
2.38880411e-01 -4.72374171e-01 8.51630151e-01 7.93567538e-01
4.13646936e-01 -8.58381867e-01 8.00655425e-01 5.95881417e-02
-1.64058769e+00 -4.24124628e-01 -2.87113875e-01 2.74252951e-01
3.00076783e-01 4.94204134e-01 -5.87954447e-02 7.98402905e-01
1.09306264e+00 8.09195757e-01 -7.49899149e-01 1.27210736e+00
-1.50984183e-01 4.97678481e-03 -4.96157199e-01 7.36080259e-02
5.51831007e-01 -1.28136605e-01 5.75909019e-01 9.65575576e-01
1.10416040e-01 2.47135386e-01 3.03343147e-01 8.92557442e-01
1.16719902e-01 -8.18358213e-02 -2.79270977e-01 3.85193974e-01
2.17056051e-01 1.60117507e+00 -1.06718946e+00 -3.07183564e-01
-6.75159633e-01 1.05695426e+00 5.64414918e-01 3.28551263e-01
-7.83821583e-01 -2.22536951e-01 5.30914783e-01 -1.24620304e-01
5.98757565e-01 1.52488872e-02 -2.18353420e-01 -9.57924187e-01
4.10046987e-02 -3.07820201e-01 5.39234757e-01 -8.92608345e-01
-1.39122272e+00 7.39192307e-01 -1.75018892e-01 -1.27903306e+00
3.03107530e-01 -5.03576875e-01 -5.74029326e-01 8.01338196e-01
-2.23067832e+00 -1.52992845e+00 -5.84146202e-01 7.12744892e-01
5.23185551e-01 1.27539143e-01 3.50328654e-01 3.38524014e-01
-7.04753518e-01 4.37738240e-01 -2.78093755e-01 2.02615336e-01
3.46790552e-01 -1.27786112e+00 1.08710323e-02 8.89563799e-01
-1.70809254e-01 2.05080450e-01 4.10304457e-01 -5.52090108e-01
-9.61032033e-01 -1.06202328e+00 5.55577934e-01 -1.55717343e-01
5.69929421e-01 2.56521460e-02 -8.68767500e-01 3.88941765e-01
1.08818822e-01 4.02579665e-01 3.61741364e-01 -3.33738387e-01
-2.58186579e-01 -2.48533323e-01 -1.02570200e+00 3.63851339e-01
9.11552548e-01 -3.55025917e-01 -6.43982530e-01 6.89572468e-02
1.03259993e+00 -3.23529422e-01 -8.63208354e-01 7.44171441e-01
5.61370432e-01 -1.57921708e+00 1.07569849e+00 -3.20354700e-01
5.20567477e-01 -4.82162088e-01 -2.31012717e-01 -8.64990234e-01
-4.54899639e-01 -2.62387127e-01 -3.53125930e-01 1.17998815e+00
8.25304464e-02 -4.70322549e-01 6.90584481e-01 5.03280222e-01
-1.85666218e-01 -1.28458858e+00 -9.28798556e-01 -3.06527793e-01
-2.26528272e-01 -1.70692444e-01 4.36101288e-01 6.60764813e-01
-2.86761284e-01 2.37638623e-01 -2.64966488e-01 2.86075681e-01
8.37734103e-01 5.35973370e-01 3.77775520e-01 -1.48043859e+00
-3.90448645e-02 -8.05120289e-01 -5.29725671e-01 -9.48852420e-01
-1.70049191e-01 -3.88002723e-01 -3.51873823e-02 -1.72837651e+00
4.71101701e-01 -4.25061822e-01 -8.15752685e-01 5.52414477e-01
-4.86477941e-01 5.87426186e-01 2.14465946e-01 1.01136006e-01
-1.04486871e+00 8.81121337e-01 1.39590895e+00 1.31593704e-01
-3.00811619e-01 -3.08564782e-01 -7.10634410e-01 1.13478351e+00
5.50693572e-01 -1.41181484e-01 -5.57238609e-02 -2.18327165e-01
5.71315642e-03 -1.02976970e-01 5.68506598e-01 -1.13221240e+00
5.33515573e-01 -1.53543577e-02 7.28187740e-01 -1.17091084e+00
4.34291005e-01 -9.29824054e-01 -4.23915923e-01 2.35595003e-01
-1.50721654e-01 1.03503667e-01 1.02715902e-01 6.27914548e-01
-3.65279615e-01 2.59992480e-01 9.77460563e-01 -3.06266844e-01
-1.09301937e+00 6.32981002e-01 6.36437759e-02 -9.51917544e-02
1.06666875e+00 -3.91475827e-01 -2.06787616e-01 -6.30667657e-02
-6.44764900e-01 4.61532235e-01 3.51687729e-01 5.81582427e-01
9.22943830e-01 -1.49592137e+00 -5.02255201e-01 3.69717240e-01
1.20379720e-02 2.96371698e-01 5.24247468e-01 1.15480363e+00
-1.50148034e-01 1.52132586e-01 -4.55496699e-01 -5.99356771e-01
-9.53813016e-01 2.85898685e-01 5.32073319e-01 -3.41055542e-01
-6.71396255e-01 1.11485672e+00 3.47859383e-01 -1.39903843e-01
3.58128130e-01 -1.70048267e-01 -5.03382623e-01 3.20387810e-01
4.74168509e-01 4.06554729e-01 -2.12745786e-01 -9.74384665e-01
-4.15624827e-01 8.89087439e-01 -2.00408220e-01 2.47278996e-02
1.48252952e+00 -4.78590041e-01 -3.25612754e-01 5.65478086e-01
1.10316789e+00 -3.22846591e-01 -1.70433187e+00 -4.96343821e-01
-3.07010144e-01 -5.26264608e-01 3.90619278e-01 -6.03743792e-01
-1.32198703e+00 7.36121356e-01 8.27743411e-01 2.78421521e-01
1.41940975e+00 3.79540808e-02 7.81745791e-01 9.49045792e-02
2.71699101e-01 -1.02906680e+00 3.41673404e-01 2.34634101e-01
7.83109665e-01 -1.49829185e+00 2.11165830e-01 -3.84948343e-01
-6.45311534e-01 8.87056053e-01 1.08201563e+00 -1.43048316e-01
8.47084761e-01 2.11291566e-01 1.22275315e-01 -2.74584085e-01
-4.17816907e-01 -5.88808954e-01 4.87089515e-01 2.96297699e-01
3.34944278e-02 -4.80241805e-01 -3.58135067e-02 6.76277339e-01
2.92024314e-01 -1.64343357e-01 1.67724401e-01 1.06024277e+00
-7.10242629e-01 -6.36133492e-01 -3.43187004e-01 3.47843498e-01
-4.41583157e-01 -3.66457850e-02 -2.65588701e-01 9.97583151e-01
4.77460533e-01 7.36106396e-01 1.67957857e-01 -4.28089738e-01
2.86825687e-01 -2.17933163e-01 1.85097426e-01 -7.81867877e-02
-5.11659801e-01 4.03801292e-01 -5.32385767e-01 -9.44405675e-01
-9.61546183e-01 -6.34799004e-01 -1.31104267e+00 1.56058788e-01
-5.13796151e-01 -5.88981695e-02 3.32715750e-01 8.36053252e-01
3.41320902e-01 7.70115137e-01 6.26750946e-01 -1.39436114e+00
-1.13212422e-01 -7.95091569e-01 -7.82097697e-01 2.51288503e-01
7.70430982e-01 -9.83797729e-01 -4.76107299e-01 -2.32657745e-01] | [9.691781997680664, -0.7652511596679688] |
3ccd59df-4e15-4c27-ad16-e88fc5edf5d8 | a-transformer-based-audio-captioning-model | 2007.00222 | null | https://arxiv.org/abs/2007.00222v2 | https://arxiv.org/pdf/2007.00222v2.pdf | A Transformer-based Audio Captioning Model with Keyword Estimation | One of the problems with automated audio captioning (AAC) is the indeterminacy in word selection corresponding to the audio event/scene. Since one acoustic event/scene can be described with several words, it results in a combinatorial explosion of possible captions and difficulty in training. To solve this problem, we propose a Transformer-based audio-captioning model with keyword estimation called TRACKE. It simultaneously solves the word-selection indeterminacy problem with the main task of AAC while executing the sub-task of acoustic event detection/acoustic scene classification (i.e., keyword estimation). TRACKE estimates keywords, which comprise a word set corresponding to audio events/scenes in the input audio, and generates the caption while referring to the estimated keywords to reduce word-selection indeterminacy. Experimental results on a public AAC dataset indicate that TRACKE achieved state-of-the-art performance and successfully estimated both the caption and its keywords. | ['Shoichiro Saito', 'Masahiro Yasuda', 'Yuma Koizumi', 'Ryo Masumura', 'Kyosuke Nishida'] | 2020-07-01 | null | null | null | null | ['audio-captioning'] | ['audio'] | [ 6.49836719e-01 -1.52630314e-01 2.94729829e-01 -1.15744218e-01
-1.82893527e+00 -7.52128184e-01 1.51445717e-01 1.76595002e-02
-1.66856110e-01 6.09564483e-01 5.63639522e-01 -1.06465437e-01
7.80514404e-02 -1.56916067e-01 -9.04358149e-01 -4.37840939e-01
1.38204083e-01 6.14829540e-01 2.13107929e-01 2.04979271e-01
7.93355778e-02 -4.69094813e-02 -1.83095658e+00 5.93884587e-01
6.44714415e-01 1.30174220e+00 7.11379290e-01 1.05098867e+00
-4.64027643e-01 7.05984771e-01 -1.05085754e+00 -1.84615329e-01
-2.49578655e-01 -5.68967760e-01 -5.52477241e-01 3.93747538e-01
4.71229553e-01 -5.95073216e-02 -2.08668113e-01 9.01541173e-01
6.64527416e-01 5.15778586e-02 4.94459957e-01 -1.73940539e+00
-1.49609551e-01 8.17629457e-01 2.51929890e-02 1.78309619e-01
5.49493849e-01 -2.50536621e-01 1.35402274e+00 -1.31743026e+00
1.54828295e-01 1.22340775e+00 4.94933605e-01 4.02013361e-01
-8.83855462e-01 -7.21286535e-01 1.47666529e-01 5.27288318e-01
-1.78966987e+00 -6.26373231e-01 8.20632637e-01 -3.08159173e-01
7.18079448e-01 6.89360797e-01 7.04459369e-01 1.02744889e+00
-2.24602371e-01 1.06109381e+00 3.30710888e-01 -4.41505283e-01
4.09580976e-01 -1.69134829e-02 -5.07308915e-02 1.49338916e-01
-4.08701420e-01 -5.34771681e-01 -8.83756042e-01 -3.68209541e-01
3.61083597e-01 -6.60766602e-01 -4.99649197e-01 2.50968635e-01
-1.28837192e+00 4.69246805e-01 -3.36760849e-01 -1.57212555e-01
-4.30970788e-01 2.61140525e-01 5.48422992e-01 -4.20213379e-02
1.94605887e-01 6.08474255e-01 -4.79584724e-01 -5.42100310e-01
-9.38049853e-01 2.62776971e-01 6.65953636e-01 1.20604658e+00
4.14381027e-01 2.20209479e-01 -5.89434922e-01 1.19877779e+00
2.24287942e-01 5.32080352e-01 4.53755826e-01 -6.99918091e-01
9.20086324e-01 1.64066195e-01 3.16274673e-01 -5.54410636e-01
-1.13319434e-01 -5.18322408e-01 -2.95326412e-01 -6.21332169e-01
1.63995594e-01 -1.99792966e-01 -1.02930748e+00 1.84984136e+00
1.38604969e-01 7.42977977e-01 -5.68935126e-02 9.14215028e-01
7.56010234e-01 1.30329311e+00 3.15245420e-01 -4.17399317e-01
1.84653211e+00 -9.05673385e-01 -1.17959702e+00 -5.81868410e-01
2.01835498e-01 -1.11126709e+00 1.11152864e+00 4.29989725e-01
-8.93103719e-01 -4.27041471e-01 -7.79392183e-01 9.22764763e-02
7.09054992e-02 5.42536914e-01 3.23875159e-01 3.91228735e-01
-7.50191212e-01 -2.73709476e-01 -3.67321670e-01 8.91430452e-02
1.47481915e-02 2.69296736e-01 4.43257168e-02 1.24531709e-01
-1.41082954e+00 3.52208287e-01 6.61615670e-01 2.48570344e-03
-1.29220390e+00 -9.29861724e-01 -9.57636774e-01 4.94675547e-01
6.51612759e-01 -3.48876029e-01 1.85069466e+00 -8.51714849e-01
-1.31542969e+00 1.15996450e-01 -4.78377521e-01 -4.19795185e-01
1.17495485e-01 -2.28199556e-01 -8.88926268e-01 2.61757821e-01
3.35409552e-01 8.15647125e-01 1.08495629e+00 -1.24621773e+00
-1.03436565e+00 1.85363948e-01 -5.43452382e-01 5.72543025e-01
-4.89337921e-01 2.60448754e-01 -1.00490272e+00 -9.35042381e-01
1.21738143e-01 -9.21480298e-01 1.74828127e-01 -2.81113297e-01
-7.40328252e-01 -4.19762939e-01 8.46505761e-01 -8.74893367e-01
1.63573253e+00 -2.53134727e+00 -1.48177549e-01 -9.24872011e-02
-2.11812869e-01 4.17535789e-02 -2.66131848e-01 3.59546155e-01
-1.84488490e-01 1.03226118e-01 -2.13998735e-01 -5.72041154e-01
-3.66115980e-02 9.78940874e-02 -9.86940801e-01 2.65091173e-02
3.17582697e-01 4.40395057e-01 -1.04634142e+00 -8.96804810e-01
9.01421010e-02 5.20765483e-01 -3.74080896e-01 5.22110760e-01
-6.40129507e-01 1.00233503e-01 -5.17742753e-01 6.38752043e-01
4.43267286e-01 -1.01781122e-01 -2.72390507e-02 -3.26919854e-01
-3.19085456e-02 6.28080845e-01 -1.41519666e+00 1.53637600e+00
-5.74306846e-01 8.66872191e-01 -2.41677975e-03 -4.92592961e-01
5.28428137e-01 9.72967207e-01 5.52303910e-01 -1.80594325e-01
-3.75022851e-02 4.18526173e-01 -4.08628434e-01 -4.87028509e-01
8.11934233e-01 7.98468813e-02 -3.92952502e-01 1.11227125e-01
9.12679136e-02 -4.42638904e-01 5.83626553e-02 3.17148000e-01
9.45231676e-01 -2.50449985e-01 -2.00236868e-02 1.01439506e-01
4.65451896e-01 1.00463361e-01 4.53574240e-01 7.94126868e-01
4.19093445e-02 1.00619864e+00 3.19612920e-01 -2.46778652e-02
-9.69247699e-01 -1.16940045e+00 6.93356842e-02 1.24200737e+00
8.50583017e-02 -8.19190145e-01 -7.81903505e-01 -3.00664514e-01
-4.95493561e-01 1.02937615e+00 -1.41896129e-01 1.63753685e-02
-6.02005959e-01 -3.40023935e-01 7.57977724e-01 3.16761792e-01
1.08813606e-01 -1.08986354e+00 -3.14948231e-01 5.43839693e-01
-8.90376031e-01 -1.52477002e+00 -1.06262612e+00 1.38653725e-01
-1.63243562e-01 -6.77424967e-01 -6.35815799e-01 -9.97530341e-01
4.07236516e-01 1.36661321e-01 1.10086632e+00 -6.39673352e-01
-2.20881790e-01 5.31607985e-01 -4.70189899e-01 -7.67712831e-01
-6.35901690e-01 -1.13240622e-01 -4.70104590e-02 4.13506091e-01
-7.20326826e-02 -2.33659357e-01 -3.41095597e-01 3.79363626e-01
-9.64643359e-01 1.62515640e-01 3.88804585e-01 6.35288477e-01
9.16799247e-01 2.11852595e-01 8.32038701e-01 -1.68014571e-01
6.75236344e-01 -4.28723931e-01 -5.81678808e-01 4.22422200e-01
-1.02108322e-01 -1.32779956e-01 7.59159505e-01 -9.33046639e-01
-8.64013195e-01 4.54979956e-01 -1.70930818e-01 -5.43648303e-01
3.12571079e-02 4.59092885e-01 -4.14434552e-01 5.78637779e-01
2.75585145e-01 7.55461872e-01 -4.25332516e-01 -5.46900094e-01
1.78336039e-01 1.20107174e+00 7.83957362e-01 -4.56185549e-01
4.32827860e-01 3.28636020e-02 -3.02403569e-01 -1.04073250e+00
-9.79246199e-01 -7.75071800e-01 3.11721146e-01 -5.59562385e-01
7.60269761e-01 -1.42252004e+00 -4.26183194e-01 2.65301853e-01
-1.53961492e+00 3.20458561e-01 -3.91971499e-01 5.78918040e-01
-6.21711314e-01 3.34331810e-01 -2.86484182e-01 -1.19975328e+00
-3.74713868e-01 -1.31192124e+00 1.72544336e+00 -7.30925202e-02
-5.31595469e-01 -2.02414379e-01 -1.52709752e-01 4.23861831e-01
7.96440467e-02 -3.77961606e-01 9.71103489e-01 -8.71745527e-01
-5.61048985e-01 -2.58458108e-01 -3.70849185e-02 1.88964218e-01
2.30831467e-02 -2.80176371e-01 -1.30161750e+00 4.66872193e-02
-1.84600458e-01 -1.08520038e-01 5.01622438e-01 3.03571939e-01
1.55055082e+00 -4.79341656e-01 -6.59446344e-02 3.73292789e-02
1.03497410e+00 5.35170555e-01 5.33276081e-01 -1.14123382e-01
6.59782231e-01 2.77753294e-01 9.18290257e-01 7.77475238e-01
2.26015225e-01 1.13538790e+00 5.07351220e-01 1.85633153e-01
-3.71480793e-01 -6.18333399e-01 4.80134070e-01 1.04447925e+00
7.92213261e-01 -8.98224592e-01 -8.17003965e-01 1.03485429e+00
-1.65016747e+00 -6.42881274e-01 -4.24008109e-02 2.05255795e+00
1.03396678e+00 -1.07783549e-01 5.82070462e-02 5.05763292e-01
1.20025003e+00 1.28802685e-02 -3.70707542e-01 -1.24320254e-01
-6.52052015e-02 -8.05542693e-02 1.35991588e-01 3.88915002e-01
-1.01024365e+00 8.82056177e-01 6.06683683e+00 1.37309670e+00
-9.01483119e-01 1.08922571e-01 4.77065116e-01 -1.41661108e-01
-3.18492264e-01 -1.96606845e-01 -9.41222548e-01 7.31073678e-01
1.30801070e+00 -2.62978703e-01 4.06778544e-01 7.83600807e-01
3.29917818e-01 3.31170373e-02 -1.18636334e+00 1.29933465e+00
2.13105693e-01 -1.11407661e+00 5.00012875e-01 -4.00604099e-01
3.27748597e-01 -1.75506264e-01 1.56453308e-02 2.97806412e-01
-3.86210978e-01 -6.12876177e-01 1.38043559e+00 -2.53431667e-02
1.01282084e+00 -5.68226039e-01 4.05733913e-01 1.35291368e-01
-1.55055714e+00 -1.88793868e-01 8.13166723e-02 2.88826019e-01
4.80024487e-01 5.02050757e-01 -1.48678899e+00 1.79627910e-01
5.82736552e-01 1.62677228e-01 -1.42123863e-01 1.53660989e+00
-2.09812090e-01 9.89721417e-01 -4.82031733e-01 -1.95270613e-01
1.23999692e-01 2.63793796e-01 1.06500077e+00 1.36630261e+00
7.60910749e-01 -9.99186784e-02 1.60546839e-01 5.73105037e-01
-1.92975149e-01 3.23656857e-01 6.59533311e-03 -3.52163732e-01
1.07975376e+00 8.64470959e-01 -5.52004576e-01 -3.81794214e-01
-3.75390574e-02 9.57737029e-01 -3.67671490e-01 4.31265891e-01
-1.10319698e+00 -4.07732815e-01 6.48997009e-01 -3.01094875e-02
4.37464893e-01 5.73328957e-02 9.84719545e-02 -7.47653723e-01
3.12347740e-01 -8.81912529e-01 4.73723203e-01 -1.27823555e+00
-9.31934893e-01 7.43340790e-01 6.31575808e-02 -1.62385559e+00
-2.92927831e-01 -2.78837234e-01 -4.98178065e-01 7.17791200e-01
-1.36149693e+00 -8.39452684e-01 -1.70651041e-02 4.83408093e-01
1.34156430e+00 1.04086073e-02 9.05189931e-01 5.48285604e-01
-5.44277966e-01 5.39920986e-01 -1.50965378e-01 -4.97499928e-02
6.21017098e-01 -1.00938678e+00 3.75616342e-01 7.74781704e-01
3.91277343e-01 -1.00328550e-01 1.00993586e+00 -6.07953548e-01
-1.55332828e+00 -1.44748938e+00 1.33806622e+00 -2.55181760e-01
8.68662596e-01 -6.65378809e-01 -6.21803343e-01 3.77264678e-01
-7.57815465e-02 -1.97148114e-01 6.81308329e-01 -4.00530368e-01
-1.83641225e-01 -4.11326550e-02 -5.82829833e-01 6.86633229e-01
6.60760939e-01 -9.27568376e-01 -5.59608698e-01 6.19477570e-01
1.33914089e+00 -6.00037336e-01 -3.38430375e-01 1.65380478e-01
3.36437225e-01 5.49032167e-02 1.00877500e+00 -3.07771593e-01
1.45926803e-01 -6.14119709e-01 -2.75364727e-01 -1.17199254e+00
1.38381183e-01 -9.83969688e-01 -2.20443919e-01 1.49063551e+00
7.88861811e-01 9.12268385e-02 5.13720214e-01 3.90355468e-01
-6.08512640e-01 -2.74168789e-01 -1.56711173e+00 -8.53713751e-01
-8.31711590e-01 -9.74199116e-01 6.44116402e-01 5.37339509e-01
-1.87054183e-02 5.92194557e-01 -6.78710461e-01 6.28684282e-01
4.22843516e-01 -1.45692050e-01 1.96232855e-01 -9.01701331e-01
-1.68216586e-01 1.31647754e-02 -1.89080030e-01 -1.05377293e+00
1.69494435e-01 -5.53800285e-01 7.64538527e-01 -1.52945256e+00
-1.55107200e-01 -1.94955647e-01 -1.74720645e-01 5.85150003e-01
-1.43893063e-01 6.37039840e-02 1.99615791e-01 1.37220964e-01
-8.84744704e-01 6.07343256e-01 9.74159658e-01 -5.22955120e-01
-4.07369733e-01 2.65254647e-01 -4.37225103e-01 2.70141214e-01
6.21174574e-01 -7.05115497e-01 -5.29107153e-01 -4.15759504e-01
3.28706086e-01 5.10610521e-01 2.75570661e-01 -1.11474001e+00
3.55891079e-01 6.83115199e-02 -1.90218210e-01 -8.53120446e-01
1.03285789e+00 -1.01502466e+00 3.11175823e-01 -2.38462817e-02
-6.67101860e-01 -4.58730198e-03 5.43130934e-01 7.09818542e-01
-5.00830173e-01 -3.46679300e-01 2.45389000e-01 8.17042142e-02
-6.67339563e-01 5.02764657e-02 -8.54650438e-01 1.85012966e-01
7.45010078e-01 4.69390340e-02 -1.13576345e-01 -6.94029033e-01
-6.74205720e-01 1.73933268e-01 -4.24873114e-01 5.76705635e-01
7.95322955e-01 -1.29142165e+00 -8.90846789e-01 3.23172077e-03
4.56692845e-01 -2.28882022e-02 2.45127946e-01 2.34551325e-01
-1.59268320e-01 4.88499582e-01 5.08124232e-01 -6.42531276e-01
-1.47837913e+00 2.07433268e-01 1.01390101e-01 7.44771492e-03
-3.10379148e-01 9.89899457e-01 2.03207895e-01 2.53041625e-01
7.08533168e-01 -3.12435865e-01 -2.98045129e-01 1.93139851e-01
6.80619240e-01 6.08582497e-02 1.83415741e-01 -7.38199711e-01
-3.67154479e-01 1.73259094e-01 9.82907712e-02 -4.92084771e-01
8.97511065e-01 -1.94462210e-01 7.05045015e-02 4.28188413e-01
1.21380901e+00 2.73580495e-02 -9.45921600e-01 -1.81324154e-01
-7.27159604e-02 -2.09204778e-01 3.35873574e-01 -8.97981465e-01
-5.21200359e-01 6.89845979e-01 6.26408815e-01 2.90766060e-01
1.33325970e+00 1.71174541e-01 1.04687691e+00 4.28515643e-01
1.76062912e-01 -1.16842830e+00 1.32912412e-01 6.36384487e-01
9.94220972e-01 -7.88051665e-01 -4.07644391e-01 -7.93569267e-01
-8.48513246e-01 9.09476638e-01 5.20580769e-01 6.33881211e-01
3.10785443e-01 2.52099872e-01 2.20170189e-02 4.77153175e-02
-7.12197602e-01 1.33518409e-02 5.49167454e-01 2.57059842e-01
-7.76056200e-02 2.63065636e-01 9.74103510e-02 7.70608068e-01
-3.65226239e-01 -3.45190793e-01 5.04420519e-01 6.93043470e-01
-4.83813375e-01 -9.53381121e-01 -7.07253337e-01 3.01017731e-01
-5.51995397e-01 -5.56506872e-01 -3.27306330e-01 1.05339982e-01
-8.62208083e-02 1.37049866e+00 2.75397688e-01 -5.54931402e-01
2.63159454e-01 3.55442196e-01 -1.64276868e-01 -7.77715385e-01
-5.11621535e-01 5.63675344e-01 4.42170471e-01 -2.72121191e-01
-3.84317450e-02 -5.32428265e-01 -1.24027443e+00 6.68314636e-01
-6.59193099e-01 8.25995624e-01 1.10673165e+00 9.56393957e-01
5.46419203e-01 9.04105604e-01 7.30204999e-01 -5.78893721e-01
-4.47755545e-01 -9.01820362e-01 -4.86157328e-01 1.45326942e-01
4.38328028e-01 -2.89673150e-01 -5.34604311e-01 3.40418667e-01] | [15.280731201171875, 4.920375347137451] |
a62c6e62-62cb-4dbf-8c07-a0d3398b8f2e | high-dynamic-range-image-forensics-using-cnn | 1902.10938 | null | http://arxiv.org/abs/1902.10938v1 | http://arxiv.org/pdf/1902.10938v1.pdf | High dynamic range image forensics using cnn | High dynamic range (HDR) imaging has recently drawn much attention in
multimedia community. In this paper, we proposed a HDR image forensics method
based on convolutional neural network (CNN).To our best knowledge, this is the
first time to apply deep learning method on HDR image forensics. The proposed
algorithm uses CNN to distinguish HDR images generated by multiple low dynamic
range (LDR) images from that expanded by single LDR image using inverse tone
mapping (iTM). To do this, we learn the change of statistical characteristics
extracted by the proposed CNN architectures and classify two kinds of HDR
images. Comparision results with some traditional statistical characteristics
shows efficiency of the proposed method in HDR image source identification. | ['Xiaofeng Zhu', 'Yongqing Huo'] | 2019-02-28 | null | null | null | null | ['tone-mapping', 'image-forensics', 'inverse-tone-mapping'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 1.37798771e-01 -4.09387112e-01 1.85201198e-01 -4.32710834e-02
-6.00622416e-01 -3.14426333e-01 6.48643017e-01 -3.74505639e-01
-3.65412384e-01 7.10156024e-01 3.31399515e-02 -3.54686379e-01
1.89835243e-02 -1.10596657e+00 -7.17563450e-01 -5.66139221e-01
-3.54082622e-02 -2.77117975e-02 2.48959407e-01 -4.47483569e-01
5.39748073e-01 8.14645290e-01 -1.50716972e+00 4.57881570e-01
4.17453140e-01 1.06183648e+00 1.24180801e-02 7.80800700e-01
1.28006786e-01 1.24761724e+00 -8.91803741e-01 -2.09937438e-01
5.13674676e-01 -3.84643167e-01 -7.32639313e-01 -1.58848763e-01
4.20829684e-01 -9.36816037e-01 -1.22258878e+00 1.08779025e+00
9.13369238e-01 1.61830381e-01 5.16885400e-01 -1.00351918e+00
-1.57343853e+00 6.36522770e-01 -1.04263306e+00 1.05628967e+00
5.28971672e-01 8.95881653e-02 1.82282701e-01 -5.84475040e-01
6.94905639e-01 1.31108570e+00 7.24346757e-01 2.81038195e-01
-7.56917715e-01 -1.08348882e+00 -8.66902947e-01 6.78184152e-01
-1.40028167e+00 -2.19444796e-01 1.21886170e+00 -1.28500313e-01
8.27123046e-01 -8.54273364e-02 2.94194818e-01 8.54602277e-01
2.79611975e-01 5.19324064e-01 1.86015713e+00 -3.47016513e-01
-1.18226513e-01 8.46449099e-03 1.65184736e-01 4.91255373e-01
1.50034890e-01 4.53812152e-01 -5.66068172e-01 2.30086997e-01
7.90788710e-01 -1.65611327e-01 -1.36179984e-01 6.22298062e-01
-7.57831752e-01 9.15271878e-01 1.73568651e-01 5.05448699e-01
-3.78056169e-02 2.20275596e-01 5.23416042e-01 6.57323480e-01
3.14972758e-01 1.17627144e-01 1.06418751e-01 -3.56324241e-02
-9.89832699e-01 -1.45390481e-01 3.05911720e-01 6.85721815e-01
5.09993970e-01 5.07462502e-01 1.37542978e-01 9.57482219e-01
-1.03608623e-01 7.16214538e-01 5.77579796e-01 -9.36030805e-01
1.41234517e-01 -5.60892709e-02 -4.24499005e-01 -1.31994128e+00
-1.44907027e-01 -2.96572391e-02 -8.85530412e-01 5.98570466e-01
2.49509960e-01 3.40762362e-02 -8.84552836e-01 1.09560061e+00
-3.80674936e-02 1.95965722e-01 8.93327221e-02 9.45462704e-01
1.19009769e+00 9.04964328e-01 -1.19089924e-01 6.22210931e-03
1.16611767e+00 -2.71427870e-01 -1.00924850e+00 5.13569057e-01
-7.40665123e-02 -1.03259802e+00 6.96011424e-01 7.83873320e-01
-9.00715828e-01 -9.84481096e-01 -1.26477230e+00 -2.90952832e-01
-6.10918403e-01 1.50134163e-02 3.80171478e-01 1.06846058e+00
-9.50243354e-01 7.08288074e-01 -4.99146655e-02 -2.70490706e-01
6.16213083e-01 3.30542028e-01 -4.44041908e-01 -2.04821631e-01
-1.70575511e+00 6.65150940e-01 6.21698320e-01 -1.17663294e-02
-9.16975498e-01 -4.48237866e-01 -5.66744328e-01 -2.98480123e-01
-4.64955457e-02 1.54878721e-01 6.71164751e-01 -7.91364372e-01
-1.60982525e+00 1.08161712e+00 6.73465312e-01 -5.98682642e-01
4.00000989e-01 -1.27776846e-01 -1.03420877e+00 6.95130646e-01
-1.19528838e-01 3.04358095e-01 1.03076434e+00 -1.23660088e+00
-2.84619302e-01 -3.30702454e-01 -2.11431280e-01 -4.43943530e-01
-1.32166445e-01 4.78381693e-01 -4.87170666e-02 -7.47517347e-01
-1.64148629e-01 -7.34096646e-01 6.53046131e-01 -1.28040731e-01
-4.86054391e-01 2.28454545e-01 1.48548353e+00 -1.27133179e+00
9.15230393e-01 -2.11097097e+00 -6.66198194e-01 2.89661624e-02
2.25300863e-01 6.09077632e-01 1.31880576e-02 2.16747597e-01
-4.63657588e-01 1.58441588e-01 -1.04149640e-01 3.94933999e-01
-2.21025154e-01 -3.71399432e-01 -5.56645274e-01 9.85748887e-01
6.87123761e-02 8.46384406e-01 -4.63976413e-01 -6.71641707e-01
5.63292861e-01 7.79419124e-01 1.39747947e-01 1.93412513e-01
3.73393387e-01 5.76053023e-01 -1.05529100e-01 8.54909003e-01
1.46982133e+00 3.11256796e-01 -2.11998820e-01 -7.87706852e-01
-2.89055854e-01 -4.30078149e-01 -9.76896882e-01 1.09542596e+00
-3.77818584e-01 1.37037241e+00 -5.09650230e-01 -9.69858646e-01
1.47649622e+00 1.28754675e-01 4.60646093e-01 -1.42685294e+00
5.22808194e-01 2.43498415e-01 6.01991639e-02 -8.92066360e-01
8.07609379e-01 -2.71648824e-01 7.13986345e-04 7.29723155e-01
1.38520479e-01 2.32512191e-01 -1.36631921e-01 -4.93834838e-02
8.99378598e-01 -9.96555090e-02 2.59053111e-01 4.48869243e-02
6.28707349e-01 -2.92929947e-01 1.38522387e-01 8.87283325e-01
-4.65154588e-01 9.51556861e-01 1.83524981e-01 -6.18383825e-01
-1.59339941e+00 -1.10477483e+00 -4.42900419e-01 7.40929067e-01
2.29131281e-01 5.28272092e-01 -5.38621068e-01 -1.36514932e-01
-1.56059131e-01 5.44038177e-01 -8.40848982e-01 -7.66228959e-02
-1.00862968e+00 -8.99811208e-01 1.27743471e+00 2.74705231e-01
1.39111936e+00 -1.40801525e+00 -4.70292479e-01 -2.24749520e-02
4.55229357e-03 -1.33679426e+00 -1.29267842e-01 4.01089825e-02
-4.63891625e-01 -1.06882417e+00 -7.38784790e-01 -5.98277926e-01
-1.07932925e-01 4.07206953e-01 9.07372773e-01 -8.55014026e-02
-7.50241578e-01 1.63976222e-01 -7.82409251e-01 -7.12848976e-02
-6.62686825e-01 -2.17894509e-01 -3.19167882e-01 1.18851095e-01
3.78450662e-01 -5.10561883e-01 -6.22266173e-01 -4.75205779e-02
-1.31397045e+00 -4.15246487e-01 7.51224697e-01 4.38210577e-01
4.27511036e-01 7.10253835e-01 5.50621212e-01 -8.82626712e-01
6.19614065e-01 -6.83753371e-01 -3.02239150e-01 1.21535659e-01
-3.70795369e-01 -1.49480432e-01 7.63658166e-01 -4.42962646e-01
-1.38404727e+00 -4.06554192e-01 -1.10848479e-01 -6.72135770e-01
-5.66385329e-01 -1.78251415e-01 -5.89688402e-03 -3.44303936e-01
5.91397107e-01 5.29778540e-01 -9.31908786e-02 -4.12246287e-01
4.58282679e-01 1.15570271e+00 1.03141868e+00 -4.18983161e-01
9.54952121e-01 9.65588868e-01 -2.20280662e-02 -9.64040279e-01
-3.17273229e-01 -1.87693834e-01 -4.51285183e-01 -3.87848616e-01
1.23041379e+00 -8.53786647e-01 -8.21103454e-01 9.88788426e-01
-9.94206429e-01 -1.83091581e-01 -3.14488336e-02 2.94180155e-01
-4.11476195e-01 7.55260766e-01 -1.03825319e+00 -8.54700565e-01
-4.00826782e-01 -1.00215936e+00 9.96441483e-01 1.97733656e-01
4.54912901e-01 -8.70197415e-01 8.69120657e-02 3.32657874e-01
6.55921996e-01 7.06652880e-01 6.49929464e-01 -5.33018291e-01
-4.47524309e-01 -1.60034180e-01 -7.28372872e-01 3.59860957e-01
2.28149034e-02 -9.49661434e-02 -1.25349772e+00 -1.87036544e-01
3.71413410e-01 -3.32941771e-01 9.76687253e-01 3.64911169e-01
1.52506518e+00 -1.39258772e-01 1.99256673e-01 8.31134260e-01
1.88610220e+00 6.31755114e-01 1.52885938e+00 9.21194971e-01
7.52519786e-01 2.45285749e-01 5.24351358e-01 6.62249029e-01
-2.61657625e-01 5.51773906e-01 2.05675676e-01 -3.96591902e-01
-7.35853553e-01 3.48613262e-02 4.48386520e-01 5.22502482e-01
5.27950488e-02 -3.82056296e-01 -6.81648076e-01 9.28421766e-02
-9.84753191e-01 -1.45444286e+00 -4.21878070e-01 1.65188801e+00
5.37601292e-01 -1.74673155e-01 1.04587995e-01 3.02333623e-01
1.41870737e+00 3.76155347e-01 -5.19963861e-01 -5.51750243e-01
-6.42369568e-01 7.06838012e-01 8.24113309e-01 6.23272508e-02
-1.15461218e+00 7.75219440e-01 5.93514252e+00 1.27360916e+00
-1.45146787e+00 2.82416791e-01 9.26013947e-01 2.39266679e-01
-3.02474238e-02 -2.56833553e-01 -3.85722578e-01 6.46929026e-01
8.46829832e-01 2.15005800e-01 5.76062381e-01 5.97996116e-01
1.39787629e-01 -1.18604168e-01 -3.36419046e-01 1.45923758e+00
5.47877848e-01 -1.12987363e+00 1.30645275e-01 2.40537643e-01
6.94439769e-01 -2.83064306e-01 7.63030946e-01 -5.85791953e-02
7.92374536e-02 -1.18580687e+00 5.56424201e-01 4.58944678e-01
1.29420221e+00 -1.11523175e+00 9.09005165e-01 -4.20706600e-01
-9.53543782e-01 -3.01197410e-01 -7.29593396e-01 4.66636807e-01
1.65177684e-03 6.19020820e-01 -5.32924891e-01 3.05444807e-01
1.03613544e+00 5.82675099e-01 -7.36360013e-01 6.02279961e-01
3.37259263e-01 5.25731027e-01 2.41546601e-01 6.09913826e-01
-1.23974256e-01 2.65419716e-03 5.18028080e-01 1.33342397e+00
5.10384619e-01 1.26301169e-01 -4.73242223e-01 1.00910890e+00
-2.22167507e-01 -1.26739800e-01 -9.57575321e-01 -7.91368335e-02
4.88052845e-01 1.27684820e+00 -1.06603181e+00 -3.01566631e-01
-3.07058930e-01 9.80500996e-01 -2.82848448e-01 7.95198753e-02
-1.00174606e+00 -8.85160089e-01 -7.58891255e-02 1.98234454e-01
4.24520493e-01 -2.57494718e-01 -1.70946628e-01 -8.60578954e-01
-5.82286298e-01 -8.67449760e-01 3.97988081e-01 -1.04799533e+00
-1.44082308e+00 8.55549812e-01 -2.98699308e-02 -1.44310188e+00
1.92513466e-01 -6.34180009e-01 -5.17982662e-01 4.98600066e-01
-1.84601569e+00 -1.05647790e+00 -5.64684153e-01 9.78177249e-01
4.43595171e-01 -7.76928663e-01 1.65077135e-01 6.58767402e-01
-2.46482894e-01 6.03249907e-01 3.32512319e-01 5.47099113e-01
9.74465966e-01 -1.10373878e+00 2.00383037e-01 8.41332316e-01
-3.75231832e-01 5.02712250e-01 7.47456610e-01 -7.47614861e-01
-1.25504589e+00 -1.16314220e+00 -6.12235209e-03 5.32321334e-02
3.89190316e-01 1.32289022e-01 -9.89025652e-01 2.48256058e-01
6.07002258e-01 -1.04467869e-01 7.98390031e-01 -6.96042895e-01
-7.83979475e-01 -1.90255150e-01 -1.69198406e+00 -3.70388851e-02
3.97433311e-01 -8.50706041e-01 -4.89247084e-01 -2.98235733e-02
6.41744971e-01 -8.85566920e-02 -1.14067507e+00 2.47678936e-01
4.51456487e-01 -1.38407612e+00 1.29083431e+00 4.95320633e-02
6.90347433e-01 -4.37682748e-01 -5.15626252e-01 -6.49098635e-01
-9.28746313e-02 -3.27274352e-01 -1.68519150e-02 1.36452425e+00
-3.86219531e-01 -3.69187176e-01 4.08200920e-01 2.37285011e-02
-8.20553526e-02 -4.12742514e-03 -7.29292691e-01 -8.32666337e-01
2.62468815e-01 -3.30214173e-01 6.38693571e-01 1.18514454e+00
-6.92054152e-01 -2.25385636e-01 -8.96360219e-01 9.88875180e-02
8.93901527e-01 1.15120091e-01 4.33914602e-01 -8.90925407e-01
-4.22875255e-01 -1.31390870e-01 -5.71303904e-01 -1.07204892e-01
3.09693813e-01 -6.97594762e-01 -2.39261493e-01 -1.02033770e+00
5.67275643e-01 -3.71993721e-01 -3.30065548e-01 2.97073834e-02
1.63339436e-01 1.09389687e+00 4.51182276e-01 3.53703201e-01
-5.18665493e-01 3.03255379e-01 1.17254865e+00 -2.78973997e-01
3.73924859e-02 -7.84578145e-01 -5.25343001e-01 3.59981596e-01
1.27594590e+00 -5.66176414e-01 -1.22433119e-01 -2.54236460e-01
8.17216411e-02 2.34173179e-01 5.32818735e-01 -1.36945844e+00
-8.13025236e-02 1.32912889e-01 8.93114746e-01 -6.26207948e-01
1.00695506e-01 -4.62961257e-01 4.40764755e-01 3.81015629e-01
-4.38474089e-01 2.08997540e-02 1.53962940e-01 4.72089708e-01
-2.99867988e-01 -2.03639835e-01 1.27439260e+00 -4.14241284e-01
-1.07073462e+00 2.64949739e-01 -6.41484976e-01 -2.78008375e-02
9.35170531e-01 -5.25325298e-01 -8.40425193e-01 -3.48006576e-01
-4.46437687e-01 -9.57113862e-01 3.50255042e-01 3.65436614e-01
1.06758773e+00 -1.47322369e+00 -6.54791176e-01 -5.49837723e-02
-2.55344659e-01 -6.66767299e-01 7.30832934e-01 5.03417552e-01
-1.05272484e+00 1.99541897e-01 -1.09633517e+00 -5.88587895e-02
-1.16441500e+00 8.72675896e-01 2.59270430e-01 1.25193954e-01
-8.23824406e-01 1.64553434e-01 -7.07004666e-02 -1.29275888e-01
-4.19523329e-01 4.95028526e-01 -6.62562072e-01 -9.24688950e-02
7.26470888e-01 8.96529019e-01 -5.97691871e-02 -1.18209636e+00
-3.89157943e-02 7.99814820e-01 -1.11664243e-01 -2.06946224e-01
1.55622780e+00 -3.22284490e-01 -3.24995577e-01 1.46638542e-01
2.02128530e+00 -5.11928909e-02 -7.97977984e-01 8.60874057e-02
-4.34519172e-01 -9.48762238e-01 3.47877026e-01 -5.24147213e-01
-1.63157618e+00 1.08335078e+00 1.51381123e+00 2.02160880e-01
1.25727284e+00 -1.39492109e-01 1.44566131e+00 2.34204993e-01
2.21378684e-01 -1.31494224e+00 6.70510054e-01 9.70387161e-02
6.46773696e-01 -1.20608330e+00 2.07603537e-03 -9.74249318e-02
-5.51684201e-01 1.77945817e+00 5.31972110e-01 -6.80008888e-01
6.65527105e-01 2.56188273e-01 6.99753538e-02 -2.10812807e-01
7.59573206e-02 -3.92612755e-01 -3.70728582e-01 1.21576631e+00
2.88618475e-01 -3.32238376e-01 -1.24162748e-01 -1.25967547e-01
-1.95722252e-01 1.10323496e-01 1.11545897e+00 6.31775200e-01
-5.06643534e-01 -8.58584940e-01 -8.35363626e-01 2.67031819e-01
-9.44409966e-01 -6.04048930e-03 -1.55667260e-01 9.84944820e-01
3.74199301e-01 1.03066313e+00 1.65291086e-01 -7.47106612e-01
-1.73870638e-01 -4.14739788e-01 4.97583956e-01 1.46302566e-01
-2.67069906e-01 -1.05390057e-01 -2.89329886e-01 -3.64021540e-01
-7.05177844e-01 -4.66452837e-01 -1.07281232e+00 -6.97187185e-01
-5.09253070e-02 -3.96888018e-01 7.57818818e-01 6.45807922e-01
5.86420260e-02 5.35986125e-01 9.02534902e-01 -5.78377485e-01
-4.74002399e-02 -7.19976902e-01 -1.15233302e+00 7.16953695e-01
4.66426700e-01 -6.42002165e-01 -2.88903981e-01 2.49423772e-01] | [11.003253936767578, -2.171233654022217] |
820e88b0-2782-4f74-8497-657da744c492 | implementation-of-neural-network-and-feature | 1802.06288 | null | http://arxiv.org/abs/1802.06288v1 | http://arxiv.org/pdf/1802.06288v1.pdf | Implementation of Neural Network and feature extraction to classify ECG signals | This paper presents a suitable and efficient implementation of a feature
extraction algorithm (Pan Tompkins algorithm) on electrocardiography (ECG)
signals, for detection and classification of four cardiac diseases: Sleep
Apnea, Arrhythmia, Supraventricular Arrhythmia and Long Term Atrial
Fibrillation (AF) and differentiating them from the normal heart beat by using
pan Tompkins RR detection followed by feature extraction for classification
purpose .The paper also presents a new approach towards signal classification
using the existing neural networks classifiers. | ['Amogh Raut', 'Dhruv Tyagi', 'Soumya Saxena', 'R Karthik', 'Rajesh Kumar M'] | 2018-02-17 | null | null | null | null | ['electrocardiography-ecg'] | ['methodology'] | [ 3.77442002e-01 -2.54164219e-01 -4.00818922e-02 -1.89883053e-01
-1.54017955e-02 -3.60360771e-01 -8.03998411e-02 3.18996161e-01
-4.98291731e-01 1.16925335e+00 -1.61832884e-01 -6.74086034e-01
-5.79845130e-01 -2.95621544e-01 5.19314766e-01 -5.83881497e-01
-6.07599556e-01 3.45972925e-01 -2.83013344e-01 1.74643770e-01
2.27238372e-01 9.03771818e-01 -1.18161833e+00 -4.72299457e-02
6.14854813e-01 1.06539762e+00 -5.45942724e-01 1.43068981e+00
5.67110121e-01 3.86280447e-01 -1.10081136e+00 5.02844870e-01
2.25791037e-01 -9.82083559e-01 -5.57432652e-01 -3.80638808e-01
-1.79940864e-01 -2.95885559e-02 2.49716938e-01 5.00483453e-01
9.90626097e-01 -2.42652163e-01 8.90876532e-01 -1.15277970e+00
3.69471073e-01 3.55400652e-01 -2.18996942e-01 1.13968754e+00
5.14301419e-01 -2.40151867e-01 3.82109880e-01 -6.53628826e-01
-2.62312815e-02 4.71897990e-01 1.31780303e+00 3.08386505e-01
-1.00015306e+00 -7.52916157e-01 -9.33917820e-01 1.88801229e-01
-1.57704222e+00 -2.15444371e-01 6.32947385e-01 -2.34300449e-01
1.27347898e+00 9.15803194e-01 1.39532244e+00 5.03774703e-01
6.50165558e-01 1.16849631e-01 1.39766407e+00 -5.97333491e-01
3.51309150e-01 1.03966407e-01 7.86530972e-01 4.78462309e-01
4.23148662e-01 4.03966337e-01 -1.44560277e-01 -7.14585602e-01
7.70606518e-01 1.38051108e-01 -3.66066188e-01 2.87939698e-01
-1.20218801e+00 3.35800260e-01 -3.88908476e-01 8.04110050e-01
-8.82971227e-01 -2.65253007e-01 6.61790907e-01 7.70551383e-01
-2.35901609e-01 5.31144083e-01 -7.85301507e-01 -4.18109089e-01
-1.36214626e+00 6.92270622e-02 1.01248801e+00 4.20997709e-01
4.07442540e-01 3.23231220e-01 -1.14420928e-01 5.06738186e-01
3.02446932e-01 4.09569710e-01 1.03943646e+00 -8.47229302e-01
-2.73125887e-01 4.84291643e-01 1.78102162e-02 -8.78067672e-01
-8.96690667e-01 -5.06577253e-01 -9.30919826e-01 6.41254857e-02
1.55932724e-01 -6.07758224e-01 -6.83324039e-01 7.85939395e-01
2.73815095e-02 3.30571622e-01 3.53371382e-01 2.77646065e-01
8.88588607e-01 4.61362690e-01 -5.58855897e-03 -1.07713807e+00
1.54530144e+00 -2.70088881e-01 -1.00084651e+00 4.38520670e-01
3.38499069e-01 -5.14198124e-01 1.78187922e-01 1.10064840e+00
-8.48915100e-01 -6.73129559e-01 -1.24985039e+00 5.08169472e-01
-3.28133315e-01 1.95893839e-01 4.32018280e-01 1.24894083e+00
-9.39960003e-01 8.59528720e-01 -7.07544744e-01 -4.18863147e-01
1.76836357e-01 7.79567420e-01 -1.07041501e-01 8.74785602e-01
-1.07619095e+00 1.06148684e+00 5.65786362e-01 1.60241246e-01
-5.90984970e-02 -3.68009478e-01 -6.42998397e-01 -2.98123121e-01
-4.51381475e-01 -6.82437479e-01 6.76270187e-01 -1.00273883e+00
-1.36165249e+00 7.41636217e-01 -1.62151307e-01 -7.16487825e-01
1.51481077e-01 -2.85311610e-01 -9.35195863e-01 3.64715904e-01
-4.12333846e-01 -2.03711867e-01 1.10137713e+00 -5.53368390e-01
-5.39358914e-01 -5.64616501e-01 -7.76154220e-01 -8.18294212e-02
1.24691837e-01 2.87422657e-01 5.54297209e-01 -7.27865875e-01
1.30228564e-01 -6.22247458e-01 -2.49683321e-01 -7.88630664e-01
-5.07909596e-01 -3.24933171e-01 1.00497210e+00 -7.91940987e-01
1.81247938e+00 -2.02785778e+00 -1.63675308e-01 7.55721271e-01
3.41822863e-01 6.74110711e-01 6.42527282e-01 3.41880083e-01
-4.97544438e-01 4.74076107e-04 -4.48520750e-01 4.16496754e-01
-6.44522190e-01 6.10655427e-01 1.69880882e-01 5.05207717e-01
-1.71218410e-01 5.21181285e-01 -5.04552543e-01 -6.07168972e-01
3.98034811e-01 5.17548978e-01 3.15654278e-01 3.88928168e-02
9.56922352e-01 3.96630764e-01 -3.18081379e-01 7.85716772e-01
3.49962324e-01 2.23520592e-01 3.88055034e-02 -2.98856914e-01
-2.11036846e-01 1.57963499e-01 -1.24902630e+00 9.89419639e-01
-1.27909556e-01 5.38256884e-01 -3.18541974e-01 -1.28955805e+00
1.28952324e+00 1.13116825e+00 7.71002352e-01 4.19250615e-02
6.02836192e-01 3.85280997e-01 1.79452062e-01 -8.48666191e-01
-3.18581402e-01 -4.29918528e-01 3.95381004e-01 7.38275051e-01
3.92196029e-02 3.12984079e-01 6.06739819e-02 -4.10271913e-01
9.72177923e-01 -1.26316667e-01 1.35280323e+00 -5.84036887e-01
9.97429073e-01 -2.61522770e-01 6.07383013e-01 6.63522124e-01
-6.72717869e-01 4.05249208e-01 6.64961413e-02 -1.07860577e+00
-3.84589672e-01 -6.92073822e-01 -5.61523497e-01 8.43584016e-02
-4.48762029e-01 -4.09209609e-01 -7.62505829e-01 -7.93950379e-01
-1.76592097e-01 2.00813100e-01 -4.25011963e-01 -1.92589402e-01
-7.71959484e-01 -1.20545208e+00 9.71290588e-01 4.64953542e-01
3.78636092e-01 -1.30203736e+00 -1.52631438e+00 5.41875899e-01
4.30949815e-02 -6.25723898e-01 2.24975109e-01 7.72549689e-01
-1.46047151e+00 -1.30501568e+00 -6.85610056e-01 -5.47368526e-01
2.51918107e-01 -5.04546404e-01 1.03897595e+00 4.09510165e-01
-9.31406677e-01 2.56082922e-01 -1.46392539e-01 -8.44183028e-01
-4.83079821e-01 -3.46590310e-01 4.82175559e-01 -1.24861367e-01
8.25661421e-01 -1.22022617e+00 -6.59443438e-01 1.59053639e-01
-2.04565302e-01 -8.56738985e-01 5.59778452e-01 3.33780378e-01
5.31796277e-01 1.86012268e-01 9.21160638e-01 -8.19976151e-01
9.38722134e-01 -3.09472531e-01 -1.27122626e-01 -2.31672168e-01
-1.30981719e+00 -6.29552126e-01 4.94471699e-01 -4.57661636e-02
-6.06679060e-02 8.34206566e-02 -2.89084464e-01 -1.33198187e-01
-6.94243133e-01 2.31648549e-01 1.97546199e-01 2.10067350e-03
1.00773227e+00 5.60939372e-01 6.02749102e-02 -4.06773776e-01
-3.24339569e-01 1.12282526e+00 7.79453039e-01 1.68213323e-01
5.51736593e-01 2.03477100e-01 2.74604231e-01 -1.40001976e+00
-2.14523628e-01 -8.06951940e-01 -1.07218516e+00 -1.30537018e-01
1.01767397e+00 -6.89340770e-01 -3.99357557e-01 3.82346660e-01
-8.61353099e-01 3.43198150e-01 -6.43018901e-01 8.20882320e-01
-5.63020527e-01 4.41075414e-01 -3.53787988e-01 -1.28715312e+00
-1.11285245e+00 -3.35719496e-01 2.74286509e-01 2.57419974e-01
-1.07977498e+00 -9.94523048e-01 4.23557997e-01 -1.38999745e-01
3.44236463e-01 7.10625947e-01 7.35509872e-01 -9.97288883e-01
4.73644525e-01 -5.44985354e-01 4.10088509e-01 7.55791485e-01
5.93575060e-01 2.38960028e-01 -9.12558615e-01 -6.73265941e-03
7.12626934e-01 3.27507406e-01 4.24818158e-01 6.73666239e-01
7.20680714e-01 -7.45479092e-02 -3.79870653e-01 6.51873648e-01
1.35257566e+00 1.01779366e+00 6.76159322e-01 3.04685980e-01
1.11817308e-01 -1.75480768e-01 2.90131658e-01 4.84002024e-01
-2.37295747e-01 -2.13081390e-03 -2.73290128e-01 -3.00261825e-01
3.56165409e-01 8.01282108e-01 1.41119167e-01 6.87973440e-01
-8.93474042e-01 3.20737034e-01 -8.47358525e-01 3.24890584e-01
-1.32465208e+00 -8.33702862e-01 -5.67382097e-01 2.07045388e+00
7.18501985e-01 1.55027360e-01 6.87126815e-01 1.38592339e+00
6.13167107e-01 -3.63651901e-01 -2.28885129e-01 -1.04370272e+00
6.97255507e-02 9.65693474e-01 2.55342215e-01 1.17677428e-01
-1.12405837e+00 -1.32956393e-02 8.05223846e+00 9.84418318e-02
-1.29313719e+00 -4.56908345e-02 4.94638920e-01 4.20658678e-01
6.81184292e-01 -2.98148066e-01 -5.45098245e-01 3.36151034e-01
1.50041676e+00 -2.25409511e-02 -4.72889515e-03 5.29670954e-01
4.62241322e-01 -1.63302720e-01 -1.02996671e+00 1.48664629e+00
8.22387412e-02 -9.48185921e-01 -3.14853877e-01 -1.38944581e-01
-2.65620742e-02 -1.97886869e-01 -6.45171702e-01 2.22123414e-02
-8.51361096e-01 -9.01908696e-01 -1.26164742e-02 6.97453558e-01
8.13968360e-01 -8.14813077e-01 9.32295561e-01 3.50591131e-02
-1.13890946e+00 -4.14875388e-01 -6.62563881e-03 -3.95075947e-01
-5.89347072e-02 7.83575475e-01 -4.82568890e-01 4.99120712e-01
7.42672443e-01 1.07426214e+00 -5.26315928e-01 1.25337696e+00
1.01842415e-02 1.00599480e+00 -6.63565695e-01 1.42112702e-01
-2.07892999e-01 -1.13402583e-01 9.49283600e-01 1.28870738e+00
2.97456563e-01 4.03474808e-01 -4.31640893e-02 5.29345036e-01
5.70455492e-01 2.97648847e-01 -6.60891354e-01 3.27738583e-01
3.50626379e-01 1.35928965e+00 -9.90390003e-01 -5.28280079e-01
-1.05822660e-01 6.72371149e-01 -8.76178443e-01 6.35300055e-02
-4.30353552e-01 -9.85897243e-01 -6.02477901e-02 8.15922171e-02
-3.58906873e-02 2.49860168e-01 -7.61296749e-01 -8.20882618e-01
-1.80058673e-01 -9.38530445e-01 7.04631031e-01 -2.95188099e-01
-6.70194924e-01 7.12164402e-01 5.50670438e-02 -1.24074984e+00
-4.31928277e-01 -3.68842363e-01 -1.26714838e+00 1.04943466e+00
-1.10031497e+00 -4.94608432e-01 -3.73972394e-02 9.47149575e-01
1.59491405e-01 -5.42058289e-01 1.34209871e+00 3.13175231e-01
-4.52374220e-01 2.76605070e-01 -4.10484552e-01 7.21874610e-02
3.15678954e-01 -1.70532846e+00 -1.69527248e-01 4.99494404e-01
1.79699324e-02 6.64471328e-01 7.59137094e-01 -3.21790189e-01
-8.84329975e-01 -8.14629316e-01 1.28640270e+00 -3.16857576e-01
-1.48081295e-02 3.81946951e-01 -5.16012192e-01 3.34544897e-01
3.63588601e-01 6.60940185e-02 1.26937497e+00 -1.57263279e-01
4.57005829e-01 -2.81171590e-01 -1.30675554e+00 -1.18350349e-01
9.88790020e-02 -1.12042904e-01 -1.14998662e+00 7.88178071e-02
-4.12557513e-01 5.82679920e-02 -1.21720493e+00 6.09098732e-01
1.01761985e+00 -1.09146285e+00 9.51898336e-01 -4.93979782e-01
-2.75673985e-01 -2.55378455e-01 4.79307145e-01 -7.41674364e-01
-1.12121962e-01 -1.44159448e+00 -4.04362053e-01 7.38086641e-01
3.58806849e-01 -7.77503490e-01 6.30045533e-01 4.00331318e-02
2.20197335e-01 -7.59095371e-01 -9.54448402e-01 -6.16275847e-01
-4.64854985e-01 -1.03298336e-01 1.61912993e-01 1.15048897e+00
2.05032572e-01 6.86130226e-01 -2.63914794e-01 -2.43526578e-01
3.49408537e-01 -2.40688119e-02 1.82841420e-01 -1.87267637e+00
-1.17200084e-01 -2.88487881e-01 -8.67086589e-01 1.06369160e-01
-5.26737392e-01 -5.76145589e-01 -4.66856867e-01 -1.39941728e+00
-3.85127306e-01 -2.85252064e-01 -1.10221720e+00 4.73200798e-01
-6.31739125e-02 5.21436989e-01 -3.87561411e-01 2.37523913e-02
1.35574311e-01 -4.24445033e-01 3.72823387e-01 4.03041005e-01
-7.14524031e-01 7.08954632e-01 -4.67782259e-01 9.92974520e-01
9.49415982e-01 -7.36752331e-01 -5.15389502e-01 6.97993577e-01
2.04209760e-01 3.86851639e-01 1.99533418e-01 -1.52645874e+00
-3.06410432e-01 2.91284442e-01 8.31070602e-01 -7.69265175e-01
-8.66572261e-02 -1.08731306e+00 3.47229719e-01 1.16890264e+00
-8.09399560e-02 7.08456993e-01 9.33469534e-02 4.14487839e-01
-2.06600785e-01 -4.18017983e-01 9.51428235e-01 -2.39882350e-01
-6.26012757e-02 4.62193191e-02 -1.28514731e+00 -1.97773635e-01
1.20001018e+00 -8.25910747e-01 4.90782768e-01 -1.27192855e-01
-1.59007704e+00 -1.94699407e-01 -2.03808069e-01 -1.40696704e-01
8.62258375e-01 -1.00281012e+00 -5.70202887e-01 6.43654823e-01
-1.89255908e-01 -4.01017070e-01 -2.53866166e-01 1.56305516e+00
-8.72127354e-01 3.24978769e-01 -6.09162271e-01 -6.73666477e-01
-1.89083099e+00 1.46640405e-01 9.58724320e-01 -1.25696152e-01
-9.11486626e-01 5.84015310e-01 -1.07645166e+00 3.44471723e-01
1.91879392e-01 -5.37017405e-01 -1.19001043e+00 2.45412588e-01
6.94795370e-01 8.20779026e-01 2.63799816e-01 -4.79092032e-01
-5.42412698e-01 8.44176292e-01 3.72339964e-01 3.09392720e-01
1.05810606e+00 -1.40326796e-02 -4.96698469e-01 8.27834427e-01
7.40964353e-01 3.47384736e-02 1.01403790e-02 3.80013138e-01
2.16245294e-01 1.64951190e-01 1.72658414e-02 -1.06932306e+00
-8.20903957e-01 8.18881214e-01 1.30011618e+00 6.54031456e-01
1.57409608e+00 -7.12837517e-01 8.63664210e-01 6.58157766e-01
-7.33887628e-02 -1.01737130e+00 -8.62447500e-01 -1.45707890e-01
4.46549445e-01 -5.76994359e-01 4.58890319e-01 -1.26848832e-01
-3.42514157e-01 1.87708271e+00 -3.19165140e-01 -6.21056616e-01
1.48179758e+00 2.42806792e-01 5.76087356e-01 -3.13411802e-01
-6.06725156e-01 2.31380329e-01 1.08392298e-01 9.97930527e-01
4.74474579e-01 9.43453237e-02 -1.06726682e+00 6.01642787e-01
-2.22834989e-01 5.06195188e-01 5.65566838e-01 1.15298355e+00
-5.98122895e-01 -1.06741548e+00 -3.89588326e-01 9.86639798e-01
-1.20601535e+00 -2.00918868e-01 -5.32806337e-01 8.31502259e-01
5.64824045e-01 8.80786896e-01 -2.13762045e-01 -3.03603381e-01
3.05341750e-01 6.26087010e-01 4.40650612e-01 -4.80086058e-01
-1.24014747e+00 3.19520742e-01 2.44227871e-01 -2.60000437e-01
-9.04440165e-01 -8.84147704e-01 -1.30211818e+00 5.15879869e-01
-3.03771317e-01 7.61331379e-01 6.28839672e-01 1.07916415e+00
-1.59064699e-02 5.62007785e-01 7.30208099e-01 -3.51475269e-01
-1.83080256e-01 -1.21264780e+00 -1.07969415e+00 -7.08166212e-02
9.86094475e-01 -2.74503797e-01 -5.35438359e-01 5.33735454e-01] | [14.196540832519531, 3.213219404220581] |
ba3d5298-398d-4d62-9bfa-a50515be146e | deep-learning-approaches-to-lexical | 2305.12000 | null | https://arxiv.org/abs/2305.12000v1 | https://arxiv.org/pdf/2305.12000v1.pdf | Deep Learning Approaches to Lexical Simplification: A Survey | Lexical Simplification (LS) is the task of replacing complex for simpler words in a sentence whilst preserving the sentence's original meaning. LS is the lexical component of Text Simplification (TS) with the aim of making texts more accessible to various target populations. A past survey (Paetzold and Specia, 2017) has provided a detailed overview of LS. Since this survey, however, the AI/NLP community has been taken by storm by recent advances in deep learning, particularly with the introduction of large language models (LLM) and prompt learning. The high performance of these models sparked renewed interest in LS. To reflect these recent advances, we present a comprehensive survey of papers published between 2017 and 2023 on LS and its sub-tasks with a special focus on deep learning. We also present benchmark datasets for the future development of LS systems. | ['Marcos Zampieri', 'Matthew Shardlow', 'Tharindu Ranasinghe', 'Kai North'] | 2023-05-19 | null | null | null | null | ['lexical-simplification'] | ['natural-language-processing'] | [ 1.52504995e-01 2.48732910e-01 -2.92766273e-01 -2.35016897e-01
-8.17429662e-01 -2.75071055e-01 7.28328526e-01 4.54763830e-01
-6.88132644e-01 8.08339536e-01 1.03946722e+00 -1.64252385e-01
2.08675101e-01 -4.35172558e-01 -5.24283171e-01 -1.96652383e-01
2.97035635e-01 5.14289856e-01 -5.66423237e-01 -5.43498337e-01
2.36901730e-01 4.43155318e-01 -1.05963910e+00 3.28168482e-01
1.08125293e+00 2.59368867e-01 -2.18525343e-02 6.19916439e-01
-5.15406013e-01 9.13860977e-01 -8.99863243e-01 -8.81873488e-01
4.00716737e-02 -4.08786476e-01 -9.25553620e-01 -4.86032069e-01
6.49953127e-01 -1.83050752e-01 -6.27051353e-01 9.87035394e-01
8.74651492e-01 4.61660981e-01 5.38349032e-01 -7.99730062e-01
-9.08853292e-01 1.08531559e+00 -2.99805522e-01 3.28399330e-01
4.33762819e-01 3.12247872e-02 1.36408329e+00 -9.66810584e-01
7.73425460e-01 1.68117726e+00 8.81916463e-01 6.14931643e-01
-1.11938620e+00 -6.18737698e-01 3.63915324e-01 3.65621358e-01
-1.34106803e+00 -6.30053103e-01 6.42173469e-01 -1.14513464e-01
1.53266537e+00 3.96946400e-01 7.18773901e-01 1.22329080e+00
5.00988305e-01 1.22869098e+00 6.60666108e-01 -7.17619479e-01
-1.65164247e-01 -2.36786216e-01 2.45129108e-01 1.55110955e-01
5.10414183e-01 -4.26554918e-01 -5.61972618e-01 -6.92565665e-02
-5.42969555e-02 -1.23719774e-01 4.88589704e-03 1.02341250e-01
-8.49649012e-01 9.66316283e-01 1.63301751e-01 3.66369218e-01
-3.62237602e-01 -1.24129623e-01 8.40308785e-01 4.53558475e-01
7.46230185e-01 8.79127562e-01 -3.57166618e-01 -1.85726434e-01
-1.20206904e+00 8.06172311e-01 1.01222634e+00 8.02655518e-01
3.01885396e-01 3.06386203e-01 -6.18015945e-01 1.02517879e+00
-2.04531237e-01 5.93018293e-01 5.83001316e-01 -7.28342533e-01
5.84407866e-01 3.46343964e-01 -1.95242867e-01 -8.01279247e-01
-5.84250450e-01 -4.36028391e-01 -1.21148670e+00 -4.73851860e-01
-1.59765687e-02 -2.08902732e-01 -7.03774452e-01 1.79175997e+00
-1.83223575e-01 -3.38017821e-01 -5.46628460e-02 3.21632832e-01
1.04509723e+00 5.84853232e-01 4.92722481e-01 -1.87475935e-01
9.21525121e-01 -9.58978415e-01 -1.18664467e+00 -3.27887356e-01
9.15940106e-01 -9.24658239e-01 1.19848740e+00 5.12761950e-01
-1.13181651e+00 -3.12140644e-01 -8.48924160e-01 -8.04780602e-01
-5.44331253e-01 -1.38702676e-01 6.91898942e-01 5.75507045e-01
-9.45639074e-01 6.75490201e-01 -4.39251900e-01 -6.72671676e-01
7.29649127e-01 3.00958008e-01 -2.31186271e-01 -1.50621608e-01
-1.67530942e+00 1.48993623e+00 1.48023471e-01 -1.83915913e-01
-4.39962715e-01 -9.90592062e-01 -8.96260321e-01 3.18740495e-02
1.25471860e-01 -8.90700340e-01 1.37621856e+00 -1.01133978e+00
-1.50674403e+00 9.43679273e-01 -3.16834956e-01 -8.23998094e-01
4.27449048e-01 -7.72739649e-01 -4.77480561e-01 -3.22356999e-01
1.93978190e-01 5.49966633e-01 6.58960462e-01 -5.85178614e-01
-6.26068473e-01 -1.03093266e-01 9.06287953e-02 4.55371201e-01
-3.71967882e-01 4.12504524e-01 -1.89793736e-01 -1.04727447e+00
-7.09361136e-01 -6.99965537e-01 -8.79091173e-02 -6.46179199e-01
-6.01150692e-01 -6.45260751e-01 3.88156801e-01 -1.11511004e+00
1.69668162e+00 -1.92589808e+00 5.40415823e-01 -3.86669040e-01
4.87542897e-01 8.50704491e-01 -2.99097776e-01 8.86850715e-01
-1.95031181e-01 4.51736838e-01 -2.02641368e-01 -1.12942266e+00
1.74408481e-01 3.94880511e-02 -5.16373575e-01 4.30780619e-01
1.64389028e-03 1.34876430e+00 -8.60709727e-01 -2.97064662e-01
3.06155086e-01 3.93050402e-01 -6.64138198e-01 -3.27051356e-02
-3.13767605e-02 2.27332935e-01 1.15529925e-01 4.07222360e-01
5.68846226e-01 3.62286061e-01 -1.05779275e-01 2.14251816e-01
-2.48431727e-01 8.94149780e-01 -6.10484004e-01 1.59892130e+00
-4.92908001e-01 1.03692508e+00 7.33381808e-02 -7.76952803e-01
5.30702114e-01 2.27595255e-01 4.17258590e-01 -8.57208967e-01
1.85258061e-01 1.37264252e-01 6.76669106e-02 -3.14700782e-01
7.97699690e-01 -2.24069133e-01 -2.71253228e-01 4.46206748e-01
-1.81573763e-01 -5.15987694e-01 5.41828394e-01 3.93295646e-01
9.18783069e-01 -1.68877035e-01 6.10450268e-01 -2.86643535e-01
4.44302469e-01 -1.76865965e-01 4.47243661e-01 9.14441764e-01
-1.89521059e-01 4.53727514e-01 3.69997174e-01 -6.12457693e-01
-1.15311229e+00 -8.32354844e-01 -2.99997106e-02 1.27551055e+00
-6.71633601e-01 -8.35503399e-01 -8.11543882e-01 -3.72429520e-01
2.45501816e-01 1.24773955e+00 -5.82392097e-01 -4.00783658e-01
-1.00407171e+00 -6.05202794e-01 9.18629527e-01 1.15697354e-01
1.88935474e-01 -1.53246105e+00 -2.25049123e-01 1.18185528e-01
-3.41309339e-01 -8.40921700e-01 -5.51186681e-01 -1.00082206e-02
-5.86675227e-01 -7.41730630e-01 -7.75962234e-01 -7.89275229e-01
2.68001497e-01 1.44893050e-01 1.41961777e+00 -4.37468365e-02
-1.32508636e-01 4.88757528e-02 -4.67708468e-01 -9.92270291e-01
-6.47130311e-01 7.75374293e-01 2.69950658e-01 -4.34433430e-01
5.86483061e-01 -4.25268948e-01 -9.37941521e-02 -8.80102217e-01
-9.76653874e-01 8.60503390e-02 5.82772195e-01 6.80281341e-01
3.78213644e-01 -2.96425134e-01 6.97863519e-01 -1.18894911e+00
1.26212072e+00 -3.84031624e-01 -5.44920675e-02 1.16356313e-01
-6.37981176e-01 -1.14886492e-01 8.22053730e-01 -2.21932098e-01
-9.71310854e-01 -5.54599524e-01 -4.92988765e-01 1.24966810e-02
6.68573454e-02 7.78127730e-01 -1.58948883e-01 2.18651086e-01
5.06805778e-01 1.48525581e-01 -2.21725419e-01 -8.01771045e-01
5.41958690e-01 7.43393183e-01 3.75259101e-01 -3.37870359e-01
6.55083597e-01 1.28732815e-01 -9.53416526e-02 -1.14605963e+00
-1.26645648e+00 -1.51271477e-01 -6.89064741e-01 2.17494130e-01
4.13192838e-01 -8.34350169e-01 -2.23998144e-01 5.57939649e-01
-1.33729637e+00 -2.90737271e-01 -5.70131779e-01 1.21186681e-01
-1.29768103e-01 5.97387731e-01 -6.22428000e-01 -4.03175324e-01
-9.33071792e-01 -8.40774894e-01 8.01877677e-01 1.41967267e-01
-9.34587479e-01 -1.14762330e+00 2.74456382e-01 2.61918575e-01
5.41978061e-01 1.92521475e-02 1.26369095e+00 -9.18059587e-01
1.32382959e-01 -5.26179075e-01 6.52116388e-02 5.94546199e-01
1.75335810e-01 -1.13458872e-01 -7.34900534e-01 -3.17789257e-01
-2.98716947e-02 -2.49620587e-01 1.04942930e+00 6.65900171e-01
9.28874195e-01 -3.73703390e-01 -5.90268224e-02 7.70788074e-01
8.06442678e-01 -2.10401103e-01 5.54810524e-01 5.45188248e-01
9.92975712e-01 4.32549655e-01 3.06260228e-01 3.55723709e-01
6.40070438e-01 3.94117296e-01 -1.87140778e-01 -2.25556299e-01
-5.97284913e-01 -3.47895116e-01 3.96023273e-01 1.33006704e+00
2.18865409e-01 -5.13182819e-01 -8.96167099e-01 4.75679427e-01
-1.58897364e+00 -8.98938060e-01 -1.33872792e-01 1.93734848e+00
1.11886549e+00 2.73829043e-01 -6.32015988e-02 -6.11517467e-02
5.16674936e-01 5.14387846e-01 -5.06320596e-01 -1.01093471e+00
-6.80295050e-01 4.72715348e-01 3.46626133e-01 7.63573289e-01
-1.12594497e+00 1.45144069e+00 7.37815428e+00 1.03927004e+00
-9.92583454e-01 7.08585791e-03 5.51196635e-01 -3.63982290e-01
-4.61243898e-01 -1.76189750e-01 -9.29476500e-01 2.35074252e-01
1.11564088e+00 -7.39368320e-01 6.78639293e-01 4.27511543e-01
4.17340547e-01 3.41915227e-02 -1.05562401e+00 9.06612813e-01
4.41976249e-01 -1.31443882e+00 6.20712698e-01 -3.56040984e-01
9.71243322e-01 4.00430471e-01 -3.11444956e-03 8.55424047e-01
3.14318955e-01 -1.34177530e+00 6.68233037e-01 6.65923893e-01
7.49543786e-01 -9.65745270e-01 8.69875669e-01 4.75783736e-01
-7.38437533e-01 1.95435677e-02 -3.27777803e-01 -6.91946208e-01
1.31003946e-01 5.49739003e-01 -5.58880329e-01 3.65097702e-01
4.79567289e-01 1.05051363e+00 -5.74879587e-01 8.17604840e-01
-4.06244427e-01 6.52756631e-01 -6.35125637e-02 -7.83506408e-02
4.41133305e-02 -1.29874334e-01 8.11753750e-01 1.75807106e+00
-2.89543182e-01 -2.71338463e-01 2.53323633e-02 5.60253978e-01
-5.79408944e-01 4.78330314e-01 -5.09465694e-01 -4.18131858e-01
6.64981186e-01 9.76881981e-01 -1.34335354e-01 -3.83966744e-01
-4.67123091e-01 9.23952818e-01 6.42671168e-01 3.61042827e-01
-4.36408699e-01 -7.01384306e-01 7.78138161e-01 -2.20519453e-01
-1.99709237e-02 -1.40149027e-01 -8.20786238e-01 -1.12840128e+00
-1.96302190e-01 -1.24435246e+00 2.40986794e-01 -3.30955237e-01
-1.30243039e+00 3.15763116e-01 1.75524708e-02 -5.11008799e-01
-6.10421970e-03 -2.14004532e-01 -4.88563627e-01 1.18963540e+00
-1.56210434e+00 -1.22234082e+00 2.58045107e-01 1.79273054e-01
9.33340490e-01 -3.34889591e-01 1.02423108e+00 4.06856328e-01
-7.79561698e-01 9.13408816e-01 4.29134816e-01 9.88675561e-03
9.99142349e-01 -1.15742040e+00 1.07300174e+00 8.19030225e-01
-1.77725762e-01 8.27910662e-01 8.93447101e-01 -9.38407958e-01
-1.28437471e+00 -1.21080315e+00 1.94652343e+00 -7.37373948e-01
5.77780068e-01 -6.23522103e-01 -7.03100204e-01 9.79241371e-01
5.02816498e-01 -8.21334481e-01 6.04696453e-01 2.49507576e-01
-1.02330074e-01 -1.15003966e-01 -8.24204922e-01 1.09890199e+00
8.81652236e-01 -6.42134547e-01 -9.72111762e-01 4.03696299e-01
9.15957928e-01 -4.34467196e-01 -5.54578006e-01 2.62956917e-01
4.15081024e-01 -5.04052103e-01 8.07906210e-01 -8.10372174e-01
4.23678160e-01 4.34531391e-01 1.29099101e-01 -1.52902293e+00
-7.53114283e-01 -1.03078127e+00 -1.77357927e-01 1.22389770e+00
9.01646763e-02 -2.96505958e-01 3.79645467e-01 5.70194304e-01
-5.07102668e-01 -7.59430885e-01 -9.69489872e-01 -4.26289976e-01
8.28211367e-01 -5.34826815e-01 5.95607519e-01 7.71456420e-01
-1.27020076e-01 6.92439020e-01 -5.04037917e-01 -5.97671330e-01
2.44454399e-01 -3.19545269e-01 7.74701178e-01 -1.24328709e+00
3.18355113e-01 -1.00175250e+00 9.81282964e-02 -8.81792784e-01
5.83023012e-01 -1.29506004e+00 -3.44598532e-01 -1.76310480e+00
3.10191780e-01 2.68404126e-01 -1.40647039e-01 2.50907779e-01
-4.24885273e-01 1.00072399e-01 3.38750988e-01 1.35961816e-01
-6.98895991e-01 7.96615660e-01 1.18530953e+00 -2.97086477e-01
-2.02417701e-01 -4.90560792e-02 -1.10471880e+00 6.97668433e-01
8.27648222e-01 -5.17860651e-01 -5.34521043e-03 -7.69781291e-01
4.43921506e-01 -6.58513010e-01 -2.34881386e-01 -5.72796345e-01
5.75918220e-02 -4.54091877e-02 2.19146460e-01 -6.03089631e-01
9.28815529e-02 -2.32245132e-01 -2.09620818e-01 6.45700395e-01
-6.38798892e-01 2.23845601e-01 4.52503324e-01 5.47376759e-02
-8.19412991e-02 -1.19231455e-01 8.28570843e-01 -1.09667704e-01
-5.01979947e-01 1.66656867e-01 -7.35958040e-01 4.33203429e-01
4.31295812e-01 2.33644336e-01 -1.81577783e-02 -5.45874715e-01
-4.18556571e-01 3.35768461e-01 2.15590790e-01 6.83737874e-01
3.38918239e-01 -1.13051152e+00 -1.11781228e+00 1.15300946e-01
-2.02885777e-01 -5.49717396e-02 1.51795328e-01 7.55901277e-01
-5.25669515e-01 1.02737808e+00 7.66838435e-03 1.64522588e-01
-1.08082807e+00 3.47833425e-01 2.20771849e-01 -5.89282572e-01
-8.06251585e-01 8.83598268e-01 5.79736978e-02 -5.41055381e-01
4.85690504e-01 -3.22288305e-01 -4.01640803e-01 1.56114385e-01
6.86125100e-01 6.66068614e-01 2.75289744e-01 -7.00759709e-01
-3.25516731e-01 2.20065303e-02 -4.98224020e-01 1.99426219e-01
1.47980583e+00 -2.90517777e-01 -4.55847383e-01 6.16357207e-01
1.05212843e+00 2.77516544e-01 -5.22109687e-01 -4.79317755e-01
1.76614165e-01 -5.09653576e-02 2.27899611e-01 -9.06852484e-01
-5.47268808e-01 8.21844101e-01 -4.89659309e-02 -1.01843022e-01
8.82472932e-01 -2.29088575e-01 1.14929128e+00 5.78052282e-01
-1.24769330e-01 -1.45132995e+00 -2.94140071e-01 1.38873076e+00
1.22551107e+00 -9.92380083e-01 3.27223688e-01 -2.18361197e-03
-6.59962475e-01 9.56282020e-01 2.68841416e-01 -1.49291351e-01
4.25532192e-01 1.23192713e-01 -7.96603933e-02 3.36493254e-02
-6.43459380e-01 -5.84308570e-03 3.02882314e-01 6.03944600e-01
7.47697830e-01 2.50392050e-01 -5.98082781e-01 9.57830489e-01
-6.79379165e-01 -6.27633333e-02 3.74520481e-01 6.51408553e-01
-3.83015871e-01 -1.24538386e+00 -1.34456411e-01 7.96013117e-01
-6.05773866e-01 -7.44443893e-01 -8.22622180e-01 8.88604879e-01
4.18965630e-02 8.84183049e-01 -1.28127411e-01 -1.71222195e-01
5.50125122e-01 2.72281349e-01 3.99450272e-01 -9.01354492e-01
-8.50479543e-01 -1.37392730e-01 2.43135825e-01 -2.50561118e-01
7.91777372e-02 -9.54871893e-01 -1.01471281e+00 -9.69440162e-01
1.77363083e-01 -1.42466668e-02 4.15350169e-01 1.15625000e+00
3.62938464e-01 7.03137040e-01 1.79496884e-01 -8.79146159e-01
-6.38655901e-01 -1.34231019e+00 -5.86838365e-01 3.98515284e-01
5.33110678e-01 -2.48263747e-01 -2.52179503e-01 -1.28124475e-01] | [11.06657886505127, 10.324089050292969] |
60ee634b-5e2f-474e-8898-f861cda5d979 | clickbait-sensational-headline-generation | 1909.03582 | null | https://arxiv.org/abs/1909.03582v1 | https://arxiv.org/pdf/1909.03582v1.pdf | Clickbait? Sensational Headline Generation with Auto-tuned Reinforcement Learning | Sensational headlines are headlines that capture people's attention and generate reader interest. Conventional abstractive headline generation methods, unlike human writers, do not optimize for maximal reader attention. In this paper, we propose a model that generates sensational headlines without labeled data. We first train a sensationalism scorer by classifying online headlines with many comments ("clickbait") against a baseline of headlines generated from a summarization model. The score from the sensationalism scorer is used as the reward for a reinforcement learner. However, maximizing the noisy sensationalism reward will generate unnatural phrases instead of sensational headlines. To effectively leverage this noisy reward, we propose a novel loss function, Auto-tuned Reinforcement Learning (ARL), to dynamically balance reinforcement learning (RL) with maximum likelihood estimation (MLE). Human evaluation shows that 60.8% of samples generated by our model are sensational, which is significantly better than the Pointer-Gen baseline and other RL models. | ['Chien-Sheng Wu', 'Pascale Fung', 'Peng Xu', 'Andrea Madotto'] | 2019-09-09 | clickbait-sensational-headline-generation-1 | https://aclanthology.org/D19-1303 | https://aclanthology.org/D19-1303.pdf | ijcnlp-2019-11 | ['headline-generation'] | ['natural-language-processing'] | [ 9.68793109e-02 5.00566363e-01 -2.67295152e-01 -2.92888612e-01
-1.55648959e+00 -5.00937700e-01 6.93715811e-01 1.83234408e-01
-4.46530670e-01 1.05075359e+00 9.89169300e-01 -1.18529215e-01
4.96369898e-01 -6.22654498e-01 -7.52994001e-01 -2.90948778e-01
3.86317343e-01 2.86081254e-01 -6.43653423e-02 -3.49060327e-01
5.22384822e-01 -1.93325058e-01 -1.13935173e+00 4.16352898e-01
1.14269423e+00 6.56981647e-01 3.88734967e-01 1.34792554e+00
-2.41582185e-01 1.24059153e+00 -1.29424310e+00 -5.79356015e-01
-1.60855934e-01 -8.10828090e-01 -6.42191112e-01 -2.00427577e-01
6.98595941e-01 -5.34268916e-01 -1.81865856e-01 7.69702554e-01
8.54666173e-01 3.08108836e-01 8.45479667e-01 -1.13809681e+00
-1.09556460e+00 9.76796806e-01 -6.94735587e-01 1.58031866e-01
7.76759386e-01 3.99816751e-01 1.60651338e+00 -5.62747598e-01
6.29645824e-01 1.42738593e+00 3.83343786e-01 9.38760340e-01
-1.14610922e+00 -5.26843488e-01 2.53396034e-01 -7.42072985e-02
-6.56306684e-01 -8.36649388e-02 6.56175077e-01 -3.38667154e-01
7.30920911e-01 5.72221994e-01 4.84519809e-01 1.65482271e+00
1.10373303e-01 1.51577306e+00 9.12277758e-01 -2.59487361e-01
3.62498015e-01 3.90671045e-01 -1.23878427e-01 1.45836130e-01
-1.32823080e-01 -3.35332192e-02 -8.27830434e-01 -1.46588117e-01
5.29579699e-01 -3.27069670e-01 -2.42278472e-01 3.29957843e-01
-1.07316101e+00 1.21317649e+00 5.38753629e-01 -3.12240601e-01
-6.01897836e-01 2.79455155e-01 3.95523071e-01 2.42567342e-02
6.30547941e-01 1.07423806e+00 -7.18570128e-02 -4.83867437e-01
-1.07983673e+00 7.18776822e-01 9.84603524e-01 1.06298494e+00
3.10086638e-01 6.83399513e-02 -1.09229040e+00 1.06515694e+00
1.07947886e-01 1.06298780e+00 5.64379632e-01 -8.13171387e-01
5.88299572e-01 3.13746333e-01 5.54574490e-01 -7.24663854e-01
-1.47641286e-01 -7.11469650e-01 -5.95634878e-01 2.35252585e-02
2.68023670e-01 -7.67467618e-01 -4.59738344e-01 1.64686406e+00
-2.32341304e-01 -3.50797445e-01 6.58808500e-02 9.78003085e-01
9.00188088e-01 1.15916729e+00 6.50220811e-02 -6.82542995e-02
1.08362079e+00 -1.29612267e+00 -8.69011760e-01 -3.05299163e-01
3.83041352e-01 -9.58620191e-01 1.79956079e+00 4.52323139e-01
-1.56476676e+00 -4.77571338e-01 -9.72207129e-01 -7.33426660e-02
8.90132263e-02 1.00674890e-01 4.58873138e-02 2.57368684e-01
-9.46217537e-01 4.16270256e-01 -1.50341168e-01 -6.08674325e-02
2.63867855e-01 -2.61203855e-01 3.07409704e-01 4.48086679e-01
-1.39612317e+00 6.66699409e-01 -1.87298864e-01 -5.74526191e-01
-6.85669720e-01 -7.98450232e-01 -7.02129304e-01 4.14277762e-02
2.45772451e-01 -7.79520452e-01 1.95964408e+00 -9.14098442e-01
-1.59624350e+00 4.96921569e-01 -1.37216389e-01 -6.84369087e-01
9.57856834e-01 -7.45352864e-01 -1.06544256e-01 1.74127206e-01
3.57568234e-01 9.27425385e-01 9.51242387e-01 -1.40094125e+00
-7.55863309e-01 4.40017194e-01 1.80444140e-02 5.34663737e-01
-1.82513207e-01 -3.93902622e-02 -2.03448460e-01 -8.56983662e-01
-8.31329763e-01 -5.73718548e-01 -1.95309907e-01 -4.44906682e-01
-1.02175605e+00 -6.14793241e-01 1.91774547e-01 -8.17258716e-01
1.49970531e+00 -1.86219883e+00 -1.87706918e-01 -1.14722000e-02
1.61863595e-01 4.72772904e-02 -3.14995259e-01 5.39011300e-01
4.09522384e-01 2.57914394e-01 2.51315385e-01 -4.64208603e-01
3.11072499e-01 -3.57624501e-01 -6.58479333e-01 -2.52899110e-01
3.53095144e-01 1.12888396e+00 -1.30359721e+00 -3.83471996e-01
-5.88232353e-02 6.54726326e-02 -8.58589292e-01 7.35539913e-01
-5.95237553e-01 1.54028133e-01 -4.24221128e-01 1.14887252e-01
3.20986003e-01 -4.83792841e-01 -5.30962288e-01 1.17370062e-01
-9.90960002e-02 5.03656328e-01 -5.71951807e-01 1.16700089e+00
-8.33111942e-01 7.18691289e-01 -4.09549206e-01 7.95491785e-02
1.11908758e+00 1.80345103e-02 -1.15276389e-02 -1.00906885e+00
-1.19529761e-01 1.59330592e-02 -3.55519921e-01 -4.28250998e-01
1.14487004e+00 -6.14659861e-02 -4.80892271e-01 8.17912102e-01
-1.86499849e-01 -2.18734235e-01 1.38209105e-01 8.58512461e-01
1.11411297e+00 -1.42133478e-02 -5.56840003e-02 1.52194023e-01
1.58526510e-01 -2.40919665e-01 9.61791947e-02 1.48007309e+00
-1.64435908e-01 9.04418886e-01 8.59347045e-01 1.59999773e-01
-1.31888771e+00 -1.36072814e+00 4.00142580e-01 1.43251526e+00
-4.26586792e-02 -5.29000223e-01 -1.06138802e+00 -6.62788332e-01
-1.30307734e-01 1.58629382e+00 -4.77563709e-01 -3.82501692e-01
-3.25888306e-01 -3.79590422e-01 5.63430071e-01 4.84694183e-01
2.23151654e-01 -1.51687956e+00 -7.35533357e-01 2.41682097e-01
-5.32568991e-01 -6.58747911e-01 -1.07390976e+00 -9.64300632e-02
-1.88585415e-01 -4.35855806e-01 -1.29614151e+00 -4.03166205e-01
5.20583868e-01 -1.06067114e-01 1.41125250e+00 -3.44818771e-01
-1.37569569e-02 2.22592995e-01 -5.56179225e-01 -6.23651266e-01
-8.03350925e-01 2.82090962e-01 -1.90691173e-01 -2.26406962e-01
2.74904072e-01 8.14704522e-02 -9.13829923e-01 -1.27462432e-01
-7.27581322e-01 3.20902377e-01 7.06414759e-01 1.09378231e+00
2.06660762e-01 -5.91884434e-01 1.18005514e+00 -9.59353745e-01
1.49375737e+00 -5.38325906e-01 -1.18251160e-01 7.96671361e-02
-4.08367455e-01 1.73968107e-01 9.90138173e-01 -4.40372676e-01
-1.28274810e+00 -5.14484525e-01 -3.72750014e-01 -2.21372284e-02
5.20719402e-03 2.41297156e-01 2.19249547e-01 8.27541649e-01
1.09038973e+00 2.38822609e-01 -4.34028096e-02 -2.38376930e-01
6.49471581e-01 8.79171133e-01 5.73496878e-01 -3.91610503e-01
5.46577930e-01 -1.59619123e-01 -6.97995067e-01 -5.89156449e-01
-1.36909270e+00 -6.01444244e-01 2.34290451e-01 -4.53754187e-01
6.99136615e-01 -7.84496129e-01 -8.16929996e-01 3.17028642e-01
-1.26710093e+00 -5.45483530e-01 -6.07322276e-01 1.44886687e-01
-8.50027323e-01 1.44899830e-01 -7.26845026e-01 -1.14372623e+00
-7.70068109e-01 -7.09131420e-01 1.18609130e+00 6.23483956e-01
-8.40304375e-01 -9.31124330e-01 1.62839904e-01 4.64736789e-01
6.53673470e-01 1.28331244e-01 5.35501480e-01 -7.89927363e-01
-2.72836953e-01 -3.79833758e-01 -2.82451898e-01 4.99039143e-01
-3.89082767e-02 1.15603646e-02 -9.77712214e-01 -1.38034355e-02
-4.06032801e-01 -9.18019354e-01 8.24597061e-01 4.55174714e-01
1.08512318e+00 -1.11942291e+00 2.41325915e-01 1.15110435e-01
1.03574240e+00 -1.29095167e-01 5.75025916e-01 2.03912839e-01
5.31439722e-01 5.60790122e-01 7.09819317e-01 1.01588106e+00
4.56117541e-01 3.81735831e-01 1.74917251e-01 -3.46367210e-01
-1.43491745e-01 -1.15759408e+00 6.24487996e-01 7.29564309e-01
3.64256740e-01 -7.32549787e-01 -3.67865741e-01 3.47444892e-01
-1.80818784e+00 -1.20533240e+00 4.45668288e-02 2.05676532e+00
1.27961457e+00 5.65098941e-01 5.59699357e-01 -2.82236040e-01
6.10802293e-01 3.14746141e-01 -7.22074211e-01 -8.19559574e-01
-1.16204597e-01 -1.64217707e-02 5.00489831e-01 8.04984033e-01
-7.70632923e-01 9.36595440e-01 6.25040197e+00 9.51825082e-01
-9.44721878e-01 -3.16392303e-01 7.64791727e-01 -3.18868130e-01
-7.60079980e-01 -4.65207905e-01 -9.30444479e-01 7.45171607e-01
1.08431244e+00 -5.60536683e-01 1.88546211e-01 1.11255383e+00
8.34681034e-01 -8.58198032e-02 -9.08094108e-01 7.72346318e-01
2.63695806e-01 -1.24070716e+00 1.76598176e-01 -1.79076403e-01
9.93867695e-01 -3.25893641e-01 4.33539629e-01 6.00254297e-01
8.96363020e-01 -9.88045335e-01 9.02336061e-01 8.29432309e-01
5.79021931e-01 -8.29915524e-01 7.02194870e-01 4.59710032e-01
-2.37720385e-01 1.70797005e-01 -2.56252974e-01 5.55167571e-02
6.76224172e-01 5.52115262e-01 -1.34696209e+00 -8.18878263e-02
1.08978786e-01 4.84879524e-01 -6.34708226e-01 1.05427504e+00
-4.67993557e-01 8.98697674e-01 6.74808249e-02 -9.99107480e-01
4.82527822e-01 3.03035766e-01 7.04691052e-01 1.60563409e+00
1.27210304e-01 -3.53393495e-01 1.57890067e-01 1.22872734e+00
-3.24638307e-01 4.12316442e-01 -2.26828694e-01 -6.63858131e-02
5.49714983e-01 1.38990307e+00 -2.01449364e-01 -5.14277399e-01
-3.11166681e-02 1.23013687e+00 2.10327163e-01 3.44761670e-01
-9.42709148e-01 -6.25042915e-01 1.11533388e-01 1.24708623e-01
-8.09841007e-02 3.83420080e-01 -4.54731822e-01 -8.91354442e-01
-2.38106996e-01 -9.78165925e-01 5.65211661e-02 -1.12554204e+00
-1.73403263e+00 4.89692807e-01 -2.47395486e-01 -1.11720788e+00
-5.75218737e-01 -1.40626624e-01 -1.04221320e+00 9.88767147e-01
-1.42450058e+00 -7.94554710e-01 -2.32664526e-01 1.58522606e-01
1.03968561e+00 -4.28796597e-02 3.92369568e-01 -1.90462470e-01
-3.12383175e-01 1.08860183e+00 -7.48532563e-02 -1.28190413e-01
1.00367653e+00 -1.99098051e+00 7.74341941e-01 4.45502698e-01
-1.97139427e-01 5.61361670e-01 1.25718784e+00 -6.45530164e-01
-7.12515056e-01 -1.16086662e+00 1.03349543e+00 -3.92506212e-01
6.54156327e-01 -3.24810594e-01 -6.50921524e-01 1.78754091e-01
7.31732190e-01 -8.85094941e-01 8.48235548e-01 7.53489807e-02
-2.78717011e-01 1.95177034e-01 -8.84979367e-01 1.06692219e+00
3.65494281e-01 -2.26706028e-01 -5.87343633e-01 6.60723507e-01
9.81829643e-01 -2.26984173e-01 -4.61961269e-01 -2.79991359e-01
3.35752994e-01 -8.83794844e-01 6.68789446e-01 -7.32595444e-01
1.15095723e+00 1.27384886e-01 2.76520550e-01 -1.82088959e+00
-3.64936709e-01 -9.49097693e-01 -1.64588496e-01 1.29980874e+00
7.68477321e-01 -8.68835300e-02 7.51138508e-01 5.04254818e-01
-2.73922890e-01 -7.09968805e-01 -2.45022088e-01 -6.62401259e-01
3.19605768e-01 -8.94975066e-02 2.90777057e-01 3.46364230e-01
4.24901485e-01 9.61944997e-01 -8.36640775e-01 -5.47237813e-01
6.92888558e-01 -2.46402770e-01 9.58528876e-01 -6.18410885e-01
-6.10704660e-01 -7.72304833e-01 4.17734683e-01 -1.55590665e+00
-7.96362758e-02 -7.22278774e-01 6.32695675e-01 -1.75236344e+00
3.92190307e-01 -1.33509591e-01 -1.00641005e-01 1.38939336e-01
-6.29880905e-01 -4.87452373e-02 3.73935103e-01 1.02063240e-02
-1.04642916e+00 7.28619814e-01 1.63170421e+00 4.96259984e-03
-3.45159322e-01 2.92501748e-01 -1.19079828e+00 4.80752140e-01
9.17042077e-01 -3.51852179e-02 -4.45333779e-01 1.52357921e-01
6.42187119e-01 1.43053040e-01 2.25364700e-01 -7.82717168e-01
1.41210347e-01 -2.68442184e-01 3.44648838e-01 -9.08324003e-01
1.33095339e-01 1.37947083e-01 -6.46585584e-01 2.51043379e-01
-1.47319114e+00 2.94887908e-02 -2.79747605e-01 6.47987127e-01
5.85231557e-02 -4.45585549e-01 7.03438699e-01 -2.31547490e-01
-8.93247724e-02 5.29665174e-03 -6.22841120e-01 7.42767870e-01
5.80071926e-01 1.46868974e-01 -6.23049498e-01 -1.29958737e+00
-2.82534957e-01 4.61619258e-01 2.99724072e-01 6.54627085e-01
7.71474898e-01 -1.24970078e+00 -1.05004776e+00 -1.53494179e-01
1.37221262e-01 -3.51642817e-02 2.23672822e-01 2.70986885e-01
-4.73036796e-01 2.38869667e-01 9.86788794e-02 -2.13205248e-01
-8.40810239e-01 3.09633106e-01 -9.30292755e-02 -5.51054418e-01
-5.21027505e-01 1.06916082e+00 2.37616509e-01 -2.01038599e-01
4.54700023e-01 -1.67376682e-01 -2.29283571e-01 1.00843437e-01
8.07001889e-01 6.05697215e-01 -3.60892624e-01 -2.41721883e-01
2.72730619e-01 -2.00437993e-01 -6.03255510e-01 -5.50770640e-01
7.89246142e-01 -1.81787610e-01 6.00922346e-01 5.50732315e-01
1.14473140e+00 3.14399213e-01 -1.58140051e+00 -2.15999726e-02
-5.05870394e-02 -3.24365318e-01 -2.02750102e-01 -1.29751396e+00
-4.27800089e-01 8.00803900e-01 -5.42753153e-02 5.75043917e-01
6.06161058e-01 1.11797050e-01 1.32267082e+00 1.13523088e-01
-2.40989923e-01 -1.55765045e+00 1.10040665e+00 5.42289555e-01
1.32195663e+00 -1.06868136e+00 -3.58501315e-01 1.15486108e-01
-1.53409529e+00 8.04125547e-01 9.35441315e-01 -3.64449233e-01
-6.96374476e-02 2.31569245e-01 3.08695704e-01 1.77254856e-01
-1.15165961e+00 -1.76953226e-02 3.36367160e-01 5.20066977e-01
5.40021539e-01 1.54255077e-01 -3.26565355e-01 7.39524901e-01
-9.33571696e-01 -1.51374683e-01 9.90938663e-01 4.55906749e-01
-8.11387062e-01 -6.12213731e-01 -2.76376575e-01 7.43286490e-01
-5.50632477e-01 -2.77698606e-01 -5.12204647e-01 1.43923745e-01
-5.52973688e-01 1.13342893e+00 5.01725338e-02 -2.71080881e-01
3.78273070e-01 -1.63920492e-01 4.71041799e-02 -6.43439174e-01
-7.67266810e-01 2.53874302e-01 2.41337225e-01 -7.62231275e-02
2.52717316e-01 -4.94730890e-01 -1.19609702e+00 -3.03731233e-01
-2.95223683e-01 1.93743333e-01 5.43519437e-01 4.86042887e-01
1.57721639e-01 7.03159750e-01 1.09661376e+00 -5.58642685e-01
-1.20035958e+00 -1.07504284e+00 -4.60035890e-01 9.11389887e-01
5.65286458e-01 -1.63660776e-02 -5.90916812e-01 -5.91778383e-02] | [12.192848205566406, 9.121403694152832] |
30117710-8d02-4734-b0a7-95426eb5378f | efficient-rgb-d-semantic-segmentation-for | 2011.06961 | null | https://arxiv.org/abs/2011.06961v3 | https://arxiv.org/pdf/2011.06961v3.pdf | Efficient RGB-D Semantic Segmentation for Indoor Scene Analysis | Analyzing scenes thoroughly is crucial for mobile robots acting in different environments. Semantic segmentation can enhance various subsequent tasks, such as (semantically assisted) person perception, (semantic) free space detection, (semantic) mapping, and (semantic) navigation. In this paper, we propose an efficient and robust RGB-D segmentation approach that can be optimized to a high degree using NVIDIA TensorRT and, thus, is well suited as a common initial processing step in a complex system for scene analysis on mobile robots. We show that RGB-D segmentation is superior to processing RGB images solely and that it can still be performed in real time if the network architecture is carefully designed. We evaluate our proposed Efficient Scene Analysis Network (ESANet) on the common indoor datasets NYUv2 and SUNRGB-D and show that we reach state-of-the-art performance while enabling faster inference. Furthermore, our evaluation on the outdoor dataset Cityscapes shows that our approach is suitable for other areas of application as well. Finally, instead of presenting benchmark results only, we also show qualitative results in one of our indoor application scenarios. | ['Horst-Michael Gross', 'Tim Wengefeld', 'Benjamin Lewandowski', 'Mona Köhler', 'Daniel Seichter'] | 2020-11-13 | null | null | null | null | ['thermal-image-segmentation'] | ['computer-vision'] | [ 2.19575092e-01 1.19833559e-01 4.91077960e-01 -5.14974535e-01
-2.67343462e-01 -5.39687455e-01 5.75603127e-01 1.55310392e-01
-9.25826609e-01 4.55658734e-01 -2.24854991e-01 -4.54618305e-01
-1.24164373e-01 -8.58792603e-01 -7.54275560e-01 -4.45386916e-01
-4.68068868e-02 7.75562942e-01 5.92345774e-01 -4.95855540e-01
-6.66045919e-02 7.23040164e-01 -2.00689101e+00 -1.18062332e-01
7.20927835e-01 1.08910358e+00 5.55834770e-01 8.64363134e-01
1.05849393e-01 7.27691770e-01 -3.65668923e-01 -1.21117100e-01
4.86081898e-01 1.29660904e-01 -1.09769964e+00 1.39218748e-01
3.99127901e-01 -1.34943500e-01 -1.60448670e-01 1.05417800e+00
5.17779887e-01 6.12750411e-01 1.99621111e-01 -1.27793515e+00
1.79764658e-01 2.02646464e-01 -1.14134170e-01 -1.01550659e-02
7.00590432e-01 2.22003222e-01 5.00638723e-01 -6.08300567e-01
6.30604267e-01 1.46020508e+00 8.33509862e-01 3.36375743e-01
-7.49156058e-01 3.43296938e-02 3.35206032e-01 3.12521368e-01
-1.31659532e+00 -3.90063941e-01 5.04791617e-01 -8.49753171e-02
1.21775734e+00 3.60388428e-01 8.05207908e-01 9.80726421e-01
-3.45566541e-01 9.96326089e-01 1.14105546e+00 -2.22149298e-01
5.31688750e-01 6.14681952e-02 2.32463330e-01 8.70593786e-01
1.61857843e-01 -3.01605225e-01 -1.98373377e-01 3.83524269e-01
5.75044751e-01 3.07631642e-02 -1.18282564e-01 -6.18661821e-01
-1.39253962e+00 5.62042832e-01 9.89489317e-01 1.18766926e-01
-4.66988474e-01 5.49331129e-01 2.98665881e-01 -4.64256853e-02
1.83693483e-01 1.50188848e-01 -3.92930686e-01 -3.00550342e-01
-6.35291457e-01 3.71634752e-01 8.32583129e-01 1.06926286e+00
8.46333981e-01 -2.25078180e-01 2.20401600e-01 6.45110130e-01
4.57448393e-01 8.50296855e-01 1.87239736e-01 -1.41954446e+00
4.28283870e-01 5.98407984e-01 4.22918111e-01 -9.83215630e-01
-1.04781544e+00 -1.26698971e-01 -7.62971699e-01 3.71372193e-01
6.98345363e-01 -3.03240009e-02 -1.16870046e+00 1.29790890e+00
4.83315706e-01 -8.69250596e-02 1.87898934e-01 1.13770938e+00
9.92792308e-01 3.40712726e-01 8.08096007e-02 5.09951890e-01
1.48664033e+00 -1.17085683e+00 -3.99890125e-01 -7.19680727e-01
6.44595444e-01 -3.12828332e-01 1.17642415e+00 6.11840606e-01
-7.73001075e-01 -5.99634349e-01 -1.06860673e+00 -4.60203707e-01
-8.92785728e-01 1.79480225e-01 1.06865537e+00 9.98204470e-01
-1.43742955e+00 6.35534644e-01 -1.26842046e+00 -1.17624009e+00
4.26361352e-01 5.75334132e-01 -4.71584529e-01 -2.15878472e-01
-8.46493483e-01 8.57684672e-01 4.06309426e-01 3.13592821e-01
-7.60459721e-01 -2.07140818e-02 -9.45365310e-01 -3.08979392e-01
4.29570287e-01 -8.54596496e-01 1.04229605e+00 -4.90005046e-01
-1.43153453e+00 9.52060819e-01 -6.95112720e-02 -7.18125820e-01
7.70028532e-01 -3.14918339e-01 -8.55643451e-02 4.63701874e-01
1.73536897e-01 1.16954994e+00 2.69015610e-01 -1.23563325e+00
-8.58217299e-01 -7.57717729e-01 5.10653734e-01 6.29719079e-01
-4.98613007e-02 -4.86864328e-01 -9.04277861e-01 1.56505331e-01
3.79704773e-01 -1.22945583e+00 -8.34916830e-01 9.27808285e-02
-6.60822570e-01 1.12998355e-02 6.65745139e-01 -6.17837846e-01
2.28022352e-01 -1.89362562e+00 6.97036907e-02 5.11090159e-01
-1.51689611e-02 2.50927061e-02 1.32118151e-01 -2.00627707e-02
4.99606580e-01 -1.68911144e-01 -4.67536002e-01 -9.10411954e-01
1.94895253e-01 3.67060632e-01 2.46520847e-01 6.88597977e-01
-3.09247583e-01 8.39780986e-01 -9.67987597e-01 -4.61701393e-01
9.59384739e-01 7.34003127e-01 -4.78057504e-01 -1.71455309e-01
-9.14746970e-02 5.81539392e-01 -3.98147106e-01 6.37966216e-01
7.03781843e-01 -1.62760079e-01 7.97606036e-02 1.20565444e-02
-1.74722746e-01 1.07678436e-01 -1.41865802e+00 2.19142342e+00
-5.58188558e-01 6.73473001e-01 1.18324324e-01 -8.67865622e-01
6.90378010e-01 -3.60602170e-01 4.14280474e-01 -1.04325128e+00
1.82024673e-01 2.47201964e-01 -5.00161111e-01 -4.08297509e-01
9.47993100e-01 5.02385437e-01 -2.24927232e-01 4.69435751e-02
-2.46710420e-01 -3.28130871e-01 2.26139098e-01 -1.89048890e-02
1.12045145e+00 4.29281324e-01 1.58400565e-01 -4.12789255e-01
5.36589682e-01 5.17573357e-01 -5.83385974e-02 9.49908614e-01
-4.40234661e-01 4.89895374e-01 -4.53709327e-02 -3.82271409e-01
-9.40407157e-01 -1.12913609e+00 1.24199009e-02 8.56392086e-01
8.54632735e-01 -2.12656453e-01 -9.56571341e-01 -5.11908829e-01
-1.73669025e-01 6.47963405e-01 -3.44828844e-01 4.09352064e-01
-5.70555151e-01 -7.40944982e-01 5.67843497e-01 6.44757569e-01
1.13433897e+00 -1.02997935e+00 -1.34063053e+00 -9.31076054e-03
-3.30614418e-01 -1.71068394e+00 4.91915971e-01 3.98572803e-01
-7.28529751e-01 -1.06545496e+00 -5.76655269e-01 -7.68930376e-01
5.21377146e-01 8.77394259e-01 9.58953321e-01 -7.71509558e-02
-1.74114451e-01 7.30359495e-01 -3.37718695e-01 -1.02967717e-01
5.60234785e-02 2.97882259e-01 1.60242632e-01 -5.08676231e-01
2.80098855e-01 -3.90112042e-01 -8.17184627e-01 3.58910531e-01
-5.06810129e-01 9.97774489e-03 1.94550291e-01 1.62829846e-01
6.04898393e-01 3.17244411e-01 -1.44328147e-01 -7.43689775e-01
9.94620025e-02 -1.76186144e-01 -7.06118762e-01 -6.69794008e-02
-2.52163231e-01 -1.07765250e-01 5.29131413e-01 2.69831240e-01
-1.12302911e+00 4.71922308e-01 -4.71973658e-01 7.10924938e-02
-6.86912060e-01 6.78670779e-02 -2.86697328e-01 -1.69770807e-01
7.36525416e-01 -8.92193243e-02 -1.83855072e-01 -3.98622215e-01
6.04307711e-01 7.37408936e-01 7.01845825e-01 -1.58277631e-01
5.37252426e-01 1.20984638e+00 2.10474849e-01 -1.29811025e+00
-3.57578725e-01 -8.39527607e-01 -8.76614034e-01 -3.29926997e-01
1.14095688e+00 -1.14096737e+00 -1.20590687e+00 6.21256053e-01
-1.05715156e+00 -8.80871475e-01 -6.76789433e-02 3.27222466e-01
-6.75146937e-01 3.77570421e-01 -3.46381664e-01 -9.73401785e-01
5.74064394e-03 -1.25792038e+00 1.46639955e+00 2.97519654e-01
9.48614255e-03 -1.14419556e+00 -2.08105341e-01 4.83826041e-01
2.12322697e-01 3.15074205e-01 2.55767822e-01 -2.40205005e-01
-8.18505883e-01 -4.30098735e-02 -4.85296935e-01 -3.14918645e-02
-1.23093866e-01 -4.73076701e-01 -1.32906902e+00 -9.60330199e-03
-3.30316007e-01 -1.20139003e-01 9.59677517e-01 3.06731313e-01
1.00620890e+00 1.77707657e-01 -5.51266789e-01 7.26465702e-01
1.45920038e+00 -5.19698448e-02 8.36368918e-01 8.97082150e-01
1.07922781e+00 7.38559067e-01 8.35450530e-01 2.67120659e-01
8.17602217e-01 8.60166967e-01 8.61830056e-01 -3.85613352e-01
-1.73465818e-01 1.27159134e-01 2.39630818e-01 1.59477159e-01
-3.06019694e-01 -3.59297752e-01 -1.12327611e+00 4.23770040e-01
-2.05182219e+00 -6.83724284e-01 -6.37692988e-01 2.04322672e+00
1.87321175e-02 2.15989009e-01 2.60193527e-01 3.07134658e-01
4.38452065e-01 2.15759091e-02 -3.54025424e-01 -2.27621868e-01
-2.58330464e-01 4.82328199e-02 9.89228427e-01 6.08435094e-01
-1.47150004e+00 1.25478506e+00 5.68333054e+00 4.83062595e-01
-8.17635715e-01 1.67373106e-01 3.56028557e-01 1.13066256e-01
8.92954692e-02 -3.64004582e-01 -6.43570781e-01 1.19478881e-01
7.90839255e-01 7.26437330e-01 5.27008116e-01 1.26125169e+00
3.15163940e-01 -6.40186250e-01 -8.12251568e-01 1.27063823e+00
-8.07247609e-02 -9.93687451e-01 -4.92718577e-01 -1.33354999e-02
3.35268676e-01 4.42674488e-01 -1.20248280e-01 1.40363932e-01
2.85464823e-01 -9.30189669e-01 9.80927050e-01 3.15313220e-01
2.84695774e-01 -7.08238065e-01 8.72087300e-01 3.27885926e-01
-1.26993680e+00 2.53048707e-02 -4.29044276e-01 -1.12218007e-01
3.04417610e-01 5.17179072e-01 -9.66040850e-01 5.70174277e-01
1.21287680e+00 7.13056922e-01 -1.00077248e+00 1.07691848e+00
-4.59834695e-01 -1.62264839e-01 -7.74153590e-01 -3.29784572e-01
4.10614550e-01 -1.12879522e-01 3.74540687e-01 1.33002090e+00
3.00248563e-01 -2.34657928e-01 1.90491199e-01 4.13271666e-01
3.14280272e-01 -1.94804743e-01 -6.25220895e-01 5.32987535e-01
2.69847006e-01 1.35052824e+00 -1.54914832e+00 -3.25452268e-01
-6.55770525e-02 1.49751019e+00 2.10594758e-01 3.75381708e-01
-8.27602446e-01 -3.89449805e-01 5.99337637e-01 -5.55920713e-02
3.58416855e-01 -7.32844949e-01 -4.50672597e-01 -8.05805385e-01
7.12622181e-02 -2.99152106e-01 1.48427403e-02 -1.04439044e+00
-6.00743055e-01 7.25715518e-01 -7.58881718e-02 -1.00617421e+00
-2.59891331e-01 -1.12927413e+00 3.83086801e-02 4.36895609e-01
-1.73162627e+00 -1.17667949e+00 -9.55867529e-01 8.56644750e-01
3.89935881e-01 2.94615656e-01 8.56560230e-01 3.46538574e-01
-4.99309242e-01 9.65418003e-04 -1.48404613e-02 -4.14262861e-02
9.88057256e-02 -1.56466448e+00 6.55620217e-01 1.16266382e+00
-3.43818218e-02 4.24046397e-01 8.42675269e-01 -5.90795457e-01
-1.55354452e+00 -1.32028115e+00 4.51807350e-01 -4.63557631e-01
1.67626664e-01 -6.68705225e-01 -2.39916071e-01 6.64017916e-01
-2.11170148e-02 -1.22353926e-01 2.68187106e-01 6.98093846e-02
1.74900934e-01 -1.29701598e-02 -1.33903110e+00 8.27486157e-01
1.54833066e+00 -2.78968573e-01 -1.14887021e-01 5.97375274e-01
7.29400635e-01 -7.66010642e-01 -4.18680310e-01 1.58416152e-01
4.55835909e-01 -1.44315577e+00 1.36976016e+00 7.80598521e-02
-8.01368505e-02 -5.57025313e-01 -5.27458012e-01 -1.00108254e+00
1.19998299e-01 -2.91551620e-01 7.22756833e-02 7.56849825e-01
1.62611768e-01 -6.08346522e-01 1.16905284e+00 5.59093952e-01
-3.87470067e-01 -1.52531892e-01 -9.52152073e-01 -8.30869019e-01
-7.15597391e-01 -1.10188758e+00 5.56285799e-01 4.08776015e-01
-3.69701147e-01 1.55926809e-01 -1.55616757e-02 5.16324818e-01
7.99321055e-01 -1.52756363e-01 1.14052081e+00 -1.06831062e+00
3.84542421e-02 -3.12502652e-01 -7.82893956e-01 -1.30768609e+00
5.33154141e-03 -6.83296144e-01 4.95565951e-01 -2.12461519e+00
-3.35984707e-01 -7.33553708e-01 1.76365703e-01 3.77293706e-01
8.58760849e-02 6.52478516e-01 1.48378626e-01 4.84996065e-02
-1.18516338e+00 3.68435055e-01 9.76309180e-01 -1.15600139e-01
-2.95812964e-01 1.45069687e-02 -3.38946044e-01 9.94807541e-01
7.33380854e-01 5.13632447e-02 -2.94913292e-01 -5.15790701e-01
2.75019228e-01 -4.09515530e-01 9.15295362e-01 -1.55076396e+00
3.70520711e-01 2.45777786e-01 5.01284838e-01 -8.54365826e-01
9.37688351e-01 -9.30850804e-01 -6.87441081e-02 6.91117823e-01
3.64811182e-01 -1.28298312e-01 1.72797307e-01 3.88340175e-01
2.39059523e-01 -1.12971790e-01 5.42212188e-01 -4.33899581e-01
-1.62312949e+00 1.20721329e-02 -4.70849097e-01 -4.30637121e-01
8.55725825e-01 -6.65245652e-01 -3.58530939e-01 -3.19938004e-01
-6.17490113e-01 4.24184769e-01 7.40647435e-01 2.60110140e-01
6.48460031e-01 -1.04246342e+00 -1.30774871e-01 -7.12710097e-02
8.04921985e-02 4.68566656e-01 4.01596159e-01 8.31926644e-01
-1.30956793e+00 6.77600026e-01 -2.00797915e-01 -1.05147886e+00
-1.07150435e+00 4.29153770e-01 3.40902418e-01 1.92178525e-02
-7.86263883e-01 8.07692766e-01 -5.40957116e-02 -9.21916366e-01
4.40049678e-01 -6.31359220e-01 -2.98888862e-01 -1.37996256e-01
2.72202283e-01 7.98601687e-01 3.08799654e-01 -7.59630024e-01
-8.43126476e-01 5.75497150e-01 6.73980713e-01 -2.15715900e-01
1.29072368e+00 -7.02856541e-01 7.36569008e-03 3.97696078e-01
9.10666823e-01 -3.17697048e-01 -1.28374028e+00 2.29965881e-01
-4.87298630e-02 -2.08295435e-01 -2.62764096e-02 -6.28335893e-01
-8.45956445e-01 9.05190468e-01 9.22900856e-01 3.65476638e-01
1.07559097e+00 2.26610750e-02 6.53775036e-01 1.02221417e+00
9.92701113e-01 -1.23359907e+00 -4.51278389e-02 6.20073140e-01
3.92118335e-01 -1.46272504e+00 5.03434241e-02 -7.59417236e-01
-4.78048086e-01 8.59042406e-01 4.14737761e-01 -3.11483927e-02
3.87467533e-01 5.08127809e-02 1.07688025e-01 -3.79195303e-01
1.80799842e-01 -9.09060299e-01 -8.59779865e-02 1.08397901e+00
-1.24948286e-01 2.32926652e-01 5.51658511e-01 1.04501843e-01
-6.11553907e-01 -3.54203969e-01 4.42251146e-01 1.01106787e+00
-6.37422800e-01 -7.17803299e-01 -6.09016061e-01 1.24971364e-02
-2.28215009e-02 5.83924986e-02 -3.16859335e-01 1.13279009e+00
2.28757396e-01 1.29441130e+00 7.59080499e-02 -3.43068928e-01
4.84074414e-01 -3.81854355e-01 7.51154244e-01 -3.73946369e-01
-3.54400486e-01 -1.69843137e-01 3.59353989e-01 -1.19128180e+00
-8.60075295e-01 -6.48757935e-01 -1.60128355e+00 -2.69940346e-01
1.30613923e-01 -3.54134291e-01 1.18872261e+00 1.09469974e+00
1.88197017e-01 6.29942417e-01 2.26067021e-01 -1.36983633e+00
3.36515576e-01 -6.50300026e-01 -5.32692432e-01 3.34998548e-01
1.77727222e-01 -7.58866191e-01 -4.66494113e-02 -1.91027239e-01] | [8.400199890136719, -2.184539794921875] |
b121a515-76a7-47b3-9374-ef790d3b225d | time-series-alignment-with-global-invariances | 2002.03848 | null | https://arxiv.org/abs/2002.03848v2 | https://arxiv.org/pdf/2002.03848v2.pdf | Time Series Alignment with Global Invariances | Multivariate time series are ubiquitous objects in signal processing. Measuring a distance or similarity between two such objects is of prime interest in a variety of applications, including machine learning, but can be very difficult as soon as the temporal dynamics and the representation of the time series, {\em i.e.} the nature of the observed quantities, differ from one another. In this work, we propose a novel distance accounting both feature space and temporal variabilities by learning a latent global transformation of the feature space together with a temporal alignment, cast as a joint optimization problem. The versatility of our framework allows for several variants depending on the invariance class at stake. Among other contributions, we define a differentiable loss for time series and present two algorithms for the computation of time series barycenters under this new geometry. We illustrate the interest of our approach on both simulated and real world data and show the robustness of our approach compared to state-of-the-art methods. | ['Nicolas Courty', 'Laetitia Chapel', 'Romain Tavenard', 'Rémi Flamary', 'Yann Soullard', 'Titouan Vayer'] | 2020-02-10 | null | null | null | null | ['time-series-alignment'] | ['time-series'] | [ 1.70695037e-01 -4.46897209e-01 2.77450234e-01 -2.75937378e-01
-4.46257323e-01 -8.80324841e-01 7.78781235e-01 4.56005543e-01
-5.85885406e-01 4.68465149e-01 -1.51540637e-01 5.25905415e-02
-7.19561398e-01 -4.77255464e-01 -5.16033947e-01 -9.49717879e-01
-6.09921932e-01 2.16298819e-01 8.48254934e-03 -1.83932200e-01
2.99494207e-01 7.95375049e-01 -1.21248209e+00 -3.80626142e-01
6.20251775e-01 1.08921504e+00 -7.92447180e-02 6.90037549e-01
4.84697282e-01 1.80315122e-01 -5.49018800e-01 -8.29059631e-02
5.69035649e-01 -3.64849329e-01 -4.76569623e-01 2.22679362e-01
3.53434198e-02 3.40917021e-01 -2.02964708e-01 1.04047894e+00
4.17301804e-01 5.18413365e-01 7.39918172e-01 -1.23917198e+00
-8.50131363e-02 8.18789601e-02 -4.48664010e-01 4.43687767e-01
1.35795340e-01 -2.63788193e-01 9.05778229e-01 -5.27427614e-01
5.60972989e-01 7.56081641e-01 7.02128470e-01 -9.49922130e-02
-1.77469718e+00 -1.31966025e-01 -3.72675881e-02 1.75893277e-01
-1.23661327e+00 -2.93835998e-01 1.22717166e+00 -7.79869080e-01
3.85361910e-01 3.14798176e-01 6.19871259e-01 7.17811763e-01
4.71481562e-01 1.99559614e-01 1.09818149e+00 -4.35959399e-01
5.01597285e-01 -1.71602443e-01 -2.63087619e-02 3.34940642e-01
-7.89627805e-02 3.45037989e-02 -3.14597398e-01 -2.86893159e-01
7.32796490e-01 1.99525595e-01 -2.96029627e-01 -1.02651405e+00
-1.51536667e+00 8.07275355e-01 1.11661501e-01 6.16417170e-01
-4.40501571e-01 -2.63776947e-02 2.94319510e-01 5.64554989e-01
6.22242510e-01 3.34900141e-01 -2.58637995e-01 -2.11296946e-01
-8.26466262e-01 3.52528095e-01 7.23150849e-01 5.50222754e-01
5.25139868e-01 -1.32139832e-01 1.94495618e-01 3.71876478e-01
4.87026870e-02 3.66552144e-01 4.28529084e-01 -6.95025206e-01
3.38408947e-01 1.41957045e-01 5.95543757e-02 -1.22083473e+00
-6.20567739e-01 -6.31640196e-01 -8.49185884e-01 7.37628415e-02
8.60796988e-01 -4.68218066e-02 -1.88981593e-01 2.03695941e+00
3.40855300e-01 4.45670307e-01 -2.43483126e-01 6.69593751e-01
-3.00218940e-01 5.42087376e-01 -3.87243360e-01 -7.42067337e-01
1.01913393e+00 -3.10235500e-01 -6.67766035e-01 2.18456581e-01
2.70464987e-01 -8.35777223e-01 5.00001132e-01 3.66348416e-01
-1.13397622e+00 -2.92331368e-01 -9.61130977e-01 3.47674429e-01
-2.49166712e-01 -1.16325781e-01 1.70924157e-01 3.50857377e-01
-8.12946856e-01 1.05040038e+00 -1.17106128e+00 -4.56993639e-01
-2.40643218e-01 3.21081340e-01 -3.60926807e-01 6.32683992e-01
-8.99902821e-01 7.26290107e-01 1.28267884e-01 2.27918670e-01
-3.37865233e-01 -6.00996375e-01 -6.23756111e-01 -2.12562755e-02
1.93072796e-01 -3.75267893e-01 1.10317910e+00 -8.10435355e-01
-1.39920545e+00 7.01898634e-01 -7.31262960e-04 -4.36800778e-01
9.09079969e-01 3.22948173e-02 -4.95161086e-01 5.86307794e-02
9.05205831e-02 -2.60225952e-01 1.23275077e+00 -5.73492289e-01
-2.82591701e-01 -5.45808434e-01 -1.10346628e-02 -9.16331187e-02
-1.30333424e-01 2.98171137e-02 1.11401990e-01 -9.01703894e-01
3.73560727e-01 -9.67318654e-01 -3.33939344e-01 1.87048867e-01
-9.97867957e-02 5.45077324e-02 7.62098014e-01 -4.13203597e-01
1.21839321e+00 -2.31450605e+00 5.98145187e-01 4.01952446e-01
7.48412758e-02 -1.79614037e-01 1.10339902e-01 7.34158099e-01
-5.88744044e-01 -2.36797094e-01 -5.17362237e-01 -2.73426801e-01
6.60437047e-02 -1.20840587e-01 -5.12530386e-01 1.27333701e+00
1.46433204e-01 4.40162122e-01 -1.02770293e+00 -1.20394148e-01
2.33544379e-01 4.26515758e-01 -3.60225290e-01 -1.66214991e-03
1.23110980e-01 9.79290605e-01 -4.82696056e-01 2.40474287e-02
3.54725093e-01 -3.48789729e-02 -1.72414538e-02 -2.49060661e-01
-5.23887336e-01 5.83824813e-02 -1.41176665e+00 1.68527508e+00
-4.00991470e-01 7.31850743e-01 -1.11448569e-02 -1.45326293e+00
8.48624766e-01 4.68185574e-01 9.25346673e-01 -5.14297724e-01
2.87443437e-02 2.60192901e-01 8.24631006e-02 -4.68425632e-01
2.36107439e-01 -3.43262076e-01 -1.95580930e-01 5.29986680e-01
-1.06047481e-01 -5.06653637e-02 1.66869804e-01 -3.24275792e-01
9.78319764e-01 1.39767081e-01 6.20295703e-01 -5.68607152e-01
5.74481905e-01 -5.13127029e-01 6.14946842e-01 2.80621648e-01
-2.51327723e-01 4.05325413e-01 7.95765460e-01 -3.96976143e-01
-1.18168688e+00 -1.12341464e+00 -4.54421043e-01 4.10102129e-01
-9.80928689e-02 -2.70508170e-01 -3.30362767e-01 -3.03094774e-01
-2.00978890e-02 2.98196048e-01 -6.50049448e-01 -1.94669947e-01
-8.58338952e-01 -7.49632835e-01 1.18158013e-01 1.82574064e-01
-2.47481335e-02 -6.19480968e-01 -1.00515735e+00 3.82473737e-01
-6.91446336e-03 -1.05002904e+00 -4.68170792e-01 1.52150080e-01
-1.20558643e+00 -7.81736016e-01 -6.44239426e-01 -3.01526338e-01
3.85412991e-01 -1.59676392e-02 8.48610640e-01 -5.77481091e-01
-4.48935390e-01 7.04913437e-01 -9.14698318e-02 -2.91148067e-01
-2.03219831e-01 -3.01746041e-01 3.36557686e-01 6.91294074e-01
-2.45561734e-01 -1.11846888e+00 -4.85422552e-01 3.06151688e-01
-1.15874958e+00 -3.51303130e-01 -8.25206339e-02 6.21062636e-01
6.31951034e-01 3.31841856e-01 3.91403168e-01 -3.41639817e-01
6.19111001e-01 -5.29871464e-01 -1.00385988e+00 1.96067020e-01
-1.71858087e-01 3.36912930e-01 7.27115333e-01 -5.88031113e-01
-5.41387558e-01 3.22058983e-02 4.04581904e-01 -4.10529435e-01
1.84133634e-01 6.38618290e-01 -8.49903524e-02 -9.35490206e-02
3.18148464e-01 9.27590653e-02 -2.43664766e-03 -6.22609258e-01
2.57470131e-01 1.46988571e-01 4.90895540e-01 -6.92569673e-01
9.29705262e-01 7.94192314e-01 7.13911474e-01 -1.10874832e+00
-4.70532447e-01 -5.98751903e-01 -1.05256212e+00 -2.07838193e-01
7.01904356e-01 -3.72660965e-01 -7.11162448e-01 5.24978280e-01
-1.18913507e+00 3.03481333e-02 -6.75752640e-01 8.21943462e-01
-1.01334786e+00 4.45666820e-01 -1.92392796e-01 -8.34230721e-01
1.80409104e-01 -1.00909781e+00 1.02359962e+00 -2.12172404e-01
-1.96779743e-01 -1.40668130e+00 6.47270143e-01 -3.84216309e-01
2.75510818e-01 7.91101217e-01 9.08612490e-01 -5.55416763e-01
-3.97177875e-01 -4.29520041e-01 2.51314938e-01 7.68739879e-02
3.25301975e-01 1.17907554e-01 -7.15959609e-01 -4.06754553e-01
6.02049828e-01 3.93936694e-01 5.18677294e-01 5.02568662e-01
8.92628014e-01 -2.01907262e-01 -1.29871204e-01 7.23044634e-01
1.32335055e+00 2.25438967e-01 1.23169243e-01 2.56146580e-01
3.70758653e-01 8.05756390e-01 5.62402487e-01 7.39208519e-01
-1.61022931e-01 1.23370171e+00 4.08349574e-01 3.45420599e-01
5.92608213e-01 2.00760603e-01 2.86050349e-01 1.03942883e+00
-2.53622591e-01 1.27193570e-01 -7.63034284e-01 5.47441721e-01
-1.92720830e+00 -1.18141031e+00 -9.63027216e-03 2.76320314e+00
4.24720228e-01 -3.06890476e-02 3.34569126e-01 4.85046417e-01
7.59213328e-01 3.62736523e-01 -5.15972078e-01 -1.05565801e-01
-1.60094127e-01 2.97510415e-01 3.87439221e-01 4.57482278e-01
-1.29399955e+00 -1.89891346e-02 5.57584763e+00 3.88550341e-01
-1.44531822e+00 -3.03827561e-02 1.27372518e-01 -3.76438908e-02
5.41894361e-02 1.02940448e-01 -1.67289674e-01 4.40458268e-01
8.54282677e-01 -7.26565897e-01 3.52860987e-01 2.78914452e-01
4.09358770e-01 2.25779772e-01 -1.30512333e+00 1.09867227e+00
2.36970466e-02 -8.44040215e-01 -4.74713117e-01 2.88275301e-01
4.48879987e-01 -1.93606824e-01 2.46694744e-01 -3.28570843e-01
-5.32802045e-01 -6.28041804e-01 1.04503071e+00 8.40571046e-01
3.35513741e-01 -5.52267253e-01 2.32010826e-01 3.47116798e-01
-1.42029977e+00 3.87044512e-02 -3.39428429e-03 -2.50196546e-01
5.00356972e-01 8.12539816e-01 -3.35896403e-01 7.63449192e-01
2.57461309e-01 9.37023461e-01 -4.88090277e-01 1.36932313e+00
2.02696085e-01 3.97656083e-01 -5.97619891e-01 8.52569193e-02
9.50205177e-02 -6.42076492e-01 1.20197999e+00 8.98745656e-01
5.57448447e-01 -6.85103014e-02 1.58945709e-01 7.70240843e-01
3.57795000e-01 1.84317067e-01 -5.88328242e-01 -2.22001165e-01
1.80495039e-01 1.01292443e+00 -9.33022201e-01 1.13793835e-01
-2.39580199e-01 7.34865010e-01 5.43775223e-02 3.74677688e-01
-7.28213966e-01 -3.76542389e-01 7.30433285e-01 3.38245779e-02
2.27282166e-01 -7.48275042e-01 5.82722351e-02 -1.46009529e+00
5.74513435e-01 -4.89885658e-01 4.99971688e-01 -3.22994947e-01
-1.19103241e+00 5.47276974e-01 2.98608005e-01 -1.85308683e+00
-5.28628707e-01 -4.86569077e-01 -6.12932682e-01 7.88071096e-01
-1.03880739e+00 -6.51511788e-01 8.83130208e-02 7.12164283e-01
2.13956699e-01 1.31427065e-01 5.99355578e-01 2.18195900e-01
-3.30350965e-01 9.72230732e-02 5.22800982e-01 -1.09521650e-01
6.21886253e-01 -1.30152488e+00 3.95179957e-01 9.41942811e-01
3.54481190e-01 5.61058581e-01 1.17177033e+00 -2.11018279e-01
-1.39218402e+00 -8.40799689e-01 8.60653102e-01 -2.54500896e-01
1.27273059e+00 -4.29659724e-01 -7.29591310e-01 7.14643896e-01
-1.00545175e-01 2.18830332e-01 4.91745979e-01 -1.85673118e-01
-1.76906854e-01 -3.33703727e-01 -8.69576693e-01 5.68265140e-01
7.62155056e-01 -6.90549731e-01 -5.11726916e-01 3.33948344e-01
4.00348425e-01 -1.32991999e-01 -1.01252437e+00 3.06221724e-01
6.12644315e-01 -1.07005441e+00 9.61949468e-01 -6.31973326e-01
-1.28634259e-01 -4.11367625e-01 -3.36130321e-01 -1.38356626e+00
-2.27902979e-01 -1.06254244e+00 -2.42472831e-02 1.00476301e+00
-1.75173078e-02 -9.37248707e-01 3.13076973e-01 2.99494237e-01
2.43737876e-01 -5.30966759e-01 -1.24434447e+00 -1.10260451e+00
4.50846106e-02 -5.19659162e-01 3.76166642e-01 1.01808131e+00
1.96788296e-01 1.51319072e-01 -8.34093019e-02 3.63198042e-01
9.28323627e-01 4.34552103e-01 4.28976864e-01 -1.46214938e+00
-5.90537667e-01 -6.39567077e-01 -1.12251544e+00 -6.60923123e-01
9.88920182e-02 -7.31159747e-01 -2.65497923e-01 -6.69302404e-01
-2.74818957e-01 -3.15654308e-01 -3.12773198e-01 -1.98309481e-01
2.51687050e-01 -8.40232670e-02 2.62489974e-01 2.38611549e-01
-8.52350444e-02 6.74236357e-01 9.01047051e-01 1.35027796e-01
-3.21334362e-01 4.51048613e-01 1.59664303e-01 8.10728729e-01
6.44787669e-01 -3.83428723e-01 -3.88205141e-01 -2.01181084e-01
2.31235117e-01 4.58505929e-01 6.29387259e-01 -1.15621734e+00
2.65708834e-01 -8.72483999e-02 -7.69527480e-02 -3.22714359e-01
4.30778950e-01 -9.87760484e-01 4.21771407e-01 3.34088236e-01
-3.62148196e-01 5.86108387e-01 -4.37096618e-02 7.44086742e-01
-3.63539159e-01 -2.31494218e-01 8.11169505e-01 2.50358135e-01
-2.07483858e-01 3.20280254e-01 -6.22410327e-02 3.69062498e-02
1.09595478e+00 -7.04703480e-02 8.41206014e-02 -5.36645949e-01
-9.60071862e-01 -1.17220908e-01 3.83028418e-01 2.36133873e-01
2.30674148e-01 -1.38734889e+00 -5.63064456e-01 3.46374005e-01
1.13791622e-01 -3.79795581e-01 1.82457194e-01 1.45617318e+00
-3.26328069e-01 3.90616715e-01 -2.48246849e-01 -7.74872780e-01
-1.25217605e+00 7.03973591e-01 4.68704462e-01 -3.11833620e-01
-6.55540168e-01 1.57438412e-01 2.47166932e-01 -1.99625254e-01
4.32703272e-02 -5.86298466e-01 -7.16734538e-03 3.41409028e-01
2.67409474e-01 5.02421439e-01 1.84532002e-01 -8.80184770e-01
-2.87674218e-01 9.11591053e-01 3.14275324e-01 -5.35196662e-01
1.37654126e+00 -1.27950296e-01 -2.94047296e-01 1.11881030e+00
1.48451090e+00 1.65110126e-01 -1.24758923e+00 -3.44533741e-01
2.63344526e-01 -3.99930686e-01 -3.35191876e-01 1.68191027e-02
-8.12919617e-01 8.85113716e-01 5.28673112e-01 6.87503457e-01
1.35586202e+00 -1.46845862e-01 2.85181254e-01 2.23230362e-01
4.28357869e-01 -8.34623218e-01 -1.15875483e-01 3.59763950e-01
9.47280824e-01 -7.03353405e-01 -7.30302036e-02 -3.27041656e-01
-4.50446419e-02 1.30549955e+00 -3.72090369e-01 -5.28598249e-01
1.00377440e+00 -7.46557340e-02 -8.98007080e-02 2.61135548e-02
-5.27447641e-01 -8.48659799e-02 3.95955145e-01 2.79461652e-01
4.96812165e-01 1.09140731e-01 -5.16211152e-01 1.69998124e-01
-2.84802139e-01 -3.55927497e-01 4.60963756e-01 1.10722971e+00
2.14799289e-02 -1.14221966e+00 -4.52164799e-01 9.63636022e-03
-5.20209253e-01 3.47586215e-01 2.60518747e-03 7.68035889e-01
-1.34899288e-01 6.50556087e-01 1.76652193e-01 1.37460344e-02
5.20206034e-01 6.62530884e-02 5.92166245e-01 -1.99151888e-01
-2.49140814e-01 3.88178170e-01 -4.02775556e-01 -6.51160717e-01
-7.10472405e-01 -1.28297293e+00 -7.96656787e-01 3.76695469e-02
-1.76626310e-01 1.85798258e-01 8.70034933e-01 1.04714799e+00
3.87863889e-02 3.41969728e-01 1.01626456e+00 -8.95398974e-01
-8.54869425e-01 -5.67232966e-01 -9.74293053e-01 5.17714560e-01
8.45286489e-01 -6.19515419e-01 -4.69578147e-01 2.07314283e-01] | [7.454353332519531, 3.389874219894409] |
67941c07-e89f-48e6-88ba-04b85f5ab3fb | interventional-video-grounding-with-dual | 2106.11013 | null | https://arxiv.org/abs/2106.11013v2 | https://arxiv.org/pdf/2106.11013v2.pdf | Interventional Video Grounding with Dual Contrastive Learning | Video grounding aims to localize a moment from an untrimmed video for a given textual query. Existing approaches focus more on the alignment of visual and language stimuli with various likelihood-based matching or regression strategies, i.e., P(Y|X). Consequently, these models may suffer from spurious correlations between the language and video features due to the selection bias of the dataset. 1) To uncover the causality behind the model and data, we first propose a novel paradigm from the perspective of the causal inference, i.e., interventional video grounding (IVG) that leverages backdoor adjustment to deconfound the selection bias based on structured causal model (SCM) and do-calculus P(Y|do(X)). Then, we present a simple yet effective method to approximate the unobserved confounder as it cannot be directly sampled from the dataset. 2) Meanwhile, we introduce a dual contrastive learning approach (DCL) to better align the text and video by maximizing the mutual information (MI) between query and video clips, and the MI between start/end frames of a target moment and the others within a video to learn more informative visual representations. Experiments on three standard benchmarks show the effectiveness of our approaches. Our code is available on GitHub: https://github.com/nanguoshun/IVG. | ['Wei Lu', 'Hao Zhang', 'Sicong Leng', 'Jun Liu', 'Yao Xiao', 'Rui Qiao', 'Guoshun Nan'] | 2021-06-21 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Nan_Interventional_Video_Grounding_With_Dual_Contrastive_Learning_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Nan_Interventional_Video_Grounding_With_Dual_Contrastive_Learning_CVPR_2021_paper.pdf | cvpr-2021-1 | ['video-grounding'] | ['computer-vision'] | [ 1.65345550e-01 -1.79472536e-01 -7.70583689e-01 -3.33413273e-01
-8.64045262e-01 -4.77336168e-01 6.86595023e-01 -1.29778489e-01
-5.38807400e-02 5.29009461e-01 6.00574553e-01 -1.97012201e-02
-1.26994520e-01 -3.68036866e-01 -1.12930512e+00 -6.01089776e-01
-3.10521200e-02 -2.76332527e-01 -1.52499467e-01 5.20784676e-01
2.01752052e-01 -4.05208617e-02 -1.39214456e+00 4.18308049e-01
5.93110025e-01 7.81623065e-01 3.99032801e-01 5.25382340e-01
1.45509034e-01 1.21159804e+00 -3.13819736e-01 -4.56002802e-01
6.19000159e-02 -7.10149944e-01 -4.31928515e-01 1.35230226e-02
7.09622979e-01 -5.51440239e-01 -1.00880098e+00 1.17276955e+00
2.25501925e-01 -8.43375474e-02 5.45549691e-01 -1.80028319e+00
-7.11821198e-01 7.08639681e-01 -8.79451334e-01 5.07522225e-01
7.24873364e-01 1.58610016e-01 1.13729525e+00 -1.00650001e+00
7.83480704e-01 1.46272254e+00 2.22962976e-01 2.19106123e-01
-1.28079927e+00 -1.07461631e+00 5.29592812e-01 5.64091802e-01
-1.45834923e+00 -4.78254586e-01 1.02496207e+00 -6.37612700e-01
2.35991627e-02 2.60986924e-01 5.74869454e-01 1.66889250e+00
9.53502357e-02 1.05116165e+00 1.08511770e+00 -2.01056838e-01
-9.42564905e-02 -9.69221145e-02 -2.14075804e-01 1.03896511e+00
1.21981129e-01 2.89757699e-01 -1.11987388e+00 -2.91437864e-01
8.21736038e-01 3.62951159e-01 -4.49564666e-01 -3.41434240e-01
-1.70806360e+00 8.86584103e-01 2.39112884e-01 4.00152691e-02
-5.11199355e-01 2.81761676e-01 8.64509121e-02 5.51132411e-02
2.35164866e-01 3.04268729e-02 -2.61260003e-01 1.23251453e-01
-9.85251307e-01 3.29969794e-01 4.49364096e-01 1.03277564e+00
8.48727882e-01 -1.58018515e-01 -6.10139251e-01 3.14325452e-01
5.39865792e-01 6.68664157e-01 3.43289167e-01 -9.55115378e-01
5.57519794e-01 3.87143612e-01 3.62096429e-02 -1.51818144e+00
2.87988055e-02 3.23876999e-02 -6.39185190e-01 -3.16343307e-01
4.98968601e-01 -2.17348099e-01 -6.33153200e-01 2.06144452e+00
3.34296703e-01 9.91374433e-01 -3.32399338e-01 1.07255065e+00
9.99672055e-01 7.05784917e-01 2.35687315e-01 -5.37950218e-01
1.29589605e+00 -5.13434887e-01 -8.80109608e-01 -2.56119877e-01
2.56682962e-01 -7.16568947e-01 1.19407523e+00 1.10803515e-01
-8.54276478e-01 -5.24929166e-01 -8.54319751e-01 -5.89840338e-02
6.73500896e-02 2.63750821e-01 4.87092197e-01 5.31888455e-02
-4.50045377e-01 2.93584049e-01 -7.70133674e-01 -1.73133254e-01
3.92975122e-01 5.10355830e-02 -3.20288092e-01 -2.72015959e-01
-1.28274727e+00 1.79968312e-01 1.77822247e-01 1.43596698e-02
-1.34896004e+00 -8.80852520e-01 -7.26135254e-01 -1.70333579e-01
8.53902340e-01 -6.74212515e-01 9.59283471e-01 -1.22121358e+00
-1.06615317e+00 8.90200675e-01 -5.34440994e-01 -3.29468578e-01
4.50140148e-01 -5.56979179e-01 -3.13489109e-01 3.40333939e-01
3.74237716e-01 6.42906725e-01 1.26836586e+00 -1.29212868e+00
-6.05687797e-01 -3.15391779e-01 7.17173964e-02 7.77954683e-02
-1.24384038e-01 1.16122440e-01 -8.06386471e-01 -9.92227614e-01
8.97099823e-02 -7.92877972e-01 1.51464805e-01 -2.45227907e-02
-6.73745990e-01 -1.33230016e-01 6.75851762e-01 -9.39141870e-01
1.53830898e+00 -2.26830983e+00 3.56849194e-01 8.25344101e-02
5.09745002e-01 -3.04996639e-01 -1.27535060e-01 3.59398514e-01
-3.47313702e-01 6.50955141e-02 9.23421979e-02 -6.92587420e-02
-1.33032486e-01 -1.36828190e-02 -7.05151021e-01 8.10819149e-01
7.60501623e-02 8.93418014e-01 -1.16173673e+00 -8.17842424e-01
6.06232919e-02 4.15637463e-01 -6.18896723e-01 5.34827411e-01
-2.17888728e-01 7.24658668e-01 -3.74963552e-01 7.00813949e-01
3.30434084e-01 -3.61794531e-01 2.02985480e-01 -6.17337584e-01
-6.34540021e-02 6.71689808e-02 -1.19041419e+00 1.78859055e+00
-7.96320662e-02 7.00613379e-01 -2.44751200e-01 -9.24795449e-01
6.05400980e-01 4.34123218e-01 6.27474546e-01 -5.95293343e-01
4.47140746e-02 -1.55205831e-01 -1.64295733e-01 -8.88449967e-01
1.46592769e-03 1.77215770e-01 -6.09485358e-02 2.18366981e-01
1.47999153e-01 5.48261344e-01 -1.23483269e-02 4.91274804e-01
9.04486299e-01 2.51872391e-01 3.03721845e-01 1.68928072e-01
2.27572605e-01 -2.02638328e-01 8.46020043e-01 9.61194932e-01
-2.08204210e-01 5.35197139e-01 8.22716355e-01 -2.03993291e-01
-7.72354960e-01 -1.20143878e+00 2.73551941e-01 1.04290259e+00
3.46370190e-01 -5.75315952e-01 -6.37705803e-01 -6.15476429e-01
4.68279980e-03 6.26142800e-01 -7.96521842e-01 -1.76910788e-01
-5.71926594e-01 -4.90491807e-01 3.37144077e-01 2.87416607e-01
2.74258465e-01 -8.08363438e-01 -3.62904370e-01 -1.88016742e-01
-7.27012813e-01 -1.15424693e+00 -7.91050196e-01 -3.53374779e-01
-6.26110494e-01 -1.17915654e+00 -3.68690282e-01 -3.97985369e-01
5.30516684e-01 5.10775089e-01 1.01571488e+00 6.19010180e-02
-2.02079669e-01 6.50239885e-01 -2.09092677e-01 -2.02134028e-01
-4.78301533e-02 -4.06201065e-01 1.28236786e-01 6.30353451e-01
3.92345160e-01 -3.76117826e-01 -8.16850960e-01 2.76923269e-01
-7.46637285e-01 3.49959761e-01 6.76852703e-01 8.02927315e-01
8.43042076e-01 -1.97270289e-01 2.33282909e-01 -6.84489787e-01
1.83431849e-01 -1.06430864e+00 -5.87215424e-01 2.55954742e-01
-3.76312643e-01 -1.14529230e-01 2.01583937e-01 -8.66931617e-01
-9.71513152e-01 1.61761254e-01 6.16287708e-01 -1.23504198e+00
-7.94929639e-02 6.00627244e-01 -3.99705559e-01 5.66501379e-01
2.85443842e-01 7.80002102e-02 -7.74790868e-02 -2.87528664e-01
5.12270272e-01 3.02309752e-01 6.63425207e-01 -5.05717993e-01
7.87657142e-01 7.29462206e-01 -5.39507121e-02 -5.67386091e-01
-1.05697286e+00 -4.81959760e-01 -5.05138099e-01 -3.93090546e-01
9.35960054e-01 -1.21231842e+00 -8.62001359e-01 4.36108820e-02
-1.17561245e+00 -8.74814913e-02 2.51628548e-01 8.90359282e-01
-5.52088737e-01 2.54791111e-01 -2.94718266e-01 -6.61709249e-01
1.99257955e-01 -8.46508324e-01 1.07634032e+00 1.96797132e-01
-3.55302244e-01 -7.50752687e-01 -1.96116650e-03 4.05747652e-01
-4.24429089e-01 3.98202628e-01 8.53129506e-01 -4.30210441e-01
-9.93131936e-01 -3.20390344e-01 -3.53905290e-01 -1.69543758e-01
1.39128447e-01 2.39176020e-01 -9.59054589e-01 -1.67290956e-01
3.71297784e-02 -5.38081340e-02 7.60849297e-01 6.42531574e-01
1.31927645e+00 -7.03913212e-01 -4.49076563e-01 6.19984865e-01
1.24691308e+00 3.68160270e-02 4.63559836e-01 2.29514744e-02
1.08461702e+00 5.57566822e-01 8.66565049e-01 6.29916847e-01
4.41270679e-01 7.83798695e-01 5.31282067e-01 1.42740691e-03
-1.25736400e-01 -1.03926611e+00 5.73456109e-01 5.78161895e-01
1.76522024e-02 -1.97839633e-01 -6.98000968e-01 4.59188610e-01
-2.11617827e+00 -1.23889732e+00 -1.12277187e-01 2.47667837e+00
7.46658742e-01 -1.79370686e-01 1.12186395e-01 -2.76350021e-01
1.01803946e+00 3.58392447e-01 -6.80870831e-01 5.11699677e-01
-1.69698983e-01 -3.44782889e-01 4.25066531e-01 4.17867512e-01
-1.08204019e+00 6.82206810e-01 5.45638943e+00 7.08872736e-01
-1.06320322e+00 2.83278406e-01 6.72415197e-01 -5.17683864e-01
-3.11185300e-01 2.09734365e-01 -7.23962605e-01 8.36504281e-01
7.53184378e-01 -2.98751116e-01 6.51561499e-01 3.94840062e-01
7.32486725e-01 -3.42003219e-02 -1.28031480e+00 1.16168678e+00
2.10763901e-01 -1.34459507e+00 1.73882648e-01 8.52905735e-02
6.34676516e-01 -2.54565537e-01 8.47504586e-02 1.06576450e-01
1.08221382e-01 -8.71878743e-01 1.02232194e+00 9.07795966e-01
7.60816276e-01 -3.53382230e-01 1.97851896e-01 2.45256662e-01
-9.98730004e-01 -1.61741480e-01 -1.08607918e-01 2.01527644e-02
1.06225470e-02 4.58562374e-01 -5.32185197e-01 3.88415396e-01
8.44820321e-01 9.37218606e-01 -4.95646387e-01 7.19349980e-01
-4.75414604e-01 8.87648642e-01 2.22985446e-02 1.74923792e-01
-6.87305704e-02 -4.17272896e-02 8.21436226e-01 1.00556505e+00
2.61262000e-01 4.04321179e-02 2.96880364e-01 1.07537985e+00
-1.90074548e-01 9.71692428e-02 -7.47088432e-01 -1.33855239e-01
6.94648921e-01 1.05638540e+00 -4.76592898e-01 -2.41978049e-01
-7.54802763e-01 8.39620411e-01 3.50332856e-01 6.94197178e-01
-1.28706777e+00 1.41793117e-01 5.91578543e-01 1.03001699e-01
2.29598075e-01 -8.01844895e-02 1.60015952e-02 -1.47335362e+00
5.92988506e-02 -9.85140979e-01 7.22886264e-01 -1.02134573e+00
-1.20122647e+00 5.29848896e-02 3.94184977e-01 -1.34243941e+00
-6.16240501e-02 -2.12582901e-01 -4.62379843e-01 6.74551606e-01
-1.26105165e+00 -1.05725586e+00 -3.94608259e-01 9.41094518e-01
5.49597085e-01 4.44042794e-02 3.09109390e-01 3.55226129e-01
-6.82258368e-01 4.06999052e-01 -2.92914271e-01 2.15189546e-01
1.02433407e+00 -9.19081569e-01 -2.48702377e-01 1.00196123e+00
5.12661219e-01 6.80336893e-01 7.56009698e-01 -9.01897728e-01
-1.74482965e+00 -1.05431151e+00 8.02619696e-01 -6.28641188e-01
9.35849845e-01 -3.34554285e-01 -8.56755257e-01 9.29804325e-01
2.07464054e-01 -2.52313539e-02 6.60685599e-01 -2.29935441e-02
-5.83162665e-01 -1.85225978e-01 -5.79608619e-01 8.59725416e-01
1.09411728e+00 -7.86228597e-01 -4.96278167e-01 3.50024730e-01
9.18740273e-01 -3.30777735e-01 -5.91010273e-01 2.58919954e-01
6.96106851e-01 -8.88886392e-01 9.74366546e-01 -7.75602400e-01
9.13414061e-01 -4.79685634e-01 -3.80160391e-01 -9.36383903e-01
-3.29028815e-01 -7.57113516e-01 -4.88940537e-01 1.40672743e+00
1.70187935e-01 -1.22942679e-01 5.09341240e-01 4.61158216e-01
3.91266495e-01 -5.29140234e-01 -8.29600096e-01 -4.83035654e-01
-4.00200069e-01 -5.32575786e-01 5.03630340e-01 1.11056471e+00
-1.79061830e-01 3.86452079e-01 -8.91974211e-01 6.17368877e-01
7.02231944e-01 3.41986954e-01 8.17340612e-01 -8.60942304e-01
-3.31491470e-01 -1.34650201e-01 -1.50911868e-01 -9.20460939e-01
3.31317097e-01 -7.03068972e-01 -5.70266806e-02 -9.92766798e-01
6.20677888e-01 1.38646320e-01 -4.81686294e-01 2.28665277e-01
-5.39371908e-01 -5.28529249e-02 2.52837479e-01 6.05977416e-01
-5.99010766e-01 4.79243696e-01 1.22190034e+00 -1.14579149e-01
-6.26112744e-02 -2.45728418e-01 -6.63651466e-01 8.35609138e-01
4.38559264e-01 -8.05018842e-01 -5.27414501e-01 -3.51769596e-01
2.49104694e-01 4.30404812e-01 9.79466140e-01 -3.68874848e-01
2.47241750e-01 -4.38236326e-01 3.72621864e-01 -5.57635307e-01
7.21314773e-02 -7.27142930e-01 2.05761388e-01 2.15018049e-01
-5.57205796e-01 2.02609301e-01 -1.32638603e-01 1.02534616e+00
-3.87639225e-01 -5.74747473e-03 2.86660045e-01 -5.71751595e-02
-6.49928391e-01 4.99342740e-01 -5.08101881e-02 1.90162838e-01
8.60973895e-01 1.07548758e-01 -3.36928487e-01 -6.25194132e-01
-4.73363727e-01 1.93017453e-01 2.13140413e-01 6.65343285e-01
6.51372194e-01 -1.56029058e+00 -6.28213704e-01 4.11407277e-02
1.60118937e-01 -5.13965726e-01 3.17823470e-01 1.21607578e+00
1.34041592e-01 3.19499016e-01 1.61663249e-01 -7.00818837e-01
-1.31600535e+00 9.23581898e-01 2.84425709e-02 1.17748685e-01
-4.33338046e-01 5.79929113e-01 8.22593927e-01 1.74281463e-01
3.79407853e-01 -2.26102278e-01 -1.10768780e-01 2.65299231e-01
5.60070455e-01 3.04805130e-01 -6.45963788e-01 -6.96369231e-01
-3.27842861e-01 2.51037985e-01 1.21240057e-01 -1.33882105e-01
9.03398573e-01 -3.52519780e-01 -7.57365068e-03 8.07988882e-01
1.26812387e+00 1.37449771e-01 -1.64893258e+00 -4.86828178e-01
-2.18125004e-02 -8.71664047e-01 1.51659980e-01 -5.06238639e-01
-1.08726764e+00 6.32056534e-01 5.50810575e-01 -5.54422252e-02
1.08594978e+00 2.63168335e-01 3.69517893e-01 -5.48793525e-02
-2.17879117e-02 -6.46259964e-01 3.02014261e-01 -1.97424963e-01
1.07424796e+00 -1.36776543e+00 1.00135677e-01 -4.46216464e-01
-6.77606285e-01 8.42977405e-01 4.72885996e-01 -1.07843973e-01
5.71822047e-01 -1.15802400e-01 -4.45138775e-02 -3.15532207e-01
-9.24075544e-01 -1.55986786e-01 3.92564565e-01 3.03230882e-01
3.81653547e-01 1.39038041e-01 -1.63942501e-01 8.30613852e-01
2.10839331e-01 1.62499666e-01 2.82897979e-01 5.10887027e-01
1.41999930e-01 -6.21240675e-01 -5.77758670e-01 2.38577381e-01
-4.86016840e-01 -1.69762656e-01 -3.65209907e-01 8.06783617e-01
1.39103755e-01 8.59012306e-01 1.01652704e-01 -4.51847404e-01
1.84269160e-01 -7.26772994e-02 4.10479456e-01 -3.55408669e-01
1.08201981e-01 4.48373139e-01 -2.39609912e-01 -9.95156109e-01
-7.64885724e-01 -1.08066416e+00 -8.81788015e-01 -8.51831809e-02
-1.00223377e-01 -1.62746146e-01 2.56477475e-01 8.99428666e-01
3.41314048e-01 2.74516940e-01 8.36260140e-01 -6.41221285e-01
-1.44547135e-01 -6.39233053e-01 -4.46793288e-01 6.72432482e-01
4.25057143e-01 -1.04069209e+00 -4.83577996e-01 6.12687528e-01] | [10.068399429321289, 0.7955071926116943] |
986fdc58-04ab-4a48-ac86-18dd9684372e | content4all-open-research-sign-language | 2105.02351 | null | https://arxiv.org/abs/2105.02351v1 | https://arxiv.org/pdf/2105.02351v1.pdf | Content4All Open Research Sign Language Translation Datasets | Computational sign language research lacks the large-scale datasets that enables the creation of useful reallife applications. To date, most research has been limited to prototype systems on small domains of discourse, e.g. weather forecasts. To address this issue and to push the field forward, we release six datasets comprised of 190 hours of footage on the larger domain of news. From this, 20 hours of footage have been annotated by Deaf experts and interpreters and is made publicly available for research purposes. In this paper, we share the dataset collection process and tools developed to enable the alignment of sign language video and subtitles, as well as baseline translation results to underpin future research. | ['Richard Bowden', 'Robin Nachtrab-Ribback', 'Giacomo Inches', 'Marco Giovanelli', 'Guillaume Rochette', 'Ben Saunders', 'Necati Cihan Camgoz'] | 2021-05-05 | null | null | null | null | ['sign-language-translation'] | ['computer-vision'] | [ 2.36399144e-01 3.53968859e-01 -5.27506769e-01 -4.91403013e-01
-6.93417430e-01 -7.58688569e-01 7.19560027e-01 -1.77243769e-01
-4.90326583e-01 8.42088878e-01 1.22877419e+00 -4.00638074e-01
2.50881016e-01 -4.14576769e-01 -4.51082110e-01 2.28216778e-02
9.99323577e-02 3.01356375e-01 3.33900213e-01 -3.77547950e-01
2.92199433e-01 -1.49208218e-01 -1.80186999e+00 6.02916121e-01
5.75858474e-01 5.15726388e-01 -9.73574594e-02 6.09479845e-01
-8.84878486e-02 1.04732788e+00 -5.93286097e-01 -5.30604362e-01
3.19303691e-01 -5.49222767e-01 -7.84905255e-01 -2.97744542e-01
1.04428017e+00 -9.35796082e-01 -3.27501714e-01 6.24948680e-01
9.79233205e-01 -1.83546051e-01 2.23794103e-01 -1.41238296e+00
-8.47774208e-01 6.35523558e-01 9.23723578e-02 7.41717815e-02
8.74523878e-01 3.99049670e-01 1.06135893e+00 -5.08853078e-01
1.50566196e+00 9.96033907e-01 6.40342355e-01 7.83159912e-01
-5.20721555e-01 -6.74668431e-01 5.04534654e-02 2.41891444e-01
-8.83554220e-01 -8.24619889e-01 5.48503518e-01 -5.69337606e-01
9.52461600e-01 2.31579065e-01 1.27818155e+00 1.52855051e+00
-5.30450463e-01 1.03292871e+00 1.36336505e+00 -5.14647245e-01
-8.29615816e-02 -2.15371653e-01 1.20187260e-01 5.12912989e-01
2.93537617e-01 1.41793668e-01 -9.59756851e-01 6.70493469e-02
7.50229836e-01 -9.35217381e-01 -2.40507677e-01 -5.24763111e-03
-1.38384020e+00 4.71096039e-01 -5.50371408e-02 3.03384691e-01
-2.79919021e-02 -1.28600942e-02 5.18270314e-01 6.25474751e-01
4.09816414e-01 2.19465703e-01 -1.36655897e-01 -9.58021522e-01
-8.38580847e-01 8.67260993e-01 1.11534977e+00 1.01780474e+00
4.35778648e-02 -2.69246250e-01 -7.29443952e-02 5.24388313e-01
5.22810936e-01 5.77368557e-01 1.89422131e-01 -1.20302546e+00
8.02246630e-01 6.29305542e-01 1.92527339e-01 -9.08394873e-01
-3.57921362e-01 1.78062350e-01 6.37693405e-02 2.14191914e-01
8.88198853e-01 -3.69357079e-01 -9.80080664e-01 1.59432542e+00
2.71660835e-01 1.73138663e-01 -4.32233587e-02 1.23944628e+00
1.32282197e+00 2.21751064e-01 1.48817286e-01 1.69649929e-01
1.29162025e+00 -7.99584985e-01 -6.24303937e-01 -2.68964142e-01
6.57706022e-01 -9.70380783e-01 1.26895118e+00 2.00838298e-01
-1.14931083e+00 1.00102976e-01 -7.21414328e-01 -4.03631508e-01
-3.47774088e-01 -1.46900669e-01 5.18547833e-01 6.70638859e-01
-1.06295109e+00 -3.74292322e-02 -8.15728545e-01 -8.93298388e-01
4.82282013e-01 -3.20733160e-01 -4.58434790e-01 -2.53614575e-01
-1.11413836e+00 1.28163779e+00 6.59339949e-02 -8.12694579e-02
-3.05991173e-01 -5.63500941e-01 -7.56288528e-01 -9.05071080e-01
1.81699112e-01 -6.26869977e-01 1.66986656e+00 -9.65951443e-01
-1.48797619e+00 1.33794475e+00 -7.40961358e-02 -4.06049937e-01
9.94598389e-01 -3.90442550e-01 -4.13236648e-01 4.56987247e-02
2.75014073e-01 6.36406183e-01 5.17979503e-01 -7.44926691e-01
-8.92231703e-01 -7.07056746e-02 3.63355905e-01 2.79613674e-01
-2.08549544e-01 8.56476963e-01 -4.35411096e-01 -8.27092946e-01
-3.20465267e-01 -9.00215805e-01 2.34021857e-01 1.50724828e-01
1.19764209e-02 1.57550409e-01 7.57076919e-01 -1.45832348e+00
1.33767092e+00 -1.96998549e+00 1.38712019e-01 -2.71307305e-02
1.29788518e-01 2.77057946e-01 -1.91346452e-01 5.98693132e-01
3.29724640e-01 1.57985583e-01 -6.47714734e-02 -9.93178487e-02
1.52816966e-01 2.50147402e-01 -3.63872319e-01 5.39863884e-01
-1.22656725e-01 1.03192067e+00 -1.11723936e+00 -6.04669988e-01
1.62928298e-01 4.14240748e-01 -6.85106158e-01 -1.83405161e-01
-3.36202204e-01 7.08633602e-01 -3.45057219e-01 9.31271076e-01
6.26652017e-02 9.50371921e-02 9.17502418e-02 -2.83524711e-02
-5.22160470e-01 7.55896330e-01 -9.33027625e-01 1.76909161e+00
-1.55054122e-01 1.40122950e+00 1.36726990e-01 -3.49315166e-01
4.34935182e-01 5.65318048e-01 5.05128920e-01 -6.98317766e-01
1.62520871e-01 3.45849514e-01 7.29764551e-02 -9.52915907e-01
5.45468986e-01 1.97855055e-01 -6.53343648e-02 7.46425450e-01
-4.02474135e-01 -1.53571501e-01 6.38769209e-01 -3.62758115e-02
1.17829823e+00 4.88406867e-01 1.98886350e-01 2.81501025e-01
-1.08280024e-02 6.46035910e-01 3.60682428e-01 3.88682395e-01
-5.16299605e-01 5.87607801e-01 1.84490010e-01 -3.28215450e-01
-1.47555614e+00 -8.40205669e-01 -1.31623194e-01 1.21687829e+00
-1.20137356e-01 -7.24645674e-01 -7.68971324e-01 -1.67422712e-01
4.86174189e-02 5.36406994e-01 -2.72653610e-01 6.06083930e-01
-8.95273924e-01 -2.37798169e-01 1.07752907e+00 6.31594658e-01
6.13342166e-01 -1.18024635e+00 -1.17350018e+00 1.06434681e-01
-5.27646482e-01 -1.15797961e+00 -2.25536436e-01 -9.58857656e-01
-3.00904363e-01 -1.25575542e+00 -8.05530369e-01 -8.17563534e-01
3.30346197e-01 -1.25585869e-01 9.19398010e-01 1.25567541e-01
-2.16555133e-01 5.84248960e-01 -7.12662041e-01 -5.48304558e-01
-6.18520021e-01 5.69176786e-02 -1.15172841e-01 -5.55783749e-01
4.32870328e-01 -3.90560329e-01 -2.60907203e-01 2.13471755e-01
-7.77285695e-01 4.59039181e-01 3.96395087e-01 6.38447762e-01
-1.87473763e-02 -9.12263572e-01 3.10531288e-01 -3.39994669e-01
9.68855917e-01 -2.54605174e-01 -4.99936432e-01 1.46153778e-01
-2.91834682e-01 -4.54180807e-01 -1.62841961e-01 -2.90993690e-01
-9.52476382e-01 -1.62338898e-01 -4.15171310e-02 3.25792789e-01
-2.55655438e-01 9.14403141e-01 2.65528411e-01 7.26944730e-02
6.20524585e-01 -2.57251829e-01 1.85975544e-02 -5.32935619e-01
6.45021379e-01 1.15879142e+00 8.53670299e-01 -5.49940348e-01
6.73845887e-01 3.47630501e-01 -5.08360684e-01 -9.97272015e-01
-4.71853346e-01 -1.90915674e-01 -5.71667194e-01 -7.52678156e-01
7.26983547e-01 -9.55003142e-01 -2.69767672e-01 7.43257582e-01
-9.25172806e-01 -6.69064045e-01 -2.92623550e-01 5.16872466e-01
-6.63663208e-01 1.18186802e-01 -5.41192412e-01 -6.04309976e-01
-6.33762870e-03 -7.00167656e-01 8.85036826e-01 1.55472159e-01
-8.33978534e-01 -5.98123968e-01 2.98936278e-01 7.91828513e-01
4.57035899e-01 6.89633787e-01 3.25524807e-01 -2.01122373e-01
-5.66868544e-01 -2.72215754e-01 -2.63055176e-01 5.39877824e-02
2.15772483e-02 4.14960802e-01 -8.18045378e-01 -1.65101856e-01
-7.04378068e-01 -5.70551515e-01 5.15417635e-01 3.67799848e-01
1.30172729e-01 -4.33343261e-01 -2.33022660e-01 3.15731823e-01
8.21642995e-01 -1.17163897e-01 6.21451318e-01 7.87367642e-01
5.21863520e-01 7.96567261e-01 6.28785670e-01 4.16454643e-01
1.05883479e+00 7.38787234e-01 -1.31221607e-01 2.08694592e-01
-6.34883225e-01 -6.61759436e-01 3.79263937e-01 9.61200237e-01
-7.75345087e-01 -1.52404532e-01 -1.35930848e+00 1.04912639e+00
-1.95170617e+00 -1.30349147e+00 5.06629907e-02 1.60233915e+00
9.49420869e-01 -4.86071333e-02 3.75589073e-01 -4.07679453e-02
5.33040702e-01 3.81778151e-01 -2.93516964e-01 -7.49389604e-02
-3.84695202e-01 -1.06124245e-01 5.20380795e-01 4.60923105e-01
-1.06611860e+00 1.09914231e+00 7.11553049e+00 -2.57073622e-02
-1.10270715e+00 1.21976592e-01 -9.32770520e-02 -4.05798733e-01
-2.64859408e-01 1.75795481e-01 -4.92015898e-01 4.47288305e-01
7.66418457e-01 -3.30126405e-01 5.67651510e-01 5.73926091e-01
5.99962771e-01 -8.44123401e-03 -9.11542475e-01 8.03066850e-01
9.74975377e-02 -1.47578669e+00 -4.08558726e-01 -4.69400994e-02
7.21262634e-01 5.54960966e-01 -2.23611474e-01 6.23704456e-02
6.88899934e-01 -9.63678896e-01 1.26374745e+00 5.04116416e-01
1.10102177e+00 7.43107572e-02 5.08587241e-01 1.93815589e-01
-9.44865048e-01 -1.65934563e-01 3.74606550e-01 -7.04124689e-01
8.17458272e-01 -1.46291494e-01 -9.41328406e-01 1.88794240e-01
7.93521166e-01 1.02688777e+00 -4.79361773e-01 1.24622893e+00
-5.69525123e-01 8.44989121e-01 -7.01293170e-01 -2.35295147e-01
5.91871440e-02 5.12254834e-02 8.26102674e-01 1.18229365e+00
2.24903971e-01 8.38048160e-02 1.21999852e-01 3.36764336e-01
6.22750856e-02 2.20858306e-01 -7.44708359e-01 -4.79358882e-01
7.11175978e-01 6.33420944e-01 -4.08711553e-01 -2.32848078e-01
-7.41087437e-01 5.62967420e-01 1.40761137e-01 3.56622398e-01
-6.55419648e-01 -9.80640650e-02 8.88784230e-01 4.82584625e-01
-7.55147934e-02 -4.64127362e-01 -2.85242945e-01 -1.27376390e+00
4.28351730e-01 -1.16741943e+00 3.18867356e-01 -8.89141619e-01
-1.16067898e+00 1.48212910e-01 2.75070697e-01 -1.34617209e+00
-6.81105793e-01 -4.92093861e-01 -3.91003460e-01 5.47894597e-01
-1.36082232e+00 -1.70144522e+00 -5.46560109e-01 2.88740784e-01
1.70282662e-01 -1.10904552e-01 7.69262612e-01 5.76700389e-01
-3.00500721e-01 4.61246520e-01 -1.56181827e-01 6.19364560e-01
1.13058984e+00 -6.59525990e-01 7.40379691e-01 9.18086112e-01
9.13541913e-02 5.24799645e-01 8.42889011e-01 -9.83162463e-01
-1.12362444e+00 -6.73151791e-01 1.38402855e+00 -9.48293030e-01
1.02848732e+00 -5.47801927e-02 -2.70868152e-01 9.04771328e-01
2.06145272e-01 -3.12436998e-01 4.59536433e-01 -2.41868690e-01
-2.48162866e-01 3.38968098e-01 -1.02057815e+00 7.91851819e-01
1.59399796e+00 -7.75093436e-01 -1.05841815e+00 1.31702363e-01
3.36119086e-01 -1.01262820e+00 -8.86630714e-01 2.92517066e-01
1.30491972e+00 -6.63142622e-01 7.03825891e-01 -8.90424788e-01
7.56150424e-01 -3.82926345e-01 -1.87753648e-01 -8.60652745e-01
1.93611115e-01 -7.61581898e-01 -2.89650001e-02 1.29602957e+00
4.74808574e-01 -6.49174094e-01 6.50398791e-01 8.93100381e-01
-2.68748969e-01 -3.65752220e-01 -1.14948416e+00 -4.65141416e-01
-7.36813620e-02 -8.60282242e-01 5.86947262e-01 1.08572137e+00
2.88768440e-01 -8.81103799e-02 -3.46893877e-01 -2.07239583e-01
3.65407407e-01 2.22770032e-02 1.23070812e+00 -1.09919357e+00
6.90658391e-02 -5.92788339e-01 -6.74589038e-01 -8.94190371e-01
2.40195338e-02 -8.68818164e-01 -3.17008533e-02 -2.01455069e+00
-2.36773118e-01 -2.60000139e-01 1.87167153e-01 8.03415060e-01
2.59967983e-01 5.14198422e-01 4.43354040e-01 2.27849126e-01
-2.59742588e-01 -7.88697228e-03 1.03192532e+00 -5.13893403e-02
-3.29029024e-01 -3.25858802e-01 -7.70048380e-01 8.72376025e-01
7.81490266e-01 -3.57305966e-02 -1.08130202e-01 -8.46088827e-01
2.90697157e-01 -3.77047837e-01 6.34718835e-01 -8.80865276e-01
6.34540170e-02 -3.17397892e-01 -8.45963955e-02 -4.91239190e-01
2.22563058e-01 -3.95987362e-01 1.34986222e-01 1.91885471e-01
-4.17742103e-01 2.59444248e-02 4.39318754e-02 1.00042395e-01
-2.77719975e-01 1.50529474e-01 1.53347775e-01 3.15017626e-02
-1.06373537e+00 -1.34805754e-01 -4.47753310e-01 5.67479014e-01
1.00254214e+00 -5.19525826e-01 -8.76115620e-01 -7.62634814e-01
-4.90859866e-01 3.68270844e-01 8.95864069e-01 7.61863053e-01
3.53014797e-01 -1.47234571e+00 -1.07745695e+00 -3.17309727e-03
2.71106154e-01 -1.55297562e-01 -1.98641285e-01 7.75582731e-01
-8.98176968e-01 4.39166099e-01 -7.23110259e-01 -7.93142840e-02
-1.49708629e+00 -2.10097805e-01 -8.86124671e-02 2.64364868e-01
-1.09622419e+00 6.70505583e-01 -6.88631833e-01 -3.21566880e-01
2.45893449e-01 -7.14231312e-01 -2.89171368e-01 2.08931834e-01
6.78724766e-01 4.96192753e-01 -4.34683740e-01 -1.13818264e+00
-4.05134380e-01 5.32836318e-01 5.13435364e-01 -7.05889165e-01
1.52141333e+00 -1.30862221e-01 -3.51806730e-02 4.85654145e-01
6.45911336e-01 2.81113237e-01 -1.24723721e+00 -4.65112627e-02
1.63468584e-01 -5.95221341e-01 -2.27345124e-01 -9.55164373e-01
-5.42135298e-01 3.24208200e-01 4.60922420e-01 -7.26256333e-03
8.81912112e-01 2.13999540e-01 1.05846751e+00 2.06530213e-01
3.49558562e-01 -1.44396937e+00 -5.22323728e-01 6.20721042e-01
1.04070246e+00 -1.02934873e+00 -1.07208667e-02 -2.18238473e-01
-6.38441801e-01 8.49564791e-01 3.39950949e-01 1.83151349e-01
4.83842820e-01 4.82767522e-01 6.58376515e-01 -9.31286067e-02
-5.97374320e-01 -2.84621596e-01 3.34761202e-01 8.55311751e-01
6.46237016e-01 1.61776021e-01 -7.86090016e-01 5.25550783e-01
-8.87623906e-01 7.38530695e-01 4.85090405e-01 1.36253250e+00
-2.34948441e-01 -1.13817251e+00 -3.39423418e-01 5.24641752e-01
-2.01066062e-01 -6.05726093e-02 -7.96253026e-01 8.94103885e-01
2.02789053e-01 7.77310729e-01 -3.32275070e-02 -2.38663107e-01
5.89999735e-01 2.11637750e-01 7.50070095e-01 -2.53481627e-01
-4.14042145e-01 -4.37023222e-01 1.10490084e+00 -5.33535182e-01
-1.02816975e+00 -1.03138959e+00 -1.06755817e+00 -5.06086409e-01
1.46216214e-01 -6.51009381e-01 6.36665046e-01 1.11900973e+00
2.50948638e-01 -6.35022018e-03 -4.59493548e-01 -7.94219196e-01
-1.78007141e-01 -1.00373960e+00 -1.45479357e-02 6.01564467e-01
2.59287179e-01 -6.45415485e-01 1.31845951e-01 5.32385588e-01] | [9.153910636901855, -6.471294403076172] |
5651f9d8-6404-4a54-85aa-25ba6c0a8faf | accented-text-to-speech-synthesis-with-a | 2211.03316 | null | https://arxiv.org/abs/2211.03316v1 | https://arxiv.org/pdf/2211.03316v1.pdf | Accented Text-to-Speech Synthesis with a Conditional Variational Autoencoder | Accent plays a significant role in speech communication, influencing understanding capabilities and also conveying a person's identity. This paper introduces a novel and efficient framework for accented Text-to-Speech (TTS) synthesis based on a Conditional Variational Autoencoder. It has the ability to synthesize a selected speaker's speech that is converted to any desired target accent. Our thorough experiments validate the effectiveness of our proposed framework using both objective and subjective evaluations. The results also show remarkable performance in terms of the ability to manipulate accents in the synthesized speech and provide a promising avenue for future accented TTS research. | ['Dorien Herremans', 'Berrak Sisman', 'Ambuj Mehrish', 'Jan Melechovsky'] | 2022-11-07 | null | null | null | null | ['text-to-speech-synthesis'] | ['speech'] | [-1.64755702e-01 1.92169398e-01 -1.19815491e-01 -6.81946397e-01
-5.61684132e-01 -4.22435820e-01 6.79647863e-01 -3.88811052e-01
-2.08205044e-01 1.01184309e+00 6.70582592e-01 -3.87945503e-01
3.45345736e-01 -6.10041082e-01 -3.50387961e-01 -8.89533997e-01
4.06609029e-01 3.93643647e-01 -9.08430591e-02 -6.36793435e-01
-9.37243998e-02 5.08121729e-01 -1.36056376e+00 -1.57052442e-01
7.13973522e-01 6.63217604e-01 3.52022618e-01 8.40359032e-01
-1.08217537e-01 7.98411608e-01 -1.29921556e+00 -4.93216783e-01
-2.26599440e-01 -2.95830071e-01 -6.91579819e-01 2.88413733e-01
8.71165320e-02 -8.24942216e-02 -1.76558957e-01 1.08182430e+00
6.91018641e-01 5.40986598e-01 6.37650669e-01 -9.20680761e-01
-8.57733607e-01 9.91290212e-01 2.48374045e-01 2.69648343e-01
3.82748455e-01 -7.48571903e-02 9.54052091e-01 -5.37754118e-01
3.74476373e-01 1.45557177e+00 1.30957723e-01 7.83185184e-01
-8.74947429e-01 -6.61175013e-01 9.74999368e-02 3.80475372e-02
-1.35052204e+00 -7.20003068e-01 1.02452552e+00 -2.96122521e-01
6.61593556e-01 5.50721347e-01 5.49383998e-01 1.35904849e+00
8.40346590e-02 8.26303899e-01 1.25904036e+00 -5.69787979e-01
3.58364969e-01 3.64023507e-01 3.64157446e-02 3.40841681e-01
-6.24147475e-01 2.25724369e-01 -5.63614249e-01 1.99193373e-01
6.18116319e-01 -7.15288103e-01 -4.10182327e-01 4.73353297e-01
-1.23271108e+00 1.03847110e+00 5.89025207e-02 6.05470598e-01
-4.77847010e-01 -1.53474793e-01 2.13107452e-01 1.70146927e-01
3.50363731e-01 4.02723551e-01 -2.95667291e-01 -2.90232331e-01
-8.22325826e-01 1.97404370e-01 7.04993963e-01 8.05744588e-01
1.95540845e-01 9.03702915e-01 -2.65716642e-01 1.07623124e+00
3.45681489e-01 7.49428988e-01 7.38795757e-01 -8.21653187e-01
1.45401224e-01 -2.60097049e-02 3.18409428e-02 -5.53676963e-01
-1.84213817e-01 -4.92063612e-01 -7.37974584e-01 1.86040506e-01
-4.14764658e-02 -6.44354939e-01 -1.02069330e+00 1.82623541e+00
4.60545689e-01 -3.57823893e-02 6.13465965e-01 9.28012013e-01
1.10884583e+00 1.16263270e+00 1.61882058e-01 -4.91161287e-01
1.44329298e+00 -8.73765171e-01 -1.51838028e+00 9.78285243e-05
-1.21398844e-01 -1.12638748e+00 9.45892930e-01 8.27534050e-02
-1.18379211e+00 -8.74402463e-01 -1.07259619e+00 1.15255095e-01
-3.77131492e-01 1.57350317e-01 3.64367247e-01 1.12658381e+00
-1.12890112e+00 -9.27191898e-02 -5.69612324e-01 -5.06664105e-02
-2.04121158e-01 5.40851235e-01 -8.93411785e-02 7.93138564e-01
-1.66493154e+00 8.44804704e-01 3.66529316e-01 1.27180532e-01
-5.23905814e-01 -5.06590068e-01 -9.80523765e-01 1.35173023e-01
1.73645541e-02 -6.16797507e-01 1.58491421e+00 -8.67290139e-01
-2.33065343e+00 3.65875661e-01 -4.47934121e-01 -4.92382735e-01
2.63799846e-01 1.18467130e-01 -1.10464990e+00 2.24089194e-02
-9.48533863e-02 6.57540381e-01 1.17866266e+00 -1.08538282e+00
-6.64683163e-01 -1.06517769e-01 -7.31737539e-02 4.24766928e-01
-3.62310380e-01 2.06683010e-01 -2.05462992e-01 -1.14113998e+00
-9.85631943e-02 -9.96004283e-01 -1.87165719e-02 -7.30483413e-01
-7.09112406e-01 -4.85457450e-01 9.36325192e-01 -7.25267649e-01
1.15140712e+00 -2.19586039e+00 2.74757773e-01 2.10659970e-02
-1.44726664e-01 4.94156748e-01 3.49996865e-01 3.92378420e-01
9.39357355e-02 8.45977142e-02 -3.88911776e-02 -5.02774715e-01
2.16306373e-01 2.79210061e-01 -4.24800813e-01 3.17030326e-02
1.09086074e-02 7.05876529e-01 -4.52212572e-01 -5.69655180e-01
7.45418429e-01 8.69832814e-01 -2.71194279e-01 6.20890856e-01
-1.54869258e-01 7.47629821e-01 -4.55428660e-01 3.65910381e-01
4.72175598e-01 4.52260464e-01 1.42433822e-01 -8.88263434e-02
-3.68059099e-01 3.74144286e-01 -1.09707654e+00 9.68280792e-01
-7.16423929e-01 8.59554291e-01 3.07343096e-01 -7.05538213e-01
1.16883552e+00 7.88024664e-01 -6.88939393e-02 -4.77513343e-01
2.49100253e-01 -9.17446390e-02 -4.35390789e-03 -4.39760715e-01
8.86864185e-01 -5.81412792e-01 -1.59340218e-01 7.78827146e-02
2.36673176e-01 -3.70352298e-01 6.40783310e-02 -2.27559078e-02
1.58047304e-01 -4.12381709e-01 3.20017785e-01 -4.53849226e-01
9.30520952e-01 -4.31961119e-01 3.82406682e-01 4.63757366e-01
-2.96828032e-01 3.00663203e-01 1.85491536e-02 -9.59964320e-02
-8.61219466e-01 -1.26124156e+00 -2.94112593e-01 1.11264479e+00
-1.84482545e-01 6.38894364e-02 -1.07340682e+00 -2.07748070e-01
-5.39932191e-01 1.30088735e+00 -3.77409101e-01 2.41251364e-01
-5.42940021e-01 -4.64995176e-01 6.87746227e-01 5.86842716e-01
5.20751476e-01 -1.27019477e+00 8.64639059e-02 2.62876838e-01
-5.56794822e-01 -1.24343860e+00 -7.35025108e-01 -7.61389434e-02
-4.02062446e-01 -2.60562718e-01 -1.02160287e+00 -1.34673035e+00
3.91886830e-01 -2.83077687e-01 6.52033091e-01 -4.08117831e-01
4.43699658e-01 2.03036338e-01 -2.75266021e-01 -6.58578813e-01
-1.08928216e+00 3.11519921e-01 2.69896865e-01 3.48962545e-01
1.20038956e-01 -4.21704382e-01 -2.06408903e-01 1.80293739e-01
-7.31979430e-01 -2.70802170e-01 2.53597617e-01 7.92915940e-01
3.63015115e-01 3.79430354e-01 8.58870327e-01 -5.67816317e-01
1.13036084e+00 -1.06923446e-01 -6.82044148e-01 -4.44422569e-03
-2.14575410e-01 -4.33039945e-03 8.85422111e-01 -2.82162398e-01
-1.62870395e+00 -1.36947885e-01 -9.07218218e-01 -9.98233333e-02
-4.78086114e-01 4.21411753e-01 -2.65025347e-01 1.74349815e-01
2.36247092e-01 4.86658335e-01 -5.69203347e-02 -1.60350367e-01
3.74478608e-01 1.27307642e+00 8.94991934e-01 -3.88220221e-01
7.37503767e-01 1.00228608e-01 -5.01052499e-01 -1.56222582e+00
-4.46608722e-01 -2.04998106e-01 -4.75797385e-01 -3.31511825e-01
1.17843676e+00 -9.17417586e-01 -1.05084884e+00 5.98965526e-01
-1.12368917e+00 -4.08645943e-02 1.71460174e-02 8.09912741e-01
-3.82156253e-01 1.57781288e-01 -7.55240083e-01 -9.93117034e-01
-3.54981393e-01 -1.44361544e+00 9.30348158e-01 5.84560990e-01
-3.27222943e-01 -1.36577761e+00 2.05505174e-02 7.91772008e-01
4.34520364e-01 -2.18781561e-01 8.97590458e-01 -6.68111265e-01
-2.70471841e-01 8.62861648e-02 4.77925569e-01 6.51616514e-01
2.79143363e-01 1.23736754e-01 -1.09396064e+00 -4.36082184e-02
1.65158436e-01 -8.91535208e-02 3.19296747e-01 5.44529438e-01
4.74532306e-01 -3.20865899e-01 1.55590639e-01 4.41682279e-01
8.71043444e-01 7.70605683e-01 6.36134267e-01 7.61078820e-02
5.28662741e-01 4.92927969e-01 3.16900313e-01 2.20786393e-01
2.97143519e-01 6.71361387e-01 -1.09404810e-01 -2.35174939e-01
-3.07228059e-01 -1.47058755e-01 3.44620973e-01 1.51348019e+00
4.68003191e-02 -5.55588782e-01 -5.33224523e-01 5.89863420e-01
-1.25572419e+00 -9.94353116e-01 -4.40078750e-02 1.77916098e+00
8.73036325e-01 1.23490721e-01 -5.02192304e-02 2.23953336e-01
1.15702999e+00 5.45363069e-01 -1.21837147e-01 -1.05108869e+00
-2.69662678e-01 4.91531402e-01 2.64883488e-02 1.11368680e+00
-1.09815657e+00 1.31314898e+00 7.31108236e+00 8.14125955e-01
-1.39026785e+00 -1.93370849e-01 4.34236795e-01 3.69034946e-01
-2.43661210e-01 -4.56229627e-01 -7.76183605e-01 4.97921497e-01
1.18525577e+00 -3.42422128e-01 3.73923361e-01 7.40868986e-01
4.62511927e-01 1.84470162e-01 -4.52437222e-01 6.95279419e-01
6.54638708e-02 -1.33439624e+00 1.22220464e-01 -1.94094658e-01
6.85797453e-01 -3.76103312e-01 4.79929090e-01 3.66429925e-01
3.18003923e-01 -1.02236021e+00 8.62992346e-01 1.63753211e-01
5.27880251e-01 -1.00040662e+00 7.65594959e-01 1.50214493e-01
-1.02990544e+00 3.05343568e-01 -1.08156450e-01 -2.20794510e-02
4.85898226e-01 2.66742885e-01 -1.41217756e+00 3.23020846e-01
2.47729719e-01 -7.97161181e-03 1.08305626e-02 2.48329684e-01
-4.37970161e-01 1.11612177e+00 -1.05703846e-01 -5.07517040e-01
3.33316475e-01 -2.80019104e-01 7.58881807e-01 1.32936406e+00
2.23964840e-01 2.10082099e-01 9.92954448e-02 6.63215876e-01
-2.95032263e-02 4.08419937e-01 -2.71981567e-01 -2.99168140e-01
7.13913500e-01 7.78795183e-01 -4.53842342e-01 -5.54785192e-01
-2.44041950e-01 9.67010200e-01 -1.83249936e-01 6.36857629e-01
-8.08084905e-01 -5.34987390e-01 7.22457826e-01 -3.82225245e-01
5.73810041e-01 -2.95850515e-01 -1.52367637e-01 -8.51748288e-01
-1.45253628e-01 -1.01087654e+00 6.31057248e-02 -7.95183241e-01
-9.48828876e-01 1.14133322e+00 -2.23563552e-01 -7.38424897e-01
-6.16502166e-01 -3.97510916e-01 -6.40803754e-01 1.14412689e+00
-1.27844191e+00 -1.18747938e+00 1.70973152e-01 8.03711295e-01
9.77340639e-01 -5.85819244e-01 1.04714298e+00 1.05476297e-01
-6.61880434e-01 7.50180423e-01 3.73054557e-02 2.48713151e-01
4.27687526e-01 -1.47063720e+00 5.21669924e-01 8.70572269e-01
2.10810378e-01 7.16325283e-01 1.18809867e+00 -3.76539767e-01
-1.00492644e+00 -8.44169021e-01 1.14110637e+00 3.34332101e-02
5.28199494e-01 -3.83738190e-01 -5.96608818e-01 7.64544308e-01
7.93083906e-01 -6.14982963e-01 8.79514039e-01 2.56496847e-01
2.21605018e-01 -1.31659687e-01 -1.03683817e+00 9.34831738e-01
3.67890090e-01 -7.12121487e-01 -9.99813497e-01 -9.09871981e-03
9.65763927e-01 -5.36685586e-01 -1.04153156e+00 9.09077898e-02
3.62679362e-01 -9.55539763e-01 7.82508969e-01 -3.06624800e-01
7.05590844e-02 -3.73991519e-01 -2.77047426e-01 -1.70366442e+00
-3.48118156e-01 -9.31252956e-01 2.48826057e-01 1.52311575e+00
5.18191874e-01 -6.05492890e-01 6.61195099e-01 3.69696766e-01
-3.59633535e-01 -3.38711530e-01 -9.12872314e-01 -5.45346141e-01
2.41985470e-02 -5.08513093e-01 8.95101368e-01 9.03931022e-01
-3.62401456e-02 6.11524642e-01 -6.00262105e-01 6.88316345e-01
3.17042738e-01 -1.07596360e-01 6.13183796e-01 -7.17295110e-01
-1.78171039e-01 -2.33131588e-01 -4.91849720e-01 -1.09144390e+00
5.20233870e-01 -5.97873271e-01 9.96980667e-02 -1.18123651e+00
-6.46853268e-01 3.25678363e-02 -2.51223475e-01 -4.20715809e-02
-3.01907152e-01 6.26334921e-02 1.52501076e-01 -3.64148885e-01
5.33144698e-02 9.64484453e-01 1.32596219e+00 -1.83395401e-01
-4.34200048e-01 3.63237947e-01 -5.08952618e-01 5.71897864e-01
1.12074375e+00 -1.34923756e-01 -6.29704177e-01 -1.66951373e-01
-6.99908495e-01 4.15240616e-01 -2.20074683e-01 -9.71360385e-01
1.83332995e-01 -3.55926782e-01 2.08533242e-01 -4.63784516e-01
7.49232829e-01 -4.17529374e-01 -6.99016005e-02 1.76896840e-01
-5.06235063e-01 -1.18519783e-01 1.56747222e-01 1.99819967e-01
-6.67783022e-01 -1.49778679e-01 9.40541625e-01 1.16686322e-01
-8.00250351e-01 -6.18994310e-02 -9.79628682e-01 -1.13803826e-01
8.62421572e-01 -6.87823594e-02 4.15835716e-02 -8.55464578e-01
-9.45907235e-01 -1.45392969e-01 -1.07008800e-01 7.08570838e-01
4.96407777e-01 -1.18199670e+00 -7.28068292e-01 4.92081434e-01
-1.77701429e-01 -5.23363531e-01 2.97360033e-01 1.91072285e-01
-4.28274393e-01 7.22228527e-01 -1.28580108e-01 -4.62126195e-01
-1.29687464e+00 3.36049825e-01 3.10576588e-01 2.71312207e-01
-3.81943345e-01 8.68900657e-01 2.60472685e-01 -4.52825725e-01
3.58954042e-01 -1.50365919e-01 -4.74414676e-01 -2.11675346e-01
4.86335933e-01 2.66983807e-01 -8.47417191e-02 -1.15928996e+00
-2.19877779e-01 1.74213722e-01 -1.14723332e-01 -6.84782684e-01
8.06068480e-01 -3.89062941e-01 1.16523348e-01 6.56015694e-01
1.04311204e+00 5.06019950e-01 -8.65511060e-01 2.01664761e-01
-4.24466282e-01 -1.58525139e-01 4.66764152e-01 -8.25860381e-01
-1.08888149e+00 6.37546659e-01 4.64282215e-01 3.19072545e-01
8.67872179e-01 -7.21387193e-02 1.07617617e+00 3.22259784e-01
-3.00872233e-02 -1.46604991e+00 -2.71578461e-01 4.36871648e-01
8.75529647e-01 -1.08789659e+00 -6.12393558e-01 -5.46676159e-01
-1.07633877e+00 1.07368731e+00 2.69574255e-01 2.65420645e-01
7.03625083e-01 3.54357690e-01 1.01820827e+00 2.12954968e-01
-5.46249509e-01 -1.44293174e-01 3.02245170e-01 9.74916756e-01
7.40808129e-01 5.20435989e-01 -2.42174357e-01 3.58585477e-01
-1.00471950e+00 -5.37750959e-01 5.86526930e-01 5.07007718e-01
-4.27014798e-01 -1.19623613e+00 -6.91235185e-01 -1.42511681e-01
-6.07347310e-01 -7.72873238e-02 -2.87742972e-01 6.01838291e-01
-7.37678334e-02 1.46040606e+00 4.50408692e-03 -3.66977334e-01
1.54306009e-01 3.00782830e-01 7.01358393e-02 -3.32491904e-01
-4.77922946e-01 4.66232568e-01 2.70653754e-01 8.17474797e-02
-4.56504762e-01 -6.63920105e-01 -1.21250272e+00 -3.23839605e-01
-2.26165101e-01 6.55563235e-01 9.47764993e-01 1.12877238e+00
-4.55969200e-02 8.65633845e-01 9.28943455e-01 -2.73633599e-01
-3.30969095e-01 -1.04554820e+00 -8.01963806e-01 1.85758948e-01
5.74097991e-01 -4.19301122e-01 -3.02622020e-01 2.52553195e-01] | [14.769609451293945, 6.594698429107666] |
ff4f4ec2-5c6e-4aa0-8162-5e0159f6c4ad | deep-single-image-camera-calibration-by | 2303.17166 | null | https://arxiv.org/abs/2303.17166v1 | https://arxiv.org/pdf/2303.17166v1.pdf | Deep Single Image Camera Calibration by Heatmap Regression to Recover Fisheye Images Under ManhattanWorld AssumptionWithout Ambiguity | In orthogonal world coordinates, a Manhattan world lying along cuboid buildings is widely useful for various computer vision tasks. However, the Manhattan world has much room for improvement because the origin of pan angles from an image is arbitrary, that is, four-fold rotational symmetric ambiguity of pan angles. To address this problem, we propose a definition for the pan-angle origin based on the directions of the roads with respect to a camera and the direction of travel. We propose a learning-based calibration method that uses heatmap regression to remove the ambiguity by each direction of labeled image coordinates, similar to pose estimation keypoints. Simultaneously, our two-branched network recovers the rotation and removes fisheye distortion from a general scene image. To alleviate the lack of vanishing points in images, we introduce auxiliary diagonal points that have the optimal 3D arrangement of spatial uniformity. Extensive experiments demonstrated that our method outperforms conventional methods on large-scale datasets and with off-the-shelf cameras. | ['Takayoshi Yamashita', 'Yasunori Ishii', 'Satoshi Sato', 'Nobuhiko Wakai'] | 2023-03-30 | null | null | null | null | ['camera-calibration'] | ['computer-vision'] | [-1.31858364e-01 -1.77700907e-01 -2.63386220e-01 -4.70831633e-01
-3.94810051e-01 -5.92448890e-01 3.18399370e-01 -5.95905721e-01
-3.97403359e-01 3.06239456e-01 -1.00641385e-01 -2.43819699e-01
-4.82538119e-02 -8.05707157e-01 -8.66983235e-01 -6.33124590e-01
4.63822573e-01 3.65634292e-01 2.87448943e-01 -3.30536574e-01
5.56796193e-01 6.95151269e-01 -9.21931207e-01 -4.51887608e-01
6.21305466e-01 6.21408641e-01 2.12216079e-01 2.24794090e-01
2.83635229e-01 8.76707509e-02 -2.86730051e-01 -3.67412865e-01
7.52111435e-01 1.47479493e-02 -2.87336230e-01 4.20267493e-01
9.21407104e-01 -6.14799023e-01 -5.98515034e-01 1.17837143e+00
9.46761221e-02 -7.53599927e-02 4.50737715e-01 -1.40573561e+00
-7.33681083e-01 -3.14135291e-02 -1.15901196e+00 -2.72410005e-01
1.55184925e-01 -1.90587685e-01 7.93136299e-01 -7.41344750e-01
6.23148501e-01 1.06953013e+00 6.90849245e-01 -8.03484097e-02
-7.72689581e-01 -7.21881866e-01 2.89089471e-01 9.48539153e-02
-1.79263783e+00 2.53781658e-02 1.04545796e+00 -2.63833404e-01
3.09163690e-01 1.07928604e-01 7.64380991e-01 6.56465530e-01
1.89235166e-01 2.94594049e-01 7.28038847e-01 -4.99918491e-01
-6.70429766e-02 -8.40250775e-03 -1.09364592e-01 7.81450510e-01
6.42525911e-01 -1.65360689e-01 -1.23660639e-01 1.22401826e-01
1.48995900e+00 4.68772799e-01 -2.66588151e-01 -1.33566558e+00
-1.52952600e+00 8.68545175e-01 7.48364627e-01 -2.66075462e-01
8.10494423e-02 1.17997944e-01 -3.07343483e-01 -1.88902393e-01
7.78285190e-02 4.55564797e-01 -6.08144403e-01 2.38419443e-01
-5.24437606e-01 1.47201061e-01 4.29902375e-01 1.63045180e+00
1.15415823e+00 -3.02871410e-02 6.55089319e-01 7.55410194e-01
3.78836751e-01 9.63955998e-01 1.13831557e-01 -1.20637262e+00
7.93153226e-01 6.96017623e-01 1.90780148e-01 -1.48353028e+00
-6.16020620e-01 -2.50313520e-01 -9.65178549e-01 -8.34943652e-02
5.84053218e-01 -2.82378376e-01 -6.99060977e-01 1.35567522e+00
4.42943126e-01 1.28058597e-01 -3.55983496e-01 1.13348889e+00
1.58527553e-01 5.70929348e-01 -7.36372650e-01 5.19064218e-02
1.33572125e+00 -1.01825440e+00 -3.17037642e-01 -5.53131461e-01
2.67164886e-01 -8.77166152e-01 1.02014792e+00 3.72320056e-01
-4.97308999e-01 -3.89360905e-01 -1.23563123e+00 -1.61667630e-01
-2.24438339e-01 3.79088968e-01 5.05676448e-01 6.06142640e-01
-8.09099793e-01 4.07306403e-02 -6.72415555e-01 -5.17167747e-01
-1.85854182e-01 1.70887172e-01 -4.57991630e-01 -2.03368217e-01
-7.87984490e-01 1.01833141e+00 1.55290276e-01 2.73207039e-01
-2.41449267e-01 -4.07825947e-01 -9.55454171e-01 -1.36581987e-01
6.83861732e-01 -4.74835753e-01 1.01733303e+00 -5.94639540e-01
-1.45217991e+00 5.81986547e-01 -2.78960615e-02 3.39362696e-02
5.96629798e-01 -5.67055404e-01 -2.98885942e-01 -1.71128809e-02
2.49733105e-01 6.88858628e-01 7.94495165e-01 -1.35087883e+00
-7.01053202e-01 -6.38307214e-01 8.25289562e-02 4.86521572e-01
-5.34223989e-02 -4.21643883e-01 -1.03298736e+00 -4.14785862e-01
1.03810048e+00 -1.61538959e+00 -4.91517574e-01 2.86929339e-01
-6.03111207e-01 2.90450126e-01 9.47713852e-01 -3.78423184e-01
9.07111704e-01 -2.12452245e+00 -1.73693985e-01 5.40271342e-01
-6.10500854e-03 -1.71263382e-01 6.47227019e-02 2.67676681e-01
-8.99559259e-02 -1.56812549e-01 -5.23299985e-02 9.41296965e-02
-3.60040396e-01 2.98909724e-01 -2.95400292e-01 8.92714202e-01
-1.64812446e-01 3.73034984e-01 -8.79939258e-01 -3.54224861e-01
4.97463912e-01 4.46019471e-01 -4.75598276e-01 -1.02215363e-02
3.95451844e-01 1.68794721e-01 -5.69873631e-01 4.95649397e-01
1.11252379e+00 -7.21365288e-02 1.23404622e-01 -4.72611636e-01
-4.08197373e-01 -8.20845813e-02 -1.53893566e+00 1.77991700e+00
-3.37863743e-01 6.89669907e-01 -2.83018976e-01 -5.67671716e-01
1.15294230e+00 -1.56126171e-01 3.86119425e-01 -4.43769068e-01
7.75860995e-02 -5.80216609e-02 -2.87669480e-01 -2.48760283e-01
6.62794650e-01 3.62565398e-01 -7.46856779e-02 1.83204994e-01
-4.00607020e-01 -6.61679626e-01 -1.22685879e-01 -2.86042392e-02
4.64084655e-01 2.59831369e-01 6.14290655e-01 -7.61453286e-02
3.21855903e-01 3.57656702e-02 6.97818398e-01 3.40336412e-01
3.82458675e-03 1.21835017e+00 3.40582043e-01 -9.12401378e-01
-1.30524004e+00 -9.13466156e-01 -8.77598524e-02 5.21463811e-01
7.06429899e-01 -4.32781428e-01 -8.25502038e-01 -4.01811451e-01
-5.04417904e-02 3.11841905e-01 -3.64894807e-01 7.27328435e-02
-8.25463355e-01 -5.15914023e-01 7.51652196e-02 4.13027406e-01
8.81775916e-01 -2.76320368e-01 -6.71104312e-01 -2.34669209e-01
-4.25805390e-01 -1.20894492e+00 -9.34275031e-01 -5.16540296e-02
-7.09028363e-01 -1.45613503e+00 -6.01410687e-01 -7.13543296e-01
1.11747932e+00 1.34773088e+00 6.79073453e-01 -1.76153168e-01
8.34975317e-02 2.07219318e-01 -4.04561684e-02 -2.32983992e-01
4.34562027e-01 1.02068007e-01 7.64175877e-02 -1.80743769e-01
4.04518932e-01 -4.76320893e-01 -7.74599373e-01 8.80145788e-01
-4.46231753e-01 3.41724187e-01 7.26375580e-01 5.41331589e-01
6.29923463e-01 1.90516207e-02 -2.92797148e-01 -6.43283844e-01
3.09724914e-04 -2.73307115e-01 -1.15043926e+00 2.02574626e-01
-6.76487923e-01 3.98377813e-02 5.01150608e-01 -1.69951171e-01
-8.02885354e-01 6.38192654e-01 4.55046982e-01 -2.32096553e-01
-2.66500749e-02 6.55711889e-02 -3.98088396e-01 -2.65999347e-01
4.09807682e-01 -7.04916120e-02 -2.28839740e-01 -1.97072968e-01
5.80367208e-01 5.30903339e-01 5.31445503e-01 -4.31175351e-01
1.20242345e+00 7.26300240e-01 3.38138103e-01 -1.09332085e+00
-7.99710155e-01 -6.30845487e-01 -1.26636553e+00 -1.06558315e-01
8.34912419e-01 -1.09242606e+00 -5.03215790e-01 5.41129351e-01
-1.36626506e+00 2.98105717e-01 3.40508997e-01 6.46275878e-01
-4.23492342e-01 4.45012271e-01 -1.01559795e-01 -3.83998543e-01
6.56151697e-02 -1.19581139e+00 1.01912761e+00 4.50159401e-01
5.87713048e-02 -5.37836611e-01 2.55492747e-01 2.76582778e-01
-1.08769722e-01 4.11555283e-02 6.71303630e-01 8.45523849e-02
-1.19256270e+00 -2.28710726e-01 -4.10562664e-01 1.47515327e-01
1.14690438e-01 3.82018000e-01 -6.25532269e-01 -7.15279281e-02
3.79062467e-03 3.47227931e-01 3.00378889e-01 5.04949689e-01
9.99467254e-01 -1.58143789e-01 -3.62207264e-01 1.22456825e+00
1.44822657e+00 4.07027125e-01 5.39299428e-01 9.30070341e-01
1.12542117e+00 4.91347194e-01 6.05440617e-01 2.84712523e-01
7.02161491e-01 8.04923594e-01 5.14125824e-01 -1.69562027e-01
3.58412087e-01 -6.32843316e-01 -5.84459417e-02 7.07889676e-01
-1.90499976e-01 1.11028984e-01 -8.78220916e-01 3.09217066e-01
-1.84150767e+00 -5.41769445e-01 -2.99493641e-01 2.45521212e+00
2.98614621e-01 1.52898803e-01 -2.20577776e-01 -2.61329025e-01
9.08281505e-01 3.49106222e-01 -5.82367718e-01 5.15093841e-02
-6.65081516e-02 -6.89734697e-01 1.27127945e+00 4.39170837e-01
-1.14430094e+00 7.63687491e-01 6.33905077e+00 2.90729493e-01
-1.21050560e+00 -2.60188580e-01 3.33240598e-01 2.76931882e-01
-6.17167689e-02 3.59854549e-01 -1.02299523e+00 3.08360994e-01
2.22473234e-01 3.23961467e-01 5.45668542e-01 1.24147952e+00
1.38225436e-01 -9.98343080e-02 -7.69072473e-01 1.16929126e+00
3.51159662e-01 -1.06198442e+00 -6.05217554e-02 2.05500543e-01
9.73241746e-01 1.92159876e-01 8.25402215e-02 -1.65780187e-01
3.42878878e-01 -3.76159698e-01 7.21490860e-01 1.11402251e-01
7.94708312e-01 -7.85192490e-01 4.18211132e-01 4.27990198e-01
-1.14852047e+00 3.20277102e-02 -6.83375657e-01 -1.14658318e-01
1.03469975e-01 2.73483992e-01 -1.00575268e+00 4.49718922e-01
6.73352718e-01 6.92535579e-01 -7.31274843e-01 1.03591180e+00
-6.77456439e-01 3.02043706e-02 -5.23376405e-01 2.09847361e-01
2.09873915e-01 -9.04317200e-01 2.72768110e-01 8.03789198e-01
7.10953474e-01 6.10707551e-02 1.07118115e-01 3.60625774e-01
3.31726186e-02 -8.33181068e-02 -9.28167999e-01 8.64551008e-01
7.56682098e-01 1.35904074e+00 -8.83300602e-01 6.22866340e-02
-6.13349140e-01 7.85624266e-01 1.16480879e-01 5.12260735e-01
-9.32213426e-01 -5.16178250e-01 4.30122346e-01 3.43665987e-01
3.00024062e-01 -7.64269114e-01 -2.65627474e-01 -1.36027336e+00
5.39690703e-02 -5.94286025e-01 1.78767983e-02 -1.24441600e+00
-6.82618976e-01 3.59916419e-01 2.58035541e-01 -1.63180816e+00
8.59474316e-02 -7.23343611e-01 -5.27943432e-01 3.84662122e-01
-1.37732947e+00 -1.29113412e+00 -6.78437531e-01 5.19439638e-01
5.10258794e-01 1.04281180e-01 4.33506429e-01 8.51071700e-02
-5.72494507e-01 2.49718100e-01 3.37488800e-01 3.32433969e-01
1.06966150e+00 -1.16163254e+00 5.34589529e-01 1.07540047e+00
3.17407012e-01 8.53645027e-01 4.90926683e-01 -3.41227382e-01
-1.38435698e+00 -9.75933611e-01 5.68204463e-01 -4.41410005e-01
4.37437445e-01 -3.77771258e-01 -3.98928046e-01 1.09925210e+00
4.43422049e-02 -7.22947791e-02 1.58857167e-01 5.79984076e-02
-5.28896451e-01 -6.21044815e-01 -7.00234473e-01 8.34192038e-01
9.68436360e-01 -2.25683793e-01 -4.62694496e-01 2.67187566e-01
6.51536107e-01 -6.30315661e-01 -2.47282192e-01 8.95982087e-02
7.10856199e-01 -9.33050334e-01 1.27769113e+00 1.07548358e-02
1.89782470e-01 -7.86695302e-01 -2.63533592e-01 -1.34919226e+00
-4.34843987e-01 -5.70397854e-01 4.87441480e-01 9.15382981e-01
8.56904835e-02 -6.55924976e-01 8.74672771e-01 5.86909354e-01
-4.79459316e-02 -4.10399735e-01 -8.06948125e-01 -3.53222251e-01
-2.87164509e-01 -1.04108512e-01 8.68530452e-01 8.88546288e-01
-2.47736067e-01 5.95953524e-01 -7.26840436e-01 5.66572368e-01
6.96720898e-01 1.26105800e-01 1.37575185e+00 -1.21545291e+00
2.83705205e-01 -1.19246803e-01 -4.61110115e-01 -1.60437155e+00
-1.86967835e-01 -1.82581156e-01 2.96697952e-02 -1.23102915e+00
3.65792327e-02 -5.36519229e-01 4.33350680e-03 1.32617220e-01
4.94579971e-02 3.30056101e-02 1.04996055e-01 4.57033128e-01
-4.82469290e-01 3.29940081e-01 1.38318241e+00 4.35361639e-02
-4.88414876e-02 1.67150125e-01 -7.06652820e-01 1.19236779e+00
6.16123199e-01 -4.22157317e-01 -6.06523216e-01 -8.90457153e-01
4.49292600e-01 -2.96119917e-02 1.22525454e-01 -9.56644058e-01
4.49762404e-01 -6.25809312e-01 5.67741334e-01 -1.09023547e+00
3.88838738e-01 -1.27091801e+00 2.42774352e-01 2.09734723e-01
1.17846295e-01 5.72328568e-01 -3.84342611e-01 5.67874849e-01
2.49258038e-02 -1.99906394e-01 7.65802741e-01 -2.05896482e-01
-6.43666029e-01 2.35147193e-01 -3.67881581e-02 -2.70010412e-01
1.09572279e+00 -4.42468464e-01 -5.60239673e-01 -4.63889033e-01
1.86807409e-01 2.09686354e-01 9.73605514e-01 3.80738527e-01
6.84075415e-01 -1.33856320e+00 -2.46306345e-01 5.62115312e-01
1.98768094e-01 6.33342743e-01 2.04546079e-02 6.09762430e-01
-1.08249879e+00 6.67014480e-01 -3.36231947e-01 -8.37128043e-01
-1.13365364e+00 6.05228186e-01 2.25641176e-01 1.81459531e-01
-6.37280583e-01 3.50996166e-01 4.30785030e-01 -7.57444143e-01
-2.00550854e-02 -6.52857900e-01 -9.36245397e-02 -2.82325447e-01
2.52195477e-01 5.27736425e-01 -4.70748059e-02 -8.01164329e-01
-2.59287834e-01 1.37052977e+00 9.24111530e-03 -1.24750338e-01
1.23446345e+00 -5.28442800e-01 -2.58958759e-03 2.37942725e-01
1.06223607e+00 2.97229409e-01 -1.65103376e+00 -9.08452496e-02
-2.25085229e-01 -9.03841853e-01 -1.72232583e-01 -2.14122653e-01
-1.24068952e+00 7.96521008e-01 5.15619278e-01 -1.55122146e-01
7.81694889e-01 -4.74061847e-01 5.72096705e-01 8.89741361e-01
5.26544333e-01 -1.13540769e+00 -1.58513382e-01 5.14776647e-01
6.53617263e-01 -1.32735920e+00 4.31584179e-01 -6.84017599e-01
-5.82877636e-01 1.45016742e+00 8.58054519e-01 -3.00285488e-01
6.81840837e-01 -1.88729271e-01 3.81972313e-01 -1.99944023e-02
5.29280417e-02 3.62938344e-01 1.30264029e-01 7.82476425e-01
-4.30496149e-02 5.86097278e-02 4.78970893e-02 1.09516338e-01
-3.81718099e-01 -4.03523058e-01 8.27680111e-01 8.02043796e-01
-4.69021827e-01 -8.38344336e-01 -7.91406274e-01 -2.05949038e-01
1.24157257e-01 1.61537215e-01 -1.13084558e-02 1.21266365e+00
1.48885936e-01 7.17556417e-01 3.17152530e-01 -2.86471695e-01
5.82045257e-01 -4.96939927e-01 3.87895197e-01 -3.36799532e-01
3.64473969e-01 3.19434226e-01 -2.43076116e-01 -4.75616872e-01
-3.71774256e-01 -6.14967048e-01 -1.03507638e+00 -2.38742843e-01
-5.47305763e-01 -1.84503552e-02 9.19908166e-01 7.14740932e-01
2.46310323e-01 -1.50902480e-01 9.76862133e-01 -8.42369735e-01
-3.82363468e-01 -6.58945739e-01 -5.84393442e-01 9.37400535e-02
2.94065267e-01 -7.93663383e-01 -3.90123844e-01 2.16587469e-01] | [8.033538818359375, -2.28489351272583] |
3c730c26-b705-4812-82c8-492f446e0169 | recycle-your-wav2vec2-codebook-a-speech | null | null | https://aclanthology.org/2022.coling-1.626 | https://aclanthology.org/2022.coling-1.626.pdf | Recycle Your Wav2Vec2 Codebook: A Speech Perceiver for Keyword Spotting | Speech information in a pretrained wav2vec2.0 model is usually leveraged through its encoder, which has at least 95M parameters, being not so suitable for small footprint Keyword Spotting. In this work, we show an efficient way of profiting from wav2vec2.0’s linguistic knowledge, by recycling the phonetic information encoded in its latent codebook, which has been typically thrown away after pretraining. We do so by transferring the codebook as weights for the latent bottleneck of a Keyword Spotting Perceiver, thus initializing such model with phonetic embeddings already. The Perceiver design relies on cross-attention between these embeddings and input data to generate better representations. Our method delivers accuracy gains compared to random initialization, at no latency costs. Plus, we show that the phonetic embeddings can easily be downsampled with k-means clustering, speeding up inference in 3.5 times at only slight accuracy penalties. | ['Mireia Farrús', 'Jordi Luque', 'Guillermo Cámbara'] | null | null | null | null | coling-2022-10 | ['small-footprint-keyword-spotting', 'keyword-spotting'] | ['speech', 'speech'] | [ 1.05223954e-01 4.52637106e-01 -2.67383866e-02 -1.42063335e-01
-9.26240444e-01 -8.55875552e-01 4.07112151e-01 2.68402189e-01
-8.30096483e-01 2.18198583e-01 6.50767207e-01 -7.84611702e-01
3.41853559e-01 -6.32811844e-01 -9.76212919e-01 -5.71759999e-01
1.94155630e-02 4.65320557e-01 -3.35139632e-02 1.01678237e-01
5.27575426e-02 2.26826549e-01 -1.64038420e+00 3.75540763e-01
4.52860892e-01 8.91975284e-01 6.28426611e-01 9.55022335e-01
-5.05257189e-01 4.43631470e-01 -5.74516416e-01 -4.55196202e-01
1.87315360e-01 -4.30068970e-02 -7.65866995e-01 -3.72795425e-02
4.35341477e-01 -5.62024236e-01 -4.45716023e-01 7.38421440e-01
5.29940724e-01 3.06514353e-01 5.92211008e-01 -7.74366796e-01
-9.09406424e-01 1.13940060e+00 -1.45272747e-01 4.32379171e-02
-7.24159181e-02 1.66214600e-01 1.32329750e+00 -1.15490925e+00
4.09328908e-01 1.27519822e+00 4.89792079e-01 5.77560961e-01
-1.40251780e+00 -4.87938017e-01 2.06711859e-01 2.86061555e-01
-1.50351095e+00 -8.32049966e-01 6.22568905e-01 -3.36834401e-01
1.62479293e+00 2.98656493e-01 4.24867332e-01 1.25218356e+00
-3.64499986e-01 9.07711327e-01 4.72681969e-01 -4.35111076e-01
3.93053412e-01 3.29901457e-01 -6.95319623e-02 5.16527891e-01
-5.41367643e-02 -7.39837363e-02 -5.23003995e-01 7.52737150e-02
7.47828603e-01 -2.00528413e-01 -1.56066433e-01 -3.52084041e-01
-9.63732839e-01 7.98871279e-01 4.31327373e-01 1.97508693e-01
-3.09590220e-01 5.22632957e-01 5.49475729e-01 5.01695633e-01
3.14709604e-01 5.03847539e-01 -6.70831978e-01 -6.25284731e-01
-9.83602941e-01 -2.74140775e-01 4.15388852e-01 1.05741537e+00
7.91401505e-01 2.12832570e-01 -4.05952372e-02 1.11050308e+00
2.38646269e-01 6.54072165e-01 8.09480488e-01 -9.32297647e-01
5.47322929e-01 1.29194856e-01 1.50379688e-02 -6.62772000e-01
8.08252990e-02 -2.45510995e-01 -4.66189951e-01 -2.36417919e-01
1.67365372e-01 -5.39179482e-02 -1.12955403e+00 1.68764091e+00
7.36621320e-02 1.62963033e-01 1.35128379e-01 7.30625868e-01
3.03988308e-01 1.04374933e+00 -1.60257906e-01 1.12996623e-01
1.26064324e+00 -1.11491251e+00 -5.35705030e-01 -4.16321754e-01
8.38997006e-01 -7.05932915e-01 1.63902247e+00 4.94508982e-01
-1.11550903e+00 -8.07286203e-01 -9.94089961e-01 -5.20498931e-01
-5.80172658e-01 9.38194543e-02 6.29443526e-01 8.28620613e-01
-1.19483531e+00 4.32995975e-01 -9.51500237e-01 -2.75672942e-01
2.49636903e-01 4.36536551e-01 -3.21873844e-01 -9.82415527e-02
-1.00294793e+00 8.59306991e-01 4.12723839e-01 1.60530256e-03
-1.03640079e+00 -7.82306790e-01 -1.01050317e+00 3.87060404e-01
3.83790255e-01 -3.24802011e-01 1.20562637e+00 -5.18339038e-01
-1.67213273e+00 3.49853128e-01 -3.46504837e-01 -6.35838747e-01
1.14504300e-01 -3.07746828e-01 -1.87135100e-01 -6.55309930e-02
-1.89507484e-01 8.56506824e-01 8.95673633e-01 -1.05802298e+00
-3.64547580e-01 -1.68664679e-01 -2.31171157e-02 1.96661681e-01
-7.93838739e-01 -2.22561643e-01 -1.02505767e+00 -5.28237760e-01
-8.22505727e-02 -7.24881530e-01 -1.90088525e-01 -2.76358038e-01
-3.85653824e-01 -2.51636595e-01 5.35606265e-01 -6.88435316e-01
1.26160133e+00 -2.29772949e+00 2.31719509e-01 2.46744245e-01
1.03828683e-01 5.01538098e-01 -5.38940728e-01 5.43462873e-01
2.42519677e-01 3.06686908e-01 -1.80442464e-02 -8.49533081e-01
2.56637692e-01 6.21670544e-01 -7.65756011e-01 2.88809299e-01
1.53068408e-01 9.79215622e-01 -9.30824399e-01 -2.06744969e-01
4.35112953e-01 6.89707994e-01 -9.51867402e-01 2.76018053e-01
-1.97963208e-01 -1.09391615e-01 9.45682228e-02 8.62866715e-02
5.69802821e-01 1.15514502e-01 3.07677031e-01 -1.26822755e-01
-7.44344816e-02 8.81348908e-01 -1.10351968e+00 1.91440606e+00
-1.12520134e+00 7.45283365e-01 4.76618372e-02 -1.03675103e+00
7.63824284e-01 2.98929065e-01 7.59580731e-02 -6.38704062e-01
1.53732002e-01 1.47220299e-01 -1.81873336e-01 -2.36552015e-01
5.48769355e-01 1.53425962e-01 -1.51271477e-01 4.72738206e-01
4.67534542e-01 2.19867919e-02 -2.29229867e-01 1.92799911e-01
9.35889542e-01 -1.15548350e-01 -2.24493623e-01 2.35423371e-02
1.67968944e-01 -3.51919591e-01 2.20678359e-01 1.10359013e+00
2.40690425e-01 5.43352485e-01 3.54032099e-01 -2.05880985e-01
-1.31920791e+00 -1.15021729e+00 3.10184192e-02 1.32374978e+00
-3.01897466e-01 -7.98598945e-01 -8.31553340e-01 -5.68614364e-01
1.70749560e-01 9.18140531e-01 -5.43247104e-01 -3.89763951e-01
-5.74959934e-01 -2.18031675e-01 8.47273171e-01 7.21338212e-01
-2.52944797e-01 -8.89119327e-01 -5.49868524e-01 5.06855249e-01
6.20259494e-02 -9.58085477e-01 -5.97593904e-01 6.83760941e-01
-7.41771936e-01 -3.79350871e-01 -5.98717630e-01 -5.25392711e-01
6.67662323e-01 2.67210245e-01 8.69904995e-01 -8.55681747e-02
-9.85712558e-02 1.76580772e-01 -4.14815813e-01 -1.74869955e-01
-3.83920968e-01 3.50492150e-01 3.62481892e-01 1.65166836e-02
4.50904876e-01 -5.87079644e-01 -2.45160162e-01 2.60010688e-03
-8.23465884e-01 1.24373220e-01 5.54617345e-01 9.87491369e-01
4.68865454e-01 1.16397358e-01 1.65609822e-01 -6.80574417e-01
3.30209255e-01 -4.31035876e-01 -6.60032928e-01 -6.51761517e-02
-5.52603364e-01 4.15157229e-01 9.05933142e-01 -6.09792113e-01
-6.41501248e-01 -8.38114023e-02 -3.55512679e-01 -6.69477940e-01
-1.53644726e-01 3.66672784e-01 -3.61112356e-01 3.00488085e-01
3.66435736e-01 2.45343432e-01 -6.80045560e-02 -1.07966483e+00
1.05046618e+00 1.06624997e+00 4.36112940e-01 -5.15829086e-01
8.09702933e-01 1.14941962e-01 -7.56837070e-01 -1.03538990e+00
-6.22564971e-01 -5.57643473e-01 -5.14070749e-01 2.67572820e-01
7.00824022e-01 -7.30939686e-01 -7.45537698e-01 1.09154634e-01
-1.33527708e+00 -5.45836627e-01 -4.76357192e-01 5.54817498e-01
-3.30612510e-01 1.89834833e-01 -6.02641344e-01 -8.33819151e-01
-9.77092981e-02 -1.28296244e+00 9.49432552e-01 -3.29714924e-01
-1.93327099e-01 -7.97631443e-01 -1.12000734e-01 2.94052362e-01
7.28421092e-01 -7.50258505e-01 1.03822911e+00 -8.34677994e-01
-4.58184958e-01 -6.02446459e-02 -2.06398755e-01 6.62192523e-01
9.42420587e-02 -3.29256654e-01 -1.44228172e+00 -3.45865130e-01
-3.03154707e-01 -2.69001186e-01 1.14808619e+00 4.15156186e-02
1.53566837e+00 -6.04073226e-01 -2.13537350e-01 8.81410360e-01
1.33201003e+00 7.97113478e-02 3.95346105e-01 2.30401922e-02
9.83455002e-01 4.07267064e-01 7.56589100e-02 1.97773412e-01
3.16198617e-01 6.95385277e-01 2.71083623e-01 -4.91318479e-02
-3.70295167e-01 -6.19658470e-01 4.87144828e-01 1.46668136e+00
2.03417167e-01 -2.57274359e-01 -9.26977456e-01 9.23301756e-01
-1.51169622e+00 -6.49547338e-01 4.90587443e-01 2.33420968e+00
1.10676455e+00 2.07072273e-01 -2.83987403e-01 2.37496302e-01
2.42317498e-01 1.84037834e-01 -4.54014719e-01 -9.70872462e-01
1.09973490e-01 5.87890446e-01 8.84954810e-01 1.11278415e+00
-8.76984656e-01 1.41359496e+00 6.22531128e+00 9.25910175e-01
-1.09681332e+00 2.73249209e-01 2.54297853e-01 -4.43375677e-01
-6.91354573e-01 -1.79469083e-02 -9.23395395e-01 5.07946849e-01
1.51919031e+00 1.14501856e-01 8.31071436e-01 9.09784853e-01
8.95130560e-02 3.30471158e-01 -1.23748684e+00 1.02455723e+00
5.44157773e-02 -1.28376782e+00 3.15420717e-01 2.49315649e-01
3.11542064e-01 1.89573511e-01 2.03609124e-01 6.88683331e-01
4.03028160e-01 -1.07650018e+00 7.15996087e-01 4.18386683e-02
8.93092632e-01 -9.66600358e-01 5.81494510e-01 1.91032186e-01
-9.15217876e-01 -9.47293192e-02 -9.04885173e-01 7.34552965e-02
3.42550948e-02 5.00779867e-01 -1.30273247e+00 9.37694833e-02
4.14987892e-01 7.86108226e-02 -5.12960196e-01 6.75081909e-01
-1.30957127e-01 9.97705519e-01 -4.93664116e-01 -6.70288950e-02
5.47255158e-01 1.07937895e-01 2.26036727e-01 1.51443624e+00
4.42153543e-01 -3.74913394e-01 -4.33894210e-02 7.04659045e-01
-2.76418328e-01 2.63022751e-01 -7.66913354e-01 -2.22659349e-01
8.44677448e-01 8.68586659e-01 -3.72251898e-01 -5.55591583e-01
-4.33912516e-01 1.41915393e+00 5.32247066e-01 3.20372552e-01
-8.75983417e-01 -6.04211926e-01 9.74001586e-01 -7.07606599e-02
9.95478451e-01 -4.45824981e-01 -8.95647854e-02 -1.14418256e+00
-2.09106818e-01 -7.71402955e-01 -1.69336721e-01 -6.87431753e-01
-8.35135877e-01 5.00915587e-01 -2.67171741e-01 -5.72767437e-01
-3.67982328e-01 -7.70099640e-01 -6.10382408e-02 1.06562591e+00
-1.27865839e+00 -9.77902532e-01 2.56165355e-01 2.12146893e-01
6.33372843e-01 6.39532134e-02 1.16246653e+00 4.85207111e-01
-4.50304836e-01 1.07866132e+00 2.00513989e-01 2.60894865e-01
5.79447031e-01 -1.51862550e+00 1.02437305e+00 7.99927711e-01
6.61543071e-01 9.64333892e-01 6.00399315e-01 -2.24833667e-01
-1.71902096e+00 -1.09709287e+00 1.17203355e+00 -5.57676077e-01
8.57852399e-01 -1.06411242e+00 -9.52517509e-01 7.63938844e-01
2.40607783e-01 -1.99461594e-01 8.07068765e-01 4.92431313e-01
-8.82076621e-01 -2.19114378e-01 -5.23274779e-01 7.40452886e-01
1.01921093e+00 -9.54346836e-01 -6.07780159e-01 1.79208025e-01
1.21814847e+00 -2.42815793e-01 -7.28598654e-01 -2.20803171e-01
4.40579712e-01 -3.54153991e-01 1.24303079e+00 -7.90124118e-01
5.21680675e-02 -2.09624484e-01 -5.61630309e-01 -1.49564636e+00
-2.97516674e-01 -5.48238277e-01 -3.22702765e-01 1.34989858e+00
6.43105209e-01 -3.46597314e-01 6.55905664e-01 2.15174288e-01
-1.80406779e-01 -5.14826655e-01 -9.55275953e-01 -9.09306049e-01
1.23807557e-01 -1.07372475e+00 8.75661016e-01 6.54634297e-01
-2.03749254e-01 4.37292635e-01 -4.59916800e-01 3.05144221e-01
4.35616702e-01 -2.57820785e-01 7.94359028e-01 -7.79400587e-01
-6.39242411e-01 -4.31380063e-01 -3.72224540e-01 -1.65345228e+00
8.98753852e-02 -1.06754696e+00 2.28363857e-01 -1.27431071e+00
-2.50942320e-01 -5.32826841e-01 -6.85493052e-01 6.30898535e-01
-5.35584353e-02 2.72111118e-01 2.44879112e-01 -8.14508647e-02
-2.36797914e-01 5.20241439e-01 6.86036110e-01 -2.95659602e-01
-1.76627710e-01 -4.88491625e-01 -7.30498135e-01 4.77971613e-01
8.52995157e-01 -4.62885112e-01 -6.99841499e-01 -1.16266847e+00
4.70097028e-02 -1.93290412e-01 1.18039623e-01 -1.00152552e+00
2.52802879e-01 1.49165124e-01 2.12743074e-01 -5.57531953e-01
6.98412836e-01 -7.64379561e-01 -2.19344556e-01 6.28645718e-02
-6.24416947e-01 -2.83838399e-02 3.84466201e-01 4.62794125e-01
3.75929736e-02 -3.67182463e-01 4.91530508e-01 5.27394237e-03
-6.22227192e-01 1.55688345e-01 -7.39661992e-01 -8.72522593e-03
3.90809834e-01 6.25755172e-03 -2.38229092e-02 -3.24490994e-01
-7.54618227e-01 -6.30117059e-02 5.27889609e-01 6.18977904e-01
5.73219359e-01 -1.21652544e+00 -2.54818767e-01 5.75536132e-01
1.13858007e-01 -1.16195127e-01 1.37277976e-01 4.56260294e-01
-2.28512183e-01 6.96016967e-01 1.48711562e-01 -3.42013776e-01
-1.03543890e+00 9.78869200e-01 8.50867257e-02 1.85464084e-01
-6.45976901e-01 1.29337215e+00 5.44633120e-02 -5.24832547e-01
6.13681912e-01 -6.91177785e-01 6.92874342e-02 2.36650214e-01
6.63296938e-01 9.28230286e-02 3.38308752e-01 -4.76666421e-01
-3.85120362e-01 4.53620344e-01 -1.77081719e-01 -5.62638104e-01
1.20206463e+00 -2.47661337e-01 1.16989478e-01 6.29531682e-01
1.75987256e+00 3.75734568e-01 -1.06462038e+00 -2.51647204e-01
-4.87746932e-02 -3.43989342e-01 3.61010015e-01 -5.43987870e-01
-9.66147304e-01 1.44725001e+00 3.81345272e-01 1.51735649e-01
6.89111292e-01 7.63590336e-02 1.14076376e+00 6.73672020e-01
3.16978723e-01 -1.19703150e+00 -2.41298571e-01 5.82098424e-01
4.79597718e-01 -8.63699377e-01 -5.57340682e-01 7.40930140e-02
-5.35116136e-01 8.45045507e-01 1.91964880e-01 -8.95545110e-02
5.24766624e-01 4.52342600e-01 1.60388365e-01 1.07278831e-01
-1.00720227e+00 -2.76775599e-01 5.90471700e-02 6.64675295e-01
1.64554924e-01 2.62109339e-01 3.52338970e-01 6.58475101e-01
-5.52699983e-01 -5.01938999e-01 3.94160956e-01 5.60846627e-01
-7.31571913e-01 -1.26990187e+00 -7.71604329e-02 2.89343446e-01
-2.47517377e-01 -6.60677016e-01 -1.01914190e-01 3.48981619e-01
1.53860718e-01 9.00764525e-01 4.69934434e-01 -5.96824467e-01
2.33126581e-01 3.45228732e-01 2.38071129e-01 -7.45753288e-01
-2.53951013e-01 1.74158469e-01 1.28234774e-01 -7.92662799e-01
3.68225902e-01 -3.71034831e-01 -1.08933163e+00 -2.18717232e-01
-2.14137375e-01 4.01108295e-01 9.02346849e-01 6.38272226e-01
6.59867227e-01 5.16103983e-01 5.66035092e-01 -7.60991573e-01
-6.87056065e-01 -9.36790168e-01 -2.17612907e-01 1.53262347e-01
4.99985874e-01 -5.23921847e-01 -5.06847560e-01 1.46664545e-01] | [14.262812614440918, 6.747035503387451] |
46bfc932-07ea-43ee-9e6f-3ea898bfb3af | contextual-pyramid-attention-network-for | 2004.07018 | null | https://arxiv.org/abs/2004.07018v1 | https://arxiv.org/pdf/2004.07018v1.pdf | Contextual Pyramid Attention Network for Building Segmentation in Aerial Imagery | Building extraction from aerial images has several applications in problems such as urban planning, change detection, and disaster management. With the increasing availability of data, Convolutional Neural Networks (CNNs) for semantic segmentation of remote sensing imagery has improved significantly in recent years. However, convolutions operate in local neighborhoods and fail to capture non-local features that are essential in semantic understanding of aerial images. In this work, we propose to improve building segmentation of different sizes by capturing long-range dependencies using contextual pyramid attention (CPA). The pathways process the input at multiple scales efficiently and combine them in a weighted manner, similar to an ensemble model. The proposed method obtains state-of-the-art performance on the Inria Aerial Image Labelling Dataset with minimal computation costs. Our method improves 1.8 points over current state-of-the-art methods and 12.6 points higher than existing baselines on the Intersection over Union (IoU) metric without any post-processing. Code and models will be made publicly available. | ['Clint Sebastian', 'Raffaele Imbriaco', 'Peter H. N. de With', 'Egor Bondarev'] | 2020-04-15 | null | null | null | null | ['segmentation-of-remote-sensing-imagery'] | ['miscellaneous'] | [ 4.80606288e-01 -2.45346174e-01 1.10704660e-01 -5.42054057e-01
-6.56176805e-01 -6.85626924e-01 4.00800318e-01 3.60763371e-01
-6.05301797e-01 4.58376169e-01 1.89430520e-01 -3.64180773e-01
-2.60724604e-01 -1.33552039e+00 -7.87670672e-01 -5.37227690e-01
-2.78986156e-01 4.35588323e-02 4.74178612e-01 -2.32837707e-01
-3.61098014e-02 7.97473967e-01 -1.64178669e+00 4.12039518e-01
9.09277022e-01 1.01393485e+00 2.83756882e-01 8.58298719e-01
1.87771484e-01 4.81449574e-01 -9.81384888e-02 2.23684013e-01
5.56165397e-01 -1.01484619e-01 -1.20636725e+00 1.99990854e-01
7.42721200e-01 -3.84677738e-01 -2.07963362e-01 1.06360996e+00
4.11887020e-01 1.60205007e-01 2.90292323e-01 -7.75410235e-01
-6.26946747e-01 5.24435103e-01 -7.93633759e-01 2.88215399e-01
-3.35756928e-01 5.75488620e-02 1.23425388e+00 -7.78782248e-01
2.84964085e-01 1.12611949e+00 1.03339911e+00 -1.39394999e-01
-1.21288502e+00 -5.01947284e-01 4.43527877e-01 -2.87976172e-02
-1.59425247e+00 -1.01430506e-01 2.72744238e-01 -4.28574979e-01
1.06483173e+00 2.37297431e-01 7.05883861e-01 3.86410020e-02
-3.00644726e-01 7.29628623e-01 9.78240967e-01 -2.78335124e-01
7.11850598e-02 -5.93857706e-01 2.05396563e-01 7.92404532e-01
8.64902511e-02 -2.17398331e-01 6.92275912e-02 1.63244560e-01
8.99105132e-01 2.39832163e-01 -1.34509265e-01 1.06541783e-01
-1.04406321e+00 9.77996111e-01 1.26135993e+00 3.68099779e-01
-6.70980453e-01 3.60936075e-01 -1.64963808e-02 -9.74005908e-02
7.86399186e-01 3.48038018e-01 -7.58243382e-01 4.75661069e-01
-1.29522085e+00 3.45563859e-01 2.60107756e-01 6.28228962e-01
9.44344401e-01 -1.72167510e-01 -3.61480266e-02 7.04277754e-01
1.41253427e-01 5.66730022e-01 -1.15791209e-01 -9.67539966e-01
3.57261211e-01 9.49324906e-01 6.90867603e-02 -9.63269591e-01
-7.74689257e-01 -4.89027679e-01 -1.02673364e+00 3.44317615e-01
1.84650242e-01 -1.59432247e-01 -1.43153083e+00 1.30437851e+00
2.97501177e-01 2.02024475e-01 -1.43756151e-01 8.39998126e-01
9.00555432e-01 8.03555369e-01 3.42537522e-01 5.50854325e-01
1.24094892e+00 -9.22470987e-01 -7.59660229e-02 -5.99813342e-01
4.30320531e-01 -5.47477126e-01 9.40077841e-01 9.55548510e-02
-6.99520826e-01 -7.24851012e-01 -1.00684261e+00 -1.04762711e-01
-6.87341630e-01 4.24190342e-01 6.90093160e-01 5.51993728e-01
-1.29302013e+00 8.28209221e-01 -9.58691776e-01 -6.87830150e-01
1.02695751e+00 4.17195588e-01 -2.15776384e-01 -1.69314638e-01
-8.65734041e-01 5.30603886e-01 5.32615304e-01 3.49173337e-01
-8.30526710e-01 -7.64355540e-01 -8.91083956e-01 8.46646950e-02
2.84789860e-01 -4.88146722e-01 8.92317235e-01 -8.15939307e-01
-9.22314644e-01 7.89718330e-01 5.64154461e-02 -6.12952828e-01
1.43601611e-01 -5.39074600e-01 -4.74156253e-02 6.13612793e-02
3.59943360e-01 1.25249708e+00 3.21097225e-01 -1.06021833e+00
-1.09000349e+00 -5.03999472e-01 5.25498092e-01 1.90258890e-01
8.47688783e-03 8.31791610e-02 -3.11252117e-01 -5.26525080e-01
4.07515824e-01 -1.00073779e+00 -6.40026629e-01 2.89241858e-02
-1.49299473e-01 3.45567125e-03 1.05238414e+00 -8.26622367e-01
1.08288896e+00 -1.89542782e+00 1.57637112e-02 1.93043947e-01
-5.15475124e-02 4.59490061e-01 -2.99590409e-01 2.20382720e-01
-8.82966295e-02 5.49518645e-01 -1.01878214e+00 -2.79061310e-02
-2.36097649e-01 2.66578078e-01 -1.53228521e-01 4.20057416e-01
4.70350742e-01 1.03166950e+00 -7.83559978e-01 -3.25569272e-01
4.78110492e-01 4.88250732e-01 -4.58587497e-01 -1.62177220e-01
-2.25399762e-01 4.94286746e-01 -3.13970745e-01 8.86741638e-01
8.51450801e-01 -2.57787138e-01 -1.24572016e-01 -1.62121862e-01
-4.33625013e-01 2.45961640e-02 -1.10492027e+00 1.76644266e+00
-4.62976933e-01 6.68668747e-01 8.62628147e-02 -1.06770527e+00
6.39887989e-01 -4.43298779e-02 4.78956789e-01 -3.20666015e-01
-5.33588566e-02 -1.48629978e-01 -1.21356055e-01 -1.31396517e-01
5.61487436e-01 1.53016210e-01 -3.88857536e-02 8.35786834e-02
-1.03679433e-01 -2.81462789e-01 2.72104889e-01 -8.73878822e-02
1.03383172e+00 3.26075137e-01 3.01245064e-01 -4.55722928e-01
3.91836494e-01 2.30827972e-01 3.81240278e-01 6.28891528e-01
-6.49589598e-02 8.16912830e-01 2.18550079e-02 -8.40428531e-01
-8.99621904e-01 -6.87360764e-01 -2.12664440e-01 1.19377339e+00
-1.75462535e-03 -2.35731035e-01 -8.63353252e-01 -5.89706838e-01
-6.33599088e-02 4.73706931e-01 -7.02580869e-01 4.83545035e-01
-5.33268213e-01 -1.12847388e+00 8.54162395e-01 9.78829741e-01
1.22262633e+00 -1.12386727e+00 -1.05326331e+00 3.79014462e-01
-4.00531799e-01 -1.37941349e+00 1.29572839e-01 1.74062759e-01
-9.83297408e-01 -1.08295763e+00 -4.83469754e-01 -5.13453603e-01
5.53918064e-01 6.35445952e-01 1.18814659e+00 3.23363692e-01
-5.98335624e-01 1.55270711e-01 -4.74379152e-01 -4.91398126e-01
2.56012082e-01 4.26610708e-01 -5.96152127e-01 -1.24907114e-01
3.94645691e-01 -4.86789405e-01 -7.99656093e-01 2.67073721e-01
-1.12082350e+00 3.37753035e-02 7.06198812e-01 5.98428667e-01
7.22461760e-01 4.82335538e-01 1.28173158e-01 -6.88673198e-01
2.40791142e-02 -3.17791909e-01 -7.32000232e-01 2.12143779e-01
-3.04076701e-01 -1.90239921e-01 3.30390990e-01 3.30114007e-01
-9.58028853e-01 5.42858005e-01 -1.83429688e-01 1.00633390e-01
-9.66368318e-01 6.11583650e-01 -7.51473755e-02 -1.29556775e-01
5.72018087e-01 -2.14788541e-01 -7.63088048e-01 -3.55904967e-01
4.74702001e-01 5.86643398e-01 4.15111870e-01 -3.02487016e-01
7.55125761e-01 9.07777667e-01 1.04512051e-01 -1.22491467e+00
-1.24639869e+00 -9.15618837e-01 -1.23455179e+00 -1.49279192e-01
1.24673438e+00 -1.16138661e+00 -2.02150807e-01 6.69113040e-01
-9.66524363e-01 -5.62442899e-01 -1.92358628e-01 2.60056257e-01
-1.99666291e-01 5.58549277e-02 -3.72175634e-01 -6.72167063e-01
-4.55882013e-01 -1.06924927e+00 1.36984801e+00 4.15977627e-01
1.63094833e-01 -6.77679777e-01 -1.26765385e-01 3.26310992e-01
5.24034381e-01 6.11775517e-01 5.83823800e-01 -5.61772054e-03
-8.30662906e-01 1.19769685e-02 -7.35931039e-01 5.23143947e-01
9.79644805e-02 -2.07463801e-02 -1.11817312e+00 -8.37839469e-02
-5.36032617e-01 -2.39414439e-01 1.49650860e+00 8.87965798e-01
1.30952322e+00 -2.91692447e-02 -2.18863890e-01 7.07664609e-01
1.78783989e+00 -9.85993966e-02 8.51575911e-01 4.88441676e-01
7.13084102e-01 6.96336627e-01 4.33633685e-01 3.51443529e-01
3.41620296e-01 2.11428821e-01 9.37758684e-01 -6.79792106e-01
-2.86058430e-02 6.33977354e-02 -1.22312814e-01 8.77108984e-03
-6.36542916e-01 -1.41743824e-01 -1.25521922e+00 1.01076758e+00
-2.00367641e+00 -9.48886156e-01 -5.26910603e-01 1.76786602e+00
4.23313648e-01 -1.21200524e-01 -1.70842662e-01 1.61928788e-01
5.62492430e-01 3.89241546e-01 -3.40309501e-01 1.21093541e-02
-3.02811831e-01 6.39327586e-01 1.28751457e+00 5.09386003e-01
-2.02366686e+00 1.41771483e+00 5.80693865e+00 5.04217803e-01
-8.22943151e-01 1.50971442e-01 7.32562780e-01 3.06804538e-01
2.68264323e-01 1.90919057e-01 -7.35908449e-01 -1.25759453e-01
6.98721588e-01 6.86533213e-01 3.37359101e-01 8.38019729e-01
1.91210374e-01 -4.18783545e-01 -4.98788118e-01 5.75682580e-01
-1.58470854e-01 -1.35717511e+00 -1.78915009e-01 6.39148578e-02
1.15761817e+00 7.28357434e-01 -5.17611355e-02 -1.33744422e-02
7.24261522e-01 -1.13641715e+00 5.07719338e-01 3.08625251e-01
5.87400794e-01 -8.99395645e-01 1.04449725e+00 1.09985426e-01
-1.80169106e+00 -2.19847083e-01 -5.40969491e-01 -2.61741549e-01
1.16918221e-01 5.52554071e-01 -4.86504167e-01 7.97792137e-01
1.24039602e+00 8.04895878e-01 -8.72611284e-01 1.06162798e+00
-3.91783714e-01 9.41219747e-01 -6.10604048e-01 5.54071426e-01
9.92996454e-01 -2.85026431e-01 1.37416601e-01 1.27372754e+00
2.69309103e-01 5.72025359e-01 4.98476505e-01 7.05732644e-01
-1.64550096e-01 7.32215941e-02 -6.38641775e-01 1.32361036e-02
-9.01094452e-02 1.39836407e+00 -1.31362283e+00 -3.42591375e-01
-3.41133773e-01 9.75927234e-01 5.30227348e-02 2.35892400e-01
-8.29682529e-01 -4.84430492e-01 7.43131220e-01 1.83808301e-02
7.32147515e-01 -3.87520939e-01 -4.81622279e-01 -6.98890388e-01
-2.26323992e-01 -4.78208721e-01 3.49451423e-01 -7.22478688e-01
-8.89406145e-01 6.07825100e-01 2.03243017e-01 -8.75467837e-01
4.74136859e-01 -6.05575919e-01 -5.61602175e-01 6.81035221e-01
-1.91370070e+00 -1.76638019e+00 -7.58292317e-01 2.92766243e-01
8.08308125e-01 3.52922469e-01 1.01229799e+00 1.48381680e-01
-2.22778723e-01 -2.72130251e-01 -1.04107589e-01 4.81740624e-01
2.42910251e-01 -1.17791724e+00 8.27842057e-01 1.32857859e+00
2.84464359e-01 9.51667130e-02 2.32728660e-01 -4.76606876e-01
-4.10097301e-01 -1.78888535e+00 7.36392021e-01 -2.18658268e-01
4.18498546e-01 2.05050513e-01 -7.98026741e-01 7.50759065e-01
3.20374280e-01 1.94400534e-01 5.36226392e-01 3.51944044e-02
-2.87169844e-01 -1.54889047e-01 -1.08214593e+00 2.57827520e-01
1.15814126e+00 -2.69315660e-01 -3.26238215e-01 3.66558939e-01
6.12945497e-01 -1.68225184e-01 -7.07489312e-01 7.56407797e-01
4.10399735e-01 -1.05213475e+00 1.09899068e+00 -4.13048655e-01
4.57858771e-01 -7.11912036e-01 -4.92131859e-01 -1.23406303e+00
-7.48577237e-01 2.08411533e-02 6.77447736e-01 8.38457108e-01
3.05299014e-01 -3.33697110e-01 5.71202457e-01 2.97095150e-01
-3.32487762e-01 -4.20208871e-01 -6.03701055e-01 -5.01592994e-01
2.82564722e-02 -5.87676764e-01 7.42527246e-01 7.59374797e-01
-7.85635650e-01 2.65080512e-01 -6.15763143e-02 9.16159451e-01
5.61651111e-01 2.56712586e-01 6.31195426e-01 -1.53262305e+00
1.88036174e-01 -4.78036314e-01 -4.87293422e-01 -1.01610339e+00
5.01190349e-02 -8.46581936e-01 7.11619183e-02 -2.23247719e+00
1.65720675e-02 -4.70357716e-01 -2.33540654e-01 1.01694071e+00
-1.91159144e-01 7.31609285e-01 1.38104871e-01 -2.96488702e-02
-5.63244998e-01 4.39363867e-01 6.37208343e-01 -4.68733102e-01
-1.26898274e-01 -1.27093956e-01 -4.60480243e-01 1.10877967e+00
1.30117667e+00 -4.94032949e-01 -4.47237529e-02 -8.10963452e-01
5.26975505e-02 -6.84929550e-01 7.68069088e-01 -1.50541174e+00
8.96352828e-02 -5.54581657e-02 5.70213318e-01 -1.14714885e+00
6.17458522e-02 -9.31394577e-01 3.01502682e-02 4.00027156e-01
-5.05780615e-02 8.62714127e-02 5.31835973e-01 5.16193449e-01
-2.09181920e-01 -1.85563400e-01 9.03391719e-01 -4.70074296e-01
-1.06891119e+00 4.37733471e-01 -4.26675379e-01 -1.23730816e-01
9.38268721e-01 -6.39038384e-02 -1.73272401e-01 -9.68009159e-02
-6.61155283e-01 2.98262805e-01 4.01729405e-01 1.83351055e-01
5.86773574e-01 -8.77860308e-01 -9.76777971e-01 1.55477896e-01
1.16620116e-01 5.64168870e-01 3.88586551e-01 6.33904994e-01
-1.01797593e+00 5.35725534e-01 -2.01589510e-01 -6.82975709e-01
-1.29466319e+00 1.68306958e-02 3.18026274e-01 -4.37786222e-01
-6.44972265e-01 9.73962307e-01 1.06198989e-01 -6.37548625e-01
-1.31441042e-01 -8.18522751e-01 -3.63440663e-01 9.41985026e-02
4.87547100e-01 3.95673692e-01 2.24747822e-01 -8.88789058e-01
-3.84534001e-01 9.25516665e-01 4.38859135e-01 8.24033841e-02
1.71758235e+00 -1.28862038e-01 -3.13896894e-01 1.35417014e-01
7.60258973e-01 -4.36915994e-01 -1.31446767e+00 -2.80981511e-01
-4.71324362e-02 -4.59818274e-01 5.89865029e-01 -8.58232200e-01
-1.26654553e+00 1.03330028e+00 8.72110665e-01 1.99076533e-01
1.29890299e+00 -5.20746261e-02 7.11425662e-01 5.59943974e-01
2.26968482e-01 -1.15799487e+00 -4.21825469e-01 6.84131682e-01
7.96525955e-01 -1.50087154e+00 2.41692975e-01 -3.78541529e-01
-4.56503868e-01 1.05491161e+00 4.16954398e-01 -2.39014536e-01
9.84758556e-01 1.78846508e-01 -4.54529226e-02 -3.45606893e-01
-1.58957437e-01 -1.11912429e+00 2.60357857e-01 5.85090101e-01
3.56718272e-01 3.67063493e-01 1.43531635e-01 1.48792937e-01
1.25784561e-01 -5.48503548e-02 1.51515290e-01 1.06818509e+00
-8.24900746e-01 -7.37687290e-01 -4.38774258e-01 4.03894395e-01
-5.02056360e-01 -5.40304005e-01 -2.55735606e-01 7.55083859e-01
4.94966686e-01 1.08194029e+00 2.72622555e-01 -2.08551854e-01
4.43252772e-01 -5.46831526e-02 1.57770589e-01 -6.64032578e-01
-5.07703185e-01 1.14504592e-02 3.00255753e-02 -6.19416773e-01
-1.18974793e+00 -4.99822229e-01 -1.21387565e+00 -1.34887710e-01
-4.26642299e-01 -2.67254889e-01 6.83748364e-01 8.63424659e-01
3.26394618e-01 7.09683776e-01 1.67824164e-01 -1.13479793e+00
7.93910548e-02 -1.08306384e+00 -3.91045421e-01 8.38520192e-03
1.01272419e-01 -4.43083763e-01 -2.56229043e-02 3.23659658e-01] | [9.327454566955566, -1.2885466814041138] |
966faf70-d237-4b1d-8212-454ec65c5d03 | commonsense-knowledge-from-scene-graphs-for | 2210.14162 | null | https://arxiv.org/abs/2210.14162v1 | https://arxiv.org/pdf/2210.14162v1.pdf | Commonsense Knowledge from Scene Graphs for Textual Environments | Text-based games are becoming commonly used in reinforcement learning as real-world simulation environments. They are usually imperfect information games, and their interactions are only in the textual modality. To challenge these games, it is effective to complement the missing information by providing knowledge outside the game, such as human common sense. However, such knowledge has only been available from textual information in previous works. In this paper, we investigate the advantage of employing commonsense reasoning obtained from visual datasets such as scene graph datasets. In general, images convey more comprehensive information compared with text for humans. This property enables to extract commonsense relationship knowledge more useful for acting effectively in a game. We compare the statistics of spatial relationships available in Visual Genome (a scene graph dataset) and ConceptNet (a text-based knowledge) to analyze the effectiveness of introducing scene graph datasets. We also conducted experiments on a text-based game task that requires commonsense reasoning. Our experimental results demonstrated that our proposed methods have higher and competitive performance than existing state-of-the-art methods. | ['Michiaki Tatsubori', 'Daiki Kimura', 'Tsunehiko Tanaka'] | 2022-10-19 | null | null | null | null | ['text-based-games', 'common-sense-reasoning'] | ['playing-games', 'reasoning'] | [-2.39132605e-02 -3.75476554e-02 1.74600244e-01 3.09779290e-02
1.98706407e-02 -5.38679719e-01 8.32809508e-01 2.89025158e-01
-7.09408879e-01 8.78842950e-01 3.18185031e-01 -3.51871580e-01
-2.23446175e-01 -1.27077508e+00 -5.17180860e-01 -1.88827530e-01
2.02493742e-01 4.20141578e-01 6.79093778e-01 -9.84024704e-01
5.14039457e-01 -9.90720689e-02 -1.69025958e+00 4.88273829e-01
1.22396863e+00 6.94785774e-01 5.70749998e-01 3.90920788e-01
-4.76612568e-01 1.78429294e+00 -9.14590061e-01 -7.47553945e-01
7.48251155e-02 -8.14492464e-01 -9.14100885e-01 -2.15692371e-02
-6.31780624e-02 -5.08366346e-01 -6.81859434e-01 1.43941152e+00
4.60613549e-01 4.70840663e-01 5.30755043e-01 -1.52891731e+00
-8.38608861e-01 9.03893292e-01 -3.20788115e-01 3.06049734e-01
8.84136915e-01 2.37089798e-01 8.36939931e-01 -2.01492190e-01
8.14665496e-01 1.42279112e+00 2.75380582e-01 3.68505061e-01
-3.99308473e-01 -7.92542219e-01 4.56286147e-02 9.37162161e-01
-1.30549204e+00 1.55910300e-02 1.23380136e+00 -3.07939678e-01
1.04485047e+00 2.24534675e-01 1.24919558e+00 9.55367088e-01
-1.76292676e-02 9.67147946e-01 1.44488990e+00 -4.85970378e-01
3.07468861e-01 -7.36704022e-02 -2.91140288e-01 9.14177537e-01
1.98485970e-01 1.43025458e-01 -8.17775011e-01 2.75812507e-01
1.23391736e+00 1.03873573e-02 -2.68365741e-01 -2.83926338e-01
-1.26540124e+00 8.79216075e-01 8.05625141e-01 4.17074710e-01
-3.08165848e-01 2.35549927e-01 6.01993084e-01 2.54805833e-01
1.23751871e-01 5.53990364e-01 1.43179759e-01 -5.74822426e-01
-5.50111830e-01 3.14295143e-01 7.11826622e-01 1.02298200e+00
5.58969796e-01 1.00404680e-01 -4.24281657e-02 6.72734201e-01
2.11831138e-01 4.77178395e-01 7.05276728e-01 -8.29831541e-01
5.75075746e-01 1.02609193e+00 -8.18291157e-02 -1.53331304e+00
-3.53036612e-01 -2.89815784e-01 -8.46218884e-01 9.95851308e-02
3.27987611e-01 1.12811320e-01 -6.27115488e-01 1.66043425e+00
1.43956348e-01 3.83710682e-01 2.06982076e-01 1.03864133e+00
1.42104220e+00 4.04586643e-01 1.29643917e-01 -4.93322872e-02
1.44416213e+00 -7.58537412e-01 -1.06807458e+00 -2.93305993e-01
7.06071734e-01 -3.99443179e-01 1.48637915e+00 2.94863403e-01
-7.46810675e-01 -4.94464576e-01 -1.11627364e+00 6.89050332e-02
-8.75064552e-01 -3.04327935e-01 8.79064381e-01 5.64416468e-01
-8.95329654e-01 3.96446913e-01 -2.35601410e-01 -4.59701389e-01
5.57698429e-01 -2.69297451e-01 -3.55022460e-01 -2.63848424e-01
-1.82827687e+00 1.12532151e+00 9.48368669e-01 -1.48937955e-01
-7.89908886e-01 -3.01409930e-01 -1.34115696e+00 -5.45572154e-02
1.09908319e+00 -8.03865314e-01 1.04218912e+00 -6.90352499e-01
-1.24656963e+00 9.37112987e-01 4.54073906e-01 -2.92794645e-01
6.81243122e-01 -1.88417733e-02 -2.00900182e-01 2.56449342e-01
3.09242368e-01 3.86166602e-01 3.28059405e-01 -1.26943612e+00
-4.79405493e-01 -4.18325633e-01 9.66156304e-01 7.73281217e-01
-1.87720776e-01 -2.01603711e-01 -3.81324530e-01 -6.55651927e-01
1.66563150e-02 -4.97998238e-01 -1.86561510e-01 1.70361379e-03
-4.62870836e-01 -1.19336061e-01 6.64424241e-01 -6.53974175e-01
1.22003055e+00 -1.85808289e+00 -1.45763969e-02 -4.63887528e-02
6.28160357e-01 1.93458274e-01 1.77205637e-01 6.86325014e-01
2.85124928e-01 2.13400096e-01 -5.63278645e-02 3.48755240e-01
1.96269050e-01 5.10897398e-01 -1.41760945e-01 -3.66925448e-02
-2.81917274e-01 1.18457973e+00 -1.36908340e+00 -1.11009288e+00
6.50396347e-01 -5.90127818e-02 -4.77311343e-01 1.16473593e-01
-3.67382973e-01 2.53821909e-01 -6.80737555e-01 3.86175513e-01
4.90334511e-01 -3.00788909e-01 1.88026711e-01 -1.64683059e-01
3.48993778e-01 1.76696461e-02 -1.18831193e+00 2.00950313e+00
-3.36366922e-01 6.04176104e-01 -6.79271579e-01 -1.09018314e+00
7.21553981e-01 1.40828276e-02 2.51451181e-03 -1.15039396e+00
3.39075565e-01 -3.50419551e-01 2.77145803e-01 -7.58183777e-01
6.78489864e-01 -4.09858555e-01 -2.67867237e-01 2.16216937e-01
6.02182522e-02 -8.18780482e-01 5.59859574e-01 6.22598052e-01
1.08288431e+00 2.41890565e-01 7.93609142e-01 2.27233320e-02
2.64153421e-01 3.48283321e-01 2.13310242e-01 1.03725648e+00
-3.56352210e-01 6.50573894e-02 5.43046176e-01 -2.87897676e-01
-7.76134014e-01 -9.73491311e-01 5.30750215e-01 8.44361603e-01
8.31553876e-01 -8.20709825e-01 -5.75119436e-01 -4.66206193e-01
-4.37400818e-01 8.15879524e-01 -6.06220901e-01 -3.49726975e-01
-1.22331791e-01 -2.11825207e-01 8.22554111e-01 5.55527449e-01
1.41774654e+00 -1.46593428e+00 -1.02012706e+00 6.91854581e-02
-4.95767921e-01 -1.43994367e+00 5.02769276e-02 -1.77821174e-01
-5.86394906e-01 -1.43343759e+00 -2.85712570e-01 -5.83732009e-01
2.60794193e-01 2.97871411e-01 1.34250355e+00 3.83747518e-01
-2.80217946e-01 3.67002010e-01 -8.22981477e-01 -6.92393363e-01
-2.08707333e-01 -5.69355726e-01 -3.25064421e-01 -6.61174297e-01
3.55571002e-01 -6.29186094e-01 -3.49953890e-01 5.44992685e-02
-1.05772901e+00 6.10903323e-01 4.04299021e-01 8.90163481e-01
1.04593538e-01 6.20840490e-01 2.71979809e-01 -9.26601946e-01
1.24651062e+00 -3.11915964e-01 -1.86206907e-01 2.52639323e-01
-8.35627541e-02 -1.96912456e-02 6.96860671e-01 -2.92876124e-01
-1.20945847e+00 -5.10521412e-01 1.86530888e-01 -3.44490737e-01
-1.47946313e-01 1.04387867e+00 -2.30788648e-01 1.05483919e-01
8.71705413e-01 5.91264784e-01 -1.74309656e-01 1.67332768e-01
6.02280378e-01 4.80876118e-01 4.61077720e-01 -8.28570545e-01
5.90272784e-01 2.72018611e-01 6.92137256e-02 -8.24158430e-01
-7.03126967e-01 -5.10074973e-01 -5.01938879e-01 -6.88707709e-01
8.34426582e-01 -8.35610509e-01 -1.03128886e+00 4.48998034e-01
-1.12744641e+00 -2.61485606e-01 -3.18065614e-01 4.78770465e-01
-7.43932188e-01 5.32310367e-01 -5.46548903e-01 -7.94490159e-01
1.07492439e-01 -8.83379757e-01 6.91779256e-01 4.14417565e-01
5.16005047e-02 -1.06085670e+00 -2.41095051e-02 4.81254280e-01
1.21678300e-01 4.55201715e-01 9.24181521e-01 -6.35839403e-01
-4.57059085e-01 1.41409948e-01 -3.78127486e-01 -1.13301806e-01
2.09695354e-01 -3.51221532e-01 -7.26776540e-01 3.16071272e-01
-7.94688240e-02 -7.73730099e-01 7.04430699e-01 2.14631539e-02
1.22852492e+00 -2.60249108e-01 -7.35353902e-02 1.38455063e-01
1.46781528e+00 4.02280331e-01 1.11997068e+00 5.04419386e-01
7.49595404e-01 4.41329449e-01 9.91871893e-01 7.88663268e-01
6.50431633e-01 5.31226635e-01 6.28322661e-01 -3.37285688e-03
-6.92751631e-02 -6.02433980e-01 -1.43295214e-01 7.42607832e-01
-5.34803391e-01 -2.58895129e-01 -1.10544038e+00 5.59598088e-01
-2.17937064e+00 -1.51106369e+00 2.90714186e-02 1.55522096e+00
1.16002214e+00 2.62287408e-01 -2.14950051e-02 3.37068886e-01
6.74039483e-01 2.47553661e-01 -3.86382252e-01 -1.40639544e-01
-2.50002712e-01 3.07010084e-01 4.63948622e-02 1.81948736e-01
-8.60826671e-01 1.41657221e+00 5.40618229e+00 1.40518987e+00
-5.19764185e-01 -3.51415537e-02 6.44166172e-02 2.51884580e-01
-2.94590771e-01 -7.36456513e-02 9.43524837e-02 2.23269492e-01
1.13025665e-01 -5.17024100e-01 4.87537265e-01 7.91471660e-01
1.58024635e-02 -4.63452101e-01 -7.48302460e-01 1.55353248e+00
1.63911000e-01 -1.44518161e+00 4.40836310e-01 -1.64196581e-01
4.73896235e-01 -5.56152046e-01 -1.98041037e-01 7.18621433e-01
6.97992980e-01 -1.23640907e+00 7.58738577e-01 4.89630520e-01
5.69874763e-01 -8.19806695e-01 8.61892223e-01 7.34669387e-01
-1.27968001e+00 1.92512080e-01 -4.95860159e-01 -7.18675137e-01
-1.37485743e-01 2.14767039e-01 -7.24914670e-01 9.81320500e-01
6.84582055e-01 8.83862019e-01 -7.57285893e-01 1.00394928e+00
-6.90419555e-01 2.45556295e-01 -7.36206919e-02 -5.71294188e-01
2.88460433e-01 -1.40977785e-01 3.57186854e-01 7.86115229e-01
5.99917993e-02 6.77509487e-01 3.95321459e-01 1.04120851e+00
-5.93158714e-02 1.94449723e-01 -1.14603066e+00 -2.78064877e-01
5.54627717e-01 9.26992178e-01 -9.34144676e-01 -5.15240908e-01
-3.31890345e-01 1.02638745e+00 4.27095234e-01 1.34437457e-01
-1.04690576e+00 -5.86844265e-01 2.26294443e-01 -7.78106004e-02
-1.01968840e-01 -2.11604297e-01 -1.68356180e-01 -1.13786829e+00
-6.19510598e-02 -8.98412049e-01 4.65428442e-01 -1.54627073e+00
-1.33863258e+00 3.95274311e-01 3.83811265e-01 -1.30884600e+00
-2.79332101e-01 -7.54928708e-01 -6.02151513e-01 5.32616079e-01
-1.12415493e+00 -1.35959256e+00 -7.82335520e-01 1.10491621e+00
4.28530037e-01 -2.40723178e-01 6.12266719e-01 -2.63904870e-01
4.77638468e-02 2.53546268e-01 -4.76892054e-01 4.30290103e-01
3.99752021e-01 -1.23045504e+00 5.13690747e-02 6.93160951e-01
3.42024803e-01 5.49709678e-01 7.44073808e-01 -9.68898535e-01
-1.11190784e+00 -6.66694403e-01 9.54928473e-02 -1.91784754e-01
7.14829922e-01 -5.27094603e-02 -5.89077353e-01 5.49337089e-01
4.30754542e-01 -2.63436258e-01 7.83779860e-01 -3.23735713e-03
-2.71937251e-01 4.05889332e-01 -1.16251910e+00 1.20567822e+00
1.59356618e+00 -7.34504163e-01 -1.20886159e+00 1.53036982e-01
7.78462768e-01 -6.27005577e-01 -4.45203006e-01 1.74936235e-01
1.81929246e-01 -1.09118867e+00 9.47823763e-01 -7.11453140e-01
8.81051421e-01 -5.01746893e-01 -3.14705580e-01 -1.67229426e+00
-6.33805469e-02 -1.54438019e-01 2.14699194e-01 7.92332590e-01
-9.60066915e-02 -3.91140908e-01 6.88114524e-01 2.20610201e-01
4.22715917e-02 -3.31549376e-01 -6.73354328e-01 -7.94891775e-01
-1.73752248e-01 -6.96458578e-01 5.61075032e-01 1.44650447e+00
5.69427550e-01 5.42381167e-01 -3.34684283e-01 -1.84335351e-01
4.95546788e-01 1.17895477e-01 7.70485103e-01 -1.17727971e+00
-2.59855777e-01 -4.34302479e-01 -9.67227101e-01 -8.51485968e-01
2.55727977e-01 -7.82680809e-01 -6.67389035e-02 -2.04806542e+00
5.64725339e-01 -8.78754035e-02 -3.67003097e-03 5.36734581e-01
-3.58812332e-01 1.01324268e-01 4.62722927e-01 -1.08789712e-01
-9.56216216e-01 7.63026655e-01 1.96975899e+00 -3.55734408e-01
3.62601243e-02 -7.36007631e-01 -7.64571011e-01 9.62370872e-01
8.81830573e-01 -6.51233718e-02 -1.00353181e+00 -1.34470344e-01
6.20017290e-01 2.33848333e-01 7.29971349e-01 -1.09468174e+00
5.64609468e-01 -6.80682719e-01 3.39685589e-01 -5.17627895e-01
4.29729730e-01 -7.65313923e-01 -7.21165985e-02 3.53954792e-01
-1.09782100e-01 -4.05583568e-02 4.40589577e-01 6.46316171e-01
-4.70790774e-01 -2.54945844e-01 2.98672885e-01 -6.86850369e-01
-1.47109222e+00 -3.39961238e-02 -2.92216957e-01 6.57185197e-01
1.09415543e+00 -5.70238292e-01 -6.55035794e-01 -9.71834600e-01
-4.96958077e-01 5.87273687e-02 3.79935592e-01 2.07597777e-01
1.10456908e+00 -1.45002985e+00 -5.25406837e-01 -3.26278359e-01
4.73872691e-01 -1.16895884e-01 4.56430793e-01 2.64709413e-01
-8.51633430e-01 1.35593787e-01 -8.41712534e-01 -3.87348771e-01
-1.24740791e+00 7.46721506e-01 2.99683154e-01 -2.83446074e-01
-5.68142056e-01 6.27972722e-01 4.88012165e-01 -4.61686760e-01
-1.17515773e-01 -2.87719548e-01 -6.27086937e-01 -2.53451705e-01
3.80948603e-01 9.34534445e-02 -2.73936152e-01 -5.40651441e-01
-3.47993165e-01 3.73536110e-01 2.86858320e-01 -3.61326754e-01
1.03698099e+00 2.51433291e-02 -7.18290061e-02 5.18770218e-01
4.20953304e-01 -1.50648996e-01 -5.78507066e-01 -3.05693030e-01
-4.00565058e-01 -7.90167868e-01 -1.02437072e-01 -8.94585133e-01
-7.77790546e-01 8.84740889e-01 1.40700862e-01 3.08319986e-01
1.12138653e+00 7.56833702e-02 2.90969998e-01 6.32406831e-01
9.49726641e-01 -1.10988986e+00 5.72984219e-01 6.75251901e-01
1.05481625e+00 -1.34780884e+00 3.22264470e-02 -6.94403350e-01
-1.04517102e+00 8.71159971e-01 1.01292872e+00 6.28660768e-02
5.06326497e-01 2.48713344e-02 -2.02509955e-01 -6.08655393e-01
-5.60968637e-01 -8.49323869e-01 9.49006677e-02 1.09115684e+00
2.21782118e-01 2.17889935e-01 -3.11977953e-01 9.42116857e-01
-7.26534128e-01 3.03784870e-02 8.11934650e-01 1.30609190e+00
-4.68513817e-01 -6.27064288e-01 -2.27524579e-01 4.02490973e-01
-2.77509051e-03 -4.34608430e-01 -6.96750462e-01 1.23509192e+00
1.32387489e-01 1.20604908e+00 -1.94802374e-01 -5.22553980e-01
4.21780497e-01 -3.01338851e-01 8.96166623e-01 -6.10088110e-01
-2.42630258e-01 -6.01384342e-01 2.41466150e-01 -6.86581790e-01
-6.64773822e-01 -1.25853214e-02 -1.51857209e+00 -6.51196718e-01
-2.31649414e-01 3.04741338e-02 2.49043331e-01 1.23795104e+00
-3.12719792e-01 9.81773198e-01 -1.36475071e-01 -4.04550999e-01
-6.96034282e-02 -9.21762466e-01 -7.80633628e-01 7.39925146e-01
-4.32531863e-01 -1.18101537e+00 -1.67479273e-02 3.57300527e-02] | [10.714873313903809, 1.7242745161056519] |
be3fc162-4083-4c04-8b5a-c2731f4486d3 | learnda-learnable-knowledge-guided-data | 2106.01649 | null | https://arxiv.org/abs/2106.01649v1 | https://arxiv.org/pdf/2106.01649v1.pdf | LearnDA: Learnable Knowledge-Guided Data Augmentation for Event Causality Identification | Modern models for event causality identification (ECI) are mainly based on supervised learning, which are prone to the data lacking problem. Unfortunately, the existing NLP-related augmentation methods cannot directly produce the available data required for this task. To solve the data lacking problem, we introduce a new approach to augment training data for event causality identification, by iteratively generating new examples and classifying event causality in a dual learning framework. On the one hand, our approach is knowledge-guided, which can leverage existing knowledge bases to generate well-formed new sentences. On the other hand, our approach employs a dual mechanism, which is a learnable augmentation framework and can interactively adjust the generation process to generate task-related sentences. Experimental results on two benchmarks EventStoryLine and Causal-TimeBank show that 1) our method can augment suitable task-related training data for ECI; 2) our method outperforms previous methods on EventStoryLine and Causal-TimeBank (+2.5 and +2.1 points on F1 value respectively). | ['Yuguang Chen', 'Weihua Peng', 'Jun Zhao', 'Kang Liu', 'Yubo Chen', 'Pengfei Cao', 'Xinyu Zuo'] | 2021-06-03 | null | https://aclanthology.org/2021.acl-long.276 | https://aclanthology.org/2021.acl-long.276.pdf | acl-2021-5 | ['event-causality-identification'] | ['natural-language-processing'] | [ 3.82319063e-01 4.60901678e-01 -3.65144700e-01 -2.75591344e-01
-8.59169900e-01 -5.96084177e-01 9.13245499e-01 2.31443435e-01
-2.96892136e-01 1.28483343e+00 6.26572728e-01 -4.00377065e-01
-3.00987381e-02 -9.30018663e-01 -7.96614587e-01 -3.08187276e-01
-9.13863331e-02 5.05326092e-01 3.33455771e-01 -2.22728804e-01
1.95785940e-01 1.26246110e-01 -1.27452421e+00 4.72332686e-01
9.73924160e-01 5.42078495e-01 2.52577752e-01 6.82088375e-01
-2.75452375e-01 1.09357131e+00 -5.84155858e-01 -2.43836835e-01
-1.79557115e-01 -7.30941355e-01 -1.12536800e+00 -4.02722895e-01
-2.24099591e-01 -1.14302397e-01 -2.15051100e-01 5.88386595e-01
4.13948417e-01 7.46155456e-02 6.21342540e-01 -1.47275698e+00
-4.95300829e-01 1.31207037e+00 -3.86673987e-01 5.32252312e-01
5.68628013e-01 -1.97038889e-01 1.01839590e+00 -8.69897246e-01
5.89022219e-01 1.46148658e+00 4.42673355e-01 6.64760053e-01
-8.16100180e-01 -8.31404865e-01 4.76604939e-01 3.86044741e-01
-8.87676537e-01 -3.39815706e-01 9.67755437e-01 -2.96845853e-01
9.95491922e-01 1.61814913e-01 3.47630978e-01 1.33185327e+00
-1.54389516e-01 1.03195894e+00 9.21106696e-01 -7.00441658e-01
7.08134174e-02 -5.12866788e-02 1.55464157e-01 4.16663170e-01
-7.60232806e-02 3.97494793e-01 -7.94335783e-01 -1.28821403e-01
7.66423702e-01 -4.79513854e-01 -1.93246752e-01 4.13382500e-01
-1.64710116e+00 7.63737023e-01 5.37379133e-03 2.55534351e-01
-3.40260863e-01 1.69924766e-01 6.30885661e-01 2.33437151e-01
4.46886003e-01 4.58424747e-01 -7.32055604e-01 -2.08013013e-01
-5.74316978e-01 5.60871184e-01 7.74933279e-01 9.37607825e-01
4.30136174e-01 1.63257495e-01 -5.16807795e-01 7.11358070e-01
3.04170251e-01 1.65812135e-01 7.84313083e-01 -5.23537874e-01
8.52994800e-01 6.32634938e-01 2.49857664e-01 -6.07340574e-01
-4.50964063e-01 -2.07295701e-01 -7.32942700e-01 -1.58262700e-01
4.64529932e-01 -4.75470215e-01 -7.36297190e-01 1.97489047e+00
3.54752392e-01 6.12506688e-01 4.35902268e-01 5.38403332e-01
1.11428511e+00 7.91067362e-01 3.84488046e-01 -6.78984642e-01
1.35866463e+00 -9.26843703e-01 -1.11617744e+00 -2.97783643e-01
7.90207684e-01 -6.88946724e-01 1.14445817e+00 2.27618605e-01
-1.06655991e+00 -6.39245689e-01 -1.05858362e+00 2.96000719e-01
-2.33958483e-01 9.99833271e-02 8.72814238e-01 4.20961410e-01
-5.44475019e-01 3.90402973e-01 -6.63646221e-01 -9.27873328e-02
1.38592005e-01 2.14924887e-01 -1.72389835e-01 3.14239442e-01
-1.81457436e+00 7.06988394e-01 1.14834213e+00 -3.38714868e-02
-9.93158698e-01 -9.88125443e-01 -9.12575722e-01 -6.23748489e-02
5.12908638e-01 -5.04649222e-01 1.44639075e+00 -7.26920784e-01
-1.52898717e+00 3.62648189e-01 -1.85346738e-01 -6.57247901e-01
5.53181112e-01 -4.54063922e-01 -6.56331062e-01 -1.68317303e-01
2.19433516e-01 5.34250021e-01 5.70300519e-01 -1.30105567e+00
-9.19082522e-01 -3.45717482e-02 5.58908358e-02 2.36349016e-01
-4.31286991e-01 3.03941667e-01 -2.47020751e-01 -1.14252317e+00
-6.77835494e-02 -6.80433095e-01 -3.23150843e-01 -7.05564678e-01
-5.26834488e-01 -6.62777483e-01 9.36885059e-01 -5.33055902e-01
1.61826110e+00 -1.78526318e+00 -6.55113235e-02 -1.97986275e-01
-2.26612255e-01 2.83568829e-01 -1.25874043e-01 3.55840832e-01
-5.86539030e-01 3.52546573e-01 -2.42816240e-01 -1.70118257e-01
-2.31549785e-01 3.49526137e-01 -6.47461951e-01 -2.07508266e-01
5.53206205e-01 7.99091876e-01 -1.42783034e+00 -8.14771950e-01
4.48442511e-02 -2.44776998e-02 -5.43569207e-01 4.86061037e-01
-5.52203536e-01 7.47689545e-01 -5.04989386e-01 1.74214721e-01
2.04525232e-01 -1.02316931e-01 3.23127121e-01 -1.67406410e-01
-2.14406908e-01 6.29483342e-01 -1.35269940e+00 1.64239216e+00
-7.11594045e-01 2.64144212e-01 -6.62563205e-01 -1.13238692e+00
8.94018292e-01 9.90606606e-01 3.00633073e-01 -3.06686908e-01
-1.60906687e-01 1.65663913e-01 -1.12364357e-02 -6.37106299e-01
3.43472064e-01 -1.67197883e-01 -3.03331316e-01 7.01855600e-01
9.86341909e-02 1.51670069e-01 4.97312695e-01 2.78334051e-01
1.12393332e+00 4.27552819e-01 4.01748866e-01 7.47374371e-02
7.24396765e-01 9.01857466e-02 9.24952269e-01 6.90610588e-01
8.26757774e-02 4.58825022e-01 6.04397595e-01 -4.95705098e-01
-7.48797297e-01 -9.31847155e-01 1.04305156e-01 9.46065128e-01
-1.91434145e-01 -3.90340030e-01 -4.50600982e-01 -1.16072226e+00
-4.58933145e-01 1.07663953e+00 -4.83001649e-01 -4.61316854e-02
-1.10735929e+00 -1.10321915e+00 6.42160237e-01 1.00581086e+00
5.78279436e-01 -1.57804763e+00 -3.19491863e-01 6.14059687e-01
-8.10873866e-01 -1.21090937e+00 -3.73968691e-01 1.60090670e-01
-8.58819485e-01 -9.84689116e-01 -2.98254699e-01 -7.72752881e-01
4.83157128e-01 -2.25339368e-01 1.19901717e+00 -6.22581244e-02
2.61474564e-03 -1.40474096e-01 -5.49231112e-01 -7.04106629e-01
-8.37926805e-01 5.61818928e-02 5.51424846e-02 -4.89596650e-02
9.41306725e-02 -7.03639448e-01 -2.47374505e-01 1.68624565e-01
-7.53963351e-01 2.33260527e-01 5.00238299e-01 1.00463915e+00
4.06931639e-01 2.30057865e-01 1.43636942e+00 -1.08365941e+00
7.05697656e-01 -6.32228315e-01 -4.62566853e-01 2.69939601e-01
-6.40072167e-01 1.91562489e-01 5.79131365e-01 -6.08073711e-01
-1.81793189e+00 1.64200768e-01 8.77869315e-03 1.23442309e-02
-2.17659757e-01 8.23822975e-01 -3.20629388e-01 6.15079820e-01
8.00616741e-01 1.44857317e-02 -5.69334388e-01 -3.67725104e-01
4.89382803e-01 4.34295297e-01 8.19358766e-01 -7.65954971e-01
9.07855451e-01 2.07871541e-01 -1.64520651e-01 -1.93652347e-01
-1.04536140e+00 -2.06839383e-01 -7.34190583e-01 -1.92520440e-01
8.85975301e-01 -9.29878891e-01 -3.49580526e-01 3.21196467e-01
-1.52282894e+00 -5.63620090e-01 -1.63952366e-01 7.18791723e-01
-5.40035427e-01 3.01087052e-01 -4.98310715e-01 -9.75056410e-01
-3.66600424e-01 -6.15726352e-01 8.39074612e-01 1.41082510e-01
-5.16770482e-01 -1.09448910e+00 2.05403894e-01 2.88322121e-01
-1.97616398e-01 4.00835514e-01 1.17692244e+00 -8.07026505e-01
-3.08523089e-01 5.40841892e-02 -2.07943261e-01 1.24211445e-01
2.62844890e-01 -1.31366968e-01 -9.23596919e-01 2.64134407e-01
-1.71626210e-01 -3.83164227e-01 5.50763428e-01 7.84250349e-02
1.25147820e+00 -5.20915687e-01 -4.60677654e-01 -6.88761547e-02
9.87668872e-01 4.37515765e-01 6.33329988e-01 3.01808506e-01
5.55602670e-01 5.66736519e-01 1.00249028e+00 5.51758349e-01
6.31149709e-01 5.92457891e-01 2.48082802e-01 -1.16739176e-01
-3.37609142e-01 -6.65572524e-01 3.28797758e-01 6.50956213e-01
-2.71426350e-01 -3.88508946e-01 -9.43045199e-01 9.90135550e-01
-2.20521045e+00 -1.10325277e+00 -4.73886222e-01 2.01270533e+00
1.56924582e+00 3.94467741e-01 1.13071188e-01 6.18405879e-01
7.18531311e-01 -1.30826727e-01 -1.63862973e-01 -1.37572214e-01
8.68762508e-02 2.06897333e-01 1.18257537e-01 3.81694674e-01
-1.13343406e+00 1.05964112e+00 6.03929377e+00 6.41321301e-01
-6.75804496e-01 1.49979278e-01 4.36811745e-01 1.08928405e-01
-4.48507577e-01 1.09929554e-01 -9.50679243e-01 5.06493151e-01
1.07337606e+00 -4.28983271e-01 -2.35276315e-02 4.73805666e-01
4.56405431e-01 1.41031489e-01 -1.38298583e+00 7.50016928e-01
-2.27882370e-01 -1.52906406e+00 2.15880215e-01 -3.40053022e-01
7.73729324e-01 -6.02043331e-01 -4.14508045e-01 6.13886833e-01
6.93381608e-01 -8.22913527e-01 7.62958229e-01 3.76682550e-01
5.67697644e-01 -7.32854187e-01 7.59511530e-01 4.47176397e-01
-1.28017616e+00 -7.97255263e-02 1.48930222e-01 -3.13828737e-01
5.53398252e-01 5.94910860e-01 -1.03310168e+00 9.10774171e-01
6.23384476e-01 6.26294792e-01 -4.20662582e-01 7.46832907e-01
-7.53944457e-01 1.13698447e+00 -1.69248194e-01 4.90818173e-02
-8.79135877e-02 3.43865305e-01 4.91392165e-01 1.28496587e+00
1.69434682e-01 3.06604296e-01 2.77692795e-01 7.04712570e-01
-1.51638955e-01 -1.20790945e-02 -3.84754062e-01 -3.69764282e-03
7.64332235e-01 9.58905458e-01 -5.50763369e-01 -5.02504766e-01
-3.44733983e-01 8.31997216e-01 2.40067273e-01 2.55619735e-01
-1.08622921e+00 -3.33347350e-01 -2.40253676e-02 -1.61078349e-01
-1.53135702e-01 -1.75447799e-02 -4.00153130e-01 -1.00267243e+00
4.63726670e-02 -7.61441231e-01 8.11238468e-01 -7.40267277e-01
-1.23704875e+00 4.23332602e-01 4.14727420e-01 -1.12038314e+00
-6.52332246e-01 -1.39002010e-01 -9.64610338e-01 8.07824731e-01
-1.59536219e+00 -1.34032559e+00 -2.20053166e-01 5.78045011e-01
7.65711784e-01 -1.61705166e-01 9.42186177e-01 4.00975257e-01
-6.01936936e-01 5.62235117e-01 -6.41206026e-01 2.76420623e-01
8.32369387e-01 -1.39626896e+00 4.90570515e-01 9.68750536e-01
1.39394239e-01 3.34799111e-01 5.98021567e-01 -9.24841821e-01
-9.46922541e-01 -1.33339846e+00 1.44273889e+00 -5.36476791e-01
8.26309085e-01 -2.42408231e-01 -9.79137480e-01 8.59095275e-01
2.18166709e-01 -1.74727783e-01 7.15756178e-01 3.79942566e-01
-2.70064145e-01 2.64137313e-02 -8.49307835e-01 6.64849997e-01
1.21944797e+00 -1.40496656e-01 -1.06776381e+00 5.48865974e-01
1.01332617e+00 -3.66622716e-01 -8.58802736e-01 6.56556189e-01
1.46825984e-01 -4.15541500e-01 9.84980106e-01 -8.86892140e-01
7.11441457e-01 -4.26021516e-01 2.30567709e-01 -1.32316303e+00
-1.23219028e-01 -7.89238214e-01 -2.73769021e-01 1.82684267e+00
8.34880710e-01 -5.32426238e-01 4.45369780e-01 3.48292470e-01
-3.60947877e-01 -4.48626101e-01 -7.43989050e-01 -8.38831902e-01
-2.78693922e-02 -7.11530268e-01 7.19719172e-01 1.13887203e+00
2.87534624e-01 6.52749658e-01 -3.74716729e-01 2.73980975e-01
4.49434042e-01 5.16456477e-02 6.22541308e-01 -1.22722900e+00
-2.52776653e-01 -1.76915541e-01 2.49688134e-01 -6.77236915e-01
4.61393952e-01 -7.90766418e-01 1.74506202e-01 -1.52559435e+00
1.25922844e-01 -7.95394659e-01 -2.96337098e-01 9.09903705e-01
-7.72419810e-01 -9.90194082e-02 -1.29754230e-01 7.81995803e-02
-1.53666615e-01 5.45565605e-01 1.15394211e+00 1.32830143e-01
-5.20357132e-01 4.21097353e-02 -6.43299162e-01 9.40663099e-01
8.49276721e-01 -4.98479575e-01 -7.89414465e-01 -3.12992156e-01
3.81858110e-01 3.71345520e-01 5.48770070e-01 -8.14372778e-01
1.21280797e-01 -4.08173323e-01 9.93997678e-02 -7.04567611e-01
-1.91183150e-01 -3.13714176e-01 4.59298082e-02 3.33108306e-01
-5.28192699e-01 -2.51259585e-03 4.11596328e-01 7.96021402e-01
-3.79458606e-01 -4.23228830e-01 3.56594414e-01 -2.77528644e-01
-7.77300894e-01 1.30784243e-01 -4.14770424e-01 2.97392011e-01
8.98977518e-01 2.41753966e-01 -1.70640886e-01 -2.52330989e-01
-6.36376977e-01 2.75736481e-01 -4.09155786e-01 7.19558120e-01
3.91480654e-01 -1.55955505e+00 -1.06962895e+00 -2.53308825e-02
9.54781696e-02 2.01210096e-01 4.29978594e-02 5.41995585e-01
1.33313164e-01 3.90492499e-01 1.61594421e-01 -2.00708881e-01
-1.05634367e+00 5.46282470e-01 -5.11580259e-02 -8.20851326e-01
-4.07869190e-01 8.39576483e-01 3.10665756e-01 -2.92736322e-01
1.60910904e-01 -2.38951117e-01 -5.75269759e-01 2.20879287e-01
5.15009344e-01 1.91240445e-01 8.33975449e-02 -3.05098537e-02
-2.17473850e-01 -1.45165056e-01 -8.60219151e-02 -5.12823045e-01
1.25348854e+00 1.64800256e-01 -1.30623639e-01 4.44402784e-01
6.30267024e-01 -2.01175034e-01 -1.00960863e+00 -3.29518795e-01
4.31727529e-01 -1.27099201e-01 -1.59244105e-01 -1.11427200e+00
-6.90842688e-01 6.10352099e-01 1.29087254e-01 2.23055601e-01
1.16502595e+00 1.82433166e-02 8.31953764e-01 4.53434549e-02
1.55446038e-01 -1.01766396e+00 4.51413184e-01 5.20460129e-01
1.17595959e+00 -1.16889405e+00 -2.85017818e-01 -7.18080521e-01
-5.19088864e-01 1.01659036e+00 8.26975048e-01 2.30644196e-01
4.90837455e-01 4.00992632e-01 -8.71797502e-02 6.04760386e-02
-1.03725517e+00 -6.17930181e-02 1.20146997e-01 5.49799323e-01
7.08521664e-01 -8.63664374e-02 -6.69197142e-01 1.00441802e+00
-2.00350910e-01 1.49898142e-01 5.24549484e-01 8.23015332e-01
-7.79699767e-03 -1.48562622e+00 -3.65997761e-01 2.62013584e-01
-4.16358948e-01 -3.34280699e-01 -9.62885022e-02 7.49910891e-01
3.88665460e-02 1.25685227e+00 -2.25086972e-01 -3.97193693e-02
4.57869828e-01 3.63019466e-01 3.01461041e-01 -8.53697479e-01
-5.66855431e-01 -3.99120748e-02 6.16990328e-01 -2.97748178e-01
-6.15006089e-01 -8.07380617e-01 -1.55382383e+00 3.36641669e-01
-3.69188786e-01 2.11073115e-01 2.62453109e-01 1.18315697e+00
6.71814680e-02 9.86225367e-01 4.79588330e-01 -4.96833116e-01
-3.76290143e-01 -1.11747944e+00 -1.82154914e-03 4.94008154e-01
2.29500905e-02 -7.13429630e-01 -2.12640032e-01 7.01673388e-01] | [9.181829452514648, 9.11620807647705] |
915fe2b8-a7d8-4c53-9a7a-11ec665d5e8b | perceiving-the-invisible-proposal-free-amodal | 2205.14637 | null | https://arxiv.org/abs/2205.14637v1 | https://arxiv.org/pdf/2205.14637v1.pdf | Perceiving the Invisible: Proposal-Free Amodal Panoptic Segmentation | Amodal panoptic segmentation aims to connect the perception of the world to its cognitive understanding. It entails simultaneously predicting the semantic labels of visible scene regions and the entire shape of traffic participant instances, including regions that may be occluded. In this work, we formulate a proposal-free framework that tackles this task as a multi-label and multi-class problem by first assigning the amodal masks to different layers according to their relative occlusion order and then employing amodal instance regression on each layer independently while learning background semantics. We propose the \net architecture that incorporates a shared backbone and an asymmetrical dual-decoder consisting of several modules to facilitate within-scale and cross-scale feature aggregations, bilateral feature propagation between decoders, and integration of global instance-level and local pixel-level occlusion reasoning. Further, we propose the amodal mask refiner that resolves the ambiguity in complex occlusion scenarios by explicitly leveraging the embedding of unoccluded instance masks. Extensive evaluation on the BDD100K-APS and KITTI-360-APS datasets demonstrate that our approach set the new state-of-the-art on both benchmarks. | ['Abhinav Valada', 'Rohit Mohan'] | 2022-05-29 | null | null | null | null | ['amodal-panoptic-segmentation'] | ['computer-vision'] | [ 3.40945244e-01 2.97675729e-01 -2.91570187e-01 -6.15477979e-01
-7.56288230e-01 -5.80528975e-01 6.70146286e-01 -9.90700498e-02
7.66727701e-02 5.00112832e-01 2.12513760e-01 -2.94777006e-01
6.43630549e-02 -5.64840734e-01 -8.30439866e-01 -4.99437392e-01
3.01522374e-01 6.69628978e-01 6.09209061e-01 9.53312665e-02
2.06847135e-02 4.14996654e-01 -1.70724416e+00 9.13536131e-01
8.39596629e-01 1.43248451e+00 1.47447005e-01 3.30804795e-01
-3.46013278e-01 1.11039793e+00 -4.69061077e-01 -6.02075994e-01
4.19936091e-01 2.43229598e-01 -7.30171323e-01 4.85473096e-01
1.24899435e+00 -5.73254898e-02 -2.45878831e-01 9.14081812e-01
1.87061891e-01 -2.03127228e-02 6.60164595e-01 -1.49705410e+00
-5.32584310e-01 3.32005143e-01 -1.13979053e+00 3.98025960e-01
8.90486464e-02 5.19112289e-01 1.14195967e+00 -7.69672930e-01
6.87847495e-01 1.59810424e+00 3.87252629e-01 1.50911316e-01
-1.66862440e+00 -6.64206266e-01 1.06410682e+00 4.54908282e-01
-1.23634326e+00 -4.64058489e-01 8.25594842e-01 -7.18373001e-01
9.79434967e-01 2.49847949e-01 4.46158856e-01 1.22290409e+00
1.21043973e-01 9.18876827e-01 1.39287114e+00 1.10541962e-01
8.99030939e-02 9.28311646e-02 3.68516594e-01 7.17638433e-01
1.78392112e-01 -2.26501405e-01 -6.38139307e-01 2.43377879e-01
3.86308372e-01 -9.23261121e-02 1.50244758e-01 -6.49637341e-01
-1.08453822e+00 4.87860322e-01 6.83681846e-01 -1.84405342e-01
-3.91068578e-01 3.13270569e-01 2.03061447e-01 -9.02822763e-02
7.89937317e-01 9.05313790e-02 -5.07469773e-01 4.42234248e-01
-1.08112895e+00 2.48145789e-01 6.50518000e-01 1.02914298e+00
1.10022438e+00 -2.26131380e-01 -6.96686506e-01 5.98079622e-01
4.70519274e-01 4.75183040e-01 -3.41822207e-01 -1.21411037e+00
9.67670560e-01 8.88765872e-01 1.49659188e-02 -1.17042148e+00
-4.86541480e-01 -7.23183036e-01 -4.55213934e-01 4.42934722e-01
4.91934091e-01 4.08662967e-02 -1.35306358e+00 1.84497976e+00
7.35369146e-01 6.07936621e-01 -2.71582186e-01 9.98382151e-01
9.10619557e-01 6.54752374e-01 4.53863353e-01 3.77109349e-01
1.77319777e+00 -1.59827137e+00 -4.91862327e-01 -8.97149742e-01
5.19996956e-02 -5.30469060e-01 1.06181955e+00 2.86763668e-01
-1.01001894e+00 -9.13316905e-01 -8.72421682e-01 -4.63639438e-01
-6.52032554e-01 2.18221039e-01 5.98428130e-01 5.04564524e-01
-1.05994797e+00 -6.70085177e-02 -5.26049733e-01 -4.33378555e-02
9.74657893e-01 2.62889326e-01 -2.69755542e-01 -1.08508870e-01
-8.46128464e-01 8.71297419e-01 4.36427146e-01 1.87202051e-01
-1.12068188e+00 -9.19292748e-01 -9.11561191e-01 2.26900294e-01
8.77277672e-01 -8.18730056e-01 6.24715090e-01 -9.72317338e-01
-1.18081057e+00 1.06222486e+00 -4.67101038e-01 -3.16861182e-01
5.16653538e-01 -2.45425656e-01 -4.84895825e-01 2.13132501e-01
3.89224708e-01 1.29107428e+00 1.00111759e+00 -1.68155205e+00
-9.79423463e-01 -4.43678737e-01 4.27245617e-01 4.49094683e-01
1.57626882e-01 -3.18886489e-01 -9.35190618e-01 -4.36955720e-01
2.19987929e-01 -7.00244248e-01 -2.42908135e-01 1.49042860e-01
-8.60845089e-01 -1.15624115e-01 7.43356645e-01 -7.51192212e-01
1.00286341e+00 -2.26130509e+00 2.37012848e-01 3.36801231e-01
6.57979131e-01 -1.39373720e-01 -2.54537791e-01 -2.18040332e-01
-5.11346199e-03 -2.41910964e-01 -3.19987953e-01 -7.47926354e-01
3.74694705e-01 2.27632552e-01 -5.50596714e-01 5.00511885e-01
5.08545816e-01 1.07312036e+00 -6.60546243e-01 -7.82923162e-01
5.40571809e-01 4.28009123e-01 -6.59310162e-01 1.01838551e-01
-6.52096927e-01 5.88343143e-01 -2.93497711e-01 1.06225193e+00
8.83680582e-01 -5.14022052e-01 -6.34256005e-02 -3.86141449e-01
-1.32466450e-01 2.33872369e-01 -1.44905746e+00 1.86830294e+00
-2.95978934e-01 7.29247391e-01 3.58759254e-01 -8.96666706e-01
5.44589520e-01 -2.20217690e-01 2.76813120e-01 -9.80997205e-01
-1.06526449e-01 -1.35768786e-01 -3.50550354e-01 -4.43583101e-01
4.08693224e-01 2.03567937e-01 -1.14885136e-01 8.88228118e-02
3.01037412e-02 3.34712058e-01 1.82706222e-01 2.29788750e-01
8.60285580e-01 3.43134731e-01 -1.44204319e-01 -3.33321244e-01
7.38828421e-01 -4.60286625e-02 5.68022907e-01 9.00946140e-01
-2.62575060e-01 6.24195099e-01 7.70976782e-01 -5.54717302e-01
-4.94290560e-01 -1.41672909e+00 -3.28551173e-01 1.46672833e+00
7.14857996e-01 -2.58993715e-01 -7.27841616e-01 -8.18278670e-01
2.61121064e-01 8.34580362e-01 -9.34823811e-01 2.80917108e-01
-5.28048813e-01 -6.36193931e-01 1.89280882e-01 5.10201216e-01
3.91751409e-01 -1.12309074e+00 -5.17696977e-01 3.37939113e-02
-4.02610660e-01 -1.53882980e+00 -2.67432064e-01 3.13392878e-01
-2.25455999e-01 -1.10001934e+00 -2.80052155e-01 -5.95508516e-01
6.17400348e-01 2.55505860e-01 1.30036533e+00 -3.97454262e-01
-5.17447412e-01 3.04636657e-01 2.40614742e-01 -1.39340669e-01
3.91340815e-02 2.85345912e-01 -5.31089306e-01 5.76978564e-01
1.87142298e-01 -5.50386250e-01 -8.45285237e-01 3.25729728e-01
-6.48941636e-01 5.44756413e-01 5.90544462e-01 2.65379041e-01
8.55878294e-01 -2.10486859e-01 2.84501791e-01 -1.00558114e+00
4.57259268e-03 -6.68665469e-01 -5.79831898e-01 6.25044107e-01
-2.35333607e-01 -4.48013805e-02 1.66901007e-01 -1.67664513e-01
-1.44517601e+00 1.04663730e-01 1.19943760e-01 -3.85560960e-01
-5.21793365e-01 -5.37079051e-02 -6.61102295e-01 6.54886961e-02
3.26205999e-01 -3.37233506e-02 -4.77720767e-01 -5.41128635e-01
1.03381491e+00 3.54380667e-01 7.09852636e-01 -9.42041337e-01
5.89499116e-01 9.60674942e-01 -8.10123682e-02 -5.20914912e-01
-1.32621050e+00 -5.92095137e-01 -8.55584621e-01 -4.65749085e-01
1.49596286e+00 -1.22947741e+00 -6.95385396e-01 2.43067503e-01
-1.22513199e+00 -4.62707460e-01 -3.83770674e-01 -2.02651605e-01
-4.05315489e-01 6.70992732e-02 -2.42123261e-01 -4.46575522e-01
7.62325004e-02 -1.57093751e+00 1.61154473e+00 1.30728155e-01
-3.52550000e-02 -7.79742360e-01 -2.51605183e-01 7.86155045e-01
2.31669232e-01 3.60565513e-01 9.41124618e-01 -4.39651221e-01
-1.16353464e+00 1.65589288e-01 -9.10663009e-01 3.47897075e-02
-2.59369522e-01 -3.07443231e-01 -1.43062305e+00 6.73359679e-03
-3.64965171e-01 -2.86484897e-01 1.35931981e+00 4.84555304e-01
1.56887209e+00 -1.59677267e-01 -5.20329595e-01 9.17851865e-01
1.26993430e+00 -2.70012110e-01 5.15850902e-01 2.46631339e-01
1.11907518e+00 8.68254721e-01 2.72992194e-01 1.41573474e-01
8.29723358e-01 8.09805810e-01 8.71822715e-01 -5.57747841e-01
-5.99573255e-01 -5.61149530e-02 2.14351133e-01 5.63286617e-02
3.05268586e-01 -4.57954586e-01 -7.15803146e-01 4.58100051e-01
-1.92167413e+00 -7.93968797e-01 -3.47109795e-01 1.69975805e+00
5.56218505e-01 4.07079875e-01 1.23418547e-01 -3.79179776e-01
6.86996579e-01 7.04257905e-01 -8.27669263e-01 -2.09580898e-01
-2.22688988e-01 6.98035210e-02 4.88964111e-01 7.03857064e-01
-1.45544696e+00 1.25036573e+00 5.55531168e+00 1.00742924e+00
-9.45504129e-01 3.75653535e-01 1.10988998e+00 -1.31272495e-01
-2.89690286e-01 6.48565069e-02 -1.08281863e+00 5.05318165e-01
3.61415148e-01 6.32355213e-01 3.96485955e-01 6.79377794e-01
-2.29052335e-01 -2.93417007e-01 -1.05685294e+00 7.60818243e-01
1.01102233e-01 -1.37080371e+00 1.12884387e-01 4.46453728e-02
8.33437443e-01 2.07440719e-01 2.18812034e-01 2.22287580e-01
2.97713190e-01 -9.78983939e-01 1.20961094e+00 5.92047334e-01
6.78272247e-01 -3.22054625e-01 1.89757571e-01 3.97686996e-02
-1.52730894e+00 -1.76798820e-01 1.77239012e-02 1.88855171e-01
3.44133347e-01 5.40207624e-01 -4.08609957e-01 5.09332299e-01
7.60148764e-01 8.77313077e-01 -8.89805019e-01 8.53095293e-01
-3.37678790e-01 4.18701977e-01 -2.42775232e-01 8.01718891e-01
3.79753202e-01 -1.55477509e-01 5.40475547e-01 1.54417932e+00
-2.39257634e-01 -2.64174908e-01 4.69204098e-01 1.33424294e+00
8.00837055e-02 -2.66151518e-01 -1.64535463e-01 6.76965833e-01
2.17194140e-01 1.36350882e+00 -9.51567113e-01 -5.52115142e-01
-5.22012651e-01 8.78840804e-01 5.84541023e-01 7.53211260e-01
-1.34659708e+00 3.10355693e-01 8.66677105e-01 2.78804094e-01
5.59762239e-01 1.58791356e-02 -8.29901278e-01 -1.07660329e+00
7.28367642e-02 -6.62358403e-01 5.16657591e-01 -8.13504875e-01
-1.27482736e+00 4.96151179e-01 5.62278405e-02 -9.48544681e-01
3.92300755e-01 -7.25639939e-01 -4.60368246e-01 8.79713356e-01
-1.87426126e+00 -1.50941706e+00 -4.31148648e-01 5.67680001e-01
7.02895701e-01 1.58127904e-01 3.16281050e-01 4.45166975e-01
-9.48743522e-01 3.28166485e-01 -4.28897440e-01 -2.39775717e-01
4.90339726e-01 -1.32230794e+00 3.50092590e-01 8.68596017e-01
2.06060350e-01 1.88568234e-01 4.73681718e-01 -5.48673153e-01
-7.81297922e-01 -1.33089793e+00 7.59919405e-01 -7.15080678e-01
6.69442296e-01 -7.36253083e-01 -7.97156990e-01 7.40548253e-01
2.26542145e-01 2.30309248e-01 2.92161852e-01 2.63778806e-01
-6.32795632e-01 -5.22382319e-01 -8.86525929e-01 6.30629599e-01
1.29354179e+00 -5.52545011e-01 -4.45809603e-01 3.44830185e-01
8.60534072e-01 -5.46330035e-01 -4.38027650e-01 4.18087482e-01
4.22811598e-01 -1.05846226e+00 1.58267331e+00 -5.85281312e-01
5.80021262e-01 -6.18847430e-01 -3.35998237e-01 -6.95417225e-01
-4.15026605e-01 -2.98914492e-01 -3.50823909e-01 1.32614923e+00
5.88790715e-01 -4.85882819e-01 6.70792341e-01 6.46393776e-01
-3.54847819e-01 -7.34433293e-01 -1.11238921e+00 -2.87128568e-01
-3.51147622e-01 -7.38710344e-01 5.57913303e-01 7.93357670e-01
-6.25878930e-01 3.80290776e-01 -3.68135363e-01 5.92717826e-01
8.24082136e-01 3.38793457e-01 8.21742475e-01 -1.13898659e+00
-2.59899437e-01 -5.01375675e-01 -2.76625544e-01 -1.26906443e+00
3.93520564e-01 -1.11108053e+00 -1.05686091e-01 -1.66332364e+00
2.27607131e-01 -5.08350670e-01 -5.15609801e-01 6.27097189e-01
-2.95224458e-01 3.98975879e-01 2.82323182e-01 7.04607740e-02
-1.32436454e+00 3.73901606e-01 1.34260178e+00 -4.27859455e-01
-9.14173871e-02 -2.38981336e-01 -8.88087034e-01 8.01338673e-01
4.30927902e-01 -4.20028925e-01 -5.38794160e-01 -7.78226018e-01
1.36291265e-01 -2.97455788e-01 9.10282373e-01 -1.18773401e+00
2.51730293e-01 -2.44170457e-01 5.15514791e-01 -8.68119180e-01
6.81848109e-01 -1.02073872e+00 3.86470631e-02 -1.06954193e-02
-5.30394137e-01 -1.56791881e-01 3.57804000e-01 8.87885690e-01
4.68454584e-02 4.75083143e-01 6.11660659e-01 2.06466913e-02
-1.04559600e+00 2.72483766e-01 -3.72504257e-02 3.54730308e-01
1.16060412e+00 -3.53997052e-01 -7.23507345e-01 1.23618245e-01
-9.26519990e-01 6.22596979e-01 2.13817552e-01 5.86326480e-01
2.82921404e-01 -9.68160689e-01 -5.62693119e-01 4.20473665e-01
2.42077187e-01 2.58423358e-01 6.54669225e-01 1.04572260e+00
-2.86739379e-01 4.10655588e-01 -6.18235171e-02 -9.78707254e-01
-9.90085125e-01 5.11387110e-01 4.45491344e-01 -4.85720068e-01
-7.07068980e-01 1.01874483e+00 1.13791466e+00 -3.26973051e-01
5.99690676e-01 -3.62160832e-01 -2.59096175e-01 2.77771860e-01
2.88481653e-01 3.66594195e-01 -1.41678616e-01 -8.60819876e-01
-4.12184805e-01 5.45705914e-01 1.96843464e-02 2.53451336e-03
1.08962262e+00 -4.72308367e-01 -2.30900481e-01 3.51239800e-01
9.01541471e-01 -1.66997388e-01 -1.92162311e+00 -4.11944330e-01
-1.43660367e-01 -3.15729201e-01 2.22093822e-03 -1.21980929e+00
-1.29780233e+00 1.07423747e+00 5.60882986e-01 -9.26599801e-02
1.02128017e+00 3.79034847e-01 5.14973938e-01 -1.83945093e-02
-5.35352342e-02 -1.07485080e+00 7.06625059e-02 2.66876191e-01
7.38155425e-01 -1.32870340e+00 -7.02705309e-02 -8.47445130e-01
-7.07101345e-01 7.28991628e-01 1.07838714e+00 1.04897045e-01
6.68191433e-01 1.98223904e-01 5.16328365e-02 -6.18612647e-01
-7.42020547e-01 -4.57144082e-01 9.22973752e-01 4.29786772e-01
-1.34422883e-01 2.49426425e-01 2.06560910e-01 6.92638338e-01
2.44095027e-01 -4.05141830e-01 -2.98194557e-01 3.89635146e-01
-4.77960914e-01 -7.04731524e-01 -3.77024800e-01 3.63087088e-01
-8.66305456e-02 -2.24803463e-01 -3.54732573e-01 6.87887669e-01
9.19473827e-01 8.76683950e-01 3.11041057e-01 -1.59599110e-01
2.88391918e-01 1.56306520e-01 2.46010765e-01 -4.92393583e-01
-4.90443707e-01 1.76583067e-01 1.83452889e-01 -9.26291645e-01
-4.44777071e-01 -6.85300827e-01 -1.06085479e+00 3.90302949e-02
1.55338302e-01 -4.69081491e-01 4.68149930e-01 1.11574066e+00
6.00268304e-01 9.26164508e-01 2.85034060e-01 -1.03562140e+00
1.07643649e-01 -5.02277017e-01 -3.28340620e-01 4.50434327e-01
4.16256040e-01 -8.34000528e-01 -1.70048505e-01 1.14608958e-01] | [9.556949615478516, 0.3472689986228943] |
75b3bf1d-4998-4f9c-a4cf-654db345d353 | exploring-simple-3d-multi-object-tracking-for | 2108.10312 | null | https://arxiv.org/abs/2108.10312v1 | https://arxiv.org/pdf/2108.10312v1.pdf | Exploring Simple 3D Multi-Object Tracking for Autonomous Driving | 3D multi-object tracking in LiDAR point clouds is a key ingredient for self-driving vehicles. Existing methods are predominantly based on the tracking-by-detection pipeline and inevitably require a heuristic matching step for the detection association. In this paper, we present SimTrack to simplify the hand-crafted tracking paradigm by proposing an end-to-end trainable model for joint detection and tracking from raw point clouds. Our key design is to predict the first-appear location of each object in a given snippet to get the tracking identity and then update the location based on motion estimation. In the inference, the heuristic matching step can be completely waived by a simple read-off operation. SimTrack integrates the tracked object association, newborn object detection, and dead track killing in a single unified model. We conduct extensive evaluations on two large-scale datasets: nuScenes and Waymo Open Dataset. Experimental results reveal that our simple approach compares favorably with the state-of-the-art methods while ruling out the heuristic matching rules. | ['Alan Yuille', 'Xiaodong Yang', 'Chenxu Luo'] | 2021-08-23 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Luo_Exploring_Simple_3D_Multi-Object_Tracking_for_Autonomous_Driving_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Luo_Exploring_Simple_3D_Multi-Object_Tracking_for_Autonomous_Driving_ICCV_2021_paper.pdf | iccv-2021-1 | ['3d-multi-object-tracking'] | ['computer-vision'] | [-1.43901050e-01 -3.97960007e-01 -3.03038538e-01 -4.98823583e-01
-7.09837139e-01 -6.86479568e-01 7.24990487e-01 4.22920696e-02
-6.99673355e-01 4.28833932e-01 -5.24303675e-01 -4.11419511e-01
2.99868011e-03 -5.89037776e-01 -1.08661592e+00 -5.64479649e-01
1.16671897e-01 9.65434372e-01 1.13232005e+00 1.72745794e-01
7.05969706e-03 8.36327136e-01 -2.01661921e+00 -1.40744507e-01
6.81729317e-01 1.05999327e+00 3.73557866e-01 7.24948943e-01
-2.42131412e-01 3.65791440e-01 -1.09284289e-01 -5.17325163e-01
2.27408886e-01 1.72013074e-01 -8.37408230e-02 -1.25115976e-01
1.04664874e+00 -5.81379831e-01 -7.58791789e-02 9.25328374e-01
5.80489516e-01 -8.65429118e-02 5.35030901e-01 -1.72018647e+00
1.34602552e-02 7.31734261e-02 -7.90611804e-01 2.88219392e-01
-3.62560153e-02 3.54945004e-01 7.47714162e-01 -1.05819845e+00
6.43008530e-01 1.31911385e+00 1.12543988e+00 5.92168748e-01
-9.87765014e-01 -1.12589300e+00 2.62877554e-01 3.08048010e-01
-1.62177944e+00 -5.24567127e-01 4.42833483e-01 -6.46467686e-01
7.06266582e-01 2.08279699e-01 6.22674227e-01 7.01619625e-01
7.28572458e-02 9.38615680e-01 4.98115718e-01 2.35289359e-03
1.28013328e-01 1.05858125e-01 6.75046891e-02 8.31164122e-01
6.32708907e-01 5.83832979e-01 -4.17856961e-01 -9.85063165e-02
2.71382064e-01 2.34413117e-01 4.40465569e-01 -9.73864913e-01
-9.40254450e-01 7.93771207e-01 3.70122761e-01 -2.98386723e-01
-1.76275581e-01 3.89776677e-01 2.88784772e-01 -2.81736583e-01
4.38070357e-01 -6.27134681e-01 -5.84025085e-01 3.14195789e-02
-1.20212710e+00 7.06095159e-01 2.99611270e-01 1.32131314e+00
8.10435355e-01 -3.08684379e-01 -5.30456662e-01 3.33041340e-01
5.89452326e-01 7.97417879e-01 -2.31804043e-01 -7.97844768e-01
2.93539047e-01 5.47632754e-01 4.17160600e-01 -4.29201126e-01
-4.52528954e-01 -3.31562728e-01 -5.87362610e-02 6.32799864e-01
3.44244003e-01 -1.68288752e-01 -9.83151138e-01 1.47402942e+00
1.21357322e+00 6.73848450e-01 -4.36343312e-01 7.24011660e-01
8.14510703e-01 2.44363189e-01 1.86567232e-01 4.00563776e-02
1.36383653e+00 -9.34569716e-01 -5.92286348e-01 -2.99577147e-01
6.43985033e-01 -5.42201161e-01 3.37634176e-01 -1.97671205e-02
-9.15855885e-01 -7.08934665e-01 -9.51604843e-01 -6.79490417e-02
-4.11190480e-01 2.96823263e-01 5.08425713e-01 7.23262727e-01
-7.43219674e-01 4.66937989e-01 -1.07024705e+00 -4.81299967e-01
8.69803131e-01 4.76049781e-01 -1.68882206e-01 -4.18847334e-03
-5.30869365e-01 1.13996804e+00 3.94340158e-01 5.27333329e-03
-9.83343542e-01 -9.22481418e-01 -6.89279020e-01 -4.11787480e-01
5.62582612e-01 -9.21556652e-01 1.43815863e+00 -1.68066733e-02
-1.00738156e+00 9.63606715e-01 -4.94061112e-01 -6.30430102e-01
9.12201881e-01 -5.65062702e-01 -3.07374418e-01 -2.71755069e-01
4.03988183e-01 8.78954947e-01 9.12246048e-01 -1.30110729e+00
-1.42753303e+00 -3.92989576e-01 -3.73345673e-01 -9.24922079e-02
1.39423609e-01 3.10172021e-01 -9.67809677e-01 -1.64995611e-01
-1.03877541e-02 -1.04402030e+00 -6.63135350e-02 7.91773975e-01
-1.21135168e-01 -4.79580522e-01 1.40919983e+00 -1.00446068e-01
1.07104325e+00 -2.00778747e+00 -3.00929397e-01 -2.44159639e-01
2.03443095e-01 3.93470287e-01 1.79138616e-01 2.05258634e-02
3.87326449e-01 -3.00913483e-01 1.67217880e-01 -7.53316700e-01
1.69923753e-01 1.66494071e-01 -2.75207847e-01 8.02707076e-01
2.55422354e-01 9.83951867e-01 -1.05788791e+00 -9.33740735e-01
5.67697525e-01 4.99673396e-01 -5.91002345e-01 9.22730565e-02
-4.14383054e-01 2.45820925e-01 -5.04194438e-01 8.53558898e-01
1.05007350e+00 5.25841340e-02 -3.53684425e-01 -1.18601650e-01
-6.72792673e-01 2.04592999e-02 -1.26771820e+00 1.59257114e+00
8.66369680e-02 4.62528497e-01 4.83874194e-02 -2.01582253e-01
6.26647353e-01 -1.00435533e-01 6.35744631e-01 -3.93249184e-01
2.51748860e-01 5.46440519e-02 -1.18580833e-01 -2.86620498e-01
6.67683303e-01 2.76212282e-02 6.29484281e-02 2.81128287e-02
-4.13413905e-02 3.54189146e-03 2.07592562e-01 9.54747796e-02
9.49616969e-01 6.49552822e-01 6.99380040e-02 1.10510625e-01
3.65193516e-01 2.77905643e-01 7.77741194e-01 9.00603831e-01
-5.05605400e-01 5.03194511e-01 -1.94350019e-01 -4.62832570e-01
-8.72399509e-01 -1.18061805e+00 -3.45109105e-01 1.10574353e+00
3.72945279e-01 -3.15267414e-01 -4.17831540e-01 -8.93553615e-01
5.06075799e-01 7.66557455e-01 -5.73800266e-01 3.41546461e-02
-5.75751424e-01 -4.57950205e-01 5.55756509e-01 7.40465879e-01
1.60709634e-01 -7.63001800e-01 -1.02225447e+00 1.13553777e-01
2.97485650e-01 -1.36448574e+00 -4.79649067e-01 7.19233751e-02
-7.57183969e-01 -8.32476616e-01 -3.17310393e-01 -3.55401695e-01
4.99880373e-01 6.08646870e-01 7.16568947e-01 2.25469783e-01
-4.68660772e-01 4.77579944e-02 -5.18634841e-02 -9.71486866e-01
-1.77815869e-01 -1.23867877e-01 3.13692719e-01 -1.44415691e-01
5.71473658e-01 -1.94814876e-01 -5.00289977e-01 5.60016513e-01
-2.21735314e-01 6.17644303e-02 5.75837672e-01 1.97099581e-01
9.40414310e-01 -1.46313190e-01 1.75054923e-01 -5.07880867e-01
-3.96673560e-01 -4.56071854e-01 -1.25147414e+00 1.34132534e-01
-3.61105591e-01 1.48228658e-02 -8.65390301e-02 -6.16364539e-01
-6.89713597e-01 7.98316479e-01 -1.30764768e-01 -1.13768232e+00
-1.13402799e-01 -2.87475079e-01 -2.53308803e-01 -2.36165658e-01
2.86994845e-01 4.69298027e-02 -1.45432517e-01 -6.77326739e-01
5.86199760e-01 2.90389717e-01 8.27478409e-01 -4.22717601e-01
1.25453174e+00 8.55350673e-01 5.26583940e-02 -5.60796261e-01
-9.73172605e-01 -8.15041184e-01 -9.57635820e-01 -6.04084730e-01
1.02930021e+00 -1.12530351e+00 -1.03285408e+00 3.09877574e-01
-1.33241951e+00 -7.44432136e-02 -1.64147750e-01 4.15787876e-01
-3.44109118e-01 8.37491378e-02 3.94103304e-02 -1.17066360e+00
-1.86848670e-01 -9.19275701e-01 1.62354386e+00 3.53435069e-01
7.83073381e-02 -5.16514838e-01 2.59588182e-01 1.87181249e-01
1.15727544e-01 1.72371432e-01 1.72678486e-01 -5.55620551e-01
-9.40932274e-01 -5.58201671e-01 -3.47548574e-01 -2.51918316e-01
-2.87480652e-01 3.65653396e-01 -1.16244435e+00 -2.46248633e-01
-4.84532744e-01 1.33767307e-01 9.43148315e-01 3.93319994e-01
1.04909086e+00 2.78013170e-01 -1.10781515e+00 6.87251866e-01
1.18075573e+00 -6.73387870e-02 2.49627560e-01 2.28895590e-01
8.08462560e-01 2.78958082e-01 1.14197648e+00 3.73429060e-01
6.85795009e-01 9.88006592e-01 8.10436845e-01 1.02833852e-01
-3.55854243e-01 -4.80276823e-01 3.52244377e-01 1.46708190e-01
1.79030493e-01 -6.51433542e-02 -8.98912013e-01 6.53674185e-01
-2.21097875e+00 -1.21605992e+00 -5.42241275e-01 2.27725148e+00
3.89690548e-01 5.67315042e-01 4.89188939e-01 -2.68709272e-01
8.85047436e-01 -1.51269794e-01 -6.94414854e-01 2.31591389e-01
1.93815008e-01 -1.00330874e-01 1.02995741e+00 5.03783226e-01
-1.34493685e+00 1.14859533e+00 6.02384329e+00 8.32363009e-01
-8.50159287e-01 3.55083436e-01 -3.36037457e-01 -4.12489921e-01
1.57679453e-01 1.57437861e-01 -1.90186059e+00 5.62699676e-01
8.05307865e-01 -4.67272624e-02 -1.46936312e-01 9.67956841e-01
2.25768760e-01 -1.66183040e-01 -1.20670891e+00 9.13802862e-01
-1.44476309e-01 -1.27612793e+00 -2.05834314e-01 6.53326958e-02
3.31559509e-01 6.27671778e-01 -1.49242610e-01 4.87849861e-01
4.65126932e-01 -5.02394617e-01 1.35338640e+00 5.01502633e-01
5.63306153e-01 -5.64368427e-01 3.76924157e-01 7.32947111e-01
-1.69754875e+00 -5.34150675e-02 -3.05617869e-01 1.49209738e-01
4.40872490e-01 3.09836417e-01 -9.42125320e-01 5.37608862e-01
8.89078677e-01 5.92901587e-01 -7.78072238e-01 1.60441709e+00
-4.31447886e-02 3.17018598e-01 -7.32885003e-01 9.56958905e-03
6.96494654e-02 1.42760620e-01 7.48979747e-01 1.17031276e+00
3.89462531e-01 -1.62805721e-01 4.13199514e-01 9.27075624e-01
1.71823174e-01 -4.73103493e-01 -5.46849549e-01 4.22221720e-01
8.58299136e-01 1.48203731e+00 -8.08538020e-01 -3.42074335e-01
-4.36934233e-01 5.33210337e-01 3.57194632e-01 -1.66694179e-01
-1.46721613e+00 3.82897630e-02 8.75854135e-01 4.72970068e-01
1.12164521e+00 -1.38632074e-01 -2.18117803e-01 -6.31656051e-01
6.87374920e-02 -9.49470773e-02 3.52355301e-01 -8.08900416e-01
-9.66337264e-01 2.33818933e-01 2.70527720e-01 -1.54598212e+00
-1.26808316e-01 -5.08935392e-01 -6.71553731e-01 5.89996636e-01
-1.71275783e+00 -1.59930623e+00 -3.80881935e-01 4.27925140e-01
4.76011038e-01 1.99701712e-01 2.98907727e-01 7.93632746e-01
-6.60706937e-01 5.38363338e-01 -1.34461656e-01 -5.05712852e-02
5.43416321e-01 -9.00326610e-01 6.70215309e-01 9.96982336e-01
-9.45379492e-04 3.16918105e-01 9.14772570e-01 -1.09505737e+00
-1.51609910e+00 -1.62009645e+00 8.66907537e-01 -1.19989347e+00
6.69040561e-01 -5.82774222e-01 -7.69051552e-01 8.40323925e-01
-2.71430582e-01 6.31796837e-01 1.55398086e-01 -5.87581657e-02
-9.28008333e-02 -1.78159192e-01 -1.08989930e+00 4.21452910e-01
1.34063280e+00 -1.34335384e-01 -4.90038633e-01 2.08625555e-01
7.76399851e-01 -7.47410178e-01 -2.98488557e-01 6.16185963e-01
7.79364586e-01 -5.71652353e-01 1.15092313e+00 -4.98322248e-01
-3.47453594e-01 -1.09998381e+00 -2.34068949e-02 -5.70783615e-01
-3.46715540e-01 -5.34594834e-01 -5.77058554e-01 1.33217943e+00
1.72239989e-01 -1.44368023e-01 9.26289141e-01 3.66790056e-01
-2.64933079e-01 -4.95620221e-01 -1.12763727e+00 -1.05790985e+00
-3.53002995e-01 -8.18673849e-01 7.52299905e-01 3.67330700e-01
-7.82311797e-01 2.82040447e-01 -2.92555094e-01 7.20518470e-01
1.14960515e+00 1.42935529e-01 1.40549016e+00 -1.67515826e+00
1.22027136e-01 -3.06745142e-01 -4.63985533e-01 -1.08211839e+00
1.76875759e-02 -7.80842364e-01 4.75780636e-01 -1.47614789e+00
2.17089295e-01 -6.20511055e-01 -8.47329870e-02 4.47648048e-01
-3.65119040e-01 1.33272916e-01 2.13823169e-01 4.41723168e-02
-1.14123595e+00 4.54659998e-01 7.69183636e-01 2.84296218e-02
-4.69422303e-02 3.47188383e-01 -2.29248747e-01 8.91341209e-01
4.05067772e-01 -1.04539859e+00 -5.87652139e-02 -5.16511202e-01
-1.86794642e-02 -3.07717562e-01 8.39394569e-01 -1.21997702e+00
7.62178779e-01 -1.74708024e-01 3.34351152e-01 -1.72686386e+00
5.26895344e-01 -1.14930952e+00 3.24314326e-01 5.13748527e-01
3.37544620e-01 1.40939981e-01 5.82729697e-01 5.94949424e-01
5.06305516e-01 -9.39642042e-02 7.20781326e-01 3.43613714e-01
-9.38503802e-01 7.41763949e-01 1.28830671e-01 -2.10431740e-01
1.38595164e+00 -5.55814385e-01 -1.48630321e-01 3.32281351e-01
-4.02713031e-01 7.01222181e-01 5.45083940e-01 7.98882365e-01
4.39588904e-01 -1.39219344e+00 -6.64083362e-01 2.32946113e-01
3.39518905e-01 2.32067287e-01 3.45874242e-02 8.43169153e-01
-1.16802737e-01 2.94695735e-01 -1.64468412e-03 -1.23399532e+00
-1.72290730e+00 7.68921673e-01 2.28770405e-01 5.68205602e-02
-7.36973464e-01 8.33327234e-01 9.95523334e-02 -3.68293226e-01
4.06770647e-01 -3.57928425e-01 -1.09420693e-03 -1.78746972e-02
6.28676772e-01 4.84718859e-01 6.09355941e-02 -9.17906165e-01
-8.04780781e-01 8.52413297e-01 -2.24405974e-01 1.84729192e-02
1.09148383e+00 -8.46062303e-02 4.24324632e-01 3.99445593e-01
7.07660496e-01 -7.48709217e-03 -1.66034102e+00 -1.73476130e-01
1.96655560e-02 -5.46044052e-01 1.31533682e-01 -4.23396081e-01
-8.57249916e-01 6.77102447e-01 9.34804261e-01 -2.85756558e-01
3.47441763e-01 2.07420260e-01 7.40616500e-01 2.41614744e-01
5.76812446e-01 -9.35629487e-01 -4.83767271e-01 4.49479878e-01
3.83582145e-01 -1.37258601e+00 3.48288596e-01 -4.21319664e-01
-2.72523105e-01 5.48930645e-01 8.98897827e-01 1.00444563e-01
6.28024340e-01 5.70721388e-01 -2.11614043e-01 -3.04700047e-01
-8.09599280e-01 -6.36706114e-01 3.47390234e-01 5.60512185e-01
-1.50851771e-01 -1.55135453e-01 1.28766358e-01 4.35463786e-01
3.45279346e-04 2.44165212e-01 -2.16292366e-01 1.19692254e+00
-8.37044656e-01 -8.20717812e-01 -6.08929694e-01 3.03994894e-01
-2.24980205e-01 3.22795093e-01 -2.28123546e-01 8.76365364e-01
6.68303490e-01 6.57987416e-01 3.05514842e-01 -2.99478114e-01
5.02328157e-01 4.38417755e-02 6.33395493e-01 -5.26097476e-01
-3.84169102e-01 -7.59489685e-02 -2.77321637e-02 -5.83094716e-01
-3.17333519e-01 -1.09424555e+00 -1.41743135e+00 -3.17567676e-01
-7.41158545e-01 -3.30078453e-01 8.12188566e-01 9.78017986e-01
4.44888026e-01 4.63591397e-01 2.05002517e-01 -1.26092219e+00
-4.19642419e-01 -6.01117969e-01 -3.46488059e-02 1.76707536e-01
6.49652600e-01 -1.20410299e+00 6.49211928e-02 -2.33370345e-02] | [6.672203540802002, -2.219566822052002] |
fe44b236-5983-4c36-8969-dbb0ddc8f32f | sub-lexical-dialogue-act-classification-in-a | null | null | https://aclanthology.org/W13-3915 | https://aclanthology.org/W13-3915.pdf | Sub-lexical Dialogue Act Classification in a Spoken Dialogue System Support for the Elderly with Cognitive Disabilities | null | ['Yoshihiro Fujita', 'Shinichi Onaka', 'Minoru Kamata', 'Ken Sadohara', 'Hiroaki Kojima', 'Takenobu Inoue', 'Misato Nihei', 'Takuya Narita'] | 2013-08-01 | null | null | null | ws-2013-8 | ['dialogue-act-classification'] | ['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] | [-7.353734016418457, 3.7106878757476807] |
b8a05338-2c80-4863-a527-d75c2dcee129 | towards-a-generic-multimodal-architecture-for | 2108.04343 | null | https://arxiv.org/abs/2108.04343v1 | https://arxiv.org/pdf/2108.04343v1.pdf | Towards a Generic Multimodal Architecture for Batch and Streaming Big Data Integration | Big Data are rapidly produced from various heterogeneous data sources. They are of different types (text, image, video or audio) and have different levels of reliability and completeness. One of the most interesting architectures that deal with the large amount of emerging data at high velocity is called the lambda architecture. In fact, it combines two different processing layers namely batch and speed layers, each providing specific views of data while ensuring robustness, fast and scalable data processing. However, most papers dealing with the lambda architecture are focusing one single type of data generally produced by a single data source. Besides, the layers of the architecture are implemented independently, or, at best, are combined to perform basic processing without assessing either the data reliability or completeness. Therefore, inspired by the lambda architecture, we propose in this paper a generic multimodal architecture that combines both batch and streaming processing in order to build a complete, global and accurate insight in near-real-time based on the knowledge extracted from multiple heterogeneous Big Data sources. Our architecture uses batch processing to analyze the data structures and contents, build the learning models and calculate the reliability index of the involved sources, while the streaming processing uses the built-in models of the batch layer to immediately process incoming data and rapidly provide results. We validate our architecture in the context of urban traffic management systems in order to detect congestions. | ['Dalila Chiadmi', 'Maryem Rhanoui', 'Siham Yousfi'] | 2021-08-09 | null | null | null | null | ['data-integration'] | ['knowledge-base'] | [-4.43191439e-01 1.55381730e-03 -2.33361460e-02 -3.30513686e-01
-4.53869462e-01 -3.75892401e-01 5.41971803e-01 9.49088454e-01
-2.81504869e-01 4.21617568e-01 2.42766306e-01 5.84479868e-02
-4.48914856e-01 -1.27863777e+00 -3.59278351e-01 -4.70585883e-01
-2.63194859e-01 7.82183468e-01 7.65644431e-01 -5.11359155e-01
1.24803931e-01 5.58203161e-01 -2.35375285e+00 7.37936556e-01
8.27842116e-01 1.40128732e+00 1.53702825e-01 7.48835146e-01
-6.67486727e-01 1.05411386e+00 -5.64614236e-01 -2.31747091e-01
-2.81180311e-02 1.54857591e-01 -5.70361674e-01 -8.41171518e-02
-1.02874748e-01 -6.40182793e-02 2.16972306e-01 6.36725903e-01
5.27559757e-01 -1.45440400e-01 1.94306627e-01 -1.43989360e+00
1.49114773e-01 6.52777255e-01 -1.85364708e-01 1.57406583e-01
6.64935410e-01 1.19171068e-02 7.87460864e-01 -7.99931884e-01
6.42823875e-01 1.11578751e+00 2.58804619e-01 -6.75250366e-02
-9.04031754e-01 -3.43649268e-01 7.54312947e-02 6.86912119e-01
-1.21165156e+00 -6.68163717e-01 6.00842655e-01 -4.57384259e-01
5.85358560e-01 4.60404575e-01 6.05955422e-01 7.46898472e-01
1.12651154e-01 7.46550858e-01 7.41501987e-01 -1.82684317e-01
5.71932793e-01 3.40024412e-01 3.86045396e-01 1.33819416e-01
2.88006514e-01 -2.30832204e-01 -7.94809878e-01 5.00447080e-02
-5.80638386e-02 -1.43077578e-02 1.46526352e-01 -4.80113775e-02
-1.01631713e+00 4.39862788e-01 1.44962892e-01 6.75328314e-01
-7.46328413e-01 -3.37895215e-01 6.46027386e-01 5.48850775e-01
3.19548368e-01 -2.27081642e-01 -3.36904883e-01 -3.23598117e-01
-9.67268109e-01 1.81624770e-01 9.65094268e-01 9.45767879e-01
9.63194191e-01 -9.33063254e-02 -1.40794262e-01 5.24385571e-01
4.66198236e-01 4.47585642e-01 3.19896758e-01 -6.16529942e-01
7.52942145e-01 1.03701556e+00 -2.45125573e-02 -1.19638360e+00
-8.04354131e-01 -2.39050597e-01 -9.69330728e-01 5.02166331e-01
4.13576186e-01 -8.35816935e-02 -3.47273201e-01 1.41413438e+00
6.05897605e-01 -9.12330076e-02 3.19395155e-01 5.85517883e-01
1.19399261e+00 8.77953947e-01 1.06754199e-01 -3.57866257e-01
1.38845491e+00 -5.42293847e-01 -8.67825449e-01 3.14749181e-01
6.02061927e-01 -8.59833777e-01 7.74388373e-01 8.46802115e-01
-1.49773598e+00 -8.93705904e-01 -7.79684961e-01 2.04665400e-02
-8.51706684e-01 -1.11090407e-01 -1.62850060e-02 6.08865440e-01
-1.21375597e+00 3.87583852e-01 -4.14610058e-01 -3.95121634e-01
-6.03675917e-02 1.79452091e-01 -2.34016612e-01 -2.31571384e-02
-1.22162426e+00 6.46261930e-01 5.06097078e-01 4.73874584e-02
-6.65929198e-01 -7.68056214e-01 -4.22910333e-01 2.63452470e-01
4.81918335e-01 -5.98221719e-01 7.92245150e-01 -7.98659980e-01
-1.29828238e+00 4.31172311e-01 -1.01934507e-01 -2.89877981e-01
5.53694665e-01 -3.54811959e-02 -8.01948488e-01 2.91281104e-01
-1.58475086e-01 2.27283716e-01 7.42857635e-01 -1.24905586e+00
-1.10903788e+00 -5.40173292e-01 -1.06527999e-01 -4.80590537e-02
-4.96831506e-01 1.91775203e-01 -5.14729023e-01 -2.70835668e-01
-1.58778965e-01 -2.36240432e-01 -8.54228139e-02 -2.77226895e-01
1.83884632e-02 -4.24935371e-01 1.00984323e+00 -4.37227607e-01
1.59696198e+00 -2.22107244e+00 2.11004376e-01 6.02859437e-01
4.87851351e-01 2.45689705e-01 -7.39328116e-02 9.59338665e-01
2.26624072e-01 -4.52767015e-02 1.80106387e-01 -3.44032496e-01
-1.29369661e-01 2.03909293e-01 -2.02017352e-01 -2.56218500e-02
-1.98108815e-02 5.00669539e-01 -8.29758406e-01 -7.61852562e-01
4.58220333e-01 6.60094976e-01 -2.45089084e-01 3.96873832e-01
-1.33382320e-01 3.60436529e-01 -4.77190763e-01 5.15316308e-01
7.37467945e-01 3.45565714e-02 6.69811070e-02 -3.92508745e-01
-4.41921413e-01 -2.60322899e-01 -1.81418884e+00 1.41237009e+00
-6.53781354e-01 3.76198441e-01 5.62462807e-01 -9.38644409e-01
1.13006008e+00 5.70918739e-01 8.44133675e-01 -9.61310148e-01
1.82420447e-01 3.62664998e-01 -3.84454936e-01 -9.03535903e-01
7.99542129e-01 1.90703183e-01 -3.54613946e-03 4.74881649e-01
2.76696589e-02 5.25630452e-02 7.38735259e-01 2.70912260e-01
1.09868526e+00 -3.40570390e-01 -2.52656825e-02 1.54939353e-01
9.55311239e-01 -7.71835744e-02 2.33867481e-01 4.09884691e-01
-9.60461702e-03 5.38474679e-01 6.02589548e-01 -6.15310371e-01
-1.06259513e+00 -9.86470938e-01 -5.75286848e-03 1.15822542e+00
2.21013725e-01 -9.01450872e-01 -5.00950158e-01 -2.35798344e-01
-1.47331387e-01 3.96597058e-01 -2.17624813e-01 1.69436961e-01
-2.70202249e-01 -4.91430074e-01 3.02089542e-01 1.56024665e-01
3.82885307e-01 -1.05694246e+00 -9.85422134e-01 3.38748336e-01
-3.43168288e-01 -1.05636489e+00 5.05608320e-01 -2.78856128e-01
-7.80829966e-01 -1.05658185e+00 -1.82122141e-01 -1.09187365e-01
1.26392096e-01 2.00300276e-01 1.34235680e+00 2.00856462e-01
-1.32076755e-01 4.09329265e-01 -8.51275802e-01 -4.84962553e-01
-5.89805067e-01 1.01432875e-01 -7.19248801e-02 4.99069601e-01
2.55056191e-02 -5.76208293e-01 -3.24296981e-01 3.61013770e-01
-1.35478830e+00 -7.60518759e-02 6.28533125e-01 1.79010984e-02
3.52451980e-01 5.92275083e-01 7.23874032e-01 -6.64291620e-01
7.19302952e-01 -9.61557508e-01 -4.69772100e-01 2.28052065e-01
-4.72904474e-01 -1.00636054e-02 1.00575745e+00 1.03825815e-01
-1.12720835e+00 -3.42786938e-01 -4.69170094e-01 -1.67382032e-01
-6.33599520e-01 7.20710099e-01 -3.35217625e-01 3.12772661e-01
5.66879213e-01 -1.41527131e-01 3.98739539e-02 -7.43873358e-01
3.62856567e-01 8.78996909e-01 3.31278533e-01 -4.12912488e-01
5.83830953e-01 6.81486487e-01 -4.04453389e-02 -1.03000867e+00
-3.01724076e-01 -7.61512220e-01 -7.38743186e-01 -8.80559564e-01
7.39423096e-01 -5.50770879e-01 -9.70907986e-01 5.37582695e-01
-1.12279403e+00 5.58771379e-02 -6.09360695e-01 3.69792998e-01
-3.75676662e-01 2.28921890e-01 -3.82815272e-01 -8.66354585e-01
-3.81287664e-01 -9.34334457e-01 9.80528653e-01 1.57109603e-01
8.65623131e-02 -7.04305410e-01 6.56156540e-02 3.05415004e-01
7.49051929e-01 2.13064760e-01 6.21049464e-01 -6.56716585e-01
-6.34005725e-01 -4.89714205e-01 -3.04586142e-01 7.02586398e-02
-1.51276067e-01 4.47452009e-01 -1.07938731e+00 -9.83257033e-03
9.40509897e-04 -1.42888203e-01 5.76350868e-01 6.84536621e-02
9.95729327e-01 -2.08096981e-01 -1.32912740e-01 1.27798036e-01
1.48416734e+00 -1.21379212e-01 7.48846233e-01 1.69203490e-01
3.85986209e-01 1.25951087e+00 5.98460734e-01 9.56373572e-01
7.00472534e-01 5.12856066e-01 8.26526046e-01 2.86171790e-02
-3.59736606e-02 1.92057833e-01 4.91559237e-01 9.39930856e-01
-3.11442554e-01 -3.70828688e-01 -9.82819259e-01 3.74186933e-01
-2.13088679e+00 -1.32845986e+00 -7.53782868e-01 2.41987586e+00
2.57618874e-01 4.86921847e-01 5.80038846e-01 6.83377504e-01
5.25376678e-01 1.81686848e-01 -8.76569301e-02 -5.25754809e-01
-6.36422411e-02 -8.99571329e-02 8.73884708e-02 3.32870156e-01
-6.63557410e-01 3.45899403e-01 6.15214157e+00 8.26934218e-01
-1.08542192e+00 1.18838124e-01 2.98266441e-01 -3.33122462e-01
-3.88943762e-01 -2.48248845e-01 -7.76660204e-01 7.53328443e-01
1.58611739e+00 -2.15383306e-01 7.97166601e-02 6.15838826e-01
5.70354640e-01 -3.73404413e-01 -8.65781665e-01 9.13160503e-01
-8.23743865e-02 -1.37362385e+00 3.28823589e-02 -2.78166775e-02
2.01262921e-01 2.36568779e-01 -2.28170261e-01 4.70524952e-02
-2.32388172e-02 -6.52746677e-01 9.96388495e-01 1.04730487e+00
1.66902855e-01 -7.73516238e-01 9.43402946e-01 7.15351880e-01
-1.57260323e+00 -5.09653091e-01 -2.22605526e-01 -1.00120038e-01
5.71886897e-01 1.16915178e+00 -3.15995693e-01 1.09042478e+00
9.42257702e-01 5.74464798e-01 -7.85119414e-01 1.30876398e+00
2.57244140e-01 3.68318230e-01 -6.12737715e-01 1.07403457e-01
-4.64729108e-02 -3.10125314e-02 5.66911280e-01 1.18156326e+00
3.82136822e-01 -2.01868787e-01 2.20884547e-01 4.66586709e-01
3.36264879e-01 4.99953896e-01 -6.02705479e-01 4.77945954e-01
2.97341228e-01 1.65113008e+00 -6.52990758e-01 -6.69033945e-01
-4.74029571e-01 -2.54036188e-02 1.37385532e-01 2.05199853e-01
-5.69708943e-01 -4.18056846e-01 3.94992471e-01 3.78565282e-01
6.23313859e-02 -2.52383232e-01 -3.37260604e-01 -1.01403832e+00
2.30050981e-01 -5.28204381e-01 7.77582288e-01 -8.58289957e-01
-1.18007457e+00 8.28294516e-01 1.95938706e-01 -1.37749851e+00
-4.04463679e-01 -3.07215691e-01 -6.26900256e-01 5.57071865e-01
-1.64741671e+00 -1.05354786e+00 -7.54411876e-01 1.00377297e+00
4.60691482e-01 -5.30641861e-02 5.57429671e-01 7.14056313e-01
-5.06819546e-01 -5.96710760e-03 1.00207508e-01 -3.89531940e-01
6.57416165e-01 -1.06246614e+00 -2.12745786e-01 6.27548456e-01
-1.14692010e-01 -1.28263608e-01 6.93602145e-01 -5.23199677e-01
-1.47488475e+00 -6.75272882e-01 1.07711232e+00 -3.76315594e-01
6.48811877e-01 -2.90462554e-01 -1.04758251e+00 -1.11548109e-02
2.74058342e-01 -7.77861997e-02 5.57319641e-01 1.30348340e-01
-2.52221376e-01 -9.35957968e-01 -1.15881550e+00 2.07075197e-02
5.63162982e-01 -1.51686296e-01 -4.85445827e-01 8.25779587e-02
5.51676631e-01 1.57818228e-01 -1.21965384e+00 3.13976765e-01
3.55785429e-01 -1.64905310e+00 7.73410916e-01 -3.41833264e-01
2.12571099e-01 -4.12967682e-01 -3.69950682e-02 -9.09631014e-01
-5.00010177e-02 -3.72580588e-01 -2.77736247e-01 1.46108973e+00
3.17933828e-01 -4.24236178e-01 6.00752652e-01 6.73845232e-01
-2.35876873e-01 -3.55090439e-01 -9.72461939e-01 -4.92094696e-01
-5.28600931e-01 -1.11130524e+00 8.22897434e-01 3.96218210e-01
2.28574246e-01 1.76770627e-01 -2.24867523e-01 1.12361014e-01
6.85380280e-01 1.09951898e-01 9.86717820e-01 -1.67264485e+00
1.88394308e-01 -5.57370663e-01 -6.14531219e-01 -3.42696905e-01
-3.54967564e-01 -5.24604619e-01 -2.75694340e-01 -1.82786846e+00
-3.50168675e-01 -6.01198196e-01 -2.17033058e-01 1.78165704e-01
3.53042841e-01 8.31593871e-02 3.51485878e-01 4.36137944e-01
-8.92277658e-01 1.81484193e-01 7.79574573e-01 1.36400133e-01
-4.69472438e-01 2.11260572e-01 -2.87041724e-01 6.56586230e-01
6.93465769e-01 -4.16360199e-02 -4.66732711e-01 -5.02372265e-01
9.03973579e-01 9.11089871e-03 2.05870762e-01 -1.56593132e+00
8.07268679e-01 -5.78694977e-02 1.11013219e-01 -1.03278434e+00
2.38519460e-01 -1.30920517e+00 3.27320814e-01 2.59842515e-01
-2.96823494e-02 1.71816781e-01 6.70218617e-02 4.10652995e-01
-5.78283846e-01 -1.25620604e-01 4.87490982e-01 -5.11538982e-03
-9.16413248e-01 1.99679837e-01 -5.79355776e-01 -8.69506076e-02
1.32769680e+00 -1.59006611e-01 -4.11937565e-01 -4.04567510e-01
-9.34418082e-01 4.85437334e-01 1.72848374e-01 5.23558497e-01
6.08554363e-01 -1.17843032e+00 -8.50570023e-01 5.29032111e-01
2.36268133e-01 9.28462893e-02 6.35055065e-01 9.71274912e-01
-4.11076993e-01 1.65108770e-01 -4.73191023e-01 -6.56059861e-01
-1.10787976e+00 9.77450907e-01 -1.15948170e-01 -4.08907086e-01
-4.13919330e-01 5.70515357e-02 -4.82011527e-01 -1.22651339e-01
3.80581915e-01 -2.85506576e-01 -7.44359255e-01 8.94559979e-01
9.72409964e-01 7.64590144e-01 4.47221994e-01 -8.95230711e-01
-3.29403967e-01 7.17406332e-01 4.38939005e-01 -2.28194669e-01
1.55831003e+00 -5.32048225e-01 -2.69875020e-01 7.30768025e-01
8.81925523e-01 3.65950838e-02 -6.93024576e-01 -1.57981217e-01
1.66826159e-01 -4.10081893e-01 -1.52622759e-01 -5.22823513e-01
-1.15173459e+00 9.68989372e-01 4.54003274e-01 1.08206296e+00
1.50628805e+00 1.84708297e-01 6.84308290e-01 1.60927653e-01
4.21963036e-01 -1.45322275e+00 -6.12793528e-02 3.13297004e-01
8.53821874e-01 -1.01547599e+00 -2.07014859e-01 -1.70244321e-01
-5.42895019e-01 1.43859732e+00 3.20249468e-01 3.00437659e-01
1.05725062e+00 5.15101016e-01 9.74643379e-02 -3.07987481e-01
-1.19783187e+00 -4.96110767e-01 1.40930384e-01 6.69334114e-01
1.52292296e-01 -2.52883404e-01 -4.11717027e-01 4.78575975e-01
2.80451942e-02 3.30630153e-01 5.57458341e-01 8.48895013e-01
-8.77386689e-01 -1.35710561e+00 -7.71503985e-01 3.81346405e-01
-1.19819559e-01 3.60794753e-01 -5.40818498e-02 3.67879927e-01
3.46193880e-01 1.39797902e+00 1.46566480e-01 -5.20460367e-01
7.72503793e-01 2.09421329e-02 -1.99468538e-01 -1.89470857e-01
-9.19810891e-01 3.74180339e-02 1.61032453e-01 -8.73498619e-01
-6.33205593e-01 -4.70180422e-01 -1.24903095e+00 -7.38730431e-01
1.94696605e-01 2.91816115e-01 1.07841635e+00 8.22632015e-01
6.42322958e-01 4.29292411e-01 8.17823470e-01 -8.23409736e-01
-8.97865295e-02 -7.64253497e-01 -4.83400494e-01 5.65018356e-01
3.11636329e-01 -3.10529113e-01 -1.54599547e-01 -5.81909530e-02] | [8.452091217041016, 0.19973288476467133] |
8e2afee3-3cc1-48af-9ec2-5e965ce20394 | diversify-your-vision-datasets-with-automatic | 2305.16289 | null | https://arxiv.org/abs/2305.16289v1 | https://arxiv.org/pdf/2305.16289v1.pdf | Diversify Your Vision Datasets with Automatic Diffusion-Based Augmentation | Many fine-grained classification tasks, like rare animal identification, have limited training data and consequently classifiers trained on these datasets often fail to generalize to variations in the domain like changes in weather or location. As such, we explore how natural language descriptions of the domains seen in training data can be used with large vision models trained on diverse pretraining datasets to generate useful variations of the training data. We introduce ALIA (Automated Language-guided Image Augmentation), a method which utilizes large vision and language models to automatically generate natural language descriptions of a dataset's domains and augment the training data via language-guided image editing. To maintain data integrity, a model trained on the original dataset filters out minimal image edits and those which corrupt class-relevant information. The resulting dataset is visually consistent with the original training data and offers significantly enhanced diversity. On fine-grained and cluttered datasets for classification and detection, ALIA surpasses traditional data augmentation and text-to-image generated data by up to 15\%, often even outperforming equivalent additions of real data. Code is avilable at https://github.com/lisadunlap/ALIA. | ['Trevor Darrell', 'Joseph E. Gonzalez', 'Jiezhi Yang', 'Han Zhang', 'Alyssa Umino', 'Lisa Dunlap'] | 2023-05-25 | null | null | null | null | ['image-augmentation'] | ['computer-vision'] | [ 6.22722507e-01 6.98316172e-02 2.87290573e-01 -7.21449912e-01
-3.52788240e-01 -1.02982116e+00 8.85908961e-01 2.67938554e-01
-5.92504323e-01 7.70142257e-01 3.37072536e-02 -2.26318121e-01
2.57195175e-01 -6.72710180e-01 -1.01605248e+00 -1.64587989e-01
2.44546339e-01 6.12988234e-01 5.26245013e-02 -1.77083522e-01
1.33134052e-01 6.60264969e-01 -1.87099218e+00 5.45556843e-01
6.27676606e-01 5.77337742e-01 2.38617599e-01 8.96286428e-01
-3.57703805e-01 6.15874350e-01 -8.23727727e-01 -2.81400710e-01
4.34792787e-01 -2.84049273e-01 -7.01835215e-01 3.59833151e-01
1.17704117e+00 -3.21564794e-01 -2.94968516e-01 1.02224874e+00
2.57713199e-01 -1.74129121e-02 5.33022404e-01 -1.45673442e+00
-1.15016401e+00 2.84887850e-01 -3.90476167e-01 2.73973554e-01
3.00781876e-01 5.68190575e-01 4.82002139e-01 -9.84490395e-01
8.33297551e-01 1.33035171e+00 7.48904109e-01 8.03615570e-01
-1.71464276e+00 -6.46285594e-01 2.52691031e-01 -1.74746603e-01
-1.29165816e+00 -5.90152323e-01 5.14444888e-01 -6.43174648e-01
7.95146167e-01 1.93545938e-01 3.73079330e-01 1.08516049e+00
-2.08717316e-01 3.71578157e-01 1.16864967e+00 -5.54953218e-01
-4.33734991e-02 3.30582917e-01 9.74532366e-02 7.02110648e-01
4.27639633e-01 3.42391342e-01 -5.05957305e-01 -8.12788904e-02
6.04628682e-01 -7.22584575e-02 -3.89311202e-02 -4.60873693e-01
-1.54393685e+00 6.75133586e-01 5.19731224e-01 -1.75091311e-01
-2.24611476e-01 -9.10816714e-02 4.33090836e-01 5.02252162e-01
3.46569926e-01 1.03158903e+00 -5.72836816e-01 2.46313304e-01
-7.63795853e-01 5.08906543e-01 5.74534655e-01 1.19346774e+00
9.49261129e-01 3.32725316e-01 -3.23710814e-02 8.93119156e-01
-2.73072324e-03 8.20751190e-01 5.33008933e-01 -9.64461803e-01
1.56888321e-01 7.05272913e-01 -2.51296386e-02 -7.59701252e-01
-2.95714796e-01 -2.99601108e-01 -7.10833728e-01 5.82464397e-01
5.20346940e-01 -8.75778198e-02 -1.34033775e+00 1.73314536e+00
2.39329100e-01 -1.16292901e-01 1.45948619e-01 4.48588043e-01
1.14305019e+00 4.98900741e-01 1.95691496e-01 2.49995470e-01
1.15444720e+00 -6.98952496e-01 -3.09631348e-01 -7.19123900e-01
5.66997588e-01 -8.42604160e-01 1.40715587e+00 1.29132316e-01
-7.15251029e-01 -8.10481608e-01 -9.75955904e-01 -4.54783551e-02
-8.82623255e-01 7.83329904e-02 4.60831463e-01 3.60877812e-01
-1.14309704e+00 2.66366154e-01 -3.23157012e-01 -6.39426291e-01
6.39262736e-01 6.70687109e-02 -7.12359130e-01 -2.68918335e-01
-8.35710466e-01 8.43195796e-01 7.24171698e-01 -2.29715243e-01
-9.69844043e-01 -9.85674381e-01 -1.23082018e+00 -5.95588326e-01
2.55611897e-01 -6.65305734e-01 1.42675459e+00 -1.09931850e+00
-7.14549601e-01 1.31521893e+00 7.20640868e-02 -6.11398637e-01
4.85257775e-01 -1.60502091e-01 -3.80857289e-01 -7.50612319e-02
3.34458202e-01 1.24791121e+00 1.07522285e+00 -1.41677046e+00
-6.55521989e-01 -2.48576581e-01 5.13883755e-02 7.37903789e-02
-3.27957571e-02 -1.90293044e-01 -2.59186059e-01 -1.00379515e+00
-3.21091652e-01 -9.24979389e-01 -2.82974362e-01 5.39581180e-01
-2.40755647e-01 3.28502268e-01 1.04716003e+00 -5.24062872e-01
7.88383305e-01 -2.16209435e+00 -2.68382013e-01 1.19122593e-02
3.34448576e-01 5.49679875e-01 -7.33276010e-01 3.07282746e-01
-1.02377810e-01 1.76509768e-01 -3.73729795e-01 -1.14772230e-01
-2.12502718e-01 4.52088475e-01 -4.88665015e-01 2.43148327e-01
6.27589345e-01 9.62760627e-01 -8.63754094e-01 -2.75117904e-01
4.81930196e-01 2.61627674e-01 -5.49640179e-01 2.82042205e-01
-5.68677604e-01 6.39026046e-01 -1.28243610e-01 6.49408937e-01
5.64794362e-01 -4.02656421e-02 -2.02048510e-01 -2.20613301e-01
-5.97798377e-02 -1.32214427e-01 -1.07594132e+00 1.49514353e+00
-4.43436414e-01 8.71517479e-01 -1.47385180e-01 -7.41660595e-01
1.09323168e+00 -2.19833404e-01 -3.17727625e-02 -8.09309721e-01
-1.37605712e-01 -1.37304533e-02 -5.20586185e-02 -4.82738912e-01
6.13749266e-01 2.32853461e-02 -2.29554102e-01 4.61416870e-01
4.66290377e-02 -5.47989190e-01 3.74250740e-01 3.94089729e-01
9.15038288e-01 2.38575056e-01 3.66683215e-01 -8.51877034e-02
1.92708552e-01 5.64386547e-01 3.57201248e-01 1.07676160e+00
-1.47510350e-01 6.28720999e-01 -4.63414490e-02 -6.94146693e-01
-1.53020966e+00 -9.34181511e-01 -2.49934688e-01 1.22941482e+00
-8.76529329e-03 -1.49064779e-01 -6.70627415e-01 -6.26760125e-01
3.52298945e-01 8.63796294e-01 -8.61715555e-01 -3.29793721e-01
-1.26510739e-01 -4.54515845e-01 9.69903886e-01 5.53798318e-01
5.72869003e-01 -1.39899993e+00 -7.47623265e-01 2.64953543e-02
8.10198262e-02 -1.24428082e+00 -3.78645360e-01 2.52856493e-01
-6.42392218e-01 -9.98764455e-01 -4.40622598e-01 -8.76526296e-01
1.07116330e+00 2.44035199e-01 1.43560565e+00 1.21551357e-01
-8.26733410e-01 5.31050086e-01 -4.94496912e-01 -7.23333657e-01
-8.90206218e-01 -3.41438144e-01 1.51631549e-01 -2.22141877e-01
5.73186338e-01 -2.43492782e-01 -1.53063208e-01 2.56909460e-01
-1.06024158e+00 2.27419421e-01 5.15595436e-01 1.05327559e+00
7.97257483e-01 -1.63459972e-01 5.68365932e-01 -1.21889651e+00
5.39889276e-01 -1.55487090e-01 -7.16963530e-01 2.12503448e-01
-3.76924992e-01 2.75578886e-01 6.84208930e-01 -6.37902379e-01
-9.62689459e-01 2.84932077e-01 2.26291135e-01 -5.01400292e-01
-8.26618433e-01 2.64753938e-01 -6.01889915e-04 -2.20572889e-01
1.33713281e+00 2.75295317e-01 9.23289880e-02 -4.58370060e-01
7.43973553e-01 6.17643774e-01 1.06414640e+00 -4.07686621e-01
1.13457322e+00 4.31742400e-01 -3.71139288e-01 -8.68049860e-01
-7.04810858e-01 -1.52720392e-01 -9.53865409e-01 6.29879981e-02
6.19585574e-01 -9.71711099e-01 -3.47345769e-02 6.85231209e-01
-1.00744212e+00 -5.93309402e-01 -7.23685503e-01 1.29857793e-01
-4.08233881e-01 8.81428346e-02 -1.35508537e-01 -4.30865020e-01
-8.04512128e-02 -8.17935288e-01 1.07926857e+00 2.86015749e-01
-4.56656307e-01 -8.62789571e-01 -1.24819219e-01 3.00436825e-01
3.42265785e-01 4.76246208e-01 9.00658011e-01 -7.00699866e-01
-3.02595705e-01 -3.63749534e-01 -3.49200934e-01 4.42634851e-01
3.60479921e-01 2.25389197e-01 -1.06282973e+00 -3.69117588e-01
-4.34718996e-01 -6.57827675e-01 6.30665779e-01 5.94430156e-02
1.19740474e+00 -3.12289894e-01 -2.19704747e-01 6.93327785e-01
1.13155973e+00 8.10539350e-02 4.06915456e-01 4.32584167e-01
8.06385815e-01 6.59116626e-01 7.87492931e-01 4.75779027e-01
1.21699564e-01 4.35182750e-01 4.24947381e-01 -5.09047151e-01
-4.75915313e-01 -4.90127385e-01 5.29658562e-03 -1.80660989e-02
5.04737496e-01 -7.82415494e-02 -9.83684897e-01 7.69000351e-01
-1.48029339e+00 -1.00338757e+00 -4.54465859e-02 2.17071199e+00
9.84026849e-01 1.24084935e-01 -1.59667619e-02 -2.04274178e-01
7.37021506e-01 -7.06730261e-02 -1.12891030e+00 -4.75863546e-01
-3.71005595e-01 1.14982054e-01 8.18100333e-01 4.51841325e-01
-1.33900321e+00 1.15904129e+00 6.66379547e+00 3.70992631e-01
-1.06268311e+00 -2.90343314e-01 4.74204957e-01 -1.03136413e-01
-3.53117883e-01 -3.48792285e-01 -6.87358439e-01 1.75333753e-01
6.18502617e-01 -2.29662091e-01 5.02847075e-01 8.67155433e-01
2.38109514e-01 3.22870798e-02 -1.30742383e+00 1.05031180e+00
2.16842115e-01 -1.48808324e+00 5.33172727e-01 -4.64132912e-02
6.96877301e-01 1.84015661e-01 1.11250386e-01 2.16342598e-01
8.49940300e-01 -1.08991504e+00 7.07449198e-01 4.69393015e-01
1.30654573e+00 -3.86618644e-01 5.17840028e-01 3.40680301e-01
-8.28556538e-01 1.19651584e-02 -3.93065840e-01 -1.36458397e-01
-1.97836205e-01 2.43429929e-01 -1.28595018e+00 -7.10254535e-02
7.96372712e-01 7.49733746e-01 -1.22372580e+00 8.68062973e-01
2.09384994e-03 3.51071209e-01 -3.24550718e-01 2.15720743e-01
1.01243339e-01 2.24892914e-01 5.27593911e-01 1.27665615e+00
-5.13879349e-03 -1.32870972e-01 3.05191547e-01 7.54163444e-01
-3.15582901e-01 -2.17597529e-01 -1.23240137e+00 -4.09425706e-01
7.25826204e-01 1.04156661e+00 -3.73357326e-01 -5.02673566e-01
-4.82643783e-01 1.02435577e+00 2.56276518e-01 3.61812145e-01
-5.88893056e-01 -4.13195372e-01 9.85774517e-01 1.66470334e-01
1.75485685e-01 -1.28319323e-01 -3.81028354e-01 -1.04730213e+00
-5.02626821e-02 -1.33727074e+00 3.87608349e-01 -1.13403165e+00
-1.41661894e+00 8.63184750e-01 2.07937695e-03 -1.36961758e+00
-3.73173326e-01 -7.07815111e-01 -3.01492989e-01 9.30051267e-01
-1.21791899e+00 -1.44642413e+00 -7.07460701e-01 7.10713983e-01
9.11889195e-01 -4.87951845e-01 1.05607402e+00 3.82752270e-02
-1.26512319e-01 7.23456740e-01 9.77377221e-02 2.74304688e-01
9.34263945e-01 -1.26493502e+00 9.66064572e-01 1.05194628e+00
3.89610887e-01 5.17822981e-01 8.48443389e-01 -7.32357025e-01
-1.09854817e+00 -1.68996108e+00 5.13207436e-01 -7.72070050e-01
6.04388475e-01 -3.48851830e-01 -1.14384544e+00 8.56473684e-01
-4.08303924e-02 2.16493845e-01 6.32905066e-01 -2.25973815e-01
-9.02521431e-01 -1.33778444e-02 -1.32504213e+00 6.81285620e-01
1.15817726e+00 -6.79628134e-01 -7.41277337e-01 3.74392033e-01
6.59478724e-01 -4.49220747e-01 -5.42725742e-01 3.16610038e-01
3.90873343e-01 -5.10017753e-01 9.49882388e-01 -1.00393224e+00
4.06298220e-01 -7.62858570e-01 -2.48567328e-01 -1.38281882e+00
-3.98095399e-01 -2.17266411e-01 4.49072450e-01 1.23531687e+00
4.81836319e-01 -5.92701674e-01 4.71854687e-01 7.45335698e-01
-4.74860258e-02 9.53949057e-03 -3.76462758e-01 -7.69888818e-01
8.51440281e-02 -4.03644443e-01 6.80928648e-01 1.06521118e+00
-5.27687848e-01 2.55374312e-01 -2.76630640e-01 3.00172091e-01
7.13359594e-01 1.03406280e-01 1.34074593e+00 -1.17525613e+00
-9.05441195e-02 -2.23465383e-01 -6.81672812e-01 -6.75449193e-01
-1.00303786e-02 -9.73939061e-01 1.34423181e-01 -1.40409565e+00
5.64227253e-02 -6.41623318e-01 1.49713427e-01 9.72157240e-01
-6.20044544e-02 7.27278709e-01 2.73432285e-01 2.41397962e-01
-3.65533441e-01 2.22906500e-01 1.10271597e+00 -3.42454702e-01
-1.64477348e-01 -3.30708057e-01 -1.04879105e+00 8.27049017e-01
8.26039016e-01 -3.40725183e-01 -5.24740994e-01 -7.60444403e-01
-1.75255299e-01 -5.33865988e-01 7.46187687e-01 -1.05326438e+00
-1.65902644e-01 -3.02576512e-01 7.33958840e-01 -3.37328643e-01
1.40811488e-01 -8.29385459e-01 2.30561092e-01 2.29056269e-01
-7.54943550e-01 7.49082342e-02 8.65401924e-01 4.92936999e-01
-1.89111114e-01 -7.95184448e-02 9.27707613e-01 -3.31478506e-01
-1.33386981e+00 3.13302070e-01 -3.77411574e-01 3.20828736e-01
1.08411300e+00 -4.49857622e-01 -7.03939736e-01 -2.23377749e-01
-7.27724075e-01 3.00176561e-01 8.91755044e-01 7.34288156e-01
6.23541296e-01 -1.20456576e+00 -1.00855494e+00 6.84631348e-01
7.59679735e-01 2.36540303e-01 1.77914470e-01 -1.46582797e-02
-7.86364257e-01 1.07143402e-01 -6.14152253e-01 -6.34880722e-01
-1.61460221e+00 6.59975111e-01 3.81689072e-01 2.53227055e-01
-6.64197385e-01 9.05541837e-01 5.95265627e-01 -7.18759894e-01
7.19736889e-02 -2.86305934e-01 -6.70536980e-02 -1.35613233e-01
7.81564474e-01 -2.28200182e-01 -3.15313563e-02 -8.07330012e-01
-4.02903289e-01 4.10039395e-01 -3.67272019e-01 -1.57851935e-03
1.24192977e+00 -1.06119327e-01 2.30095670e-01 2.21636519e-01
8.64167869e-01 -8.32657516e-02 -1.58229494e+00 -5.42393982e-01
-1.62535265e-01 -6.49738848e-01 -1.83969572e-01 -1.23567343e+00
-7.22991765e-01 9.30886328e-01 6.85825288e-01 3.57289091e-02
1.15163946e+00 1.32653475e-01 2.59172171e-01 6.83006525e-01
2.63630152e-01 -8.37989330e-01 3.00126877e-02 6.13698065e-01
1.24303603e+00 -1.58070219e+00 -2.20516585e-02 -1.82663515e-01
-9.61733997e-01 9.63860571e-01 9.26141202e-01 1.06094860e-01
3.48013788e-01 3.83499414e-01 6.20977104e-01 -4.32481505e-02
-6.28082454e-01 -3.27580035e-01 1.80347800e-01 1.35414839e+00
1.67454064e-01 1.06029272e-01 4.33522224e-01 4.41679433e-02
-4.77709055e-01 1.49249464e-01 6.58127129e-01 9.09207404e-01
-4.14776206e-01 -9.48081017e-01 -5.81854165e-01 7.25723207e-01
-1.44986510e-02 -2.24328190e-01 -7.27337360e-01 7.42812276e-01
4.09764946e-01 8.20253193e-01 3.22348893e-01 -2.90971786e-01
5.43473721e-01 5.76403290e-02 2.37582237e-01 -1.05372405e+00
-3.79839301e-01 -3.43580723e-01 2.56775230e-01 -3.49968314e-01
-3.20124745e-01 -7.39675403e-01 -1.04336417e+00 -2.98351616e-01
9.34684202e-02 -2.25477740e-01 5.32134235e-01 7.60295868e-01
5.36429286e-01 4.65557992e-01 2.07708314e-01 -9.14398253e-01
-2.70427078e-01 -8.08240294e-01 -2.90721953e-01 7.97467232e-01
6.38817608e-01 -5.23038685e-01 -1.48992777e-01 7.72792459e-01] | [9.96374225616455, 1.7049845457077026] |
3e02c187-f28f-446a-9058-661add5d1ef2 | diving-deep-into-clickbaits-who-use-them-to | 1703.09400 | null | http://arxiv.org/abs/1703.09400v1 | http://arxiv.org/pdf/1703.09400v1.pdf | Diving Deep into Clickbaits: Who Use Them to What Extents in Which Topics with What Effects? | The use of alluring headlines (clickbait) to tempt the readers has become a
growing practice nowadays. For the sake of existence in the highly competitive
media industry, most of the on-line media including the mainstream ones, have
started following this practice. Although the wide-spread practice of clickbait
makes the reader's reliability on media vulnerable, a large scale analysis to
reveal this fact is still absent. In this paper, we analyze 1.67 million
Facebook posts created by 153 media organizations to understand the extent of
clickbait practice, its impact and user engagement by using our own developed
clickbait detection model. The model uses distributed sub-word embeddings
learned from a large corpus. The accuracy of the model is 98.3%. Powered with
this model, we further study the distribution of topics in clickbait and
non-clickbait contents. | ['Naeemul Hassan', 'Md Main Uddin Rony', 'Mohammad Yousuf'] | 2017-03-28 | null | null | null | null | ['clickbait-detection'] | ['natural-language-processing'] | [-4.55518156e-01 -5.88359125e-03 -3.83582383e-01 -2.03658622e-02
-7.75094748e-01 -4.71587509e-01 8.97578597e-01 6.24737263e-01
-6.53745234e-01 4.51972663e-01 5.66873968e-01 -6.06634915e-01
-1.22508638e-01 -7.17816591e-01 -4.51256007e-01 -2.56570783e-02
3.28240812e-01 1.50231853e-01 8.55874538e-01 -1.92628488e-01
8.03093076e-01 -2.15295583e-01 -1.48513508e+00 2.10020483e-01
8.14739048e-01 1.06356096e+00 2.28932500e-01 5.53778946e-01
-6.20797455e-01 6.70104444e-01 -6.11495912e-01 -5.19264638e-01
4.41528745e-02 -1.40142469e-02 -6.20661497e-01 -1.81167141e-01
7.20388412e-01 -4.41805750e-01 -3.98105383e-01 9.11686718e-01
2.38398969e-01 -2.19819441e-01 2.56376445e-01 -8.46341252e-01
-7.49725759e-01 4.87986207e-01 -4.53334779e-01 5.00195146e-01
1.57356352e-01 -9.15217251e-02 1.49536574e+00 -5.75386882e-01
7.69344032e-01 7.78952897e-01 5.24410188e-01 -1.88547909e-01
-9.73937392e-01 -6.21956289e-01 -1.08623505e-01 2.74140341e-03
-1.09688628e+00 1.61581546e-01 6.04109406e-01 -8.76030922e-01
3.85383636e-01 5.38360536e-01 8.39897096e-01 1.32813108e+00
2.70060450e-01 8.15709949e-01 1.36715531e+00 -3.52334946e-01
4.19661985e-04 6.45533979e-01 8.87471318e-01 2.47093782e-01
6.44954681e-01 -4.59468216e-01 -4.69131649e-01 -3.31785798e-01
2.78252870e-01 2.65143603e-01 3.41227539e-02 7.77889639e-02
-8.05477917e-01 1.14684498e+00 3.03590536e-01 7.64376760e-01
-3.05038363e-01 1.25016481e-01 3.84507746e-01 4.18356583e-02
9.50868428e-01 5.37493229e-01 -1.80251822e-01 -8.11911404e-01
-8.97088766e-01 5.10627806e-01 9.20498967e-01 5.29377759e-01
6.89433575e-01 -5.94768524e-01 -5.19529320e-02 1.00482821e+00
5.05419314e-01 1.21532768e-01 7.94092298e-01 -3.77118677e-01
6.06702864e-01 1.01633203e+00 3.16675276e-01 -1.41371834e+00
-1.38116807e-01 -7.55531669e-01 -6.83763102e-02 -2.18347013e-01
6.92616642e-01 -1.20229773e-01 -4.08418030e-01 9.29001033e-01
1.07172564e-01 -3.37407976e-01 -1.09238672e+00 5.22999167e-01
1.80341139e-01 5.46718299e-01 2.52506565e-02 1.62002414e-01
1.52428997e+00 -7.04222441e-01 -1.13116121e+00 -3.82782698e-01
6.50938809e-01 -1.20871758e+00 1.24190581e+00 3.07678252e-01
-6.96657956e-01 -3.35412264e-01 -9.65286911e-01 -2.41612822e-01
-8.92825067e-01 -1.39731586e-01 5.34423709e-01 1.01369345e+00
-2.53660649e-01 3.71060044e-01 -4.83894140e-01 -7.13427901e-01
6.82434201e-01 -3.15845728e-01 1.43068999e-01 -5.79090603e-02
-1.06610966e+00 8.70481968e-01 6.46746680e-02 -5.14942408e-01
-2.16856822e-01 -9.63189006e-01 2.00774387e-01 -8.65746960e-02
6.58325374e-01 -1.34466588e-02 1.32482910e+00 -6.83547139e-01
-8.34010124e-01 6.55028284e-01 1.65092677e-01 -3.12550157e-01
7.81747460e-01 -8.50123048e-01 -5.83764255e-01 -2.72453398e-01
1.51688352e-01 -1.95042104e-01 6.12498701e-01 -7.72857308e-01
-8.46465051e-01 -4.18982744e-01 -1.54693261e-01 -4.61557060e-01
-9.28000271e-01 2.92802483e-01 -4.11816895e-01 -4.45025384e-01
-5.06149791e-02 -7.33490407e-01 1.37790471e-01 -2.17167392e-01
-7.07946837e-01 -5.75716436e-01 7.91547239e-01 -1.03153229e+00
2.19871807e+00 -2.13755345e+00 -7.10802913e-01 1.77317977e-01
6.34871483e-01 2.17366308e-01 4.10203815e-01 1.15132618e+00
4.73604351e-01 8.11212659e-01 6.07463419e-01 2.67740600e-02
4.18298215e-01 -3.91748011e-01 -3.32171470e-01 2.94764191e-01
-2.67563879e-01 8.10007870e-01 -6.69570684e-01 -1.54606447e-01
-1.41686097e-01 1.39730632e-01 -6.94283366e-01 1.88741326e-01
-2.26987451e-01 -3.43285650e-01 -1.01677942e+00 4.41023588e-01
5.96001565e-01 -6.09892249e-01 7.36194029e-02 5.15712984e-02
-8.21090341e-01 6.96457267e-01 -5.78876555e-01 8.68665159e-01
-3.44142020e-01 1.17197084e+00 -4.05232608e-01 -1.87259480e-01
7.48108089e-01 -2.43143469e-01 2.80835718e-01 -7.51429737e-01
2.53407598e-01 3.10735017e-01 1.79554150e-01 -7.49845445e-01
7.40213215e-01 2.61871427e-01 -5.10371523e-03 7.63899028e-01
-4.13186431e-01 4.36688393e-01 6.96625710e-02 4.19610977e-01
1.29550004e+00 -3.26440781e-01 3.92469019e-01 -2.38687545e-01
-1.22295260e-01 2.72443872e-02 -1.67124167e-01 7.86638975e-01
-3.32698345e-01 2.73306608e-01 7.07721353e-01 -3.01400572e-01
-1.24325407e+00 -6.02917314e-01 -2.83004940e-01 1.31404972e+00
-1.95561484e-01 -7.33664215e-01 -7.38352001e-01 -6.86014473e-01
3.73002291e-01 8.65358651e-01 -5.33464372e-01 2.44205937e-01
-1.31177217e-01 -5.46170533e-01 2.64537245e-01 5.73154464e-02
7.78158009e-01 -4.93869960e-01 -3.81700218e-01 3.36196423e-01
-1.69999495e-01 -9.40656960e-01 -5.04732072e-01 -3.74277890e-01
-4.61638033e-01 -1.08178091e+00 -6.52083933e-01 -3.39559138e-01
1.51103795e-01 2.84472257e-01 7.04385221e-01 1.28543722e-02
-4.38647300e-01 2.56847143e-01 -4.23783034e-01 -4.21597749e-01
-8.32782760e-02 4.76896495e-01 -4.55680341e-01 1.55347317e-01
8.01041782e-01 -1.19645745e-01 -5.18174946e-01 4.05106962e-01
-9.41055357e-01 -4.21084881e-01 3.43408734e-01 4.47762638e-01
-3.05022448e-01 -3.23041826e-02 5.76272130e-01 -1.12633443e+00
1.09599793e+00 -8.69570076e-01 -3.91108096e-01 -1.49501711e-01
-7.00284123e-01 -4.77615356e-01 -2.47400507e-01 -5.25351107e-01
-8.53562415e-01 -8.72641802e-01 -2.38263011e-01 5.30057371e-01
-2.09017433e-02 8.30631196e-01 3.69423896e-01 1.93004027e-01
7.30723441e-01 -4.23125178e-03 1.03182338e-01 -9.47787762e-01
6.34565651e-02 1.19045615e+00 -4.64186102e-01 1.88194588e-01
9.48196650e-01 2.59455800e-01 -7.85681725e-01 -1.26033592e+00
-1.24515021e+00 -9.09292459e-01 -3.02817905e-03 -5.24029255e-01
8.89404297e-01 -4.72714037e-01 -8.65129352e-01 2.82003433e-01
-1.04572690e+00 5.20603172e-02 2.32905850e-01 4.77865726e-01
1.87991217e-01 3.96617264e-01 -5.29046714e-01 -1.01183438e+00
2.88407564e-01 -6.84018433e-01 3.65723372e-01 4.36650999e-02
-6.39397204e-01 -9.83351111e-01 2.80476660e-01 9.07217205e-01
9.79173541e-01 2.24979408e-02 8.07083726e-01 -1.44526339e+00
-9.49139297e-01 -8.43228757e-01 -5.09126723e-01 3.77443284e-01
1.79609284e-01 5.14201038e-02 -9.04536426e-01 1.08729407e-01
-8.76507685e-02 -2.64225215e-01 5.33299387e-01 3.16763707e-02
1.11649323e+00 -3.80061060e-01 -4.27793801e-01 -4.26922590e-01
1.44344604e+00 -1.82461783e-01 6.82856917e-01 1.06814623e+00
3.66134614e-01 7.37851381e-01 5.89904606e-01 9.22580123e-01
1.95697039e-01 5.13947964e-01 6.47863269e-01 5.57134390e-01
9.15221721e-02 -7.75808215e-01 6.06085137e-02 1.24549043e+00
9.75072011e-03 -1.90526485e-01 -1.05015731e+00 3.93128306e-01
-1.43393338e+00 -8.33454311e-01 -5.89625239e-01 2.16651320e+00
5.21157563e-01 5.93066871e-01 2.52006501e-01 7.42045045e-02
5.61435223e-01 4.71164972e-01 3.48034315e-02 -1.65709838e-01
3.43554974e-01 1.43089946e-02 8.65424871e-01 2.44328544e-01
-7.54917145e-01 3.98371279e-01 5.81776190e+00 9.85495090e-01
-7.35863566e-01 2.79507667e-01 4.73444372e-01 3.07319015e-01
-3.67773890e-01 -9.40222293e-02 -1.15712273e+00 9.79795158e-01
9.40342724e-01 -3.48039806e-01 1.49360925e-01 1.04930639e+00
5.93842924e-01 -2.63894349e-01 -4.53543127e-01 5.84130645e-01
1.53793469e-01 -1.48527980e+00 -3.46762240e-01 5.95748246e-01
5.39828241e-01 1.08741954e-01 2.68031001e-01 5.20609915e-01
3.03851021e-03 -5.01547337e-01 5.52932501e-01 3.12063664e-01
1.92469075e-01 -2.34284639e-01 7.42130399e-01 4.62380737e-01
-3.73120815e-01 -4.29880500e-01 -8.24347660e-02 -2.05919921e-01
3.13303530e-01 9.69698429e-01 -8.50282907e-01 -1.76861498e-03
4.89288837e-01 5.88183105e-01 -1.12582338e+00 1.47595489e+00
2.69995332e-01 1.11430013e+00 -1.61345422e-01 -8.90152216e-01
5.07528007e-01 -1.95708469e-01 3.86796325e-01 9.72004235e-01
3.06772053e-01 -8.43418419e-01 -1.01960115e-01 6.30719006e-01
-2.45384172e-01 3.82868588e-01 -4.21409279e-01 -9.83177900e-01
5.70675910e-01 1.19350970e+00 -4.11214650e-01 -9.72460955e-02
-7.28638887e-01 5.87134957e-01 2.11445048e-01 1.88556910e-01
-9.30371344e-01 -6.16660237e-01 5.95780075e-01 1.02298331e+00
3.14400375e-01 -4.82728362e-01 -2.33499542e-01 -7.15026915e-01
1.76932395e-01 -7.16938496e-01 -4.01453190e-02 -6.26803637e-01
-1.47561610e+00 2.08390862e-01 -2.66748875e-01 -9.26162899e-01
3.96203011e-01 -6.02093399e-01 -5.11230707e-01 6.97642803e-01
-1.44996035e+00 -8.87490392e-01 -5.09037733e-01 -2.15490565e-01
6.18435919e-01 1.08990356e-01 2.43715197e-01 8.17106605e-01
-5.20499229e-01 2.79470861e-01 3.34952205e-01 5.85782416e-02
7.35135019e-01 -1.09647262e+00 2.11911425e-01 1.19508043e-01
-2.47102797e-01 1.23959279e+00 8.49507093e-01 -8.87008071e-01
-1.16248405e+00 -7.40231395e-01 1.52887201e+00 -6.39005661e-01
1.87524581e+00 -5.17076492e-01 -7.57535279e-01 8.02575469e-01
5.82620978e-01 -6.00455046e-01 1.05234945e+00 6.15429461e-01
-3.78880411e-01 -5.21391854e-02 -6.92047000e-01 8.22351217e-01
6.65451288e-01 -5.48476040e-01 -8.22997510e-01 6.69582009e-01
5.56115866e-01 4.39398319e-01 -7.74801254e-01 -7.23477602e-01
7.98099101e-01 -1.06854177e+00 5.88404953e-01 -7.24929869e-01
6.39120162e-01 3.24305445e-01 -2.00223908e-01 -1.00351942e+00
-3.25394452e-01 -6.39970601e-01 1.52484655e-01 1.62925267e+00
4.13501173e-01 -1.13257349e+00 8.35071683e-01 6.62762880e-01
2.36314777e-02 -4.67981249e-01 -9.41091537e-01 -6.53683305e-01
-9.65488702e-02 -4.52866018e-01 2.55299568e-01 8.28159273e-01
3.86721059e-03 2.03769073e-01 -2.29513705e-01 -3.33977163e-01
3.40053618e-01 -2.47353628e-01 1.04037642e+00 -1.44053125e+00
-2.33927161e-01 -3.94979596e-01 -3.21336299e-01 -1.38792014e+00
-5.74974537e-01 -5.58139741e-01 -4.77123320e-01 -1.39130104e+00
3.80917162e-01 -3.51751357e-01 -2.43376613e-01 -1.36809260e-01
6.65326342e-02 1.20917946e-01 -2.19326112e-02 2.84682214e-01
-6.64383173e-01 1.64065555e-01 1.08543015e+00 8.33943114e-02
2.08972648e-01 5.51271066e-02 -9.02662635e-01 6.37632549e-01
7.60215223e-01 -4.51548010e-01 7.05556795e-02 4.98660542e-02
9.07462001e-01 -5.60490131e-01 1.69621751e-01 -8.15603733e-01
-4.64383624e-02 -8.87419004e-03 -1.29661560e-01 -6.25560760e-01
1.66905597e-01 -8.62166762e-01 -1.53599724e-01 3.15524131e-01
-6.72306478e-01 1.93095356e-02 -2.58340836e-01 8.25408697e-01
-2.65084594e-01 -5.50624192e-01 3.32092583e-01 8.23600739e-02
-1.99394315e-01 3.80233899e-02 -9.23484266e-01 2.73240447e-01
8.26640964e-01 -1.72904417e-01 -7.36061633e-01 -4.92796481e-01
-3.06757569e-01 -2.43685931e-01 5.03928475e-02 5.83388209e-01
9.69140679e-02 -1.14518094e+00 -2.70753592e-01 -1.03155069e-01
3.96240234e-01 -8.77038360e-01 1.69505998e-01 9.89678383e-01
-5.85731149e-01 7.22559988e-01 6.36389256e-02 -4.68183756e-02
-1.04088748e+00 1.26289725e-01 -3.25318158e-01 -2.34365076e-01
-2.89092720e-01 4.26630110e-01 -3.05091828e-01 9.43155885e-02
5.13044465e-03 -4.18957472e-01 -1.59656569e-01 6.65280163e-01
6.07450247e-01 9.55959499e-01 -1.10323563e-01 -6.62863404e-02
1.09657288e-01 -6.26896620e-02 -4.35126454e-01 -2.51604300e-02
1.34344292e+00 -2.68895805e-01 -1.81764513e-01 9.68525469e-01
1.61551976e+00 5.47277451e-01 -6.53319895e-01 4.90386412e-02
6.61472321e-01 -1.10079432e+00 2.10479051e-01 -6.86217666e-01
-4.67690915e-01 7.69510269e-01 3.83250654e-01 1.30684984e+00
-1.38866184e-02 4.10954133e-02 9.50902700e-01 1.48862168e-01
8.89416039e-02 -1.50453138e+00 5.25578916e-01 3.45431805e-01
6.24715269e-01 -1.08954847e+00 1.38719335e-01 -4.34256285e-01
-3.33973855e-01 8.72770846e-01 3.00722510e-01 -9.50708240e-02
1.05619621e+00 -3.71811002e-01 -1.11857103e-02 -2.69609034e-01
-5.10798573e-01 5.46582490e-02 2.09816154e-02 8.06186423e-02
7.16795504e-01 -9.71544981e-02 -9.54022467e-01 6.46480620e-01
-2.25039534e-02 6.65800273e-02 7.02440321e-01 8.17925990e-01
-9.69366133e-01 -9.64008927e-01 1.57142300e-02 1.03961551e+00
-1.19657314e+00 -2.40455177e-02 -3.43450427e-01 1.11374843e+00
-2.49029577e-01 8.72517526e-01 1.20999403e-01 -2.75265902e-01
2.06421733e-01 3.36481929e-01 -2.68885642e-01 -4.85040754e-01
-8.04298759e-01 8.42059478e-02 6.76313519e-01 -3.32737237e-01
2.79915687e-02 -7.46173024e-01 -3.40794027e-01 -4.24783707e-01
-6.17475986e-01 4.59667817e-02 1.19521928e+00 5.60831785e-01
3.79976660e-01 4.75501001e-01 6.84660316e-01 4.11033211e-03
-8.55753660e-01 -1.36969435e+00 -7.91890562e-01 5.96034646e-01
6.09537102e-02 -6.94741845e-01 -8.27157736e-01 -3.89406472e-01] | [7.75368070602417, 9.78248119354248] |
e15aec1a-e61a-4f82-b462-bcaae86f119e | disentangled-recurrent-wasserstein-1 | 2101.07496 | null | https://arxiv.org/abs/2101.07496v1 | https://arxiv.org/pdf/2101.07496v1.pdf | Disentangled Recurrent Wasserstein Autoencoder | Learning disentangled representations leads to interpretable models and facilitates data generation with style transfer, which has been extensively studied on static data such as images in an unsupervised learning framework. However, only a few works have explored unsupervised disentangled sequential representation learning due to challenges of generating sequential data. In this paper, we propose recurrent Wasserstein Autoencoder (R-WAE), a new framework for generative modeling of sequential data. R-WAE disentangles the representation of an input sequence into static and dynamic factors (i.e., time-invariant and time-varying parts). Our theoretical analysis shows that, R-WAE minimizes an upper bound of a penalized form of the Wasserstein distance between model distribution and sequential data distribution, and simultaneously maximizes the mutual information between input data and different disentangled latent factors, respectively. This is superior to (recurrent) VAE which does not explicitly enforce mutual information maximization between input data and disentangled latent representations. When the number of actions in sequential data is available as weak supervision information, R-WAE is extended to learn a categorical latent representation of actions to improve its disentanglement. Experiments on a variety of datasets show that our models outperform other baselines with the same settings in terms of disentanglement and unconditional video generation both quantitatively and qualitatively. | ['Xuan Zhang', 'Li Erran Li', 'Ligong Han', 'Martin Renqiang Min', 'Jun Han'] | 2021-01-19 | disentangled-recurrent-wasserstein | https://openreview.net/forum?id=O7ms4LFdsX | https://openreview.net/pdf?id=O7ms4LFdsX | iclr-2021-1 | ['unconditional-video-generation'] | ['computer-vision'] | [ 3.20880294e-01 1.52904674e-01 -3.82013083e-01 -1.66585401e-01
-4.88345146e-01 -6.63336337e-01 8.56067598e-01 -7.44679332e-01
-2.82050576e-03 8.00282657e-01 6.57666266e-01 -2.01601058e-01
-1.86998665e-01 -7.51018763e-01 -8.25662553e-01 -1.03753173e+00
-6.47201389e-03 4.21712250e-01 -3.23542565e-01 -8.67907181e-02
-3.25022727e-01 3.29413503e-01 -1.33506560e+00 1.56213477e-01
7.19088137e-01 3.97747308e-01 7.51569271e-02 8.58446360e-01
4.04617697e-01 8.96296382e-01 -4.13809985e-01 -2.44238630e-01
3.42287034e-01 -8.00155342e-01 -4.58849639e-01 4.29702371e-01
1.91162854e-01 -5.74235559e-01 -8.20815861e-01 7.48904705e-01
2.63657600e-01 1.74157649e-01 1.04577303e+00 -1.55119741e+00
-1.13724637e+00 6.08579636e-01 -5.62260747e-01 1.27772227e-01
6.41848668e-02 2.77497441e-01 1.29799592e+00 -8.48551989e-01
7.18351245e-01 1.23537314e+00 5.47880381e-02 7.95634508e-01
-1.81834507e+00 -8.34309757e-01 1.80432275e-01 2.95516104e-02
-1.07389426e+00 -2.28044480e-01 8.85442853e-01 -7.36652136e-01
6.86214209e-01 2.72340536e-01 7.57518709e-01 1.73038805e+00
1.38882250e-01 1.13429654e+00 1.00115025e+00 -2.26953477e-01
9.83471572e-02 -2.15256661e-01 -6.48225248e-02 7.17417836e-01
4.14492846e-01 4.19234157e-01 -7.44877100e-01 -1.29310666e-02
1.31583929e+00 2.54020214e-01 -2.65660971e-01 -7.18005955e-01
-1.61262584e+00 1.06860137e+00 2.06899896e-01 7.39769563e-02
-3.42759341e-01 1.69197887e-01 2.48017818e-01 4.53753769e-01
4.94914621e-01 3.56536299e-01 -2.59103119e-01 -1.40843540e-01
-5.38858294e-01 2.45921075e-01 5.67241311e-01 9.71929729e-01
6.84173703e-01 4.83544499e-01 -3.96680176e-01 4.73551989e-01
3.67238700e-01 6.14071667e-01 6.41540885e-01 -8.79116833e-01
7.58541584e-01 4.92164522e-01 7.42837414e-02 -9.25875962e-01
1.30169466e-01 -1.47406727e-01 -1.19785166e+00 2.58847088e-01
2.40228951e-01 -4.89789128e-01 -9.79677320e-01 2.14690614e+00
-1.14768393e-01 4.67910528e-01 2.31309235e-01 8.90206218e-01
5.84325552e-01 8.69804025e-01 -2.05294967e-01 -3.25756401e-01
1.01729393e+00 -8.42826962e-01 -1.02355850e+00 -7.34225810e-02
4.21283990e-02 -4.17823136e-01 8.64815533e-01 2.21279711e-01
-1.00988817e+00 -5.49342573e-01 -1.13831294e+00 -1.00927152e-01
1.19888812e-01 2.37795785e-01 7.47119963e-01 2.86192238e-01
-8.13766003e-01 4.07735646e-01 -1.26364887e+00 4.97700572e-02
3.51983815e-01 2.14548081e-01 -6.34118378e-01 1.19880386e-01
-1.25463521e+00 5.37334740e-01 3.12462002e-01 9.89686139e-03
-1.29993021e+00 -5.41467547e-01 -1.03874052e+00 3.20607349e-02
5.27717590e-01 -1.10926425e+00 9.72162068e-01 -6.98770940e-01
-1.78725159e+00 4.09316331e-01 -6.45757765e-02 -2.90865779e-01
6.17958248e-01 -3.61116141e-01 -2.94994712e-01 -8.84973034e-02
-1.07397037e-02 5.65853357e-01 1.15315676e+00 -1.28042519e+00
-1.98318958e-02 -1.98283657e-01 1.80028319e-01 3.98513049e-01
-3.11646342e-01 -4.29616779e-01 -1.16150148e-01 -9.75002885e-01
1.37082413e-02 -1.10080838e+00 -1.87504083e-01 -7.50980750e-02
-5.10149360e-01 -1.13336988e-01 8.77447903e-01 -6.51216507e-01
1.03755093e+00 -2.09281111e+00 1.00525594e+00 -2.07454890e-01
6.04492426e-01 -9.62078422e-02 -2.28642061e-01 5.59320688e-01
-4.17409152e-01 2.44270861e-01 3.40331951e-03 -5.36946416e-01
7.43090585e-02 5.55061579e-01 -6.80999875e-01 3.08840871e-01
3.01261902e-01 1.04526532e+00 -1.08250368e+00 -1.34016335e-01
1.29286274e-01 4.78783011e-01 -5.62019587e-01 7.20808446e-01
-2.33363286e-01 6.88673437e-01 -3.96353543e-01 -1.36654023e-02
2.94992894e-01 -4.77155626e-01 3.20526719e-01 -2.09336951e-01
2.10484326e-01 1.42605007e-01 -1.03529596e+00 1.81866491e+00
-5.12489617e-01 8.73547792e-01 -5.86126387e-01 -7.85250366e-01
7.37828493e-01 7.19395876e-01 4.88862157e-01 -1.35096729e-01
2.76423693e-02 -2.51828998e-01 7.21367821e-02 -5.49282134e-01
3.75722021e-01 -3.28508973e-01 -1.09725043e-01 9.36549246e-01
3.82513463e-01 1.35614499e-01 8.51090401e-02 5.93520820e-01
8.46921921e-01 5.14046907e-01 3.76223475e-01 9.97302681e-02
-1.91531852e-02 -7.08680451e-01 6.69901609e-01 6.18797004e-01
1.35050267e-01 7.62636900e-01 7.42171586e-01 -2.38466233e-01
-1.24749577e+00 -1.55950105e+00 3.34139496e-01 8.57240736e-01
7.59752467e-02 -4.27899122e-01 -5.22625864e-01 -6.64027214e-01
-2.01739743e-01 8.54315221e-01 -1.01952183e+00 -4.31019008e-01
-4.91444796e-01 -7.09125221e-01 3.24926764e-01 7.11003006e-01
1.28786251e-01 -9.57323015e-01 -4.79257166e-01 -6.23353943e-02
-3.42378825e-01 -9.53182697e-01 -8.12635064e-01 -3.80702130e-02
-9.65934575e-01 -7.69648314e-01 -7.14123726e-01 -2.16686398e-01
8.15740228e-01 4.58804965e-01 8.91176581e-01 -5.38393378e-01
-1.62815496e-01 2.42242143e-01 -2.20365822e-01 -9.03218538e-02
-4.43435311e-01 -1.39691502e-01 2.09240481e-01 3.90087396e-01
1.03028648e-01 -9.03209448e-01 -5.35490751e-01 3.00145626e-01
-1.20851946e+00 7.62091517e-01 6.76022768e-01 1.20067894e+00
3.67363513e-01 -3.47746015e-01 4.09450591e-01 -7.77014434e-01
6.69779599e-01 -6.86766207e-01 -3.24211359e-01 2.21280947e-01
-5.01933634e-01 6.09723985e-01 4.26210850e-01 -8.61787558e-01
-1.18616307e+00 -2.35292435e-01 5.61915219e-01 -9.42783475e-01
-1.01060696e-01 5.08123398e-01 -3.64326596e-01 8.50554645e-01
5.30562937e-01 3.98506671e-01 1.78362504e-01 -2.77950495e-01
7.53183782e-01 2.90071964e-01 2.23109305e-01 -5.87305963e-01
9.67739105e-01 4.03080732e-01 -1.80657014e-01 -5.34237206e-01
-8.00070584e-01 -3.65551151e-02 -7.81630754e-01 -1.44034103e-01
9.98592615e-01 -9.11591768e-01 -4.74439234e-01 3.60650361e-01
-1.18869197e+00 -2.23210320e-01 -5.59817791e-01 6.63967371e-01
-8.83786440e-01 2.10236534e-01 -5.58139265e-01 -6.80551946e-01
5.35419816e-03 -1.10951078e+00 1.10641956e+00 8.54098052e-03
-3.90232027e-01 -1.14304364e+00 3.89866024e-01 4.29096520e-01
2.02325955e-02 6.59830928e-01 1.00275648e+00 -4.05825377e-01
-7.56734014e-01 -7.96775073e-02 2.08033212e-02 3.90067548e-01
5.23309410e-01 -2.38941796e-02 -7.68856823e-01 -3.75963539e-01
-2.98026744e-02 -4.32122737e-01 8.86831343e-01 2.30288371e-01
9.43800569e-01 -8.11731279e-01 5.59742674e-02 4.64022160e-01
1.02064896e+00 1.58287153e-01 6.91369712e-01 -2.22021222e-01
1.01421094e+00 3.28696907e-01 1.19564056e-01 4.47584450e-01
2.88864881e-01 6.17158413e-01 3.88391227e-01 1.15172826e-01
-9.63722989e-02 -6.40524566e-01 8.04293931e-01 1.21109986e+00
-5.76227844e-01 -4.43998635e-01 -4.28710729e-01 4.11529630e-01
-2.15646672e+00 -1.34340787e+00 1.79368943e-01 2.09932137e+00
8.97666395e-01 -2.84404475e-02 1.73009321e-01 -7.90049061e-02
6.23729944e-01 5.72958112e-01 -7.92724073e-01 -3.92364115e-02
-8.40229169e-02 5.21401688e-02 8.72440562e-02 5.54320097e-01
-8.08873236e-01 6.34600043e-01 5.93795252e+00 6.01707697e-01
-8.88773918e-01 1.45594999e-01 3.63428593e-01 -3.66152078e-01
-7.18983769e-01 -6.68180883e-02 -3.01390499e-01 4.17631537e-01
7.73088634e-01 -4.84169751e-01 5.46375334e-01 5.16345203e-01
2.45731011e-01 4.46208030e-01 -1.62132943e+00 8.96307826e-01
8.40100423e-02 -1.28386140e+00 5.52506208e-01 3.45430553e-01
1.11851060e+00 -3.04448307e-01 5.33338189e-01 2.92357445e-01
8.93898964e-01 -1.20897186e+00 7.09361851e-01 7.17035651e-01
7.17343390e-01 -6.21576488e-01 4.03919816e-01 6.10027194e-01
-8.44826102e-01 1.90184519e-01 -6.94138333e-02 -1.66475520e-01
2.63635099e-01 2.86945164e-01 -6.40971661e-01 7.56000459e-01
1.37915194e-01 1.09012401e+00 -2.41562679e-01 2.49397397e-01
-5.45661449e-01 6.27645433e-01 1.16447747e-01 1.62823200e-01
9.68509167e-02 -5.57264149e-01 7.94024885e-01 7.13786185e-01
1.95347160e-01 1.64224699e-01 9.02819335e-02 1.15211308e+00
-1.18360803e-01 -3.74782979e-01 -8.23656023e-01 -4.01249379e-01
1.38820693e-01 9.61988091e-01 -3.35529774e-01 -2.89523542e-01
-2.69341320e-01 1.17002416e+00 4.13060665e-01 8.17580700e-01
-8.11661601e-01 7.90086091e-02 8.72446954e-01 -1.46740258e-01
2.79027998e-01 -6.02676451e-01 -1.28168821e-01 -1.82032359e+00
-8.32879022e-02 -7.88596332e-01 2.93217123e-01 -8.34096134e-01
-1.36859751e+00 5.76961637e-01 4.33866233e-01 -1.48603594e+00
-7.48109221e-01 -3.57528895e-01 -4.87858087e-01 8.71668875e-01
-8.52368116e-01 -1.28912508e+00 -1.18625738e-01 5.67746758e-01
8.21916938e-01 -3.82104427e-01 9.87090528e-01 -1.80254802e-01
-6.98352277e-01 4.58986670e-01 2.56990552e-01 1.18802391e-01
4.14288163e-01 -1.37267303e+00 3.11799407e-01 9.25033331e-01
6.31791890e-01 8.55834424e-01 8.57794583e-01 -5.99780977e-01
-1.33557487e+00 -1.11726248e+00 4.56902236e-01 -6.24989033e-01
8.10412407e-01 -6.61047816e-01 -6.16359949e-01 1.21996665e+00
4.76232737e-01 -1.15338145e-02 1.05723405e+00 1.42229889e-02
-6.65081680e-01 2.08807305e-01 -5.89858949e-01 1.00767589e+00
1.17102563e+00 -7.28123784e-01 -5.70289493e-01 1.47554383e-01
1.02440739e+00 -3.61409158e-01 -8.08075428e-01 2.23727494e-01
7.61526167e-01 -9.26028550e-01 8.62587094e-01 -1.11009002e+00
9.89679456e-01 -1.63977623e-01 -8.32508951e-02 -1.64627647e+00
-5.42011738e-01 -8.64395738e-01 -7.54638433e-01 9.47881818e-01
3.59558553e-01 -3.91222894e-01 6.00335419e-01 5.50613999e-01
2.81544149e-01 -8.64527464e-01 -6.34939015e-01 -7.76976287e-01
2.75905579e-02 -2.47313425e-01 5.47981262e-01 9.37134027e-01
-2.49504253e-01 5.73522866e-01 -1.04766560e+00 1.82841122e-01
6.68076038e-01 2.95490593e-01 8.48725796e-01 -8.92734349e-01
-7.33277380e-01 -1.92981660e-01 -5.52156270e-01 -1.26059198e+00
3.29987407e-01 -9.05741394e-01 -1.95083499e-01 -1.35480785e+00
5.38429677e-01 9.58573595e-02 -4.08318728e-01 4.01329517e-01
-3.39580446e-01 3.21531645e-03 2.76916713e-01 4.00199652e-01
-3.20739180e-01 1.10310042e+00 1.63152969e+00 -2.49064803e-01
-1.01443306e-01 -5.62238432e-02 -6.87796354e-01 5.88790059e-01
4.88541871e-01 -4.44213450e-01 -9.91828918e-01 -5.62018633e-01
1.47760034e-01 4.50800121e-01 4.73060906e-01 -4.31247264e-01
-2.85515100e-01 -5.68982005e-01 2.97569662e-01 -3.12860906e-01
3.76313657e-01 -7.49288678e-01 5.56972027e-01 2.36138538e-01
-7.05059111e-01 2.10339520e-02 -2.24814221e-01 1.02145243e+00
-9.58650857e-02 1.24605939e-01 3.35804284e-01 4.75282222e-02
-2.87761867e-01 6.59025729e-01 -3.94061983e-01 -7.52699003e-02
1.08326733e+00 -1.81373000e-01 -2.58524060e-01 -6.85020149e-01
-1.02792406e+00 9.51209478e-03 2.03372270e-01 8.58984590e-01
7.94788420e-01 -1.88341296e+00 -8.07579935e-01 3.83412063e-01
9.19163078e-02 1.07705019e-01 4.66336817e-01 4.48724717e-01
-9.94679183e-02 2.19463989e-01 -3.50808471e-01 -5.24656892e-01
-1.16449797e+00 4.83057588e-01 -5.43341078e-02 -5.78793466e-01
-6.67806864e-01 5.45192778e-01 5.99711835e-01 -2.73250192e-01
-1.09064795e-01 -2.73601919e-01 -1.53970465e-01 3.58091295e-02
3.80814999e-01 3.72358054e-01 -5.34451425e-01 -7.14374244e-01
1.84964657e-01 1.90435767e-01 -2.53278226e-01 -4.19018477e-01
1.42258310e+00 -1.69103779e-02 1.04729973e-01 1.00888121e+00
1.38385820e+00 -2.76958585e-01 -1.89214098e+00 -3.03290457e-01
-5.76747656e-01 -5.74014843e-01 -2.01877579e-01 -4.15723532e-01
-1.13155866e+00 9.83969748e-01 3.33603680e-01 2.12540343e-01
8.06435406e-01 1.71108525e-02 5.02390742e-01 9.24919322e-02
2.00950414e-01 -5.08386731e-01 8.73684287e-01 2.65355676e-01
1.27563608e+00 -1.04448283e+00 -8.23234692e-02 -1.87458009e-01
-9.32328880e-01 9.22019899e-01 7.42844582e-01 -2.90440381e-01
4.52086002e-01 6.40227944e-02 -1.95087999e-01 -4.68565598e-02
-1.01219642e+00 1.59884840e-01 6.55672908e-01 5.41911483e-01
3.16174209e-01 3.67048651e-01 5.41161820e-02 5.73969603e-01
-3.48797381e-01 -1.04335517e-01 6.06755435e-01 6.49213195e-01
1.68067336e-01 -1.02667594e+00 -1.36485798e-02 3.22941631e-01
9.06968722e-04 1.53216973e-01 -1.21110745e-01 7.24559307e-01
-7.08887950e-02 6.47564232e-01 1.34345055e-01 -4.73254085e-01
-5.48342466e-02 -7.72883184e-04 6.57655716e-01 -7.34792292e-01
8.70809704e-02 1.24268107e-01 -2.81285256e-01 -3.95110577e-01
-6.08994067e-01 -7.05025136e-01 -8.60104144e-01 -1.41610086e-01
-3.76646519e-01 1.68149676e-02 1.50188684e-01 9.42953348e-01
3.59653562e-01 7.43083894e-01 8.02408457e-01 -1.01265669e+00
-7.82573640e-01 -9.57202852e-01 -5.89899540e-01 7.90912867e-01
6.42998219e-01 -7.42234945e-01 -3.01810473e-01 5.54888248e-01] | [10.918984413146973, 0.45350998640060425] |
99b51221-0cd5-46d5-b6a4-34dd830bc249 | dhsegment-a-generic-deep-learning-approach | 1804.10371 | null | https://arxiv.org/abs/1804.10371v2 | https://arxiv.org/pdf/1804.10371v2.pdf | dhSegment: A generic deep-learning approach for document segmentation | In recent years there have been multiple successful attempts tackling document processing problems separately by designing task specific hand-tuned strategies. We argue that the diversity of historical document processing tasks prohibits to solve them one at a time and shows a need for designing generic approaches in order to handle the variability of historical series. In this paper, we address multiple tasks simultaneously such as page extraction, baseline extraction, layout analysis or multiple typologies of illustrations and photograph extraction. We propose an open-source implementation of a CNN-based pixel-wise predictor coupled with task dependent post-processing blocks. We show that a single CNN-architecture can be used across tasks with competitive results. Moreover most of the task-specific post-precessing steps can be decomposed in a small number of simple and standard reusable operations, adding to the flexibility of our approach. | ['Benoit Seguin', 'Sofia Ares Oliveira', 'Frederic Kaplan'] | 2018-04-27 | null | null | null | null | ['document-layout-analysis'] | ['computer-vision'] | [ 4.58950251e-01 -2.04011753e-01 2.86266029e-01 -3.47108006e-01
-1.00155008e+00 -8.96485925e-01 8.82061720e-01 3.04014564e-01
-7.57153094e-01 4.41985369e-01 7.79527202e-02 -5.19430518e-01
-1.96796149e-01 -5.52069426e-01 -6.81803048e-01 -4.64819998e-01
-1.55912116e-01 1.60021320e-01 2.28976801e-01 -2.40623340e-01
6.06209576e-01 8.25639069e-01 -1.63454914e+00 7.25992143e-01
4.95782435e-01 1.05582988e+00 3.67787600e-01 1.24968970e+00
-3.56495380e-01 5.42606413e-01 -8.93283665e-01 -4.78258520e-01
3.96292239e-01 -1.10748723e-01 -7.30159581e-01 7.23156109e-02
9.28963006e-01 -2.88973153e-01 -4.05819006e-02 7.25267470e-01
7.28887856e-01 1.41782705e-02 5.43541372e-01 -9.38344896e-01
-5.43009043e-01 6.42519832e-01 -5.68965971e-01 2.93034047e-01
1.51957840e-01 8.46348852e-02 9.49097037e-01 -7.29351044e-01
7.64119148e-01 9.52578783e-01 9.42750990e-01 3.47650111e-01
-1.16559577e+00 -1.05017304e-01 2.52152205e-01 8.14710781e-02
-1.03666127e+00 -4.27295536e-01 8.70128930e-01 -4.11758155e-01
1.31900764e+00 4.66132432e-01 5.00527740e-01 1.21478355e+00
1.09080143e-01 8.36902678e-01 1.10808158e+00 -6.86212301e-01
6.44896477e-02 -2.14065284e-01 2.82626182e-01 5.70673048e-01
2.60136575e-01 -5.62137127e-01 -3.33371937e-01 2.14403123e-01
6.63625658e-01 -3.31421614e-01 -1.50799096e-01 1.75061356e-02
-1.18912268e+00 5.03035009e-01 2.51011103e-01 6.85256660e-01
-6.69572949e-02 3.30793262e-01 8.28532815e-01 3.41605365e-01
3.92560691e-01 5.99236012e-01 -8.12659025e-01 -2.02070206e-01
-1.47769856e+00 6.37553811e-01 8.75933290e-01 1.08255100e+00
3.42116445e-01 -3.28799598e-02 -3.48526865e-01 8.36883545e-01
-1.71570495e-01 3.18953558e-03 4.14899915e-01 -7.87571490e-01
8.23339164e-01 3.50728542e-01 -2.86203809e-02 -1.02327955e+00
-7.70258427e-01 -5.05677223e-01 -7.27067292e-01 1.49486288e-01
5.58928370e-01 -4.50476818e-02 -1.30618191e+00 1.30490208e+00
-9.51064974e-02 -4.72643703e-01 -3.21043789e-01 5.55087090e-01
5.09168386e-01 7.48339474e-01 5.08221723e-02 -3.41091864e-02
1.56089306e+00 -1.16700065e+00 -6.21843398e-01 -2.73105502e-01
5.04795611e-01 -9.49477136e-01 1.30681705e+00 8.44363809e-01
-1.36699533e+00 -5.11983752e-01 -1.39534616e+00 -7.44026244e-01
-8.93268824e-01 4.76524025e-01 4.76594359e-01 6.57028913e-01
-1.06942320e+00 8.07739973e-01 -6.43612504e-01 -3.25519323e-01
3.87462676e-01 4.95590895e-01 -2.43886009e-01 2.59646684e-01
-5.97701073e-01 9.48321939e-01 5.04435956e-01 3.57289553e-01
-3.63169521e-01 -7.28678107e-01 -4.44734842e-01 4.02316570e-01
3.94312829e-01 -7.41301239e-01 1.32844639e+00 -8.91800404e-01
-1.40355885e+00 8.52826715e-01 1.95561826e-01 -3.79746646e-01
9.61742938e-01 -3.94330680e-01 -2.51155555e-01 -6.14861026e-02
-3.23930085e-01 5.58445692e-01 1.10385299e+00 -1.16303968e+00
-7.06148744e-01 -3.37400913e-01 -3.52972522e-02 4.34279218e-02
-7.22432613e-01 1.37252361e-01 -8.59662473e-01 -1.09199476e+00
-2.64118850e-01 -8.24573338e-01 -1.62102997e-01 2.70703044e-02
-4.42560315e-01 6.89242855e-02 7.29822397e-01 -7.63065398e-01
1.38588810e+00 -2.03997540e+00 2.61605799e-01 4.61179316e-02
2.67252102e-02 2.71288395e-01 -2.95242250e-01 4.49029177e-01
2.51175407e-02 2.87216634e-01 -4.36176002e-01 -8.44454885e-01
2.66212732e-01 4.33007479e-02 -4.80990708e-01 2.90335447e-01
3.15169692e-01 9.25675273e-01 -6.55120730e-01 -4.95861083e-01
1.24619320e-01 5.29378474e-01 -1.68473795e-01 -1.78870723e-01
-4.17938769e-01 -1.51244134e-01 -1.78525388e-01 5.70205688e-01
7.44439483e-01 -1.40829861e-01 2.01095134e-01 -1.46561757e-01
-3.52123469e-01 3.16815436e-01 -1.31130767e+00 2.04407382e+00
-6.45305097e-01 1.16080105e+00 2.60977685e-01 -9.65029120e-01
6.81833684e-01 1.27283424e-01 5.36903068e-02 -7.60696530e-01
3.12361866e-01 3.53317708e-01 -2.11677462e-01 -2.84935236e-01
1.07406282e+00 3.51040453e-01 -1.12668119e-01 3.12768191e-01
2.82279551e-01 -1.31980583e-01 5.31353474e-01 -1.37947261e-01
1.16573942e+00 3.57180089e-01 2.01830253e-01 -3.74467105e-01
4.91505951e-01 2.79815853e-01 2.33804118e-02 9.34620500e-01
9.78170782e-02 1.00875354e+00 4.52800006e-01 -8.07836831e-01
-1.50029469e+00 -6.56082630e-01 -1.41249999e-01 1.35147893e+00
-5.60287237e-01 -5.94306648e-01 -8.26202750e-01 -4.19142604e-01
-2.04523847e-01 3.72321546e-01 -7.23199904e-01 5.43638766e-01
-1.12110639e+00 -8.98566723e-01 5.68717659e-01 6.49075091e-01
3.51969332e-01 -1.13096416e+00 -1.02267349e+00 3.54968131e-01
3.70382607e-01 -1.09887350e+00 -2.89626837e-01 7.98710406e-01
-1.02360773e+00 -6.22705758e-01 -1.02544224e+00 -6.98451519e-01
4.91964161e-01 1.79868996e-01 1.16903603e+00 1.85341179e-01
-4.74477917e-01 -1.34065922e-03 -1.71216473e-01 -4.79216456e-01
-8.34670737e-02 8.77056420e-01 -5.56253672e-01 -3.23976576e-01
2.14638282e-02 -5.27377725e-01 -4.77200896e-01 -4.20554042e-01
-1.03795886e+00 1.72127396e-01 5.99602580e-01 6.33282363e-01
2.77214766e-01 -8.08710307e-02 7.13071823e-02 -9.90102172e-01
1.06791162e+00 4.97802533e-02 -6.61516249e-01 4.40770447e-01
-5.28135777e-01 1.07109025e-01 9.28124070e-01 -3.63744915e-01
-9.47816670e-01 1.78806379e-01 1.04860729e-03 -4.59233448e-02
-2.68209815e-01 5.97461224e-01 -3.20124812e-02 1.92957267e-01
6.40204966e-01 6.88613206e-02 -4.82166529e-01 -6.54144585e-01
5.82942784e-01 4.94385481e-01 7.16077745e-01 -6.05039358e-01
7.97913790e-01 4.32792515e-01 1.21303787e-02 -9.23087239e-01
-7.64376104e-01 -2.11718425e-01 -8.63621175e-01 -2.13774145e-01
8.33348751e-01 -6.85915411e-01 -3.82681936e-01 6.29254460e-01
-1.54305601e+00 -4.06005412e-01 -3.37554812e-02 -2.07323104e-01
-4.80577677e-01 2.97046721e-01 -5.74845910e-01 -6.23232305e-01
-5.52122653e-01 -1.11988771e+00 1.21729279e+00 1.00704916e-01
-3.59785914e-01 -7.93354988e-01 1.77173063e-01 8.06597099e-02
5.34101665e-01 3.16760182e-01 1.00679421e+00 -5.59911668e-01
-6.29072011e-01 -3.44977766e-01 -3.61916006e-01 1.45922944e-01
-3.23308229e-01 5.73820829e-01 -1.26513851e+00 -1.47032082e-01
-2.68950015e-01 -1.04522360e-02 1.18963945e+00 4.31471288e-01
1.38569021e+00 -2.21667230e-01 -1.61533952e-01 8.64054143e-01
1.60912132e+00 1.64321214e-01 8.12785685e-01 9.64085937e-01
8.80681038e-01 6.47706211e-01 6.22289889e-02 3.04290444e-01
8.35603476e-02 5.56364000e-01 -8.24558586e-02 -2.05886707e-01
-8.31285715e-02 1.87901288e-01 1.51266873e-01 6.95821702e-01
-2.36052886e-01 -5.82186341e-01 -1.07677376e+00 7.43025899e-01
-1.86400735e+00 -8.14751625e-01 -3.79447758e-01 1.65408862e+00
7.15412140e-01 3.96181822e-01 4.72798385e-02 4.60117400e-01
3.39950472e-01 4.08924222e-01 -7.95273632e-02 -9.36540604e-01
-1.37354508e-01 5.27042568e-01 8.17267001e-01 1.35874629e-01
-1.31549954e+00 7.90770769e-01 6.96912432e+00 1.01326501e+00
-1.26268649e+00 -1.39624491e-01 5.14019012e-01 -2.53609955e-01
-1.67797804e-01 -1.65658325e-01 -7.16885924e-01 2.23041624e-01
9.38268185e-01 2.36580804e-01 2.02685878e-01 8.56739461e-01
-1.56527907e-01 -7.05435798e-02 -1.03287089e+00 8.42999518e-01
1.51847117e-02 -1.63747895e+00 1.78105906e-02 -2.89519448e-02
5.84319055e-01 -1.78841814e-01 1.29057452e-01 -2.56833266e-02
-7.62257054e-02 -9.83789384e-01 1.14031482e+00 4.70007718e-01
4.70381558e-01 -7.36815393e-01 3.58409524e-01 -2.66742762e-02
-1.08067596e+00 -2.35965610e-01 -6.63865358e-02 -3.78140993e-02
1.07851923e-01 5.45629382e-01 -3.77265513e-01 5.96657455e-01
9.44276512e-01 3.17497700e-01 -9.45442259e-01 9.98021066e-01
-1.93096476e-03 2.08323330e-01 -3.10823292e-01 -5.93611524e-02
4.94055271e-01 -1.83278508e-02 3.10128421e-01 1.85181797e+00
4.89747733e-01 -4.57273394e-01 -2.52770245e-01 5.75162530e-01
-1.18023440e-01 1.55526742e-01 -3.98814857e-01 -1.68942377e-01
7.45966956e-02 1.59392297e+00 -1.33679628e+00 -4.82388109e-01
-5.01720190e-01 1.09467185e+00 4.12512392e-01 2.24936947e-01
-6.58326983e-01 -8.05505335e-01 1.73966497e-01 -3.41247395e-03
6.96493804e-01 -7.05858469e-01 -9.93093789e-01 -1.12207699e+00
4.47458386e-01 -7.98019588e-01 3.16632509e-01 -8.07906270e-01
-9.94692206e-01 8.04156065e-01 -7.47517124e-02 -9.41450596e-01
-3.11150644e-02 -1.05814779e+00 -7.23721385e-01 7.23319530e-01
-1.42266989e+00 -1.24718463e+00 -3.16159338e-01 2.54251570e-01
7.75435328e-01 -7.62328580e-02 6.63347900e-01 3.69198650e-01
-7.43221462e-01 4.61343259e-01 1.19847767e-01 1.49219051e-01
7.92633772e-01 -1.59594226e+00 7.96265662e-01 1.11720967e+00
1.92540586e-01 7.16185272e-01 7.21825838e-01 -3.04470807e-01
-1.36183929e+00 -6.87054694e-01 1.14350390e+00 -4.44534928e-01
8.51591468e-01 -8.33915114e-01 -1.02627778e+00 4.00569946e-01
6.29850149e-01 -4.36246425e-01 2.80351549e-01 1.47043511e-01
-3.74938011e-01 -1.54627234e-01 -6.30077660e-01 7.50735104e-01
9.31940734e-01 -5.78955650e-01 -4.27009732e-01 2.23844498e-01
4.37008739e-01 -4.67842638e-01 -6.34818912e-01 -4.28682268e-02
6.72774076e-01 -7.83962727e-01 1.06721067e+00 -3.12755644e-01
8.14123213e-01 -2.31030256e-01 1.58353627e-01 -8.55540991e-01
-1.36079445e-01 -7.86833107e-01 -1.45834878e-01 1.33252311e+00
5.51112950e-01 -2.97554042e-02 5.52455723e-01 5.48085928e-01
-3.03562641e-01 -4.16719466e-01 -7.25212395e-01 -6.31840289e-01
4.65701334e-02 -5.81303716e-01 4.65778321e-01 7.93356836e-01
-3.48396361e-01 3.66927534e-01 -4.23450351e-01 -1.37358531e-01
1.56585008e-01 1.08426653e-01 7.39562988e-01 -1.18351686e+00
-2.39130646e-01 -9.45230186e-01 8.62351968e-04 -7.26595223e-01
-1.97470993e-01 -6.34526193e-01 1.31298378e-01 -1.61886501e+00
-6.10930771e-02 -1.04010828e-01 -1.20410070e-01 4.57142293e-01
1.03898793e-02 1.91540867e-01 4.28736210e-01 1.54270276e-01
-2.97104180e-01 -1.78348701e-02 9.62834299e-01 -3.00658107e-01
-2.04621688e-01 -4.60860580e-01 -6.04128659e-01 5.77908874e-01
7.35003889e-01 -4.06985551e-01 -2.60947287e-01 -9.01999950e-01
6.27421677e-01 -2.92781502e-01 3.37720782e-01 -1.14515114e+00
3.53470743e-01 1.10358402e-01 7.10452378e-01 -8.22739840e-01
1.28418639e-01 -6.68591976e-01 -1.38637006e-01 1.43694744e-01
-4.05115157e-01 5.09146929e-01 5.24091601e-01 2.88778037e-01
-9.33090746e-02 -5.70468187e-01 4.32550550e-01 -3.05783153e-01
-7.37036884e-01 -7.89038390e-02 -5.50221920e-01 -2.94836611e-01
6.46339715e-01 -3.90822560e-01 -5.19651771e-01 1.80935651e-01
-6.24761522e-01 -1.15218781e-01 4.00644600e-01 5.14480114e-01
1.66747272e-01 -8.25604856e-01 -4.41589892e-01 -1.91601768e-01
-1.49279773e-01 6.65711164e-02 9.48511809e-02 6.17019832e-01
-9.54573512e-01 6.79231465e-01 -6.12044275e-01 -2.85688043e-01
-1.26851797e+00 7.80716062e-01 7.14235902e-02 -5.92681408e-01
-6.85203075e-01 7.69625425e-01 -3.41734856e-01 6.58082888e-02
3.79601419e-01 -6.96141839e-01 -1.68453529e-01 5.80502331e-01
5.21349669e-01 3.68371964e-01 7.33076155e-01 -1.28433391e-01
-2.03182697e-01 6.89615846e-01 -2.94792950e-01 -4.49866265e-01
1.67089045e+00 1.54286623e-01 -9.36213732e-02 4.35803473e-01
1.29255438e+00 -1.38860151e-01 -1.27278757e+00 2.26381764e-01
5.25046647e-01 -1.64838478e-01 3.53763998e-01 -8.22766542e-01
-9.99294758e-01 1.00435269e+00 4.07139719e-01 4.89342332e-01
1.46465671e+00 -5.20046473e-01 6.39788747e-01 5.48009932e-01
-3.70828398e-02 -1.59474576e+00 -2.37622678e-01 5.37319362e-01
1.06966507e+00 -8.94863427e-01 4.29085404e-01 -1.72035635e-01
-3.12759250e-01 1.57166576e+00 2.48750582e-01 -3.11523199e-01
3.67254972e-01 6.61770940e-01 -8.31077918e-02 -2.35379413e-01
-7.64541328e-01 -1.74031422e-01 6.18664265e-01 4.06282693e-01
7.02932239e-01 -1.79480150e-01 -4.97892350e-01 4.20942485e-01
-3.41830134e-01 -5.83173521e-02 4.61453229e-01 1.42286623e+00
-2.33758241e-01 -1.35235095e+00 -3.00132513e-01 3.43485862e-01
-5.67731321e-01 -2.96050429e-01 -6.66138887e-01 9.88649786e-01
4.82928604e-02 3.38250190e-01 2.98241317e-01 -1.78185757e-02
3.18034977e-01 2.30350837e-01 7.60951757e-01 -2.92664886e-01
-1.28438497e+00 3.65510523e-01 2.44729042e-01 -3.47683638e-01
-4.15090650e-01 -6.06487989e-01 -7.66156614e-01 -2.24026158e-01
-4.22622589e-03 -5.52436650e-01 9.91745830e-01 7.99374759e-01
3.34959120e-01 7.94995546e-01 -2.01157164e-02 -1.09780121e+00
-2.80258119e-01 -8.42240453e-01 -1.28946394e-01 1.08613186e-01
4.38355297e-01 -1.53190404e-01 1.92442983e-02 5.56299686e-01] | [11.66201114654541, 2.6945865154266357] |
b480ff78-3a70-49de-bb12-fd1b75b88715 | can-large-language-models-reason-about | 2207.08143 | null | https://arxiv.org/abs/2207.08143v3 | https://arxiv.org/pdf/2207.08143v3.pdf | Can large language models reason about medical questions? | Although large language models (LLMs) often produce impressive outputs, it remains unclear how they perform in real-world scenarios requiring strong reasoning skills and expert domain knowledge. We set out to investigate whether GPT-3.5 (Codex and InstructGPT) can be applied to answer and reason about difficult real-world-based questions. We utilize two multiple-choice medical exam questions (USMLE and MedMCQA) and a medical reading comprehension dataset (PubMedQA). We investigate multiple prompting scenarios: Chain-of-Thought (CoT, think step-by-step), zero- and few-shot (prepending the question with question-answer exemplars) and retrieval augmentation (injecting Wikipedia passages into the prompt). For a subset of the USMLE questions, a medical expert reviewed and annotated the model's CoT. We found that InstructGPT can often read, reason and recall expert knowledge. Failure are primarily due to lack of knowledge and reasoning errors and trivial guessing heuristics are observed, e.g.\ too often predicting labels A and D on USMLE. Sampling and combining many completions overcome some of these limitations. Using 100 samples, Codex 5-shot CoT not only gives close to well-calibrated predictive probability but also achieves human-level performances on the three datasets. USMLE: 60.2%, MedMCQA: 62.7% and PubMedQA: 78.2%. | ['Ole Winther', 'Christoffer Egeberg Hother', 'Valentin Liévin'] | 2022-07-17 | null | null | null | null | ['multiple-choice-qa'] | ['natural-language-processing'] | [ 2.93454409e-01 7.95346677e-01 -9.43022594e-02 -2.37239510e-01
-1.46201515e+00 -6.16001725e-01 3.69392514e-01 7.59113550e-01
-5.27254522e-01 8.30426753e-01 3.68718147e-01 -9.28165197e-01
-5.98136961e-01 -5.61691046e-01 -7.69113123e-01 3.33749317e-02
4.55519140e-01 1.01148999e+00 2.39426538e-01 -3.01738709e-01
2.92146891e-01 -1.73985153e-01 -1.09805167e+00 7.87792861e-01
1.60533881e+00 5.10532200e-01 2.90624082e-01 1.14856863e+00
-4.12427366e-01 1.56821287e+00 -6.82765126e-01 -9.74422753e-01
-2.24905148e-01 -4.37567085e-01 -1.41525590e+00 -2.88526297e-01
7.93857872e-01 -1.85429707e-01 -1.87029749e-01 8.45763981e-01
5.58445454e-01 2.29809389e-01 6.33892834e-01 -7.53342390e-01
-9.07080889e-01 6.89983130e-01 -1.41958922e-01 5.65581560e-01
9.92085159e-01 4.90321726e-01 7.93510914e-01 -5.76407254e-01
8.07318151e-01 1.29217505e+00 8.39384079e-01 6.63955748e-01
-1.23144197e+00 -2.29006678e-01 -2.59146035e-01 2.60022432e-01
-9.40560579e-01 -2.29012027e-01 -1.42347991e-01 -4.52499300e-01
1.23131609e+00 4.27294940e-01 2.46539727e-01 1.26993811e+00
5.15247822e-01 8.57661664e-01 1.36060047e+00 -6.46737278e-01
2.60132819e-01 3.70020598e-01 7.74018168e-01 9.22911763e-01
1.08246699e-01 -2.10242048e-01 -3.99275661e-01 -4.58348751e-01
3.18009079e-01 -2.18703151e-01 -4.96566117e-01 5.26379824e-01
-1.22170031e+00 7.69476712e-01 2.85784930e-01 1.10429414e-01
-5.31266809e-01 -2.14428231e-01 8.50485116e-02 5.65708995e-01
2.61178743e-02 1.28045285e+00 -7.43739784e-01 -6.51820242e-01
-8.43640447e-01 4.65205401e-01 1.31031287e+00 1.03853047e+00
2.20206931e-01 -6.49924457e-01 -6.75800800e-01 9.20234025e-01
-1.81878746e-01 6.12728357e-01 6.28206849e-01 -1.08739519e+00
6.96309090e-01 5.19143403e-01 1.19622678e-01 -8.09895158e-01
-6.23809636e-01 -3.63667309e-01 -3.72745007e-01 -3.67962778e-01
8.58049631e-01 -2.55460054e-01 -1.03379893e+00 1.55613875e+00
-1.01141015e-03 -1.65926203e-01 2.88539648e-01 6.53539121e-01
1.32622361e+00 5.78016400e-01 6.57963634e-01 -3.08776181e-02
1.95597374e+00 -1.12860024e+00 -9.33419704e-01 -6.04006350e-01
9.07555997e-01 -8.62595499e-01 1.47367418e+00 5.10564804e-01
-1.40750825e+00 -4.67590034e-01 -5.18163502e-01 -3.79546404e-01
-2.35872775e-01 -8.16237703e-02 3.59286994e-01 3.23677748e-01
-9.89533901e-01 4.10304993e-01 -4.89060998e-01 -4.61572915e-01
2.38735318e-01 -1.18575796e-01 -1.82962120e-01 -8.43827367e-01
-1.47324157e+00 1.34785807e+00 1.56040952e-01 -5.52046061e-01
-8.40823114e-01 -1.10357630e+00 -7.11283088e-01 2.05698416e-01
6.94754839e-01 -9.95083988e-01 1.87759888e+00 -6.03269458e-01
-1.13037229e+00 8.23235214e-01 -9.75863039e-02 -4.67618585e-01
5.86623609e-01 -4.72517222e-01 -4.57992196e-01 5.36300421e-01
3.30224842e-01 6.43224359e-01 2.91153818e-01 -6.90643609e-01
-3.09871614e-01 1.70748383e-02 2.87196547e-01 2.26812184e-01
8.93124416e-02 -1.79107249e-01 -3.60180080e-01 -3.78128737e-01
-3.28512415e-02 -6.72775745e-01 -2.54203647e-01 -1.63884565e-01
-5.68427563e-01 -3.52465659e-01 -2.40895256e-01 -1.06084347e+00
1.38064933e+00 -1.66678476e+00 -1.76904961e-01 -7.29985610e-02
4.46113408e-01 4.56466407e-01 -3.56346607e-01 6.62957668e-01
-1.17207300e-02 2.09076807e-01 -2.39287511e-01 2.72496760e-01
-1.43038496e-01 1.47590339e-01 -1.92906782e-01 -2.68062651e-01
1.65312469e-01 1.09519720e+00 -1.23165190e+00 -7.90820897e-01
-1.88243166e-02 1.91203430e-02 -7.14787841e-01 3.04559857e-01
-7.38361239e-01 1.23030625e-01 -3.81849736e-01 5.94333827e-01
1.81966990e-01 -8.45574081e-01 -1.30901918e-01 1.20740756e-01
4.39459383e-01 4.72053349e-01 -7.38546550e-01 1.75727880e+00
-4.77403522e-01 3.30684870e-01 -2.78624475e-01 -2.47030646e-01
4.26562488e-01 4.24816310e-01 -2.35402524e-01 -9.40495849e-01
-6.56178370e-02 3.11522871e-01 2.46912554e-01 -1.40270567e+00
3.78719538e-01 -2.77583122e-01 2.63034943e-02 4.81387258e-01
2.97486126e-01 -1.76904306e-01 2.91935056e-01 7.61067331e-01
1.77091467e+00 -3.77821207e-01 4.35525239e-01 -3.38557631e-01
3.30599487e-01 6.01512790e-01 9.70898122e-02 1.42025280e+00
-5.42801954e-02 4.70950335e-01 6.17028415e-01 -1.51839495e-01
-6.17785692e-01 -1.06892467e+00 -9.71182510e-02 1.15488112e+00
-1.94865867e-01 -5.24509907e-01 -8.08572054e-01 -8.43957603e-01
-9.33853909e-03 1.68871057e+00 -5.48012793e-01 -3.23631555e-01
-1.87784702e-01 -3.40079963e-01 7.68680155e-01 4.17412072e-01
3.27652067e-01 -1.09123957e+00 -7.01121628e-01 4.15547341e-01
-6.24156117e-01 -1.16039705e+00 -3.74033302e-01 1.35539636e-01
-7.21742392e-01 -1.47378051e+00 -6.90820873e-01 -5.35206318e-01
7.25291729e-01 -1.37128547e-01 1.81689048e+00 3.71552467e-01
-3.42714369e-01 9.14117754e-01 -5.07831991e-01 -3.32396984e-01
-7.80474544e-01 1.34397224e-01 -3.93095702e-01 -8.69145751e-01
6.82737052e-01 5.26299700e-03 -5.96477449e-01 5.02091981e-02
-9.18335974e-01 2.13987261e-01 1.03112078e+00 9.36986864e-01
2.44084850e-01 -4.45945263e-01 5.77552795e-01 -1.45213163e+00
1.18858206e+00 -7.75065899e-01 -7.04645962e-02 7.54361987e-01
-7.99567521e-01 1.20758429e-01 6.38939857e-01 -6.14836216e-01
-1.15600634e+00 -6.13414943e-01 -3.64834607e-01 -2.55336583e-01
-3.40535611e-01 8.26983213e-01 4.18039799e-01 2.32350901e-01
1.40057099e+00 5.18487729e-02 -7.32522160e-02 -3.75769734e-01
3.18383783e-01 5.70756853e-01 5.43962598e-01 -9.44265068e-01
4.33220178e-01 -2.84218699e-01 -5.32417595e-01 -4.95329589e-01
-1.43523610e+00 -2.99143791e-01 3.59726436e-02 -5.50232008e-02
1.06444168e+00 -7.33744323e-01 -8.25104713e-01 -1.50016829e-01
-1.05961120e+00 -6.68684423e-01 -4.37902272e-01 4.71222013e-01
-3.95550519e-01 1.20558538e-01 -1.10095191e+00 -7.40109146e-01
-4.66484725e-01 -1.03074694e+00 7.34105289e-01 5.01667500e-01
-7.93158650e-01 -1.03816378e+00 -3.35202063e-03 8.19186032e-01
4.77039903e-01 -1.37899488e-01 1.47701228e+00 -1.32222390e+00
-3.40386122e-01 -1.76624388e-01 -1.43265009e-01 1.87511090e-02
-3.63894135e-01 -5.46894073e-01 -8.26986074e-01 4.09736112e-02
1.31321430e-01 -9.11530554e-01 5.34844279e-01 2.11932808e-01
1.19707799e+00 -4.58941102e-01 -2.68615156e-01 2.17788219e-02
1.35413802e+00 1.13322943e-01 5.56318998e-01 4.65678759e-02
3.02997768e-01 6.98784709e-01 5.99119425e-01 2.05570906e-02
5.92248976e-01 1.53571337e-01 -2.61018425e-01 1.91887408e-01
-9.43823531e-02 -4.03095782e-01 -1.66412257e-02 9.49175775e-01
4.13754791e-01 -3.51723939e-01 -1.60247838e+00 5.75091779e-01
-1.47170949e+00 -6.52277648e-01 -3.00730735e-01 1.79562962e+00
1.27835155e+00 3.68278831e-01 -4.68506694e-01 -3.09366167e-01
1.06734768e-01 -3.67381483e-01 -5.62099218e-01 -4.04111445e-01
1.74032301e-01 4.80164409e-01 2.01658800e-01 7.71226645e-01
-4.11456704e-01 7.07759261e-01 6.14188719e+00 9.31569517e-01
-3.91480654e-01 2.90854663e-01 8.33104074e-01 1.84390366e-01
-5.33015013e-01 -1.64326727e-01 -6.66054785e-01 3.22618723e-01
1.42503786e+00 -4.27643694e-02 3.83945823e-01 5.62644243e-01
-2.45406106e-01 -5.58803439e-01 -1.26645684e+00 7.37250209e-01
3.07811558e-01 -1.23171210e+00 2.48749144e-02 -5.11577964e-01
6.97469234e-01 -2.76222706e-01 -1.34608030e-01 9.70645785e-01
6.71936333e-01 -1.25695539e+00 2.92768836e-01 9.80660617e-01
7.17175841e-01 -3.46698105e-01 9.21482503e-01 8.49978983e-01
-2.58214772e-01 -8.97759721e-02 -3.36994678e-01 -3.69622698e-03
1.57208648e-02 5.30387402e-01 -1.22429848e+00 4.27699655e-01
5.88833511e-01 3.66815776e-02 -1.03372407e+00 1.03934181e+00
-4.32741374e-01 9.19635177e-01 -1.92043304e-01 -4.19873804e-01
3.07555169e-01 3.06930602e-01 3.83800983e-01 1.25502610e+00
3.79524976e-02 8.39533567e-01 1.83320448e-01 9.57415640e-01
-6.48215637e-02 1.01123847e-01 -1.53432846e-01 -2.42958531e-01
5.91210783e-01 1.05681980e+00 -5.78655899e-01 -7.14835167e-01
-4.27093476e-01 7.80846000e-01 4.84703302e-01 3.98421228e-01
-6.89765394e-01 -4.20093924e-01 1.01930924e-01 1.36204332e-01
-1.96470886e-01 5.60433030e-01 -2.27096766e-01 -1.07210839e+00
-2.02982545e-01 -1.62819088e+00 7.23034978e-01 -1.24736488e+00
-1.67501044e+00 6.00527167e-01 1.11001477e-01 -7.66894817e-01
-2.64247626e-01 -7.26351738e-01 -3.57498556e-01 9.13975239e-01
-1.32784164e+00 -4.44533527e-01 -5.25903881e-01 5.32746851e-01
7.81416535e-01 4.10918295e-02 1.01790977e+00 2.74693251e-01
-2.30033800e-01 5.17850697e-01 -3.13744962e-01 5.20607010e-02
9.34929132e-01 -1.57491004e+00 2.22423106e-01 3.53545159e-01
-6.01867139e-02 1.05813110e+00 9.62637842e-01 -8.36797476e-01
-1.31034327e+00 -7.79582262e-01 1.17168188e+00 -1.11863983e+00
4.70841229e-01 1.57730922e-01 -1.42819476e+00 7.09298193e-01
3.74181867e-01 -4.30026352e-01 9.90193188e-01 1.51984811e-01
-2.57000208e-01 3.18104863e-01 -1.36128676e+00 7.99811244e-01
6.61790371e-01 -6.29865646e-01 -1.42856681e+00 9.83650386e-01
9.58006799e-01 -8.34370792e-01 -1.01549590e+00 1.07446775e-01
1.36012331e-01 -6.93316162e-01 7.95643032e-01 -1.11653638e+00
9.31311846e-01 1.37454376e-01 1.08249806e-01 -1.26259029e+00
-1.37869760e-01 -3.54246080e-01 -9.94899869e-02 8.12073469e-01
7.81518519e-01 -4.48502243e-01 3.91137570e-01 1.16324294e+00
-2.78109878e-01 -1.01503372e+00 -5.60632348e-01 -3.26223820e-01
3.61024171e-01 -4.35205132e-01 2.70356059e-01 1.09231639e+00
3.55764210e-01 5.65004170e-01 2.12373257e-01 8.97063594e-03
2.91986316e-01 -2.06508726e-01 3.79323065e-01 -1.15974236e+00
-5.43409526e-01 -1.27224803e-01 3.62509817e-01 -1.06636417e+00
-1.73039183e-01 -9.17309701e-01 1.69812873e-01 -1.91433585e+00
3.64083052e-01 -1.70467868e-01 4.13490981e-02 5.95503390e-01
-8.00745130e-01 -4.83358711e-01 7.53202364e-02 -9.76484269e-02
-9.01384056e-01 -1.52746886e-02 1.41763663e+00 7.69933453e-04
7.69315884e-02 -1.91182733e-01 -8.51393700e-01 6.28697872e-01
4.65502083e-01 -5.67727149e-01 -6.09560370e-01 -3.69841129e-01
5.36039233e-01 8.96982133e-01 3.26452553e-01 -1.00249839e+00
6.33024037e-01 4.09665853e-02 4.63623255e-01 -2.98349231e-01
8.71817842e-02 -5.24010360e-01 -2.05677509e-01 6.87686086e-01
-1.02138162e+00 4.19972658e-01 4.08255905e-01 5.27104676e-01
-1.01872571e-02 -6.93102956e-01 4.64858711e-01 -6.84061766e-01
-4.51013982e-01 -2.46659949e-01 -5.47626853e-01 9.54481542e-01
6.59413338e-01 2.26327509e-01 -8.55288267e-01 -6.30669057e-01
-8.54021966e-01 7.04730093e-01 2.79803779e-02 2.68093944e-01
6.82638407e-01 -6.65980816e-01 -7.23937809e-01 -1.19315103e-01
3.05627376e-01 -4.79766428e-02 5.69951713e-01 1.08065617e+00
-7.60027587e-01 6.98006809e-01 1.00169569e-01 -4.92163509e-01
-9.40851569e-01 5.82866967e-01 3.75135422e-01 -6.36861265e-01
-5.40043890e-01 1.18217373e+00 -2.32007325e-01 -7.61462450e-01
4.06259567e-01 -5.70492089e-01 -1.28809422e-01 -4.10914198e-02
7.15261638e-01 3.06881189e-01 5.81261143e-02 2.56750762e-01
1.84032843e-02 1.40190437e-01 -4.53462452e-01 -1.90792248e-01
8.38447392e-01 1.00920781e-01 2.00665399e-01 4.51431632e-01
7.06168056e-01 -2.69986745e-02 -5.37450790e-01 -2.49833599e-01
2.86890060e-01 -1.33348882e-01 -3.16871822e-01 -1.85038435e+00
-1.92596823e-01 8.57175291e-01 2.74511099e-01 1.40459523e-01
7.34634042e-01 1.64410248e-01 8.02866817e-01 9.97237623e-01
2.12475240e-01 -8.59252632e-01 3.09644908e-01 5.16313434e-01
7.38850713e-01 -1.40388370e+00 -6.55276403e-02 -2.74422139e-01
-9.79854465e-01 8.97628307e-01 1.01910746e+00 3.13639909e-01
2.08787531e-01 -1.12253495e-01 3.50080580e-01 -6.33549988e-01
-1.23516703e+00 1.87110856e-01 4.67671007e-01 1.83390215e-01
6.12834573e-01 -9.16074663e-02 -3.56986970e-01 8.34927320e-01
-4.46576774e-01 1.36334494e-01 5.83726764e-01 9.09952700e-01
-4.01006937e-01 -5.42330503e-01 -3.56160849e-01 1.10253739e+00
-5.73506534e-01 -5.01337349e-01 -2.39988983e-01 7.67184317e-01
-2.06965759e-01 1.14595008e+00 -2.44817510e-01 -6.45610765e-02
5.01398265e-01 6.00077212e-01 3.24175298e-01 -9.43279743e-01
-1.12964654e+00 -4.24576819e-01 4.67238903e-01 -5.28756142e-01
1.57246262e-01 -3.28619391e-01 -1.18205750e+00 -1.82439551e-01
-3.22269887e-01 5.08109331e-01 7.16649145e-02 1.02969098e+00
3.56749356e-01 8.27580094e-01 -4.31532383e-01 3.77360910e-01
-1.02449775e+00 -1.30368364e+00 -9.07458887e-02 6.33376181e-01
3.53151441e-01 -3.11667144e-01 -4.13510114e-01 -9.04910266e-02] | [8.985782623291016, 8.410384178161621] |
df7f6327-e13f-4768-8f14-71624b1e8900 | deep-occlusion-aware-instance-segmentation | 2103.12340 | null | https://arxiv.org/abs/2103.12340v1 | https://arxiv.org/pdf/2103.12340v1.pdf | Deep Occlusion-Aware Instance Segmentation with Overlapping BiLayers | Segmenting highly-overlapping objects is challenging, because typically no distinction is made between real object contours and occlusion boundaries. Unlike previous two-stage instance segmentation methods, we model image formation as composition of two overlapping layers, and propose Bilayer Convolutional Network (BCNet), where the top GCN layer detects the occluding objects (occluder) and the bottom GCN layer infers partially occluded instance (occludee). The explicit modeling of occlusion relationship with bilayer structure naturally decouples the boundaries of both the occluding and occluded instances, and considers the interaction between them during mask regression. We validate the efficacy of bilayer decoupling on both one-stage and two-stage object detectors with different backbones and network layer choices. Despite its simplicity, extensive experiments on COCO and KINS show that our occlusion-aware BCNet achieves large and consistent performance gain especially for heavy occlusion cases. Code is available at https://github.com/lkeab/BCNet. | ['Chi-Keung Tang', 'Yu-Wing Tai', 'Lei Ke'] | 2021-03-23 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Ke_Deep_Occlusion-Aware_Instance_Segmentation_With_Overlapping_BiLayers_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Ke_Deep_Occlusion-Aware_Instance_Segmentation_With_Overlapping_BiLayers_CVPR_2021_paper.pdf | cvpr-2021-1 | ['amodal-instance-segmentation', 'real-time-instance-segmentation', 'occlusion-handling'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 1.91477656e-01 4.06382054e-01 -2.91299939e-01 -3.24407697e-01
-4.86442715e-01 -6.27146602e-01 3.44606996e-01 -1.51541933e-01
-3.00645232e-01 3.51481825e-01 -2.17552036e-01 -2.42955089e-01
4.07355666e-01 -6.40779257e-01 -6.88630044e-01 -6.15349889e-01
8.13800693e-02 3.67584676e-01 9.21924889e-01 2.26788223e-01
-1.75546296e-02 4.94485080e-01 -1.29400396e+00 4.97478664e-01
9.31839883e-01 1.04896712e+00 7.95596838e-02 7.49891996e-01
-2.97927409e-01 4.76854324e-01 -5.16977966e-01 -4.12376255e-01
6.11389339e-01 2.88598426e-02 -7.18903661e-01 3.00119787e-01
9.13724065e-01 -3.64498585e-01 -2.52364129e-01 9.91102517e-01
2.52057463e-01 -1.66864321e-01 5.17553806e-01 -1.12324727e+00
-4.92794126e-01 5.15166759e-01 -1.03258562e+00 2.07039684e-01
-3.71879965e-01 3.44852597e-01 8.47355664e-01 -8.81334305e-01
6.04890585e-01 1.20075428e+00 6.48392558e-01 5.07127225e-01
-1.57504332e+00 -7.11904526e-01 8.05378556e-01 -2.27957800e-01
-1.33726668e+00 -4.72339481e-01 7.50415504e-01 -6.49885535e-01
5.54245710e-01 1.31603792e-01 6.82176232e-01 7.64123976e-01
1.06319750e-03 1.15209544e+00 9.29218352e-01 -1.19132236e-01
2.40019131e-02 5.47263809e-02 5.16206026e-01 6.42580211e-01
4.89526033e-01 -5.63083217e-02 6.59678085e-03 -3.86206210e-02
1.00941098e+00 -8.73020291e-02 -2.71697462e-01 -5.36061347e-01
-7.37468898e-01 5.65842450e-01 4.81325388e-01 8.86775479e-02
-9.27681699e-02 2.32271105e-01 5.30335419e-02 -1.26848176e-01
5.39618850e-01 9.32748765e-02 -5.19956052e-01 4.80956197e-01
-1.14125991e+00 2.25890651e-01 7.27985382e-01 1.17315388e+00
9.34628308e-01 -1.09428160e-01 -2.73345172e-01 8.64016473e-01
2.50017554e-01 7.81522319e-02 -1.15501890e-02 -1.10792553e+00
2.83216119e-01 7.77649999e-01 1.06720552e-02 -6.93847179e-01
-3.71251941e-01 -5.31063080e-01 -5.35469174e-01 4.50048596e-01
7.52364039e-01 -3.51489097e-01 -1.49862003e+00 1.55955029e+00
6.79189861e-01 4.07898515e-01 -3.87869656e-01 1.04317641e+00
1.09053302e+00 2.65987635e-01 3.08415830e-01 1.27551764e-01
1.57773066e+00 -1.34940732e+00 -4.17200625e-01 -5.59803009e-01
3.16753149e-01 -7.54720628e-01 8.87014687e-01 1.29158497e-01
-1.30336332e+00 -6.59060240e-01 -1.04142642e+00 -5.93864918e-01
-2.25400388e-01 2.77487397e-01 8.06501150e-01 2.81932533e-01
-1.04128861e+00 4.05303776e-01 -9.49285865e-01 1.43780559e-01
9.48401868e-01 5.43116510e-01 -7.29046017e-02 3.53335729e-03
-7.24202871e-01 3.42853308e-01 2.21913561e-01 3.98316801e-01
-9.09161687e-01 -8.62063825e-01 -8.50184798e-01 5.46486564e-02
5.05551577e-01 -6.06615841e-01 1.23905110e+00 -1.04380178e+00
-1.19904816e+00 1.18538225e+00 -3.18217963e-01 -3.78023237e-01
9.34233427e-01 -2.92527169e-01 1.24705069e-01 1.00256957e-01
6.94641620e-02 1.21780360e+00 8.30615819e-01 -1.57284260e+00
-7.53438234e-01 -3.61711532e-01 1.91901639e-01 1.99743658e-01
2.64695257e-01 -1.10642090e-01 -1.22681713e+00 -5.41975260e-01
4.88957852e-01 -7.40189672e-01 -2.92825311e-01 5.05389810e-01
-8.61929297e-01 -1.62742496e-01 1.03240824e+00 -5.35931468e-01
1.12299514e+00 -2.15478468e+00 1.23993936e-03 -1.82859097e-02
5.73472142e-01 3.28855425e-01 -2.17294410e-01 -1.52532935e-01
2.88165957e-02 2.11320490e-01 -5.65356910e-01 -7.81983495e-01
-2.97848314e-01 1.94292560e-01 -2.34943330e-01 4.54164475e-01
5.16606569e-01 1.02096069e+00 -5.43167055e-01 -6.74859166e-01
2.65447438e-01 5.70422292e-01 -4.65956390e-01 9.17815939e-02
-4.75153953e-01 4.98531342e-01 -5.23366451e-01 8.20137620e-01
1.08985090e+00 -4.46941763e-01 3.05812538e-01 -3.45159233e-01
-2.66167790e-01 2.68717736e-01 -1.23374224e+00 1.51777124e+00
6.82916418e-02 7.57969260e-01 6.51795328e-01 -5.12955964e-01
5.16705871e-01 1.32989109e-01 3.11405480e-01 -5.01677155e-01
2.62286097e-01 -5.61262928e-02 8.03794414e-02 -3.77715915e-01
2.60078996e-01 1.15910180e-01 2.94199169e-01 1.36713862e-01
-8.52921307e-02 -1.09546289e-01 3.13768804e-01 2.44286656e-01
7.84331381e-01 4.26330030e-01 1.52837262e-02 -4.14442629e-01
3.00127417e-01 -7.60371238e-02 1.08654583e+00 7.33996511e-01
-5.36915123e-01 1.11137271e+00 7.98577070e-01 -5.69489717e-01
-7.68083334e-01 -9.37400520e-01 -5.48198640e-01 9.01845872e-01
7.41301715e-01 -1.79511070e-01 -1.04307401e+00 -6.53671801e-01
2.73619950e-01 3.50178450e-01 -6.94713056e-01 3.39815736e-01
-9.43319857e-01 -8.61409009e-01 4.51305926e-01 5.19642234e-01
5.28634727e-01 -1.10781956e+00 -5.44381201e-01 1.45165026e-01
-1.54894024e-01 -1.49034297e+00 -5.90099037e-01 1.24212377e-01
-7.93239057e-01 -1.38286793e+00 -5.43383658e-01 -9.53981936e-01
7.07389534e-01 2.86117375e-01 1.23213565e+00 5.83346784e-01
-5.03146052e-01 -1.47619203e-01 1.82299212e-01 -3.76597673e-01
-5.64876944e-02 1.83649480e-01 -5.99696577e-01 2.48033088e-02
1.33226007e-01 -6.18887782e-01 -9.87639248e-01 4.24466014e-01
-7.56148815e-01 5.30658782e-01 3.84600222e-01 2.86118388e-01
7.78624713e-01 -3.26726764e-01 1.82533115e-01 -1.24110246e+00
9.39106345e-02 -1.86545119e-01 -9.22415376e-01 2.28096500e-01
-2.08621994e-01 -2.96081990e-01 8.91828984e-02 -6.36700928e-01
-1.03369379e+00 2.07717970e-01 -6.87456727e-02 -4.66632724e-01
-5.45618355e-01 -3.25933546e-01 -3.88177961e-01 -6.16298877e-02
3.16024035e-01 -2.73555934e-01 -2.63760328e-01 -6.60551071e-01
3.64591777e-01 2.46896759e-01 6.11804843e-01 -7.41723955e-01
6.36625350e-01 8.63095224e-01 -4.68008459e-01 -6.60369277e-01
-8.72503996e-01 -3.30349535e-01 -9.15918887e-01 -1.66939244e-01
1.14471781e+00 -8.92352462e-01 -6.32798970e-01 6.53406382e-01
-1.36596847e+00 -8.95847917e-01 -3.93105328e-01 -3.01786000e-03
-1.67445242e-01 1.43450350e-01 -9.67279077e-01 -7.16635287e-01
-1.49537608e-01 -1.44522893e+00 1.28311706e+00 5.44086158e-01
5.23174144e-02 -7.53046274e-01 -2.71139771e-01 4.87776399e-01
3.89936417e-02 4.38038111e-01 7.67796576e-01 -3.02982628e-01
-1.10196614e+00 6.68381527e-02 -6.09472752e-01 1.23794325e-01
9.57941040e-02 4.18041825e-01 -1.27447248e+00 -1.80114448e-01
-2.26982027e-01 -1.71635017e-01 1.16880655e+00 8.91005874e-01
1.24947035e+00 8.96282494e-02 -6.04239643e-01 9.77677584e-01
1.36942768e+00 1.62322477e-01 7.80812383e-01 3.20911780e-02
9.98768449e-01 6.44234598e-01 1.26848236e-01 -1.77027211e-01
3.79256666e-01 6.15282774e-01 3.89724702e-01 -7.30581999e-01
-6.85749233e-01 -2.20342726e-02 -5.87606765e-02 2.98126668e-01
1.71768680e-01 -3.74426425e-01 -7.46055126e-01 5.37311614e-01
-1.81241274e+00 -4.23665643e-01 -3.11785012e-01 1.87368655e+00
9.76095915e-01 4.62878376e-01 1.39229044e-01 -2.97534078e-01
9.31037724e-01 3.33635449e-01 -8.96742702e-01 8.65661353e-03
-2.08705232e-01 1.78800933e-02 4.26062346e-01 8.75306487e-01
-1.34314275e+00 1.32235026e+00 6.17969608e+00 8.71721983e-01
-1.07003510e+00 1.48826078e-01 1.15628064e+00 -1.92150190e-01
-1.14510633e-01 2.23763138e-01 -1.02807736e+00 4.10905123e-01
3.69681120e-02 5.69547832e-01 -3.00574768e-03 6.55895233e-01
6.08158223e-02 -1.79627940e-01 -9.81439948e-01 4.28387612e-01
-4.61871415e-01 -1.33511472e+00 -1.17674276e-01 4.67662439e-02
8.35164487e-01 2.54503846e-01 -4.14383225e-02 -9.26157925e-03
4.03956681e-01 -8.24507892e-01 1.01275527e+00 1.68064162e-01
7.66008914e-01 -2.64137149e-01 2.30672821e-01 1.33894637e-01
-1.48315001e+00 1.44668192e-01 -2.56330043e-01 3.74879502e-02
2.65971422e-01 7.03618944e-01 -3.34090352e-01 2.16064379e-01
6.68015897e-01 5.44763505e-01 -4.22143906e-01 1.03835332e+00
-4.70208615e-01 5.44783950e-01 -4.09092098e-01 8.15254390e-01
1.75294876e-01 -4.34825599e-01 6.03457272e-01 1.26354587e+00
-5.23008704e-01 2.82506585e-01 3.93042207e-01 1.27826834e+00
-2.77296931e-01 -1.41597793e-01 -1.82400003e-01 4.75806743e-01
4.53871965e-01 1.47244334e+00 -1.29589677e+00 -4.73335445e-01
-5.19583225e-01 7.15304136e-01 6.52343452e-01 7.41403818e-01
-1.03224528e+00 5.37969843e-02 9.33018327e-01 4.23439890e-01
5.90827107e-01 -1.58758029e-01 -6.59989595e-01 -1.14132309e+00
9.47687402e-02 -6.13486409e-01 3.07080179e-01 -4.36987221e-01
-1.05816305e+00 5.49676895e-01 -1.38654402e-02 -7.70427704e-01
5.68552494e-01 -6.12285137e-01 -8.47486734e-01 8.18397701e-01
-1.69843793e+00 -1.22154641e+00 -4.04621810e-01 1.50069162e-01
7.32022047e-01 6.45616293e-01 1.79644316e-01 4.25489128e-01
-1.06215048e+00 4.69040692e-01 -4.53931779e-01 4.59190905e-01
3.64724874e-01 -1.21895325e+00 6.45205557e-01 9.54811335e-01
3.66529003e-02 5.29793024e-01 3.46695244e-01 -8.84992957e-01
-6.35652602e-01 -1.18596470e+00 4.61367607e-01 -4.36624676e-01
2.85588264e-01 -7.36478329e-01 -1.23737752e+00 7.35670924e-01
6.76894635e-02 3.80976915e-01 1.67129397e-01 -1.10930964e-01
-4.52707976e-01 1.34785011e-01 -9.63761568e-01 8.54634643e-01
1.21033275e+00 -2.24423021e-01 -1.65325359e-01 2.95608699e-01
1.01148570e+00 -7.04319596e-01 -5.74511290e-01 6.54902399e-01
6.47435665e-01 -1.15625608e+00 1.07401681e+00 -5.63799500e-01
3.09724331e-01 -5.24944484e-01 1.77548543e-01 -4.26968694e-01
-1.61088750e-01 -6.01181328e-01 -2.00329363e-01 1.10823786e+00
5.61390817e-01 -5.33119917e-01 1.12048888e+00 9.35937643e-01
-2.22977832e-01 -1.10048127e+00 -6.27717912e-01 -3.87506336e-01
2.75109738e-01 -3.24706882e-01 4.97189641e-01 7.67288744e-01
-6.98022604e-01 1.11447871e-01 -2.28874628e-02 3.76753807e-01
6.60273612e-01 3.52138817e-01 7.49712348e-01 -1.23192561e+00
-3.34138691e-01 -6.43521667e-01 2.89893616e-03 -1.43594527e+00
-1.82208493e-02 -7.14134812e-01 7.79213160e-02 -1.47673094e+00
2.23511651e-01 -8.62588525e-01 -1.13946699e-01 6.55952394e-01
-4.72408384e-01 5.66645443e-01 2.56429583e-01 3.19778681e-01
-4.45959240e-01 1.80013880e-01 1.54784155e+00 -1.68111756e-01
-5.13532996e-01 -2.51802262e-02 -7.17646360e-01 1.14904010e+00
7.08559096e-01 -5.01880407e-01 -4.19190705e-01 -7.17921138e-01
-4.34528649e-01 -1.78437382e-01 5.20497561e-01 -7.52010942e-01
1.26851633e-01 -6.67189658e-02 4.99726206e-01 -5.59324145e-01
2.83046991e-01 -5.38608134e-01 1.67087704e-01 2.26019934e-01
-2.23446772e-01 -2.86594003e-01 4.83065873e-01 4.16085929e-01
5.68269640e-02 -1.05080888e-01 9.71842885e-01 -2.53387362e-01
-5.17422140e-01 6.79043412e-01 -2.11393293e-02 2.73939580e-01
9.90365446e-01 -4.93392706e-01 -3.80818546e-01 1.34254619e-01
-9.05702651e-01 4.61500555e-01 6.30710840e-01 3.31068397e-01
3.27597171e-01 -7.35789776e-01 -4.79587376e-01 3.75547528e-01
-3.44343334e-01 8.50737453e-01 3.01376641e-01 8.77451718e-01
-6.44021094e-01 -5.37491869e-03 2.93299794e-01 -7.60014474e-01
-1.21433413e+00 2.48174205e-01 7.58393347e-01 -1.99655458e-01
-9.57066298e-01 1.16163349e+00 1.15723646e+00 -2.39164129e-01
5.00061333e-01 -4.87891704e-01 -2.17322260e-02 -9.24053863e-02
3.14576834e-01 1.24128617e-01 -1.98696479e-01 -5.48841774e-01
-2.77069330e-01 6.68926001e-01 -5.21141887e-01 2.76955515e-02
9.91011977e-01 -7.26559386e-03 -1.11214772e-01 1.40538037e-01
9.55888152e-01 -1.52467966e-01 -1.95370579e+00 -9.98832732e-02
-1.07243985e-01 -2.44199812e-01 -1.49102425e-02 -5.06508529e-01
-1.47791255e+00 9.23411787e-01 3.90181959e-01 6.17024153e-02
8.62404764e-01 6.78383186e-02 8.33163440e-01 -2.18197227e-01
8.15501288e-02 -8.74205768e-01 -1.15119003e-01 2.56990671e-01
6.53680742e-01 -1.11041331e+00 5.15583120e-02 -1.09345257e+00
-3.92020375e-01 7.37933874e-01 1.16893888e+00 -3.40142280e-01
7.36705720e-01 6.31856918e-01 3.69182825e-01 -4.17046040e-01
-6.19627237e-01 -2.56260812e-01 4.60502923e-01 3.99521828e-01
3.44979554e-01 5.17492816e-02 -1.88859671e-01 4.13652956e-01
1.04974940e-01 -3.50545108e-01 2.43997857e-01 8.89140785e-01
-3.31598848e-01 -1.08322287e+00 -1.83466911e-01 3.97078097e-01
-5.20887971e-01 -1.40932113e-01 -4.29232568e-01 1.17106473e+00
6.87071860e-01 6.96809769e-01 4.98089910e-01 9.55256447e-02
1.72079280e-01 -4.74065095e-02 3.63783956e-01 -7.89049208e-01
-7.70588517e-01 5.93070030e-01 -1.11931905e-01 -7.03235030e-01
-3.42382580e-01 -5.46858430e-01 -1.49805987e+00 3.68548627e-03
-6.11502230e-01 -4.57181931e-02 4.18552995e-01 7.67412424e-01
5.15137196e-01 6.78725421e-01 5.82328327e-02 -1.13566411e+00
1.13516159e-01 -7.15659916e-01 -4.34398681e-01 3.76015246e-01
4.82842416e-01 -6.81610882e-01 -2.81308383e-01 1.75916493e-01] | [9.420957565307617, 0.06944147497415543] |
de0e5f0d-15ba-438a-8389-607d3b6a5cbb | dense-but-efficient-videoqa-for-intricate | 2210.10300 | null | https://arxiv.org/abs/2210.10300v1 | https://arxiv.org/pdf/2210.10300v1.pdf | Dense but Efficient VideoQA for Intricate Compositional Reasoning | It is well known that most of the conventional video question answering (VideoQA) datasets consist of easy questions requiring simple reasoning processes. However, long videos inevitably contain complex and compositional semantic structures along with the spatio-temporal axis, which requires a model to understand the compositional structures inherent in the videos. In this paper, we suggest a new compositional VideoQA method based on transformer architecture with a deformable attention mechanism to address the complex VideoQA tasks. The deformable attentions are introduced to sample a subset of informative visual features from the dense visual feature map to cover a temporally long range of frames efficiently. Furthermore, the dependency structure within the complex question sentences is also combined with the language embeddings to readily understand the relations among question words. Extensive experiments and ablation studies show that the suggested dense but efficient model outperforms other baselines. | ['Eun-Sol Kim', 'Wooyoung Kang', 'Jihyeon Lee'] | 2022-10-19 | null | null | null | null | ['video-question-answering'] | ['computer-vision'] | [-1.62622020e-01 -2.96449780e-01 3.80545110e-02 -6.42176330e-01
-7.37915397e-01 -6.02629542e-01 4.35282826e-01 -5.61284065e-01
-1.47483110e-01 3.68211299e-01 8.16074073e-01 -4.40257080e-02
-9.56235602e-02 -5.31742513e-01 -8.49128246e-01 -5.64380348e-01
9.15123597e-02 1.07462674e-01 3.17509621e-01 -3.01716298e-01
3.98871973e-02 6.72949627e-02 -1.40571809e+00 6.91799939e-01
7.82783031e-01 1.07203972e+00 5.42238235e-01 6.84679210e-01
-5.39734185e-01 1.24886584e+00 -6.05355382e-01 -5.26883423e-01
6.40246421e-02 -6.68229282e-01 -9.52404976e-01 4.15231645e-01
8.11106145e-01 -9.29436386e-01 -7.74396122e-01 1.05855989e+00
1.21471912e-01 4.63365585e-01 2.52160162e-01 -1.22772515e+00
-1.17457581e+00 2.75073528e-01 -2.56882071e-01 6.35000288e-01
7.97511637e-01 6.10291839e-01 1.33581913e+00 -1.08034658e+00
7.03460336e-01 1.62814367e+00 2.20884234e-01 4.29248899e-01
-5.92463315e-01 -5.22785068e-01 6.68192029e-01 9.22663391e-01
-1.11261940e+00 -2.60511935e-01 1.00774837e+00 -2.16558635e-01
8.67469907e-01 2.42292613e-01 9.85901773e-01 1.30869317e+00
-7.12602958e-03 1.08904624e+00 6.59119189e-01 2.24075094e-01
1.29163321e-02 -3.16401184e-01 8.50559175e-02 1.02738190e+00
-2.65015632e-01 -2.88488060e-01 -5.50171614e-01 1.44371733e-01
7.60875523e-01 4.11567122e-01 -5.59371769e-01 -4.46481347e-01
-1.37824774e+00 1.02109826e+00 6.14811838e-01 2.96771049e-01
-3.42276365e-01 3.91436011e-01 4.14994121e-01 2.85676330e-01
2.12182924e-01 3.70116621e-01 -3.87764066e-01 -2.02054039e-01
-6.25242114e-01 2.36256063e-01 4.71048713e-01 1.26297545e+00
7.80410767e-01 3.86476330e-02 -4.85735685e-01 4.74031568e-01
5.19968629e-01 3.91594708e-01 2.91869700e-01 -1.30997968e+00
8.84979427e-01 9.87509489e-01 -1.38308376e-01 -1.31377292e+00
-7.88714439e-02 -2.34045610e-02 -6.77833021e-01 -3.86022538e-01
4.68505591e-01 1.11654371e-01 -8.02211523e-01 1.53083301e+00
4.75882739e-01 3.34182650e-01 -5.28321154e-02 1.25256991e+00
1.11817193e+00 1.24208164e+00 1.47202164e-01 -1.50710598e-01
1.72847116e+00 -1.36256301e+00 -1.11650419e+00 -2.30378896e-01
2.52377123e-01 -4.10685450e-01 1.67819393e+00 -9.37677845e-02
-1.15301454e+00 -7.79368460e-01 -8.35985482e-01 -9.55931485e-01
-1.95103467e-01 -2.31265098e-01 5.57039917e-01 1.26199126e-01
-8.95605683e-01 -3.95477265e-02 -5.78289628e-01 -2.60005176e-01
7.25267589e-01 4.09880653e-02 -3.76169294e-01 -6.54983222e-01
-1.45801413e+00 4.32541490e-01 1.30321845e-01 3.46425265e-01
-1.06900930e+00 -6.48470402e-01 -1.21412051e+00 2.13842899e-01
6.69289112e-01 -8.69798422e-01 1.17783105e+00 -1.38041079e+00
-1.44942892e+00 4.91619259e-01 -4.49961931e-01 -8.47496167e-02
3.76343787e-01 -5.89483261e-01 -1.52462602e-01 1.01402485e+00
1.15425847e-01 7.43006706e-01 1.13240528e+00 -9.52055335e-01
-3.61428350e-01 -1.98322251e-01 7.80853093e-01 5.61336935e-01
-3.22890878e-01 1.21943347e-01 -7.51661956e-01 -8.16278756e-01
-2.49956101e-01 -5.18392801e-01 -3.11406776e-02 4.43996102e-01
1.19486172e-02 -3.07308465e-01 1.09838414e+00 -9.99440134e-01
1.16134048e+00 -2.19479799e+00 5.09028673e-01 -2.51261681e-01
1.81825146e-01 3.09316795e-02 -4.51369584e-01 4.22866553e-01
5.23948250e-03 -9.18249190e-02 -1.28161147e-01 -4.05145921e-02
-8.49434063e-02 5.05604863e-01 -5.84316373e-01 3.02403927e-01
5.37709594e-01 1.46303737e+00 -1.30881703e+00 -6.18015349e-01
1.81019753e-01 4.74788249e-01 -8.09761882e-01 5.81728995e-01
-5.64491808e-01 4.09727067e-01 -9.54505205e-01 7.01324940e-01
4.38181132e-01 -5.43718278e-01 -1.53108463e-01 -5.59726715e-01
3.50971490e-01 1.99045226e-01 -6.20360792e-01 2.23670864e+00
-1.14116408e-01 6.84221864e-01 -4.03669477e-02 -1.17613459e+00
5.03145874e-01 3.47982943e-01 3.44196677e-01 -7.42120922e-01
5.41069992e-02 -3.32368284e-01 -1.40655831e-01 -1.24185658e+00
3.42908323e-01 6.32853806e-02 3.61372493e-02 3.31351608e-01
3.29843372e-01 -1.90179169e-01 2.34986246e-01 4.63098586e-01
9.57361221e-01 3.83280396e-01 -1.48294112e-02 -1.32699192e-01
8.30337405e-01 -2.63743494e-02 4.67496306e-01 2.62156516e-01
-5.84796190e-01 7.72368670e-01 6.67667508e-01 -7.45184839e-01
-8.39840829e-01 -1.10071015e+00 3.88065755e-01 1.14140856e+00
5.34596801e-01 -5.22070229e-01 -6.33860528e-01 -1.00393581e+00
-2.96038210e-01 3.41566056e-01 -7.80302584e-01 -2.51539946e-01
-7.64078319e-01 -7.48845711e-02 5.58809787e-02 6.01928413e-01
8.18270206e-01 -1.29538000e+00 -4.00138289e-01 -5.36365947e-03
-7.12305903e-01 -1.36339414e+00 -1.00756872e+00 -6.56072140e-01
-5.83002746e-01 -1.28874576e+00 -5.97881556e-01 -1.04477108e+00
5.90192378e-01 7.26747811e-01 1.25232077e+00 2.45349795e-01
-1.55199021e-01 7.95118928e-01 -7.65632570e-01 3.35719697e-02
1.84288681e-01 -3.88970882e-01 -4.80079830e-01 3.73668969e-01
4.87523049e-01 -3.37090105e-01 -8.79611671e-01 1.25234425e-01
-1.08470178e+00 1.56373769e-01 4.08569604e-01 8.83145154e-01
6.25834823e-01 -2.64976144e-01 5.19966006e-01 -5.57029188e-01
4.29552317e-01 -7.20832109e-01 -2.44431511e-01 4.95998055e-01
2.94932038e-01 5.03356755e-02 6.63895845e-01 -5.79888225e-01
-1.24989998e+00 -8.83079544e-02 -1.09113671e-01 -7.87567317e-01
-8.70084986e-02 3.45112413e-01 -5.34373105e-01 3.52953464e-01
4.97269928e-02 3.87870461e-01 2.80935913e-02 -5.18832542e-02
6.98187649e-01 2.14508012e-01 3.20697218e-01 -5.15037000e-01
9.19384480e-01 7.13727176e-01 -3.07379842e-01 -7.71306038e-01
-1.17966723e+00 -4.66699451e-01 -4.74764258e-01 -4.04577971e-01
1.40728188e+00 -1.15566194e+00 -5.57720959e-01 2.34306559e-01
-1.30489612e+00 -2.74553657e-01 -3.89367193e-01 3.24592859e-01
-7.16915905e-01 7.23549426e-01 -7.19659209e-01 -3.61383557e-01
-2.32653335e-01 -1.25172353e+00 1.24449825e+00 2.84308225e-01
-2.50306446e-02 -8.23906779e-01 -2.17167228e-01 9.23740327e-01
1.05618559e-01 -3.18156481e-02 8.19144547e-01 -2.12776259e-01
-1.25163829e+00 1.43852532e-01 -5.32992423e-01 3.24602574e-01
1.41225159e-01 6.35215268e-03 -6.02156699e-01 -1.48328781e-01
3.37946296e-01 -4.98610854e-01 7.69186378e-01 1.89394772e-01
1.67300308e+00 -4.48413104e-01 2.01157287e-01 6.97965860e-01
1.08556283e+00 1.76455811e-01 7.77903080e-01 -3.58208865e-02
9.93798137e-01 7.48228371e-01 7.83395410e-01 3.79722239e-03
8.65423024e-01 3.05815220e-01 6.36102855e-01 1.19139083e-01
-1.18753418e-01 -4.37537402e-01 5.71455002e-01 1.03651464e+00
6.81013018e-02 -1.25323668e-01 -5.51625431e-01 5.82273185e-01
-1.83846188e+00 -1.37836075e+00 9.58597660e-02 1.29601645e+00
6.91952109e-01 -3.65499556e-01 -3.71909961e-02 -2.57682711e-01
6.31254792e-01 6.36574984e-01 -4.58144218e-01 7.50839850e-03
-8.99522454e-02 -1.64500117e-01 -3.00271869e-01 3.55228513e-01
-9.83978808e-01 7.54551053e-01 6.21507692e+00 6.10721111e-01
-7.09040403e-01 9.44565609e-02 4.64862227e-01 -2.67161041e-01
-6.62448883e-01 -5.57433739e-02 -3.09595257e-01 5.16016185e-01
5.53632081e-01 -8.14799592e-02 4.28812891e-01 5.58996141e-01
1.15774430e-01 1.10286944e-01 -9.15381014e-01 1.14839005e+00
4.75594938e-01 -1.51032102e+00 5.72674572e-01 -3.31697822e-01
7.19058752e-01 -3.13890070e-01 6.56880960e-02 4.44087893e-01
1.57572627e-02 -9.41766024e-01 7.78505921e-01 6.52640164e-01
5.71231365e-01 -5.92113018e-01 4.96248037e-01 1.07390575e-01
-1.46991003e+00 -3.99397373e-01 -4.17260736e-01 -1.26153901e-01
4.87289697e-01 8.62731338e-02 -1.12658786e-02 5.85382044e-01
1.04645967e+00 1.25252509e+00 -6.99778080e-01 6.40963197e-01
-3.01983356e-01 5.14736354e-01 7.76329711e-02 -1.88124031e-01
5.96998394e-01 -3.87632579e-01 3.36989552e-01 9.19913709e-01
3.10307115e-01 5.66914916e-01 -7.59307221e-02 7.35409081e-01
-1.14542902e-01 5.83074763e-02 -5.97718060e-01 -2.16041982e-01
6.19676560e-02 1.17431617e+00 -4.24207717e-01 -4.53556567e-01
-1.01579630e+00 1.02511692e+00 4.56869870e-01 7.06862450e-01
-1.03313708e+00 -8.06223080e-02 8.28068376e-01 -5.02968170e-02
5.64587951e-01 -3.40937823e-01 4.15943295e-01 -1.71453047e+00
3.25910717e-01 -1.07620859e+00 6.89748406e-01 -1.24792302e+00
-1.48997676e+00 4.92746651e-01 1.46339918e-02 -1.21786523e+00
4.01021242e-02 -5.64953327e-01 -5.64748585e-01 4.95390832e-01
-1.70053267e+00 -1.25105011e+00 -7.60368705e-01 1.14126003e+00
1.26838064e+00 2.98869349e-02 4.04676318e-01 3.11936468e-01
-4.25128758e-01 2.17057660e-01 -3.13950241e-01 3.76549006e-01
4.99316007e-01 -1.11054730e+00 2.51859277e-01 9.18170631e-01
2.64688492e-01 4.94049728e-01 3.33040893e-01 -2.91176766e-01
-1.77236366e+00 -1.17500842e+00 4.81374294e-01 -6.45362377e-01
9.19936240e-01 -4.90196913e-01 -1.18029952e+00 8.81894588e-01
6.60777450e-01 3.65483284e-01 5.88226914e-01 -3.93077701e-01
-6.56584740e-01 -1.88062623e-01 -7.58496106e-01 6.92660868e-01
1.35320842e+00 -1.00233924e+00 -1.12772691e+00 4.26199675e-01
1.15248001e+00 -2.09939480e-01 -5.75538993e-01 1.84981972e-01
4.40142214e-01 -8.63331437e-01 1.14591563e+00 -8.97355258e-01
7.61152208e-01 -3.95286918e-01 -2.64379531e-01 -8.69109929e-01
-3.13769281e-01 -5.84312916e-01 -5.24780571e-01 1.11924851e+00
-1.42318562e-01 -1.55670092e-01 5.02804101e-01 4.45595235e-01
-1.05021454e-01 -6.98338151e-01 -8.71321082e-01 -3.54171813e-01
-1.28415987e-01 -2.60741264e-01 5.12482703e-01 9.11109746e-01
-1.28521487e-01 6.79887891e-01 -4.10447001e-01 1.42858908e-01
3.17212343e-01 4.10653919e-01 6.66650116e-01 -6.89435840e-01
-2.10538015e-01 -2.45540991e-01 -4.83675212e-01 -1.67606509e+00
3.13337773e-01 -5.56617856e-01 2.29839124e-02 -1.70188880e+00
3.55550021e-01 2.40629673e-01 -1.38344988e-01 -3.08256317e-02
-6.66482329e-01 1.67475976e-02 1.51869163e-01 -3.73756140e-02
-1.21558726e+00 1.02072084e+00 1.83198273e+00 -3.27277958e-01
2.40478933e-01 -6.43638790e-01 -5.01871943e-01 6.16886735e-01
3.76966119e-01 -6.64174259e-02 -9.33296859e-01 -9.51169491e-01
3.40436071e-01 2.21146107e-01 4.90541756e-01 -5.95650613e-01
-1.76081937e-02 -4.44597542e-01 4.09014106e-01 -6.83576882e-01
5.29652596e-01 -1.04709876e+00 -2.42775813e-01 2.09328279e-01
-3.81539047e-01 4.65435386e-01 -9.24965143e-02 9.20263648e-01
-6.69006646e-01 1.46499678e-01 3.80146235e-01 -2.08891332e-01
-1.11932361e+00 7.82435238e-01 -1.96077824e-01 4.52849895e-01
1.38012993e+00 -4.56211120e-02 -3.66357893e-01 -7.76308715e-01
-6.41128659e-01 7.17812717e-01 1.10969909e-01 8.99013638e-01
1.03969717e+00 -1.53592646e+00 -5.20442247e-01 1.86289057e-01
1.32103115e-01 3.42092544e-01 8.25081170e-01 6.25009656e-01
-7.99302161e-01 2.71089345e-01 -3.46866548e-01 -4.91895288e-01
-1.16793478e+00 1.04013681e+00 3.56498659e-01 -3.85440700e-03
-8.27451766e-01 1.07675195e+00 8.68149698e-01 -1.02471247e-01
2.15800732e-01 -6.48605824e-01 -4.55197990e-01 2.23524347e-01
8.79484236e-01 -3.44091021e-02 -7.18601108e-01 -7.79172778e-01
-2.49064848e-01 6.91329479e-01 7.56485462e-02 1.77649066e-01
1.23293078e+00 -6.20276272e-01 -1.50771454e-01 3.62680197e-01
1.38840079e+00 -2.81037241e-01 -1.75075674e+00 -3.03936422e-01
-2.62766808e-01 -7.63273835e-01 -3.78939033e-01 -1.72244936e-01
-1.29853225e+00 1.26564085e+00 1.50314957e-01 2.15966403e-01
1.27817452e+00 2.84485430e-01 9.67349291e-01 3.52730870e-01
1.12693168e-01 -7.43437350e-01 9.01055932e-01 4.33469236e-01
1.33093941e+00 -1.24627566e+00 -1.42791703e-01 -4.62713689e-01
-8.06764781e-01 1.12572563e+00 9.25422788e-01 -3.76852483e-01
6.02533638e-01 -3.33027422e-01 1.83411852e-01 -4.13811564e-01
-9.61195767e-01 -2.84352452e-01 3.60144943e-01 6.15380704e-01
-8.94973055e-02 -4.04205412e-01 -2.06363425e-02 6.47943437e-01
1.99087188e-01 -2.14038804e-01 3.78849208e-01 6.90949380e-01
-3.47541571e-01 -6.17093086e-01 -1.42584473e-01 1.99360669e-01
-3.93586755e-01 -8.42470750e-02 -1.88017026e-01 7.74422586e-01
1.01961672e-01 9.31628883e-01 2.66074777e-01 -1.09788559e-01
3.04290265e-01 1.09411173e-01 4.40170437e-01 -4.94525135e-01
-3.69956255e-01 -1.70720026e-01 -1.10534437e-01 -1.13707125e+00
-7.59824574e-01 -6.95286036e-01 -1.26699686e+00 5.78953512e-02
4.31409180e-02 2.87375804e-02 6.93691000e-02 1.23305535e+00
3.57030720e-01 5.86288095e-01 4.33817565e-01 -4.93873328e-01
-2.78003097e-01 -6.97617412e-01 -3.18375736e-01 8.26212108e-01
7.15972424e-01 -6.02608979e-01 -3.35186660e-01 3.94247949e-01] | [10.407377243041992, 1.0591470003128052] |
76cebbc2-cab6-4b7a-b65c-a7184c15b3c0 | the-open-images-dataset-v4-unified-image | 1811.00982 | null | https://arxiv.org/abs/1811.00982v2 | https://arxiv.org/pdf/1811.00982v2.pdf | The Open Images Dataset V4: Unified image classification, object detection, and visual relationship detection at scale | We present Open Images V4, a dataset of 9.2M images with unified annotations for image classification, object detection and visual relationship detection. The images have a Creative Commons Attribution license that allows to share and adapt the material, and they have been collected from Flickr without a predefined list of class names or tags, leading to natural class statistics and avoiding an initial design bias. Open Images V4 offers large scale across several dimensions: 30.1M image-level labels for 19.8k concepts, 15.4M bounding boxes for 600 object classes, and 375k visual relationship annotations involving 57 classes. For object detection in particular, we provide 15x more bounding boxes than the next largest datasets (15.4M boxes on 1.9M images). The images often show complex scenes with several objects (8 annotated objects per image on average). We annotated visual relationships between them, which support visual relationship detection, an emerging task that requires structured reasoning. We provide in-depth comprehensive statistics about the dataset, we validate the quality of the annotations, we study how the performance of several modern models evolves with increasing amounts of training data, and we demonstrate two applications made possible by having unified annotations of multiple types coexisting in the same images. We hope that the scale, quality, and variety of Open Images V4 will foster further research and innovation even beyond the areas of image classification, object detection, and visual relationship detection. | ['Tom Duerig', 'Jordi Pont-Tuset', 'Matteo Malloci', 'Ivan Krasin', 'Alexander Kolesnikov', 'Vittorio Ferrari', 'Shahab Kamali', 'Jasper Uijlings', 'Hassan Rom', 'Alina Kuznetsova', 'Neil Alldrin', 'Stefan Popov'] | 2018-11-02 | null | null | null | null | ['visual-relationship-detection'] | ['computer-vision'] | [ 6.93843216e-02 2.04693392e-01 -3.58833015e-01 -2.91852921e-01
-3.61761391e-01 -9.97473121e-01 6.19398057e-01 1.87113732e-01
-3.91017854e-01 2.93908924e-01 2.58508511e-03 -2.40914747e-01
-1.52342394e-01 -5.21224141e-01 -7.14584172e-01 -4.51398909e-01
-3.86914343e-01 4.06590641e-01 6.21066630e-01 -3.60995978e-02
-3.96641083e-02 3.52590978e-01 -1.88887632e+00 5.62637627e-01
4.06032681e-01 1.37516320e+00 2.97864288e-01 7.22016156e-01
-1.25565857e-01 8.99613321e-01 -5.44031620e-01 -5.74429095e-01
3.80396396e-01 3.90509784e-01 -1.07739210e+00 5.70508897e-01
1.01998079e+00 -2.47758850e-01 -1.78593516e-01 1.19898462e+00
-3.99474055e-02 -3.26004066e-02 7.94149280e-01 -1.79535520e+00
-1.27312338e+00 4.76913601e-01 -7.25281477e-01 1.48614973e-01
4.12654698e-01 2.00099096e-01 1.10914791e+00 -1.04890406e+00
8.68896306e-01 1.25245929e+00 5.65754592e-01 3.82716686e-01
-1.00580823e+00 -6.94608927e-01 2.01521352e-01 2.21689656e-01
-1.75757110e+00 -1.67022571e-01 1.48423076e-01 -7.74953127e-01
8.88460994e-01 4.99460369e-01 7.24459529e-01 8.68604004e-01
-4.63286042e-03 5.56132197e-01 8.05503428e-01 -2.89855838e-01
-3.63746956e-02 5.44913948e-01 3.01316649e-01 7.71797240e-01
3.48171443e-01 -3.01282942e-01 -1.53794706e-01 9.44439322e-02
7.02556849e-01 2.45207816e-01 -8.73387232e-02 -5.12159705e-01
-1.32807541e+00 5.74147522e-01 9.15101171e-01 5.93761243e-02
1.68049425e-01 2.64685214e-01 4.37153339e-01 1.75401166e-01
2.45105386e-01 3.44021261e-01 -3.60368818e-01 3.67024064e-01
-5.21553040e-01 4.50749509e-02 3.45149785e-01 1.68886602e+00
8.20976019e-01 -4.29934144e-01 -2.93825362e-02 1.04800105e+00
2.44061917e-01 5.68124592e-01 1.60119832e-01 -1.37079024e+00
3.97543877e-01 9.68356967e-01 9.78160724e-02 -1.22143519e+00
-4.50041145e-01 -1.69624463e-01 -8.67251575e-01 1.59708902e-01
4.06545490e-01 4.43083763e-01 -9.10118937e-01 1.29911411e+00
3.24845791e-01 -4.19415623e-01 -1.63626149e-01 9.18752551e-01
1.33636308e+00 4.68237281e-01 1.21792980e-01 2.86447734e-01
1.94222128e+00 -1.18434775e+00 -5.56807935e-01 -1.71536744e-01
6.35408878e-01 -8.05163205e-01 1.30685949e+00 3.12793702e-01
-8.53778422e-01 -7.50407636e-01 -8.68956506e-01 -3.09029847e-01
-9.74764049e-01 9.68298838e-02 7.59612799e-01 4.26252067e-01
-9.83408272e-01 9.96951829e-04 2.01192014e-02 -5.58675945e-01
8.28643680e-01 1.32904053e-01 -6.17194831e-01 -2.44474411e-01
-9.57446933e-01 6.79957747e-01 5.43588042e-01 -2.07236752e-01
-9.55094039e-01 -4.94744062e-01 -9.07130897e-01 -1.08848125e-01
5.53856909e-01 -3.77064914e-01 9.22260106e-01 -7.67805696e-01
-4.81596917e-01 1.49404895e+00 4.05579865e-01 -3.29068214e-01
4.15498912e-01 1.07800690e-02 -5.56972742e-01 4.50790554e-01
5.20520866e-01 1.31438828e+00 7.39603341e-01 -1.47570086e+00
-1.02641118e+00 -1.34191692e-01 5.39844215e-01 -5.70333935e-02
-6.49152875e-01 2.73675710e-01 -8.09972405e-01 -5.89463592e-01
-5.56048378e-02 -9.93763745e-01 -1.99229315e-01 4.17358875e-01
-3.55145603e-01 -2.49295369e-01 8.69674981e-01 -1.59314245e-01
1.09448099e+00 -2.45211554e+00 -2.39517212e-01 -5.48069887e-02
6.31414592e-01 1.85985938e-02 -2.91250110e-01 1.20824009e-01
-3.46319862e-02 4.65588123e-01 -5.09276390e-02 -2.66067386e-01
-1.38171101e-02 4.44282144e-01 -3.74901116e-01 4.11501080e-01
7.32526630e-02 1.09620488e+00 -9.80634034e-01 -1.01126027e+00
2.31313959e-01 1.39282987e-01 -3.69507104e-01 1.98518601e-03
-2.05260083e-01 1.43146083e-01 1.12484246e-02 9.51109469e-01
6.90007567e-01 -7.27201402e-01 -2.74127033e-02 -4.90954041e-01
-8.18280056e-02 -4.75484461e-01 -1.33344638e+00 1.31620443e+00
-1.11295484e-01 9.95674372e-01 -5.91725335e-02 -4.78707910e-01
7.11620927e-01 8.40100050e-02 4.87921476e-01 -5.48229337e-01
8.71795043e-02 -3.15719023e-02 -1.43430755e-01 -6.40530527e-01
8.18872273e-01 4.87065107e-01 -3.19642365e-01 3.36908937e-01
1.98255051e-02 -2.00842664e-01 5.24294198e-01 6.38183713e-01
7.52338767e-01 -2.08654046e-01 3.91056865e-01 -1.94487497e-01
3.19463313e-01 2.83789843e-01 2.27283537e-01 9.95828927e-01
-3.94740909e-01 6.50678098e-01 4.40133363e-01 -6.63903296e-01
-1.22310436e+00 -1.01030564e+00 -6.88177407e-01 1.17015386e+00
5.98433852e-01 -6.51351213e-01 -4.36673939e-01 -7.45962858e-01
2.49899566e-01 -3.64382938e-03 -8.37992430e-01 2.39581868e-01
-1.08159244e-01 -3.73084038e-01 6.50302351e-01 5.91962755e-01
6.07211053e-01 -8.98049593e-01 -6.93887711e-01 -4.01634097e-01
-2.13656232e-01 -1.65053082e+00 -4.37317818e-01 6.52060881e-02
-3.93974423e-01 -1.47581387e+00 -5.18382907e-01 -1.03423214e+00
8.89505804e-01 6.28102422e-01 1.39570355e+00 1.79468423e-01
-7.07377315e-01 7.46398985e-01 -5.46148419e-01 -3.01484764e-01
-2.91034400e-01 -3.48609805e-01 1.12241648e-01 -5.02817146e-02
1.51718214e-01 -4.88776043e-02 -6.22425079e-01 8.06516171e-01
-9.25409079e-01 1.66384980e-01 5.05160689e-01 4.99365211e-01
6.18011713e-01 2.24367484e-01 1.15982397e-02 -8.53370070e-01
-6.09357506e-02 -5.54386079e-01 -6.59681857e-01 4.34443474e-01
-4.60501820e-01 -5.09929657e-01 1.54114604e-01 -5.63885272e-01
-7.34877229e-01 2.43768655e-03 5.65657318e-01 -5.31511426e-01
-3.62499058e-01 -8.52361694e-03 -6.36190549e-02 -9.45405066e-02
7.31642365e-01 -2.56618947e-01 -2.78571308e-01 -2.54183680e-01
9.59307194e-01 8.06804121e-01 7.27500498e-01 -1.38947427e-01
7.64169097e-01 7.35716343e-01 -8.30904394e-02 -7.00656354e-01
-1.14418471e+00 -8.54561090e-01 -8.33445430e-01 -5.14807343e-01
1.19763029e+00 -1.09274089e+00 -7.68762529e-01 3.07996690e-01
-1.05515790e+00 -3.09039265e-01 -2.49522045e-01 8.85813870e-03
-4.12351042e-01 3.54335248e-01 -5.99809051e-01 -6.63497329e-01
9.54712089e-03 -9.53810871e-01 1.21074414e+00 -6.38234615e-02
-2.13569552e-01 -7.61226118e-01 -6.09083414e-01 6.95721686e-01
4.38676141e-02 1.80615574e-01 7.34143853e-01 -2.69504666e-01
-7.98695445e-01 -9.72861052e-02 -8.98272753e-01 3.81308526e-01
-3.65904793e-02 2.11899072e-01 -9.71130371e-01 -2.30918340e-02
-5.48927784e-01 -5.64525902e-01 9.43820000e-01 1.50402889e-01
1.47416198e+00 -2.91167051e-01 -6.66940093e-01 5.06463408e-01
1.34218025e+00 -5.46629075e-04 5.27973652e-01 3.72082829e-01
1.13506627e+00 8.35302889e-01 9.06904817e-01 3.66819769e-01
3.35337490e-01 6.48341000e-01 9.24082339e-01 -2.94920534e-01
-3.24382126e-01 8.69835764e-02 6.40134811e-02 3.18955362e-01
-9.68256518e-02 -2.49107778e-01 -1.13912630e+00 5.21244764e-01
-1.76161265e+00 -9.23715472e-01 -6.55919075e-01 1.72093010e+00
7.26489484e-01 2.32374310e-01 1.26794398e-01 1.10620961e-01
7.62374818e-01 2.77427342e-02 -3.62836927e-01 1.18869744e-01
-3.72471273e-01 -4.12064791e-01 7.73316085e-01 2.22590685e-01
-1.38204443e+00 7.77126610e-01 7.08635712e+00 8.52937937e-01
-6.03179812e-01 1.08615920e-01 5.84542632e-01 -9.15204734e-02
-5.97671568e-02 1.19944420e-02 -9.34407234e-01 2.76800066e-01
3.90277147e-01 2.47013420e-01 1.82653323e-01 1.12668419e+00
-2.65778005e-01 -3.63615483e-01 -1.20983458e+00 1.37066543e+00
1.94778338e-01 -1.52372527e+00 2.67457575e-01 2.19112128e-01
8.08642745e-01 -1.11437060e-01 2.13426828e-01 1.72247306e-01
1.77729011e-01 -9.97363925e-01 1.12686777e+00 4.47975814e-01
1.09056652e+00 -4.01509851e-01 5.41507244e-01 1.07909463e-01
-1.46899462e+00 -4.20583427e-01 -3.83417547e-01 -1.04042232e-01
-1.35534614e-01 4.09490883e-01 -5.69249630e-01 2.00015932e-01
1.40858006e+00 1.09716606e+00 -1.29008067e+00 9.05156612e-01
5.27558103e-02 2.10349131e-02 -6.94591925e-02 8.30137879e-02
1.32316038e-01 4.26584110e-03 2.77041197e-01 1.21883011e+00
-2.10823029e-01 -2.92854514e-02 4.97361690e-01 7.18163729e-01
-1.51361704e-01 -1.49250314e-01 -6.29234433e-01 1.60919130e-01
5.50232768e-01 1.58186305e+00 -1.25096774e+00 -5.52959979e-01
-6.14212513e-01 8.67631733e-01 2.19686002e-01 2.72223711e-01
-1.16916108e+00 -2.16489390e-01 5.81945240e-01 3.21259916e-01
3.16153407e-01 -1.29080042e-01 -3.65353860e-02 -9.71026421e-01
-1.01260744e-01 -6.39285505e-01 8.11292589e-01 -1.23102474e+00
-1.46029139e+00 5.65459490e-01 3.08163881e-01 -1.30739486e+00
3.10264260e-01 -1.04392350e+00 1.08553976e-01 1.43711388e-01
-1.20028770e+00 -1.29880333e+00 -7.39151776e-01 5.63687980e-01
4.77159023e-01 -2.94739995e-02 8.16318989e-01 6.54978693e-01
-6.08329356e-01 3.85797501e-01 -1.18135866e-02 4.82749164e-01
8.79070938e-01 -1.29779267e+00 2.07823992e-01 4.46701586e-01
3.18645120e-01 3.74468833e-01 3.95622313e-01 -5.01600862e-01
-9.30565298e-01 -1.32937789e+00 4.56032515e-01 -1.14552307e+00
7.92407811e-01 -7.07882166e-01 -6.95587516e-01 7.81902313e-01
2.10553482e-01 6.22818649e-01 6.32482946e-01 4.33208309e-02
-7.73402035e-01 -1.53615579e-01 -9.20084417e-01 5.64721942e-01
1.47239029e+00 -4.72883075e-01 -2.33309403e-01 6.49219036e-01
9.78655756e-01 -3.47475111e-01 -1.04321587e+00 4.53281581e-01
6.54832482e-01 -9.11080182e-01 1.25917447e+00 -5.02465487e-01
3.39204550e-01 -5.54358602e-01 -5.47904551e-01 -5.59587538e-01
-4.88930613e-01 -5.81931742e-03 -1.94315389e-01 1.29706621e+00
2.45606467e-01 -4.40902084e-01 4.44473207e-01 5.08909166e-01
-8.71788487e-02 -8.11207116e-01 -4.39744532e-01 -1.09049487e+00
-3.29941630e-01 -7.30000496e-01 5.21053493e-01 9.48566020e-01
-2.85056293e-01 2.71010280e-01 -2.52808809e-01 1.51638329e-01
7.05307245e-01 2.85913527e-01 8.60823572e-01 -1.22610724e+00
-8.94281417e-02 -6.96814060e-01 -7.66214013e-01 -9.59698021e-01
-1.56394675e-01 -8.12092423e-01 -3.57292295e-01 -1.55485177e+00
6.59995735e-01 -7.94975579e-01 -1.01960801e-01 7.27177441e-01
4.37225699e-02 1.14486682e+00 3.30785841e-01 4.90424991e-01
-1.37054384e+00 6.93188012e-02 1.25738525e+00 -6.09105587e-01
2.48615861e-01 -5.71777403e-01 -6.16899788e-01 9.16137695e-01
4.11387265e-01 -2.90497601e-01 -4.41485316e-01 -4.27257836e-01
3.95628214e-01 -6.57662392e-01 8.27265143e-01 -1.05116713e+00
7.42780268e-02 -2.25804389e-01 6.61169052e-01 -7.88062930e-01
3.39578837e-01 -1.00599027e+00 2.13348866e-01 2.46864080e-01
-3.25491846e-01 -1.56271860e-01 2.09364742e-01 6.47439003e-01
-1.85044333e-01 -1.32670954e-01 7.66657352e-01 -3.48387331e-01
-1.42330623e+00 3.47483695e-01 -1.18279316e-01 2.76025593e-01
1.47152722e+00 -4.28663105e-01 -6.02197111e-01 -1.54103830e-01
-9.60499227e-01 5.28117597e-01 6.36488795e-01 7.85893083e-01
6.06679261e-01 -1.52562952e+00 -4.09257770e-01 1.05181793e-02
9.14773762e-01 1.39966279e-01 3.44346404e-01 4.75643665e-01
-4.27838922e-01 2.41117910e-01 -1.41815051e-01 -1.14032972e+00
-1.42702317e+00 1.06877053e+00 1.69407040e-01 2.31259644e-01
-7.41153359e-01 8.13585877e-01 5.22960901e-01 -1.80349350e-01
5.23736954e-01 -4.20460731e-01 -4.02084827e-01 4.10144806e-01
6.41901612e-01 3.21850270e-01 -2.67019749e-01 -8.54698300e-01
-5.28933525e-01 7.91382313e-01 4.09900621e-02 3.64998072e-01
9.85003829e-01 -5.14483273e-01 -2.36160830e-01 5.00894666e-01
1.14522767e+00 -1.25017330e-01 -1.15576482e+00 -1.60631567e-01
-8.80362540e-02 -6.76257253e-01 -3.83578598e-01 -7.53922403e-01
-9.20350671e-01 4.83334690e-01 6.00135922e-01 6.12274289e-01
1.03748190e+00 7.17927814e-01 4.09837514e-01 1.38801843e-01
5.07553339e-01 -9.56760585e-01 5.71086466e-01 5.02701283e-01
9.54674602e-01 -1.65189242e+00 2.46197894e-01 -8.25277627e-01
-6.57491624e-01 8.29804182e-01 8.82530630e-01 2.15787753e-01
6.54368103e-01 2.50454664e-01 1.63314447e-01 -4.72582668e-01
-6.23656392e-01 -5.20144761e-01 4.83101338e-01 7.68922865e-01
3.13704796e-02 1.74150005e-01 3.09324354e-01 2.98278391e-01
-7.68334493e-02 -6.42278969e-01 4.96745378e-01 5.96702635e-01
-4.10689384e-01 -7.37906098e-01 -6.27724051e-01 3.40998739e-01
-2.46761534e-02 -7.51186237e-02 -4.12545890e-01 9.76407409e-01
7.79875994e-01 1.07224584e+00 3.88207406e-01 -3.24185789e-01
2.83714682e-01 -4.24932152e-01 4.71550584e-01 -7.14725733e-01
-2.10469320e-01 -3.40753078e-01 7.07281306e-02 -5.50419807e-01
-5.93212724e-01 -3.30713272e-01 -1.30837178e+00 -8.39088783e-02
-4.21368420e-01 -1.37939319e-01 5.15285194e-01 3.59688193e-01
4.92644638e-01 4.40773875e-01 3.22221279e-01 -6.59121037e-01
8.78113434e-02 -7.22058475e-01 -6.24886394e-01 7.95263469e-01
3.18339407e-01 -8.80337536e-01 -3.19243640e-01 4.69498932e-01] | [9.98962688446045, 1.5758353471755981] |
788d0aa6-a78a-471a-a18b-f60ee1a70683 | lake-ice-monitoring-with-webcams-and-crowd | 2002.07875 | null | https://arxiv.org/abs/2002.07875v2 | https://arxiv.org/pdf/2002.07875v2.pdf | Lake Ice Monitoring with Webcams and Crowd-Sourced Images | Lake ice is a strong climate indicator and has been recognised as part of the Essential Climate Variables (ECV) by the Global Climate Observing System (GCOS). The dynamics of freezing and thawing, and possible shifts of freezing patterns over time, can help in understanding the local and global climate systems. One way to acquire the spatio-temporal information about lake ice formation, independent of clouds, is to analyse webcam images. This paper intends to move towards a universal model for monitoring lake ice with freely available webcam data. We demonstrate good performance, including the ability to generalise across different winters and different lakes, with a state-of-the-art Convolutional Neural Network (CNN) model for semantic image segmentation, Deeplab v3+. Moreover, we design a variant of that model, termed Deep-U-Lab, which predicts sharper, more correct segmentation boundaries. We have tested the model's ability to generalise with data from multiple camera views and two different winters. On average, it achieves intersection-over-union (IoU) values of ~71% across different cameras and ~69% across different winters, greatly outperforming prior work. Going even further, we show that the model even achieves 60% IoU on arbitrary images scraped from photo-sharing web sites. As part of the work, we introduce a new benchmark dataset of webcam images, Photi-LakeIce, from multiple cameras and two different winters, along with pixel-wise ground truth annotations. | ['Laura Leal-Taixe', 'Emmanuel Baltsavias', 'Manu Tom', 'Konrad Schindler', 'Rajanie Prabha', 'Mathias Rothermel'] | 2020-02-18 | null | null | null | null | ['webcam-rgb-image-classification', 'lake-ice-detection', 'lake-detection', 'change-detection-for-remote-sensing-images', 'lake-ice-detection', 'segmentation-of-remote-sensing-imagery', 'remote-sensing-image-classification', 'the-semantic-segmentation-of-remote-sensing'] | ['computer-code', 'computer-vision', 'computer-vision', 'miscellaneous', 'miscellaneous', 'miscellaneous', 'miscellaneous', 'miscellaneous'] | [ 8.71278625e-03 -3.46492529e-01 3.50989223e-01 -6.04254961e-01
-6.13637686e-01 -1.03600073e+00 4.65838879e-01 -1.69990629e-01
-4.92178231e-01 4.57403183e-01 -2.11652771e-01 -3.59030217e-01
5.72577000e-01 -8.23001027e-01 -9.09013331e-01 -7.37634897e-01
-2.62335837e-01 3.73237729e-01 3.06418955e-01 -4.29409385e-01
-1.04218572e-01 3.47808838e-01 -1.72736955e+00 1.51837140e-01
9.18540716e-01 8.12400997e-01 4.50944066e-01 1.13774168e+00
-3.91448252e-02 1.16622739e-01 -2.13370144e-01 1.70928575e-02
5.55743098e-01 -2.13599280e-01 -8.37705672e-01 2.15436965e-01
1.00941122e+00 -1.42099842e-01 9.37602744e-02 9.80765641e-01
4.33394536e-02 -2.98734009e-01 4.25371319e-01 -1.03475869e+00
-1.04528129e-01 -6.62188381e-02 -4.39047128e-01 3.85073096e-01
-2.59506665e-02 6.61884129e-01 9.99543786e-01 -4.53254640e-01
5.69847941e-01 1.06106961e+00 9.91964102e-01 2.62753636e-01
-1.21144068e+00 -6.97959065e-01 3.19881827e-01 -1.62179843e-01
-1.16192603e+00 -3.09197605e-01 1.85485095e-01 -6.40278041e-01
1.03808355e+00 4.61887300e-01 1.07878709e+00 6.45673513e-01
1.14604121e-03 6.77810550e-01 1.58917809e+00 -3.33089620e-01
2.05822185e-01 -7.39378929e-02 1.15125887e-01 5.51402211e-01
3.63919944e-01 2.07844153e-01 -3.89155418e-01 1.19529106e-01
4.48553175e-01 -1.38288820e-02 -5.58331966e-01 -1.80289924e-01
-9.63200152e-01 6.31119132e-01 7.67814219e-01 -3.38696671e-04
4.51313443e-02 2.22359240e-01 1.06241360e-01 3.97197932e-01
7.26837695e-01 1.98288336e-01 -9.46430206e-01 7.00916648e-02
-1.10771537e+00 5.19511998e-01 8.50582600e-01 9.02215600e-01
1.16407073e+00 -2.82853365e-01 6.98785603e-01 3.12525094e-01
8.41279626e-01 1.24159253e+00 1.13254569e-01 -1.00656879e+00
5.13949215e-01 5.99758685e-01 2.67646343e-01 -4.30888921e-01
-6.15919173e-01 7.29321763e-02 -7.50592470e-01 5.81381321e-01
5.32813191e-01 -3.08417767e-01 -1.36320615e+00 1.19538116e+00
2.11369812e-01 2.48242021e-01 2.09896356e-01 1.00936770e+00
7.73192108e-01 7.87861466e-01 -3.71463150e-02 1.73937529e-01
1.46811461e+00 -6.22066438e-01 -3.40007246e-01 -8.14976990e-01
5.72873056e-01 -4.15597826e-01 1.12345862e+00 1.61424190e-01
-3.91589075e-01 -3.36548895e-01 -1.07071376e+00 2.13911682e-01
-8.54373276e-01 -1.30624339e-01 6.15602255e-01 7.05430090e-01
-1.31719422e+00 6.38016045e-01 -1.17109644e+00 -9.32295084e-01
2.52789319e-01 7.17871860e-02 -3.27753216e-01 1.40048519e-01
-1.13949811e+00 7.47530520e-01 1.92856252e-01 6.46267772e-01
-9.96424019e-01 -8.22930992e-01 -9.69904840e-01 -1.63660154e-01
-1.45657271e-01 -4.27719623e-01 1.14016938e+00 -9.98951375e-01
-1.14893913e+00 1.22704709e+00 -1.66402787e-01 -5.03624320e-01
5.88999987e-01 -4.19510484e-01 -2.40901634e-01 9.61534157e-02
4.94397730e-02 1.02015996e+00 3.44654113e-01 -1.53493631e+00
-8.98887575e-01 -4.86292601e-01 -8.04615915e-02 3.30185443e-01
2.77239740e-01 -1.68562293e-01 -5.24201989e-01 1.93149019e-02
1.10682346e-01 -1.28048050e+00 -4.08408821e-01 1.83157623e-01
-4.09230255e-02 2.17898265e-01 1.05092382e+00 -7.19531238e-01
6.73921227e-01 -1.97853374e+00 -9.17660221e-02 1.27602637e-01
-3.35481837e-02 2.29568064e-01 4.56599966e-02 8.93901736e-02
1.42980382e-01 3.64749640e-01 -1.05603456e+00 -4.72922742e-01
-2.19753444e-01 6.93233609e-01 -2.33663499e-01 7.10732043e-01
2.46049911e-01 8.12197268e-01 -8.08581471e-01 -2.76502132e-01
4.63346213e-01 6.39862344e-02 -1.81083098e-01 1.20878786e-01
-4.75214005e-01 6.53076053e-01 5.11937849e-02 8.15549076e-01
1.21748877e+00 -1.67509280e-02 3.06219041e-01 3.65158170e-01
-3.97070199e-01 -1.23629726e-01 -1.05529284e+00 1.53515005e+00
-3.35475355e-01 1.13445020e+00 5.44616342e-01 -5.22622287e-01
9.40888047e-01 1.11888230e-01 1.46173716e-01 -7.95694530e-01
-1.36745557e-01 2.53522843e-01 -3.25245678e-01 -6.56743526e-01
3.94849360e-01 -1.50268629e-01 1.90225728e-02 1.69486068e-02
-2.42500007e-01 -6.23046875e-01 4.00738083e-02 8.04745313e-03
6.65257633e-01 4.22950119e-01 -2.27244914e-01 -7.99310088e-01
2.20081061e-01 4.14890021e-01 4.50431466e-01 7.24637568e-01
-6.19619131e-01 1.06135035e+00 1.88820764e-01 -1.00253046e+00
-1.15142274e+00 -1.03985786e+00 -4.42281455e-01 7.33598053e-01
5.07840037e-01 -1.06904507e-01 -7.43409991e-01 -4.07308310e-01
1.85273245e-01 2.05410212e-01 -7.24611223e-01 5.12026370e-01
-5.43768585e-01 -1.30569434e+00 6.29152715e-01 5.80763280e-01
1.02504432e+00 -1.20876980e+00 -1.14680099e+00 1.97225809e-02
-2.88293213e-01 -1.24103332e+00 7.88228437e-02 4.70282018e-01
-1.20306993e+00 -1.50576019e+00 -5.29986382e-01 -4.53585476e-01
3.58110011e-01 1.86868533e-01 1.62681413e+00 -2.70192362e-02
-2.31571451e-01 2.52866805e-01 -3.94729137e-01 -5.63972890e-01
-1.97051525e-01 1.17929123e-01 -2.93548912e-01 -3.98551881e-01
5.95527291e-01 -2.98587829e-01 -8.00417423e-01 3.69967759e-01
-1.01265359e+00 1.42021487e-02 1.75231516e-01 2.88014084e-01
5.34875810e-01 -4.44909781e-01 -2.19634667e-01 -9.09245849e-01
-1.70203969e-01 -4.81203645e-01 -1.14670157e+00 1.26167938e-01
-5.07953107e-01 -3.58706564e-02 1.92727268e-01 4.18506265e-01
-1.09208596e+00 6.96686506e-01 -8.12810808e-02 -1.33312106e-01
-6.81626856e-01 3.90819281e-01 -9.49614346e-02 -8.23465064e-02
6.84437394e-01 2.08028406e-01 2.07576193e-02 -4.61012721e-01
3.46591622e-01 7.85595238e-01 7.26262093e-01 -3.18180948e-01
8.26162040e-01 1.16710281e+00 -2.17944220e-01 -1.20852304e+00
-5.98840415e-01 -9.51579154e-01 -9.65138793e-01 -3.59328359e-01
1.24701381e+00 -1.45722640e+00 -5.37019968e-01 1.09489584e+00
-8.52091253e-01 -9.50359702e-01 3.02068710e-01 6.99316710e-02
-2.51854092e-01 2.75332242e-01 -3.41033697e-01 -8.34757328e-01
-5.40168583e-01 -1.00704300e+00 1.37223113e+00 5.35535932e-01
1.25546828e-01 -9.78798926e-01 4.44530308e-01 3.14739853e-01
3.31818432e-01 7.56467998e-01 7.74248764e-02 1.50651157e-01
-8.59920561e-01 1.52206168e-01 -2.61518478e-01 3.41892362e-01
7.28552267e-02 4.57581252e-01 -1.46661294e+00 -3.59510154e-01
-2.49404311e-01 -9.63787809e-02 1.57927191e+00 5.06803691e-01
5.57098627e-01 4.15124707e-02 -4.15886164e-01 1.10350239e+00
1.81672144e+00 -8.16474259e-02 9.72029567e-01 9.05093968e-01
7.29914784e-01 4.87661213e-01 4.16317105e-01 2.14015812e-01
6.32135630e-01 2.69614041e-01 1.13964927e+00 -4.41594422e-01
3.34762216e-01 3.95151138e-01 4.19363648e-01 3.28155458e-01
-2.15569377e-01 -8.05174410e-02 -1.18956184e+00 9.96593475e-01
-1.83365798e+00 -5.85370183e-01 -7.07491100e-01 2.13273048e+00
4.61833328e-01 -5.48878834e-02 -1.09608002e-01 -3.83452743e-01
3.84923369e-01 4.38289702e-01 -5.22202432e-01 -4.61495668e-01
-4.31638360e-01 1.67517245e-01 1.19728458e+00 6.43594444e-01
-1.67989182e+00 1.04854202e+00 6.76322603e+00 -2.08727583e-01
-1.14290297e+00 -3.12611870e-02 6.93504870e-01 1.58050865e-01
-4.96232174e-02 1.37025520e-01 -9.09930348e-01 3.36307585e-01
1.19704044e+00 5.47445297e-01 3.33345771e-01 6.24659657e-01
4.42806304e-01 -5.33194780e-01 -6.41890943e-01 3.98285270e-01
-2.43123338e-01 -1.31578255e+00 -4.72665399e-01 -2.30608825e-02
9.74718869e-01 1.10904527e+00 -3.94237190e-01 -6.48827851e-03
7.21531332e-01 -9.29966092e-01 6.91659033e-01 5.94227254e-01
6.84198081e-01 -3.71716052e-01 8.97844851e-01 1.47017092e-01
-1.64775944e+00 1.08206458e-01 -3.66819948e-01 -2.58885235e-01
-1.77153096e-01 3.87344301e-01 -4.82589513e-01 4.95208234e-01
1.72569346e+00 8.40447903e-01 -6.57399535e-01 9.92143214e-01
-1.94700897e-01 7.45544612e-01 -1.01180172e+00 2.62477815e-01
5.79935551e-01 -5.21316588e-01 2.13490739e-01 1.48344374e+00
1.30799398e-01 1.12684682e-01 -5.27694961e-03 6.75365806e-01
5.00937412e-03 -2.78601110e-01 -7.93966651e-01 4.10528809e-01
7.35889450e-02 1.29615533e+00 -7.54322350e-01 -4.08808410e-01
-4.66193765e-01 9.52294946e-01 -1.12842508e-01 3.91882747e-01
-6.30204618e-01 -1.64640561e-01 1.39579272e+00 -5.17116040e-02
4.48068589e-01 -4.64754075e-01 -3.32535923e-01 -1.22465682e+00
7.79902115e-02 -5.16017199e-01 2.78719157e-01 -1.01688945e+00
-1.05429840e+00 5.06195843e-01 3.51019539e-02 -1.17941821e+00
2.21152723e-01 -8.58811975e-01 -8.79961491e-01 1.00743127e+00
-2.17329884e+00 -1.26044989e+00 -9.67422366e-01 1.83436677e-01
4.26523179e-01 5.73578477e-01 6.96379304e-01 -3.28807496e-02
-2.64809489e-01 -1.23147532e-01 6.59646988e-01 2.26610214e-01
7.09290922e-01 -1.83357501e+00 9.63487387e-01 1.08005011e+00
1.25553086e-01 3.65376323e-02 7.61193871e-01 -5.21819592e-01
-1.21725881e+00 -1.62337649e+00 6.31523252e-01 -8.28033149e-01
5.41152298e-01 -4.03930694e-01 -1.15711546e+00 8.66810441e-01
3.59662592e-01 2.87203431e-01 1.83993056e-01 -7.52867907e-02
-2.19476759e-01 -2.70125151e-01 -1.00250375e+00 3.20191085e-01
7.48350680e-01 -4.63815600e-01 -4.28943574e-01 4.62431729e-01
5.57616770e-01 -6.93407893e-01 -6.67879164e-01 4.58501756e-01
6.75315499e-01 -1.40092432e+00 6.20799005e-01 -1.01658858e-01
4.52465236e-01 -7.32398212e-01 -2.28195161e-01 -1.39805806e+00
4.08558315e-03 -1.68681353e-01 4.78802204e-01 8.97637844e-01
5.03272593e-01 -5.35820782e-01 9.50157106e-01 5.30467808e-01
-3.61179262e-01 -1.12079889e-01 -9.62816596e-01 -6.51057780e-01
4.43383098e-01 -5.61666548e-01 5.52797794e-01 8.48639429e-01
-4.34306473e-01 -3.01670641e-01 -1.06779888e-01 9.58039522e-01
5.64105332e-01 4.36860710e-01 9.65575278e-01 -1.49110556e+00
1.05757020e-01 -2.10115492e-01 -2.21919939e-01 -8.59796941e-01
-1.54678538e-01 -6.32613063e-01 5.45401692e-01 -1.64288473e+00
-6.92045093e-02 -2.67811030e-01 8.15179572e-03 6.60123587e-01
-1.95201918e-01 6.18106842e-01 2.19582245e-01 5.26789784e-01
-3.48653406e-01 2.84621239e-01 9.20910060e-01 -2.03328341e-01
-2.47120515e-01 -2.54493922e-01 1.28545426e-02 8.91082585e-01
8.89231443e-01 -4.33675408e-01 2.56286621e-01 -6.09152615e-01
3.18494856e-01 -3.18391860e-01 5.02909303e-01 -1.23678887e+00
-9.34138298e-02 -5.77815883e-02 3.80876452e-01 -6.85021758e-01
-2.03263061e-03 -1.01565039e+00 3.68965179e-01 5.81968963e-01
2.31746063e-01 7.18533099e-02 6.02013946e-01 6.34126484e-01
-2.61235803e-01 2.61450484e-02 9.10155058e-01 -6.32773995e-01
-1.20120549e+00 9.35176611e-02 -4.82287645e-01 -1.28096938e-01
7.67707407e-01 -2.23720521e-01 -5.28373480e-01 -7.35016763e-02
-6.14936292e-01 8.41128469e-01 9.50398088e-01 3.77964944e-01
1.64486751e-01 -5.55767357e-01 -7.63545573e-01 3.89206827e-01
6.15152180e-01 4.90726531e-01 6.64416924e-02 5.84001303e-01
-1.42886543e+00 2.98847139e-01 4.86015119e-02 -1.18737161e+00
-1.30147469e+00 -1.63009718e-01 1.02541542e+00 -3.02638207e-02
-9.43537772e-01 7.61533856e-01 1.95629433e-01 -8.63173366e-01
-2.86242098e-01 -1.03719270e+00 -6.72823042e-02 4.77681607e-02
3.45562935e-01 -1.20173655e-01 2.43488342e-01 -6.00645721e-01
-4.75855231e-01 8.71327281e-01 4.46110338e-01 9.81974080e-02
1.33109391e+00 -4.45812196e-01 -1.84306979e-01 6.44578278e-01
8.37911665e-01 -6.44252360e-01 -1.91928971e+00 -7.75303133e-03
1.03505939e-01 -4.41418648e-01 -1.62260626e-02 -8.17933798e-01
-1.48995543e+00 9.67286646e-01 1.23624444e+00 1.47221938e-01
1.04582131e+00 -7.33100548e-02 6.00425184e-01 5.05420744e-01
2.93399453e-01 -8.39611888e-01 -7.08335280e-01 6.46529794e-01
4.98299003e-01 -1.87876129e+00 -9.12644789e-02 -1.34995535e-01
-5.97984254e-01 1.14280951e+00 6.71862245e-01 -1.57525778e-01
7.61695981e-01 3.83159965e-01 8.29603136e-01 -3.87628078e-01
-4.48224753e-01 -5.04336059e-01 -1.52021870e-01 6.33381724e-01
2.53668845e-01 3.90689045e-01 2.86315829e-01 -9.51632708e-02
-5.52731901e-02 4.39245487e-03 6.64739966e-01 9.08453584e-01
-7.21717000e-01 -6.41884804e-01 -7.21955895e-01 3.40947509e-01
-3.69449630e-02 -2.98481703e-01 -3.20794195e-01 8.57231855e-01
2.38911927e-01 9.84288156e-01 5.80654323e-01 -3.42688449e-02
1.74782932e-01 1.88081667e-01 8.35661739e-02 -2.73211241e-01
-7.20927954e-01 -1.34488404e-01 -2.58195656e-03 -6.64144278e-01
-9.42868114e-01 -8.09080422e-01 -1.02346087e+00 -3.72910470e-01
-1.79954678e-01 4.63111401e-02 8.52480173e-01 1.03094006e+00
2.03079984e-01 1.73164174e-01 5.32656789e-01 -1.29946411e+00
2.25043803e-01 -1.09811747e+00 -8.71128023e-01 4.05503541e-01
5.65908015e-01 -2.34068796e-01 -7.58922696e-01 5.48517048e-01] | [9.50558853149414, -1.5695388317108154] |
280b5082-76e9-4289-8bce-da129655b743 | sliding-mode-theory-under-feedback | 2109.05464 | null | https://arxiv.org/abs/2109.05464v1 | https://arxiv.org/pdf/2109.05464v1.pdf | Sliding-mode theory under feedback constraints and the problem of epidemic control | One of the most important branches of nonlinear control theory is the so-called sliding-mode. Its aim is the design of a (nonlinear) feedback law that brings and maintains the state trajectory of a dynamic system on a given sliding surface. Here, dynamics becomes completely independent of the model parameters and can be tuned accordingly to the desired target. In this paper we study this problem when the feedback law is subject to strong structural constraints. In particular, we assume that the control input may take values only over two bounded and disjoint sets. Such sets could be also non perfectly known a priori. An example is a control input allowed to switch only between two values. Under these peculiarities, we derive the necessary and sufficient conditions that guarantee sliding-mode control effectiveness for a class of time-varying continuous-time linear systems that includes all the stationary state-space linear models. Our analysis covers several scientific fields. It is only apparently confined to the linear setting and allows also to study an important set of nonlinear models. We describe fundamental examples related to epidemiology where the control input is the level of contact rate among people and the sliding surface permits to control the number of infected. For popular epidemiological models we prove the global convergence of control schemes based on the introduction of severe restrictions, like lockdowns, to contain epidemic. This greatly generalizes previous results obtained in the literature by casting them within a general sliding-mode theory. | ['Gianluigi Pillonetto', 'Mauro Bisiacco'] | 2021-09-12 | null | null | null | null | ['epidemiology'] | ['medical'] | [ 4.33258384e-01 3.31531167e-01 -4.52398062e-01 3.62723291e-01
1.83291420e-01 -6.43524826e-01 5.82920730e-01 3.12384337e-01
-3.38117152e-01 9.04592037e-01 -4.91905332e-01 -3.73123795e-01
-6.05592430e-01 -7.35786915e-01 -8.51210058e-01 -1.02895510e+00
-3.89966816e-01 3.79094809e-01 5.99713504e-01 -8.36227953e-01
-3.08470824e-03 5.49922109e-01 -1.12872326e+00 -5.13124943e-01
8.40830803e-01 6.47404850e-01 1.29443020e-01 7.27276266e-01
3.31494182e-01 7.76848346e-02 -5.64804792e-01 3.45667213e-01
3.30042779e-01 -4.87602890e-01 -4.51186329e-01 3.82628709e-01
-2.79987365e-01 7.89165422e-02 -2.88986284e-02 1.09595942e+00
1.82456374e-01 1.14891969e-01 6.48922026e-01 -1.26778567e+00
-3.59500080e-01 8.39483514e-02 -9.18089226e-02 2.09209234e-01
-4.36537229e-02 3.51588279e-02 3.55757922e-01 -4.72995713e-02
7.87376106e-01 8.28728080e-01 4.96236295e-01 7.56282866e-01
-1.28139746e+00 -2.25329474e-01 1.78345665e-01 -1.57507397e-02
-1.27114487e+00 -1.74808905e-01 3.39297920e-01 -7.17948914e-01
2.49474183e-01 6.16755068e-01 7.05449522e-01 6.13934994e-01
5.83142161e-01 9.69543755e-02 1.01171398e+00 -4.41999942e-01
4.09971625e-01 3.57267529e-01 1.87619954e-01 5.37864745e-01
6.86323225e-01 9.31094363e-02 4.26558167e-01 -2.83756226e-01
1.10565126e+00 2.09717959e-01 -5.82176626e-01 -4.93333131e-01
-9.99502242e-01 8.09285581e-01 2.25491777e-01 8.24299753e-01
-4.04086590e-01 -4.16186363e-01 -1.96021888e-03 7.69032538e-01
3.08928519e-01 4.79762495e-01 -3.23829502e-01 3.94877255e-01
-4.03695107e-01 2.19280094e-01 1.03856587e+00 6.70906782e-01
2.88375676e-01 -3.24193649e-02 -5.60289286e-02 2.73771703e-01
-1.00215346e-01 7.99353063e-01 -7.12178648e-02 -6.00727677e-01
1.68723017e-01 5.83476305e-01 6.20213568e-01 -9.13582683e-01
-4.76593882e-01 -5.48880756e-01 -9.98738587e-01 2.05927029e-01
7.56761551e-01 -3.88783157e-01 -6.65322185e-01 1.86699843e+00
5.24005949e-01 9.52521861e-02 -1.31913155e-01 7.54053354e-01
-1.69471160e-01 7.73714721e-01 -3.61323714e-01 -1.06584370e+00
1.00937355e+00 -1.51599288e-01 -9.89027858e-01 5.84712438e-02
3.65736753e-01 -3.70492786e-01 6.38448596e-01 2.94101357e-01
-1.01331365e+00 1.74982414e-01 -1.01444304e+00 8.30904424e-01
-2.97049522e-01 1.19617492e-01 -4.69994247e-01 2.41144106e-01
-1.01656353e+00 6.96002424e-01 -8.50602508e-01 -7.29718447e-01
-3.56348187e-01 5.27256072e-01 -1.68340847e-01 4.05301094e-01
-1.46175337e+00 9.77464080e-01 1.18955977e-01 4.00584579e-01
-3.89703184e-01 -4.02495474e-01 -3.76075745e-01 -1.72808051e-01
7.05707073e-01 -5.97734392e-01 9.06762123e-01 -1.09491420e+00
-1.49980330e+00 5.21692455e-01 -7.05561042e-02 -3.74695778e-01
1.06468427e+00 2.16315478e-01 -2.69681245e-01 1.37626812e-01
-1.80051122e-02 -4.66406703e-01 9.54543293e-01 -1.02144372e+00
-1.35573313e-01 -4.30211067e-01 1.15487598e-01 -4.75451574e-02
-6.27278268e-01 -8.84739459e-02 -2.09120780e-01 -3.89031321e-01
-2.62447774e-01 -1.25749910e+00 -5.39276719e-01 2.79792696e-02
-4.48765993e-01 -1.31063566e-01 9.41525340e-01 -3.08743447e-01
1.36934435e+00 -1.84901536e+00 5.75411379e-01 3.01262498e-01
-6.82300478e-02 5.30666053e-01 2.80714005e-01 9.09516215e-01
1.13459937e-01 4.12069112e-02 -6.83040321e-01 2.39863947e-01
-4.41740185e-01 1.05520219e-01 -3.08378458e-01 1.13001084e+00
2.50813216e-01 3.62106800e-01 -6.86174512e-01 -2.32710227e-01
1.62093624e-01 4.57901388e-01 -4.41540033e-01 1.54873997e-01
-1.48233488e-01 7.57229745e-01 -8.57116699e-01 -2.76950803e-02
5.15696883e-01 -2.79341042e-01 2.82863349e-01 4.79807794e-01
-6.93219185e-01 -3.59442264e-01 -1.30767202e+00 5.16516864e-01
-3.55865419e-01 3.32915336e-01 7.62758732e-01 -1.33179533e+00
6.64429009e-01 5.69441617e-01 5.18383920e-01 5.40643446e-02
5.92081070e-01 2.09540382e-01 9.01824385e-02 -6.58208549e-01
-1.26705348e-01 -5.89119017e-01 7.93502778e-02 -1.10625908e-01
-5.72587550e-01 1.04013152e-01 1.41249448e-01 -1.64658904e-01
1.01431060e+00 -8.09283078e-01 6.34971619e-01 -8.21557939e-01
9.19410586e-01 4.91553359e-02 5.19764125e-01 4.10265654e-01
-5.73298372e-02 -8.12048540e-02 8.48633230e-01 2.06349716e-01
-1.06500447e+00 -7.51504183e-01 -5.87112188e-01 5.98695695e-01
4.44680750e-01 2.20566884e-01 -6.08317375e-01 -5.49076265e-03
2.10183561e-01 1.73611015e-01 -1.13475418e+00 -3.79407346e-01
-8.22851419e-01 -6.60778761e-01 7.10375048e-03 -9.54567045e-02
2.30730727e-01 -6.47407413e-01 -6.26644492e-01 3.30738962e-01
3.54741454e-01 -7.49068677e-01 -4.02052075e-01 -2.63158027e-02
-9.52795327e-01 -1.45240951e+00 -1.02861357e+00 -7.21126676e-01
9.92453456e-01 -1.50615245e-01 3.64983380e-01 1.05360955e-01
-2.13076305e-02 3.80893379e-01 -7.21787363e-02 -3.90779287e-01
-8.33778620e-01 6.32635504e-02 2.46883735e-01 3.39536458e-01
-4.04216379e-01 -3.89891192e-02 -3.39528620e-01 7.78653204e-01
-1.17590952e+00 -3.08431208e-01 2.27887869e-01 7.68898129e-01
4.35187489e-01 1.82595938e-01 8.54352534e-01 -7.58618414e-01
6.41180873e-01 -6.72961950e-01 -9.42131460e-01 2.66503066e-01
-4.35991168e-01 1.03048891e-01 1.00861371e+00 -8.51835966e-01
-6.27847493e-01 1.22948818e-01 2.41313294e-01 -2.18796998e-01
1.85484007e-01 2.02036336e-01 -8.43132436e-02 -1.36982098e-01
4.31014001e-01 4.23819155e-01 4.91591752e-01 -4.17074651e-01
2.25900225e-02 6.54587328e-01 2.54661947e-01 -2.01744020e-01
8.74308705e-01 8.13095510e-01 5.00375688e-01 -1.51702762e+00
-8.84915888e-02 -3.75883818e-01 -6.85785115e-01 -4.19107050e-01
5.94923258e-01 -1.81988344e-01 -1.05950499e+00 4.45066243e-01
-9.58487153e-01 -5.61827302e-01 -3.26428860e-01 3.20941746e-01
-7.16583908e-01 -1.24207456e-02 -6.24907196e-01 -1.30321276e+00
4.54620719e-02 -7.17226386e-01 4.70609844e-01 3.45653445e-02
9.53397602e-02 -1.35367858e+00 4.17291462e-01 -5.17960250e-01
6.11416042e-01 6.66905463e-01 6.11047864e-01 -3.38040411e-01
-2.31978625e-01 -4.81680006e-01 4.47853774e-01 2.18696892e-01
1.84685215e-01 -1.98800955e-02 -8.07855129e-02 -5.40687323e-01
5.00469446e-01 4.25540388e-01 6.65080965e-01 7.85935879e-01
2.93400675e-01 -8.29577088e-01 -8.05556297e-01 1.63549140e-01
1.54741836e+00 5.39993167e-01 6.42897785e-02 2.14463532e-01
3.72529835e-01 1.08069861e+00 4.55931216e-01 3.67420346e-01
-9.72971693e-02 9.27921474e-01 4.94342417e-01 -2.33653039e-02
8.81741583e-01 2.64422774e-01 2.20703214e-01 5.86962223e-01
-2.42838249e-01 -2.22321689e-01 -7.09560037e-01 7.38161862e-01
-1.79186475e+00 -9.30617094e-01 -4.32131529e-01 2.73199773e+00
9.02338803e-01 1.25324026e-01 5.27472377e-01 1.79496601e-01
1.33044517e+00 -2.05112010e-01 -6.26901448e-01 -3.23136032e-01
-5.61265461e-02 -3.20410818e-01 9.82172906e-01 8.73231590e-01
-9.89482582e-01 2.35535845e-01 6.02127981e+00 2.67864525e-01
-1.46022999e+00 1.10319801e-01 1.39586508e-01 -3.45035419e-02
9.83064398e-02 -2.49900952e-01 -7.17451930e-01 5.32095253e-01
9.97029483e-01 -5.88166475e-01 1.48996204e-01 5.08576870e-01
6.95664942e-01 1.08737044e-01 -5.72116494e-01 1.06163576e-01
-4.37914521e-01 -1.02268326e+00 -5.05528450e-01 4.23494339e-01
8.08533669e-01 -4.74116474e-01 8.80883858e-02 -1.73894450e-01
-3.12365517e-02 -6.21756673e-01 2.50382155e-01 5.02098739e-01
8.36714745e-01 -5.91491401e-01 5.55886149e-01 8.37365806e-01
-1.19332778e+00 -2.76828051e-01 -8.07310194e-02 -2.68887490e-01
4.69742775e-01 4.38107729e-01 -5.56599498e-01 3.05263728e-01
5.38812540e-02 5.76857746e-01 1.25681236e-01 1.11888671e+00
2.71888226e-01 4.99632180e-01 -6.02441788e-01 -4.17893320e-01
2.21897602e-01 -3.14216286e-01 9.27165389e-01 9.43224311e-01
1.72955379e-01 3.34448755e-01 3.45226862e-02 6.04007661e-01
3.93679082e-01 2.41387919e-01 -9.77032721e-01 2.24954173e-01
2.03835309e-01 6.91931367e-01 -8.53770375e-01 -2.89394647e-01
-2.80095786e-02 6.89175010e-01 -2.02583328e-01 2.24359944e-01
-5.75419664e-01 -5.61296165e-01 6.98812366e-01 8.97862494e-01
2.63270497e-01 -3.15041721e-01 -9.42996442e-02 -1.09769344e+00
9.52705741e-02 -9.61788446e-02 2.75201023e-01 2.88355649e-01
-1.00651491e+00 4.70910609e-01 4.13613200e-01 -1.42536521e+00
-3.47232342e-01 -3.45671803e-01 -7.54511952e-01 6.25854254e-01
-1.06306541e+00 -4.52063978e-01 3.07490945e-01 7.72630095e-01
2.38424376e-01 3.09643269e-01 2.74662703e-01 2.36475199e-01
-7.09832847e-01 -9.63316823e-04 7.86947846e-01 -2.25778520e-01
4.31385189e-01 -1.21959651e+00 -2.63400614e-01 8.80177796e-01
-9.26969945e-01 8.71716976e-01 1.41752243e+00 -8.29929113e-01
-1.25169730e+00 -1.09483457e+00 9.25154984e-01 -1.28227562e-01
1.04240787e+00 -4.28497583e-01 -1.31965339e+00 6.17067873e-01
2.70195771e-02 7.61473998e-02 -2.78495193e-01 -5.31297088e-01
5.84739089e-01 1.73790008e-02 -1.03946865e+00 4.35187429e-01
6.09371662e-01 6.56127408e-02 -3.19685638e-01 3.65584284e-01
5.36543071e-01 -1.83141768e-01 -7.94770062e-01 3.32816005e-01
3.29874516e-01 -3.52685571e-01 8.15070033e-01 -8.90339255e-01
-8.72078389e-02 -3.29843432e-01 4.54555452e-01 -1.30207479e+00
-1.46653056e-01 -1.04380655e+00 8.05345103e-02 8.02611887e-01
6.24959171e-02 -1.08698976e+00 3.26192111e-01 2.07483411e-01
2.13823676e-01 -1.10171592e+00 -1.26898384e+00 -1.21009111e+00
3.60949725e-01 2.90759087e-01 -1.13069832e-01 9.92016017e-01
4.02387291e-01 1.10799268e-01 -5.15471756e-01 4.37707782e-01
6.00643456e-01 -2.50925571e-01 2.96394616e-01 -1.38901126e+00
-1.95215434e-01 -4.87812698e-01 -2.91543871e-01 -7.40860879e-01
1.85787585e-02 -3.37942660e-01 1.78666800e-01 -1.41146028e+00
-3.82940650e-01 -4.49070811e-01 -3.43971550e-02 1.43755489e-04
-3.41495611e-02 -1.53251082e-01 4.15823571e-02 4.94112581e-01
-3.84990796e-02 3.95382464e-01 1.44303524e+00 2.56061554e-01
-7.93907166e-01 8.41768265e-01 -2.05045551e-01 5.38161099e-01
8.86822581e-01 -3.19361150e-01 -5.16885698e-01 2.58916020e-01
1.86287791e-01 6.73968196e-01 2.65534937e-01 -5.84525943e-01
2.73176283e-01 -7.06194401e-01 -5.32288134e-01 -6.35136589e-02
8.42000842e-02 -8.91793609e-01 2.61209399e-01 1.21316314e+00
-3.52244318e-01 -5.83944060e-02 -2.95523927e-02 8.59424531e-01
1.87721580e-01 -1.39841080e-01 1.30496907e+00 2.64249057e-01
-3.16381827e-02 1.06536180e-01 -7.06259727e-01 -4.70853373e-02
1.48005617e+00 -4.06538658e-02 -3.37590396e-01 -4.14414257e-01
-1.08446383e+00 2.99872428e-01 4.49757874e-01 1.74272865e-01
1.52088627e-01 -9.72402573e-01 -5.34688294e-01 2.16715887e-01
-1.66427687e-01 -4.09946144e-01 2.89496988e-01 1.34905124e+00
-2.98351526e-01 7.79263616e-01 -2.52341837e-01 -4.73792911e-01
-1.22716427e+00 7.78179109e-01 7.33657420e-01 -1.04029598e-02
-4.72275585e-01 1.24832019e-01 1.28295735e-01 5.54620959e-02
-1.86511591e-01 -4.46848750e-01 -4.07135308e-01 -6.64280653e-02
4.30537194e-01 5.16084135e-01 -1.75398141e-01 -7.57921040e-01
-4.71884400e-01 7.06628621e-01 4.09552872e-01 -1.31310254e-01
1.32584441e+00 -2.97174037e-01 -1.22259595e-01 7.43947148e-01
1.23232949e+00 -1.76550411e-02 -1.35141587e+00 -1.40705511e-01
-1.93573639e-01 4.14266512e-02 -3.67373705e-01 -2.13868007e-01
-8.34139526e-01 4.80226159e-01 3.74646604e-01 1.05390477e+00
9.91869748e-01 1.27953306e-01 3.76970977e-01 1.58903748e-01
2.50645608e-01 -8.25746477e-01 -3.80089223e-01 2.74286568e-01
1.03920305e+00 -7.80520558e-01 -3.07041585e-01 -6.02023363e-01
-2.31988609e-01 1.16344178e+00 3.05803120e-01 -7.28452563e-01
1.00770843e+00 3.71392906e-01 -3.34005713e-01 2.22958252e-01
-8.13986063e-01 -2.18071327e-01 2.01135203e-01 3.56955200e-01
1.42295510e-01 1.59300789e-01 -1.17712569e+00 1.17034458e-01
3.16870719e-01 2.27358431e-01 8.08283925e-01 8.45802605e-01
-7.46060133e-01 -9.85045552e-01 -8.08083892e-01 3.12253177e-01
-5.38859010e-01 5.63246608e-01 -2.16359779e-01 1.14435303e+00
-1.67161003e-01 8.59829307e-01 9.58961099e-02 6.47602156e-02
6.72989726e-01 -2.34100282e-01 2.52473384e-01 -4.59909678e-01
-4.31789517e-01 2.46064495e-02 -3.44313085e-01 -2.34295696e-01
-3.44644845e-01 -6.72448158e-01 -1.12273407e+00 -3.98785532e-01
-6.58651769e-01 3.63783449e-01 4.31509495e-01 1.04351759e+00
-2.94024516e-02 5.31663954e-01 7.48884678e-01 -5.27503788e-01
-8.45831990e-01 -8.07095945e-01 -8.65764439e-01 -2.04008780e-02
1.04442799e+00 -7.30992794e-01 -6.88688099e-01 8.09656382e-02] | [5.519758224487305, 2.8589084148406982] |
fab2acca-c398-4e96-b266-15fec57fd581 | llql-logistic-likelihood-q-learning-for | 2307.02345 | null | https://arxiv.org/abs/2307.02345v1 | https://arxiv.org/pdf/2307.02345v1.pdf | LLQL: Logistic Likelihood Q-Learning for Reinforcement Learning | Currently, research on Reinforcement learning (RL) can be broadly classified into two categories: online RL and offline RL. Both in online and offline RL, the primary focus of research on the Bellman error lies in the optimization techniques and performance improvement, rather than exploring the inherent structural properties of the Bellman error, such as distribution characteristics. In this study, we analyze the distribution of the Bellman approximation error in both online and offline settings. We find that in the online environment, the Bellman error follows a Logistic distribution, while in the offline environment, the Bellman error follows a constrained Logistic distribution, where the constrained distribution is dependent on the prior policy in the offline data set. Based on this finding, we have improved the MSELoss which is based on the assumption that the Bellman errors follow a normal distribution, and we utilized the Logistic maximum likelihood function to construct $\rm LLoss$ as an alternative loss function. In addition, we observed that the rewards in the offline data set should follow a specific distribution, which would facilitate the achievement of offline objectives. In our numerical experiments, we performed controlled variable corrections on the loss functions of two variants of Soft-Actor-Critic in both online and offline environments. The results confirmed our hypothesis regarding the online and offline settings, we also found that the variance of LLoss is smaller than MSELoss. Our research provides valuable insights for further investigations based on the distribution of Bellman errors. | ['Yu Guang Wang', 'Bingxin Zhou', 'Outongyi Lv'] | 2023-07-05 | null | null | null | null | ['q-learning', 'reinforcement-learning-1', 'offline-rl'] | ['methodology', 'methodology', 'playing-games'] | [-4.17016238e-01 1.10812038e-01 -4.51045960e-01 -2.12497532e-01
-6.98400378e-01 -5.29862940e-01 3.11909206e-02 1.45443141e-01
-8.19191158e-01 1.05706835e+00 -2.54079372e-01 -5.17024338e-01
-4.47665453e-01 -4.54923958e-01 -7.49541163e-01 -8.36500168e-01
-3.51373255e-01 9.93818194e-02 -1.53798848e-01 -6.56233802e-02
4.16227460e-01 3.67736757e-01 -1.14318895e+00 -7.01143622e-01
8.47973228e-01 1.31387901e+00 1.64581910e-01 5.79508364e-01
1.56590194e-01 7.03727663e-01 -9.11244333e-01 -2.67853349e-01
6.50290847e-01 -5.30319273e-01 -3.70673418e-01 -1.87849060e-01
-3.23763788e-01 -6.56038821e-01 -5.25760412e-01 1.11828589e+00
5.70076764e-01 5.76013327e-01 4.12151575e-01 -1.53196752e+00
-3.62174779e-01 5.59620976e-01 -5.32270551e-01 2.19456345e-01
1.30865157e-01 2.09599271e-01 1.08304894e+00 -2.77709603e-01
1.22309111e-01 1.06470776e+00 5.87937772e-01 2.77162969e-01
-8.44061732e-01 -7.82344282e-01 5.13924360e-01 1.05914252e-03
-1.39679372e+00 -2.93528497e-01 5.21571815e-01 -1.39034554e-01
6.06773198e-01 -1.77526981e-01 6.67732298e-01 7.18756914e-01
5.92221200e-01 9.17136788e-01 1.12224448e+00 -5.40442646e-01
4.82879221e-01 3.79979402e-01 -1.51908591e-01 6.85629308e-01
2.27897510e-01 6.48045182e-01 -4.46989723e-02 -1.97847039e-02
1.01358485e+00 -1.71331570e-01 -1.09855570e-01 -2.59226948e-01
-5.57016969e-01 1.10805583e+00 1.77954912e-01 8.42464790e-02
-3.56043696e-01 4.52045679e-01 3.16049665e-01 5.19919574e-01
2.31942639e-01 4.50250566e-01 -2.97432005e-01 -4.70224112e-01
-4.52963889e-01 2.03027353e-01 8.64679813e-01 8.78611386e-01
4.02477920e-01 2.15087205e-01 -2.11315036e-01 5.80806434e-01
4.64772731e-01 6.49270236e-01 3.19492936e-01 -1.24174678e+00
9.14064288e-01 2.70998832e-02 6.21604443e-01 -9.07020092e-01
-3.46054494e-01 -5.52355170e-01 -2.05804810e-01 3.56234938e-01
1.04238153e+00 -6.65190697e-01 -3.64223450e-01 1.88931882e+00
1.28712967e-01 -3.57011795e-01 1.60652533e-01 1.03130162e+00
-2.69430041e-01 5.43089688e-01 -2.64115930e-01 -4.78854150e-01
7.79259026e-01 -8.35286736e-01 -9.99158561e-01 -3.52895670e-02
4.51222390e-01 -5.78419745e-01 1.28540599e+00 3.91412556e-01
-1.20297658e+00 -2.41143003e-01 -1.10107005e+00 4.38643515e-01
2.22079754e-01 6.33667931e-02 5.09871721e-01 6.81623280e-01
-7.16180265e-01 8.22834730e-01 -8.85810614e-01 -8.18591788e-02
9.90781784e-02 2.64043719e-01 3.02563131e-01 2.17873484e-01
-1.11951268e+00 9.49905932e-01 2.36264199e-01 1.90501645e-01
-1.04008186e+00 -4.15016711e-01 -5.11888087e-01 1.99494243e-01
8.09115171e-01 -1.22736543e-01 1.54385900e+00 -1.11652982e+00
-2.07392931e+00 7.66158700e-02 2.34138772e-01 -6.30876422e-01
7.02775538e-01 -4.29632235e-03 -3.36517803e-02 1.30867317e-01
-2.62077183e-01 2.19779648e-02 7.33594775e-01 -1.12007952e+00
-6.81352258e-01 -2.16142997e-01 4.26090926e-01 3.61564785e-01
1.66007709e-02 -1.85921088e-01 1.05942301e-02 -5.21864772e-01
-3.66176277e-01 -8.72900069e-01 -1.49076819e-01 -9.89161059e-02
-2.21007243e-02 -1.98128253e-01 3.80176514e-01 -5.20521939e-01
1.32603335e+00 -2.28878164e+00 -3.80800575e-01 4.75336313e-01
-1.94010898e-01 5.71395867e-02 3.05082090e-02 5.88253021e-01
2.81508774e-01 7.28425905e-02 1.70890927e-01 -3.63496505e-02
3.16256016e-01 2.51320571e-01 -5.00238061e-01 7.94269443e-01
-2.28571683e-01 4.35006857e-01 -9.81077611e-01 -2.90152550e-01
-1.20967023e-01 -6.89424500e-02 -5.36551297e-01 3.66333485e-01
-5.70967384e-02 2.05374375e-01 -5.38754404e-01 3.10039341e-01
2.30625808e-01 1.49512244e-02 2.80099750e-01 3.05528641e-01
-3.45252275e-01 9.49256122e-02 -1.30713475e+00 1.04581439e+00
-7.41138875e-01 6.01202309e-01 2.72719711e-01 -1.11961508e+00
1.12001157e+00 2.14570150e-01 6.13004684e-01 -7.85411358e-01
4.14023012e-01 1.11788623e-01 1.50688678e-01 -2.49162525e-01
4.72062618e-01 -2.79462785e-01 1.06926255e-01 5.78511178e-01
-1.79842338e-01 1.92946434e-01 1.13953613e-01 -1.45007327e-01
7.80997932e-01 1.79730296e-01 2.07334563e-01 -1.94561854e-01
-2.88365241e-02 -2.09292397e-01 4.43346113e-01 9.99809921e-01
-5.79856277e-01 -4.13154662e-02 8.71371269e-01 1.56377122e-01
-8.31683278e-01 -1.08645988e+00 -1.39023691e-01 8.77092600e-01
3.43369305e-01 -1.10984974e-01 -4.79442984e-01 -7.90914774e-01
2.74105877e-01 8.90088022e-01 -3.22833687e-01 -2.02011272e-01
-2.06036031e-01 -4.79814947e-01 6.52729750e-01 5.35640895e-01
3.56629640e-01 -7.02249527e-01 -7.90604174e-01 2.78402984e-01
1.47181243e-01 -8.92462075e-01 -8.90371382e-01 2.94852018e-01
-6.71033323e-01 -1.16803634e+00 -5.66441417e-01 -2.66456723e-01
5.76389909e-01 4.64218520e-02 5.21082759e-01 -2.89857805e-01
1.59607064e-02 8.25639427e-01 -3.31940711e-01 -6.42065406e-01
-2.03499004e-01 -2.73230821e-01 2.14960471e-01 -6.32688999e-02
-2.01228142e-01 -2.09321693e-01 -6.84249818e-01 4.30100918e-01
-5.47636986e-01 -7.10258543e-01 4.42028493e-01 8.28117907e-01
4.63496357e-01 5.15750408e-01 8.44291449e-01 -3.24954748e-01
9.77519512e-01 -4.65483516e-01 -1.19165862e+00 3.19583714e-01
-1.07861269e+00 3.06705356e-01 1.09892285e+00 -7.18281686e-01
-9.99316216e-01 -3.93724084e-01 6.73442334e-02 -5.74722171e-01
3.66544962e-01 4.14968133e-01 5.12051471e-02 -6.12585396e-02
1.07381962e-01 -6.94744568e-03 4.74324346e-01 -1.15213186e-01
1.29012227e-01 6.78238630e-01 -4.47208285e-02 -9.17931378e-01
3.20358604e-01 1.66788340e-01 7.03442022e-02 -7.21588969e-01
-8.69784176e-01 -1.39666557e-01 1.77933767e-01 -3.51050794e-01
5.49636364e-01 -5.89880586e-01 -1.45315051e+00 2.31166810e-01
-6.22274220e-01 -6.66801989e-01 -6.87135398e-01 8.85142386e-01
-9.45539594e-01 3.89221132e-01 -6.17744923e-01 -1.66400123e+00
-2.96038538e-02 -1.05535138e+00 3.83790970e-01 4.32754636e-01
9.00277793e-02 -1.31696105e+00 5.85161336e-02 -7.00763837e-02
3.74137759e-01 1.42537449e-02 7.00833023e-01 -5.95325589e-01
-1.95552543e-01 -2.18127340e-01 -1.62625052e-02 4.93957102e-01
-8.13673064e-02 -8.40747263e-03 -3.69698405e-01 -6.85237050e-01
2.10447565e-01 -5.54234147e-01 2.68809259e-01 5.02156675e-01
1.18377638e+00 -5.58356702e-01 2.04991221e-01 3.99159491e-01
1.51747620e+00 8.96359146e-01 4.09027934e-01 5.20325541e-01
6.56145290e-02 4.40997154e-01 1.16418147e+00 1.07330334e+00
3.89266849e-01 5.43284655e-01 4.86725479e-01 2.24779859e-01
6.65774226e-01 -6.69590831e-01 8.53868484e-01 5.08634806e-01
1.21780738e-01 -3.21582228e-01 -4.84159797e-01 1.93614170e-01
-1.96235096e+00 -9.64468300e-01 4.55826014e-01 2.63125300e+00
8.60640824e-01 2.00980037e-01 3.92837405e-01 -1.37907848e-01
6.43949747e-01 2.88349111e-03 -8.42226148e-01 -7.26172268e-01
1.91633984e-01 1.64274424e-01 1.08871901e+00 6.20348930e-01
-6.68552756e-01 5.32905936e-01 6.94922113e+00 9.19662774e-01
-1.11835277e+00 -1.15657374e-01 5.47889471e-01 -2.26440117e-01
6.68200140e-04 1.65259853e-01 -9.69442010e-01 7.23551214e-01
1.02210796e+00 -4.25684273e-01 7.18802094e-01 8.32487464e-01
6.19721830e-01 -4.18604493e-01 -8.57410908e-01 6.53975785e-01
-3.89627814e-01 -5.84791005e-01 -6.15498960e-01 2.59092867e-01
4.58971947e-01 -3.29440832e-01 1.24993078e-01 6.10599995e-01
5.18813252e-01 -8.73091519e-01 8.52262080e-01 5.88675439e-01
5.22181988e-01 -1.08618283e+00 8.05031896e-01 5.06366670e-01
-8.47933531e-01 -4.89209026e-01 -3.43019634e-01 -2.81334341e-01
-4.16267663e-02 3.68608922e-01 -7.15504348e-01 3.37536991e-01
2.88001597e-01 2.92244017e-01 1.01417713e-01 1.12044561e+00
-2.95240253e-01 6.42513096e-01 -3.44424248e-01 -2.67769605e-01
3.39466214e-01 -5.48108220e-01 3.41403425e-01 7.13691235e-01
1.63206741e-01 -4.51163389e-02 5.25815010e-01 8.63686383e-01
2.46688321e-01 1.08091086e-01 -2.35381737e-01 -3.71115118e-01
5.95485628e-01 8.49242866e-01 -3.56679916e-01 1.99415535e-01
-2.66375363e-01 3.69404227e-01 3.34420770e-01 5.20034254e-01
-1.22610426e+00 -7.63925195e-01 4.95193630e-01 -1.95137471e-01
3.21593940e-01 -4.54439163e-01 -7.32073188e-02 -8.33365500e-01
-1.75419182e-03 -6.28940105e-01 3.51599783e-01 -2.31704235e-01
-1.15821183e+00 -1.76965259e-02 2.15024561e-01 -8.80798578e-01
-2.73331732e-01 -4.29716885e-01 -6.07170701e-01 5.56898534e-01
-1.30422890e+00 -1.14664435e-01 3.54677796e-01 5.83915889e-01
2.68060893e-01 3.19442227e-02 2.91118383e-01 2.86529064e-01
-9.79618907e-01 9.91074741e-01 6.85343862e-01 6.24305122e-02
5.61698854e-01 -1.00300729e+00 -4.94702458e-01 4.90263224e-01
-2.11320490e-01 4.14072484e-01 7.40008891e-01 -4.39278603e-01
-1.45134854e+00 -6.54400766e-01 6.65587485e-02 2.98644751e-01
9.17937577e-01 1.15355672e-02 -4.42566931e-01 6.72424853e-01
-1.28080115e-01 -1.25729784e-01 4.00949240e-01 4.94194822e-03
2.12599471e-01 -2.93868244e-01 -1.31277251e+00 5.80059230e-01
6.78544343e-01 -2.31193647e-01 -6.05209097e-02 1.49729952e-01
5.23178756e-01 -4.25986558e-01 -1.03498888e+00 4.28304374e-02
7.59508252e-01 -7.61548340e-01 4.26304609e-01 -7.00671673e-01
9.75113213e-02 8.57372135e-02 -1.85758993e-01 -1.52932239e+00
6.56471327e-02 -7.09455073e-01 -2.93428868e-01 1.11422479e+00
3.60535890e-01 -1.07304037e+00 5.83885849e-01 7.10570633e-01
-7.37220794e-02 -1.13054860e+00 -1.08883941e+00 -1.29795396e+00
1.83954149e-01 -2.70303667e-01 1.18641831e-01 2.84033239e-01
9.93710593e-04 -1.95928514e-01 -2.47339532e-01 1.00524880e-01
5.22286117e-01 1.30573586e-01 5.96677780e-01 -3.53341073e-01
-8.14308107e-01 -4.65347022e-01 4.12699655e-02 -1.40562272e+00
5.08824527e-01 -4.07751143e-01 3.06298465e-01 -1.07907736e+00
-2.23200485e-01 -9.00839329e-01 -4.62580353e-01 2.42026761e-01
9.77278873e-02 -4.86730486e-01 4.99766171e-01 5.16643152e-02
-6.77958965e-01 9.94807720e-01 1.31006265e+00 2.34912530e-01
-4.88685608e-01 4.81810838e-01 -4.84006256e-01 6.97123587e-01
9.96430993e-01 -5.01044989e-01 -5.82527399e-01 -1.23558238e-01
1.04866073e-01 5.85502863e-01 -1.05335638e-01 -5.35963416e-01
9.53980349e-03 -6.74783409e-01 -1.47799104e-01 -1.86798424e-01
1.82561025e-01 -9.88449097e-01 -3.21912825e-01 4.46604222e-01
-4.92534757e-01 1.18533382e-02 5.15700988e-02 7.03145862e-01
1.72528252e-02 -4.94973242e-01 8.19468260e-01 7.88067505e-02
-6.55834600e-02 1.17790513e-01 -5.58773994e-01 4.51878786e-01
1.18687439e+00 -4.67403996e-04 -1.07330486e-01 -8.95364046e-01
-4.61453348e-01 6.11595809e-01 2.33673275e-01 -8.02965388e-02
5.23672581e-01 -1.00186110e+00 -1.37033328e-01 8.02164301e-02
-4.43595409e-01 -4.14885163e-01 -1.66591614e-01 1.19668794e+00
-2.40009233e-01 4.28881139e-01 5.93806580e-02 -1.67120680e-01
-6.45412266e-01 3.64227355e-01 6.23966277e-01 -4.75679278e-01
-2.84316957e-01 4.22212720e-01 -2.35529348e-01 -6.40819818e-02
8.37077618e-01 -3.83832425e-01 1.50170386e-01 -8.53243619e-02
3.01403224e-01 8.29206586e-01 -3.99261415e-01 -1.75835472e-02
-3.53798121e-02 2.11516961e-01 1.99314252e-01 -3.71510327e-01
1.00147939e+00 -4.79845673e-01 4.78553355e-01 6.91288710e-01
1.00255167e+00 1.66414380e-01 -1.80663013e+00 9.12860557e-02
-4.01608981e-02 -7.14924932e-01 9.16067362e-02 -7.05509126e-01
-1.10272086e+00 3.69137257e-01 5.61100721e-01 3.53413522e-01
9.91802037e-01 -4.26762313e-01 6.75792933e-01 2.34995350e-01
6.86594546e-01 -1.45830107e+00 1.29085064e-01 7.47895420e-01
6.21090293e-01 -1.07633388e+00 1.59277990e-02 1.81211218e-01
-9.15995419e-01 1.00990474e+00 5.29727042e-01 -4.42881823e-01
5.44904590e-01 1.75703213e-01 -1.03435628e-01 3.42585295e-01
-7.25485861e-01 -8.43921900e-02 -1.51050061e-01 2.59084255e-01
2.68875539e-01 8.75734165e-02 -6.17869735e-01 8.11320245e-01
-1.90835223e-01 -8.14092010e-02 5.94670475e-01 1.00382185e+00
-4.12676632e-01 -1.10024333e+00 -4.70167130e-01 2.10076556e-01
-5.90312481e-01 2.28259802e-01 -3.91120417e-03 1.04391897e+00
-3.52057457e-01 1.17928493e+00 1.77087009e-01 -1.29377827e-01
4.15556580e-01 -2.22135559e-01 7.47340441e-01 -1.93507761e-01
-4.24728692e-01 -1.11183643e-01 5.88025860e-02 -5.62949479e-01
-9.79995877e-02 -5.44361591e-01 -1.59986937e+00 -5.41733861e-01
-6.99390590e-01 3.97691280e-01 8.98487985e-01 1.12003624e+00
6.25792444e-02 3.17291051e-01 1.18978333e+00 -4.47196811e-01
-1.55798113e+00 -5.74495256e-01 -9.79275227e-01 -8.15111101e-02
2.84114182e-01 -8.24717343e-01 -7.53020823e-01 -7.41979659e-01] | [4.219833850860596, 2.55479097366333] |
f627518f-a27a-43dd-8abf-5e5af50e985b | aist-an-interpretable-attention-based-deep | 2012.08713 | null | https://arxiv.org/abs/2012.08713v2 | https://arxiv.org/pdf/2012.08713v2.pdf | AIST: An Interpretable Attention-based Deep Learning Model for Crime Prediction | Accuracy and interpretability are two essential properties for a crime prediction model. Because of the adverse effects that the crimes can have on human life, economy and safety, we need a model that can predict future occurrence of crime as accurately as possible so that early steps can be taken to avoid the crime. On the other hand, an interpretable model reveals the reason behind a model's prediction, ensures its transparency and allows us to plan the crime prevention steps accordingly. The key challenge in developing the model is to capture the non-linear spatial dependency and temporal patterns of a specific crime category while keeping the underlying structure of the model interpretable. In this paper, we develop AIST, an Attention-based Interpretable Spatio Temporal Network for crime prediction. AIST models the dynamic spatio-temporal correlations for a crime category based on past crime occurrences, external features (e.g., traffic flow and point of interest (POI) information) and recurring trends of crime. Extensive experiments show the superiority of our model in terms of both accuracy and interpretability using real datasets. | ['Tanzima Hashem', 'Yeasir Rayhan'] | 2020-12-16 | null | null | null | null | ['crime-prediction'] | ['miscellaneous'] | [ 4.25440297e-02 -8.97617415e-02 -2.86135137e-01 -5.69241881e-01
2.06554011e-01 -2.59655058e-01 5.05595565e-01 5.44572413e-01
-1.35840744e-01 5.09847343e-01 6.54394209e-01 -4.72752810e-01
-7.38887548e-01 -1.20007086e+00 -2.44660303e-01 -3.32657576e-01
-3.89132440e-01 2.94028699e-01 1.68974131e-01 -1.46808073e-01
4.18483764e-01 6.50482953e-01 -1.22094822e+00 3.80776674e-01
9.60742593e-01 6.09731972e-01 1.21047042e-01 2.80397058e-01
5.29600941e-02 1.22512722e+00 -2.87412196e-01 -4.47708637e-01
-1.48614466e-01 -1.38110712e-01 -9.36150193e-01 -1.00431472e-01
-3.45823228e-01 -2.97682345e-01 -5.74786484e-01 9.35980856e-01
-2.22315460e-01 1.12880670e-01 8.91586363e-01 -1.41013122e+00
-8.23933780e-01 6.07683241e-01 -6.37754261e-01 6.49582148e-01
3.30536425e-01 4.80712764e-02 8.59255254e-01 -5.18002212e-01
2.91261107e-01 1.33564782e+00 6.14373326e-01 2.23742828e-01
-1.20365810e+00 -4.63307112e-01 4.88622427e-01 5.28987706e-01
-1.37155151e+00 -2.56890535e-01 8.66402984e-01 -7.29034722e-01
9.60209131e-01 5.87433696e-01 7.11296916e-01 8.09756517e-01
3.36288333e-01 4.64684159e-01 8.69236767e-01 -2.53010869e-01
1.39235660e-01 -1.31538883e-01 4.69994694e-01 4.46680367e-01
2.38775998e-01 -1.58266295e-02 -3.70816022e-01 -7.14169145e-02
4.89659250e-01 6.65479660e-01 9.88553241e-02 4.80302095e-01
-4.98859465e-01 7.32117414e-01 6.36532009e-01 3.86037946e-01
-4.96378303e-01 1.26532346e-01 3.69094849e-01 -2.51037069e-02
9.71570671e-01 1.68595806e-01 -2.27987409e-01 -2.37113014e-01
-6.85374796e-01 1.57138914e-01 1.61084890e-01 3.71412665e-01
7.46006012e-01 -2.31122047e-01 -2.53445208e-01 7.01470554e-01
2.00244009e-01 4.52323139e-01 -2.22073317e-01 -5.82746983e-01
9.14572239e-01 1.07059181e+00 -1.84862688e-01 -1.84742785e+00
-3.98165315e-01 4.29532398e-03 -1.16884041e+00 -1.11722708e-01
2.63241917e-01 9.42687839e-02 -6.55602753e-01 1.61285567e+00
7.44959339e-02 4.78817791e-01 -4.36377376e-01 7.21280992e-01
2.28900746e-01 8.41778755e-01 5.10439515e-01 -1.71353042e-01
1.13121414e+00 -1.54604360e-01 -9.03926671e-01 -3.67019802e-01
7.06735373e-01 -2.02047437e-01 8.12693357e-01 -1.23629138e-01
-7.94112325e-01 -1.67303205e-01 -3.00565541e-01 6.21923953e-02
-5.03846645e-01 -1.83510527e-01 7.23427773e-01 3.28007966e-01
-6.12069845e-01 7.50488997e-01 -9.18135345e-01 -2.93741137e-01
7.39133000e-01 2.34132722e-01 -2.56693989e-01 -1.24336459e-01
-1.17200482e+00 9.31001842e-01 4.22348440e-01 3.20134729e-01
-1.08773053e-01 -6.34223878e-01 -7.32621431e-01 2.90948331e-01
6.88332245e-02 -1.24545738e-01 4.51280028e-01 -7.04278886e-01
-5.09218276e-01 4.97098356e-01 -6.04547620e-01 -2.63264149e-01
2.96653479e-01 5.26708402e-02 -5.75715840e-01 -3.66153792e-02
5.56149662e-01 -3.98040371e-04 2.54519552e-01 -8.18136752e-01
-1.03468740e+00 -7.54447997e-01 9.84788164e-02 -7.81075954e-02
-7.08127856e-01 3.27503383e-01 -2.98279196e-01 -5.35527587e-01
1.79510847e-01 -7.53673494e-01 -2.76615590e-01 -2.84036160e-01
-5.15127659e-01 -3.68539661e-01 1.16170609e+00 -1.02460825e+00
1.85692585e+00 -2.09045029e+00 -1.15351558e-01 5.43945909e-01
4.48929429e-01 2.75066406e-01 5.33240363e-02 5.27600765e-01
-1.98802248e-01 5.08613646e-01 -2.92341352e-01 -2.08849192e-01
-2.42992207e-01 4.25939173e-01 -4.63249773e-01 5.26923418e-01
5.82363367e-01 7.42244244e-01 -1.06876647e+00 -3.84258211e-01
4.50812161e-01 3.32193375e-01 -4.62776214e-01 1.06008947e-02
2.54002541e-01 4.15953219e-01 -9.02616143e-01 3.36771250e-01
5.51367700e-01 -1.97122529e-01 -3.73106939e-03 4.00770307e-01
-2.43425235e-01 4.12480265e-01 -9.72422957e-01 6.14101410e-01
-3.71387392e-01 9.15780008e-01 -3.94713521e-01 -1.09275615e+00
9.10764337e-01 2.97150493e-01 3.73791754e-01 -1.14023817e+00
-3.74164104e-01 -1.30777657e-01 -1.28932923e-01 -7.21463382e-01
3.25120926e-01 2.20137462e-01 2.40652394e-02 7.70021081e-01
-7.29529321e-01 5.32304406e-01 2.40935951e-01 1.44103333e-01
1.18636596e+00 -4.82807666e-01 3.48252594e-01 -2.97904640e-01
3.71850282e-01 4.55898568e-02 8.26737881e-01 6.16753757e-01
-3.91268320e-02 3.35100085e-01 7.47267067e-01 -1.14282227e+00
-1.02366590e+00 -7.17610598e-01 -2.05471307e-01 7.09982872e-01
6.33373670e-03 -3.59092742e-01 -6.73961043e-01 -5.01043439e-01
-1.80098772e-01 8.97261679e-01 -9.94333804e-01 -1.76479623e-01
-8.13239396e-01 -9.70055938e-01 2.99659580e-01 6.56848609e-01
5.96318543e-01 -9.75531638e-01 -6.87034905e-01 2.38344535e-01
-4.72371370e-01 -8.51998210e-01 -2.47642696e-01 -4.09685373e-01
-7.78255105e-01 -1.25572741e+00 1.30398065e-01 -2.92026669e-01
9.32188451e-01 4.14859682e-01 8.69474292e-01 7.04334080e-01
-4.23615985e-02 -4.43162285e-02 -3.51431996e-01 -4.97155190e-01
-6.31614476e-02 -1.38741761e-01 2.23554987e-02 2.54636586e-01
6.62208080e-01 -7.96237111e-01 -6.38157308e-01 4.49611917e-02
-8.35359216e-01 3.41328323e-01 -9.09755528e-02 4.78123218e-01
3.27295959e-01 7.67140150e-01 3.57795894e-01 -8.35205495e-01
8.50004971e-01 -8.66717935e-01 -4.08287764e-01 4.62948710e-01
-6.43588722e-01 -2.23000914e-01 6.88524425e-01 -1.42668098e-01
-1.12635112e+00 -1.03266157e-01 1.51028782e-01 1.46065608e-01
-3.35969001e-01 6.93602443e-01 -1.67587828e-02 5.39002478e-01
3.56852561e-01 2.00762138e-01 -7.16887593e-01 -4.26843345e-01
-1.67082265e-01 6.61815405e-01 3.05560619e-01 -4.62825447e-01
6.49591386e-01 7.09340751e-01 2.72683233e-01 -7.78025568e-01
-7.51144648e-01 -4.77842927e-01 -1.03249574e+00 -4.05174702e-01
7.97523975e-01 -4.40410525e-01 -8.67636561e-01 2.90618300e-01
-1.62944460e+00 -2.40464006e-02 1.47593066e-01 1.51745394e-01
-8.61438960e-02 2.03715965e-01 -4.10994470e-01 -1.33859551e+00
-1.18621208e-01 -7.63694167e-01 7.27869272e-01 9.54122096e-02
-6.35640502e-01 -1.50178015e+00 8.11351463e-02 2.50356108e-01
-1.14397332e-02 6.12248003e-01 1.24202609e+00 -2.46156782e-01
-5.13707638e-01 -2.97498345e-01 -6.20941818e-01 -1.70712426e-01
4.90054220e-01 2.59277105e-01 -8.85843217e-01 2.48565704e-01
-1.70838028e-01 5.45242846e-01 7.42004693e-01 5.65357447e-01
1.50436020e+00 -9.65027153e-01 -4.73843902e-01 3.99728656e-01
1.22388089e+00 4.00731742e-01 9.15347755e-01 3.76076549e-01
7.92723060e-01 9.98639226e-01 5.88525295e-01 4.33432609e-01
7.15498984e-01 7.82732308e-01 4.52122837e-01 -2.94962525e-01
3.97905916e-01 -6.01897776e-01 -2.49369666e-02 3.20104331e-01
-6.35409117e-01 -1.92459926e-01 -1.47197247e+00 9.00058150e-01
-2.33099818e+00 -1.60434616e+00 -7.52628565e-01 2.00490212e+00
1.96578950e-01 -1.07960431e-02 1.23524658e-01 5.81826150e-01
7.37429142e-01 -3.00239250e-02 -1.80457085e-01 -6.31688893e-01
3.91602106e-02 -6.49916753e-02 3.37347955e-01 5.31997025e-01
-9.51115072e-01 9.05805290e-01 6.09892416e+00 7.63349771e-01
-9.32368100e-01 5.47874384e-02 1.28581667e+00 1.84550360e-01
-3.09179276e-01 8.52020904e-02 -3.84162515e-01 8.07662845e-01
8.16733003e-01 -3.73043358e-01 5.18137693e-01 4.80711937e-01
1.00151706e+00 -1.93269759e-01 -9.89932418e-01 6.26003683e-01
-4.52721030e-01 -1.61560369e+00 -5.98302223e-02 2.68220156e-01
3.66778314e-01 -3.93208504e-01 1.17507495e-01 -2.24022314e-01
3.51453960e-01 -1.29824233e+00 8.15695226e-01 8.48475873e-01
4.74909246e-01 -1.03771091e+00 5.38990736e-01 7.58073986e-01
-1.33105624e+00 -3.87494206e-01 -1.77410066e-01 -8.10474873e-01
2.05985829e-01 5.17692506e-01 -6.51626706e-01 2.67572165e-01
8.95565212e-01 1.24329340e+00 -4.83888119e-01 6.62898183e-01
-4.00172502e-01 8.42179060e-01 -1.16585940e-01 1.01321749e-01
4.20990825e-01 -4.37237084e-01 2.90512651e-01 1.26609814e+00
2.15325177e-01 7.97203541e-01 -2.32696861e-01 1.16510820e+00
3.71016294e-01 -2.65061080e-01 -1.09081507e+00 2.37391531e-01
4.81067985e-01 8.31868947e-01 -7.54891336e-01 -2.08024005e-03
-2.14281529e-01 5.39004982e-01 7.03389227e-01 3.30698490e-01
-6.06048584e-01 -3.23446281e-02 9.46560860e-01 5.83278418e-01
-2.60635048e-01 -7.10597411e-02 -7.93840885e-01 -9.28223670e-01
1.96148813e-01 -3.47626150e-01 5.06860733e-01 -4.91140634e-01
-1.17975199e+00 4.34595197e-01 1.60053909e-01 -9.55700099e-01
-1.48682833e-01 -3.00763726e-01 -1.30236721e+00 1.02336085e+00
-1.39379036e+00 -1.21627939e+00 8.20785314e-02 6.22990906e-01
4.34754550e-01 7.97405541e-02 7.25268781e-01 1.10413626e-01
-9.55345869e-01 1.04655623e-01 -1.25554070e-01 5.71630776e-01
-3.19540709e-01 -8.47361743e-01 6.49122059e-01 1.26802027e+00
1.08744293e-01 6.84424222e-01 5.99575639e-01 -8.21260095e-01
-7.22799480e-01 -1.37598002e+00 1.77584231e+00 -8.08971286e-01
8.27864349e-01 -2.79166669e-01 -9.72304761e-01 6.46203697e-01
-4.25659299e-01 -2.08927348e-01 6.43108666e-01 5.00934243e-01
-6.43521473e-02 -1.59437433e-01 -1.13294613e+00 8.01472485e-01
1.17293227e+00 -5.34991801e-01 -4.14482862e-01 5.01855433e-01
3.65038514e-01 1.50991201e-01 -5.49924433e-01 5.45272157e-02
2.22263843e-01 -1.21414471e+00 9.97157991e-01 -9.33178186e-01
6.88700497e-01 9.02572572e-02 1.27715111e-01 -1.16655540e+00
-8.41107368e-01 -7.92412087e-02 5.00231422e-02 1.28194952e+00
3.84415865e-01 -5.86823046e-01 5.85081995e-01 1.48852134e+00
2.31954053e-01 -5.91899812e-01 -1.20465231e+00 -6.09541535e-01
-7.53202289e-02 -9.69483852e-01 1.30206156e+00 1.10234118e+00
4.38112170e-01 6.76836371e-02 -7.00474739e-01 5.39561450e-01
5.72283924e-01 -7.33251721e-02 3.85520160e-01 -1.31699407e+00
2.47690409e-01 -3.40443194e-01 -6.58432066e-01 -4.51224685e-01
2.19385073e-01 -4.63637382e-01 -6.92057550e-01 -1.51178706e+00
5.75408638e-01 -7.20659673e-01 -1.44668370e-01 8.17662418e-01
-2.74753451e-01 -5.34659624e-02 2.30477259e-01 3.02219480e-01
-4.58918095e-01 2.31639534e-01 8.50912631e-01 -1.95392087e-01
-2.52101332e-01 2.78907388e-01 -6.30837440e-01 1.06048083e+00
8.58232796e-01 -7.84236908e-01 -4.66106981e-01 -8.75181377e-01
3.16762060e-01 2.71986187e-01 7.60615468e-01 -6.63292825e-01
3.27065080e-01 -9.74889398e-01 3.07763785e-01 -5.19304574e-01
2.93227166e-01 -1.09001422e+00 2.57066488e-01 4.63032126e-01
-2.89111823e-01 3.45674396e-01 1.11060619e-01 5.70797980e-01
-2.38242149e-01 1.30228270e-02 2.63376236e-01 1.44219354e-01
-6.18829310e-01 6.64176404e-01 -5.63129961e-01 -1.91965938e-01
1.07877243e+00 -4.80530560e-01 -3.82406563e-01 -4.39237028e-01
-6.14008605e-01 4.08186823e-01 1.24016829e-01 5.25989890e-01
8.42725456e-01 -1.35393047e+00 -7.60289967e-01 2.05699310e-01
-2.73445239e-05 -2.98843235e-01 2.92601556e-01 5.33365250e-01
-3.58415216e-01 6.65177584e-01 -1.51930749e-01 -2.70548463e-01
-1.43430531e+00 3.85322332e-01 1.55337408e-01 -5.00721216e-01
-6.41724885e-01 2.39298001e-01 2.35748544e-01 -1.50874481e-01
-2.76164979e-01 -1.74813271e-01 -6.83253706e-01 -3.17360371e-01
8.99304509e-01 7.00644135e-01 -2.82228380e-01 -1.07156873e+00
-4.56978649e-01 3.57635915e-01 2.47812629e-01 2.67855942e-01
1.99386585e+00 -3.49720955e-01 -4.22187090e-01 2.68974096e-01
7.50747323e-01 -3.98949891e-01 -1.11478853e+00 -1.22747459e-01
3.31649035e-01 -1.07373655e+00 1.38354450e-01 -4.77235585e-01
-1.32572973e+00 1.05006480e+00 1.59643099e-01 6.17399037e-01
1.28138328e+00 -1.32748127e-01 5.54118276e-01 5.77292740e-02
3.31080735e-01 -1.30270255e+00 -2.89039522e-01 4.19952035e-01
7.12873101e-01 -1.14886665e+00 -7.86224082e-02 -6.24414325e-01
-6.18446231e-01 1.10816884e+00 4.24323857e-01 4.34009321e-02
8.28135252e-01 5.60224894e-03 -4.06695366e-01 -4.75887388e-01
-6.23212099e-01 -1.16688954e-02 3.63757670e-01 7.31157601e-01
1.77975208e-01 4.54964787e-01 -3.62407938e-02 5.69656909e-01
-8.46803375e-03 -1.81724340e-01 3.05894703e-01 4.07630146e-01
-3.57165277e-01 -7.48422801e-01 -5.20222127e-01 5.68692446e-01
-3.46325725e-01 1.50288679e-02 -3.76914024e-01 6.99819565e-01
4.15449828e-01 1.32942247e+00 3.25816482e-01 -3.71751219e-01
3.88353825e-01 -3.70232791e-01 -1.91936329e-01 -5.20586133e-01
-3.65534782e-01 -5.91387331e-01 5.72336800e-02 -5.66963494e-01
-2.35564321e-01 -9.07735169e-01 -1.07436681e+00 -1.10668194e+00
4.44119200e-02 -1.90752447e-02 2.61448473e-01 1.33073318e+00
3.24536800e-01 2.82005131e-01 8.89630914e-01 -3.32583040e-01
3.48584086e-01 -6.16118133e-01 -7.15526223e-01 7.19197094e-01
3.04983914e-01 -6.10749483e-01 -1.12225860e-01 2.00236365e-02] | [6.713070869445801, 1.9809898138046265] |
ef5d4dd8-99ec-4c81-9574-9e6a7cfe9443 | effectively-leveraging-multi-modal-features | 2203.13281 | null | https://arxiv.org/abs/2203.13281v1 | https://arxiv.org/pdf/2203.13281v1.pdf | Effectively leveraging Multi-modal Features for Movie Genre Classification | Movie genre classification has been widely studied in recent years due to its various applications in video editing, summarization, and recommendation. Prior work has typically addressed this task by predicting genres based solely on the visual content. As a result, predictions from these methods often perform poorly for genres such as documentary or musical, since non-visual modalities like audio or language play an important role in correctly classifying these genres. In addition, the analysis of long videos at frame level is always associated with high computational cost and makes the prediction less efficient. To address these two issues, we propose a Multi-Modal approach leveraging shot information, MMShot, to classify video genres in an efficient and effective way. We evaluate our method on MovieNet and Condensed Movies for genre classification, achieving 17% ~ 21% improvement on mean Average Precision (mAP) over the state-of-the-art. Extensive experiments are conducted to demonstrate the ability of MMShot for long video analysis and uncover the correlations between genres and multiple movie elements. We also demonstrate our approach's ability to generalize by evaluating the scene boundary detection task, achieving 1.1% improvement on Average Precision (AP) over the state-of-the-art. | ['Huayan Wang', 'Jiayi Liu', 'Xin Miao', 'Bryan A. Plummer', 'Yiwen Gu', 'Zhongping Zhang'] | 2022-03-24 | null | null | null | null | ['boundary-detection', 'genre-classification'] | ['computer-vision', 'computer-vision'] | [ 3.24919522e-01 -4.86112088e-01 -3.25867563e-01 -1.95920125e-01
-9.33818758e-01 -5.44297576e-01 5.36679924e-01 2.74741530e-01
-2.19580129e-01 4.22133476e-01 3.39544475e-01 1.46983191e-01
1.35989547e-01 -5.16166747e-01 -5.09421408e-01 -4.47120696e-01
-9.31572616e-02 -1.18709520e-01 4.15555298e-01 -1.38127608e-02
4.67834294e-01 1.53899984e-02 -1.78071404e+00 8.73370707e-01
7.11457908e-01 1.41332424e+00 2.31104344e-01 6.80581689e-01
-1.86033905e-01 8.77124071e-01 -7.18839884e-01 -4.99985486e-01
5.85817769e-02 -5.27490199e-01 -6.20851338e-01 3.46569091e-01
5.01034796e-01 -3.50455642e-01 -3.02484602e-01 9.62962925e-01
2.67341584e-01 3.38311791e-01 6.69935524e-01 -9.58857894e-01
-3.29222769e-01 4.61349666e-01 -9.45781589e-01 3.73470813e-01
8.27826798e-01 -1.32963091e-01 1.25740087e+00 -9.15430903e-01
6.00448787e-01 1.10084867e+00 5.85598648e-01 2.43185461e-01
-1.04998171e+00 -7.34835505e-01 3.40173751e-01 6.53082550e-01
-1.49052453e+00 -6.42190218e-01 7.77765334e-01 -5.02841592e-01
8.52110326e-01 2.43781015e-01 5.07698119e-01 1.05555236e+00
8.68804231e-02 1.05767798e+00 7.45339751e-01 -4.23281819e-01
1.78518161e-01 -2.76525188e-02 -1.07851222e-01 4.97502357e-01
-1.83633789e-01 -4.77065355e-01 -8.64263773e-01 5.23689166e-02
5.32791734e-01 -4.51593660e-02 -4.35025632e-01 6.43812269e-02
-1.21300280e+00 6.27019525e-01 1.55812001e-03 1.11154027e-01
-2.98507631e-01 -1.30976826e-01 8.49574208e-01 3.13283324e-01
7.34345734e-01 5.71093440e-01 -1.44547299e-01 -6.37214065e-01
-1.13396490e+00 2.28269503e-01 7.23888218e-01 8.29890072e-01
1.29809394e-01 -1.38061598e-01 -2.56029218e-01 1.25654256e+00
-4.00620364e-02 6.11413866e-02 4.21910584e-01 -1.10394096e+00
6.47553444e-01 4.57944512e-01 -3.72165651e-03 -1.23264933e+00
-2.18797639e-01 -3.05633694e-01 -8.31355929e-01 -3.32580537e-01
3.79499614e-01 2.60829628e-01 -6.03971422e-01 1.45605981e+00
1.08823858e-01 3.38443249e-01 -1.78971037e-01 1.04749000e+00
8.80386055e-01 9.09648001e-01 5.83529379e-03 -7.31537402e-01
1.42679632e+00 -1.03670216e+00 -7.59027302e-01 -4.77102678e-03
3.71779144e-01 -8.21776152e-01 1.28956497e+00 7.68821120e-01
-9.07431781e-01 -6.38483107e-01 -1.02262235e+00 1.67642370e-01
5.13246059e-02 1.96643531e-01 2.44054526e-01 2.61023462e-01
-4.36928600e-01 7.16858208e-01 -7.80096531e-01 -4.83141780e-01
5.27077317e-01 2.10319646e-02 -2.62492478e-01 -8.93773511e-02
-9.30047333e-01 3.36099267e-01 2.60727018e-01 -3.10906291e-01
-7.16082811e-01 -6.56987071e-01 -6.71577871e-01 1.46622583e-01
7.06003964e-01 -2.59371668e-01 1.28115451e+00 -1.09644806e+00
-1.44268966e+00 5.09726346e-01 -2.28517979e-01 -4.04166132e-01
3.13535094e-01 -3.40095013e-01 -6.92980051e-01 7.20496714e-01
6.40056506e-02 4.57561761e-01 8.89378726e-01 -8.24986398e-01
-8.86404157e-01 -2.74870545e-01 4.58162487e-01 3.67634684e-01
-8.43316317e-01 2.83172309e-01 -8.15003693e-01 -1.18667138e+00
3.69209982e-02 -8.83614779e-01 1.60415307e-01 -6.14792220e-02
-2.05197752e-01 -3.44292998e-01 7.78399825e-01 -7.30068326e-01
1.80406117e+00 -2.42546153e+00 2.80250162e-01 -2.07853138e-01
1.18157580e-01 9.02566165e-02 2.99858898e-02 4.22095746e-01
3.47508907e-01 4.99020852e-02 -3.47740464e-02 -4.02172446e-01
-2.35702157e-01 -9.54728499e-02 -3.81279767e-01 3.56655270e-01
-6.47532269e-02 4.30803776e-01 -8.14243078e-01 -7.12509096e-01
3.79810482e-02 2.11194873e-01 -5.75408936e-01 1.20348580e-01
-2.26637572e-01 4.77134049e-01 -2.19977498e-01 7.64697731e-01
2.04286724e-01 -4.17251498e-01 2.68611729e-01 -3.36941451e-01
-3.83181348e-02 3.03743899e-01 -1.07395470e+00 1.92271543e+00
-3.38843614e-01 9.18511748e-01 -1.88531965e-01 -1.02599180e+00
4.64443713e-01 4.86924022e-01 5.12779653e-01 -6.19909346e-01
-4.54479866e-02 -4.77697104e-02 -1.35486498e-01 -6.79473162e-01
8.23743463e-01 -2.63667908e-02 -1.82472140e-01 3.28405976e-01
-3.89958778e-03 2.96215922e-01 5.58450699e-01 2.45626643e-01
1.03652048e+00 1.61334127e-01 3.32569242e-01 3.73558700e-02
5.60613573e-01 1.03016095e-02 6.57480001e-01 7.35301435e-01
-1.37403354e-01 8.56852353e-01 5.59698761e-01 -1.96964189e-01
-8.33261907e-01 -6.47270858e-01 -5.97669408e-02 1.38146532e+00
3.35379153e-01 -1.02687585e+00 -7.20670044e-01 -6.76831841e-01
-2.83068389e-01 5.13359547e-01 -4.17502314e-01 -3.90248932e-02
-4.33065444e-01 -5.65282762e-01 5.23027062e-01 5.64320147e-01
5.35439432e-01 -7.96852887e-01 -4.37632680e-01 1.70273244e-01
-5.30077577e-01 -1.53065908e+00 -4.42111820e-01 -5.21454990e-01
-7.40937889e-01 -1.03618693e+00 -6.82804644e-01 -4.65408385e-01
2.91375399e-01 5.52929342e-01 9.45392668e-01 -2.27469355e-01
-9.20877978e-02 3.44754577e-01 -8.15080941e-01 -1.74457714e-01
-3.67984027e-01 1.50136501e-01 3.03002149e-01 3.17802697e-01
2.65919745e-01 -5.58685422e-01 -7.95125425e-01 5.25869906e-01
-8.29911470e-01 2.96187013e-01 2.47457862e-01 6.48120105e-01
6.57791436e-01 1.69697583e-01 7.13002384e-01 -8.48543704e-01
4.98080641e-01 -5.17511129e-01 -2.46221453e-01 2.45869637e-01
-4.40922379e-01 -4.51570570e-01 9.35049057e-01 -6.48738027e-01
-9.58874822e-01 -1.21313512e-01 3.49129736e-03 -8.08762014e-01
-1.68261573e-01 6.13826513e-01 4.94807363e-02 8.70439634e-02
4.33620006e-01 2.50568479e-01 -1.92612290e-01 -7.07465827e-01
-1.76802762e-02 9.98259783e-01 5.55306554e-01 -4.17362869e-01
2.18423247e-01 3.86480778e-01 -1.60871908e-01 -1.06880152e+00
-1.06775272e+00 -8.11734974e-01 -3.54174107e-01 -6.28629804e-01
8.11095953e-01 -1.13481653e+00 -4.92612123e-01 4.57201779e-01
-1.00043881e+00 1.98498875e-01 4.23486531e-02 5.92483699e-01
-5.06815374e-01 6.89534724e-01 -8.34502459e-01 -6.78097248e-01
-2.49176234e-01 -1.12893522e+00 1.07887638e+00 8.02799910e-02
-3.22384447e-01 -6.32685840e-01 -2.17488140e-01 6.08414114e-01
7.72186294e-02 1.40613034e-01 8.68413031e-01 -5.72190106e-01
-3.39427739e-01 -1.95250764e-01 -1.97401911e-01 2.79400527e-01
-5.77505678e-02 1.37005299e-02 -9.64817822e-01 -2.64263481e-01
-1.53235495e-01 -3.12056541e-01 9.49711919e-01 2.02834219e-01
1.56753302e+00 -2.34886959e-01 -9.23391283e-02 3.54296654e-01
1.19323039e+00 1.09599069e-01 5.16010940e-01 2.49983266e-01
7.92844355e-01 5.64229488e-01 1.09698319e+00 7.39508390e-01
3.15748394e-01 1.12044489e+00 3.17889035e-01 5.08126259e-01
-8.76459926e-02 -1.54632688e-01 4.45063561e-01 9.27936316e-01
-3.32841098e-01 -3.50311339e-01 -7.37098813e-01 4.63907629e-01
-2.04419732e+00 -1.19395816e+00 1.90521590e-02 2.24094534e+00
6.88702345e-01 2.30152249e-01 3.59320164e-01 4.63790745e-01
7.34083176e-01 3.18554252e-01 -3.73340040e-01 -2.21417978e-01
4.11573909e-02 -2.06801280e-01 7.42700249e-02 -1.50879338e-01
-1.42784679e+00 7.73098350e-01 5.99611473e+00 1.30718386e+00
-1.07163024e+00 1.39373496e-01 5.96573770e-01 -4.96317834e-01
2.57238239e-01 -3.03347945e-01 -6.28751576e-01 7.03670025e-01
8.15850556e-01 -9.90823936e-03 4.45311934e-01 7.34733701e-01
3.35104525e-01 -3.34983230e-01 -1.12199593e+00 1.30726171e+00
4.50950474e-01 -1.38475192e+00 1.43383756e-01 -1.15052564e-02
7.60105729e-01 -2.88104385e-01 -1.32926047e-01 3.70810837e-01
-6.57579541e-01 -6.36265039e-01 9.40549672e-01 2.51439452e-01
1.02894258e+00 -7.56648898e-01 5.32574058e-01 3.77891630e-01
-1.39719009e+00 -2.90317267e-01 -2.78793305e-01 -2.63046592e-01
2.56342381e-01 4.27604556e-01 -4.75191265e-01 4.86619622e-01
8.47938716e-01 1.06676877e+00 -4.79583800e-01 1.07749784e+00
1.18013611e-02 7.80364990e-01 -9.15444568e-02 1.81691915e-01
3.04084960e-02 6.55710744e-03 7.91112542e-01 1.33504486e+00
3.33503455e-01 2.54621893e-01 4.41933513e-01 1.69113040e-01
-2.82977879e-01 3.55794042e-01 -3.91796529e-01 -1.70807257e-01
4.15681034e-01 1.17299426e+00 -9.83585954e-01 -4.38628286e-01
-6.67334139e-01 9.65019345e-01 1.88370596e-03 1.59324437e-01
-1.00922322e+00 -3.69969010e-01 6.75536752e-01 2.53165245e-01
5.57191133e-01 -1.44927055e-01 -8.01374465e-02 -1.40318811e+00
1.61952674e-01 -9.83020723e-01 4.67447311e-01 -6.05322301e-01
-1.17119479e+00 5.53853869e-01 -5.06744124e-02 -1.79563475e+00
-3.04720521e-01 -2.81113446e-01 -2.54072279e-01 9.70498398e-02
-1.02441788e+00 -9.01769996e-01 -3.01870644e-01 4.73677784e-01
1.23586214e+00 -2.72621721e-01 7.66788423e-01 5.76635778e-01
-7.15828300e-01 7.55338550e-01 2.53544599e-01 8.30997247e-03
9.05499399e-01 -9.08218443e-01 -3.72737609e-02 8.01882982e-01
4.43571538e-01 2.54139960e-01 7.57050693e-01 -3.80320668e-01
-1.44942486e+00 -1.11962044e+00 5.46651721e-01 -1.75026447e-01
7.02470660e-01 -3.28064948e-01 -7.91385829e-01 1.59538373e-01
-4.96438239e-04 -1.52828261e-01 9.54928041e-01 2.42409050e-01
-4.92128432e-01 -2.54301190e-01 -7.15961218e-01 5.82434177e-01
1.15023041e+00 -6.72010422e-01 -4.42542642e-01 2.00743511e-01
6.26879394e-01 -3.81086975e-01 -1.01999664e+00 4.90104824e-01
8.79893422e-01 -1.03338158e+00 8.42869520e-01 -4.13130730e-01
1.02679372e+00 -2.53775567e-01 -3.76489729e-01 -9.93288219e-01
-1.93539321e-01 -4.26502287e-01 -6.03616714e-01 1.36314404e+00
1.64330393e-01 3.35344598e-02 5.54037154e-01 1.25177532e-01
-8.98941085e-02 -9.40485120e-01 -8.70770693e-01 -7.61425614e-01
-5.28772831e-01 -8.18996072e-01 1.78242326e-01 8.71330202e-01
2.05992192e-01 5.08840203e-01 -7.29603946e-01 -8.64844099e-02
3.97970915e-01 4.33612376e-01 6.69863045e-01 -1.05829835e+00
-4.33635831e-01 -4.80004549e-01 -6.72282755e-01 -1.19324219e+00
9.83278155e-02 -5.32572925e-01 -1.76067241e-02 -1.31893611e+00
4.78188187e-01 -7.56871104e-02 -4.31702405e-01 1.75190970e-01
-1.08523928e-01 8.47817421e-01 4.51882362e-01 6.58386350e-01
-1.38395739e+00 3.97003829e-01 1.02656901e+00 -1.33281112e-01
-1.88057333e-01 2.16644146e-02 -6.12325251e-01 9.14002419e-01
5.12333751e-01 -2.21370563e-01 -3.87701720e-01 -2.68093556e-01
2.78591782e-01 2.62098134e-01 1.86458267e-02 -1.22068822e+00
1.03694901e-01 -1.22581907e-02 2.73334563e-01 -4.80066836e-01
5.48985243e-01 -5.17381132e-01 -3.85069214e-02 -1.09283373e-01
-3.93972754e-01 -8.70203897e-02 1.20049335e-01 9.22920167e-01
-6.07519627e-01 -5.92473056e-03 5.11814296e-01 7.42944628e-02
-8.50852609e-01 1.31400391e-01 -4.51754034e-01 4.64795604e-02
9.92036879e-01 -3.27431500e-01 -2.66709358e-01 -5.51554561e-01
-6.76109850e-01 9.16559771e-02 6.27082646e-01 6.74262762e-01
6.23601079e-01 -1.17008197e+00 -6.04703844e-01 -1.91402826e-02
4.19379711e-01 -3.40495348e-01 4.16630298e-01 1.12399769e+00
-5.10935426e-01 2.69272745e-01 -1.00652492e-02 -7.91047990e-01
-1.57533216e+00 4.53138351e-01 -1.79051042e-01 -2.32634574e-01
-6.73172772e-01 6.37298465e-01 4.61582154e-01 4.92104322e-01
5.09923458e-01 -7.59563083e-03 -6.29248202e-01 3.70726377e-01
8.09767485e-01 5.23621857e-01 2.95334477e-02 -7.39920914e-01
-2.57177502e-01 6.32595778e-01 -2.75826871e-01 7.79182389e-02
1.20816517e+00 -3.84634495e-01 9.98357460e-02 7.59717941e-01
1.05477333e+00 2.11258717e-02 -1.15712047e+00 -1.58930749e-01
-9.83924717e-02 -7.31364667e-01 2.54242271e-02 -5.58686852e-01
-9.62363482e-01 8.07461441e-01 3.51040930e-01 4.57965195e-01
1.43309271e+00 8.08672011e-02 9.42061722e-01 2.24069074e-01
4.29635763e-01 -1.28847051e+00 1.61807165e-01 3.95166814e-01
7.54390001e-01 -1.32267547e+00 2.30928928e-01 -5.23246884e-01
-8.46878171e-01 1.13200903e+00 5.16963661e-01 -3.70910182e-03
4.50165868e-01 -1.21745862e-01 -2.39977986e-01 5.84892277e-03
-8.24750483e-01 7.37666860e-02 5.76379359e-01 -7.77963698e-02
4.65836406e-01 1.23841716e-02 -4.32608873e-01 8.69649470e-01
1.29969165e-01 -2.90850643e-02 4.01735544e-01 7.98508227e-01
-3.63619268e-01 -8.58441472e-01 -2.85878748e-01 7.12433040e-01
-1.08112991e+00 3.46964002e-02 -2.55026400e-01 3.87298256e-01
-1.08019318e-02 1.14152360e+00 1.90172255e-01 -7.22215772e-01
1.94760844e-01 -8.04551169e-02 5.16959369e-01 -5.97936034e-01
-5.17125845e-01 4.31553692e-01 2.87215441e-01 -7.30299473e-01
-6.82120621e-01 -7.52198875e-01 -8.54550183e-01 -3.01331699e-01
-2.69873112e-01 4.87304851e-02 4.75786388e-01 1.00186586e+00
4.57200021e-01 5.52148759e-01 6.93669081e-01 -9.64764953e-01
-3.77756566e-01 -1.04812336e+00 -6.16211414e-01 6.11778080e-01
2.64373049e-02 -7.36927748e-01 -2.19774082e-01 2.74036705e-01] | [10.153129577636719, 0.524202287197113] |
ce905218-f6ce-400b-b0ec-3927d976ca51 | discriminative-deep-dyna-q-robust-planning | 1808.09442 | null | http://arxiv.org/abs/1808.09442v2 | http://arxiv.org/pdf/1808.09442v2.pdf | Discriminative Deep Dyna-Q: Robust Planning for Dialogue Policy Learning | This paper presents a Discriminative Deep Dyna-Q (D3Q) approach to improving
the effectiveness and robustness of Deep Dyna-Q (DDQ), a recently proposed
framework that extends the Dyna-Q algorithm to integrate planning for
task-completion dialogue policy learning. To obviate DDQ's high dependency on
the quality of simulated experiences, we incorporate an RNN-based discriminator
in D3Q to differentiate simulated experience from real user experience in order
to control the quality of training data. Experiments show that D3Q
significantly outperforms DDQ by controlling the quality of simulated
experience used for planning. The effectiveness and robustness of D3Q is
further demonstrated in a domain extension setting, where the agent's
capability of adapting to a changing environment is tested. | ['Yun-Nung Chen', 'Shang-Yu Su', 'Jingjing Liu', 'Jianfeng Gao', 'Xiujun Li'] | 2018-08-28 | discriminative-deep-dyna-q-robust-planning-1 | https://aclanthology.org/D18-1416 | https://aclanthology.org/D18-1416.pdf | emnlp-2018-10 | ['task-completion-dialogue-policy-learning'] | ['natural-language-processing'] | [-4.10014123e-01 3.17221642e-01 1.45302325e-01 -1.80876240e-01
-8.00467908e-01 -7.35831857e-01 1.01474404e+00 -1.29935026e-01
-8.71191084e-01 7.11534679e-01 4.58453536e-01 -3.21213186e-01
-1.04277998e-01 -5.57388425e-01 -5.43845475e-01 -4.15110141e-01
-3.50761324e-01 4.73341614e-01 2.20524907e-01 -5.05242527e-01
2.13028789e-01 3.42714190e-01 -1.32207382e+00 6.24113865e-02
9.31643724e-01 7.72665381e-01 5.20198286e-01 1.00311792e+00
3.47268552e-01 7.38462627e-01 -9.57989573e-01 -8.44833208e-04
5.03550231e-01 -6.01439536e-01 -8.59659612e-01 9.06002969e-02
-1.52376324e-01 -8.05440664e-01 -3.97022277e-01 7.60030031e-01
9.27036524e-01 6.69644654e-01 4.17469144e-01 -1.18645990e+00
-2.98398018e-01 2.76628464e-01 3.56875539e-01 3.02718192e-01
7.43602753e-01 7.43624926e-01 7.71688938e-01 -4.55905735e-01
7.76876807e-01 1.61140764e+00 2.76531398e-01 7.59996116e-01
-9.98942614e-01 -1.89518914e-01 1.96244553e-01 -7.75766596e-02
-8.77012610e-01 -4.05126214e-01 3.75386059e-01 -2.81437039e-01
1.33149004e+00 -1.75814927e-01 8.13365698e-01 1.40837240e+00
4.28503841e-01 1.13652158e+00 9.65055645e-01 -2.96192110e-01
7.72695601e-01 8.12593699e-02 -4.33087528e-01 5.28698206e-01
-6.52566314e-01 9.23602819e-01 -3.69047403e-01 -1.61945835e-01
9.23408926e-01 -5.78852594e-01 -1.04446173e-01 -3.99653584e-01
-1.14592063e+00 8.74164999e-01 2.32212886e-01 2.55044773e-02
-5.94299078e-01 5.88305853e-02 7.58465528e-01 8.39351058e-01
2.16331601e-01 1.12992716e+00 -5.66045344e-01 -8.85059118e-01
-2.39640698e-01 7.49282539e-01 8.07956696e-01 9.64786172e-01
7.00652599e-02 3.92701328e-01 -7.84606576e-01 7.39641488e-01
1.41622216e-01 5.29605329e-01 8.29316914e-01 -1.58762479e+00
3.55602592e-01 2.83918351e-01 6.43090010e-01 -3.63107830e-01
-6.64322019e-01 -2.10444942e-01 -4.69608121e-02 4.92766768e-01
3.83058220e-01 -4.03246313e-01 -7.18749166e-01 1.94354820e+00
2.90435314e-01 -9.32442769e-02 6.68139696e-01 1.10155272e+00
5.62897205e-01 7.70079017e-01 2.35776365e-01 -1.90364838e-01
8.25655878e-01 -1.18593681e+00 -6.33798480e-01 1.50594339e-02
8.11010599e-01 -2.38806978e-01 1.44051564e+00 4.61431354e-01
-1.30254912e+00 -8.42684865e-01 -7.97650337e-01 1.87890664e-01
-1.59414619e-01 -2.25418806e-01 4.84562486e-01 3.05434883e-01
-1.21406949e+00 6.93656027e-01 -8.00687909e-01 -3.64644468e-01
-1.49730831e-01 3.52079064e-01 -2.47510076e-01 1.90771267e-01
-1.68054044e+00 1.16215026e+00 6.49369657e-01 -2.57719994e-01
-1.47134840e+00 -2.07115844e-01 -1.00889230e+00 -5.46860620e-02
5.23988605e-01 -5.43959439e-01 1.97317982e+00 -8.41016829e-01
-2.41711283e+00 3.13735098e-01 5.20546079e-01 -5.66502690e-01
8.01036358e-01 -3.84979516e-01 -4.14239377e-01 2.48687342e-01
1.20215841e-01 9.79022920e-01 6.75928056e-01 -9.38222468e-01
-6.26491427e-01 -1.50838261e-02 5.04739225e-01 8.46230388e-01
2.88610339e-01 -3.77526402e-01 -3.90461296e-01 -4.58260298e-01
-5.63455522e-01 -1.12606859e+00 -4.74548161e-01 -3.27106535e-01
-1.10687986e-01 -5.91678202e-01 4.17037159e-01 -4.75971431e-01
8.45597744e-01 -2.26866198e+00 3.13477397e-01 -3.81488609e-03
-2.53448963e-01 6.42263710e-01 -9.00215745e-01 7.18654633e-01
4.72492903e-01 -2.54996747e-01 9.46038589e-02 -3.36460292e-01
2.60632813e-01 5.07612526e-01 -1.19860381e-01 1.23541512e-01
4.15524602e-01 9.61299479e-01 -1.23405063e+00 -1.62713811e-01
3.35640669e-01 2.26199940e-01 -6.74355805e-01 7.67713130e-01
-6.62959635e-01 6.68661416e-01 -4.67519373e-01 1.59149438e-01
2.10315362e-01 1.90371603e-01 9.83905122e-02 4.65474099e-01
-1.26296997e-01 6.48909092e-01 -9.87734020e-01 1.99265695e+00
-7.82016695e-01 1.95999697e-01 1.04198098e-01 -3.85703117e-01
8.42515886e-01 5.64637601e-01 1.93571344e-01 -1.43325484e+00
2.57418603e-01 -1.22579634e-01 2.02492654e-01 -6.55102611e-01
9.25056338e-01 6.92182332e-02 -2.69880742e-01 5.25738597e-01
1.88124657e-01 -4.03065801e-01 1.72786996e-01 2.61646509e-01
1.03866184e+00 5.52056491e-01 3.25583935e-01 -9.16558206e-02
1.37984127e-01 -1.94972262e-01 5.61629653e-01 1.01436520e+00
-7.56812453e-01 3.23580623e-01 5.99262714e-01 -1.72811911e-01
-1.05476058e+00 -1.07402599e+00 2.46885285e-01 1.48063052e+00
2.21148923e-01 -3.56730931e-02 -7.33731329e-01 -1.00093865e+00
1.27778845e-02 1.22929120e+00 -5.35002410e-01 -3.72149020e-01
-4.53274995e-01 -3.00920248e-01 4.03423667e-01 6.66858494e-01
4.59959030e-01 -1.66630912e+00 -1.12829876e+00 5.29991865e-01
-5.98837957e-02 -1.12982309e+00 -4.95895296e-01 2.54468024e-01
-6.05844796e-01 -7.51943111e-01 -6.36347234e-01 -3.18005621e-01
7.70168379e-02 -3.71661633e-02 1.12810111e+00 -3.71137232e-01
3.82632941e-01 5.50005734e-01 -7.37611890e-01 -5.09954542e-02
-9.59393859e-01 2.35226899e-02 2.85638213e-01 -7.39017248e-01
-1.42894104e-01 -3.46747309e-01 -5.34454525e-01 3.51159096e-01
-8.98320794e-01 -1.47901565e-01 5.24866462e-01 1.02117896e+00
1.50267914e-01 -1.35200530e-01 9.32521224e-01 -6.46910429e-01
1.33867502e+00 -3.85103106e-01 -6.19246900e-01 -1.45950645e-01
-7.72093892e-01 3.28048289e-01 7.53661633e-01 -6.33960068e-01
-1.01711571e+00 -2.94155866e-01 -4.21642721e-01 -1.71099916e-01
4.41628415e-03 5.10681093e-01 -2.30738640e-01 2.80890703e-01
5.73591292e-01 8.38799477e-02 1.64236903e-01 -7.62710571e-02
6.81400120e-01 5.77498078e-01 5.33721983e-01 -4.17681992e-01
1.96095422e-01 -2.03800440e-01 -4.16873246e-01 -3.42655659e-01
-4.41273093e-01 -3.41714382e-01 -1.93666995e-01 -2.34573275e-01
7.03907251e-01 -8.74074578e-01 -9.07785475e-01 3.72915775e-01
-1.14780140e+00 -1.14878190e+00 -5.04753709e-01 6.38750374e-01
-1.12114370e+00 2.47105971e-01 -6.99143529e-01 -9.40051019e-01
-1.71334907e-01 -1.33127153e+00 9.08154368e-01 2.90365726e-01
-2.39836544e-01 -1.00152564e+00 3.52480561e-01 7.87793938e-03
1.72735229e-01 2.44096480e-02 5.29708266e-01 -7.90293157e-01
-1.77390605e-01 1.91077276e-03 -4.47588265e-02 4.30827498e-01
-1.36329219e-01 -3.16990465e-01 -8.19481730e-01 -5.61199665e-01
-6.87432662e-02 -8.39639604e-01 4.43212152e-01 1.57726958e-01
4.71867025e-01 -4.59140211e-01 3.63125652e-01 3.25007945e-01
9.25096154e-01 6.62680030e-01 7.12609947e-01 5.90328217e-01
3.01111042e-01 4.26057160e-01 1.22227859e+00 7.57679880e-01
5.01449585e-01 8.88235688e-01 5.94978273e-01 8.96496698e-02
-2.38279756e-02 -4.14796948e-01 7.27558613e-01 6.33695483e-01
1.30806983e-01 -5.09641707e-01 -4.63304430e-01 4.74813014e-01
-1.86070263e+00 -8.54518116e-01 5.43820322e-01 1.93151975e+00
7.55148292e-01 3.94967943e-01 5.54898918e-01 -1.63157031e-01
2.22280487e-01 2.30775550e-01 -8.02429378e-01 -9.66411471e-01
1.14220627e-01 1.81000847e-02 1.81524917e-01 6.66607916e-01
-9.09423053e-01 1.11387336e+00 6.66558552e+00 6.60918534e-01
-8.42539191e-01 2.60984618e-02 7.07778707e-02 2.11151317e-01
-5.44435270e-02 -4.77069348e-01 -2.50547051e-01 5.67474723e-01
1.05167854e+00 1.21487156e-01 5.11819065e-01 1.03240120e+00
6.51313007e-01 -3.63672495e-01 -1.03793240e+00 4.48003232e-01
-3.91603678e-01 -8.20527077e-01 -2.77181894e-01 -2.41576955e-02
4.78495181e-01 3.78674082e-03 3.04469794e-01 1.08582544e+00
6.97983146e-01 -8.34857762e-01 7.36014307e-01 3.17529768e-01
5.82214355e-01 -9.27406192e-01 8.65184784e-01 6.31509483e-01
-6.15334392e-01 -3.03252161e-01 -2.31237307e-01 -2.94297367e-01
1.81340829e-01 -1.62877619e-01 -1.51796389e+00 3.71628344e-01
2.22834319e-01 2.46655732e-01 -2.38133654e-01 7.22499430e-01
-4.26597178e-01 4.68171120e-01 5.43558300e-02 -2.07659543e-01
8.27668965e-01 4.89965603e-02 7.30119407e-01 1.32285845e+00
6.08872361e-02 1.21870436e-01 6.05519354e-01 6.13424361e-01
1.98564753e-01 -1.75221890e-01 -5.98776639e-01 -1.28994420e-01
6.49743974e-01 8.12512279e-01 -3.07162732e-01 -1.73942715e-01
1.47201642e-01 1.28624356e+00 3.05915445e-01 3.52315634e-01
-5.63920617e-01 -1.78786233e-01 6.48465395e-01 -2.76753306e-01
3.88058335e-01 -4.29714769e-01 4.11977410e-01 -6.53712332e-01
-2.37401381e-01 -1.12057340e+00 1.75347641e-01 -8.01816404e-01
-9.23210204e-01 8.35820913e-01 -7.13214651e-02 -1.39128375e+00
-1.00916445e+00 -2.89777160e-01 -2.67465413e-01 6.69844329e-01
-1.37409365e+00 -8.75908434e-01 1.66998419e-03 3.29423070e-01
7.83961892e-01 -3.72353107e-01 9.79072452e-01 -2.22960159e-01
-3.39671403e-01 5.57618737e-01 3.54133368e-01 -1.83533579e-01
8.52813900e-01 -1.59982109e+00 9.24524665e-01 4.70393062e-01
-3.48357916e-01 1.72213301e-01 8.37586701e-01 -4.74664062e-01
-1.16905534e+00 -1.01671708e+00 2.98021644e-01 -3.90442759e-01
3.45435679e-01 -1.61679298e-01 -8.06282282e-01 3.02027255e-01
1.72364354e-01 -2.96890825e-01 3.05457205e-01 -1.55167550e-01
-1.47155821e-01 1.75659373e-01 -1.42243540e+00 5.77359378e-01
8.08199406e-01 -5.88868022e-01 -7.07559824e-01 8.44670609e-02
1.07951033e+00 -5.81156373e-01 -8.61729383e-01 2.42057323e-01
3.52052420e-01 -1.12178183e+00 5.86078405e-01 -4.78056699e-01
1.58873677e-01 3.75052802e-02 1.95459239e-02 -1.86409771e+00
-2.64653414e-01 -9.78704095e-01 -2.13770255e-01 6.65042162e-01
-5.90705778e-03 -4.57496822e-01 5.11484563e-01 4.57442671e-01
-3.90755087e-01 -7.00318635e-01 -9.26959515e-01 -1.12728465e+00
1.05571516e-01 -2.42965415e-01 5.37925363e-01 6.06009603e-01
2.25953594e-01 3.51608098e-01 -5.43227077e-01 3.78344348e-03
-4.81544398e-02 -2.57001191e-01 7.79786229e-01 -6.50175929e-01
-6.05378091e-01 -1.34327307e-01 -3.02775532e-01 -1.31944346e+00
3.03642094e-01 -4.04067576e-01 5.29241741e-01 -1.45236516e+00
-3.94589454e-01 -3.77441347e-01 -3.01854104e-01 6.73595145e-02
-3.09371173e-01 -2.97362566e-01 5.85510850e-01 5.99585585e-02
-9.96001065e-01 1.08548987e+00 1.76132715e+00 1.20215960e-01
-8.32298219e-01 2.09845424e-01 -2.62713581e-01 3.55496943e-01
7.20116794e-01 -8.35274607e-02 -8.23526144e-01 -3.18003684e-01
-6.36226907e-02 7.56005883e-01 1.38089657e-01 -9.05274868e-01
-1.34172678e-01 -1.90157548e-01 1.79606806e-02 -2.76891351e-01
5.36049426e-01 -3.76334101e-01 -4.47150409e-01 5.12181044e-01
-7.09461868e-01 2.65053958e-01 3.52988899e-01 6.74264669e-01
-9.30122808e-02 -1.41465604e-01 7.82883525e-01 -1.23350747e-01
-1.02271616e+00 -7.09566521e-03 -8.43671501e-01 2.32751012e-01
8.64180744e-01 -1.92496270e-01 -8.68506134e-02 -7.89298773e-01
-7.35305607e-01 6.94484174e-01 3.13606113e-01 4.08368945e-01
7.35676706e-01 -1.14361632e+00 -5.35072625e-01 9.56851896e-03
2.14816153e-01 -1.49378732e-01 2.00718507e-01 4.58225697e-01
-1.79518640e-01 3.88385445e-01 -5.18381119e-01 -2.52497971e-01
-7.97802925e-01 6.37031674e-01 3.99631023e-01 -4.78001863e-01
-6.67877853e-01 7.28877246e-01 -1.88646913e-02 -7.70365953e-01
6.19625151e-01 -3.76380384e-01 -2.78000891e-01 -2.31169701e-01
4.25107419e-01 4.30146456e-01 -7.00428188e-02 -2.82191306e-01
-8.43097121e-02 -3.07989746e-01 -2.38934070e-01 -8.12951446e-01
9.87833083e-01 -2.51211524e-01 7.36090660e-01 3.28039616e-01
8.16292346e-01 -4.19193953e-01 -2.05352855e+00 -1.03074551e-01
7.21706543e-03 -1.87254235e-01 -3.22040059e-02 -1.39532840e+00
-4.70194548e-01 6.00916207e-01 7.70257294e-01 1.26574546e-01
9.25261855e-01 -2.76085079e-01 8.92419755e-01 7.37179637e-01
4.54106659e-01 -1.50759792e+00 7.14902878e-01 8.95202041e-01
1.00570405e+00 -1.23053300e+00 -4.10263389e-01 3.96955997e-01
-1.36230409e+00 9.40693915e-01 7.02654779e-01 -9.54350382e-02
1.26212269e-01 4.57827263e-02 4.04132158e-01 8.19661841e-02
-9.42447424e-01 -3.65269095e-01 -3.76459621e-02 8.62713993e-01
3.35424207e-02 1.83307886e-01 -1.27014518e-01 3.79156500e-01
-1.49705395e-01 4.66330796e-02 6.08905733e-01 1.03761458e+00
-3.88724416e-01 -1.06534922e+00 7.19632208e-02 -5.57162091e-02
7.32065067e-02 2.24867150e-01 -2.69852698e-01 8.73049200e-01
4.82946075e-02 1.02306950e+00 3.71040441e-02 -4.64510828e-01
6.33892298e-01 -1.62453040e-01 4.45434421e-01 -6.69064283e-01
-1.02286327e+00 1.22505821e-01 4.25777495e-01 -9.30942178e-01
-2.77587891e-01 -4.98539865e-01 -1.41676950e+00 -1.24073792e-02
-1.37185723e-01 2.98727095e-01 5.37125349e-01 8.57952178e-01
3.15522999e-01 6.00974798e-01 9.21371758e-01 -7.47966826e-01
-1.13616335e+00 -1.14866197e+00 -5.96195698e-01 4.70822364e-01
5.60243249e-01 -7.56053627e-01 -1.40090749e-01 -5.08515596e-01] | [4.083180904388428, 1.7200183868408203] |
5ec0f125-28d2-4c9d-a813-d722ef12492d | deep-autoencoding-gaussian-mixture-model-for | null | null | https://openreview.net/forum?id=BJJLHbb0- | https://openreview.net/pdf?id=BJJLHbb0- | Deep Autoencoding Gaussian Mixture Model for Unsupervised Anomaly Detection | Unsupervised anomaly detection on multi- or high-dimensional data is of great importance in both fundamental machine learning research and industrial applications, for which density estimation lies at the core. Although previous approaches based on dimensionality reduction followed by density estimation have made fruitful progress, they mainly suffer from decoupled model learning with inconsistent optimization goals and incapability of preserving essential information in the low-dimensional space. In this paper, we present a Deep Autoencoding Gaussian Mixture Model (DAGMM) for unsupervised anomaly detection. Our model utilizes a deep autoencoder to generate a low-dimensional representation and reconstruction error for each input data point, which is further fed into a Gaussian Mixture Model (GMM). Instead of using decoupled two-stage training and the standard Expectation-Maximization (EM) algorithm, DAGMM jointly optimizes the parameters of the deep autoencoder and the mixture model simultaneously in an end-to-end fashion, leveraging a separate estimation network to facilitate the parameter learning of the mixture model. The joint optimization, which well balances autoencoding reconstruction, density estimation of latent representation, and regularization, helps the autoencoder escape from less attractive local optima and further reduce reconstruction errors, avoiding the need of pre-training. Experimental results on several public benchmark datasets show that, DAGMM significantly outperforms state-of-the-art anomaly detection techniques, and achieves up to 14% improvement based on the standard F1 score. | ['Cristian Lumezanu', 'Wei Cheng', 'Qi Song', 'Daeki Cho', 'Bo Zong', 'Martin Renqiang Min', 'Haifeng Chen'] | 2018-01-01 | null | null | null | iclr-2018-1 | ['unsupervised-anomaly-detection-with-specified-5', 'unsupervised-anomaly-detection-with-specified-4', 'unsupervised-anomaly-detection-with-specified-7', 'unsupervised-anomaly-detection-with-specified-6', 'unsupervised-anomaly-detection-with-specified'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [-3.87643963e-01 -2.52496656e-02 1.98568881e-01 -2.10553274e-01
-6.33863747e-01 7.19411895e-02 5.23281455e-01 5.85593507e-02
-4.82625872e-01 2.64748991e-01 -6.54027238e-02 -1.87175021e-01
-8.69554132e-02 -6.57075047e-01 -5.74224830e-01 -1.11681294e+00
6.82364777e-02 6.66459143e-01 -1.78471372e-01 4.35890973e-01
5.88296205e-02 4.83472377e-01 -1.37642336e+00 -3.54895085e-01
1.00626564e+00 1.30536938e+00 1.56112835e-02 4.15873289e-01
-2.93000966e-01 6.83326721e-01 -4.25566316e-01 -4.05654967e-01
1.81720536e-02 -2.80932724e-01 -2.82792270e-01 3.21658134e-01
1.28093600e-01 -6.26353145e-01 -5.45106530e-01 1.26627541e+00
3.58588398e-01 4.02726501e-01 1.03994608e+00 -1.12274384e+00
-6.24473512e-01 1.17825694e-01 -8.35769355e-01 2.00749278e-01
-3.80038619e-01 1.09298401e-01 7.10793138e-01 -1.01093280e+00
-9.12277997e-02 1.03406334e+00 5.49708545e-01 4.53645945e-01
-1.45432782e+00 -5.14881670e-01 5.46380542e-02 4.39780578e-02
-1.59630573e+00 -3.40744138e-01 9.58334744e-01 -5.40248811e-01
9.42173362e-01 -3.17454249e-01 4.72123682e-01 1.16723251e+00
1.48097172e-01 8.50024760e-01 2.83169240e-01 -1.60109356e-01
6.85434759e-01 6.69096559e-02 -1.00905418e-01 6.71931446e-01
2.92778254e-01 -1.50675699e-01 -1.31865710e-01 -4.66874629e-01
9.25512075e-01 3.83173466e-01 -9.98185948e-03 -7.80496776e-01
-5.90520918e-01 1.04703236e+00 4.82751429e-02 2.27689236e-01
-8.11265707e-01 -2.60653961e-02 4.25278276e-01 -1.94166414e-02
6.29812539e-01 6.23703152e-02 -1.62377909e-01 -3.29097062e-01
-9.26909626e-01 1.14668272e-01 5.75553477e-01 3.57525200e-01
7.26073742e-01 5.90325594e-01 2.74638027e-01 1.15265834e+00
7.75648475e-01 4.38092411e-01 8.52959454e-01 -7.60306239e-01
2.79535323e-01 7.72066116e-01 -2.06282973e-01 -9.83820558e-01
-2.05407262e-01 -5.62800825e-01 -1.38889527e+00 1.76493883e-01
2.40923062e-01 -3.83522153e-01 -8.51229668e-01 1.63509119e+00
5.66406012e-01 6.01436794e-01 1.61519438e-01 9.05519068e-01
-2.37830952e-02 7.85228431e-01 9.46795866e-02 -4.90115620e-02
1.02356219e+00 -6.81221187e-01 -6.51854336e-01 -3.70663315e-01
8.33665729e-01 -3.40500832e-01 8.37606847e-01 5.62076390e-01
-8.60881388e-01 -2.90394455e-01 -1.11327589e+00 1.68299571e-01
-1.37040570e-01 2.54061997e-01 4.41999853e-01 4.81589764e-01
-7.05886126e-01 5.29089987e-01 -1.39400506e+00 2.47373935e-02
7.19066083e-01 3.20636094e-01 -5.13100088e-01 3.61285992e-02
-7.46418774e-01 5.34258366e-01 5.59135258e-01 1.56228587e-01
-8.77221406e-01 -5.37805557e-01 -1.11781740e+00 2.45900065e-01
1.96748614e-01 -5.41258276e-01 9.84512508e-01 -5.39282262e-01
-1.69370735e+00 4.13973033e-01 -1.77422557e-02 -7.52889514e-01
1.60439774e-01 -6.04318798e-01 -5.41418970e-01 2.33934931e-02
-2.24629745e-01 3.15405935e-01 1.31515622e+00 -8.59436452e-01
-5.15411794e-01 -5.05045652e-01 -8.23880792e-01 8.27239305e-02
-7.53994465e-01 -2.05212831e-01 -4.74872589e-01 -7.23913848e-01
4.34558183e-01 -5.74607909e-01 -2.90828556e-01 -1.21128969e-01
-2.95419127e-01 -1.89142227e-01 1.10552084e+00 -7.52506077e-01
1.39110887e+00 -2.45708942e+00 3.08980048e-01 3.66114795e-01
3.41905206e-01 3.87676418e-01 1.82446435e-01 -3.89147587e-02
-1.72960445e-01 -3.59949440e-01 -7.33245134e-01 -9.16637003e-01
1.77038714e-01 2.46818736e-01 -3.88188839e-01 8.80903065e-01
4.58886713e-01 4.88292813e-01 -6.24360621e-01 -2.08720177e-01
4.62895751e-01 7.35522687e-01 -8.47852111e-01 4.63276327e-01
-8.34763944e-02 3.51492137e-01 -2.14478523e-01 5.38550258e-01
8.10934305e-01 -2.71774471e-01 -5.64549444e-03 1.21089466e-01
3.11033458e-01 8.83845985e-02 -1.48603284e+00 1.78902864e+00
-2.85835236e-01 2.84428924e-01 1.21818490e-01 -1.41215634e+00
1.06097817e+00 2.25484103e-01 6.76011980e-01 -5.16235471e-01
1.55632928e-01 2.80849099e-01 -5.82813546e-02 -1.17761202e-01
2.48753011e-01 -4.08279970e-02 -6.96075708e-02 5.50181150e-01
4.60704327e-01 3.02813083e-01 -4.29136790e-02 4.58610877e-02
8.83807898e-01 -1.41142709e-02 1.09648950e-01 -1.62003059e-02
5.96370995e-01 -5.92312753e-01 5.84518075e-01 4.78003561e-01
-5.89358527e-03 5.43369710e-01 5.32691479e-01 -3.29433113e-01
-1.19286728e+00 -1.28615332e+00 -3.17839473e-01 7.02114522e-01
-2.85839468e-01 -3.53426099e-01 -8.18588912e-01 -7.78134525e-01
4.14772220e-02 8.86750162e-01 -4.69215572e-01 -6.68203175e-01
-4.74585831e-01 -1.20396650e+00 5.23295403e-01 4.79435116e-01
4.53992248e-01 -8.09862554e-01 -2.72334963e-01 2.55329043e-01
8.50719884e-02 -1.03407907e+00 -1.13915719e-01 2.52207577e-01
-1.17429709e+00 -6.28562808e-01 -6.86049700e-01 -3.70056957e-01
7.39988744e-01 -2.62507349e-01 7.88690209e-01 -3.47164392e-01
-2.57505387e-01 1.97877854e-01 9.92096514e-02 -1.81588650e-01
-2.27157831e-01 -2.49842145e-02 5.12786984e-01 4.21089530e-01
8.69351327e-01 -1.03852642e+00 -5.71229696e-01 1.50400132e-01
-1.07118225e+00 -4.59519356e-01 7.06570327e-01 9.24061358e-01
6.81986988e-01 2.00610310e-01 5.33119321e-01 -4.46384400e-01
4.46316808e-01 -9.35082555e-01 -7.81400681e-01 -3.76293868e-01
-7.13987768e-01 3.07128578e-01 6.16047561e-01 -3.66487235e-01
-9.65067744e-01 1.16298487e-02 -4.93145764e-01 -1.12252200e+00
-4.59575117e-01 2.81020463e-01 -3.19157153e-01 3.89352322e-01
4.61092770e-01 5.64123750e-01 3.57700914e-01 -8.17852139e-01
2.18301401e-01 6.60932899e-01 6.31104589e-01 -4.36301082e-01
6.40933156e-01 3.86938155e-01 -1.75916418e-01 -9.30971146e-01
-5.91158628e-01 -5.41178942e-01 -5.88381588e-01 1.62759662e-01
7.62004673e-01 -1.08753192e+00 -4.37843978e-01 8.49556565e-01
-7.62840092e-01 -9.35717523e-02 -4.42630231e-01 7.33944118e-01
-5.30301571e-01 6.92071438e-01 -5.84644318e-01 -1.01053679e+00
-3.89622986e-01 -9.64780867e-01 9.43233967e-01 1.42585948e-01
-2.27703545e-02 -1.06378472e+00 3.07909876e-01 -1.65680964e-02
3.71143013e-01 -9.45581198e-02 9.25254226e-01 -1.05160296e+00
-2.19196156e-01 -5.50702691e-01 -4.38043252e-02 8.37082803e-01
3.07758097e-02 -2.58811295e-01 -9.87288535e-01 -2.58004665e-01
4.77138042e-01 -2.60271668e-01 9.14584816e-01 4.10604268e-01
1.58742499e+00 -3.36297929e-01 -5.71008250e-02 8.97698998e-01
1.21184170e+00 -8.33097696e-02 7.68466473e-01 2.22464308e-01
8.66600871e-01 2.93966442e-01 1.92504779e-01 6.97843075e-01
2.30329379e-01 3.47043067e-01 6.31819844e-01 2.51450062e-01
3.95158470e-01 -1.70605615e-01 4.71040308e-01 9.37237680e-01
4.45030451e-01 -5.25120459e-02 -8.87908161e-01 4.95272666e-01
-1.91671288e+00 -7.96844959e-01 2.99230725e-01 2.42990208e+00
5.56673408e-01 7.94065073e-02 3.41428429e-01 1.55492648e-01
5.97628653e-01 1.07623503e-01 -8.49050879e-01 -1.94747403e-01
6.18808605e-02 -4.49692644e-02 1.09093366e-02 2.40550637e-01
-1.34045827e+00 7.10509479e-01 5.40847397e+00 9.36366022e-01
-1.08387518e+00 -6.58902898e-02 6.52076125e-01 -1.06257841e-01
1.62043676e-01 -3.05137277e-01 -7.24920213e-01 8.06907356e-01
1.31242394e+00 2.52588153e-01 3.37071538e-01 1.22627556e+00
-1.05403267e-01 -4.32548299e-02 -8.14087451e-01 1.23185718e+00
-1.20358251e-01 -9.07070220e-01 6.80156574e-02 5.02095342e-01
3.89841139e-01 2.46262446e-01 2.57893443e-01 5.58651209e-01
2.33104184e-01 -8.37822795e-01 3.09451371e-01 6.38794184e-01
3.21632355e-01 -1.15125060e+00 7.67113984e-01 6.68024063e-01
-7.15979993e-01 -2.40110219e-01 -6.56595886e-01 2.25337252e-01
7.37683326e-02 9.55503643e-01 -4.91026878e-01 1.61192238e-01
6.04165375e-01 5.06538391e-01 -2.04486668e-01 9.04342294e-01
9.92065594e-02 8.08781862e-01 -6.22588336e-01 3.84953499e-01
2.11508900e-01 -6.38774514e-01 7.35339046e-01 1.04396474e+00
4.39700484e-01 -3.38021576e-01 -1.82609465e-02 1.15466309e+00
-5.08820787e-02 3.96425761e-02 -4.16250497e-01 -3.49760443e-01
3.63559753e-01 1.22683644e+00 -3.28194052e-01 -2.35706002e-01
-4.59989369e-01 1.17931390e+00 6.17310047e-01 3.15504313e-01
-7.65371799e-01 -3.22749197e-01 9.62763429e-01 3.90250161e-02
5.93011677e-01 -3.43813181e-01 -5.26026338e-02 -1.43790960e+00
1.46874711e-01 -7.55463004e-01 4.70429003e-01 -8.25718231e-03
-1.49618566e+00 3.92736077e-01 -1.77911758e-01 -1.15926516e+00
-8.11949372e-01 -5.58668375e-01 -7.71817684e-01 9.47340250e-01
-1.18439198e+00 -5.59939206e-01 4.61144932e-02 5.06893039e-01
3.50180328e-01 -5.36632657e-01 9.77089524e-01 4.88130271e-01
-1.27041173e+00 6.85641766e-01 6.06896043e-01 3.01450908e-01
4.59512919e-01 -1.39962339e+00 4.65048164e-01 1.00851691e+00
1.16098560e-01 4.92878169e-01 3.85417670e-01 -4.93099481e-01
-1.17776585e+00 -1.16246450e+00 2.22622871e-01 -1.41166449e-01
7.02287853e-01 -3.91395807e-01 -1.44279504e+00 4.55734521e-01
-2.58722901e-01 3.14283311e-01 9.65221822e-01 9.47491974e-02
-2.59934962e-01 8.63932148e-02 -1.11323035e+00 3.74336302e-01
4.53313947e-01 -5.66276670e-01 -3.11817586e-01 -6.40493557e-02
2.99629599e-01 -2.80413687e-01 -8.79191458e-01 2.83551574e-01
2.28248730e-01 -1.01254904e+00 9.71779048e-01 -5.66172898e-01
3.05806994e-01 -2.17041925e-01 -2.99213499e-01 -1.24898660e+00
-3.53893280e-01 -3.29984426e-01 -1.31414092e+00 1.18871951e+00
8.82451534e-02 -6.46936953e-01 8.94899428e-01 6.69413090e-01
-1.17287755e-01 -1.05735838e+00 -9.72227514e-01 -5.94106019e-01
1.22179590e-01 -6.62249565e-01 4.57243115e-01 7.61012435e-01
-1.91662684e-01 2.78013647e-01 -4.11764890e-01 4.03602749e-01
9.01725709e-01 -3.25385839e-01 7.80419409e-01 -1.34113133e+00
-4.85291392e-01 -6.25089407e-01 -6.91464722e-01 -1.20679569e+00
2.91350931e-01 -5.80251694e-01 -1.62701756e-02 -1.08004475e+00
-9.33105871e-02 -1.03670217e-01 -5.49327791e-01 1.89352304e-01
-2.05597594e-01 -4.40256074e-02 -3.73730958e-01 2.84845263e-01
-6.76300645e-01 1.20001781e+00 3.93289179e-01 5.99043928e-02
-3.08649927e-01 3.60182207e-03 -6.29915178e-01 1.06992233e+00
7.78490365e-01 -5.62691033e-01 -3.66710901e-01 -3.29660684e-01
-1.77306533e-01 -3.56134087e-01 3.55282068e-01 -1.34267604e+00
2.27183163e-01 1.87180251e-01 7.43282437e-01 -5.01142800e-01
5.13470054e-01 -8.33748281e-01 -2.94472992e-01 1.29639179e-01
1.70903429e-01 -4.64875475e-02 3.21100503e-01 9.44071770e-01
-2.89705485e-01 -3.13721776e-01 9.07377124e-01 3.03960413e-01
-5.31316996e-01 6.01096928e-01 -4.45496827e-01 -6.38825968e-02
9.13647652e-01 -4.25844342e-02 2.49524996e-01 -2.63716996e-01
-7.87809014e-01 4.42736559e-02 4.60290968e-01 2.54937500e-01
6.88409090e-01 -1.52083766e+00 -6.02610886e-01 6.64893985e-01
-5.15748039e-02 5.01104593e-01 5.36156893e-01 1.00371659e+00
-3.10203463e-01 7.77347982e-02 6.97736740e-02 -8.95220876e-01
-4.02488023e-01 4.60657984e-01 5.22734940e-01 -3.56100887e-01
-9.61277604e-01 7.02451766e-01 2.70219922e-01 -4.20024455e-01
4.74315107e-01 1.36573985e-01 1.88449193e-02 -1.05683111e-01
7.12321699e-01 3.74382317e-01 1.15689412e-01 -6.32849455e-01
-1.86984152e-01 8.21895674e-02 -4.53461885e-01 -1.16386013e-02
1.40296829e+00 -1.91854000e-01 5.33418506e-02 4.71572399e-01
1.36808443e+00 -3.09275001e-01 -1.51259232e+00 -4.71865475e-01
1.34656588e-02 -3.07460666e-01 6.20363474e-01 -1.53010994e-01
-1.09507346e+00 1.13395941e+00 6.65816545e-01 5.97029850e-02
1.05730820e+00 -1.62663132e-01 9.98994112e-01 5.79273462e-01
-2.74367213e-01 -1.16943634e+00 2.62102693e-01 5.96146107e-01
4.87416685e-01 -1.19472373e+00 -2.00184330e-01 2.82672018e-01
-5.95659852e-01 9.62312639e-01 8.52148652e-01 -5.09535611e-01
9.03484881e-01 2.78147280e-01 -1.00905970e-01 -2.03493699e-01
-5.44004083e-01 1.61537543e-01 4.11808938e-01 3.39788467e-01
1.31402433e-01 -2.86955893e-01 5.38980126e-01 9.40721810e-01
1.42272055e-01 -4.42903847e-01 -5.47437891e-02 6.76359057e-01
-5.95557749e-01 -8.53446484e-01 -2.89744169e-01 7.13600934e-01
-5.30473530e-01 6.80603236e-02 4.60795611e-02 4.07949567e-01
-2.16857985e-01 4.13518012e-01 6.35522366e-01 -2.86959946e-01
2.35650055e-02 4.85084832e-01 4.39030305e-02 -3.22398722e-01
1.48070216e-01 3.58429581e-01 -5.69358528e-01 -5.54308593e-01
2.71105111e-01 -8.12930465e-01 -1.22465324e+00 -2.75544703e-01
-4.00025815e-01 2.82520324e-01 7.82387495e-01 1.09347641e+00
7.18760073e-01 4.39818412e-01 5.98979235e-01 -8.03520203e-01
-9.52053130e-01 -1.15093791e+00 -7.91209519e-01 2.44283944e-01
4.94151980e-01 -6.84531450e-01 -5.49226463e-01 -3.21004421e-01] | [7.638636112213135, 2.359224557876587] |
99cb5e8a-abaa-4dc7-bcae-fdad81cdbda5 | towards-explainable-collaborative-filtering | 2304.13937 | null | https://arxiv.org/abs/2304.13937v1 | https://arxiv.org/pdf/2304.13937v1.pdf | Towards Explainable Collaborative Filtering with Taste Clusters Learning | Collaborative Filtering (CF) is a widely used and effective technique for recommender systems. In recent decades, there have been significant advancements in latent embedding-based CF methods for improved accuracy, such as matrix factorization, neural collaborative filtering, and LightGCN. However, the explainability of these models has not been fully explored. Adding explainability to recommendation models can not only increase trust in the decisionmaking process, but also have multiple benefits such as providing persuasive explanations for item recommendations, creating explicit profiles for users and items, and assisting item producers in design improvements. In this paper, we propose a neat and effective Explainable Collaborative Filtering (ECF) model that leverages interpretable cluster learning to achieve the two most demanding objectives: (1) Precise - the model should not compromise accuracy in the pursuit of explainability; and (2) Self-explainable - the model's explanations should truly reflect its decision-making process, not generated from post-hoc methods. The core of ECF is mining taste clusters from user-item interactions and item profiles.We map each user and item to a sparse set of taste clusters, and taste clusters are distinguished by a few representative tags. The user-item preference, users/items' cluster affiliations, and the generation of taste clusters are jointly optimized in an end-to-end manner. Additionally, we introduce a forest mechanism to ensure the model's accuracy, explainability, and diversity. To comprehensively evaluate the explainability quality of taste clusters, we design several quantitative metrics, including in-cluster item coverage, tag utilization, silhouette, and informativeness. Our model's effectiveness is demonstrated through extensive experiments on three real-world datasets. | ['Xing Xie', 'Yunjun Gao', 'Lu Chen', 'Mingqi Wu', 'Xiting Wang', 'Jing Yao', 'Jianxun Lian', 'Yuntao Du'] | 2023-04-27 | null | null | null | null | ['collaborative-filtering'] | ['miscellaneous'] | [-1.96378157e-01 -2.52854247e-02 -6.36908650e-01 -7.33766079e-01
-3.73387218e-01 -4.48082000e-01 1.96384862e-01 2.03939602e-01
1.72471300e-01 3.28713983e-01 8.63132954e-01 -3.23178858e-01
-6.22684777e-01 -6.44343317e-01 -2.74296850e-01 -4.13635969e-01
-1.26831785e-01 3.36919308e-01 -2.99637735e-01 7.56166950e-02
2.57757187e-01 -5.09319343e-02 -1.52169955e+00 7.08381474e-01
1.36286962e+00 9.72013354e-01 1.24238238e-01 3.00344586e-01
-2.24162474e-01 6.64190471e-01 -9.07984972e-02 -6.04381442e-01
1.11857809e-01 -4.29028898e-01 -3.60187709e-01 5.62379509e-02
-1.18944727e-01 -3.33175100e-02 1.12182364e-01 8.78266513e-01
1.46601304e-01 3.60880375e-01 7.16318190e-01 -1.15442479e+00
-1.44800019e+00 1.28193402e+00 -4.73109365e-01 -3.22158724e-01
2.06094041e-01 -2.36215472e-01 1.68662441e+00 -1.10576403e+00
9.03882757e-02 1.23392773e+00 7.38101065e-01 3.91808152e-01
-1.07705486e+00 -7.43031979e-01 8.15470576e-01 8.20497349e-02
-1.12064254e+00 -7.07922727e-02 5.95662117e-01 -4.93212402e-01
5.94919086e-01 5.67534208e-01 8.33320916e-01 6.06711686e-01
-1.16579987e-01 9.17900741e-01 7.91643322e-01 -3.59319538e-01
4.82791573e-01 5.15125692e-01 4.16591734e-01 4.36339378e-01
3.87593895e-01 6.94087371e-02 -4.53789085e-01 -3.77967805e-01
5.58688879e-01 7.56075382e-01 -3.96539241e-01 -3.67476255e-01
-8.96476209e-01 1.26950383e+00 6.71086848e-01 1.75320148e-01
-4.06682432e-01 -6.70194700e-02 -1.81263864e-01 1.67983487e-01
5.97463310e-01 7.59803951e-01 -5.76778948e-01 1.67715698e-01
-7.04089046e-01 5.07074632e-02 7.35696733e-01 7.34435141e-01
5.55510581e-01 -1.07106343e-01 -1.14566274e-01 6.66765571e-01
9.00948405e-01 3.29953581e-01 5.59268892e-01 -7.19599307e-01
-2.02143984e-03 8.16134155e-01 2.51070589e-01 -1.22741854e+00
-2.27309898e-01 -7.91635096e-01 -8.61580014e-01 -1.08682692e-01
-8.27923492e-02 -1.05425917e-01 -5.63964307e-01 1.56829309e+00
2.85527259e-01 1.24590479e-01 -1.74983785e-01 1.09016347e+00
6.17669284e-01 4.81532216e-01 1.25986427e-01 -2.30710864e-01
1.03180146e+00 -1.07619131e+00 -6.28242612e-01 -1.20389402e-01
5.33237636e-01 -8.03632319e-01 1.15624964e+00 4.64300752e-01
-7.92385936e-01 -7.55868614e-01 -7.64403045e-01 1.86343193e-01
-2.25519831e-03 3.31158280e-01 1.27285266e+00 7.06158876e-01
-6.51863158e-01 5.62971830e-01 -5.11051834e-01 3.44912522e-02
3.97619516e-01 5.19539416e-01 5.36267571e-02 -1.12602606e-01
-1.11427629e+00 2.88852185e-01 -1.78723067e-01 -8.92551057e-03
-2.30308861e-01 -7.29766190e-01 -3.80021870e-01 4.32863712e-01
2.57719934e-01 -1.02891076e+00 1.02628613e+00 -1.05885530e+00
-1.21262753e+00 -1.04826398e-01 -2.55868047e-01 -1.58955812e-01
-5.44277914e-02 -3.28059167e-01 -7.73679614e-01 -5.58691919e-01
1.22306451e-01 3.60508680e-01 4.70698029e-01 -1.44620097e+00
-9.79289055e-01 -3.62030387e-01 1.27158716e-01 2.32792482e-01
-5.69879472e-01 -3.92356157e-01 -4.40329701e-01 -7.03360319e-01
2.44847059e-01 -1.02422988e+00 -5.98684251e-01 -1.09754734e-01
-2.42601424e-01 -3.12629074e-01 1.90439180e-01 -2.28803918e-01
1.57090068e+00 -2.25569463e+00 -1.39046624e-01 6.19288266e-01
5.02211690e-01 5.24765253e-02 -1.06623255e-01 3.01349312e-01
1.43152595e-01 3.77541125e-01 4.47353452e-01 -4.79952812e-01
2.81647414e-01 1.61576331e-01 -4.38102216e-01 7.23543689e-02
-2.52001792e-01 9.19377863e-01 -9.70879555e-01 2.79936474e-02
1.38723329e-01 4.71531093e-01 -1.24613738e+00 5.56656793e-02
-1.97629467e-01 2.24364400e-01 -7.10377693e-01 3.87654155e-01
3.67655963e-01 -6.20321333e-01 4.31973100e-01 -4.05761749e-01
-6.89230412e-02 3.23972672e-01 -1.37823439e+00 1.30294263e+00
-4.71194208e-01 1.52737275e-01 -2.80357957e-01 -4.32563752e-01
8.80065560e-01 4.96682711e-02 5.04567802e-01 -4.26936775e-01
-3.02339736e-02 5.75954542e-02 9.26505998e-02 -2.82948703e-01
7.24677920e-01 2.29375213e-02 2.20875181e-02 8.70612681e-01
-4.62608278e-01 7.59004295e-01 -1.41796678e-01 4.38484579e-01
6.55693889e-01 -2.62321293e-01 2.76846051e-01 -2.33324602e-01
3.80965583e-02 -6.99537322e-02 6.85024858e-01 8.10810447e-01
2.20248610e-01 4.99062568e-01 -1.55100882e-01 -4.08568799e-01
-6.40061021e-01 -8.58125627e-01 1.38888866e-01 1.27323949e+00
3.18595290e-01 -7.36548901e-01 -1.14725262e-01 -8.14782381e-01
4.12195742e-01 8.35480511e-01 -9.35741782e-01 -3.03926438e-01
1.74657613e-01 -5.40650368e-01 -8.15286413e-02 7.91539967e-01
-1.51298987e-02 -6.45673335e-01 1.37291104e-01 1.82413816e-01
-2.69452661e-01 -3.35902184e-01 -1.02461088e+00 1.47315651e-01
-7.31421232e-01 -1.03280342e+00 -1.43891409e-01 -5.90093017e-01
8.69230330e-01 9.13109004e-01 1.12489414e+00 4.29252565e-01
3.49143177e-01 9.17660818e-02 -8.22902977e-01 -3.51851493e-01
8.29154179e-02 -8.15594941e-02 4.70719755e-01 1.56333655e-01
6.45672560e-01 -4.79327112e-01 -9.33769286e-01 7.92819083e-01
-6.51194155e-01 7.48138269e-03 5.65316439e-01 7.24883795e-01
6.42372370e-01 2.70231605e-01 6.86211765e-01 -1.18305290e+00
8.94325197e-01 -7.32800841e-01 3.37163714e-04 3.33991885e-01
-1.47548532e+00 8.21636617e-02 6.75024807e-01 -6.44018054e-01
-9.64721680e-01 -1.85964644e-01 -1.77601680e-01 -2.28708118e-01
1.56249166e-01 8.80298615e-01 -1.97458714e-01 2.88657069e-01
6.93219543e-01 -2.11858869e-01 -1.66322634e-01 -7.62988627e-01
8.76638651e-01 8.96197081e-01 1.59407243e-01 -3.73920709e-01
6.14623964e-01 2.95691907e-01 -6.81443155e-01 -1.05720542e-01
-1.21519363e+00 -9.00833368e-01 -2.91002125e-01 -2.56103813e-04
3.76738757e-01 -1.03638148e+00 -8.10213745e-01 -5.05883336e-01
-5.15009999e-01 1.01323068e-01 -3.96890551e-01 8.32082152e-01
3.30473669e-02 3.84585857e-01 -2.59281307e-01 -9.75104511e-01
-4.41822052e-01 -9.16256785e-01 5.20045936e-01 3.26687723e-01
-6.44091010e-01 -1.08189034e+00 1.41247511e-01 5.21615446e-01
4.08621043e-01 -2.45849624e-01 1.00722480e+00 -6.89552128e-01
-3.93721342e-01 -8.96399617e-02 -1.43083572e-01 6.44799098e-02
4.06147540e-01 -5.21047264e-02 -7.04342961e-01 -3.12935174e-01
-3.09523165e-01 3.44944894e-02 8.84891391e-01 5.91857851e-01
9.36519504e-01 -5.79510450e-01 -4.63729978e-01 5.26317477e-01
1.14342344e+00 3.02437842e-01 3.09943318e-01 -8.00873488e-02
7.22310603e-01 4.86845136e-01 7.31617212e-01 5.59509039e-01
6.27156377e-01 6.12582088e-01 3.55677575e-01 -2.12955549e-01
2.16256678e-02 -6.16895318e-01 3.36780250e-01 1.04860938e+00
-3.29200327e-02 -3.95298421e-01 -2.15631709e-01 3.20781201e-01
-2.32176232e+00 -1.00750148e+00 -3.65480155e-01 2.07911801e+00
3.87692899e-01 -7.28365406e-02 2.84311444e-01 9.62616727e-02
4.79009390e-01 -4.24578488e-01 -6.53546095e-01 -2.97901809e-01
3.89758050e-02 -3.42320979e-01 4.73913476e-02 5.01410782e-01
-7.17143178e-01 6.73801184e-01 5.59000826e+00 4.66731310e-01
-7.08714128e-01 2.38467548e-02 5.79997063e-01 -2.22653255e-01
-1.17325616e+00 8.73960629e-02 -7.62099564e-01 5.53930223e-01
4.45758730e-01 -1.09147571e-01 6.08315289e-01 1.04909182e+00
4.39913779e-01 3.39232594e-01 -1.02531314e+00 7.45307684e-01
-3.96712199e-02 -1.50897765e+00 2.21538901e-01 3.15512925e-01
1.06998241e+00 -1.76675782e-01 4.03779298e-01 4.43470240e-01
8.46118569e-01 -7.94066131e-01 7.04656839e-01 5.13811290e-01
4.77671057e-01 -8.71746838e-01 6.93192959e-01 2.24533305e-01
-1.28617048e+00 -5.07002890e-01 -6.08174086e-01 -3.05407375e-01
1.82727531e-01 9.79418755e-01 -5.47015905e-01 5.29408276e-01
6.45305514e-01 9.47868764e-01 -1.55596614e-01 1.03937280e+00
-2.42672876e-01 8.61744583e-01 -2.96654794e-02 -3.03608239e-01
1.85365021e-01 -3.18162084e-01 -8.66603553e-02 9.39933062e-01
4.61358249e-01 4.03527945e-01 3.86523873e-01 8.57392967e-01
-3.49511318e-02 3.99681062e-01 -8.54950100e-02 -2.36719921e-01
8.81392002e-01 1.26997697e+00 -5.45969009e-01 -2.83926595e-02
-3.96267146e-01 6.34688079e-01 9.11876410e-02 2.35165089e-01
-7.09441364e-01 -3.13351825e-02 1.14624536e+00 1.18708231e-01
4.89225417e-01 6.13429435e-02 -6.36090875e-01 -1.36697638e+00
-4.11462605e-01 -8.73356879e-01 4.35618013e-01 -5.49210787e-01
-1.66480410e+00 5.05234480e-01 -6.24900341e-01 -1.21500540e+00
-6.77311569e-02 -2.42005885e-01 -6.55090749e-01 8.93771768e-01
-1.13199747e+00 -1.12940443e+00 -3.08927745e-01 5.15573740e-01
4.18183476e-01 -2.08337694e-01 8.95771503e-01 3.82389039e-01
-3.99126112e-01 8.02977979e-01 4.93279725e-01 -1.89923868e-01
6.79576933e-01 -1.22425377e+00 3.16050947e-01 5.80580413e-01
6.07755899e-01 1.27235520e+00 6.12499774e-01 -7.36144602e-01
-1.26790333e+00 -1.16084540e+00 9.62306738e-01 -6.60423279e-01
4.67253506e-01 -2.00796515e-01 -7.11590052e-01 4.12567794e-01
-4.24863070e-01 -4.53583479e-01 1.60129201e+00 1.17961907e+00
-4.23328310e-01 -1.72285270e-02 -8.40452611e-01 5.12145579e-01
9.57809389e-01 -2.69951671e-01 -2.51090556e-01 2.58875459e-01
7.37583280e-01 2.15154424e-01 -9.94136870e-01 -5.98730566e-03
1.00593162e+00 -8.37261379e-01 1.00890517e+00 -8.23026001e-01
3.90550375e-01 -5.66219091e-01 -3.58868778e-01 -1.53807712e+00
-1.34328568e+00 -3.67026687e-01 -2.78111339e-01 1.24289703e+00
7.13640392e-01 -2.68719018e-01 8.85349751e-01 8.89782488e-01
-2.13857308e-01 -1.06975126e+00 2.15677428e-04 -3.29439044e-01
-3.20438713e-01 -5.12739956e-01 1.17053998e+00 1.10805547e+00
3.12855601e-01 6.06674433e-01 -7.28513300e-01 2.91614711e-01
4.96749490e-01 7.58528888e-01 7.45006561e-01 -1.62269628e+00
-6.95095956e-01 -3.62240493e-01 2.46983513e-01 -1.34074306e+00
-3.11816484e-01 -9.72928464e-01 -1.49804652e-01 -1.87534094e+00
3.98009419e-01 -1.04393089e+00 -6.29119933e-01 5.78214705e-01
-4.61887151e-01 1.08579598e-01 1.30105451e-01 7.75765300e-01
-9.60290849e-01 4.20093030e-01 1.11589181e+00 6.11042604e-02
-4.67560530e-01 3.42573315e-01 -1.70095670e+00 5.82886159e-01
5.50320208e-01 -3.93917948e-01 -8.31499815e-01 -4.08185989e-01
7.08122671e-01 -2.73501992e-01 -2.08159104e-01 -4.59009647e-01
2.53238559e-01 -2.88833648e-01 5.75082123e-01 -3.01353753e-01
2.70428974e-02 -8.12915266e-01 5.52194893e-01 2.77782261e-01
-6.69045508e-01 1.22400671e-01 -4.81230795e-01 1.03992629e+00
-3.37565457e-03 1.88510507e-01 2.61539787e-01 1.16177849e-01
-5.50298750e-01 4.59280133e-01 -1.82349592e-01 -3.99441063e-01
5.86929083e-01 -2.23642185e-01 -1.07117109e-01 -6.34644270e-01
-7.95135081e-01 3.74022067e-01 4.84305739e-01 7.02421784e-01
5.27952075e-01 -1.53175831e+00 -5.23985565e-01 2.82150477e-01
2.76747227e-01 -4.46993262e-01 3.40519398e-01 6.68140411e-01
3.29105288e-01 4.82659847e-01 2.74045378e-01 -2.56891489e-01
-1.13661110e+00 6.36736155e-01 -1.60340384e-01 -2.15494946e-01
-3.01562935e-01 9.59694922e-01 3.27936381e-01 -4.88202542e-01
2.94627637e-01 -3.53424370e-01 -4.25403744e-01 -9.34734717e-02
7.11812377e-01 2.30296150e-01 -2.09496319e-01 -2.67039716e-01
-1.62742004e-01 2.34172001e-01 -2.65306234e-01 3.73000205e-01
1.36832452e+00 -4.68808323e-01 2.69167185e-01 2.06133798e-01
6.96085274e-01 3.68338317e-01 -1.13033187e+00 -3.78855139e-01
-2.15308324e-01 -6.55445755e-01 3.34367216e-01 -1.21679723e+00
-1.13580418e+00 5.40654242e-01 4.79910463e-01 3.75353426e-01
8.69259238e-01 -2.56072115e-02 8.31533670e-01 4.96560149e-02
4.00833398e-01 -7.93627501e-01 7.59778172e-02 1.07161246e-01
3.93480539e-01 -1.08481014e+00 1.10340863e-02 -3.54634434e-01
-9.14106727e-01 7.72260487e-01 3.82146776e-01 3.05524707e-01
8.54145527e-01 -1.09610543e-01 1.19802952e-01 -1.49937168e-01
-8.88284206e-01 -1.34637654e-01 7.53342509e-01 4.20381427e-01
6.96842670e-01 4.80157256e-01 -1.72999054e-01 1.60059583e+00
-1.66417345e-01 7.43858069e-02 4.28022258e-03 3.45505953e-01
-7.03295887e-01 -1.13012815e+00 4.45853285e-02 8.76864851e-01
-2.49275297e-01 -2.63472646e-01 -5.79883575e-01 2.70747453e-01
3.57195824e-01 1.33343971e+00 -2.16363356e-01 -1.09749782e+00
1.46813110e-01 -4.53705490e-01 -1.04512367e-02 -7.54949450e-01
-6.53493583e-01 1.55182868e-01 1.05248531e-02 -4.36962992e-01
-2.33760014e-01 -5.63454390e-01 -1.16220593e+00 -6.05545938e-01
-9.91858304e-01 9.42078054e-01 5.89799762e-01 8.98617089e-01
8.53054762e-01 3.62006426e-01 9.87058759e-01 -4.00182098e-01
-5.29249668e-01 -6.57811344e-01 -7.33237684e-01 5.26302874e-01
-1.92657515e-01 -6.84410453e-01 -4.98967052e-01 -1.32288396e-01] | [9.895395278930664, 5.637152194976807] |
e264a585-452f-4834-b8fb-e0e5bbeaae2c | outline-then-details-syntactically-guided | 2305.00909 | null | https://arxiv.org/abs/2305.00909v3 | https://arxiv.org/pdf/2305.00909v3.pdf | Outline, Then Details: Syntactically Guided Coarse-To-Fine Code Generation | For a complicated algorithm, its implementation by a human programmer usually starts with outlining a rough control flow followed by iterative enrichments, eventually yielding carefully generated syntactic structures and variables in a hierarchy. However, state-of-the-art large language models generate codes in a single pass, without intermediate warm-ups to reflect the structured thought process of "outline-then-detail". Inspired by the recent success of chain-of-thought prompting, we propose ChainCoder, a program synthesis language model that generates Python code progressively, i.e. from coarse to fine in multiple passes. We first decompose source code into layout frame components and accessory components via abstract syntax tree parsing to construct a hierarchical representation. We then reform our prediction target into a multi-pass objective, each pass generates a subsequence, which is concatenated in the hierarchy. Finally, a tailored transformer architecture is leveraged to jointly encode the natural language descriptions and syntactically aligned I/O data samples. Extensive evaluations show that ChainCoder outperforms state-of-the-arts, demonstrating that our progressive generation eases the reasoning procedure and guides the language model to generate higher-quality solutions. Our codes are available at: https://github.com/VITA-Group/ChainCoder. | ['Zhangyang Wang', 'Dejia Xu', 'Yihan Xi', 'Kevin Wang', 'Ajay Kumar Jaiswal', 'S P Sharan', 'Wenqing Zheng'] | 2023-04-28 | null | null | null | null | ['program-synthesis'] | ['computer-code'] | [ 5.46072237e-02 2.26241961e-01 -3.55164498e-01 -5.02023935e-01
-9.32837009e-01 -8.12999725e-01 5.11779845e-01 2.69642293e-01
2.61361510e-01 2.23236352e-01 6.55106723e-01 -8.24548721e-01
3.97291273e-01 -8.95514965e-01 -1.00108325e+00 1.18662966e-02
2.10145861e-01 3.04131240e-01 8.82631913e-03 -2.04065159e-01
4.00012195e-01 1.11650251e-01 -1.47847414e+00 8.57079029e-01
1.05947721e+00 3.30440909e-01 2.72058487e-01 7.34547794e-01
-6.25394404e-01 1.31963658e+00 -2.50526518e-01 -7.15177715e-01
1.05269171e-01 -4.96422499e-01 -1.10950089e+00 2.30816349e-01
1.49657741e-01 -3.54688436e-01 -4.06952798e-02 8.90469730e-01
2.74425671e-02 -2.14752868e-01 2.03496948e-01 -1.03643548e+00
-7.69879937e-01 1.29297268e+00 -7.02229083e-01 -2.96454579e-01
5.05866826e-01 5.10248005e-01 1.49063444e+00 -1.04066896e+00
6.64089203e-01 1.17036009e+00 6.54389620e-01 4.63520139e-01
-1.64488173e+00 -4.89892602e-01 2.95014828e-01 -2.96057463e-01
-1.32481062e+00 -4.84759212e-01 6.42075837e-01 -7.99941421e-01
1.34850144e+00 1.69273093e-01 7.24595368e-01 8.47993493e-01
2.76943058e-01 1.07080507e+00 5.59656620e-01 -3.69211376e-01
2.12506726e-01 -1.16130136e-01 2.60898411e-01 1.05567384e+00
-1.61923619e-03 -3.44497487e-02 -3.37272495e-01 -2.89886057e-01
5.17996848e-01 5.39005473e-02 -8.03214312e-03 -4.67188567e-01
-1.07053041e+00 8.13990116e-01 4.95841235e-01 1.23462431e-01
-2.43614644e-01 3.72040927e-01 5.60312510e-01 1.44691631e-01
1.93892326e-02 6.07109249e-01 -4.86109525e-01 -2.67534167e-01
-1.43624699e+00 6.38238311e-01 1.04136813e+00 1.46232057e+00
9.49957490e-01 -1.46933988e-01 -3.92948151e-01 6.97079122e-01
4.08170253e-01 5.57538383e-02 2.95156658e-01 -1.06630385e+00
7.09990859e-01 9.85414445e-01 -1.16668306e-01 -6.16070390e-01
-2.31331468e-01 -7.07542419e-01 -5.27656555e-01 1.10094574e-04
2.16323152e-01 5.70558906e-02 -5.92498541e-01 1.55660367e+00
1.09906539e-01 -3.91343206e-01 -1.13836005e-01 6.76619470e-01
4.36687022e-01 8.02345932e-01 7.98453987e-02 1.91790551e-01
1.57846761e+00 -1.52822459e+00 -2.65125155e-01 -5.04509568e-01
9.83826160e-01 -7.21362948e-01 1.52109683e+00 5.10844648e-01
-1.56909370e+00 -5.61775744e-01 -8.44035804e-01 -6.45274222e-01
1.89848170e-01 2.56788194e-01 6.99126720e-01 2.39984244e-01
-1.06785810e+00 4.41193670e-01 -8.55643094e-01 -3.06788925e-02
4.81105626e-01 -5.91556244e-02 9.46883187e-02 -5.51865548e-02
-5.57742536e-01 3.44767004e-01 3.82547140e-01 -1.65380940e-01
-1.18757784e+00 -1.26098144e+00 -1.00337827e+00 2.73587883e-01
4.77283359e-01 -1.09791076e+00 1.88148415e+00 -8.10058117e-01
-1.48834205e+00 9.11458790e-01 -4.31337804e-01 -4.47541684e-01
4.53031301e-01 -3.81279588e-01 1.93021834e-01 -3.61132354e-01
1.72189713e-01 6.35967731e-01 5.92684150e-01 -1.27122867e+00
-7.33331680e-01 -7.43019655e-02 2.98899204e-01 -1.01049237e-01
1.95347751e-03 3.21702212e-01 -8.80672991e-01 -6.74263477e-01
-1.48514077e-01 -8.11412513e-01 -4.07655627e-01 -1.66334406e-01
-6.53142989e-01 -8.34595114e-02 7.61988387e-02 -7.45117128e-01
1.96973920e+00 -2.21683145e+00 5.81963181e-01 1.32973477e-01
5.56670547e-01 -3.12066942e-01 -1.52864456e-01 5.77519774e-01
-1.82207599e-01 2.98555464e-01 -4.72034067e-01 -4.86232907e-01
4.50821102e-01 -5.93744479e-02 -7.47894824e-01 1.10726088e-01
2.56920755e-01 1.06514359e+00 -1.06682324e+00 -4.61257756e-01
-9.02438164e-02 1.17751554e-01 -1.49536705e+00 5.33329546e-01
-8.48341346e-01 1.16845302e-01 -3.88146222e-01 6.56754553e-01
3.56167495e-01 -6.21474266e-01 2.65679002e-01 -1.29587770e-01
-4.27076191e-01 7.82169938e-01 -7.68373072e-01 2.20866084e+00
-9.24588680e-01 4.23944592e-01 6.76860511e-02 -4.82304603e-01
1.01920295e+00 -8.96104649e-02 -1.49351060e-02 -4.00868684e-01
-8.04312229e-02 2.67241538e-01 -7.40605593e-02 -4.89995062e-01
7.31168866e-01 1.73034608e-01 -5.26285887e-01 5.71207821e-01
-1.78441107e-01 -5.91460466e-01 4.32814777e-01 5.33826590e-01
1.22482145e+00 6.64125085e-01 4.55172390e-01 -2.50872970e-01
5.20339668e-01 3.26783806e-01 5.46075463e-01 7.27092624e-01
3.82996827e-01 5.19452512e-01 1.04075015e+00 -4.58951026e-01
-1.50067854e+00 -9.51872170e-01 2.15249673e-01 1.38331604e+00
-4.17027116e-01 -1.30915225e+00 -1.05552471e+00 -4.91990477e-01
-6.60925508e-02 1.09907019e+00 -4.10971284e-01 7.28252754e-02
-8.39011610e-01 -3.41967821e-01 5.94693840e-01 6.36691689e-01
2.44980454e-01 -8.24448049e-01 -8.14359903e-01 3.65170062e-01
1.64865497e-02 -7.51618981e-01 -7.17790246e-01 2.98446417e-01
-7.15082407e-01 -7.04801738e-01 -2.48040304e-01 -7.77065575e-01
8.06327403e-01 -1.41324922e-01 1.73325813e+00 5.54274380e-01
-2.41088390e-01 -1.41314954e-01 -3.13405633e-01 1.46502465e-01
-9.63770032e-01 3.61894459e-01 -6.26179695e-01 -4.51548606e-01
1.56042412e-01 -6.94320858e-01 -4.32776958e-01 -3.51764224e-02
-8.60740900e-01 8.76724184e-01 6.68695271e-01 7.89145589e-01
6.06195688e-01 -1.64701939e-01 -2.86957204e-01 -1.12200463e+00
5.69667518e-01 -5.56956828e-01 -9.15426672e-01 2.64645815e-01
-2.58144438e-01 5.44309437e-01 1.08370721e+00 -9.14507657e-02
-1.01097727e+00 1.28659427e-01 -1.46346956e-01 -2.62136638e-01
-7.35023692e-02 7.07437336e-01 -1.62233919e-01 5.37793338e-01
8.52594137e-01 3.69869739e-01 -2.20590323e-01 -4.80040163e-01
7.24865079e-01 4.81095880e-01 6.44342721e-01 -1.31102037e+00
9.06275153e-01 -1.62301455e-02 -4.10463870e-01 -2.80809879e-01
-7.55809903e-01 -6.99787308e-03 -4.19029921e-01 2.06930310e-01
7.04428494e-01 -1.02909338e+00 -4.84589189e-01 2.74658084e-01
-1.40026093e+00 -9.67531741e-01 -3.00728917e-01 -3.70074213e-01
-6.00830793e-01 6.62295520e-02 -9.18448269e-01 -3.69825810e-01
-3.83145571e-01 -1.68088853e+00 1.24147272e+00 -3.25084627e-02
-7.50740588e-01 -6.69268429e-01 -8.54186863e-02 4.45475012e-01
4.63780671e-01 -1.76828966e-01 1.30513954e+00 -3.81710589e-01
-9.85797882e-01 2.26317689e-01 -2.43327230e-01 -2.34563649e-02
-1.94967255e-01 3.45384270e-01 -6.02409422e-01 1.02232331e-02
-2.91846693e-01 -3.24296385e-01 4.12430555e-01 -1.95202470e-01
1.54053164e+00 -6.07563794e-01 -2.59060562e-01 1.02900946e+00
1.38775742e+00 -5.74923567e-02 5.65253973e-01 4.12202209e-01
8.42868984e-01 4.57745284e-01 1.58488646e-01 6.18241906e-01
8.02566886e-01 4.86429036e-01 2.32560381e-01 7.38220811e-02
-1.64770573e-01 -8.35868955e-01 4.50466871e-01 9.28356230e-01
5.03875852e-01 7.96299279e-02 -1.44967759e+00 6.38880849e-01
-1.76173007e+00 -6.18315220e-01 -1.50576949e-01 1.69780958e+00
1.31918740e+00 2.06981555e-01 1.51666194e-01 -1.84655324e-01
2.14477524e-01 8.04895386e-02 -3.78553182e-01 -6.07824743e-01
4.44853276e-01 1.53838113e-01 3.03905964e-01 7.32453465e-01
-5.93583584e-01 1.24953091e+00 5.88788891e+00 7.76872516e-01
-1.02781963e+00 -4.55110520e-02 6.22116864e-01 -1.73301443e-01
-1.11769807e+00 6.17227495e-01 -9.49449539e-01 3.39690179e-01
9.79255140e-01 -4.69165444e-01 8.31156969e-01 9.67169940e-01
2.19941273e-01 2.10394070e-01 -1.50822616e+00 5.56792855e-01
-2.45141983e-01 -1.66709137e+00 2.81519443e-01 -9.69777778e-02
6.86459124e-01 -1.45462295e-02 -1.99346632e-01 5.48564494e-01
8.43803763e-01 -9.29656148e-01 1.33288717e+00 3.02238017e-01
5.42312205e-01 -5.17728567e-01 6.01591729e-02 4.94826406e-01
-1.29856074e+00 -3.00060630e-01 -3.19754370e-02 -1.99044585e-01
1.23006858e-01 5.75984895e-01 -6.51411772e-01 2.98226237e-01
5.16943514e-01 6.49782956e-01 -7.21665025e-01 6.76252127e-01
-5.97735047e-01 5.22311866e-01 8.35302100e-02 -1.24446929e-01
2.29270056e-01 -4.10968885e-02 2.84025908e-01 1.62748075e+00
2.57804483e-01 7.42480233e-02 3.82971913e-01 1.65438461e+00
-1.86884061e-01 9.84013081e-02 -2.33962476e-01 -1.01380281e-01
6.42466486e-01 1.13921583e+00 -5.31915784e-01 -5.36413789e-01
-6.33276343e-01 8.27348769e-01 6.82829201e-01 2.96242356e-01
-9.44483697e-01 -3.62387568e-01 4.52387780e-01 3.93212348e-01
3.84152919e-01 -2.61112571e-01 -7.08679616e-01 -1.33383405e+00
8.27042088e-02 -1.44954157e+00 1.20904408e-01 -9.39357221e-01
-7.06693351e-01 6.56650484e-01 6.01757467e-02 -9.78143334e-01
-3.36884946e-01 -3.39274406e-01 -5.90116084e-01 9.06580448e-01
-1.23859620e+00 -1.16545653e+00 -2.24529758e-01 9.88543183e-02
8.86935294e-01 -4.91554802e-03 6.78742349e-01 1.31925702e-01
-6.88082814e-01 6.70292974e-01 -3.25864941e-01 1.89739823e-01
2.11899042e-01 -1.36955929e+00 1.08006835e+00 1.19111872e+00
-1.61157206e-01 1.33022344e+00 8.45275521e-01 -6.21708453e-01
-1.85003710e+00 -1.06343293e+00 1.01673126e+00 -4.00443345e-01
1.22446370e+00 -7.14381635e-01 -9.11625683e-01 8.37222219e-01
2.61628956e-01 -8.93989801e-02 6.13091171e-01 9.06027630e-02
-8.10655177e-01 4.53784540e-02 -5.53326845e-01 8.27240348e-01
1.10602927e+00 -7.43676186e-01 -6.86876714e-01 9.38935429e-02
1.06142116e+00 -8.50143909e-01 -8.51198554e-01 -1.19129650e-01
4.88735855e-01 -9.71611798e-01 7.41948843e-01 -5.01107395e-01
1.39658356e+00 -6.07731640e-01 -2.55667567e-01 -1.03108311e+00
-4.71820414e-01 -1.08183277e+00 -1.50922701e-01 1.33838665e+00
7.82967746e-01 -1.38766840e-01 6.00986302e-01 6.37934446e-01
-5.74571729e-01 -9.11935806e-01 -2.90180147e-02 -3.73059809e-01
2.63881356e-01 -6.60475016e-01 9.20391917e-01 5.77726305e-01
4.26349759e-01 5.44783711e-01 -4.12899517e-02 -4.18099836e-02
5.74336171e-01 5.96968949e-01 1.12443638e+00 -7.28961825e-01
-9.74182606e-01 -7.12715030e-01 4.09472346e-01 -1.50396895e+00
2.27664277e-01 -1.26536310e+00 2.02533871e-01 -1.56422937e+00
4.41228569e-01 -5.04546881e-01 3.40240955e-01 7.54899442e-01
-1.71372250e-01 -3.18783820e-01 3.36197823e-01 2.86160529e-01
-6.24832869e-01 3.24285120e-01 1.01727092e+00 -1.81593716e-01
-3.38414848e-01 -3.88809174e-01 -1.24186122e+00 6.68702602e-01
5.07613957e-01 -4.30173039e-01 -4.12455350e-01 -9.06569541e-01
7.20224082e-01 4.23777193e-01 2.98222065e-01 -9.45463419e-01
2.63373256e-01 -2.09789470e-01 -8.85238275e-02 -4.05830085e-01
-2.79669046e-01 -6.09993160e-01 3.97580773e-01 4.12079453e-01
-8.19297016e-01 2.85275877e-01 2.69869000e-01 -1.11122712e-01
-3.67003269e-02 -3.85041177e-01 5.57940900e-01 -3.51663530e-01
-4.94314194e-01 1.53332889e-01 -3.07445526e-01 2.04560623e-01
7.07485259e-01 7.16156363e-02 -3.53325665e-01 9.15932283e-02
-4.10463423e-01 3.67064714e-01 9.79127407e-01 4.59991693e-01
1.99012876e-01 -9.64310825e-01 -6.69238091e-01 3.18397194e-01
1.98456466e-01 3.81314784e-01 -4.11245190e-02 5.39870679e-01
-8.28481078e-01 2.52917826e-01 1.51902750e-01 -5.41456938e-01
-8.94321263e-01 8.03189933e-01 2.40154117e-01 -5.74615359e-01
-6.96159482e-01 8.29719961e-01 4.52456176e-01 -4.73021597e-01
6.72757551e-02 -9.07441795e-01 4.19840246e-01 -2.39049017e-01
5.99823833e-01 -8.08111578e-02 -1.22354187e-01 -1.63803592e-01
-3.06721479e-01 4.00594115e-01 -3.17784160e-01 -3.80141437e-02
1.39525545e+00 4.81159575e-02 -6.32701099e-01 1.93188593e-01
1.15807581e+00 2.99224138e-01 -1.33185136e+00 -2.51039952e-01
1.90857142e-01 -1.58959344e-01 -1.06515072e-01 -7.62755632e-01
-8.15434456e-01 8.57684672e-01 -4.25792187e-01 2.43774019e-02
1.16142118e+00 1.53439268e-01 7.28414476e-01 2.86234438e-01
4.77724761e-01 -5.61962724e-01 1.50030583e-01 7.26570785e-01
9.48332667e-01 -6.27356291e-01 -1.35494277e-01 -5.19019485e-01
-5.46416402e-01 1.17111528e+00 6.83914304e-01 -8.01227465e-02
2.12601960e-01 9.88045871e-01 -3.10730010e-01 -5.77518418e-02
-1.20747256e+00 2.34061256e-01 4.30161692e-02 9.75975320e-02
9.30734813e-01 2.97544580e-02 -1.82574302e-01 1.07720709e+00
-6.87298536e-01 4.59532917e-01 6.39596641e-01 1.10195112e+00
-2.76547849e-01 -1.32343209e+00 -3.19195509e-01 2.89632410e-01
-3.49090397e-01 -5.66531539e-01 -6.24641106e-02 4.75468099e-01
2.21837801e-03 4.62928474e-01 4.37405296e-02 -3.76474679e-01
4.72759902e-01 3.44915092e-02 2.79302686e-01 -9.78335083e-01
-1.08024228e+00 1.44937485e-01 1.97425202e-01 -8.61607254e-01
2.96671957e-01 -6.39671564e-01 -1.44309223e+00 -5.22673726e-01
2.92742014e-01 1.90624252e-01 2.51329303e-01 6.25399649e-01
6.68130577e-01 9.51819658e-01 2.91365236e-01 -8.48146498e-01
-5.64241707e-01 -4.69888240e-01 1.00487970e-01 3.65097523e-01
4.17135030e-01 -1.12247467e-03 -1.65454730e-01 5.55168867e-01] | [7.726959228515625, 7.862443447113037] |
d5d3e75e-9e73-4f10-8e8d-c472a372edc8 | robustsense-defending-adversarial-attack-for | 2204.01560 | null | https://arxiv.org/abs/2204.01560v2 | https://arxiv.org/pdf/2204.01560v2.pdf | SecureSense: Defending Adversarial Attack for Secure Device-Free Human Activity Recognition | Deep neural networks have empowered accurate device-free human activity recognition, which has wide applications. Deep models can extract robust features from various sensors and generalize well even in challenging situations such as data-insufficient cases. However, these systems could be vulnerable to input perturbations, i.e. adversarial attacks. We empirically demonstrate that both black-box Gaussian attacks and modern adversarial white-box attacks can render their accuracies to plummet. In this paper, we firstly point out that such phenomenon can bring severe safety hazards to device-free sensing systems, and then propose a novel learning framework, SecureSense, to defend common attacks. SecureSense aims to achieve consistent predictions regardless of whether there exists an attack on its input or not, alleviating the negative effect of distribution perturbation caused by adversarial attacks. Extensive experiments demonstrate that our proposed method can significantly enhance the model robustness of existing deep models, overcoming possible attacks. The results validate that our method works well on wireless human activity recognition and person identification systems. To the best of our knowledge, this is the first work to investigate adversarial attacks and further develop a novel defense framework for wireless human activity recognition in mobile computing research. | ['Lihua Xie', 'Han Zou', 'Jianfei Yang'] | 2022-04-04 | null | null | null | null | ['person-identification'] | ['computer-vision'] | [ 4.83989418e-01 -6.79410324e-02 -7.69532397e-02 -3.45251933e-02
-5.40586531e-01 -6.12444341e-01 4.78695899e-01 -2.94981033e-01
-2.84948021e-01 8.68918955e-01 6.83497265e-02 -4.12706465e-01
2.82629002e-02 -9.39546645e-01 -8.82345796e-01 -8.79144073e-01
-1.73190340e-01 -3.59874934e-01 1.96606722e-02 -7.56779015e-02
-3.54308426e-01 5.68120539e-01 -1.10249245e+00 -5.88311031e-02
6.68414056e-01 1.14799702e+00 -8.65686059e-01 6.83194697e-01
6.18366063e-01 8.08015585e-01 -1.31119823e+00 -5.41191101e-01
3.80014271e-01 -1.03997707e-01 -2.53563553e-01 -4.82666016e-01
3.63405123e-02 -4.18574184e-01 -7.68005967e-01 1.06975901e+00
9.53529894e-01 -1.93198413e-01 3.66823047e-01 -1.69527614e+00
-5.13577700e-01 5.90198636e-01 -3.16027701e-01 2.64078881e-05
6.09659016e-01 2.66491443e-01 2.52562672e-01 -3.13877732e-01
-1.55684620e-01 7.80160666e-01 1.12716651e+00 9.42647278e-01
-7.83158302e-01 -1.18994009e+00 7.21754655e-02 -9.95005742e-02
-1.63742089e+00 -5.42029023e-01 8.58734787e-01 -7.27192163e-02
7.03238428e-01 6.43965721e-01 5.16206086e-01 2.16214490e+00
3.79944175e-01 7.88620770e-01 7.58663893e-01 -7.63530582e-02
4.89303887e-01 -8.82914960e-02 -1.57061238e-02 4.39739883e-01
6.25515461e-01 5.35260558e-01 -4.22259927e-01 -6.14255190e-01
3.21882546e-01 2.83555627e-01 -3.91692340e-01 -1.97059289e-02
-1.01422179e+00 3.79624158e-01 3.69781762e-01 2.00215161e-01
-3.82348418e-01 4.20439482e-01 4.56436932e-01 1.03856944e-01
6.21245019e-02 1.46493495e-01 -4.26746666e-01 -3.55599105e-01
-5.39335012e-01 1.56531930e-01 9.31084394e-01 7.67742634e-01
-1.00532807e-02 5.15729845e-01 -1.42249897e-01 4.34808135e-01
3.02953094e-01 1.18494070e+00 4.16774541e-01 -6.32755697e-01
5.76403379e-01 2.08193570e-01 7.38435313e-02 -1.28873324e+00
-4.35441285e-01 -6.84302390e-01 -1.34811509e+00 -6.91231713e-02
4.58400518e-01 -8.78852844e-01 -4.67933983e-01 1.96309173e+00
4.43273969e-02 8.57405543e-01 2.30425850e-01 4.57306266e-01
5.38882375e-01 3.24401081e-01 3.01096261e-01 -6.71723559e-02
1.01867795e+00 -3.72874767e-01 -7.66793072e-01 -2.90704250e-01
4.37535763e-01 -1.81567803e-01 6.87983871e-01 6.04996264e-01
-7.74334729e-01 -4.61472720e-01 -1.49084353e+00 7.61690259e-01
-3.89013529e-01 -2.42787600e-01 8.41225207e-01 1.84311640e+00
-5.76312244e-01 1.95834473e-01 -1.07312226e+00 -1.75373703e-02
8.45140815e-01 5.87241471e-01 -3.21269602e-01 1.94137350e-01
-1.68138874e+00 4.46309805e-01 3.90166230e-02 2.04346627e-01
-8.97747755e-01 -5.69571137e-01 -7.92336941e-01 -1.08542897e-01
3.15937728e-01 -6.41508400e-01 1.09782231e+00 -7.95200288e-01
-1.28541756e+00 3.67916524e-01 8.98237899e-02 -1.03723931e+00
5.64835370e-01 -3.19860309e-01 -1.08948016e+00 -2.19389483e-01
-2.67004758e-01 -8.22599381e-02 8.46969068e-01 -9.33663726e-01
-1.25215665e-01 -5.42861819e-01 6.38307929e-02 -3.61306310e-01
-8.92969966e-01 -2.13735122e-02 -7.43269995e-02 -9.27773058e-01
-3.53637934e-01 -1.03574240e+00 -3.83425802e-01 -1.76319703e-01
-6.73651278e-01 3.87872636e-01 7.56771564e-01 -4.50761080e-01
1.31120741e+00 -2.04549241e+00 -7.16815293e-01 5.83618402e-01
1.44282132e-01 7.42325664e-01 1.28146872e-01 2.98153400e-01
5.88282123e-02 2.23211393e-01 -6.79356009e-02 -1.03361703e-01
1.41296744e-01 3.37551028e-01 -5.16633689e-01 7.69162297e-01
-1.52195364e-01 1.08240199e+00 -6.18800998e-01 6.19352385e-02
9.58006606e-02 8.26627135e-01 -4.13478583e-01 -5.65334149e-02
3.92957151e-01 3.18742603e-01 -6.18501425e-01 9.04810488e-01
8.60832751e-01 1.28869429e-01 9.66227129e-02 -6.97577670e-02
6.45858169e-01 -1.38033237e-02 -1.22873545e+00 1.08877063e+00
-4.17008668e-01 4.83830273e-01 -6.03297539e-02 -1.14107192e+00
8.19831967e-01 4.12203699e-01 6.14602566e-01 -5.81108868e-01
6.03900015e-01 3.74403894e-02 -1.71576999e-02 -2.53060490e-01
2.53062323e-02 3.02875161e-01 -6.11140907e-01 3.63648474e-01
-3.10919046e-01 6.78526580e-01 -6.67798340e-01 2.85535445e-03
1.53692961e+00 -4.11379099e-01 4.14416850e-01 4.75992709e-02
6.01937056e-01 -8.23010802e-01 6.67391300e-01 1.38448274e+00
-6.98029578e-01 1.33787438e-01 -6.66219601e-03 -3.12570602e-01
-3.30177009e-01 -1.40154099e+00 -3.04293409e-02 8.51238728e-01
4.03924644e-01 -3.44259411e-01 -9.46509361e-01 -8.55160952e-01
1.62735447e-01 2.40906641e-01 -5.38316429e-01 -6.59439504e-01
-4.84001011e-01 -9.43870544e-01 1.88798010e+00 1.03011036e+00
1.09161019e+00 -6.53839886e-01 -3.97526473e-01 1.36570334e-01
-2.17628807e-01 -1.29669249e+00 -2.65784949e-01 5.75440265e-02
-3.44768763e-01 -1.10391903e+00 -4.86336380e-01 -3.59271765e-01
4.03505623e-01 1.71945900e-01 5.84283948e-01 1.30217255e-03
-1.36705801e-01 5.30182123e-01 -1.53566882e-01 -8.88202369e-01
-3.55986834e-01 -3.09633277e-02 6.94487214e-01 2.21359059e-01
7.15014040e-01 -6.91801786e-01 -5.58610678e-01 5.08752346e-01
-7.68688679e-01 -7.66314805e-01 2.58236170e-01 7.27870107e-01
8.89703184e-02 4.97069985e-01 6.60611928e-01 -6.33331358e-01
8.77241373e-01 -5.01118660e-01 -1.60866663e-01 2.64136523e-01
-3.77966851e-01 -2.10103855e-01 6.49758577e-01 -9.32336688e-01
-7.62653172e-01 7.23466352e-02 -5.93216836e-01 -7.66772181e-02
-3.78537029e-01 2.15115309e-01 -8.65565956e-01 -5.21988213e-01
1.10201061e+00 3.36538404e-01 -3.82540256e-01 -1.28719345e-01
4.65791263e-02 9.16567326e-01 8.66992652e-01 -5.67794323e-01
1.23380291e+00 6.83539748e-01 4.22034971e-03 -8.69830370e-01
-4.95974302e-01 -1.61169112e-01 -4.64456379e-02 -2.05736518e-01
5.75097680e-01 -1.11681557e+00 -1.11917448e+00 1.13031185e+00
-9.82316494e-01 -1.21660315e-01 4.41939048e-02 2.61202097e-01
-1.30503684e-01 5.85071504e-01 -4.91758376e-01 -1.07741535e+00
-4.77025777e-01 -8.19858849e-01 7.01173544e-01 2.86921859e-01
-3.19388181e-01 -1.04811323e+00 -1.68840572e-01 4.53774005e-01
4.92633402e-01 7.05021381e-01 -1.19332828e-01 -1.00897992e+00
-2.34684631e-01 -6.38673842e-01 4.41949397e-01 4.11277145e-01
1.61519900e-01 -3.58836144e-01 -1.35521793e+00 -4.09383506e-01
1.97004184e-01 -9.77951810e-02 4.57613856e-01 2.10726187e-01
1.48138618e+00 -7.55157888e-01 -6.81915045e-01 1.03144169e+00
8.24366748e-01 2.26154745e-01 1.07704127e+00 2.48553783e-01
6.98280990e-01 -6.96118101e-02 2.59487331e-01 5.35648882e-01
7.43169198e-03 7.02828228e-01 5.52048385e-01 -2.66303658e-01
3.45811933e-01 -3.87725949e-01 6.36059105e-01 -1.35629922e-02
3.54029536e-02 -5.96810043e-01 -8.97147417e-01 1.58938721e-01
-1.66664481e+00 -1.21148980e+00 1.00426655e-02 2.10610557e+00
6.23254299e-01 4.55109477e-01 2.02332601e-01 6.69734418e-01
5.06264210e-01 2.07894862e-01 -6.11481607e-01 -2.10485354e-01
-3.17372739e-01 4.23671693e-01 9.40302312e-01 2.58672446e-01
-1.46633363e+00 5.59307396e-01 6.17605639e+00 6.40341103e-01
-1.00036883e+00 2.68503726e-01 3.86478484e-01 -1.34589493e-01
4.04930189e-02 -6.62658691e-01 -6.60204232e-01 8.12631011e-01
1.09034967e+00 2.23921742e-02 1.32724479e-01 1.04546189e+00
3.36245336e-02 4.72543359e-01 -8.00974071e-01 1.27637053e+00
3.26089650e-01 -1.05510187e+00 -2.40170985e-01 3.24158579e-01
5.88921547e-01 -1.22306705e-01 2.62366802e-01 4.15083021e-01
3.51089031e-01 -1.15075243e+00 3.12526941e-01 4.04561490e-01
5.09031773e-01 -9.24048543e-01 9.59764242e-01 6.00920081e-01
-1.14922822e+00 -2.75407732e-01 5.76723516e-02 -2.94514418e-01
5.16700260e-02 3.39532763e-01 -4.84078050e-01 3.38644862e-01
5.63070536e-01 3.85401249e-01 -6.26555979e-01 7.10802257e-01
-2.82623529e-01 9.44371760e-01 -4.39628661e-01 -2.64185190e-01
-2.33919531e-01 6.17501497e-01 4.73636836e-01 1.11199367e+00
3.73121470e-01 -1.13186516e-01 9.05096754e-02 4.00755614e-01
-1.19162656e-01 -3.71287525e-01 -9.63208854e-01 1.59114331e-01
8.69757175e-01 5.47661901e-01 -1.01831630e-01 5.62424622e-02
-2.24776506e-01 1.05719566e+00 -3.08922410e-01 5.40488183e-01
-1.21883571e+00 -5.50171196e-01 9.45109606e-01 1.55269057e-01
-1.42203853e-01 -2.06547335e-01 -5.37898242e-01 -1.31323659e+00
2.15107590e-01 -1.11085987e+00 3.40449870e-01 -2.99088985e-01
-1.24419057e+00 1.48382872e-01 -2.91006804e-01 -1.15134239e+00
-3.20107043e-01 -5.59732139e-01 -7.35441804e-01 6.00580096e-01
-8.86030436e-01 -1.30667531e+00 -2.66995877e-01 1.17594910e+00
-1.92943260e-01 -5.56539774e-01 1.06001663e+00 4.65537995e-01
-6.21159911e-01 1.63722229e+00 7.43001252e-02 7.78768718e-01
3.21946979e-01 -7.36896455e-01 6.07371867e-01 1.46640480e+00
2.79890597e-01 1.02665925e+00 6.19824290e-01 -5.83616972e-01
-1.29319644e+00 -1.19467139e+00 4.55688238e-01 -5.68991661e-01
5.33711970e-01 -7.61753023e-01 -7.05593944e-01 7.04600215e-01
-2.35041499e-01 2.15933263e-01 1.15745902e+00 7.10189417e-02
-6.26034379e-01 -4.02557820e-01 -1.37217867e+00 7.86963701e-01
1.17271256e+00 -7.01386750e-01 -2.83947766e-01 6.12321310e-02
4.54550534e-01 -2.56083012e-01 -8.66229415e-01 6.84970915e-01
8.91235590e-01 -8.68259728e-01 1.43476856e+00 -5.35200417e-01
-4.95090365e-01 -1.38878450e-01 -2.63993114e-01 -9.65076923e-01
4.58495282e-02 -1.16023874e+00 -1.02323699e+00 1.21915090e+00
2.01607466e-01 -9.76055086e-01 9.68710303e-01 6.61285520e-01
2.41687045e-01 -3.98147583e-01 -1.08107471e+00 -1.11233807e+00
2.95211468e-02 -9.96632993e-01 1.10275948e+00 8.39431465e-01
3.05582378e-02 -2.67030954e-01 -7.86376417e-01 7.98525095e-01
7.16234565e-01 -8.87754261e-01 1.10166585e+00 -1.12563872e+00
-5.25196671e-01 -1.40135720e-01 -7.97208369e-01 -1.02522767e+00
8.29748735e-02 -2.75138855e-01 -1.48276433e-01 -7.33075738e-01
-3.16707104e-01 -3.26279402e-01 -7.20677614e-01 6.78759158e-01
-1.22144617e-01 5.91030300e-01 -8.28005746e-02 -1.14044450e-01
-5.23057342e-01 2.97359139e-01 4.57998633e-01 -5.17409384e-01
-1.72937885e-01 6.97798789e-01 -1.24786890e+00 8.43239069e-01
1.42010152e+00 -3.03554624e-01 -5.58062792e-01 -5.76086938e-02
1.55109271e-01 -3.00168484e-01 7.71472454e-01 -1.58557546e+00
2.49580830e-01 -1.01394951e-01 6.09880149e-01 -2.63279136e-02
2.70904094e-01 -1.10741282e+00 1.17841877e-01 8.10067713e-01
-6.54966906e-02 -3.05802852e-01 3.47990185e-01 8.93975735e-01
2.82861173e-01 4.26810235e-01 5.22333920e-01 4.27381516e-01
-2.80011684e-01 3.35637361e-01 -5.53148568e-01 -1.85040683e-01
1.15092385e+00 -4.56272423e-01 -5.17721832e-01 -6.81809247e-01
-6.78379893e-01 -1.23122744e-01 1.44563481e-01 6.39225662e-01
6.12812936e-01 -1.52447605e+00 -3.81277680e-01 5.71386456e-01
1.89605191e-01 -5.78435183e-01 2.71947801e-01 5.09886205e-01
-1.67008758e-01 2.36413151e-01 -1.10257603e-02 -3.62859964e-01
-1.49317956e+00 5.34222066e-01 6.66399479e-01 -1.10937275e-01
-3.03516895e-01 7.17594385e-01 -1.62875056e-01 -3.71363938e-01
8.20273936e-01 -1.55483857e-01 2.47401390e-02 -5.98748565e-01
9.85534549e-01 3.41335744e-01 1.00939199e-01 -6.64295793e-01
-8.13031852e-01 2.89600432e-01 1.64490491e-01 7.37873614e-02
7.89496601e-01 -1.44592039e-02 6.29955173e-01 -2.22155228e-01
9.65789676e-01 3.63728106e-01 -1.16703677e+00 -1.75337091e-01
-3.38815570e-01 -2.97695607e-01 -1.87932134e-01 -9.86424863e-01
-1.01192832e+00 7.58189201e-01 9.94003057e-01 4.49008554e-01
1.27778280e+00 -4.38172221e-01 1.19792616e+00 6.72388136e-01
6.59783065e-01 -8.95462573e-01 5.25865443e-02 2.55146772e-01
3.45052868e-01 -1.00755870e+00 -1.86573759e-01 -3.07925284e-01
-3.06317657e-01 4.81763780e-01 5.71631610e-01 6.57813810e-03
9.22363281e-01 7.18542159e-01 1.46463633e-01 2.72888243e-01
-1.40512168e-01 1.47489145e-01 1.24655411e-01 1.48212087e+00
-8.38468224e-02 2.07497418e-01 2.27310453e-02 1.32039380e+00
-3.35516572e-01 -9.86631773e-03 2.70220906e-01 8.57597768e-01
-1.71427324e-01 -1.08733177e+00 -7.24830627e-01 1.09416783e-01
-7.99089134e-01 2.13806957e-01 -5.36068261e-01 5.69177568e-01
3.72779816e-01 1.33219707e+00 -4.12894309e-01 -9.31381464e-01
3.27554792e-01 -1.28027707e-01 2.07100257e-01 3.66539843e-02
-6.51279449e-01 -4.29087311e-01 4.26839665e-02 -5.84957302e-01
-3.36784035e-01 -4.50325698e-01 -9.23139215e-01 -6.25068963e-01
-4.25273888e-02 -8.24054405e-02 4.81180936e-01 9.53070998e-01
3.77720147e-01 6.41618848e-01 7.01387644e-01 -3.75110775e-01
-8.72394621e-01 -6.67820454e-01 -5.65307379e-01 4.68040526e-01
3.18809301e-01 -6.00236654e-01 -2.79338896e-01 -1.94398463e-01] | [13.607733726501465, 5.654835224151611] |
7ea950c3-6cb7-47eb-a7bc-ee4924971fe1 | exploring-overcomplete-representations-for | 2010.10661 | null | https://arxiv.org/abs/2010.10661v1 | https://arxiv.org/pdf/2010.10661v1.pdf | Exploring Overcomplete Representations for Single Image Deraining using CNNs | Removal of rain streaks from a single image is an extremely challenging problem since the rainy images often contain rain streaks of different size, shape, direction and density. Most recent methods for deraining use a deep network following a generic "encoder-decoder" architecture which captures low-level features across the initial layers and high-level features in the deeper layers. For the task of deraining, the rain streaks which are to be removed are relatively small and focusing much on global features is not an efficient way to solve the problem. To this end, we propose using an overcomplete convolutional network architecture which gives special attention in learning local structures by restraining the receptive field of filters. We combine it with U-Net so that it does not lose out on the global structures as well while focusing more on low-level features, to compute the derained image. The proposed network called, Over-and-Under Complete Deraining Network (OUCD), consists of two branches: overcomplete branch which is confined to small receptive field size in order to focus on the local structures and an undercomplete branch that has larger receptive fields to primarily focus on global structures. Extensive experiments on synthetic and real datasets demonstrate that the proposed method achieves significant improvements over the recent state-of-the-art methods. | ['Vishal M. Patel', 'Jeya Maria Jose Valanarasu', 'Rajeev Yasarla'] | 2020-10-20 | null | null | null | null | ['single-image-deraining'] | ['computer-vision'] | [-1.20178767e-01 -1.74719021e-01 6.67461395e-01 -4.77097869e-01
-1.88792408e-01 -1.17649272e-01 3.65502909e-02 -3.11953366e-01
-3.20601672e-01 8.93724799e-01 1.94485307e-01 1.68253258e-02
1.85784727e-01 -9.62248921e-01 -9.57466066e-01 -9.51966763e-01
-2.32203588e-01 -1.05682403e-01 4.30863500e-01 -2.55782545e-01
-1.24300141e-02 5.24038970e-01 -1.47858322e+00 3.14163178e-01
1.04863477e+00 7.91088045e-01 5.33380091e-01 6.26873553e-01
2.72990428e-02 1.31752813e+00 -5.10885417e-01 2.15202749e-01
3.48230213e-01 -4.56648946e-01 -4.32254761e-01 1.89587459e-01
8.06991458e-01 -6.66843534e-01 -4.59471613e-01 1.21419954e+00
5.02479017e-01 1.43002316e-01 5.02445519e-01 -2.32897595e-01
-6.50748610e-01 1.22161955e-01 -6.67028368e-01 7.20895410e-01
-3.49350750e-01 -1.47274300e-01 6.99582994e-01 -1.02564943e+00
4.69994128e-01 1.30218363e+00 5.96739888e-01 2.20725551e-01
-9.75801289e-01 -5.75383902e-01 4.58533198e-01 1.53452605e-01
-1.27970886e+00 -1.13775104e-01 6.20395124e-01 -3.83559942e-01
7.69182920e-01 1.67051867e-01 5.19679368e-01 5.31246603e-01
4.26908970e-01 5.57311833e-01 1.21875226e+00 -1.25743359e-01
8.43029469e-02 -1.35153219e-01 4.44878370e-01 7.57008016e-01
4.69560623e-01 7.85788521e-02 1.54328704e-01 2.61138976e-01
8.99437368e-01 2.99588352e-01 -7.36676335e-01 -3.05207865e-03
-6.08170271e-01 8.98290575e-01 1.02420151e+00 3.67222995e-01
-6.64398789e-01 -9.89898518e-02 -8.39715973e-02 4.97519940e-01
5.94512463e-01 1.47332430e-01 -5.50443232e-01 6.68681681e-01
-1.04080629e+00 5.30602992e-01 6.29962683e-01 5.78612864e-01
1.22948563e+00 4.00102794e-01 -1.89504683e-01 9.24666703e-01
1.66316077e-01 7.11895823e-01 2.90342599e-01 -5.21005929e-01
5.26072681e-01 5.23757696e-01 1.72028005e-01 -9.84096289e-01
-4.39757705e-01 -7.25416780e-01 -1.36349308e+00 6.52428806e-01
5.85145392e-02 -4.37149882e-01 -1.55811369e+00 1.45255280e+00
1.56736970e-01 3.82378131e-01 1.73062757e-01 1.33697677e+00
1.04516041e+00 1.19622087e+00 -1.79310516e-01 -3.20542008e-01
1.20222473e+00 -1.04829586e+00 -7.57665575e-01 -5.69363415e-01
9.03285816e-02 -8.48611355e-01 5.44914901e-01 2.75601804e-01
-1.00899863e+00 -8.41054261e-01 -1.21702933e+00 -2.00225219e-01
-3.37015241e-01 3.05663526e-01 2.36025080e-01 3.95316929e-02
-9.10042822e-01 6.71290457e-01 -6.28115237e-01 -1.48806348e-01
3.44304413e-01 2.66372144e-01 -3.39723021e-01 -4.61522669e-01
-1.23428917e+00 8.82440925e-01 4.24708903e-01 8.46612155e-01
-1.19225502e+00 -4.73963857e-01 -8.83152962e-01 2.82587260e-01
5.65201007e-02 -5.13493598e-01 7.31104851e-01 -1.16110075e+00
-1.18529427e+00 3.75656545e-01 -4.96247672e-02 -3.68190140e-01
1.34989396e-01 -5.32420397e-01 -3.39092642e-01 1.49086028e-01
-2.00065091e-01 3.07445496e-01 1.37689435e+00 -1.33098495e+00
-8.69076252e-01 -3.57787222e-01 1.68600947e-01 4.11562741e-01
7.64534101e-02 -3.14142972e-01 -2.42453054e-01 -8.85499001e-01
2.52998918e-01 -7.13312685e-01 -3.59994590e-01 -2.95640767e-01
-1.39191717e-01 3.53481561e-01 1.10168338e+00 -1.01586449e+00
1.18396509e+00 -2.08163023e+00 2.60714382e-01 5.43166287e-02
4.49783832e-01 7.04143107e-01 -2.04684764e-01 1.13070585e-01
-2.27187365e-01 -3.17948312e-01 -5.05739152e-01 -4.60100062e-02
-5.86961746e-01 4.41344559e-01 -1.22139603e-01 6.90021455e-01
2.55778909e-01 3.17083478e-01 -5.43780208e-01 -2.13595316e-01
3.67090374e-01 6.37312531e-01 -4.79355961e-01 6.52282953e-01
-3.01597510e-02 3.56024981e-01 -3.87119621e-01 4.46263969e-01
1.53675985e+00 5.02497144e-02 -1.49321496e-01 -2.23845795e-01
-4.21890616e-01 -8.25861916e-02 -1.38499165e+00 1.01661253e+00
-4.06644106e-01 4.12057132e-01 7.73229957e-01 -9.69682336e-01
1.02682126e+00 2.87819356e-02 -5.08822389e-02 -5.70811689e-01
2.65043139e-01 1.75275385e-01 9.51950625e-03 -8.33804131e-01
5.09273112e-02 -3.83303732e-01 4.80468065e-01 -1.29020691e-01
1.30767867e-01 7.47848228e-02 5.40204570e-02 -5.95137514e-02
9.48379517e-01 -4.97044250e-02 7.22876713e-02 -4.40034539e-01
6.95829093e-01 -3.78675491e-01 8.39757979e-01 6.46487892e-01
-2.15197797e-03 9.24089313e-01 1.36207551e-01 -7.67724276e-01
-8.91620278e-01 -8.46200883e-01 -1.93250731e-01 8.79205406e-01
2.64864743e-01 9.96573195e-02 -7.98024714e-01 -5.87282419e-01
-7.76416510e-02 1.91149861e-01 -9.01771903e-01 1.99054647e-02
-8.60000193e-01 -1.07928026e+00 8.83015394e-02 3.79384756e-01
1.04166496e+00 -1.38244331e+00 -4.14449126e-01 2.47341871e-01
-1.22472338e-01 -9.76995111e-01 -2.65370876e-01 5.83060980e-01
-1.04357445e+00 -9.67918515e-01 -1.07423222e+00 -1.17676365e+00
9.24182713e-01 4.97924268e-01 1.05703056e+00 4.49914820e-02
-2.61006981e-01 -5.24684191e-01 -5.08988678e-01 -4.70609933e-01
2.93921590e-01 -5.74312024e-02 -4.87497181e-01 2.86155641e-01
1.83844700e-01 -7.93418109e-01 -7.94514120e-01 -1.90315209e-02
-1.08486331e+00 -2.61798292e-01 1.09709537e+00 1.03337133e+00
5.26279271e-01 2.28946924e-01 2.62416989e-01 -1.19798207e+00
2.72272557e-01 -5.62891722e-01 -6.16657972e-01 -4.97046188e-02
-3.63218412e-02 9.91485864e-02 9.71083701e-01 1.18351586e-01
-1.30516088e+00 3.66410017e-02 -2.31803924e-01 -6.26941144e-01
-1.50501639e-01 4.13436890e-01 -1.82683945e-01 -3.30722004e-01
6.11435115e-01 4.08658385e-01 -1.68227971e-01 -8.62938702e-01
9.61789712e-02 5.63114345e-01 3.90697867e-01 -7.09122196e-02
1.02763224e+00 5.94132900e-01 -2.08046451e-01 -1.02849317e+00
-1.16939080e+00 -5.69065034e-01 -6.30273223e-01 2.70233992e-02
8.62378001e-01 -1.23013270e+00 -2.29924187e-01 7.25216031e-01
-1.04204321e+00 -1.63208798e-01 -4.40268330e-02 4.64457214e-01
8.33694339e-02 3.05919796e-01 -7.82686889e-01 -7.07980573e-01
-5.93005419e-01 -8.93399894e-01 7.86992073e-01 5.77299654e-01
6.61022782e-01 -8.13564062e-01 1.75444871e-01 -8.47714469e-02
5.68989336e-01 2.24021152e-01 6.95625782e-01 -2.54713446e-02
-7.58122087e-01 1.39490217e-01 -4.29116845e-01 1.02603805e+00
2.16686055e-01 -1.98713869e-01 -9.43694413e-01 -5.72821081e-01
4.21452522e-01 -2.87362665e-01 1.53569901e+00 6.89935207e-01
9.12153482e-01 -3.82773399e-01 1.71068572e-02 9.97610331e-01
1.87141931e+00 4.25133519e-02 9.83337402e-01 2.65107810e-01
7.37189949e-01 5.41542053e-01 5.78314543e-01 3.49519670e-01
1.52661324e-01 -2.32360885e-02 7.76201248e-01 -6.17623389e-01
-2.79630363e-01 2.07880124e-01 2.84658343e-01 9.39401269e-01
-4.55861300e-01 -1.97130874e-01 -3.18505406e-01 7.69357920e-01
-1.85955811e+00 -9.63806391e-01 -1.19291633e-01 2.00044608e+00
5.68578541e-01 2.39646062e-02 -4.52240527e-01 -1.41124174e-01
7.98732638e-01 5.89591742e-01 -3.53552759e-01 -4.55176294e-01
-3.30877721e-01 4.74456429e-01 6.22424304e-01 7.81799793e-01
-1.33689880e+00 9.47675943e-01 5.91738558e+00 5.93823552e-01
-1.18588698e+00 -9.70338359e-02 4.92261410e-01 1.34207845e-01
6.26495108e-02 -3.27850319e-02 -8.85475874e-01 4.92685467e-01
3.38722080e-01 5.92792094e-01 3.80758077e-01 6.43062949e-01
4.63093281e-01 -2.10875362e-01 -3.61007392e-01 7.82260120e-01
6.27835095e-02 -1.03621054e+00 2.05504760e-01 -2.97683358e-01
8.07115674e-01 3.06388944e-01 -2.08091572e-01 4.36327904e-01
2.03762710e-01 -1.07902145e+00 4.25791591e-01 7.98112273e-01
4.74159688e-01 -7.76179433e-01 1.18313956e+00 3.89381349e-01
-1.17995858e+00 -1.41801834e-01 -1.13367784e+00 -3.37409854e-01
-2.52994031e-01 1.08650148e+00 -3.01748186e-01 5.97003877e-01
1.12054276e+00 7.57634759e-01 -3.64337891e-01 9.98428404e-01
-3.82492691e-01 6.90111399e-01 -3.86738390e-01 4.16589826e-01
5.70935845e-01 -4.80747283e-01 5.27999759e-01 1.47113419e+00
2.98734814e-01 5.37039816e-01 2.04636186e-01 4.98013228e-01
-4.02563028e-02 -5.55422269e-02 -5.72328627e-01 5.78091145e-01
-8.16158578e-02 1.26904237e+00 -2.98033476e-01 -4.12061900e-01
-5.30967414e-01 8.63946855e-01 4.81647432e-01 7.02533543e-01
-7.08462358e-01 -8.36669505e-01 7.01544046e-01 1.57332450e-01
9.42465007e-01 -1.08534694e-01 -2.70295423e-02 -1.38273048e+00
8.74080956e-02 -9.26019669e-01 2.38690093e-01 -6.28849328e-01
-1.13486600e+00 1.05275500e+00 -4.69724238e-01 -1.21912181e+00
3.13538134e-01 -6.66881621e-01 -7.30736673e-01 1.21944761e+00
-2.11871409e+00 -1.07845199e+00 -8.50887060e-01 8.69048476e-01
6.60721600e-01 1.35416210e-01 3.01104754e-01 5.42112768e-01
-5.67253232e-01 1.75939277e-01 4.60428774e-01 1.91281080e-01
7.43549049e-01 -1.19509554e+00 -2.81552553e-01 1.18969822e+00
-2.62962490e-01 3.88715655e-01 8.12847853e-01 -6.40073717e-01
-8.81977916e-01 -1.42559195e+00 8.60450089e-01 2.11368322e-01
1.29548892e-01 -2.81131655e-01 -1.29690361e+00 6.01185501e-01
2.55977839e-01 4.63459849e-01 3.15713789e-03 -7.09202979e-03
-1.94787398e-01 -6.21894836e-01 -1.09602082e+00 9.65638310e-02
6.78505063e-01 -3.20786759e-02 -8.38245809e-01 2.61180818e-01
3.35103959e-01 -4.42528307e-01 -4.91135180e-01 6.39998615e-01
1.56581566e-01 -1.33718252e+00 7.59182930e-01 -3.17648411e-01
6.20798588e-01 -5.82742453e-01 -2.08001524e-01 -1.48057556e+00
-8.94782305e-01 -2.94946462e-01 -7.68764690e-02 1.02861893e+00
9.73199680e-02 -5.11176288e-01 7.36735404e-01 -3.45279634e-01
-3.74117434e-01 -6.87585890e-01 -6.05323911e-01 -3.72994244e-01
2.59653758e-02 2.97496438e-01 2.80554026e-01 7.34908164e-01
-7.53568947e-01 5.60163438e-01 -7.70923018e-01 7.91657090e-01
6.97030425e-01 5.67636728e-01 4.44215894e-01 -1.16583586e+00
-1.08649649e-01 2.07695052e-01 -2.97857434e-01 -1.29330504e+00
-2.06623033e-01 -3.87907624e-01 4.68817174e-01 -1.57709146e+00
2.86863089e-01 -1.25335842e-01 -3.73064458e-01 4.69098061e-01
-4.41370308e-01 3.46343577e-01 1.93533137e-01 3.33820164e-01
-2.90321887e-01 8.25799644e-01 1.69756162e+00 -1.91898912e-01
-2.29591995e-01 3.07840705e-02 -3.50708663e-01 9.96195853e-01
6.27122819e-01 -4.58989799e-01 -2.51413673e-01 -8.70948255e-01
-2.69097894e-01 -8.20612386e-02 2.48540789e-01 -1.27635694e+00
5.33229411e-02 -2.63488255e-02 7.39906013e-01 -5.98853171e-01
2.28523895e-01 -7.76315629e-01 -2.88212985e-01 4.49530274e-01
9.29005817e-02 -2.67373592e-01 3.52248512e-02 5.25246382e-01
-8.64289939e-01 -2.34062567e-01 1.41146421e+00 -4.93346065e-01
-7.64639318e-01 5.03331482e-01 -5.12282431e-01 -1.86019912e-01
7.80708909e-01 -6.43290207e-02 -2.83960551e-01 -3.11351001e-01
-1.07672393e+00 3.20805997e-01 -3.48990336e-02 5.57795316e-02
7.32280016e-01 -8.23957801e-01 -1.06571472e+00 3.96790266e-01
-4.63487774e-01 3.18151742e-01 6.10517740e-01 5.16556561e-01
-1.04567587e+00 2.77727693e-01 -5.33944547e-01 -2.01097980e-01
-1.16429222e+00 4.58915740e-01 5.92579722e-01 -4.59958881e-01
-8.80768359e-01 9.94297504e-01 1.00700510e+00 -2.40627706e-01
9.31411982e-02 -5.44997215e-01 -7.64327466e-01 1.69066265e-02
6.39840841e-01 2.05715403e-01 1.82250246e-01 -5.91497242e-01
-1.63469315e-01 7.73466289e-01 -3.44543397e-01 4.92721260e-01
1.67852175e+00 -2.45132819e-01 -4.35231715e-01 1.46818846e-01
1.13556349e+00 5.32548763e-02 -1.50564122e+00 -2.26350829e-01
-6.94072068e-01 -4.65822667e-01 4.38764572e-01 -6.37806416e-01
-1.65855682e+00 1.10942531e+00 7.76619256e-01 1.52816087e-01
1.58077741e+00 -5.67207813e-01 8.38273227e-01 5.36721885e-01
-4.64347303e-02 -8.74773383e-01 -1.44820184e-01 1.02499497e+00
9.71833646e-01 -1.10746789e+00 7.42225125e-02 -4.65028495e-01
-3.35717380e-01 1.14518809e+00 8.78245890e-01 -9.99931335e-01
1.03494322e+00 4.67959434e-01 2.37409592e-01 -2.78404891e-01
-4.37697411e-01 -4.72000688e-01 -1.66052971e-02 3.47563297e-01
2.97319949e-01 -2.28104398e-01 -3.80840003e-01 5.04337072e-01
2.39602938e-01 1.44575089e-01 5.78824461e-01 8.56823504e-01
-1.04479790e+00 -6.19893193e-01 -7.72988200e-01 5.62474847e-01
-6.40299380e-01 -3.63319606e-01 8.23422819e-02 6.56379938e-01
7.10463405e-01 6.74977005e-01 8.41689780e-02 -1.11035116e-01
4.10387278e-01 -5.46192646e-01 3.54597330e-01 -6.21576130e-01
-5.11053622e-01 2.88213104e-01 -9.34935436e-02 -3.60069573e-01
-4.84900922e-01 -2.84848779e-01 -9.79545355e-01 -1.88587829e-01
-2.60721207e-01 2.63473570e-01 1.03831083e-01 8.70361209e-01
1.53174803e-01 7.05489218e-01 7.89841533e-01 -1.08793795e+00
-3.90353888e-01 -1.33207345e+00 -1.15564466e+00 2.54477888e-01
1.00689459e+00 -4.75848049e-01 -5.36986411e-01 1.83515251e-01] | [10.927396774291992, -3.2249531745910645] |
50f354f9-e869-414d-82ee-a0491ac6a6b1 | a-first-look-at-dataset-bias-in-license-plate | 2208.10657 | null | https://arxiv.org/abs/2208.10657v2 | https://arxiv.org/pdf/2208.10657v2.pdf | A First Look at Dataset Bias in License Plate Recognition | Public datasets have played a key role in advancing the state of the art in License Plate Recognition (LPR). Although dataset bias has been recognized as a severe problem in the computer vision community, it has been largely overlooked in the LPR literature. LPR models are usually trained and evaluated separately on each dataset. In this scenario, they have often proven robust in the dataset they were trained in but showed limited performance in unseen ones. Therefore, this work investigates the dataset bias problem in the LPR context. We performed experiments on eight datasets, four collected in Brazil and four in mainland China, and observed that each dataset has a unique, identifiable "signature" since a lightweight classification model predicts the source dataset of a license plate (LP) image with more than 95% accuracy. In our discussion, we draw attention to the fact that most LPR models are probably exploiting such signatures to improve the results achieved in each dataset at the cost of losing generalization capability. These results emphasize the importance of evaluating LPR models in cross-dataset setups, as they provide a better indication of generalization (hence real-world performance) than within-dataset ones. | ['David Menotti', 'Eduardo Luz', 'Valter Estevam', 'Marcelo Santos', 'Rayson Laroca'] | 2022-08-23 | null | null | null | null | ['license-plate-recognition'] | ['computer-vision'] | [ 2.39036173e-01 -5.08705676e-01 -1.51127130e-01 -4.02490407e-01
-9.58825946e-01 -9.01471674e-01 7.94159055e-01 4.95973118e-02
-5.03027797e-01 5.91237664e-01 -1.73759565e-01 -2.38715738e-01
-2.14600265e-01 -6.05107605e-01 -7.72356331e-01 -8.66776109e-01
1.56897426e-01 4.04700398e-01 3.21554273e-01 1.84622854e-02
6.76156461e-01 7.84057319e-01 -1.67278218e+00 2.45321035e-01
6.70576632e-01 7.21298754e-01 4.29154113e-02 3.27548146e-01
1.87327296e-01 6.59440339e-01 -8.03029001e-01 -9.29891229e-01
4.27763313e-01 -1.04903460e-01 -4.76467937e-01 3.19705397e-01
7.79033065e-01 9.33321565e-02 -5.19792676e-01 9.91986513e-01
3.27495933e-01 -7.71028101e-02 6.77724183e-01 -1.25490117e+00
-4.91950542e-01 3.79791856e-01 -4.72038627e-01 3.17378640e-01
3.12252645e-03 2.71093577e-01 1.13794923e+00 -7.92916954e-01
6.65963352e-01 6.14294767e-01 8.76073360e-01 3.80398601e-01
-1.33315206e+00 -6.44048452e-01 -1.88676029e-01 1.48299515e-01
-1.58018589e+00 -4.85073924e-01 7.71557271e-01 -5.78100860e-01
4.34603453e-01 3.94317240e-01 1.12790294e-01 1.17248833e+00
2.19868049e-02 9.77571487e-01 1.38799763e+00 -2.91994542e-01
7.96018615e-02 5.65841198e-01 1.63835838e-01 1.00081928e-01
4.34662312e-01 2.12790929e-02 -5.83027482e-01 7.77557269e-02
5.29328704e-01 -2.02698514e-01 -3.74680489e-01 -3.95167023e-01
-1.17595625e+00 7.70635366e-01 1.83404565e-01 4.32314336e-01
-3.36388536e-02 -2.15202555e-01 3.78893435e-01 3.74488026e-01
3.96079630e-01 5.88790059e-01 -4.05271798e-01 -1.24560691e-01
-1.09424770e+00 2.70217866e-01 6.76169872e-01 7.68790722e-01
7.23215163e-01 -8.10673833e-02 1.96406379e-01 1.10038197e+00
1.45367393e-02 6.25825167e-01 6.22048199e-01 -3.00658256e-01
6.72726989e-01 7.02228367e-01 -2.97446214e-02 -1.14290094e+00
-2.13890940e-01 -6.65782988e-01 -5.10039687e-01 1.25064552e-01
7.40364671e-01 2.34667614e-01 -5.94402075e-01 1.29338646e+00
-4.06867325e-01 2.33061031e-01 2.68308103e-01 6.35413051e-01
5.02240539e-01 3.89516503e-01 -9.75110605e-02 1.54617131e-01
1.02031648e+00 -6.47195280e-01 -1.26560390e-01 -3.88190836e-01
6.64387524e-01 -8.81466091e-01 8.79013956e-01 4.89162028e-01
-2.74836898e-01 -4.32792068e-01 -8.95243108e-01 3.43280464e-01
-6.02135777e-01 6.57707155e-01 3.22264791e-01 7.72868752e-01
-6.42501593e-01 4.73048389e-01 -1.36598781e-01 -5.57037234e-01
4.95561510e-01 3.30695689e-01 -7.77521193e-01 -3.63896787e-01
-8.79412293e-01 9.67320502e-01 3.35113525e-01 1.61902204e-01
-6.51899099e-01 -5.79972982e-01 -4.23366934e-01 -1.91206440e-01
5.40644407e-01 2.93381333e-01 8.44745338e-01 -9.21910822e-01
-1.01832664e+00 1.32894301e+00 2.49671966e-01 -5.83938420e-01
8.45346332e-01 -3.33789289e-02 -9.15253222e-01 -1.71910331e-01
7.57929403e-03 2.75710374e-01 8.33887339e-01 -1.43811738e+00
-7.05996275e-01 -4.99615043e-01 -4.89968024e-02 -3.02805841e-01
6.50045797e-02 4.12801653e-02 -2.64729112e-01 -7.39555180e-01
1.84206381e-01 -1.20362997e+00 2.11752415e-01 -3.70568097e-01
-1.40988857e-01 -7.31756389e-02 7.41589725e-01 -5.68880498e-01
9.85665739e-01 -2.41973305e+00 -4.33639079e-01 2.48541176e-01
-1.24993742e-01 6.39727712e-01 1.18647590e-01 4.41774815e-01
-7.24212751e-02 2.67325252e-01 -3.69785666e-01 -1.08869478e-01
-1.46951497e-01 2.32738733e-01 -6.46537125e-01 7.94799626e-01
4.01129067e-01 6.31148040e-01 -5.44508815e-01 -1.98377252e-01
3.12108278e-01 2.84203947e-01 -2.66366482e-01 -2.12490819e-02
1.75883308e-01 2.71344841e-01 -2.33053729e-01 6.85134351e-01
7.61218607e-01 3.71939205e-02 3.55550684e-02 -8.89837965e-02
-2.94437259e-01 -1.53042644e-01 -1.26913059e+00 9.07093942e-01
-3.14413160e-01 1.27956843e+00 -4.71801221e-01 -7.35925972e-01
1.35862970e+00 1.60342574e-01 1.44362837e-01 -7.53630877e-01
-6.98374361e-02 5.92134535e-01 1.41015530e-01 -5.24498284e-01
7.80430198e-01 -1.74601585e-01 -6.95174113e-02 1.63804010e-01
-1.12557851e-01 1.79583460e-01 2.84415424e-01 -2.96437591e-01
6.44666255e-01 9.73721594e-03 5.33081964e-02 -3.47263366e-01
8.86678457e-01 6.89150468e-02 4.25118655e-01 8.84244859e-01
-2.60623097e-01 1.07582879e+00 5.70207894e-01 -4.18270171e-01
-1.04654920e+00 -8.56398284e-01 -5.54948092e-01 8.06635618e-01
1.22174695e-01 -3.25509831e-02 -3.86418909e-01 -7.77325809e-01
1.73111886e-01 7.27747202e-01 -6.83315814e-01 -6.47734180e-02
-7.23759055e-01 -9.57263410e-01 1.04798138e+00 4.19379860e-01
7.75434315e-01 -6.80957377e-01 -3.78378332e-01 5.49392551e-02
-9.63589102e-02 -1.22742450e+00 -6.79689348e-02 4.01064567e-02
-7.88994431e-01 -1.33498991e+00 -7.59224117e-01 -5.68900347e-01
6.01152718e-01 3.46749753e-01 1.15672696e+00 1.12924408e-02
-7.26666152e-02 3.88058841e-01 -3.96504760e-01 -4.54247296e-01
-7.55619645e-01 1.29642218e-01 -2.36050822e-02 5.87911725e-01
6.22363567e-01 -7.73836151e-02 -8.03953707e-02 9.01566863e-01
-1.01686096e+00 -5.27975440e-01 7.87183583e-01 5.23278236e-01
1.79665878e-01 2.10803792e-01 5.04899979e-01 -9.91343319e-01
4.63760525e-01 -3.83709550e-01 -8.54097545e-01 4.72889960e-01
-5.30061603e-01 -1.50581196e-01 4.60475028e-01 -2.42195547e-01
-9.18181062e-01 -6.20739572e-02 -8.04129615e-02 -3.28754127e-01
-4.93833750e-01 3.97398263e-01 -1.44306123e-01 -2.55479515e-01
6.66804373e-01 6.49259388e-01 8.80206823e-02 -6.11941636e-01
-1.15637161e-01 8.25046659e-01 5.60763955e-01 -6.16866529e-01
8.36792529e-01 2.62496740e-01 9.28831995e-02 -1.09343016e+00
-3.66959691e-01 -7.25410700e-01 -7.12000787e-01 -4.74184483e-01
5.15019953e-01 -8.48972797e-01 -5.88222146e-01 8.19021881e-01
-6.56895638e-01 7.68952817e-02 -1.05549805e-01 5.61061502e-01
-3.12036574e-01 2.70234227e-01 6.33021519e-02 -8.25467408e-01
5.44515550e-01 -1.33457053e+00 7.73439109e-01 1.18934073e-01
7.38378018e-02 -1.00580311e+00 -4.59732823e-02 5.85762203e-01
4.14478958e-01 4.17576022e-02 6.48268223e-01 -1.20718336e+00
-6.15387201e-01 -8.27442110e-01 -2.44080171e-01 5.65131009e-01
-5.29674776e-02 3.34369987e-01 -1.47715068e+00 -1.19479254e-01
-1.50474057e-01 -7.41598755e-02 9.70270157e-01 -6.05568849e-02
8.46398532e-01 7.36910999e-02 -4.21830341e-02 7.71787912e-02
1.83116233e+00 -6.69689327e-02 8.96407962e-01 6.22997224e-01
4.16588902e-01 5.42603493e-01 4.75257486e-01 2.15123460e-01
-9.00706500e-02 8.84584486e-01 3.65178764e-01 1.94410548e-01
-8.91254097e-02 -3.19198370e-01 4.69463080e-01 2.63861179e-01
-3.09381574e-01 -3.70545417e-01 -1.21825290e+00 4.93961722e-01
-1.52154696e+00 -9.37418640e-01 -5.31617880e-01 2.56229424e+00
2.46147797e-01 2.09646583e-01 2.90343761e-01 4.30562526e-01
8.07374656e-01 1.95730373e-01 -1.11202598e-01 -1.01504929e-01
-6.44469321e-01 -4.50620025e-01 8.88227522e-01 2.12766573e-01
-1.18953526e+00 6.65231049e-01 6.07544422e+00 7.18348503e-01
-1.51916897e+00 -3.98326933e-01 6.21400297e-01 3.49410474e-01
1.13062501e-01 -1.77951232e-01 -9.52268243e-01 5.10453522e-01
7.45764911e-01 -6.71654493e-02 2.50689108e-02 9.51746821e-01
-8.07641074e-02 -3.04753304e-01 -1.27252984e+00 1.08848250e+00
4.17183578e-01 -1.25094903e+00 9.53378603e-02 3.51085007e-01
5.55045903e-01 1.23057306e-01 3.34527433e-01 3.74912679e-01
-3.76383454e-01 -1.03979647e+00 6.04541719e-01 3.50659430e-01
4.44621295e-01 -6.74690783e-01 1.12775183e+00 3.63667995e-01
-6.73009753e-01 -1.13711476e-01 -5.16401529e-01 1.47342175e-01
-2.31169060e-01 4.26936001e-01 -1.00595784e+00 6.57263219e-01
4.67485338e-01 8.18636119e-01 -1.11934555e+00 1.32587683e+00
-4.57555763e-02 7.79945493e-01 -2.95774974e-02 6.39336184e-02
2.91349292e-01 -1.85719803e-01 6.08228087e-01 1.28214872e+00
2.13602021e-01 -4.88470674e-01 -3.19542706e-01 6.82273269e-01
-1.41278252e-01 1.50336340e-01 -7.75698125e-01 7.70375729e-02
9.07413065e-02 9.62380946e-01 -7.24138498e-01 1.87578961e-01
-6.88786983e-01 5.49171507e-01 -5.65962791e-02 1.97276652e-01
-8.63796771e-01 -3.64027694e-02 5.99221528e-01 2.15097830e-01
3.78690451e-01 5.20510823e-02 -1.97067559e-01 -1.20025051e+00
1.69493556e-01 -8.47191155e-01 3.93341511e-01 -4.84331280e-01
-1.45446956e+00 4.36096072e-01 -9.49608907e-02 -1.62834799e+00
7.67311454e-02 -1.11609614e+00 -3.23873460e-01 7.22958803e-01
-1.52352428e+00 -1.06127048e+00 -2.18437109e-02 3.97971898e-01
4.83142823e-01 -5.15554368e-01 5.56332707e-01 4.93730515e-01
-7.01914608e-01 6.12713277e-01 5.55279732e-01 5.65128386e-01
8.86419296e-01 -9.16628838e-01 5.01452275e-02 1.00795639e+00
4.41593885e-01 5.89886487e-01 7.22148418e-01 -3.76254231e-01
-1.24112439e+00 -1.09017789e+00 9.33697045e-01 -7.87606359e-01
6.04366839e-01 -1.63200408e-01 -1.12676227e+00 7.21846461e-01
-2.08638996e-01 9.02654827e-02 8.88351977e-01 1.44493937e-01
-5.47755599e-01 -3.47155035e-01 -1.20294940e+00 4.66949344e-01
4.09867644e-01 -6.03500128e-01 -6.27044261e-01 2.02810504e-02
-1.73023880e-01 -1.66573852e-01 -8.45525563e-01 2.01221362e-01
5.47424734e-01 -1.21955478e+00 8.86668503e-01 -4.94086415e-01
4.03680146e-01 -5.91866612e-01 -4.70650971e-01 -9.69326913e-01
-5.77654280e-02 4.36496213e-02 5.14952779e-01 1.56336081e+00
6.87271297e-01 -7.29101062e-01 8.01105857e-01 9.36822832e-01
8.00174549e-02 -4.44664240e-01 -9.89262760e-01 -1.18582010e+00
1.46515578e-01 -6.67807579e-01 5.08749306e-01 8.61753464e-01
-4.92274880e-01 -7.59490579e-02 -5.39654136e-01 4.25129414e-01
5.12070358e-01 1.71549901e-01 1.05927062e+00 -1.40483928e+00
8.19081143e-02 -5.14365673e-01 -9.22147155e-01 -6.48568273e-01
2.33404219e-01 -9.78999078e-01 -1.98482517e-02 -8.11393917e-01
1.66103333e-01 -5.27400136e-01 -2.95786023e-01 1.02581300e-01
1.32983588e-02 7.66605794e-01 4.45270061e-01 5.51279962e-01
-4.60619003e-01 -8.86951014e-02 7.16364741e-01 -1.67736128e-01
1.10877566e-01 3.97706062e-01 -5.62984049e-01 7.24805415e-01
9.10346389e-01 -5.73137164e-01 -5.30163161e-02 -2.66844809e-01
2.08583415e-01 -3.97811145e-01 7.07666278e-01 -1.29835188e+00
2.92199492e-01 4.91111614e-02 3.82705897e-01 -4.69007641e-01
-1.18435407e-02 -1.20088279e+00 1.46746352e-01 2.34640852e-01
-3.97381514e-01 -2.04382345e-01 2.74872363e-01 5.34491122e-01
-4.08323556e-01 -5.86799145e-01 8.44522476e-01 -8.84274915e-02
-1.28397632e+00 1.80851854e-02 -3.35677624e-01 1.75125390e-01
1.02743471e+00 -4.88048524e-01 -2.36244440e-01 -1.09394146e-02
-4.32285279e-01 -2.79452145e-01 6.71861053e-01 6.12044096e-01
2.73201197e-01 -9.39967275e-01 -9.25520182e-01 3.85182053e-01
7.88193643e-01 -3.87011409e-01 2.27995396e-01 6.64817274e-01
-6.09109998e-01 5.55544376e-01 -2.13871554e-01 -8.53718102e-01
-1.38027537e+00 4.25942302e-01 3.33692372e-01 -9.79954675e-02
-4.29839492e-01 5.16551316e-01 -2.05542311e-01 -4.00147796e-01
3.71357314e-02 2.35630479e-02 -1.47017539e-01 3.78067791e-01
3.91622990e-01 6.37255132e-01 3.79276365e-01 -1.04096425e+00
-4.25829560e-01 7.02570617e-01 -1.65996909e-01 2.26836860e-01
1.35389829e+00 5.83574027e-02 1.15914434e-01 5.97390056e-01
1.13756573e+00 4.53767955e-01 -1.13288403e+00 -2.19964117e-01
3.55115741e-01 -9.15570498e-01 -2.95263112e-01 -7.28210866e-01
-9.04889345e-01 6.68098152e-01 4.12783831e-01 3.42807919e-01
7.11742282e-01 -6.27043694e-02 7.75876194e-02 4.23863590e-01
5.64318299e-01 -1.11827075e+00 -2.94494152e-01 3.83801848e-01
8.48408341e-01 -1.55285132e+00 1.04031041e-01 -3.52679253e-01
-8.78833532e-01 1.36471617e+00 2.58779615e-01 -1.71636507e-01
2.76668191e-01 -4.15943302e-02 2.26526737e-01 2.20607072e-01
-1.63473994e-01 -7.71085694e-02 2.32483432e-01 6.48541987e-01
3.74414176e-01 -1.62308868e-02 -5.73533736e-02 3.74864548e-01
-6.58511184e-03 1.31439850e-01 7.80871034e-01 6.12998843e-01
-1.71229929e-01 -1.23016083e+00 -5.09501994e-01 4.77390766e-01
-5.44182718e-01 1.21417604e-01 -4.18170959e-01 1.37826908e+00
1.88108787e-01 7.46850550e-01 -8.97577256e-02 -3.38296920e-01
4.57028389e-01 2.56933898e-01 2.50438720e-01 -2.05091357e-01
-6.42679095e-01 -3.52602333e-01 4.82422188e-02 -9.61519107e-02
-5.60533822e-01 -1.02427328e+00 -4.92044121e-01 -3.06958258e-01
-3.06362122e-01 7.14439005e-02 6.45825267e-01 8.03458750e-01
9.20010358e-02 1.55203760e-01 7.02098370e-01 -4.91397381e-01
-6.03537679e-01 -6.52158976e-01 -9.31967080e-01 6.64419830e-01
1.67727113e-01 -5.46266079e-01 -6.19836688e-01 2.04751104e-01] | [9.839425086975098, -4.9075422286987305] |
e3e41dcb-e834-41c6-ba07-d456ab948317 | learning-reasoning-patterns-for-relational | null | null | https://aclanthology.org/2022.findings-acl.129 | https://aclanthology.org/2022.findings-acl.129.pdf | Learning Reasoning Patterns for Relational Triple Extraction with Mutual Generation of Text and Graph | Relational triple extraction is a critical task for constructing knowledge graphs. Existing methods focused on learning text patterns from explicit relational mentions. However, they usually suffered from ignoring relational reasoning patterns, thus failed to extract the implicitly implied triples. Fortunately, the graph structure of a sentence’s relational triples can help find multi-hop reasoning paths. Moreover, the type inference logic through the paths can be captured with the sentence’s supplementary relational expressions that represent the real-world conceptual meanings of the paths’ composite relations. In this paper, we propose a unified framework to learn the relational reasoning patterns for this task. To identify multi-hop reasoning paths, we construct a relational graph from the sentence (text-to-graph generation) and apply multi-layer graph convolutions to it. To capture the relation type inference logic of the paths, we propose to understand the unlabeled conceptual expressions by reconstructing the sentence from the relational graph (graph-to-text generation) in a self-supervised manner. Experimental results on several benchmark datasets demonstrate the effectiveness of our method. | ['Yongfeng Huang', 'Yunqi Zhang', 'Yubo Chen'] | null | null | null | null | findings-acl-2022-5 | ['relational-reasoning'] | ['natural-language-processing'] | [ 3.90158370e-02 7.99499393e-01 -6.35637224e-01 -6.11207902e-01
-2.89624780e-01 -5.40954590e-01 4.74222988e-01 4.43802893e-01
5.34314036e-01 3.98672342e-01 3.60831112e-01 -7.74416566e-01
-3.82323563e-01 -1.62166727e+00 -1.06953800e+00 7.43646771e-02
-1.18733473e-01 4.03882295e-01 2.39894569e-01 -3.83182853e-01
-2.65958607e-02 9.23796892e-02 -1.23465466e+00 8.70984137e-01
8.66937578e-01 9.18374658e-01 -3.16200078e-01 3.18933010e-01
-7.76147485e-01 1.86755490e+00 -2.00463787e-01 -9.85499859e-01
-1.52004078e-01 -4.72382963e-01 -1.28279293e+00 -3.85221839e-02
1.67315662e-01 -2.01508790e-01 -6.80210769e-01 1.06502819e+00
-5.28539717e-01 -1.72265172e-01 3.70503873e-01 -1.51273859e+00
-9.96827185e-01 1.48652732e+00 -2.77631432e-01 -7.69230872e-02
8.01267862e-01 -2.49408022e-01 1.74626899e+00 -6.92371190e-01
8.45343530e-01 1.44611800e+00 6.78209603e-01 3.17490011e-01
-8.12640846e-01 -4.38957393e-01 2.19827279e-01 4.56541866e-01
-1.31869733e+00 -2.68704832e-01 1.04128051e+00 -2.88782686e-01
1.21796572e+00 2.06401378e-01 5.48297048e-01 7.04513788e-01
-6.51729554e-02 9.23475087e-01 6.38097346e-01 -4.09842342e-01
-2.66732603e-01 9.05200243e-02 6.40068650e-01 1.35037136e+00
3.78357470e-01 -3.84312600e-01 -5.63142776e-01 -8.36341362e-03
5.32185316e-01 -5.76640144e-02 -7.37537742e-02 -2.31531888e-01
-1.02712667e+00 7.05937088e-01 8.27806354e-01 3.51707935e-01
-3.40112932e-02 2.76272178e-01 3.85256201e-01 4.41844314e-01
2.09327877e-01 1.81123301e-01 -7.11928248e-01 3.50245267e-01
-2.04489946e-01 1.52910337e-01 1.17960489e+00 1.40985346e+00
1.00942969e+00 -4.26874220e-01 -1.20382033e-01 3.91576558e-01
6.62425101e-01 1.97842106e-01 -3.64321321e-02 -7.10601807e-01
8.87179792e-01 1.53496683e+00 -2.37675935e-01 -1.39449298e+00
-4.15204465e-01 -7.39288926e-02 -6.47307336e-01 -8.37113619e-01
3.28400135e-01 1.69540104e-02 -4.59126621e-01 1.54581141e+00
5.91630399e-01 9.75815952e-02 5.15271962e-01 5.77119052e-01
1.27764833e+00 4.70609337e-01 -1.64963916e-01 9.30363592e-03
1.50944614e+00 -9.18809056e-01 -8.26688051e-01 -1.75687894e-01
1.13204467e+00 -7.65688866e-02 8.69497597e-01 -5.62015548e-02
-7.84943461e-01 -3.05640489e-01 -1.00321662e+00 -4.58953798e-01
-4.56008852e-01 -2.75387913e-02 1.03879404e+00 1.72122449e-01
-6.23171926e-01 5.52198768e-01 -6.37500167e-01 -1.96083590e-01
5.65201223e-01 1.40549183e-01 -3.56358945e-01 -3.51761371e-01
-1.57917583e+00 6.90622985e-01 7.98886657e-01 4.24129784e-01
-4.31900203e-01 -6.87974095e-01 -1.31887650e+00 2.71816522e-01
1.11714566e+00 -8.29572082e-01 1.00672305e+00 -6.30281270e-01
-9.88949060e-01 6.68605328e-01 -3.62180799e-01 -4.26718116e-01
7.33791813e-02 -4.98269759e-02 -5.92234850e-01 2.01494426e-01
3.32181543e-01 5.48042879e-02 3.02313864e-01 -1.24504697e+00
-4.79623705e-01 -5.03018916e-01 7.54587650e-01 -6.91828430e-02
-1.14190795e-01 -1.40705571e-01 -5.73635936e-01 -1.32719994e-01
4.99278426e-01 -4.89191830e-01 1.65468961e-01 -3.24662864e-01
-1.04630721e+00 -6.64910316e-01 6.95667982e-01 -4.51011300e-01
1.31354356e+00 -1.91395891e+00 5.89595875e-03 3.85481834e-01
5.69475889e-01 -2.52264351e-01 1.94610029e-01 6.12158358e-01
-1.95778161e-01 3.21600437e-01 -7.89727196e-02 3.06219965e-01
2.25229651e-01 6.28778934e-01 -6.78804934e-01 -1.07903942e-01
4.74477917e-01 1.42948246e+00 -1.09815729e+00 -6.10749304e-01
-2.05011830e-01 -7.19827227e-03 -4.08859104e-01 2.28386194e-01
-8.68826687e-01 -6.15588352e-02 -6.91274643e-01 6.41467333e-01
5.05611658e-01 -6.95291460e-01 8.37089121e-01 -6.91515326e-01
4.17478710e-01 6.93907917e-01 -8.68654966e-01 1.47656393e+00
-3.52790356e-01 2.57444143e-01 -5.03608048e-01 -1.10551059e+00
9.06142056e-01 5.85209988e-02 3.89249206e-01 -7.05759406e-01
-1.45277143e-01 -4.23558094e-02 -5.47363386e-02 -8.80918860e-01
3.50436211e-01 -3.82171482e-01 -1.73085883e-01 5.57041764e-01
1.52103633e-01 3.96535285e-02 4.51338351e-01 7.98389256e-01
1.17748761e+00 2.36427471e-01 2.49704525e-01 1.03549302e-01
7.02812672e-01 2.21264362e-01 5.96043050e-01 3.81554186e-01
4.82255310e-01 -1.03262819e-01 1.30413556e+00 -7.54460752e-01
-5.53942084e-01 -1.07982326e+00 2.97613204e-01 5.13640225e-01
2.60606140e-01 -1.15005624e+00 -3.47550213e-01 -1.15297234e+00
-1.02060497e-01 7.90280342e-01 -5.34052372e-01 -3.71374309e-01
-6.66225791e-01 -4.25357223e-01 7.85007656e-01 6.13089025e-01
4.73839819e-01 -9.15055156e-01 1.33235201e-01 -8.46693572e-03
-6.07038677e-01 -1.65544331e+00 8.97557959e-02 -1.22871786e-01
-6.58401489e-01 -1.82503998e+00 6.77046359e-01 -7.48754382e-01
9.97106254e-01 4.92753237e-02 1.32367289e+00 6.04759395e-01
2.44948953e-01 2.24335983e-01 -4.10294116e-01 -4.62538116e-02
-3.71149063e-01 -3.80964112e-03 -3.79927933e-01 5.17044254e-02
6.24098241e-01 -3.72474700e-01 -3.66215408e-02 -5.39631732e-02
-9.36930060e-01 3.38807166e-01 3.38332921e-01 5.21905124e-01
5.79558730e-01 8.71385157e-01 3.76828194e-01 -1.67691350e+00
6.61442161e-01 -6.77200377e-01 -5.34794152e-01 7.73559868e-01
-6.34310663e-01 4.25068736e-01 8.17422450e-01 1.27914235e-01
-1.25460362e+00 -2.11427823e-01 1.25762120e-01 -2.27046877e-01
7.17753023e-02 1.21977830e+00 -6.62505031e-01 5.30745089e-01
3.20602208e-01 9.14648548e-02 -4.27588582e-01 -6.11174330e-02
7.81363130e-01 1.95374534e-01 3.62173945e-01 -1.25950289e+00
9.77738798e-01 3.69193256e-01 2.82158971e-01 -3.48778456e-01
-1.52079117e+00 -3.60393152e-02 -6.76954865e-01 1.10529259e-01
7.50259161e-01 -7.79500604e-01 -9.14367795e-01 1.59571636e-02
-1.36916208e+00 -3.28556120e-01 -2.51453936e-01 -2.84992680e-02
-3.07994604e-01 3.99982214e-01 -7.44915962e-01 -7.26295650e-01
-1.89993337e-01 -8.30444336e-01 8.24441254e-01 7.77166337e-02
-2.38947302e-01 -1.19297683e+00 -2.09320039e-01 4.96451378e-01
-3.41561317e-01 2.66558230e-01 1.69691634e+00 -7.05320001e-01
-1.03320658e+00 -5.10165691e-02 -7.24662244e-01 -9.22706127e-02
4.89868164e-01 5.23428954e-02 -5.98109126e-01 5.12636840e-01
-2.95156300e-01 -2.91371703e-01 4.11486983e-01 -2.15521112e-01
1.13948691e+00 -8.50430071e-01 -4.24270272e-01 3.81263614e-01
1.38140559e+00 -4.91732731e-02 5.95604122e-01 -7.81637356e-02
1.18471563e+00 8.39507461e-01 2.75884032e-01 -6.10987470e-03
1.13560605e+00 1.59473717e-01 2.69729763e-01 2.79863268e-01
-8.50779563e-02 -1.05791092e+00 2.06351608e-01 7.89946675e-01
6.54669181e-02 -1.58450399e-02 -1.05457461e+00 3.85535240e-01
-2.05610561e+00 -9.68888819e-01 -7.87995934e-01 1.53480613e+00
9.74733829e-01 3.53915572e-01 -2.33674675e-01 1.27128005e-01
4.14887011e-01 7.86525197e-03 -3.99541855e-01 -1.39758721e-01
-1.17080696e-01 1.90935389e-03 5.25974901e-04 6.44451439e-01
-7.14152157e-01 1.10681307e+00 5.36019373e+00 3.70923191e-01
-5.30916929e-01 -1.69101968e-01 2.29368761e-01 2.87038863e-01
-1.03070962e+00 6.52898371e-01 -7.60223508e-01 3.21078598e-02
7.52570570e-01 -2.95668602e-01 4.88937020e-01 5.94374895e-01
-3.47093761e-01 1.29289255e-01 -1.52471912e+00 5.37285805e-01
-5.61002530e-02 -1.61450326e+00 4.40937310e-01 -1.82674438e-01
4.37213570e-01 -4.61370975e-01 -5.16099334e-01 5.58432519e-01
5.85473597e-01 -9.28346038e-01 6.14880323e-01 6.42557204e-01
6.03356719e-01 -6.79426789e-01 6.36565924e-01 2.73760378e-01
-1.63379335e+00 -1.13695510e-01 -2.25337565e-01 -1.97911650e-01
-1.32956505e-01 9.02772009e-01 -9.08470631e-01 1.33136845e+00
5.39415419e-01 1.21846771e+00 -6.39494359e-01 7.59715214e-03
-8.94703209e-01 3.24170172e-01 -7.53264055e-02 -1.19296148e-01
1.18489027e-01 -3.53893191e-01 4.09320518e-02 9.41692352e-01
-3.70138464e-03 3.59126657e-01 1.03875652e-01 1.39655840e+00
-6.04164064e-01 -1.85808644e-01 -1.01153576e+00 -5.84133148e-01
3.00822645e-01 1.20814347e+00 -7.31317163e-01 -4.72759396e-01
-7.81643808e-01 5.41256070e-01 8.73326063e-01 4.80431199e-01
-6.40793562e-01 -4.36596870e-01 3.53307396e-01 1.28579184e-01
1.29069909e-01 -1.48549691e-01 -2.21221805e-01 -1.46278548e+00
3.99299860e-01 -7.01963603e-01 7.69748747e-01 -1.06693769e+00
-1.49720359e+00 3.43968838e-01 1.67725429e-01 -7.36297905e-01
-6.51337625e-03 -5.82661271e-01 -6.42592609e-01 6.51090086e-01
-1.38999605e+00 -1.55224895e+00 -1.02962814e-01 8.12122047e-01
6.24533817e-02 1.19980872e-01 7.53794730e-01 4.26142961e-02
-5.31260014e-01 3.82266909e-01 -1.00204313e+00 7.76032448e-01
1.88995916e-02 -1.27538800e+00 2.78053463e-01 9.66697037e-01
4.76686180e-01 1.24362338e+00 3.00559789e-01 -8.69253814e-01
-2.04730606e+00 -1.09979486e+00 1.24736547e+00 -5.87400675e-01
1.35122323e+00 -3.33825767e-01 -1.12180948e+00 1.50233400e+00
7.75195658e-02 -2.74022315e-02 8.19393277e-01 6.22988105e-01
-8.79363716e-01 -2.49304116e-01 -7.31842756e-01 7.25886524e-01
1.49131787e+00 -9.79746580e-01 -9.04414356e-01 5.03678143e-01
1.33754015e+00 -5.15026331e-01 -1.01624823e+00 4.30990726e-01
2.84621954e-01 -7.02322364e-01 8.12274814e-01 -1.10742426e+00
1.08258736e+00 -5.59058964e-01 -2.72823900e-01 -9.46740687e-01
-3.80673259e-02 -3.52556467e-01 -7.99986959e-01 1.48342371e+00
7.94588804e-01 -4.50194359e-01 7.00333297e-01 8.06995690e-01
-1.39214601e-02 -8.98745537e-01 -3.57127726e-01 -2.40244865e-01
-2.35106632e-01 -7.83368528e-01 1.00065374e+00 1.22880018e+00
6.84920907e-01 1.00605047e+00 -3.89463976e-02 5.12897491e-01
6.17225051e-01 7.95725524e-01 6.42869711e-01 -1.14182496e+00
-2.71228194e-01 -1.71710283e-01 -1.65395722e-01 -8.65693688e-01
7.15507090e-01 -1.43573916e+00 -2.95119196e-01 -2.07090902e+00
1.35931909e-01 -4.98079479e-01 -3.91312279e-02 8.04981649e-01
-1.50022969e-01 -5.01610756e-01 -1.21656984e-01 -6.26426190e-02
-5.92989445e-01 3.60553861e-01 1.55951095e+00 -5.90254843e-01
1.87446028e-01 -3.04988950e-01 -1.25022483e+00 7.23690212e-01
5.45220077e-01 -3.94520104e-01 -1.01872516e+00 -5.99311352e-01
1.17974305e+00 4.23648953e-01 3.94895643e-01 -3.88388872e-01
5.88483155e-01 -4.90475208e-01 -1.22273974e-02 -5.72136223e-01
-4.78890352e-02 -9.08918202e-01 1.66188270e-01 2.68043131e-01
-4.45881486e-01 -7.38428980e-02 -1.38794154e-01 6.06757224e-01
-3.86948496e-01 -1.06289357e-01 5.32298610e-02 -3.05450499e-01
-5.98485112e-01 3.37048233e-01 2.70679861e-01 2.33882278e-01
7.37272263e-01 2.64679551e-01 -6.05134785e-01 -1.26944661e-01
-7.61205196e-01 5.09481609e-01 1.36189550e-01 4.25287753e-01
8.44285846e-01 -1.39591372e+00 -2.80231059e-01 2.39277273e-01
3.49777013e-01 6.24049246e-01 8.18481967e-02 8.40147018e-01
-3.95147681e-01 2.37839788e-01 2.57860601e-01 -1.82147712e-01
-9.13231313e-01 9.27448809e-01 4.95789140e-01 -5.30277252e-01
-7.61646807e-01 5.90359092e-01 2.92266477e-02 -7.53375709e-01
-9.22753438e-02 -8.00430059e-01 -3.00012439e-01 -1.62429124e-01
3.95898908e-01 -8.92547667e-02 1.82874722e-03 -3.55096906e-01
-4.05898333e-01 4.04131472e-01 1.59765524e-03 3.33851516e-01
1.21165931e+00 2.42765043e-02 -9.06443357e-01 5.24822176e-01
9.59755182e-01 1.74474478e-01 -5.53715229e-01 -6.36506140e-01
5.28960407e-01 -3.16980362e-01 -4.31046277e-01 -4.98278260e-01
-1.12173307e+00 6.07269347e-01 -8.48957062e-01 4.93494034e-01
7.75648177e-01 4.70308185e-01 8.82467210e-01 9.17286932e-01
3.62673610e-01 -6.00842178e-01 5.91401383e-02 7.36624897e-01
6.94314599e-01 -1.05680597e+00 5.31786215e-03 -1.13383365e+00
-5.65336823e-01 1.35646033e+00 8.32000732e-01 1.60955712e-01
7.34205127e-01 3.77059579e-01 -1.77307099e-01 -8.69383752e-01
-8.37048411e-01 -1.93007380e-01 2.80730784e-01 4.51204985e-01
4.12205368e-01 1.61480725e-01 1.73296496e-01 9.11027789e-01
-3.96605581e-01 1.51983932e-01 4.67038929e-01 7.03654170e-01
-9.51578990e-02 -1.18416762e+00 1.51096508e-01 5.06702840e-01
-7.78578073e-02 -1.84131444e-01 -8.40984523e-01 7.33994246e-01
5.14252596e-02 1.11483824e+00 -2.09433079e-01 -6.55251205e-01
4.64592516e-01 1.83728531e-01 6.76339686e-01 -7.91246712e-01
-2.85199553e-01 -5.42082787e-01 6.34115636e-01 -5.88139653e-01
-5.43173015e-01 -2.71088421e-01 -1.98078787e+00 -5.22621870e-01
-4.94121909e-02 3.17860276e-01 -1.54915042e-02 1.47560823e+00
1.62691727e-01 8.43821228e-01 3.68949294e-01 4.01245981e-01
2.99965274e-02 -6.32638276e-01 -4.57231760e-01 6.78911388e-01
6.63515925e-03 -4.93859440e-01 4.97798398e-02 1.28621355e-01] | [9.10258674621582, 7.915329456329346] |
fb7dff59-1b31-4f3c-8a57-231a0afc9268 | semi-supervised-stance-detection-of-tweets | 2201.00614 | null | https://arxiv.org/abs/2201.00614v2 | https://arxiv.org/pdf/2201.00614v2.pdf | Semi-supervised Stance Detection of Tweets Via Distant Network Supervision | Detecting and labeling stance in social media text is strongly motivated by hate speech detection, poll prediction, engagement forecasting, and concerted propaganda detection. Today's best neural stance detectors need large volumes of training data, which is difficult to curate given the fast-changing landscape of social media text and issues on which users opine. Homophily properties over the social network provide strong signal of coarse-grained user-level stance. But semi-supervised approaches for tweet-level stance detection fail to properly leverage homophily. In light of this, We present SANDS, a new semi-supervised stance detector. SANDS starts from very few labeled tweets. It builds multiple deep feature views of tweets. It also uses a distant supervision signal from the social network to provide a surrogate loss signal to the component learners. We prepare two new tweet datasets comprising over 236,000 politically tinted tweets from two demographics (US and India) posted by over 87,000 users, their follower-followee graph, and over 8,000 tweets annotated by linguists. SANDS achieves a macro-F1 score of 0.55 (0.49) on US (India)-based datasets, outperforming 17 baselines (including variants of SANDS) substantially, particularly for minority stance labels and noisy text. Numerous ablation experiments on SANDS disentangle the dynamics of textual and network-propagated stance signals. | ['Tanmoy Chakraborty', 'Soumen Chakrabarti', 'Samiya Caur', 'Subhabrata Dutta'] | 2022-01-03 | null | null | null | null | ['propaganda-detection'] | ['natural-language-processing'] | [-9.91404951e-02 4.12458688e-01 -6.72095358e-01 -6.10915124e-01
-8.58911693e-01 -7.92197287e-01 1.12815726e+00 5.06629109e-01
-4.76117641e-01 8.20801675e-01 8.89911234e-01 -3.58886003e-01
4.73903000e-01 -9.79047418e-01 -3.99642825e-01 -4.11118865e-01
3.61840948e-02 5.22107601e-01 2.54220795e-02 -7.35925019e-01
2.63465345e-01 -2.95384943e-01 -8.35082591e-01 5.48244059e-01
9.61518884e-01 6.06629550e-01 -7.37599194e-01 4.40048903e-01
-1.36641964e-01 1.26216471e+00 -7.45031953e-01 -7.93840110e-01
-1.32175028e-01 -4.89977062e-01 -7.31106997e-01 -4.70404446e-01
7.95939147e-01 -7.19645172e-02 -3.81192863e-01 9.58477318e-01
5.98653495e-01 -3.84175807e-01 6.87853217e-01 -1.04279828e+00
-6.71318114e-01 1.44327927e+00 -9.62482274e-01 5.36776721e-01
8.14548656e-02 4.31627929e-02 1.57071483e+00 -7.95925379e-01
9.44301248e-01 1.53464139e+00 9.41550434e-01 2.65475810e-01
-1.39464664e+00 -1.00111330e+00 1.78180173e-01 -3.99824888e-01
-6.58710599e-01 -4.34134990e-01 1.00210917e+00 -7.23434508e-01
2.80225009e-01 2.42751554e-01 6.96074069e-01 1.99297619e+00
3.03647798e-02 1.06927001e+00 1.69433331e+00 1.70483634e-01
-1.10061556e-01 3.39896023e-01 9.45323646e-01 7.12357640e-01
2.66294807e-01 -1.25671804e-01 -9.61799026e-01 -7.70860970e-01
-1.74162552e-01 -3.34409624e-01 9.43071172e-02 3.59614342e-01
-1.02030265e+00 1.64844096e+00 8.56671333e-01 2.93163747e-01
-7.03393444e-02 -6.77700564e-02 6.68370605e-01 5.71751535e-01
1.34174490e+00 6.51823521e-01 -6.28164560e-02 -8.14788267e-02
-1.24584126e+00 3.87284666e-01 1.10142469e+00 1.62708983e-01
6.75628960e-01 8.14296454e-02 -6.10170662e-02 6.55279338e-01
1.21861383e-01 8.87105286e-01 2.37268135e-01 -5.40079176e-01
7.21655786e-01 6.90582156e-01 -1.49479166e-01 -1.56919122e+00
-6.97533309e-01 -8.98838758e-01 -6.96897745e-01 -2.77536035e-01
6.26428246e-01 -6.94440544e-01 -3.40570152e-01 1.80634534e+00
3.66103709e-01 -2.32944429e-01 -3.89803737e-01 5.78832686e-01
9.15826559e-01 4.76250023e-01 -7.60634392e-02 -1.95309728e-01
1.20477915e+00 -6.82516515e-01 -5.81293762e-01 -4.82984662e-01
6.67352021e-01 -8.55279028e-01 1.08118427e+00 9.09377784e-02
-1.00170040e+00 -7.02547655e-02 -9.52220142e-01 3.68890390e-02
-6.00656509e-01 -5.31841815e-01 4.85152304e-01 5.46398282e-01
-5.71637094e-01 7.70094335e-01 -3.93975317e-01 -1.16315491e-01
8.77005696e-01 -8.56526420e-02 1.44785896e-01 3.64984691e-01
-1.70059788e+00 9.86986637e-01 -2.91008919e-01 -1.09450012e-01
-6.62648439e-01 -7.45030224e-01 -5.84124982e-01 -5.71668565e-01
1.85800016e-01 -3.15473139e-01 1.03871644e+00 -1.14554250e+00
-1.04103744e+00 1.37782145e+00 -1.56836793e-01 -8.28945398e-01
8.92062366e-01 -4.63738918e-01 -4.94394779e-01 -1.04291543e-01
5.51494658e-01 -2.97151543e-02 1.00264049e+00 -8.97761106e-01
-2.42915466e-01 -3.18513066e-01 -1.22181714e-01 -1.44745689e-02
-3.99695426e-01 2.50235200e-01 7.66349018e-01 -4.15099829e-01
-1.15315549e-01 -9.36379969e-01 2.28767887e-01 -6.34510338e-01
-9.62749481e-01 -3.87807935e-01 1.10621536e+00 -5.44759333e-01
1.31997693e+00 -1.77768230e+00 -1.36709914e-01 4.76615369e-01
8.67873371e-01 2.38390684e-01 1.86286926e-01 5.73800087e-01
1.82044923e-01 3.43280882e-01 1.18059270e-01 -4.47850138e-01
2.30830535e-01 -1.10398412e-01 -6.29963696e-01 1.02458036e+00
-2.61970330e-02 1.06701529e+00 -1.22149861e+00 -4.14679706e-01
-3.21349829e-01 2.16171369e-01 -6.00218892e-01 -1.90525398e-01
-2.71220982e-01 3.80213737e-01 -4.19554412e-01 6.54850781e-01
2.47856617e-01 -5.50004423e-01 2.59465873e-01 -1.22362087e-02
-2.03471810e-01 9.59659517e-01 -1.69972628e-01 8.40821445e-01
-2.48339310e-01 1.22768068e+00 2.93040395e-01 -7.52657950e-01
1.08158004e+00 -1.04454011e-01 2.16972798e-01 -6.58512712e-01
6.26838863e-01 4.25226986e-01 2.21312791e-01 -1.05745688e-01
3.82370412e-01 -1.61482826e-01 -5.84322453e-01 1.06880820e+00
-3.39017272e-01 2.17256576e-01 -4.34421413e-02 7.65416205e-01
1.05406392e+00 -4.30469245e-01 9.56465751e-02 -6.80212259e-01
1.07036620e-01 6.56655505e-02 3.01325351e-01 8.95593584e-01
-4.70135570e-01 2.41246670e-01 1.01580787e+00 -4.89876240e-01
-1.07353485e+00 -9.11606431e-01 -1.25914901e-01 1.72673690e+00
-2.89011329e-01 -5.05501807e-01 -6.58303082e-01 -9.16614413e-01
3.07244331e-01 5.93751431e-01 -1.06939757e+00 3.80952433e-02
-1.02872431e+00 -1.12674963e+00 7.88047791e-01 -8.22516382e-02
5.42540669e-01 -9.41346467e-01 -4.13105376e-02 1.32908329e-01
-6.37207985e-01 -8.83953273e-01 -3.63494188e-01 1.21001817e-01
-2.41714999e-01 -1.37283611e+00 -2.84965962e-01 -4.75594640e-01
1.84764311e-01 5.63079007e-02 1.47426856e+00 1.09090157e-01
3.33870709e-01 -3.61716241e-01 -1.45455346e-01 -4.99916792e-01
-7.16520429e-01 5.46959639e-01 2.35536024e-01 1.19499780e-01
6.14126742e-01 -7.38194227e-01 -4.95193273e-01 1.99631885e-01
-2.68394232e-01 -1.76478371e-01 5.60190901e-02 9.67841804e-01
-1.63824737e-01 -5.51224709e-01 8.39566886e-01 -1.79498672e+00
1.03611362e+00 -9.98573720e-01 -1.97487101e-01 -3.11261475e-01
-5.52585959e-01 -3.63934934e-01 5.44932485e-01 -3.44288260e-01
-7.29449689e-01 -8.04088175e-01 -4.40022312e-02 3.38676602e-01
3.34094614e-01 6.04186356e-01 5.78841984e-01 3.76833558e-01
1.36484849e+00 -3.16786826e-01 1.35544345e-01 -3.02726358e-01
4.24365580e-01 9.01625991e-01 3.62052530e-01 -3.19182932e-01
1.14907444e+00 6.35133326e-01 -5.41100144e-01 -1.09075022e+00
-1.93364894e+00 -3.96979272e-01 -3.21229100e-01 -1.53815106e-01
6.46689475e-01 -9.79307413e-01 -7.61609852e-01 5.91641963e-01
-1.00286925e+00 -3.94626647e-01 4.40860055e-02 8.42781924e-03
-2.00316794e-02 -2.51827738e-03 -1.14822805e+00 -6.41386211e-01
-7.18986511e-01 -5.40960073e-01 5.94163537e-01 -2.00882196e-01
-1.07394099e+00 -1.25679564e+00 3.82140338e-01 6.96581662e-01
6.23429835e-01 9.56116855e-01 7.55485952e-01 -1.16916656e+00
1.56244859e-01 -2.26220220e-01 -3.10565948e-01 2.07786143e-01
-4.76658791e-02 2.14438975e-01 -1.16952658e+00 -2.58868515e-01
-1.38763666e-01 -9.36011851e-01 1.19347811e+00 4.03921396e-01
4.87285405e-01 -7.83790231e-01 -2.38124683e-01 1.77328542e-01
7.25990236e-01 -8.07007372e-01 3.20112705e-02 3.89707327e-01
6.50805354e-01 6.64398491e-01 2.42290378e-01 3.09733510e-01
2.96869397e-01 1.47480175e-01 3.80851179e-01 -2.88736969e-01
-8.50708503e-03 -6.74661040e-01 7.08544493e-01 1.04535258e+00
2.75832146e-01 -1.47828117e-01 -1.01400363e+00 3.42861593e-01
-1.80282903e+00 -1.48987007e+00 -5.83879411e-01 1.56608784e+00
1.22041965e+00 9.03268695e-01 6.99509978e-01 9.95161086e-02
7.30379522e-01 1.00474203e+00 -4.65416968e-01 -5.35647988e-01
-6.08165622e-01 9.61446911e-02 5.01028895e-01 1.13786566e+00
-1.35800409e+00 1.02289736e+00 6.06951475e+00 5.51281929e-01
-1.17231655e+00 3.42058152e-01 8.15276206e-01 -5.00031233e-01
-5.45742571e-01 -3.26351821e-01 -9.47763979e-01 6.38019383e-01
1.16039073e+00 -2.21956417e-01 -2.35108677e-02 8.61754656e-01
4.27546173e-01 8.77962187e-02 -7.69882441e-01 4.68811035e-01
1.43847004e-01 -1.71372080e+00 -4.81369972e-01 2.60968238e-01
1.10170591e+00 9.04738605e-01 4.90186095e-01 3.87707204e-01
1.03129613e+00 -9.99176145e-01 7.89517403e-01 -8.97148401e-02
5.42240143e-01 -7.22210050e-01 6.22889638e-01 6.06918991e-01
-4.59091127e-01 -1.11131206e-01 -8.89827758e-02 -4.39539343e-01
1.32069483e-01 1.18952096e+00 -9.65147793e-01 -3.87448311e-01
2.91181177e-01 1.10463798e+00 -5.27194023e-01 -8.19029301e-05
-5.44233799e-01 1.42114508e+00 -4.67168331e-01 -5.34837604e-01
9.18933570e-01 1.43247098e-01 7.64281034e-01 1.56722713e+00
-5.39480388e-01 -1.33088827e-01 1.82805359e-01 9.01750624e-01
-5.99290609e-01 -9.47946459e-02 -7.72799909e-01 -2.86774635e-01
6.16803825e-01 1.35768843e+00 -4.18535113e-01 -5.08989811e-01
-9.97293890e-02 3.71107936e-01 7.21257150e-01 2.04870909e-01
-8.40612352e-01 -9.51950550e-02 7.03014076e-01 6.82137847e-01
-5.07040441e-01 -1.53383225e-01 -3.78510833e-01 -1.18002355e+00
-5.17708063e-01 -1.18930864e+00 2.62846053e-01 -1.43980235e-01
-1.77664053e+00 4.71839190e-01 -3.77172559e-01 -5.01020074e-01
-3.84218812e-01 -2.81893969e-01 -9.61586893e-01 8.02721918e-01
-1.45266807e+00 -1.20801616e+00 -1.73132680e-02 4.61974978e-01
2.94211149e-01 -1.42030358e-01 5.44084370e-01 1.57255992e-01
-6.09162211e-01 6.13457859e-01 -6.27888814e-02 8.28264892e-01
9.25223529e-01 -1.29407895e+00 6.91533089e-01 4.34130073e-01
1.49973100e-02 7.96138525e-01 9.70230341e-01 -7.96855032e-01
-7.31220543e-01 -1.11910295e+00 1.40111971e+00 -9.68868613e-01
1.54776752e+00 -7.33304679e-01 -7.17574358e-01 8.00117552e-01
4.35537368e-01 -4.20172244e-01 1.02300906e+00 8.75974655e-01
-1.04203677e+00 2.69902498e-01 -8.92290354e-01 6.41102016e-01
1.00471544e+00 -9.02319729e-01 -7.84441948e-01 8.82065177e-01
5.55863261e-01 -2.77510792e-01 -4.01607126e-01 -3.70173715e-02
6.25296175e-01 -1.04691875e+00 6.69269443e-01 -9.27761018e-01
1.01488042e+00 3.87113005e-01 6.86318241e-03 -1.46308815e+00
-3.67864788e-01 -9.03289080e-01 -2.23654628e-01 1.06143010e+00
7.31169641e-01 -7.61847496e-01 1.03203678e+00 1.68562710e-01
1.53198868e-01 -7.67490029e-01 -4.59374070e-01 -2.56354153e-01
5.98242164e-01 4.70485887e-04 9.40179452e-02 1.59114206e+00
3.07161421e-01 1.14932489e+00 -6.02063179e-01 -3.91608715e-01
9.09930050e-01 4.81220633e-01 8.33397031e-01 -1.58122075e+00
-1.15767516e-01 -7.86624432e-01 2.61905313e-01 -8.05819571e-01
4.85286653e-01 -1.11877310e+00 -4.70972538e-01 -1.13090932e+00
3.06891471e-01 -5.02691031e-01 1.67625815e-01 2.34923691e-01
-1.94585021e-03 6.85693502e-01 -6.71328679e-02 4.04688418e-01
-6.13717020e-01 1.94257453e-01 1.08969569e+00 -3.59481066e-01
-1.60485357e-02 1.13750525e-01 -1.03957999e+00 1.32247794e+00
9.83344197e-01 -6.67425931e-01 -1.72105879e-02 -1.71748012e-01
9.13571835e-01 -4.61789489e-01 2.69178092e-01 -3.20833027e-01
-5.03307395e-02 -1.90866832e-02 2.59215832e-01 -8.50378275e-01
3.50670964e-01 1.16786063e-01 -6.37801349e-01 5.90279162e-01
-7.94099867e-01 -9.81226265e-02 -4.17348295e-01 6.46104813e-01
-1.64878190e-01 3.63263041e-01 1.09043980e+00 -3.23731333e-01
1.13410108e-01 1.14501491e-01 -3.41688782e-01 1.01134491e+00
4.01210099e-01 2.50295639e-01 -1.17617583e+00 -6.58434868e-01
-7.07823634e-01 2.74263859e-01 2.00837985e-01 3.12313437e-01
1.12365305e-01 -1.05318367e+00 -1.31687129e+00 -1.18782029e-01
-2.05399111e-01 -4.45455343e-01 -2.53463149e-01 1.13469243e+00
-1.92043021e-01 9.29691717e-02 2.58890003e-01 -4.37056184e-01
-1.18948674e+00 -5.63736483e-02 1.86241046e-01 -3.27236593e-01
-4.20534849e-01 1.14659417e+00 -2.89979950e-02 -6.41126275e-01
-8.16882029e-02 1.81098863e-01 -6.41386583e-02 8.86179090e-01
6.43778980e-01 4.06735629e-01 -1.56821817e-01 -8.60888660e-01
-2.35949308e-01 6.72111381e-03 -2.78854430e-01 -4.19564254e-04
1.25279665e+00 -6.72066286e-02 -2.32007712e-01 8.33840787e-01
1.54796290e+00 5.99476635e-01 -6.32615328e-01 -6.67249203e-01
1.25430927e-01 -2.31374517e-01 -9.13820416e-02 -7.08301008e-01
-7.02013671e-01 9.03790116e-01 -4.30367440e-01 7.47796059e-01
-1.24611318e-01 1.43463135e-01 1.00568187e+00 2.90144503e-01
-2.65493877e-02 -1.39665210e+00 4.80139464e-01 9.92565274e-01
6.84290826e-01 -1.40462089e+00 1.72248825e-01 -5.75367771e-02
-6.23347938e-01 6.70639694e-01 3.83000582e-01 -4.42916334e-01
9.05088246e-01 2.00640127e-01 3.23610008e-01 -6.07193291e-01
-7.67066061e-01 1.42066419e-01 -4.10140567e-02 1.34287536e-01
7.86796212e-01 2.73119122e-01 -3.97470117e-01 3.75220895e-01
-9.26793694e-01 -8.27052832e-01 4.94024962e-01 4.59389597e-01
-8.97691846e-01 -5.98444164e-01 -1.81960121e-01 7.14422882e-01
-8.42595518e-01 -3.60305637e-01 -1.08333552e+00 7.55862057e-01
-1.44209012e-01 1.07294691e+00 -4.11397330e-02 -5.16572714e-01
-1.23788759e-01 8.34001973e-02 -2.11896747e-01 -6.14798427e-01
-1.05906487e+00 -1.54414803e-01 9.04515028e-01 -5.12980342e-01
-4.59246129e-01 -7.39742339e-01 -8.19231451e-01 -1.36045146e+00
-2.99339890e-01 3.50802273e-01 4.32043701e-01 8.47895324e-01
1.17684584e-02 1.22427093e-02 9.36192453e-01 -5.83005667e-01
-7.96194911e-01 -1.11277306e+00 -3.50317001e-01 7.00314522e-01
6.96800232e-01 -3.99079770e-01 -8.05466771e-01 -5.78942180e-01] | [8.733528137207031, 10.199392318725586] |
f6198630-5f5d-418e-b25f-92e018f77e44 | directional-sparse-filtering-using-weighted | 2102.00196 | null | https://arxiv.org/abs/2102.00196v3 | https://arxiv.org/pdf/2102.00196v3.pdf | Directional Sparse Filtering using Weighted Lehmer Mean for Blind Separation of Unbalanced Speech Mixtures | In blind source separation of speech signals, the inherent imbalance in the source spectrum poses a challenge for methods that rely on single-source dominance for the estimation of the mixing matrix. We propose an algorithm based on the directional sparse filtering (DSF) framework that utilizes the Lehmer mean with learnable weights to adaptively account for source imbalance. Performance evaluation in multiple real acoustic environments show improvements in source separation compared to the baseline methods. | ['Andy W. H. Khong', 'Ching-Hui Ooi', 'Anh H. T. Nguyen', 'Karn Watcharasupat'] | 2021-01-30 | directional-sparse-filtering-using-weighted-1 | https://arxiv.org/abs/2102.00196 | https://arxiv.org/abs/2102.00196 | null | ['audio-source-separation', 'multi-speaker-source-separation'] | ['audio', 'speech'] | [ 4.60703135e-01 -5.30445516e-01 1.43102422e-01 -2.59496957e-01
-1.40437639e+00 -6.41450346e-01 4.25330579e-01 -3.37694079e-01
1.40284777e-01 4.49842006e-01 8.48532319e-01 -6.98740557e-02
-5.44655740e-01 7.06166029e-02 -3.92884135e-01 -9.58679557e-01
-2.24152908e-01 -7.89151415e-02 -3.73789705e-02 -8.88838433e-03
-1.50604360e-02 1.96312368e-01 -1.49050057e+00 1.02559626e-01
8.79568756e-01 9.49578583e-01 3.79117578e-01 1.03423059e+00
-3.55113707e-02 6.23197794e-01 -9.25135970e-01 1.24900892e-01
4.88773048e-01 -8.13677013e-01 8.77441019e-02 -2.28394140e-02
4.83749419e-01 2.21075155e-02 -2.82866359e-01 1.46536207e+00
8.68516922e-01 3.16760242e-01 5.75464249e-01 -9.86190438e-01
-2.98309326e-02 1.04570329e+00 -3.33049685e-01 6.66358709e-01
3.21507543e-01 -4.70182210e-01 8.27034771e-01 -1.35985887e+00
-5.96076213e-02 1.20448124e+00 9.98627424e-01 1.00227281e-01
-1.23414242e+00 -8.05822849e-01 -2.14277953e-01 2.45654508e-01
-1.40125275e+00 -1.44593143e+00 1.24229431e+00 -2.03349099e-01
7.59926260e-01 4.32941139e-01 1.98646873e-01 9.50306535e-01
-4.35233623e-01 6.08537674e-01 9.59485471e-01 -7.96555638e-01
3.73691648e-01 -2.21622035e-01 -4.83512767e-02 1.05567984e-01
2.72935063e-01 2.95386702e-01 -1.27155089e+00 -4.59927976e-01
4.35182899e-01 -6.81628287e-01 -5.98271012e-01 -1.63474143e-01
-1.10595453e+00 6.54868901e-01 4.82595228e-02 4.27601725e-01
-3.47842127e-01 1.61809817e-01 8.10628682e-02 2.77672380e-01
6.45594180e-01 4.41191852e-01 -2.29601279e-01 -1.77231692e-02
-1.61287189e+00 2.28618667e-01 9.06199634e-01 6.88581824e-01
3.12639832e-01 9.84862387e-01 -8.85839015e-02 1.20657825e+00
7.18127131e-01 1.03564024e+00 5.85991323e-01 -1.22961807e+00
3.50173235e-01 -2.79737592e-01 2.05095008e-01 -1.02092898e+00
-1.28859118e-01 -1.02250767e+00 -8.12182188e-01 4.07855324e-02
3.81229192e-01 -2.92114288e-01 -9.34064686e-01 1.77098572e+00
3.85735691e-01 8.90510201e-01 4.63552177e-01 8.45554411e-01
5.80969751e-01 6.57103956e-01 -5.93662202e-01 -5.80203056e-01
7.47346878e-01 -9.90815103e-01 -1.20304704e+00 -5.67103744e-01
-4.21795011e-01 -1.19519329e+00 3.99397403e-01 7.47663975e-01
-1.14497793e+00 -4.51919854e-01 -1.29156876e+00 3.48226398e-01
1.53903902e-01 8.00657347e-02 2.70290464e-01 1.25911105e+00
-1.04919219e+00 2.35985279e-01 -6.20633602e-01 2.59545416e-01
-6.68647140e-02 -3.84959951e-03 -8.79457071e-02 7.54142329e-02
-1.07074785e+00 4.31697667e-01 -1.28165483e-01 2.92003006e-01
-1.08087885e+00 -9.05842304e-01 -8.22368562e-01 5.08493334e-02
6.57565445e-02 -1.02927551e-01 1.29769719e+00 -8.46948922e-01
-1.73459375e+00 1.32002393e-02 -6.48451209e-01 -8.04051697e-01
1.91619962e-01 -4.30137515e-01 -9.01604831e-01 3.35851282e-01
1.16029859e-01 -2.24939153e-01 1.76821685e+00 -1.35622311e+00
-3.67258191e-01 -4.12788540e-02 -7.15670288e-01 2.48008952e-01
-2.51504928e-01 2.12461531e-01 -1.42435223e-01 -1.30055535e+00
7.35680401e-01 -6.22414708e-01 -2.35019699e-01 -5.61540425e-01
-3.99824202e-01 3.35682690e-01 6.45769060e-01 -1.03152752e+00
1.27764738e+00 -2.67033887e+00 1.01547770e-01 4.94184822e-01
-1.11672305e-01 1.51288942e-01 -1.37793019e-01 4.26411688e-01
-2.33027861e-01 -6.49047077e-01 -4.51544225e-01 -6.29886329e-01
2.48038899e-02 -1.88854679e-01 -7.24931121e-01 6.28010690e-01
-2.98172206e-01 -9.92166921e-02 -7.29597151e-01 -2.17504308e-01
3.30449753e-02 7.12619185e-01 -5.72583914e-01 5.13663352e-01
4.23492789e-01 3.65936249e-01 2.51607209e-01 4.84035373e-01
8.30184221e-01 2.42907271e-01 1.21567756e-01 -3.85272950e-01
-9.28115845e-02 5.33986509e-01 -1.84687889e+00 1.78806663e+00
-3.99752200e-01 7.06843853e-01 1.18102598e+00 -7.35062599e-01
6.77556634e-01 7.93870091e-01 3.65085185e-01 -2.04263325e-03
-1.04810290e-01 7.16782689e-01 1.59004360e-01 -1.04357071e-01
9.44281965e-02 -2.84189522e-01 3.43880564e-01 1.22952789e-01
4.70700234e-01 -4.61516887e-01 8.55776668e-02 2.49904811e-01
9.23001587e-01 -4.29701298e-01 2.78793365e-01 -6.18381917e-01
6.21055543e-01 -6.09995782e-01 5.87785542e-01 8.86161566e-01
-2.70811856e-01 7.62859941e-01 -2.41087288e-01 4.22093153e-01
-4.72838610e-01 -1.35154963e+00 -1.35906264e-01 1.03245091e+00
-2.02264547e-01 -3.50766808e-01 -8.63149107e-01 -1.54837659e-02
-3.30178663e-02 7.77142465e-01 -1.22581065e-01 -4.56024595e-02
-6.11721575e-01 -6.49666607e-01 6.80168927e-01 4.33860391e-01
6.09135069e-03 -2.18761876e-01 -2.78552055e-01 3.80084515e-01
-5.22970498e-01 -1.06996477e+00 -6.02747440e-01 6.35923505e-01
-5.38987756e-01 -5.82029760e-01 -7.40020573e-01 -5.61096787e-01
3.46363276e-01 7.05983639e-01 5.86808681e-01 -5.40523946e-01
1.14920080e-01 4.60692167e-01 -3.38202208e-01 -7.40014255e-01
-4.84839529e-01 -4.85325485e-01 5.54791093e-01 4.78366315e-01
-1.81916386e-01 -8.86756837e-01 -4.94743258e-01 2.66401440e-01
-5.53730249e-01 -3.79912704e-01 3.25476021e-01 6.60899878e-01
1.89368442e-01 5.51749051e-01 1.02308357e+00 -2.68664420e-01
8.07517767e-01 -6.53176308e-01 -4.73513722e-01 -1.44461513e-01
-6.50265753e-01 -2.60617703e-01 4.49916452e-01 -4.80014831e-01
-1.49893975e+00 -8.66871048e-03 -1.88249260e-01 -4.49539542e-01
9.67061520e-02 2.76253939e-01 -3.33788693e-01 -2.65540838e-01
7.22127199e-01 2.38991007e-01 -2.35252813e-01 -8.73749495e-01
3.87151003e-01 7.98153400e-01 8.25634122e-01 -1.89424574e-01
1.04769623e+00 4.68276352e-01 -1.64511412e-01 -1.15306842e+00
-6.33566082e-01 -8.60191345e-01 -2.49454185e-01 3.14172804e-02
2.63287872e-01 -1.19477105e+00 1.24709994e-01 7.03801930e-01
-1.03900778e+00 -2.02216685e-01 -2.90363431e-01 1.00878084e+00
-3.69560063e-01 3.12004447e-01 -3.48013461e-01 -1.28285861e+00
-3.35336924e-01 -8.71899486e-01 8.78101051e-01 -1.18593849e-01
-2.30605319e-01 -7.64844418e-01 2.79366940e-01 1.78255096e-01
9.65319037e-01 -3.04124862e-01 3.97676766e-01 -6.56754255e-01
-2.13236317e-01 -1.43578842e-01 3.59135628e-01 5.62169373e-01
4.01919484e-01 -4.64662671e-01 -1.42144907e+00 -3.02543998e-01
6.27071619e-01 1.75372630e-01 7.89114594e-01 8.00182641e-01
4.43363547e-01 -4.69866693e-01 -1.83072060e-01 8.05244982e-01
1.11832285e+00 1.69652089e-01 9.89362597e-02 -2.65641659e-01
5.00846922e-01 4.45209801e-01 8.15878883e-02 5.80952883e-01
-7.47233108e-02 5.38419545e-01 -5.06536700e-02 9.94830951e-02
-6.91146791e-01 -1.28473461e-01 4.33288544e-01 1.31085789e+00
4.26344782e-01 -2.09373474e-01 -5.50117433e-01 8.38209331e-01
-1.22033310e+00 -1.03808141e+00 -1.35122702e-01 2.23752499e+00
1.03614855e+00 -7.76677728e-02 -3.43472784e-04 9.57768261e-01
5.72811127e-01 4.97118205e-01 -3.99738312e-01 -7.20942998e-03
-3.74365538e-01 2.33031467e-01 5.33449829e-01 9.72276270e-01
-9.39204514e-01 3.28209311e-01 8.00770473e+00 9.64657307e-01
-9.10817802e-01 4.58538443e-01 -1.77599922e-01 -2.77816564e-01
-2.87699193e-01 -2.84737736e-01 -5.80976486e-01 3.59706759e-01
1.15711045e+00 -3.01469773e-01 9.02744710e-01 4.40443456e-01
2.85740763e-01 1.12374909e-01 -6.74351752e-01 1.19723177e+00
4.73852664e-01 -8.21823537e-01 -5.12514472e-01 -3.84835511e-01
7.94755101e-01 2.17604652e-01 1.55918658e-01 -2.62492508e-01
2.10008249e-01 -8.01863194e-01 1.13616657e+00 2.79942751e-01
6.46303952e-01 -4.57177073e-01 1.37044176e-01 2.38568097e-01
-1.30393386e+00 -2.42272481e-01 -4.92219552e-02 8.79398659e-02
5.02896190e-01 1.19219851e+00 -1.07124448e+00 4.04708475e-01
7.02426910e-01 3.71107966e-01 -2.61399150e-02 1.35507011e+00
-4.17991370e-01 1.24625778e+00 -5.94480336e-01 5.90381682e-01
-2.28613183e-01 9.39888582e-02 1.32128036e+00 1.36853445e+00
6.87011302e-01 -1.69357568e-01 -1.47294402e-01 3.13794583e-01
6.24256916e-02 2.66806670e-02 -3.46740246e-01 1.13393672e-01
6.92557156e-01 9.39392507e-01 -5.15666842e-01 -1.15237841e-02
-1.81386590e-01 6.74447715e-01 -3.61805707e-01 7.59811938e-01
-3.73378724e-01 -3.50481987e-01 8.21280837e-01 -1.75293550e-01
6.49178743e-01 -4.01644945e-01 -2.13260874e-01 -1.04311025e+00
-5.79197481e-02 -1.28099740e+00 2.20539615e-01 -5.37527561e-01
-1.29790878e+00 7.08467662e-01 -3.41458581e-02 -1.10868824e+00
-4.85127509e-01 -2.02323824e-01 -6.69885933e-01 9.63847876e-01
-1.63572049e+00 -7.02365637e-01 2.69985288e-01 8.26832414e-01
8.29404652e-01 -3.51041466e-01 8.07256579e-01 3.79824936e-01
-2.77078092e-01 4.60244358e-01 6.21366858e-01 -2.99231917e-01
8.40962231e-01 -1.37300861e+00 3.54095757e-01 1.42342639e+00
6.14812851e-01 8.25196028e-01 1.25184119e+00 -4.78064775e-01
-1.36079407e+00 -7.78537631e-01 7.41289675e-01 -3.13928008e-01
6.42178178e-01 -6.10022068e-01 -8.39549601e-01 3.43195945e-01
5.61623514e-01 1.28471935e-02 1.14840817e+00 -6.71398789e-02
-5.68167508e-01 -4.60035622e-01 -1.02991867e+00 -2.90059093e-02
7.11665213e-01 -6.37341857e-01 -8.24189901e-01 2.28994057e-01
6.34409964e-01 -3.91865879e-01 -2.90228188e-01 2.21936673e-01
4.40745085e-01 -9.27835584e-01 1.35398734e+00 -1.20636456e-01
-3.98907810e-01 -7.07286417e-01 -7.47762740e-01 -1.81156278e+00
-3.65809649e-01 -1.29018593e+00 -4.45508152e-01 1.39532948e+00
3.73459995e-01 -5.33627212e-01 2.40928918e-01 1.38357267e-01
-8.50289837e-02 2.27423474e-01 -1.23216128e+00 -1.03485036e+00
-4.31057721e-01 -8.99720967e-01 4.60735977e-01 6.01698816e-01
-2.22386606e-02 3.19339395e-01 -7.26925433e-01 8.33700061e-01
1.22828329e+00 4.43241931e-02 3.14693004e-01 -1.07856858e+00
-6.69785023e-01 -3.13521326e-01 -2.69500678e-03 -9.31935430e-01
1.83832645e-01 -5.59427321e-01 6.45051599e-01 -1.35765147e+00
-5.93235612e-01 -2.48335868e-01 -5.56434870e-01 -2.01441646e-01
-1.80699855e-01 2.93289810e-01 1.37043595e-01 1.47366554e-01
-2.71802038e-01 5.11508167e-01 3.19812715e-01 -5.24848253e-02
-7.43638650e-02 3.89987975e-01 -8.09410095e-01 9.46652830e-01
4.07674342e-01 -7.00137317e-01 -6.75879836e-01 -6.55574024e-01
1.14052080e-01 1.37095138e-01 7.81489089e-02 -1.48589170e+00
6.48446560e-01 1.83903113e-01 1.25226960e-01 -3.51090938e-01
5.87141156e-01 -9.81238365e-01 1.96422055e-01 9.69044119e-02
-5.06242096e-01 -4.88573700e-01 1.50599748e-01 9.17862475e-01
-5.07073164e-01 -2.83124715e-01 7.18548119e-01 2.52091199e-01
-1.79465085e-01 -3.64376277e-01 -5.03622591e-01 2.06001505e-01
3.05278897e-01 2.62156755e-01 1.22131795e-01 -7.64821708e-01
-4.50238883e-01 -3.01954865e-01 -2.32529014e-01 1.96022540e-01
6.34626389e-01 -1.32576251e+00 -9.87883627e-01 6.29926920e-01
-3.90684992e-01 -4.60559636e-01 3.03171754e-01 8.41099858e-01
1.23028960e-02 3.17495942e-01 3.93150121e-01 -2.07718208e-01
-1.02949905e+00 3.58418286e-01 3.90421033e-01 2.03886837e-01
-2.18307957e-01 1.38600683e+00 1.72484368e-01 -1.32840395e-01
5.33801973e-01 -1.54243773e-02 -1.32481322e-01 7.71008953e-02
1.09348249e+00 8.83816957e-01 3.11816245e-01 -9.75925028e-01
-6.52050078e-01 4.12173003e-01 6.84731364e-01 -7.43883669e-01
1.18597960e+00 -6.00065231e-01 -2.32038274e-01 6.64828062e-01
1.12108731e+00 1.16616929e+00 -1.14626670e+00 -4.20211971e-01
-7.81486258e-02 -8.98114085e-01 4.58163917e-01 -9.32940781e-01
-8.27897370e-01 7.33715355e-01 7.02906311e-01 5.56500912e-01
1.34487104e+00 -3.71247470e-01 6.47710800e-01 9.22813863e-02
2.88021535e-01 -8.25464964e-01 -9.96008441e-02 3.74562502e-01
1.02785945e+00 -8.18092406e-01 -1.72294721e-01 -4.52485144e-01
-2.09162772e-01 9.10566211e-01 -1.89000025e-01 -8.94208103e-02
1.09861827e+00 7.25479901e-01 6.77347124e-01 1.95487231e-01
-2.90404975e-01 -4.94182371e-02 6.64255679e-01 6.93872571e-01
2.29900360e-01 9.45872162e-03 7.00306073e-02 7.57572830e-01
-2.59108156e-01 -4.69964921e-01 2.32778400e-01 4.73416388e-01
-6.34370625e-01 -1.04735947e+00 -1.09756911e+00 2.21405298e-01
-7.66726673e-01 -5.25381923e-01 -1.64652810e-01 -4.74982172e-01
-1.15027130e-01 1.83981287e+00 -3.59113157e-01 -7.76986703e-02
3.04726422e-01 3.88106346e-01 1.86251059e-01 -3.69556814e-01
-3.29144478e-01 9.73613143e-01 1.05669156e-01 -3.19248557e-01
-7.58442342e-01 -8.59904408e-01 -8.81736994e-01 3.66830558e-01
-5.87078035e-01 5.44650018e-01 8.98206592e-01 7.65372813e-01
2.79476255e-01 5.93496680e-01 8.75541747e-01 -9.01281834e-01
-8.36254954e-01 -1.07405400e+00 -8.84426951e-01 5.74083999e-02
1.03244114e+00 -5.32075226e-01 -1.04834461e+00 1.87871560e-01] | [15.172135353088379, 5.729889392852783] |
c779b828-292f-4f65-a472-2fbb1668bb81 | dualposenet-category-level-6d-object-pose-and | 2103.06526 | null | https://arxiv.org/abs/2103.06526v3 | https://arxiv.org/pdf/2103.06526v3.pdf | DualPoseNet: Category-level 6D Object Pose and Size Estimation Using Dual Pose Network with Refined Learning of Pose Consistency | Category-level 6D object pose and size estimation is to predict full pose configurations of rotation, translation, and size for object instances observed in single, arbitrary views of cluttered scenes. In this paper, we propose a new method of Dual Pose Network with refined learning of pose consistency for this task, shortened as DualPoseNet. DualPoseNet stacks two parallel pose decoders on top of a shared pose encoder, where the implicit decoder predicts object poses with a working mechanism different from that of the explicit one; they thus impose complementary supervision on the training of pose encoder. We construct the encoder based on spherical convolutions, and design a module of Spherical Fusion wherein for a better embedding of pose-sensitive features from the appearance and shape observations. Given no testing CAD models, it is the novel introduction of the implicit decoder that enables the refined pose prediction during testing, by enforcing the predicted pose consistency between the two decoders using a self-adaptive loss term. Thorough experiments on benchmarks of both category- and instance-level object pose datasets confirm efficacy of our designs. DualPoseNet outperforms existing methods with a large margin in the regime of high precision. Our code is released publicly at https://github.com/Gorilla-Lab-SCUT/DualPoseNet. | ['Yuanqing Li', 'Kui Jia', 'Songcen Xu', 'Zhihao LI', 'Zewei Wei', 'Jiehong Lin'] | 2021-03-11 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Lin_DualPoseNet_Category-Level_6D_Object_Pose_and_Size_Estimation_Using_Dual_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Lin_DualPoseNet_Category-Level_6D_Object_Pose_and_Size_Estimation_Using_Dual_ICCV_2021_paper.pdf | iccv-2021-1 | ['6d-pose-estimation-using-rgbd'] | ['computer-vision'] | [-9.81022939e-02 2.96049803e-01 7.16815367e-02 -6.79585636e-01
-8.77837300e-01 -5.51191270e-01 4.23251152e-01 -2.34498858e-01
-7.00639561e-02 2.79383212e-01 -5.56799443e-03 2.13647261e-01
1.55419856e-01 -5.80090761e-01 -1.40796578e+00 -5.47812045e-01
1.33437291e-01 9.09927607e-01 4.73271906e-01 1.62110224e-01
1.22869276e-01 5.84710062e-01 -1.48894119e+00 3.55330318e-01
4.16569740e-01 1.42190361e+00 3.69228989e-01 5.74177206e-01
3.46444577e-01 4.42956656e-01 -1.76730052e-01 -6.81180537e-01
6.16800725e-01 2.12168619e-01 -5.14206290e-01 3.08476150e-01
9.69085693e-01 -6.67137265e-01 -2.28628233e-01 8.14000010e-01
5.53873837e-01 -2.50218958e-01 4.59629536e-01 -1.11491442e+00
-4.87265080e-01 3.18130940e-01 -6.05879903e-01 -2.67207801e-01
1.87325910e-01 3.01908880e-01 9.72416818e-01 -1.25439954e+00
6.71556473e-01 1.30025673e+00 8.99720669e-01 3.44432563e-01
-1.23250616e+00 -6.07610226e-01 4.41454321e-01 -5.99138299e-03
-1.48815286e+00 -4.37335968e-01 7.88771510e-01 -4.94132221e-01
8.41839612e-01 -3.92121915e-03 4.87532109e-01 1.20874512e+00
1.58209562e-01 6.90544248e-01 7.91305304e-01 -2.11759806e-02
6.89439923e-02 1.81975886e-01 -2.78332740e-01 8.92655730e-01
1.57901153e-01 -7.24630654e-02 -4.82174814e-01 -2.64466330e-02
1.05782402e+00 5.90485483e-02 -2.76543170e-01 -1.15732443e+00
-1.13270402e+00 5.74219763e-01 6.48397386e-01 -2.88337082e-01
-2.01806769e-01 3.55616003e-01 2.30695963e-01 1.76620651e-02
6.56914711e-01 2.97179520e-01 -9.63521957e-01 1.43007576e-01
-5.74202538e-01 4.03879613e-01 7.35564232e-01 1.37007284e+00
6.71748698e-01 -2.89444923e-01 -2.43707627e-01 6.28728986e-01
7.03237355e-01 6.54414356e-01 2.64154561e-02 -8.18985462e-01
6.95698082e-01 6.48092866e-01 1.91931296e-02 -7.88967907e-01
-3.09315622e-01 -9.24818754e-01 -5.28766632e-01 1.54249460e-01
2.67244130e-01 2.07153231e-01 -1.00621164e+00 1.72333193e+00
6.62191510e-01 3.40594620e-01 -4.52546924e-01 1.10618508e+00
6.21233404e-01 3.88901442e-01 -1.69546828e-01 3.16069454e-01
1.43688309e+00 -9.95053351e-01 -1.01109050e-01 -3.22984368e-01
3.59839141e-01 -8.16743255e-01 7.80174255e-01 3.41981798e-01
-1.18511462e+00 -6.51579261e-01 -1.19866586e+00 -3.65132630e-01
-9.70297754e-02 5.59985936e-01 4.49041545e-01 2.71857381e-01
-8.71376455e-01 6.82953596e-01 -9.31905985e-01 2.29919348e-02
5.43442488e-01 6.37479722e-01 -4.00677681e-01 1.03295706e-01
-6.95902646e-01 9.54545498e-01 2.40926772e-01 3.29552501e-01
-1.02229834e+00 -9.09474254e-01 -9.74230826e-01 -6.62132651e-02
5.72880328e-01 -1.03089070e+00 1.32600331e+00 -7.43439496e-01
-1.47024119e+00 1.02032661e+00 3.12176310e-02 -2.50182867e-01
8.88623416e-01 -5.75323641e-01 1.84429556e-01 -1.95037443e-02
1.39748037e-01 7.03386486e-01 1.15429497e+00 -1.49628067e+00
-2.22749501e-01 -7.62511611e-01 1.13337435e-01 3.57223094e-01
-4.12709676e-02 -3.80130470e-01 -1.02410209e+00 -3.57960343e-01
4.75194126e-01 -9.92520273e-01 -5.99884018e-02 5.84251881e-01
-4.50869888e-01 -1.85547456e-01 7.47261047e-01 -6.55365586e-01
5.87139070e-01 -2.16828156e+00 3.54411542e-01 4.15527262e-02
3.34029496e-01 2.84008570e-02 -1.99839845e-01 5.68605475e-02
-1.42556593e-01 -3.14908087e-01 -1.15288809e-01 -8.26103091e-01
6.73279017e-02 4.29450646e-02 -1.65220767e-01 6.91682100e-01
5.01963437e-01 1.03412378e+00 -5.26450217e-01 -3.26385856e-01
2.36115590e-01 5.83123386e-01 -9.99589205e-01 3.66011798e-01
-4.92316604e-01 4.33684736e-01 -6.87212586e-01 6.95792079e-01
1.03737772e+00 -5.68874657e-01 1.48541406e-01 -6.95347130e-01
2.07802225e-02 4.36483949e-01 -1.31873631e+00 2.03901362e+00
-3.48847210e-01 1.10169910e-01 1.70218930e-01 -7.40664899e-01
8.10218453e-01 1.21487841e-01 4.79412019e-01 -4.64956731e-01
2.39004642e-01 1.92115933e-01 -3.26081425e-01 -2.71341771e-01
1.36889488e-01 1.92502495e-02 6.50298223e-02 1.09285340e-01
4.05482829e-01 -3.36407810e-01 -2.89449960e-01 -1.11339539e-01
7.68757284e-01 7.23814249e-01 7.65345246e-02 -7.28585199e-02
4.99891013e-01 -4.67109323e-01 4.49658960e-01 3.19830984e-01
1.72434449e-02 9.56160665e-01 4.01659101e-01 -4.83632416e-01
-1.34131455e+00 -1.24818313e+00 -4.33096021e-01 9.36828136e-01
3.37944984e-01 -4.83392805e-01 -4.92879957e-01 -8.03415596e-01
4.46898043e-01 2.00045556e-01 -6.12831354e-01 -1.14082985e-01
-7.47243047e-01 -3.84417534e-01 1.52646542e-01 7.77968347e-01
3.68322372e-01 -6.48275018e-01 -3.66781503e-01 -1.41253101e-03
1.67993605e-01 -1.27800977e+00 -5.32426357e-01 4.15423512e-01
-9.13332343e-01 -1.03876340e+00 -4.87751931e-01 -7.75376976e-01
8.10359418e-01 -5.66166714e-02 1.07664871e+00 -2.14487717e-01
-1.28881916e-01 2.12521076e-01 -1.46938562e-01 -4.25253332e-01
6.47081155e-03 1.99651018e-01 -5.03778607e-02 1.09433671e-02
-3.86636890e-02 -7.04700530e-01 -7.40995944e-01 3.94146264e-01
-4.84789282e-01 4.28439140e-01 7.77516305e-01 6.21219337e-01
7.64034927e-01 -5.42771339e-01 3.09363436e-02 -5.67912519e-01
-1.85642064e-01 -3.01567137e-01 -8.04967821e-01 1.43747509e-01
-3.51474434e-01 3.07989120e-01 2.79741198e-01 -2.77529359e-01
-9.14830029e-01 5.41102886e-01 -2.61832714e-01 -8.81492138e-01
-1.02167595e-02 -1.46821439e-02 -4.47422266e-01 -9.55431536e-02
3.40121746e-01 1.16313316e-01 1.02945380e-01 -8.45013559e-01
2.83843279e-01 3.11756968e-01 3.92420828e-01 -6.94844007e-01
1.11140490e+00 3.62500995e-01 -6.47127032e-02 -2.31359899e-01
-9.84994948e-01 -2.45095998e-01 -8.01922262e-01 -1.58300579e-01
7.82896876e-01 -1.25815427e+00 -8.60753596e-01 5.60372949e-01
-1.35521400e+00 -1.95108756e-01 -2.32946783e-01 4.21985477e-01
-7.03482568e-01 2.40165330e-02 -5.28189182e-01 -3.65603805e-01
-1.28965214e-01 -1.28856802e+00 1.89912033e+00 -7.82130137e-02
1.64744124e-01 -5.51536679e-01 -1.10194452e-01 5.22587895e-01
8.55287686e-02 1.30060017e-01 5.26305616e-01 -5.91562569e-01
-1.11336195e+00 -1.10465042e-01 -3.35117906e-01 5.67808151e-01
-1.89243183e-01 -1.90763056e-01 -1.11458766e+00 -4.31415737e-01
2.18320236e-01 -5.48089147e-01 6.26426816e-01 4.61891949e-01
1.49663651e+00 -1.55334145e-01 -2.99284011e-01 1.10070658e+00
1.51742041e+00 -1.37446731e-01 4.15478349e-01 1.43411443e-01
9.56313968e-01 3.13524067e-01 5.07959068e-01 5.55126846e-01
5.41995108e-01 9.43703711e-01 8.74361694e-01 1.10572435e-01
-1.74984723e-01 -3.75886232e-01 3.23635876e-01 7.58164108e-01
-1.45865157e-01 -7.30757788e-02 -7.81245172e-01 9.11587477e-02
-1.67267847e+00 -5.45804799e-01 1.75171662e-02 2.22622943e+00
8.74636590e-01 2.63502628e-01 -1.04582794e-01 -3.54112178e-01
6.40312791e-01 1.58424228e-01 -8.31895232e-01 7.66272321e-02
3.00167441e-01 1.40278161e-01 6.08756542e-01 3.64124954e-01
-1.19881260e+00 7.34880269e-01 5.40669250e+00 6.35568142e-01
-1.11650789e+00 8.27307180e-02 6.42177045e-01 -2.57956624e-01
-2.17325389e-01 -4.32895422e-02 -1.14106321e+00 3.95379573e-01
4.63588744e-01 2.95066804e-01 8.85457024e-02 1.22542334e+00
-1.99692383e-01 1.53367728e-01 -1.60350883e+00 9.15886104e-01
1.76342204e-01 -1.18234634e+00 -6.37876540e-02 7.94596449e-02
5.02431870e-01 2.26665065e-01 1.40547857e-01 2.02273324e-01
-9.63803679e-02 -8.22899520e-01 1.27144122e+00 4.59880590e-01
9.18772995e-01 -4.02013034e-01 6.10174894e-01 3.37807298e-01
-1.29814243e+00 2.77508539e-03 -2.05129310e-01 1.81112498e-01
4.19301391e-02 4.41205770e-01 -9.21245396e-01 7.19809413e-01
6.46524847e-01 8.58557224e-01 -7.38173783e-01 8.72317493e-01
-6.07477278e-02 1.55525908e-01 -4.42115784e-01 3.60396266e-01
-1.34707764e-01 8.58766437e-02 5.15764594e-01 8.15726161e-01
3.06221783e-01 -2.22141162e-01 2.20025107e-01 1.08411348e+00
-1.58364400e-01 -2.37832054e-01 -2.88488686e-01 3.41010004e-01
5.03343463e-01 1.20267498e+00 -4.18247521e-01 -6.22090921e-02
-3.56272578e-01 1.00014246e+00 5.42471886e-01 -2.65289582e-02
-1.22471309e+00 2.84276843e-01 6.61668420e-01 4.04180825e-01
7.59932041e-01 -1.98599592e-01 -4.39900309e-01 -1.23219085e+00
5.22656798e-01 -6.92942083e-01 -3.00262123e-02 -8.98883820e-01
-1.24975204e+00 4.23622906e-01 -3.57280970e-02 -1.47585189e+00
-9.31006484e-03 -1.00392997e+00 -2.62016505e-01 9.38993216e-01
-1.36507785e+00 -1.47537184e+00 -3.90495002e-01 3.42451602e-01
6.37333572e-01 1.12968661e-01 5.06062567e-01 2.60720521e-01
-3.96101683e-01 6.09684408e-01 -3.37965876e-01 1.35493986e-02
8.30294490e-01 -1.19887209e+00 3.86722326e-01 3.92388612e-01
-8.37930664e-02 4.92173851e-01 6.15997910e-01 -5.68954766e-01
-1.67939329e+00 -1.15423286e+00 5.00438035e-01 -8.13383341e-01
4.25026298e-01 -8.55259955e-01 -6.21098399e-01 8.40319991e-01
-1.91073194e-01 3.22139651e-01 1.79424465e-01 8.63848701e-02
-5.50136507e-01 -2.57197738e-01 -8.62729609e-01 2.53582239e-01
1.30010903e+00 -5.25303721e-01 -5.10217488e-01 3.88476223e-01
8.92746985e-01 -9.17646587e-01 -1.00053728e+00 6.71098113e-01
7.22490907e-01 -9.03502941e-01 1.16922915e+00 -2.74779588e-01
6.98179603e-01 -4.84922498e-01 -1.47554249e-01 -1.00250077e+00
-4.11511809e-01 -1.05237067e-01 -2.83153921e-01 9.25337195e-01
3.21001142e-01 -3.89307410e-01 8.89611244e-01 2.91481137e-01
-5.11039138e-01 -1.25142181e+00 -8.85683239e-01 -6.64253175e-01
-1.43662810e-01 -3.75048399e-01 6.33158147e-01 4.26924169e-01
-7.32919693e-01 4.47060049e-01 -2.40056023e-01 4.74057674e-01
6.78022623e-01 1.99743137e-01 1.01619136e+00 -1.19550502e+00
-6.06780052e-01 -2.04499811e-01 -6.22569561e-01 -1.37048101e+00
7.20864236e-02 -8.50436628e-01 1.73702359e-01 -1.15628946e+00
4.16448146e-01 -2.71871120e-01 2.71331631e-02 4.81999665e-01
4.74848673e-02 2.21712798e-01 2.04779133e-01 1.52211115e-01
-7.13547230e-01 8.66027474e-01 1.60445130e+00 2.29908866e-04
1.84098303e-01 -1.96688473e-02 -5.06317616e-01 7.37012446e-01
4.38071698e-01 -3.31757277e-01 -4.33626175e-01 -6.48152053e-01
2.07910836e-01 5.70109189e-02 8.26659560e-01 -1.07824278e+00
1.27257884e-01 1.24175422e-01 6.65421307e-01 -8.16394269e-01
5.84442794e-01 -9.83881772e-01 1.93389818e-01 2.90744126e-01
-2.64054388e-01 -1.04253121e-01 2.08191782e-01 5.32099664e-01
5.27014509e-02 1.32011235e-01 8.01900208e-01 -1.80279613e-02
-5.01197159e-01 8.18376958e-01 5.84113717e-01 -1.41430080e-01
1.08986413e+00 -2.96705037e-01 -1.81003273e-01 2.92581338e-02
-8.42928052e-01 3.50942075e-01 6.31352842e-01 6.07399225e-01
5.48245609e-01 -1.38052034e+00 -6.77479982e-01 4.07155126e-01
2.63568044e-01 6.13119483e-01 1.72945365e-01 9.66901302e-01
-4.74786699e-01 3.06867778e-01 -9.70634520e-02 -9.61137891e-01
-1.12102473e+00 3.95585060e-01 5.63455462e-01 -2.11787641e-01
-6.03697896e-01 1.17983091e+00 6.89043105e-01 -6.33608818e-01
3.98437381e-01 -4.19287652e-01 2.33876660e-01 -1.91083089e-01
1.83270290e-01 2.87357066e-02 3.29001427e-01 -5.81449747e-01
-5.05021632e-01 6.79012477e-01 -1.96334749e-01 1.77729338e-01
1.66484606e+00 7.35794101e-03 -3.93932536e-02 4.07443881e-01
1.43348336e+00 -2.34461859e-01 -1.93089378e+00 -1.51011810e-01
-3.16482663e-01 -4.42637742e-01 -2.25427747e-01 -7.42241204e-01
-1.07406449e+00 7.24238575e-01 6.18160844e-01 -3.86645913e-01
8.70398521e-01 3.40978116e-01 4.54051465e-01 1.98689938e-01
5.03280282e-01 -9.46479261e-01 3.86569232e-01 6.00400627e-01
1.30111241e+00 -1.31576741e+00 -6.31137416e-02 -7.21754432e-01
-3.86366814e-01 9.62251782e-01 1.18479884e+00 -3.86593401e-01
6.86870635e-01 5.46482205e-01 -4.05346215e-01 -4.20530975e-01
-8.91057491e-01 1.14130281e-01 6.05570614e-01 2.96753764e-01
4.51043129e-01 2.23810263e-02 1.61658272e-01 6.72250926e-01
-1.99032068e-01 -2.01028854e-01 -2.03052923e-01 7.65570402e-01
-2.81435430e-01 -8.85800779e-01 -3.12709689e-01 4.66435701e-01
-7.87190944e-02 1.11563198e-01 -2.50134468e-01 6.67310655e-01
4.96021807e-01 -3.18383262e-03 2.40269080e-01 -4.65646774e-01
4.10347134e-01 8.88890922e-02 8.21049511e-01 -8.50578308e-01
-3.93340766e-01 1.29379362e-01 8.15659110e-03 -1.04067731e+00
-1.51127934e-01 -7.90556073e-01 -1.00579774e+00 -4.75976653e-02
-4.00400281e-01 -3.29382062e-01 6.34525120e-01 8.90438437e-01
4.43823487e-01 5.28845608e-01 6.61784410e-01 -1.35766864e+00
-1.10262144e+00 -9.22992051e-01 -5.72441161e-01 4.30825621e-01
3.46150815e-01 -9.51878488e-01 -1.84487060e-01 8.22301880e-02] | [7.641714096069336, -2.716809034347534] |
43c38974-9c63-4060-9e09-89a0258ac3ec | interaction-aware-motion-planning-for | 2212.11819 | null | https://arxiv.org/abs/2212.11819v2 | https://arxiv.org/pdf/2212.11819v2.pdf | Interaction-Aware Motion Planning for Autonomous Vehicles with Multi-Modal Obstacle Uncertainties in Multi-Vehicle Scenarios | This paper proposes an interaction-aware motion-planning method for an autonomous vehicle in uncertain multi-vehicle traffic environments. An interaction-aware motion-prediction model is used to predict the behaviors of surrounding vehicles. The multi-modal prediction uncertainties, containing both the maneuver and trajectory uncertainties of surrounding vehicles, are considered in the method for resilient motion planning of the ego vehicle. Based on the prediction of the surrounding vehicles, an optimal reference trajectory of the ego vehicle is computed by model predictive control (MPC) to follow the time-varying reference targets and avoid collisions with obstacles. A trade-off between the performance and robustness of the method can be achieved by tuning a safety-awareness parameter in the MPC. The efficiency of the method is illustrated in challenging highway-driving simulation scenarios and a driving scenario from a recorded traffic dataset. | ['Erik Frisk', 'Björn Olofsson', 'Jian Zhou'] | 2022-12-22 | null | null | null | null | ['motion-prediction', 'motion-planning'] | ['computer-vision', 'robots'] | [-1.02523334e-01 4.16329086e-01 -1.23330556e-01 -2.95914441e-01
-4.45510864e-01 -2.38698110e-01 7.69111991e-01 -1.04950167e-01
-1.95766881e-01 6.40218735e-01 -4.01932597e-02 -5.17746508e-01
-5.18041611e-01 -8.29736114e-01 -4.82640356e-01 -8.92452180e-01
-1.89748317e-01 2.92422146e-01 6.04095340e-01 -3.18132281e-01
1.88744113e-01 7.60341048e-01 -1.66081572e+00 -2.68110394e-01
9.34354365e-01 7.29934692e-01 6.44488037e-01 7.75641203e-01
2.73436993e-01 4.62062985e-01 1.35103747e-01 -4.28560227e-02
2.78754175e-01 1.61780462e-01 -1.34117857e-01 3.15825135e-01
-3.56185406e-01 -2.21797228e-01 -6.62096679e-01 8.99380684e-01
1.40863642e-01 7.67258763e-01 5.85202038e-01 -1.70182741e+00
6.00166440e-01 7.41158649e-02 -2.89706755e-02 5.97783849e-02
-3.51282388e-01 6.69819295e-01 2.81876951e-01 -5.02697408e-01
5.20076036e-01 1.28562856e+00 1.64573371e-01 4.73555148e-01
-9.68181729e-01 -2.56465554e-01 2.63647825e-01 8.75307500e-01
-1.58466518e+00 -4.68419164e-01 6.37642384e-01 -6.14277184e-01
7.71747649e-01 1.85106233e-01 5.26026368e-01 4.47941780e-01
1.10331285e+00 2.39508748e-01 3.93032074e-01 7.17567578e-02
4.54442680e-01 4.06612046e-02 -1.52507320e-01 2.12513342e-01
1.92186624e-01 7.09099650e-01 9.33795944e-02 2.70774998e-02
-1.10808074e-01 -2.31710464e-01 1.57140672e-01 -4.14139569e-01
-8.64902079e-01 7.30593383e-01 7.72970691e-02 -2.22464234e-01
-8.21590543e-01 2.01664940e-01 3.28295916e-01 5.50417900e-02
3.22155714e-01 -2.08684787e-01 -7.28146508e-02 -4.70820777e-02
-7.19761729e-01 7.94767797e-01 3.43180746e-01 1.18222630e+00
8.13840449e-01 5.28567851e-01 -3.49174321e-01 2.68425047e-01
4.77293909e-01 8.44681144e-01 -2.43697703e-01 -1.50423515e+00
4.98722494e-01 3.46197754e-01 6.35824084e-01 -1.05866742e+00
-4.27312136e-01 -1.97822988e-01 -4.27365124e-01 6.03056729e-01
1.70220435e-02 -4.51367766e-01 -6.81255102e-01 1.52799773e+00
7.55951405e-01 5.91741264e-01 3.19360107e-01 8.77188981e-01
-2.65645627e-02 9.42774236e-01 3.48403931e-01 -5.38642764e-01
8.22929263e-01 -6.21905684e-01 -1.03015411e+00 -3.51024866e-01
6.73978567e-01 -5.12074888e-01 2.77233869e-02 -6.08718991e-02
-9.84003365e-01 -6.02401078e-01 -1.09177613e+00 5.93693674e-01
-2.42179036e-01 -2.25721076e-01 -4.08506006e-01 2.17574403e-01
-7.01936483e-01 3.94148588e-01 -7.63463378e-01 -1.52224839e-01
-1.37966275e-01 4.25813705e-01 -9.30378139e-02 -7.82791302e-02
-1.34285343e+00 1.34259892e+00 4.78778780e-01 1.87786639e-01
-1.30700338e+00 -7.38133967e-01 -8.67538750e-01 -1.56468362e-01
5.52625656e-01 -5.99400997e-01 9.10278976e-01 -5.22331953e-01
-1.44714439e+00 -3.39812785e-02 -3.36803854e-01 -5.66712677e-01
7.78134882e-01 8.60920176e-02 -7.43713558e-01 1.15006633e-01
1.26497492e-01 3.97480607e-01 6.84411645e-01 -1.31130660e+00
-1.09128344e+00 -2.36322377e-02 -2.05315053e-01 5.90967357e-01
4.97389644e-01 -4.62697864e-01 -4.51445162e-01 -6.80470839e-02
-2.15637580e-01 -1.46065414e+00 -8.21292758e-01 -3.35111082e-01
-2.11947083e-01 -7.04488903e-02 1.19837177e+00 -3.83082390e-01
1.09548688e+00 -2.21781325e+00 1.48215264e-01 4.31780905e-01
-3.02958190e-01 2.25284874e-01 -2.52961010e-01 5.88495851e-01
4.40283716e-01 -3.95795256e-01 9.79683995e-02 -1.61817685e-01
-1.37151584e-01 5.92586815e-01 -3.91728878e-01 6.19563401e-01
8.09634849e-02 5.31234205e-01 -7.13082850e-01 -3.68101150e-01
7.52314508e-01 6.05820239e-01 -2.49956653e-01 3.46389323e-01
-3.85062933e-01 5.75850785e-01 -8.85680795e-01 9.27668884e-02
8.35262418e-01 7.38183081e-01 -3.87041247e-03 2.16182828e-01
-7.45446801e-01 -3.46187174e-01 -1.25616682e+00 6.61194444e-01
-4.96864647e-01 5.76016068e-01 5.27031958e-01 -6.72221065e-01
8.42231095e-01 1.98428780e-01 6.48109376e-01 -6.51920915e-01
3.53972882e-01 7.56367743e-02 8.38622171e-03 -6.70378327e-01
8.24909985e-01 1.39933363e-01 -5.11489883e-02 -8.49178284e-02
-7.21988678e-01 -3.07166308e-01 6.30547553e-02 -2.58826409e-02
9.04229045e-01 -1.53210580e-01 7.37603605e-02 -4.22717780e-01
9.72906649e-01 4.53648776e-01 8.45241308e-01 2.53106505e-01
-6.26453996e-01 -1.05978623e-01 6.23122528e-02 -3.85771841e-01
-1.18122458e+00 -8.11411738e-01 -1.13768600e-01 5.21911919e-01
7.47355521e-01 2.19141126e-01 -5.77469885e-01 -1.02566302e-01
1.27559513e-01 1.34416902e+00 -3.97629470e-01 -4.98049527e-01
-8.70923579e-01 -2.97990829e-01 -2.19635084e-01 1.33499905e-01
2.77477592e-01 -3.68829221e-01 -7.03717828e-01 6.38923526e-01
2.67957207e-02 -1.38110495e+00 -3.30254346e-01 -4.06907469e-01
-3.92196566e-01 -1.10144830e+00 8.98880586e-02 -3.17398489e-01
6.00093663e-01 6.04777217e-01 -6.17315236e-04 2.45417282e-02
2.64136016e-01 4.75528061e-01 1.01080805e-01 -4.83947098e-01
-1.05324376e+00 -4.01062161e-01 3.33103895e-01 5.02290368e-01
9.60837901e-02 -1.67008296e-01 -5.31291246e-01 7.48600304e-01
-5.74610889e-01 2.50254273e-01 1.71731010e-01 4.74881142e-01
8.53066027e-01 5.41816711e-01 7.64076173e-01 -4.10513431e-02
2.63128549e-01 -8.21366906e-01 -1.03458142e+00 -3.76622081e-02
-5.14522076e-01 -1.75009087e-01 6.46944165e-01 -5.14289856e-01
-1.19515312e+00 4.73368734e-01 9.27814767e-02 -4.88320589e-01
-2.31094688e-01 2.27675423e-01 -4.88376379e-01 -2.38414168e-01
-4.70069759e-02 2.31446236e-01 2.81215757e-01 1.92758307e-01
3.62709552e-01 4.53475893e-01 6.67294502e-01 -2.28562579e-01
1.02499938e+00 5.01633227e-01 6.90265536e-01 -9.47179973e-01
3.23804244e-02 -3.07368577e-01 -5.30647457e-01 -1.06397748e+00
8.23132634e-01 -1.02156007e+00 -7.92600691e-01 2.09479347e-01
-1.04802716e+00 -3.94123167e-01 -4.65223640e-02 7.37558246e-01
-1.12420952e+00 2.10828811e-01 7.10324869e-02 -1.28778505e+00
2.36076191e-01 -1.49287140e+00 5.33999443e-01 1.41902149e-01
1.84944216e-02 -9.53145325e-01 1.16292469e-01 1.84258237e-01
4.73960429e-01 5.01765609e-01 8.21178555e-01 -2.87362069e-01
-1.05212784e+00 -3.24080080e-01 1.86140671e-01 -1.25018954e-01
-4.01913881e-01 1.64264038e-01 -4.04311866e-01 -2.71290392e-01
-6.46922588e-02 6.10112786e-01 3.55690837e-01 5.02107441e-01
4.95842159e-01 -5.86783648e-01 -8.95611286e-01 1.80639103e-01
1.58980143e+00 6.73983276e-01 5.72585881e-01 5.29319048e-01
4.58342224e-01 1.10951483e+00 1.39034700e+00 4.18083251e-01
8.34378362e-01 9.82249618e-01 1.10866761e+00 6.50498569e-01
1.96278498e-01 -1.09785080e-01 3.00286621e-01 2.35706463e-01
1.51067391e-01 -2.48327553e-01 -9.08466101e-01 9.16899323e-01
-2.18760538e+00 -1.16696286e+00 -5.14628410e-01 2.20286727e+00
-7.64537007e-02 1.97885185e-01 -5.88856451e-02 -1.27000753e-02
1.00259781e+00 8.17832444e-03 -6.53629541e-01 -6.84350848e-01
2.05301166e-01 -1.03141320e+00 1.02817571e+00 9.07719731e-01
-9.35844123e-01 7.22023129e-01 5.88654470e+00 8.98038030e-01
-8.65559876e-01 -2.81767491e-02 3.94260049e-01 -7.97568783e-02
-2.80809194e-01 2.26256132e-01 -1.03682530e+00 5.84455907e-01
1.60681832e+00 -7.91086495e-01 2.78316736e-01 9.35472608e-01
1.32615161e+00 -2.91726619e-01 -5.20219088e-01 2.06052184e-01
-4.72667038e-01 -1.30868435e+00 -2.73523390e-01 1.87277317e-01
6.93733692e-01 1.27038032e-01 7.99925700e-02 1.81234747e-01
1.91750154e-01 -4.91841286e-01 9.42870796e-01 7.92590976e-01
1.76541880e-01 -1.33960462e+00 6.46363914e-01 7.80953765e-01
-1.30064428e+00 -5.83897352e-01 -5.83683997e-02 1.21415079e-01
1.08685279e+00 1.84528306e-01 -8.38950098e-01 5.14975488e-01
1.10685714e-01 4.03116524e-01 -7.10528716e-02 1.12355042e+00
1.50056317e-01 1.92627728e-01 -1.45244956e-01 -1.30701989e-01
5.43344319e-01 -3.10927868e-01 1.44670284e+00 9.07426119e-01
4.25159216e-01 2.79865026e-01 4.17874306e-01 5.41014194e-01
8.57738435e-01 -8.40179324e-02 -9.80769336e-01 4.61893886e-01
5.65969765e-01 1.06337941e+00 -1.24685049e-01 -1.37198761e-01
-1.22147366e-01 -1.10462360e-01 -1.82097062e-01 6.43490076e-01
-1.11391234e+00 -1.54657751e-01 1.25104392e+00 4.34371382e-01
3.79389673e-01 -6.92332089e-01 -2.43551776e-01 -2.34853268e-01
-2.26619080e-01 -1.03048319e-02 -1.32423535e-01 -4.99509335e-01
-5.06578088e-01 5.74694753e-01 4.18690681e-01 -1.64966297e+00
-5.51702797e-01 -1.89110786e-01 -9.59818661e-01 7.47830987e-01
-1.70986640e+00 -9.65108275e-01 4.02132683e-02 3.41387302e-01
6.13464355e-01 -3.27163249e-01 7.69559070e-02 -2.95034777e-02
-6.74519897e-01 -2.25748435e-01 3.01599950e-01 -7.66577065e-01
6.48841858e-02 -5.10740519e-01 1.17126554e-01 1.12517571e+00
-1.14004171e+00 -8.27957466e-02 1.41273737e+00 -8.89213681e-01
-1.73353004e+00 -1.88487220e+00 8.82752597e-01 -2.25147558e-03
7.01161742e-01 2.59470642e-01 -8.21036816e-01 3.76628339e-01
1.01703480e-01 -1.06910087e-01 -7.04278275e-02 -8.87464225e-01
4.07065421e-01 -1.57574728e-01 -1.17206728e+00 1.03270674e+00
4.61613387e-01 -2.54063718e-02 -1.27647236e-01 3.87348048e-02
7.73671806e-01 -3.25818241e-01 -5.76579690e-01 8.28236401e-01
3.51871848e-01 -2.49875501e-01 7.93864727e-01 -3.18310350e-01
-3.88051420e-01 -6.77798748e-01 -3.02343667e-01 -1.34129918e+00
-3.95705730e-01 -6.59987926e-01 -1.95124090e-01 7.50194967e-01
2.82057762e-01 -4.89040107e-01 4.50223535e-01 9.75323796e-01
-5.18005550e-01 -5.77640533e-01 -1.72769296e+00 -8.85411263e-01
1.79448351e-02 -8.70809376e-01 4.82892752e-01 1.86747387e-02
-1.20678134e-01 -1.45586640e-01 -4.89558518e-01 9.65616405e-01
7.98335254e-01 -2.59436905e-01 8.72661531e-01 -7.08197296e-01
4.85188723e-01 -3.27776551e-01 -5.64118743e-01 -5.98895907e-01
7.45384037e-01 -4.05380189e-01 5.32856226e-01 -1.38065493e+00
-2.60471761e-01 -2.81870335e-01 1.92684576e-01 -9.04972553e-02
-6.94427788e-02 -4.51709330e-01 2.61694163e-01 -8.50147083e-02
-4.60037172e-01 9.08326507e-01 1.03542042e+00 -2.03559116e-01
-4.34715778e-01 5.63561499e-01 1.72629252e-01 4.94052619e-01
9.46829557e-01 -5.25699496e-01 -7.53303170e-01 -6.58497438e-02
-3.45354944e-01 7.98038006e-01 3.37405145e-01 -1.08203256e+00
5.58698535e-01 -9.98570085e-01 -6.75021827e-01 -1.27000868e+00
6.17363453e-01 -1.29217792e+00 8.64842951e-01 8.10645223e-01
-2.52110720e-01 1.67652406e-02 4.42734510e-01 1.15349936e+00
-7.09909499e-02 -1.19172165e-03 1.10592782e+00 4.99819845e-01
-1.13125455e+00 5.33332050e-01 -1.20669866e+00 -4.89639848e-01
1.90396047e+00 -2.23941475e-01 -1.10227473e-01 -4.55858886e-01
-6.37325704e-01 9.61210489e-01 1.54064253e-01 7.03217268e-01
7.15992987e-01 -1.33497500e+00 -7.53618956e-01 1.31143183e-01
-4.51693200e-02 -3.55507046e-01 9.34714615e-01 8.20882797e-01
-1.98311985e-01 6.55007303e-01 -2.24024341e-01 -5.22895694e-01
-1.04414332e+00 8.52251053e-01 5.66174090e-01 1.59961730e-01
-4.11592782e-01 -2.76990414e-01 2.19862789e-01 -9.84108970e-02
-1.86418474e-01 1.85646206e-01 -4.07681376e-01 -3.29403967e-01
7.74585962e-01 9.93734956e-01 -1.39633849e-01 -1.49662876e+00
-3.41804117e-01 3.97034943e-01 4.43115056e-01 -2.81541348e-01
8.55519235e-01 -9.18249309e-01 4.13701296e-01 2.80658871e-01
8.80858898e-01 -2.73976415e-01 -1.79249990e+00 4.52534333e-02
-3.48994210e-02 -3.66487235e-01 3.78677607e-01 -2.90025175e-01
-9.31436121e-01 4.39132363e-01 5.36450982e-01 -1.19554773e-01
6.64948881e-01 -4.34899569e-01 7.34070182e-01 -3.47233862e-02
5.80719292e-01 -1.36063135e+00 -4.86010462e-01 7.66391754e-01
8.99409831e-01 -9.25744295e-01 -2.21638605e-01 -5.09534359e-01
-9.51170385e-01 9.55688655e-01 8.21346879e-01 -9.74648148e-02
9.00247574e-01 2.07504064e-01 7.10548973e-03 3.46017092e-01
-1.29975510e+00 -3.50570261e-01 2.19245911e-01 5.79374313e-01
-7.04032063e-01 4.02892977e-01 -5.63094378e-01 2.48209670e-01
3.91729981e-01 -2.84898192e-01 7.20256031e-01 6.46296680e-01
-9.03582931e-01 -5.74919939e-01 -4.76802319e-01 -1.71483219e-01
1.73176557e-01 6.51953340e-01 5.17425597e-01 8.01469386e-01
2.25409910e-01 1.57988489e+00 1.43443272e-01 -5.97756982e-01
5.16710699e-01 -1.49562344e-01 -4.50168014e-01 -1.67851627e-01
-5.23173809e-03 -6.45765886e-02 4.41113293e-01 -6.82852268e-01
-6.45716414e-02 -7.89851427e-01 -1.50131536e+00 -5.72663009e-01
-1.07976601e-01 1.52267262e-01 8.61108720e-01 1.17519403e+00
6.94197893e-01 3.29004765e-01 1.15123737e+00 -9.42847013e-01
-5.57259142e-01 -3.71066302e-01 -2.97748715e-01 -1.88508704e-01
4.12501872e-01 -7.83850491e-01 -3.98022026e-01 -3.54583174e-01] | [5.513972282409668, 1.5954967737197876] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.