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
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
9ab79aa4-944e-4a74-98d4-7983b5f89a1b
|
conditional-mutual-information-for
|
2305.14133
| null |
https://arxiv.org/abs/2305.14133v1
|
https://arxiv.org/pdf/2305.14133v1.pdf
|
Conditional Mutual Information for Disentangled Representations in Reinforcement Learning
|
Reinforcement Learning (RL) environments can produce training data with spurious correlations between features due to the amount of training data or its limited feature coverage. This can lead to RL agents encoding these misleading correlations in their latent representation, preventing the agent from generalising if the correlation changes within the environment or when deployed in the real world. Disentangled representations can improve robustness, but existing disentanglement techniques that minimise mutual information between features require independent features, thus they cannot disentangle correlated features. We propose an auxiliary task for RL algorithms that learns a disentangled representation of high-dimensional observations with correlated features by minimising the conditional mutual information between features in the representation. We demonstrate experimentally, using continuous control tasks, that our approach improves generalisation under correlation shifts, as well as improving the training performance of RL algorithms in the presence of correlated features.
|
['Stefano V. Albrecht', 'Josiah P. Hanna', 'Kevin Sebastian Luck', 'Trevor McInroe', 'Mhairi Dunion']
|
2023-05-23
| null | null | null | null |
['disentanglement', 'continuous-control']
|
['methodology', 'playing-games']
|
[ 3.64923686e-01 1.94859043e-01 9.71593056e-03 2.91031669e-03
-4.56674397e-01 -7.26837397e-01 8.45807910e-01 -3.22722867e-02
-6.46650791e-01 1.11372364e+00 2.65089154e-01 3.12769681e-01
-6.37101471e-01 -7.18081355e-01 -4.83974934e-01 -1.12874401e+00
-7.35852778e-01 5.20432055e-01 -2.73382932e-01 -2.04799145e-01
1.05588503e-01 4.11523342e-01 -1.81833756e+00 4.27243970e-02
4.92541462e-01 3.19689542e-01 1.31112784e-01 9.07220364e-01
6.69410884e-01 7.81742513e-01 -1.01613319e+00 2.14584976e-01
5.24021924e-01 -5.02246439e-01 -4.38903838e-01 -3.48444320e-02
-5.41325882e-02 -1.72214463e-01 -3.79134506e-01 6.24851644e-01
3.09594125e-01 4.47228819e-01 6.82098985e-01 -1.38362932e+00
-3.48731101e-01 4.08148646e-01 -2.50049531e-01 1.48809180e-01
3.46546531e-01 5.63890874e-01 1.04468691e+00 -2.17840701e-01
4.06411916e-01 1.22258103e+00 3.60132217e-01 7.41234422e-01
-1.87735868e+00 -7.86760032e-01 2.18505219e-01 -5.46104126e-02
-8.71893048e-01 -5.31667233e-01 5.98693132e-01 -3.88826728e-01
1.22420096e+00 4.97694582e-01 7.55224526e-01 1.42758358e+00
2.79073656e-01 5.97104490e-01 1.34038043e+00 -2.53788292e-01
3.77428532e-01 1.39566865e-02 -4.05940652e-01 5.76296151e-01
6.61571205e-01 1.19929075e+00 -8.70297253e-01 -2.48147815e-01
7.51868248e-01 -4.82785590e-02 -3.01492900e-01 -1.08284807e+00
-1.45249474e+00 1.01782238e+00 3.78696471e-01 2.49670073e-01
-3.95458400e-01 1.70800120e-01 3.83456200e-01 9.03741956e-01
2.28998497e-01 1.38289547e+00 -6.72293007e-01 -2.26724908e-01
-3.52790445e-01 5.51252961e-01 6.58744395e-01 6.01740837e-01
9.34828103e-01 2.10889190e-01 1.19086511e-01 3.41728002e-01
3.20694417e-01 4.70446557e-01 8.33943367e-01 -9.49334443e-01
5.70335805e-01 6.21192515e-01 4.07042950e-02 -9.28324342e-01
-5.00534415e-01 -4.69399124e-01 -5.44644773e-01 9.00084376e-01
2.57486373e-01 -5.03038883e-01 -4.55992579e-01 2.19209886e+00
-8.43429659e-03 -1.06278006e-02 5.37434578e-01 8.56325090e-01
2.62338109e-02 7.92217180e-02 -2.53312618e-01 -2.69794762e-01
6.38583302e-01 -4.74782556e-01 -6.86961532e-01 -7.36061990e-01
7.10606337e-01 -3.30222964e-01 7.75617182e-01 3.61709535e-01
-8.12181592e-01 -2.38605633e-01 -1.38366711e+00 4.51155901e-01
-3.15723509e-01 -4.24535275e-01 1.02303898e+00 4.93614554e-01
-5.47451973e-01 1.06335068e+00 -1.03611577e+00 -1.90992262e-02
4.18884009e-01 8.07670414e-01 -9.33583260e-01 1.31097794e-01
-1.20983171e+00 1.22508907e+00 4.93984312e-01 5.21349795e-02
-8.97597849e-01 -4.28043246e-01 -9.84386444e-01 7.55709186e-02
4.83802170e-01 -6.34494007e-01 9.78779793e-01 -8.90285850e-01
-1.64787698e+00 1.82078034e-01 5.27475834e-01 -1.92806870e-01
5.25738001e-01 -3.96159589e-01 -1.49518341e-01 -2.64713198e-01
-1.04864597e-01 2.88100928e-01 1.06843030e+00 -1.11315906e+00
-2.06598192e-01 -6.58168197e-01 -1.59887806e-01 6.13841593e-01
2.52221733e-01 -5.78033686e-01 8.18248153e-01 -3.74135494e-01
6.14852970e-03 -1.07354927e+00 -5.15590608e-01 -2.76953727e-01
-1.74895585e-01 2.72830397e-01 7.61396170e-01 3.57116908e-02
4.77063924e-01 -2.17968249e+00 6.29411280e-01 2.03995720e-01
3.60077113e-01 6.13268092e-03 -4.98984456e-01 6.00479484e-01
-3.92514974e-01 -6.00921176e-02 5.25256880e-02 -2.46844664e-01
2.06779078e-01 7.77851999e-01 -3.83551180e-01 7.73521602e-01
5.45711577e-01 5.64468563e-01 -1.28023446e+00 1.99407130e-01
1.63754329e-01 3.17217946e-01 -6.37452126e-01 4.63950664e-01
-3.16619761e-02 6.98639095e-01 -4.68552828e-01 -4.09491599e-01
2.07907900e-01 2.96031982e-01 3.59064460e-01 3.54534298e-01
1.61948442e-01 5.90196371e-01 -1.23432887e+00 1.59891391e+00
-6.12982869e-01 8.32587481e-01 -3.31803292e-01 -8.88963044e-01
9.81312871e-01 3.57616633e-01 4.49405104e-01 -5.54365039e-01
1.23451808e-02 -1.58821687e-01 5.90104997e-01 -5.53794146e-01
3.78271580e-01 -2.16576129e-01 -2.05893070e-01 7.48389661e-01
2.86880016e-01 -3.48981023e-01 -8.64557773e-02 5.78867234e-02
1.43512046e+00 3.36693883e-01 4.51466709e-01 -1.46942183e-01
-2.52632815e-02 -3.79381657e-01 4.66893673e-01 8.09140027e-01
2.34786025e-03 3.00229549e-01 8.06442261e-01 -4.74327266e-01
-8.80876958e-01 -1.23385119e+00 1.82739839e-01 8.36638689e-01
-3.02722394e-01 -5.08303702e-01 6.04265928e-03 -8.52093995e-01
5.11992238e-02 8.15943718e-01 -1.06054914e+00 -8.28167737e-01
-3.21349353e-01 -8.31426859e-01 4.65113133e-01 3.03607821e-01
-8.86491984e-02 -9.64093328e-01 -1.36133146e+00 3.21147889e-02
2.58699834e-01 -5.32377183e-01 1.00804798e-01 1.09481788e+00
-8.00788820e-01 -1.36697972e+00 2.89890189e-02 -1.01134658e-01
7.71904528e-01 4.05733958e-02 9.47809458e-01 -2.30386421e-01
-4.09406692e-01 2.62208283e-01 -1.23592719e-01 -1.66966856e-01
-3.82229149e-01 -2.92507529e-01 4.18132514e-01 -4.64915425e-01
1.51413351e-01 -7.46261597e-01 -2.53099233e-01 2.93089181e-01
-8.17487836e-01 -1.34232072e-02 6.78118706e-01 1.30228829e+00
-1.70255810e-01 7.88105093e-03 6.33194983e-01 -7.75931537e-01
9.88393307e-01 -4.68232363e-01 -4.75154161e-01 -6.81391954e-02
-5.39234400e-01 9.58444476e-01 4.87238765e-01 -8.38724911e-01
-7.28380799e-01 2.23782342e-02 7.19868124e-01 -2.10897341e-01
-3.00953805e-01 3.05531740e-01 -1.58721969e-01 2.12068185e-01
1.08366573e+00 2.46260921e-03 3.39372635e-01 -1.53279811e-01
5.64419687e-01 2.05772579e-01 -8.80272016e-02 -5.33214390e-01
8.34806919e-01 8.81936867e-03 2.73407221e-01 -4.07092243e-01
-4.75347996e-01 3.77554446e-02 -7.55825460e-01 1.73332676e-01
3.39741826e-01 -8.80485892e-01 -8.05094361e-01 -1.53213084e-01
-6.87991381e-01 -3.35620642e-01 -7.31321692e-01 8.69569182e-01
-1.14375973e+00 -7.51766860e-02 -1.19452681e-02 -8.04147959e-01
5.27541101e-01 -1.29592919e+00 6.95993125e-01 -1.59074008e-01
-7.01926589e-01 -8.52769554e-01 7.52240002e-01 -8.70901439e-03
2.63883561e-01 4.70129937e-01 9.52652395e-01 -9.44604397e-01
-2.81241894e-01 -1.56902120e-01 3.41154784e-01 -3.44264368e-03
4.43779439e-01 -2.18024626e-01 -1.01524854e+00 -5.38279772e-01
-1.83337256e-02 -6.92640662e-01 8.00682247e-01 -2.56410986e-01
2.63837755e-01 -7.37028599e-01 -1.10220522e-01 1.11431845e-01
9.42145944e-01 -3.34392749e-02 2.63915688e-01 3.88919920e-01
2.09830597e-01 8.57021272e-01 3.72342795e-01 5.08352280e-01
-2.82911956e-01 6.11597717e-01 5.70981205e-01 3.67094189e-01
3.86137038e-01 -3.24362725e-01 6.58019066e-01 4.71816212e-01
-3.76728512e-02 1.33943096e-01 -3.74093801e-01 1.87378690e-01
-1.96673954e+00 -1.11150587e+00 2.58563668e-01 2.35315681e+00
7.68455505e-01 3.67942173e-03 1.03548981e-01 3.63968283e-01
2.16269165e-01 2.30861723e-01 -8.36374521e-01 -5.94243586e-01
7.58991316e-02 2.05152795e-01 4.01161104e-01 4.63896096e-01
-7.59055495e-01 4.24970329e-01 6.53982067e+00 2.31602527e-02
-7.93819189e-01 -3.48401457e-01 -6.36547618e-03 -5.42273819e-01
-1.20509632e-01 -8.49646777e-02 -1.47197964e-02 1.35752812e-01
6.98395908e-01 -1.86025620e-01 7.84456670e-01 6.22858942e-01
-3.05166524e-02 -2.40199953e-01 -1.66312885e+00 7.97762215e-01
-4.82038558e-02 -5.82435489e-01 -2.45788574e-01 2.59688348e-01
6.65581524e-01 7.89473578e-02 5.29087365e-01 4.78500724e-01
1.01762247e+00 -1.34968579e+00 3.05729717e-01 3.40557396e-01
2.20036790e-01 -9.28625047e-01 5.13151050e-01 7.12504745e-01
-3.32954228e-01 -1.69279948e-01 -3.57510090e-01 -6.30158186e-01
-7.18658507e-01 3.14066522e-02 -1.08502221e+00 3.56807321e-01
6.00189380e-02 6.44967437e-01 -6.53795242e-01 7.45677531e-01
-4.74258989e-01 1.56136274e-01 -2.58206487e-01 -1.42940924e-01
1.12195246e-01 -2.06481472e-01 7.55003452e-01 4.50298190e-01
-1.16298944e-01 -2.42917001e-01 -9.12689492e-02 8.74217033e-01
5.65205038e-01 -4.12709564e-01 -1.12892687e+00 -1.12176321e-01
4.85929213e-02 8.75709116e-01 -3.69976372e-01 1.70970008e-01
-7.57449046e-02 9.28461313e-01 5.64896882e-01 5.45996904e-01
-3.63435209e-01 -1.65122580e-02 1.11925185e+00 -5.16430795e-01
1.14260443e-01 -4.02119935e-01 9.84223634e-02 -1.40522265e+00
-2.24913046e-01 -1.01127303e+00 2.09413305e-01 -4.63025391e-01
-1.24380863e+00 6.74600303e-01 9.60162207e-02 -1.14381266e+00
-8.37172210e-01 -2.82231897e-01 -2.51560926e-01 7.55062819e-01
-1.00910103e+00 -5.79165399e-01 2.58223206e-01 5.20560324e-01
2.97917068e-01 -4.65998858e-01 1.42842090e+00 -4.64400619e-01
-5.74748755e-01 3.17889184e-01 4.11791146e-01 -1.58041313e-01
6.50861323e-01 -1.40837026e+00 1.43619493e-01 3.46248537e-01
6.02262855e-01 8.06313634e-01 1.07712507e+00 -3.54226768e-01
-1.26493645e+00 -7.31371045e-01 3.00570607e-01 -5.73923528e-01
8.21568429e-01 -6.54830754e-01 -5.95179737e-01 8.03287923e-01
4.34823297e-02 1.51014224e-01 1.05989039e+00 5.18463433e-01
-7.39649117e-01 3.73096019e-01 -1.15157664e+00 8.73010576e-01
9.18458879e-01 -6.13361776e-01 -9.39081788e-01 1.40172690e-01
4.27077085e-01 -1.28544256e-01 -6.78586721e-01 1.39446437e-01
7.81668544e-01 -1.03579962e+00 7.33021379e-01 -1.23743153e+00
2.67497748e-01 -9.94985700e-02 -1.06473349e-01 -2.13215256e+00
-3.83760065e-01 -8.28813434e-01 -6.86978325e-02 6.65563643e-01
3.36622894e-01 -8.08611393e-01 7.06608236e-01 5.99105179e-01
5.58089554e-01 -6.03534043e-01 -1.03649342e+00 -8.62945318e-01
3.72256637e-02 -2.23791167e-01 5.43264747e-01 8.93662095e-01
5.30566573e-01 5.75804710e-01 -2.94900090e-01 1.55674353e-01
5.10454953e-01 6.18128153e-03 7.39519119e-01 -1.17549896e+00
-7.63134003e-01 -2.88379312e-01 -8.80216479e-01 -5.38781703e-01
4.57673758e-01 -6.04996681e-01 1.42528594e-01 -7.63075292e-01
-1.05974413e-02 -5.78247428e-01 -2.32207239e-01 5.77427685e-01
-8.17874521e-02 -1.64273545e-01 4.49275881e-01 -3.70161645e-02
-5.68929255e-01 9.80371416e-01 1.25262749e+00 -8.74403715e-02
-2.98668832e-01 1.20795406e-01 -5.42902589e-01 6.09253645e-01
9.62311208e-01 -8.97774816e-01 -7.79990196e-01 -2.37274885e-01
5.68028390e-01 3.41939442e-02 4.43004698e-01 -9.79228914e-01
-2.22201079e-01 -3.72538358e-01 7.51407266e-01 1.17325239e-01
5.67742646e-01 -1.26120079e+00 2.08902583e-01 6.67836130e-01
-8.79214823e-01 2.38612279e-01 2.44308293e-01 7.36156046e-01
7.92392716e-02 -2.54761279e-01 4.11528200e-01 -1.83058128e-01
-8.42104703e-02 -1.85670167e-01 -8.11563373e-01 4.05430654e-03
9.94611442e-01 -1.40532628e-01 -4.22062397e-01 -3.40375245e-01
-8.86552691e-01 2.19263434e-01 4.73329008e-01 5.52875102e-01
7.07149088e-01 -1.16078377e+00 -5.03036737e-01 8.10433865e-01
8.12864453e-02 5.70588000e-02 -6.77421466e-02 3.53539616e-01
1.43894702e-02 1.92722172e-01 -6.80526316e-01 -1.53496847e-01
-1.25012422e+00 6.33503139e-01 3.22550744e-01 -4.11027461e-01
-4.50691342e-01 5.62453866e-01 2.24330232e-01 -5.99863291e-01
1.14215026e-02 -3.41878831e-01 -2.42643416e-01 9.12955999e-02
3.89690697e-01 1.94562182e-01 -8.41681659e-02 -5.15440106e-01
-1.15849264e-01 1.69976220e-01 -2.42435217e-01 -4.19513613e-01
1.60388839e+00 1.29321560e-01 2.93430120e-01 7.53785968e-01
1.19509423e+00 -1.11262292e-01 -1.66397631e+00 -8.42141807e-02
-4.11561690e-02 -7.83537924e-01 3.22314687e-02 -8.31009984e-01
-7.68659472e-01 7.18192339e-01 5.35151660e-01 2.09743217e-01
7.47044146e-01 -2.82258302e-01 -1.74650401e-01 7.92565107e-01
3.94659132e-01 -6.54110372e-01 4.82276201e-01 3.91616285e-01
1.04016066e+00 -1.00100493e+00 4.62149680e-01 1.00800157e-01
-8.86767268e-01 1.20758486e+00 5.86558163e-01 -6.41840279e-01
2.39737615e-01 2.93945760e-01 -1.32544830e-01 -2.54353225e-01
-1.29367876e+00 -2.43013382e-01 6.72980100e-02 1.11112785e+00
1.84901822e-02 1.17427528e-01 1.88477576e-01 1.31799206e-01
-5.13222098e-01 -5.71134329e-01 7.09909320e-01 9.21168745e-01
-1.45343512e-01 -1.21366107e+00 -1.01393335e-01 3.56275827e-01
-8.27968493e-02 2.81364739e-01 -4.98929769e-01 9.34944570e-01
1.07336845e-02 9.83078182e-01 2.47419223e-01 -4.03697520e-01
2.74383932e-01 5.46270199e-02 8.77565503e-01 -7.87305772e-01
-6.16705298e-01 -2.26078227e-01 -2.16323603e-02 -7.84449995e-01
-3.68284434e-01 -9.24013734e-01 -9.78019118e-01 1.63170353e-01
-6.33475780e-01 2.94640511e-01 5.24930656e-01 9.64644313e-01
2.33195201e-01 7.57175267e-01 8.59588444e-01 -8.60808730e-01
-9.45464015e-01 -9.23753262e-01 -5.72228670e-01 3.61117393e-01
9.61814702e-01 -1.09902108e+00 -5.68732977e-01 -4.43729132e-01]
|
[4.142155170440674, 1.8440216779708862]
|
f81300dd-3793-4ea6-8c18-403a741fa4d0
|
towards-end-to-end-unsupervised-speech
|
2204.02492
| null |
https://arxiv.org/abs/2204.02492v2
|
https://arxiv.org/pdf/2204.02492v2.pdf
|
Towards End-to-end Unsupervised Speech Recognition
|
Unsupervised speech recognition has shown great potential to make Automatic Speech Recognition (ASR) systems accessible to every language. However, existing methods still heavily rely on hand-crafted pre-processing. Similar to the trend of making supervised speech recognition end-to-end, we introduce wav2vec-U 2.0 which does away with all audio-side pre-processing and improves accuracy through better architecture. In addition, we introduce an auxiliary self-supervised objective that ties model predictions back to the input. Experiments show that wav2vec-U 2.0 improves unsupervised recognition results across different languages while being conceptually simpler.
|
['Alexei Baevski', 'Michael Auli', 'Wei-Ning Hsu', 'Alexander H. Liu']
|
2022-04-05
| null | null | null | null |
['unsupervised-speech-recognition']
|
['speech']
|
[ 2.63692081e-01 2.36644045e-01 -1.82990536e-01 -7.98929155e-01
-9.23271120e-01 -2.89941221e-01 7.76194155e-01 -2.02773675e-01
-5.50970495e-01 3.85971278e-01 6.93157673e-01 -9.47564304e-01
4.57476109e-01 -3.21460456e-01 -3.14139694e-01 -3.33919138e-01
6.25923723e-02 3.98233026e-01 -1.78798974e-01 -3.78697038e-01
-2.04998642e-01 2.21633285e-01 -1.16455019e+00 3.29693168e-01
3.31549108e-01 8.39285195e-01 1.23117991e-01 9.69980359e-01
-2.96637714e-01 9.49993670e-01 -5.25661707e-01 -3.87453675e-01
1.36436343e-01 -3.43172342e-01 -8.20819914e-01 4.34051245e-01
1.27281949e-01 -2.01434642e-01 -7.98669040e-01 6.30391419e-01
5.58452189e-01 6.00724339e-01 4.34563220e-01 -6.99673474e-01
-9.07853305e-01 8.46260369e-01 -6.03850633e-02 2.50367701e-01
3.71311128e-01 1.31879717e-01 1.23531461e+00 -1.18939424e+00
3.06948513e-01 1.28599513e+00 4.69023973e-01 8.01274002e-01
-1.29670608e+00 -4.66653913e-01 2.29185462e-01 2.46546492e-01
-1.33819127e+00 -1.07403767e+00 8.00378799e-01 -2.29828149e-01
1.67529583e+00 4.07936752e-01 5.45601696e-02 1.42131889e+00
-3.85574222e-01 1.10191977e+00 7.81527758e-01 -6.32371485e-01
2.14214861e-01 1.08734131e-01 3.60836118e-01 2.93128222e-01
-4.70180362e-01 2.40849078e-01 -5.75003743e-01 2.86104977e-01
3.76104891e-01 -9.62658077e-02 -2.32329667e-01 -5.00775762e-02
-1.15128362e+00 7.93565512e-01 9.23094824e-02 3.57224911e-01
-2.15167537e-01 3.03031933e-02 4.33070481e-01 6.51153326e-01
4.96435136e-01 2.87648588e-01 -6.97494984e-01 -4.93689895e-01
-1.24807751e+00 -5.07242322e-01 5.97852707e-01 7.53496230e-01
6.50401056e-01 9.48500216e-01 -1.49177179e-01 1.42063320e+00
4.76173133e-01 4.90561694e-01 1.06218350e+00 -3.54401469e-01
4.09166276e-01 1.54885516e-01 -3.41344506e-01 -4.66917872e-01
-3.33612375e-02 -5.51795065e-01 -7.65112102e-01 4.46094871e-02
-1.55001700e-01 -1.35576785e-01 -1.31504726e+00 1.27268672e+00
-1.58758745e-01 5.72651885e-02 5.01456797e-01 7.92680085e-01
7.91952968e-01 1.03431356e+00 -4.51681428e-02 -2.17420403e-02
8.05206597e-01 -1.47579730e+00 -9.24878359e-01 -4.66837227e-01
6.16334260e-01 -7.58485198e-01 1.17694092e+00 3.91539365e-01
-9.73463416e-01 -6.56656682e-01 -9.04294312e-01 -1.87125847e-01
-4.62349713e-01 1.66160226e-01 3.12700868e-01 1.07353795e+00
-1.30496240e+00 1.94488108e-01 -8.66635442e-01 -3.21500689e-01
1.89313173e-01 2.83840716e-01 -6.16805911e-01 8.63507465e-02
-1.05976975e+00 1.02645612e+00 3.29621494e-01 2.85942554e-02
-8.90361249e-01 -3.51669490e-01 -1.03916824e+00 5.60239255e-02
4.08542544e-01 -1.85931504e-01 1.63809490e+00 -1.12439036e+00
-2.12493515e+00 4.88268942e-01 -5.76055825e-01 -8.64616334e-01
1.14357732e-01 -5.67308627e-02 -1.03681386e+00 -2.54063815e-01
-5.40857613e-01 5.93932450e-01 9.17698681e-01 -1.10091829e+00
-4.34310675e-01 -9.10393447e-02 -4.83113736e-01 1.75989226e-01
-7.15954781e-01 2.15959802e-01 -4.56614137e-01 -9.53723788e-01
-1.23145534e-02 -5.05432487e-01 -3.17537218e-01 -5.40739179e-01
-3.06867510e-01 -3.28506291e-01 8.85484099e-01 -8.78867626e-01
1.43125594e+00 -2.37290788e+00 -1.01842478e-01 3.61296356e-01
9.85444263e-02 6.59701109e-01 -4.02830184e-01 6.30550504e-01
-2.81952649e-01 5.77047095e-02 -2.00854674e-01 -7.42059171e-01
6.43808320e-02 6.06435359e-01 -5.64515710e-01 2.73939192e-01
3.83936077e-01 8.34796965e-01 -8.35823298e-01 2.95729656e-02
5.15284717e-01 6.74891293e-01 -5.50109863e-01 2.45759398e-01
-5.66977412e-02 2.42618918e-02 2.50296175e-01 5.03181398e-01
2.94834226e-01 1.31147563e-01 2.09707350e-01 4.04150724e-01
-1.61965013e-01 1.03927350e+00 -9.50704575e-01 1.39152253e+00
-7.45820642e-01 9.69546199e-01 1.48168370e-01 -1.24204433e+00
1.22200847e+00 5.97374558e-01 7.86828771e-02 -6.15384996e-01
1.21282235e-01 1.96559221e-01 6.95016384e-02 -1.07283272e-01
5.08861899e-01 -1.69517547e-01 2.34356195e-01 2.96577990e-01
5.17816842e-01 -1.36804610e-01 -1.76501185e-01 2.12120831e-01
1.14350557e+00 -4.51098055e-01 2.32790947e-01 6.84946477e-02
6.75088227e-01 -1.98305115e-01 2.76039809e-01 7.33454525e-01
-1.33808047e-01 8.72736037e-01 -1.91664472e-01 -3.03602934e-01
-9.70135570e-01 -1.11505210e+00 1.40097439e-01 1.32790971e+00
-5.94936967e-01 -5.42165756e-01 -4.14284408e-01 -7.50478923e-01
-3.87061667e-03 1.00878346e+00 -3.30044776e-01 1.86740365e-02
-1.74901530e-01 1.67137478e-02 7.66504884e-01 7.15748906e-01
-4.47843373e-02 -9.69949663e-01 3.23450476e-01 3.77714962e-01
1.14298977e-01 -1.16672516e+00 -6.82744145e-01 5.46840608e-01
-6.38273954e-01 -1.53216183e-01 -8.86319578e-01 -8.96295071e-01
4.30609941e-01 6.60684288e-01 9.12342668e-01 -1.29162073e-01
1.55653015e-01 5.48477650e-01 -7.06122994e-01 -3.48675400e-01
-6.14658415e-01 1.53726175e-01 4.87774163e-01 3.77120763e-01
5.95703959e-01 -4.04909611e-01 1.79459125e-01 3.09240848e-01
-6.77601576e-01 -1.05407432e-01 4.73644465e-01 1.06070435e+00
5.17268002e-01 -1.22334369e-01 6.43860102e-01 -5.67046404e-01
6.04647994e-01 -3.03994358e-01 -1.48613393e-01 2.25004867e-01
-6.72403872e-01 5.37912324e-02 7.47989058e-01 -4.97783333e-01
-8.44901502e-01 2.08025888e-01 -6.53566539e-01 -7.45567501e-01
-3.20536047e-01 7.87164032e-01 -2.86571831e-01 2.86454529e-01
5.26377738e-01 6.35801435e-01 2.16341391e-01 -7.07825899e-01
6.65783644e-01 1.45261705e+00 6.50390685e-01 4.97449301e-02
8.82279336e-01 2.09571850e-02 -1.04765773e+00 -1.76759994e+00
-5.18756032e-01 -8.43071580e-01 -4.22725469e-01 -1.01747923e-01
5.79789281e-01 -1.05412793e+00 -1.77594855e-01 2.95414120e-01
-1.02281606e+00 -5.44356227e-01 -3.70915323e-01 8.01124394e-01
-2.36498922e-01 5.54403484e-01 -3.39765280e-01 -1.21145952e+00
-3.98497283e-01 -9.58195984e-01 7.30493784e-01 -3.54083866e-01
-4.52654928e-01 -9.92818177e-01 1.04973003e-01 6.30979180e-01
8.05841684e-01 -1.03832757e+00 2.02000841e-01 -1.19282889e+00
-1.41069219e-01 -2.46770799e-01 4.21630256e-02 9.94945228e-01
3.39837670e-01 -1.08304046e-01 -1.45822883e+00 -3.19760174e-01
-3.07930470e-01 -2.73587614e-01 1.08653986e+00 -2.19169888e-03
1.15711832e+00 -5.06958723e-01 2.09880963e-01 5.14743388e-01
6.59141243e-01 3.33080679e-01 5.60862958e-01 8.29065964e-02
6.68981552e-01 3.90618205e-01 -1.32486671e-01 9.62928385e-02
1.96787104e-01 7.22269833e-01 -1.00920215e-01 -2.27864116e-01
-4.31560695e-01 -5.04888058e-01 8.76189351e-01 1.67251337e+00
3.34519804e-01 -4.77746069e-01 -1.15129411e+00 7.18593359e-01
-1.43068743e+00 -1.03937101e+00 1.98995337e-01 2.07715273e+00
8.38872910e-01 2.19555035e-01 1.89823717e-01 4.33901101e-01
4.32982594e-01 3.31417531e-01 -2.88233399e-01 -6.81220710e-01
-3.66378129e-01 5.45548916e-01 4.12145644e-01 1.05650187e+00
-9.84948218e-01 1.28350544e+00 6.92751598e+00 9.99795139e-01
-1.42306471e+00 2.78054535e-01 3.08017045e-01 -8.34123269e-02
-5.99659860e-01 -1.70944810e-01 -6.75270677e-01 1.90116033e-01
1.42082679e+00 -2.26529226e-01 5.59802532e-01 9.10165846e-01
3.43387455e-01 5.74552894e-01 -9.90078330e-01 1.12800562e+00
2.35598683e-01 -1.18500257e+00 2.28316605e-01 -5.10349758e-02
4.36894387e-01 6.04100883e-01 4.61743735e-02 6.35328233e-01
5.42434096e-01 -1.09073842e+00 6.96455121e-01 3.01266182e-02
8.69165361e-01 -5.86796403e-01 5.63772023e-01 2.18533784e-01
-8.69676888e-01 1.70093793e-02 -2.29794651e-01 -2.80264050e-01
2.88698882e-01 5.05479872e-01 -1.48045325e+00 2.67554879e-01
2.61338025e-01 6.44419849e-01 -3.62957329e-01 6.75448120e-01
-3.88566762e-01 1.37320352e+00 -3.51248026e-01 -1.30487770e-01
5.35639107e-01 1.00611784e-01 5.07634342e-01 1.79419565e+00
4.85134088e-02 -6.79712445e-02 1.41969666e-01 1.15605325e-01
-1.86980784e-01 3.24268103e-01 -8.08069885e-01 -5.42752981e-01
4.16046560e-01 6.29661262e-01 -2.48143807e-01 -5.51104903e-01
-6.52008593e-01 1.13192260e+00 3.84109437e-01 4.85507280e-01
-2.33228385e-01 -4.25586104e-01 9.73021269e-01 -3.88446008e-03
5.69815576e-01 -8.21365952e-01 -3.48939240e-01 -1.36757922e+00
-2.61120886e-01 -9.74157512e-01 5.53902946e-02 -6.98059082e-01
-1.24530602e+00 8.87909651e-01 -8.02810848e-01 -1.03699553e+00
-5.05481005e-01 -8.00771773e-01 -5.40108025e-01 8.41700971e-01
-1.53106737e+00 -1.00768495e+00 3.39266419e-01 4.76647615e-01
1.26620793e+00 -6.74348354e-01 1.18507040e+00 2.96549827e-01
-8.12580407e-01 1.15856791e+00 3.81214917e-01 5.50429583e-01
6.28074586e-01 -1.19662261e+00 1.00242913e+00 1.13848424e+00
8.57712984e-01 6.79383993e-01 4.74976063e-01 -4.00875866e-01
-1.48448896e+00 -1.08155525e+00 1.37942374e+00 -4.73733485e-01
9.69836950e-01 -5.28548539e-01 -9.41217601e-01 7.18134582e-01
4.51415360e-01 -1.34199560e-01 9.94614959e-01 5.07116616e-01
-8.29236805e-01 -1.32421628e-01 -4.06984329e-01 6.46837652e-01
8.19477379e-01 -1.22807908e+00 -9.25730824e-01 1.07505061e-01
9.33397710e-01 2.46932376e-02 -5.55631757e-01 5.11889644e-02
4.16786581e-01 -3.30852717e-01 9.09950733e-01 -9.86610115e-01
1.74114347e-01 -1.49910137e-01 -6.16225660e-01 -1.56134224e+00
-3.21061850e-01 -8.65847588e-01 -2.79237956e-01 1.17281163e+00
9.93138850e-01 -5.70604503e-01 6.03445709e-01 3.46298039e-01
-5.60023069e-01 -4.11426693e-01 -1.02282965e+00 -1.24677157e+00
-1.70922861e-01 -1.26313603e+00 3.22020710e-01 1.06782281e+00
2.06239805e-01 5.68490982e-01 -5.55419087e-01 1.77004024e-01
3.00459564e-01 -4.73512471e-01 8.75363648e-01 -8.28778565e-01
-4.18160200e-01 -7.01285779e-01 -6.95236862e-01 -1.53688979e+00
1.89359799e-01 -1.13376820e+00 2.85289973e-01 -1.36007893e+00
-4.98978257e-01 -2.98703555e-02 -6.35115385e-01 8.11957896e-01
7.61747882e-02 1.64953247e-01 1.16613835e-01 1.02455601e-01
-4.59108502e-01 7.87775457e-01 4.91381407e-01 -4.36422586e-01
-4.04995799e-01 -3.58801819e-02 -4.98436272e-01 3.99681151e-01
1.02164495e+00 -1.59711346e-01 -3.46347362e-01 -8.39335382e-01
-4.63386208e-01 -1.16819061e-01 -1.24478422e-01 -9.75878239e-01
3.99383068e-01 5.65218739e-02 1.19079025e-02 -5.67018390e-01
4.93342400e-01 -6.39790833e-01 -3.35093886e-01 -6.65562749e-02
-6.41626656e-01 -2.18153566e-01 1.42071515e-01 4.52945679e-01
-5.81420660e-01 -6.19510822e-02 5.81006765e-01 1.81639344e-01
-8.87984276e-01 4.84243296e-02 -1.02244294e+00 -1.51760191e-01
4.09828901e-01 -2.83272535e-01 8.17839429e-02 -8.53359640e-01
-8.84227991e-01 1.67632475e-01 1.83528513e-02 8.43865693e-01
7.47185886e-01 -1.15874958e+00 -7.83738196e-01 7.69047678e-01
2.46998549e-01 -4.68507379e-01 9.29041430e-02 4.85791057e-01
-2.39423271e-02 6.27866864e-01 3.17149669e-01 -3.08775306e-01
-1.28672194e+00 3.86686742e-01 7.10687488e-02 2.22087592e-01
-5.23607850e-01 1.07837212e+00 -2.72686929e-01 -8.13402295e-01
7.99813807e-01 -3.23590368e-01 -1.54169545e-01 6.80822833e-03
7.89130986e-01 1.49073556e-01 4.44007218e-01 -8.67361546e-01
-3.47031593e-01 -1.45674601e-01 -3.04250896e-01 -6.17081046e-01
1.38692379e+00 -2.04506651e-01 4.26918983e-01 6.99110389e-01
1.25028551e+00 2.44561121e-01 -7.87244558e-01 -4.10645008e-01
1.18600726e-01 -1.67590097e-01 6.27859235e-01 -9.24499512e-01
-1.07043588e+00 1.25777447e+00 2.73147106e-01 2.41736814e-01
6.91272378e-01 -1.64448723e-01 8.63008082e-01 8.16340148e-01
9.17103440e-02 -1.27153671e+00 -2.31603652e-01 1.16969025e+00
7.23708689e-01 -1.20063949e+00 -3.99785012e-01 -8.61837342e-02
-8.35289717e-01 9.46430743e-01 3.25845420e-01 3.01175565e-01
7.73932993e-01 3.59761715e-01 6.46028221e-01 3.59583050e-01
-9.22237098e-01 -4.92947012e-01 3.58229518e-01 7.68195331e-01
6.89039946e-01 3.70851278e-01 1.37949467e-01 7.75341511e-01
-4.38555509e-01 -8.81855637e-02 4.20857579e-01 8.12378049e-01
-4.73216146e-01 -1.35888112e+00 -2.71249473e-01 3.69483322e-01
-3.27470630e-01 -3.73274028e-01 -6.21195197e-01 1.75627738e-01
-5.35864651e-01 1.42664015e+00 1.20354004e-01 -9.39088583e-01
4.55348462e-01 5.16157627e-01 -1.84475005e-01 -8.44874024e-01
-5.48093200e-01 3.17301810e-01 3.76063347e-01 -3.47900867e-01
5.90766408e-02 -4.27124441e-01 -9.80738223e-01 -1.56130251e-02
-4.73138630e-01 2.98304349e-01 7.79995024e-01 8.50978971e-01
5.11390150e-01 4.68035847e-01 9.41437483e-01 -7.55016208e-01
-7.30799496e-01 -1.26453912e+00 -4.15783435e-01 1.01318784e-01
6.14826798e-01 -2.48865172e-01 -4.24782336e-01 2.15332493e-01]
|
[14.425724029541016, 6.66486120223999]
|
d28a5ce9-39d9-476b-9885-2f97d21cb2bc
|
meshmvs-multi-view-stereo-guided-mesh-1
|
2010.08682
| null |
https://arxiv.org/abs/2010.08682v3
|
https://arxiv.org/pdf/2010.08682v3.pdf
|
MeshMVS: Multi-View Stereo Guided Mesh Reconstruction
|
Deep learning based 3D shape generation methods generally utilize latent features extracted from color images to encode the semantics of objects and guide the shape generation process. These color image semantics only implicitly encode 3D information, potentially limiting the accuracy of the generated shapes. In this paper we propose a multi-view mesh generation method which incorporates geometry information explicitly by using the features from intermediate depth representations of multi-view stereo and regularizing the 3D shapes against these depth images. First, our system predicts a coarse 3D volume from the color images by probabilistically merging voxel occupancy grids from the prediction of individual views. Then the depth images from multi-view stereo along with the rendered depth images of the coarse shape are used as a contrastive input whose features guide the refinement of the coarse shape through a series of graph convolution networks. Notably, we achieve superior results than state-of-the-art multi-view shape generation methods with 34% decrease in Chamfer distance to ground truth and 14% increase in F1-score on ShapeNet dataset.Our source code is available at https://git.io/Jmalg
|
['Zuozhuo Dai', 'Ping Tan', 'Qingkun Su', 'Siyu Zhu', 'Zhiwen Fan', 'Rakesh Shrestha']
|
2020-10-17
|
meshmvs-multi-view-stereo-guided-mesh
|
https://openreview.net/forum?id=pULTvw9X313
|
https://openreview.net/pdf?id=pULTvw9X313
| null |
['3d-shape-generation']
|
['computer-vision']
|
[-8.61709341e-02 1.97460279e-01 3.55161637e-01 -5.19332469e-01
-7.56139815e-01 -6.60320163e-01 7.46874928e-01 4.00547124e-02
5.76567128e-02 3.61220956e-01 2.55491108e-01 1.96112007e-01
2.57599413e-01 -1.50473464e+00 -8.70643795e-01 -5.96620023e-01
3.30117792e-01 1.03934789e+00 2.85046577e-01 -1.64538115e-01
2.22749799e-01 8.81982625e-01 -1.64297581e+00 6.47950590e-01
5.26549816e-01 1.12847328e+00 1.57145426e-01 6.96734488e-01
-3.69208127e-01 5.22091761e-02 9.29050297e-02 -4.01104003e-01
4.70920473e-01 2.81022806e-02 -6.42592549e-01 1.87139302e-01
8.30345809e-01 -5.18962681e-01 -2.95914561e-01 8.22464526e-01
5.90168357e-01 5.52342683e-02 8.11941326e-01 -8.95584464e-01
-9.26252306e-01 2.98095018e-01 -6.31973982e-01 -3.08879435e-01
3.26691270e-01 1.74888447e-01 1.09723997e+00 -1.34033525e+00
9.02547419e-01 1.35229123e+00 4.44402397e-01 6.24427497e-01
-1.38112974e+00 -6.00635827e-01 1.05397984e-01 -3.88544023e-01
-1.37047851e+00 -2.34912708e-01 1.23263443e+00 -6.61497533e-01
8.35203707e-01 1.40075400e-01 1.10606420e+00 6.48925781e-01
2.50258952e-01 6.01469457e-01 9.26332712e-01 -1.20434668e-02
1.17281988e-01 -2.32383326e-01 -6.59316361e-01 1.12478757e+00
1.34934047e-02 3.23525310e-01 -4.75345969e-01 -3.46396595e-01
1.36751831e+00 1.64110839e-01 8.33171755e-02 -9.12301779e-01
-1.17580450e+00 7.84333110e-01 7.80212820e-01 -1.12019740e-01
-4.28752720e-01 3.25190455e-01 -1.19163565e-01 -3.54491234e-01
9.02241528e-01 2.45884582e-01 -4.37238216e-01 3.03434938e-01
-7.12147415e-01 4.67540652e-01 4.24458742e-01 9.62355733e-01
1.02195978e+00 1.88478827e-01 -2.41241250e-02 6.87862039e-01
6.91754341e-01 7.73293674e-01 -1.29671186e-01 -1.14517760e+00
4.35200840e-01 1.15362990e+00 -4.39065173e-02 -8.81471217e-01
-3.94486755e-01 -9.72906724e-02 -7.30627835e-01 4.36269045e-01
2.83333629e-01 1.68714181e-01 -1.30924237e+00 1.32224369e+00
6.95263505e-01 -4.57001813e-02 -4.64815944e-01 9.73927498e-01
1.12165165e+00 4.92998719e-01 -2.09015384e-01 4.47507501e-01
9.59401906e-01 -5.13129175e-01 1.05746828e-01 5.75964376e-02
2.05772430e-01 -7.67294288e-01 8.69513333e-01 1.80155799e-01
-1.46171093e+00 -5.69481373e-01 -7.81754553e-01 -2.30194166e-01
-3.16988051e-01 5.90235181e-02 5.57209611e-01 3.79398078e-01
-1.26806271e+00 6.14383638e-01 -8.51036131e-01 8.92652869e-02
8.14968944e-01 4.22452122e-01 -3.51193249e-01 -1.22881174e-01
-7.57988572e-01 4.47073340e-01 1.05868485e-02 -1.75029963e-01
-8.58783007e-01 -1.02533031e+00 -1.07702863e+00 -1.50740296e-01
4.98955660e-02 -1.13846874e+00 9.40613866e-01 -5.15264273e-01
-1.30771577e+00 1.28276908e+00 -5.00360578e-02 1.20303094e-01
4.95442748e-01 1.19703591e-01 3.31149369e-01 1.51917934e-01
1.01377271e-01 1.19831300e+00 7.41722763e-01 -1.78571939e+00
-4.53125089e-01 -7.12782741e-01 2.02698588e-01 3.65020573e-01
4.74388987e-01 -7.81693161e-01 -4.38277125e-01 -5.01038194e-01
6.43996358e-01 -8.34348202e-01 -2.29957610e-01 3.85065675e-01
-4.75697666e-01 -1.21484570e-01 6.10659182e-01 -4.00815248e-01
4.17575568e-01 -1.91710854e+00 2.18968302e-01 4.06672597e-01
4.49879020e-01 -4.39248294e-01 -1.74236059e-01 1.29926845e-01
1.20281994e-01 2.55407572e-01 -3.17700982e-01 -4.09122646e-01
-1.40461780e-03 3.66481990e-02 -2.13942692e-01 4.83659804e-01
1.33673996e-01 1.14437521e+00 -9.51928854e-01 -2.72160590e-01
5.45122027e-01 7.89579809e-01 -8.02622616e-01 3.28820080e-01
-5.93060493e-01 6.93581522e-01 -5.23893893e-01 4.96007383e-01
7.61446238e-01 -4.90203917e-01 -1.91371709e-01 -4.11749363e-01
7.94733614e-02 2.76366144e-01 -8.82480264e-01 2.15726495e+00
-4.70894516e-01 7.08326772e-02 -3.54851782e-02 -4.45225596e-01
1.03418744e+00 2.13627070e-01 8.79214466e-01 -5.17226756e-01
6.47768080e-02 -1.95710547e-02 -2.93567419e-01 6.46189004e-02
4.02552575e-01 -3.75669807e-01 1.03543870e-01 6.52513921e-01
-1.04035161e-01 -9.31426883e-01 -3.41304481e-01 2.35427186e-01
4.37204272e-01 5.39746284e-01 -9.05845761e-02 -2.69234795e-02
2.76983619e-01 -1.73026323e-01 2.39292994e-01 2.14577958e-01
4.35054243e-01 1.20039201e+00 2.77405262e-01 -9.02180851e-01
-1.50478733e+00 -1.64533794e+00 -6.74370304e-02 6.98681772e-01
1.59692243e-01 -2.20436230e-01 -5.01801491e-01 -5.90077639e-01
3.46832782e-01 6.80162370e-01 -8.53561461e-01 1.11836441e-01
-5.51429451e-01 -4.02879238e-01 1.00888751e-01 6.30525351e-01
2.64987260e-01 -9.12365794e-01 -6.02465391e-01 -2.09712144e-03
-3.61823407e-03 -1.04397762e+00 -6.16507411e-01 -2.90779889e-01
-1.20143044e+00 -1.12726784e+00 -6.66765213e-01 -5.08198321e-01
1.09818006e+00 4.58717942e-02 1.24569404e+00 3.82006794e-01
-4.22924697e-01 5.72522700e-01 -4.80799600e-02 -1.45018548e-02
-4.14218545e-01 -7.87518993e-02 -1.82037026e-01 7.65716285e-03
-6.95669279e-02 -9.02510643e-01 -8.22468519e-01 3.03403046e-02
-8.20077658e-01 8.09652805e-01 2.65219033e-01 6.57769620e-01
1.11590755e+00 -4.85747427e-01 5.22462241e-02 -9.52189028e-01
-1.41862221e-03 -2.72426188e-01 -8.06602657e-01 -1.30414009e-01
-4.74552512e-01 1.77400067e-01 3.82139862e-01 -4.38057147e-02
-1.05542803e+00 3.47878993e-01 -2.77949274e-01 -6.68952048e-01
-3.68964076e-01 -5.55201918e-02 -8.22486281e-02 5.75978011e-02
4.80781138e-01 4.00438070e-01 7.62846600e-03 -3.69090706e-01
6.61408961e-01 2.30133772e-01 5.09479828e-02 -8.33542943e-01
7.76486397e-01 9.70719099e-01 1.41462833e-01 -5.66975772e-01
-6.84157848e-01 3.77822295e-02 -1.02484763e+00 -3.34310353e-01
1.04057074e+00 -9.29436088e-01 -6.12042129e-01 5.12313843e-01
-1.28895855e+00 -5.09658754e-01 -1.51598722e-01 6.94913715e-02
-8.95157516e-01 6.44930080e-02 -5.55831552e-01 -5.23700714e-01
-4.98262942e-01 -1.26608789e+00 1.69538462e+00 2.52608508e-02
6.99764937e-02 -8.55930030e-01 -1.25455126e-01 5.09217381e-01
1.62446752e-01 5.62498033e-01 1.32897782e+00 -6.91968352e-02
-9.98159468e-01 8.77734274e-02 -2.87925035e-01 3.66879031e-02
2.43168250e-01 2.88395192e-02 -1.05948901e+00 -6.00697249e-02
-4.42649096e-01 -1.62391052e-01 5.20178854e-01 7.83520937e-01
1.40587080e+00 -1.94348861e-02 -2.95148700e-01 9.63513315e-01
1.53438079e+00 8.52789283e-02 4.36248451e-01 -2.61099100e-01
1.20290279e+00 6.14853024e-01 2.49806978e-03 6.34724319e-01
7.03462362e-01 6.50635064e-01 6.29001319e-01 2.27308795e-02
-3.41989309e-01 -6.61723852e-01 -2.37380609e-01 8.13006222e-01
-1.79386184e-01 3.52150835e-02 -1.10163212e+00 5.81574798e-01
-1.37294102e+00 -5.70011914e-01 -1.82017893e-01 2.18553495e+00
7.17269778e-01 4.56328019e-02 -3.24322730e-02 -4.41376418e-01
5.57265520e-01 2.19346449e-01 -7.87812114e-01 -1.06457703e-01
1.42629862e-01 2.44418383e-01 2.07630366e-01 7.24535763e-01
-8.66462767e-01 1.00779498e+00 5.38393211e+00 4.77153003e-01
-1.16346323e+00 -7.77370632e-02 1.05401778e+00 -2.86389738e-01
-9.91362870e-01 -1.53776407e-01 -5.05540252e-01 1.40269056e-01
3.40014905e-01 9.62514281e-02 4.96039599e-01 6.91404641e-01
9.89010110e-02 -1.04667611e-01 -1.20280349e+00 1.16197109e+00
-4.88939183e-03 -1.70902002e+00 6.00638926e-01 2.73371637e-01
1.10618365e+00 3.77930343e-01 4.27873172e-02 -2.85238087e-01
5.98117769e-01 -1.13352346e+00 9.15410399e-01 7.86232471e-01
1.17118824e+00 -7.40429819e-01 1.62951246e-01 2.61034071e-01
-1.39317715e+00 5.09780824e-01 -3.79650325e-01 7.94887319e-02
2.54108191e-01 5.86988151e-01 -7.75198936e-01 5.79979300e-01
6.46115601e-01 6.36221707e-01 -3.77275616e-01 7.01826096e-01
-7.08250776e-02 6.53632060e-02 -3.02889466e-01 2.61595964e-01
7.64675811e-02 -3.42698425e-01 4.65389162e-01 5.72060764e-01
3.99813771e-01 2.46840745e-01 1.43432260e-01 1.52683759e+00
-1.42149359e-01 -2.26529092e-02 -8.97006035e-01 1.89037815e-01
3.93196255e-01 1.28380525e+00 -9.57340539e-01 -3.70454401e-01
-3.37690294e-01 8.60112548e-01 3.53181303e-01 3.03073376e-01
-5.61730921e-01 2.23915890e-01 6.52332962e-01 5.84527910e-01
3.65332484e-01 -2.62656510e-01 -8.49020481e-01 -1.03429711e+00
-1.58221021e-01 -3.38586241e-01 -5.12214974e-02 -1.23903215e+00
-1.42973447e+00 6.28595471e-01 5.44700883e-02 -1.24515712e+00
-1.03256010e-01 -5.33838391e-01 -3.30775440e-01 1.09248936e+00
-1.10817349e+00 -1.42783380e+00 -3.57860118e-01 2.74792254e-01
5.21662056e-01 -6.18625432e-02 8.54839265e-01 -1.51468396e-01
3.25819165e-01 1.08534195e-01 -3.83659124e-01 2.92431056e-01
2.72772402e-01 -1.33925593e+00 9.27058399e-01 2.50164777e-01
2.89412200e-01 2.38659143e-01 1.11055069e-01 -9.59358275e-01
-1.44219804e+00 -1.13212180e+00 4.44017321e-01 -6.77458584e-01
1.68651104e-01 -3.80170882e-01 -6.93355381e-01 4.37032163e-01
-1.45869404e-01 2.86044061e-01 3.84938210e-01 -2.80307114e-01
-4.11626011e-01 1.20433554e-01 -1.23948097e+00 5.33608079e-01
1.30029702e+00 -5.27928591e-01 -1.98833197e-01 4.45194868e-03
5.86664796e-01 -1.01209807e+00 -1.01301837e+00 3.29341531e-01
6.73235178e-01 -1.06894815e+00 1.21910203e+00 -5.49587369e-01
8.46288145e-01 -2.61758208e-01 -4.10616845e-01 -1.36473584e+00
-3.95537168e-01 -4.28136467e-04 1.17323153e-01 6.62568450e-01
3.29064876e-01 -3.51036847e-01 1.09987962e+00 7.20340431e-01
-1.28877014e-01 -1.23425257e+00 -9.61357296e-01 -1.31638646e-01
3.34022969e-01 -5.15938640e-01 9.09936726e-01 6.63987160e-01
-5.66660762e-01 8.60717744e-02 2.17268988e-01 1.27214655e-01
8.64766896e-01 6.80179954e-01 6.89410865e-01 -1.47196412e+00
-5.41877784e-02 -4.58109498e-01 -3.51849973e-01 -1.11949110e+00
4.85742688e-02 -1.36439145e+00 -1.36428013e-01 -1.89627385e+00
2.27192417e-01 -5.81053317e-01 1.73793465e-01 4.05343533e-01
1.83274880e-01 4.77835625e-01 2.10612729e-01 -1.21910430e-01
-8.35793018e-02 8.62967849e-01 1.95619822e+00 -1.59526527e-01
-9.42821950e-02 -2.50704050e-01 -5.26474476e-01 8.80926192e-01
6.11422360e-01 -3.30698192e-01 -4.19569612e-01 -8.49441707e-01
3.28451186e-01 3.12105119e-01 6.16398215e-01 -5.88985205e-01
-2.60337181e-02 -3.39517683e-01 9.44703758e-01 -9.70714748e-01
7.42699862e-01 -7.11655080e-01 3.33095759e-01 2.19506487e-01
-1.43610746e-01 2.51103431e-01 1.93701342e-01 5.30890703e-01
3.22072774e-01 3.29548389e-01 8.41923416e-01 -6.49448872e-01
-3.22562486e-01 9.60244775e-01 7.57688880e-02 1.16522469e-01
8.06068718e-01 -4.82304722e-01 -3.83220874e-02 -3.66097569e-01
-1.01873136e+00 -1.82437729e-02 9.48812485e-01 5.07935524e-01
1.19675195e+00 -1.67421567e+00 -7.79367626e-01 6.39031470e-01
-2.81849969e-02 7.08260477e-01 3.71960521e-01 4.81803030e-01
-6.24634445e-01 1.98130310e-01 -2.53068328e-01 -8.45568001e-01
-9.23922539e-01 2.90978495e-02 6.17061496e-01 1.01238862e-01
-9.46871877e-01 9.29597020e-01 7.49655128e-01 -8.84614885e-01
-2.02739462e-01 -5.28313041e-01 1.40745178e-01 -2.89043695e-01
1.17578484e-01 1.26395583e-01 4.12056185e-02 -8.34554970e-01
-2.38279223e-01 9.44783330e-01 1.82612538e-01 -3.49782228e-01
1.77592838e+00 1.18622281e-01 -1.87040597e-01 4.37376380e-01
1.15519655e+00 6.00355566e-02 -1.67623425e+00 -1.21769927e-01
-5.12374580e-01 -6.48053586e-01 1.46659106e-01 -7.37609267e-01
-1.56893575e+00 9.89396274e-01 4.85976875e-01 -1.46156609e-01
7.06426561e-01 2.61443973e-01 8.68310452e-01 -1.86215624e-01
6.87786400e-01 -7.14978218e-01 1.97700933e-01 4.70915675e-01
1.18711758e+00 -1.07222712e+00 -1.24492329e-02 -5.17269850e-01
-4.43604201e-01 1.18414116e+00 8.79670024e-01 -5.67589760e-01
9.60654199e-01 2.96473473e-01 -1.19953260e-01 -7.49257982e-01
-6.38872683e-01 5.54954633e-02 6.73411489e-01 5.81653833e-01
4.53680724e-01 2.64561981e-01 2.48263672e-01 3.34994227e-01
-3.81618530e-01 -3.15006912e-01 3.15127969e-01 4.16222572e-01
-2.70562738e-01 -9.68867898e-01 -3.15484017e-01 7.75787830e-01
-4.00638394e-02 -8.66643786e-02 -4.49732363e-01 3.71222466e-01
2.19990045e-01 2.37824678e-01 3.96968126e-01 -5.67713380e-01
3.15696627e-01 8.21052343e-02 8.51700544e-01 -1.08463240e+00
-2.93697894e-01 1.28649369e-01 -2.97102153e-01 -7.46405244e-01
-2.56030977e-01 -6.02314591e-01 -1.58150780e+00 -3.67883354e-01
1.71760276e-01 -3.60829771e-01 4.41340834e-01 7.71766067e-01
5.73953032e-01 3.23757619e-01 5.64211488e-01 -1.47200489e+00
6.80035446e-03 -5.39328456e-01 -5.07911026e-01 5.52233875e-01
1.59389138e-01 -8.43302906e-01 -9.96091068e-02 1.52147070e-01]
|
[8.811355590820312, -3.5410094261169434]
|
6f691f2e-11ca-4dd9-9336-5088edfb5aca
|
a-strong-transfer-baseline-for-rgb-d-fusion
|
2210.00843
| null |
https://arxiv.org/abs/2210.00843v2
|
https://arxiv.org/pdf/2210.00843v2.pdf
|
Early or Late Fusion Matters: Efficient RGB-D Fusion in Vision Transformers for 3D Object Recognition
|
The Vision Transformer (ViT) architecture has established its place in computer vision literature, however, training ViTs for RGB-D object recognition remains an understudied topic, viewed in recent literature only through the lens of multi-task pretraining in multiple vision modalities. Such approaches are often computationally intensive, relying on the scale of multiple pretraining datasets to align RGB with 3D information. In this work, we propose a simple yet strong recipe for transferring pretrained ViTs in RGB-D domains for 3D object recognition, focusing on fusing RGB and depth representations encoded jointly by the ViT. Compared to previous works in multimodal Transformers, the key challenge here is to use the attested flexibility of ViTs to capture cross-modal interactions at the downstream and not the pretraining stage. We explore which depth representation is better in terms of resulting accuracy and compare early and late fusion techniques for aligning the RGB and depth modalities within the ViT architecture. Experimental results in the Washington RGB-D Objects dataset (ROD) demonstrate that in such RGB -> RGB-D scenarios, late fusion techniques work better than most popularly employed early fusion. With our transfer baseline, fusion ViTs score up to 95.4% top-1 accuracy in ROD, achieving new state-of-the-art results in this benchmark. We further show the benefits of using our multimodal fusion baseline over unimodal feature extractors in a synthetic-to-real visual adaptation as well as in an open-ended lifelong learning scenario in the ROD benchmark, where our model outperforms previous works by a margin of >8%. Finally, we integrate our method with a robot framework and demonstrate how it can serve as a perception utility in an interactive robot learning scenario, both in simulation and with a real robot.
|
['Hamidreza Kasaei', 'Georgios Tziafas']
|
2022-10-03
| null | null | null | null |
['3d-object-recognition']
|
['computer-vision']
|
[ 3.54751885e-01 2.82906350e-02 1.85294390e-01 -2.37447307e-01
-1.21586430e+00 -6.61591351e-01 7.32995689e-01 -2.29747489e-01
-5.63119173e-01 2.37861350e-01 -3.83784180e-04 -2.43131846e-01
-6.99873418e-02 -2.75970280e-01 -1.07663441e+00 -9.00380790e-01
1.96517736e-01 5.92081130e-01 1.37234524e-01 -3.30071270e-01
-1.16701834e-01 5.75036883e-01 -1.83622539e+00 3.66770297e-01
3.60669732e-01 1.41227579e+00 2.72772640e-01 7.33399034e-01
2.01957852e-01 5.48441172e-01 -3.70455444e-01 -3.30459267e-01
7.26451933e-01 -7.81535804e-02 -7.21745551e-01 3.63861382e-01
8.30674708e-01 -2.81902015e-01 -1.56021044e-01 6.44786000e-01
8.24426532e-01 -1.13268584e-01 7.85673618e-01 -1.33632445e+00
-5.02114534e-01 7.84979761e-02 -7.47027397e-01 -1.56033427e-01
4.98688519e-01 4.51214433e-01 9.17509556e-01 -1.02070379e+00
4.17885602e-01 1.32158697e+00 7.10689366e-01 4.58313763e-01
-1.48385477e+00 -3.36227000e-01 -1.48563669e-03 2.16652706e-01
-1.08215404e+00 -4.12239820e-01 6.96555912e-01 -3.66905689e-01
1.47439516e+00 -1.35345489e-01 5.87965190e-01 1.18849874e+00
3.08641195e-01 9.84377742e-01 1.33937633e+00 -4.78181720e-01
-3.15255448e-02 -8.78990069e-02 -1.64028689e-01 6.29130900e-01
-8.34681764e-02 3.43457669e-01 -7.21416235e-01 3.63811016e-01
7.71966279e-01 -2.10616902e-01 -2.20682293e-01 -7.40370870e-01
-1.29940295e+00 6.62646532e-01 8.41873050e-01 1.25597278e-02
-5.07727563e-01 4.04274613e-01 2.60313749e-01 4.17062491e-01
2.02187583e-01 2.40775228e-01 -5.24765015e-01 -4.93591651e-02
-5.32894194e-01 6.08488359e-02 5.66543818e-01 8.60360146e-01
9.57033098e-01 -1.12026311e-01 -3.08478847e-02 8.68906796e-01
6.87165439e-01 7.85606980e-01 1.87609836e-01 -1.09252179e+00
4.74675387e-01 5.02882361e-01 -1.57533213e-01 -3.74541163e-01
-5.43351531e-01 -2.08735093e-01 -6.46442950e-01 8.30412447e-01
5.96837819e-01 2.26173941e-02 -1.42175055e+00 1.66640735e+00
3.48919362e-01 -7.07738250e-02 3.73396337e-01 1.16195214e+00
8.42103243e-01 4.26674753e-01 -3.66936298e-03 1.35404512e-01
1.42033148e+00 -7.77465641e-01 -1.76014885e-01 -5.05453289e-01
4.01189238e-01 -8.32184196e-01 1.02534962e+00 5.91513872e-01
-1.13636971e+00 -7.62237370e-01 -1.13042665e+00 -3.49529237e-01
-4.12694812e-01 1.02696538e-01 6.29060090e-01 6.23678207e-01
-1.37960875e+00 3.05501431e-01 -9.24963653e-01 -6.81097507e-01
3.66586894e-01 6.18660867e-01 -7.04716086e-01 -4.01134193e-01
-7.71697760e-01 1.32478118e+00 3.77588123e-01 1.65798426e-01
-1.13464797e+00 -5.44425309e-01 -8.71167898e-01 -3.19185466e-01
1.72765553e-01 -1.06260562e+00 1.33194578e+00 -8.55280399e-01
-1.57804441e+00 9.96364594e-01 1.53403744e-01 -3.80671173e-01
3.94991040e-01 -2.63541758e-01 1.60985410e-01 2.92604893e-01
-1.60048425e-01 1.15805387e+00 9.15954173e-01 -1.61910641e+00
-6.71342313e-01 -5.65216482e-01 2.26952046e-01 5.24888754e-01
5.89393824e-03 -5.26828289e-01 -6.82519317e-01 -2.17194766e-01
2.55071700e-01 -1.05806160e+00 -1.27946302e-01 1.17845990e-01
-3.50791961e-02 -1.17572807e-01 7.74385154e-01 -2.98478544e-01
6.82074800e-02 -2.11206627e+00 5.87066412e-01 1.05685607e-01
1.05801467e-02 -3.96128856e-02 -2.61179835e-01 3.14708561e-01
-1.15294881e-01 -4.07889545e-01 -2.89745033e-01 -8.52083087e-01
3.21617395e-01 5.94424129e-01 -2.64604241e-01 5.99881411e-01
4.58399445e-01 1.03813100e+00 -6.01802468e-01 -2.78610259e-01
5.12210667e-01 7.35329092e-01 -4.32599247e-01 2.94550776e-01
-1.04592420e-01 5.23241162e-01 4.13998663e-02 1.02293348e+00
5.13306499e-01 -1.29269153e-01 -1.21547751e-01 -5.41817605e-01
1.00665167e-01 -5.13192602e-02 -9.67381895e-01 2.29394078e+00
-4.43385601e-01 5.68350971e-01 8.54168534e-02 -1.02830803e+00
7.84499645e-01 1.54727504e-01 4.90722060e-01 -1.06111228e+00
1.86864555e-01 3.03781182e-01 -1.95187539e-01 -4.97630775e-01
4.28398967e-01 -4.02685344e-01 -2.11970106e-01 2.84160137e-01
5.06241918e-01 -8.05400789e-01 -1.77157491e-01 -5.07543981e-02
1.16703212e+00 5.65141618e-01 -8.14104974e-02 2.96293885e-01
5.03441691e-02 7.67259821e-02 -2.09248848e-02 6.38335943e-01
-2.77803779e-01 9.12788987e-01 1.57842040e-01 1.29970312e-01
-1.01033938e+00 -1.29083228e+00 -1.45542458e-01 1.00492156e+00
3.01846504e-01 8.20459276e-02 -2.22454011e-01 -6.23905718e-01
2.99733669e-01 5.80581903e-01 -6.38373673e-01 -2.89312780e-01
-3.15767705e-01 -6.67510450e-01 8.07503462e-01 6.42135561e-01
5.17057300e-01 -8.17314625e-01 -9.77590740e-01 3.05533949e-02
3.76000963e-02 -1.47262704e+00 1.48582399e-01 8.95508468e-01
-7.80543923e-01 -9.56074417e-01 -9.71077859e-01 -6.08720720e-01
2.64887899e-01 4.36262846e-01 9.34604824e-01 -5.29502094e-01
-1.83253378e-01 1.21240699e+00 -3.82233679e-01 -3.47523749e-01
-2.86749363e-01 -1.30015165e-01 -7.14298263e-02 -2.34446615e-01
1.80773601e-01 -5.15885890e-01 -5.72097719e-01 2.95974612e-01
-1.04577696e+00 1.59737796e-01 1.16787851e+00 7.45175838e-01
3.61116886e-01 -6.50082111e-01 3.78776193e-01 -8.97423550e-02
1.44758299e-01 -3.61747414e-01 -3.97747189e-01 2.40769073e-01
-5.67682624e-01 1.20524280e-01 -4.76730429e-02 -3.90094727e-01
-8.84973288e-01 4.59886819e-01 -2.37200722e-01 -8.71067226e-01
-3.44004631e-01 3.76196265e-01 -1.57692045e-01 -3.72516841e-01
6.50061309e-01 1.27290413e-01 2.55725831e-01 -2.92277187e-01
6.79454505e-01 4.51936096e-01 7.63128400e-01 -7.41569638e-01
7.21455634e-01 4.93552804e-01 3.50867093e-01 -5.24842560e-01
-4.54510957e-01 -5.25837660e-01 -7.64758408e-01 -3.61352772e-01
9.49153543e-01 -1.19609725e+00 -9.44862664e-01 6.16091490e-01
-1.08902633e+00 -6.38621747e-01 -2.82184184e-01 5.06623745e-01
-9.04287338e-01 3.27169955e-01 -4.51749921e-01 -6.61106467e-01
4.58691306e-02 -1.49407828e+00 1.64816511e+00 1.49805248e-01
2.61940360e-01 -8.38028312e-01 1.09716356e-01 5.15246987e-01
1.99917376e-01 2.73342103e-01 8.29773009e-01 -2.86462158e-01
-6.37231767e-01 6.78855181e-02 -5.49641073e-01 4.23070669e-01
-2.15549543e-02 -2.98657000e-01 -1.52724135e+00 -2.05763236e-01
-2.23632008e-01 -8.30352008e-01 1.22182000e+00 1.67478427e-01
4.23729986e-01 2.94540524e-01 -1.14306331e-01 6.22900307e-01
1.52340722e+00 -1.65833116e-01 6.81900680e-01 4.45726275e-01
9.34895456e-01 5.42761445e-01 5.69118381e-01 2.31243744e-01
7.49167025e-01 6.76491737e-01 9.80972528e-01 -2.27780163e-01
-4.57107037e-01 1.19486518e-01 7.18408525e-01 4.33374524e-01
-3.21661569e-02 -2.23363191e-01 -1.24022317e+00 4.97623920e-01
-1.86307502e+00 -6.26509786e-01 1.12095535e-01 2.00959039e+00
7.49119639e-01 1.99121758e-01 1.21219911e-01 2.47962058e-01
9.96345580e-02 -2.13190526e-01 -6.00687921e-01 -2.86584020e-01
-3.81534457e-01 3.25914383e-01 6.62478149e-01 3.15565825e-01
-1.04144454e+00 6.69807017e-01 5.77229071e+00 4.79998350e-01
-1.23249853e+00 2.11553946e-02 2.68229157e-01 -1.94532901e-01
-6.73931912e-02 -1.17940389e-01 -6.05198920e-01 -2.47647300e-01
9.33492184e-01 4.91015941e-01 3.11139226e-01 5.92430949e-01
-2.15549394e-01 -4.53802347e-01 -1.52869582e+00 1.25170386e+00
3.72091621e-01 -8.62740457e-01 -5.40393963e-02 9.58414972e-02
5.03737450e-01 5.45122504e-01 3.97411078e-01 4.95075434e-01
2.97772497e-01 -1.16401219e+00 1.04458308e+00 4.66200918e-01
6.59029961e-01 -3.52206498e-01 6.51599944e-01 1.29529551e-01
-1.14718413e+00 -2.07578018e-01 -9.21611413e-02 2.23767951e-01
9.25861001e-02 2.76160538e-01 -1.09161568e+00 9.05054450e-01
9.22545969e-01 8.86550367e-01 -8.70677710e-01 1.10664022e+00
-1.55339748e-01 8.36356208e-02 -5.96804857e-01 3.67987037e-01
4.06510979e-01 1.33302778e-01 3.90722662e-01 1.24528539e+00
3.47928494e-01 -2.39483658e-02 7.02264011e-02 5.10476172e-01
1.10693961e-01 -3.71951252e-01 -7.56397963e-01 2.66884595e-01
-9.00888294e-02 1.27605116e+00 -4.94504809e-01 -1.89910661e-02
-5.35332024e-01 1.11965430e+00 2.30844647e-01 5.65897763e-01
-9.02463615e-01 8.45469609e-02 5.68422616e-01 -3.50778103e-01
7.06271470e-01 -4.58668828e-01 -4.66330439e-01 -8.81738782e-01
-5.93173169e-02 -7.81217098e-01 3.95118535e-01 -1.26880610e+00
-1.40390921e+00 4.35663790e-01 1.88170999e-01 -1.36623681e+00
-3.84162039e-01 -1.24755454e+00 -1.05825663e-01 8.12272429e-01
-1.82181990e+00 -1.79204130e+00 -5.25824487e-01 7.98181057e-01
3.34718525e-01 1.06338583e-01 7.18324006e-01 2.72208512e-01
-2.00204343e-01 4.65709835e-01 -1.37704462e-01 -2.68794507e-01
9.71778154e-01 -1.38618076e+00 7.89911207e-03 6.77336097e-01
2.83351839e-01 1.55125052e-01 6.68236315e-01 -1.89783141e-01
-1.99937654e+00 -8.13213050e-01 1.24340303e-01 -7.25170016e-01
4.16153580e-01 -3.58429641e-01 -7.36809254e-01 7.01689720e-01
4.46871698e-01 3.20515893e-02 4.38531309e-01 -2.49786675e-03
-6.51900530e-01 -1.83308318e-01 -1.10820460e+00 3.31544667e-01
9.74978209e-01 -7.04242587e-01 -7.26630509e-01 5.12326695e-02
5.77769160e-01 -6.65036976e-01 -1.08389664e+00 6.90400183e-01
8.05070221e-01 -1.16407371e+00 1.22945857e+00 3.55855152e-02
4.12319541e-01 -4.59279954e-01 -7.88203061e-01 -1.34541547e+00
1.25395149e-01 -1.73821196e-01 -1.15772821e-02 1.19429898e+00
2.99697250e-01 -4.57830995e-01 4.94990528e-01 3.79127681e-01
-5.78288198e-01 -6.68623328e-01 -1.34453821e+00 -4.85683769e-01
8.40133801e-02 -8.82002711e-01 7.72577003e-02 5.55461884e-01
-3.18386674e-01 5.35062850e-01 1.40495747e-02 2.54024714e-01
6.72249377e-01 -5.53502142e-02 1.06638777e+00 -9.68862474e-01
-4.34942633e-01 -6.86232567e-01 -6.17126048e-01 -1.12701273e+00
1.21516578e-01 -9.88931298e-01 4.02266234e-01 -1.66464543e+00
1.42095564e-02 -4.26430285e-01 -2.99141824e-01 8.20869029e-01
1.32282645e-01 5.77533185e-01 5.11916995e-01 1.25261784e-01
-5.54478526e-01 8.14341426e-01 1.12978435e+00 -4.28021878e-01
2.53627747e-02 -4.52442765e-01 -3.97060037e-01 4.31400895e-01
2.83616096e-01 -6.80210814e-03 -3.22396189e-01 -6.54364645e-01
7.14844391e-02 -4.13630530e-02 9.74421501e-01 -1.09475517e+00
3.54580551e-01 1.99961394e-01 6.67944312e-01 -5.57696939e-01
9.03504372e-01 -1.12224829e+00 5.69351390e-02 1.38097689e-01
1.03803374e-01 5.17792888e-02 7.77682602e-01 4.67543751e-01
-4.58948873e-02 2.46985942e-01 6.90414846e-01 8.18193480e-02
-1.08324456e+00 -9.03762970e-03 -2.53395997e-02 -2.46577591e-01
1.00388587e+00 -5.51549017e-01 -4.49524313e-01 -1.90652996e-01
-6.64970994e-01 2.43053168e-01 5.54091632e-01 4.81312662e-01
8.26595187e-01 -1.18490434e+00 -6.10936761e-01 3.14930558e-01
4.92046803e-01 1.37414515e-01 7.21421540e-02 1.16542697e+00
-1.71106175e-01 2.30115920e-01 -5.01394987e-01 -1.48018539e+00
-1.06848145e+00 3.83840352e-01 4.76902783e-01 -4.40219603e-03
-4.75942701e-01 9.71151769e-01 2.27064759e-01 -4.42739338e-01
4.86145526e-01 -5.51468909e-01 9.36163068e-02 1.49471819e-01
-9.79698449e-02 -8.63779336e-02 2.83520192e-01 -7.45777965e-01
-5.72620511e-01 9.25889611e-01 1.84444204e-01 -5.54591775e-01
1.38277566e+00 -3.47968429e-01 2.17609525e-01 7.88974345e-01
1.21670079e+00 -5.32353282e-01 -1.62324846e+00 -1.81604370e-01
-1.51677683e-01 -1.78690791e-01 8.02812055e-02 -9.96893823e-01
-9.00275171e-01 1.04929817e+00 1.10341144e+00 -5.44362180e-02
1.31452096e+00 4.34123605e-01 4.19558465e-01 4.59998012e-01
4.51472253e-01 -6.36254370e-01 4.13872629e-01 5.85770905e-01
1.02172577e+00 -1.59415781e+00 -3.22496518e-02 1.28790252e-02
-8.90041947e-01 1.11889875e+00 5.89307845e-01 -2.72681303e-02
3.27867925e-01 1.13726929e-01 3.67952228e-01 -2.65880913e-01
-7.96790183e-01 -6.92230225e-01 4.60108817e-01 8.12842607e-01
2.33267784e-01 -3.55719566e-01 6.82588518e-01 1.66124597e-01
1.14121757e-01 -1.99887156e-01 1.90564290e-01 1.02523518e+00
-2.16834530e-01 -1.02040124e+00 -5.01680613e-01 -4.92238812e-03
-9.62559506e-02 1.36308402e-01 -3.39747548e-01 1.02755010e+00
2.08916873e-01 9.33494806e-01 -4.83699292e-02 -5.14752090e-01
6.44719899e-01 1.15078703e-01 1.22782195e+00 -4.45452601e-01
-6.98407471e-01 2.84417063e-01 1.44338431e-02 -7.39327013e-01
-8.45327318e-01 -8.37346852e-01 -1.24155390e+00 7.67414868e-02
-1.58490568e-01 -4.43534017e-01 9.67945039e-01 1.14064598e+00
2.35218823e-01 6.31289959e-01 2.90713161e-01 -1.66604996e+00
-4.18309301e-01 -8.90682280e-01 -2.78736949e-01 3.02417696e-01
6.93879902e-01 -9.76810873e-01 -2.99375832e-01 8.29236209e-02]
|
[8.855507850646973, -1.883314847946167]
|
313f4013-221e-426e-bd51-bb388fd25414
|
benchmarking-automated-machine-learning
|
2304.14735
| null |
https://arxiv.org/abs/2304.14735v1
|
https://arxiv.org/pdf/2304.14735v1.pdf
|
Benchmarking Automated Machine Learning Methods for Price Forecasting Applications
|
Price forecasting for used construction equipment is a challenging task due to spatial and temporal price fluctuations. It is thus of high interest to automate the forecasting process based on current market data. Even though applying machine learning (ML) to these data represents a promising approach to predict the residual value of certain tools, it is hard to implement for small and medium-sized enterprises due to their insufficient ML expertise. To this end, we demonstrate the possibility of substituting manually created ML pipelines with automated machine learning (AutoML) solutions, which automatically generate the underlying pipelines. We combine AutoML methods with the domain knowledge of the companies. Based on the CRISP-DM process, we split the manual ML pipeline into a machine learning and non-machine learning part. To take all complex industrial requirements into account and to demonstrate the applicability of our new approach, we designed a novel metric named method evaluation score, which incorporates the most important technical and non-technical metrics for quality and usability. Based on this metric, we show in a case study for the industrial use case of price forecasting, that domain knowledge combined with AutoML can weaken the dependence on ML experts for innovative small and medium-sized enterprises which are interested in conducting such solutions.
|
['Christian Tutschku', 'Alexandre Beiderwellen-Bedrikow', 'Dennis Klau', 'Marc-André Zöller', 'Horst Stühler']
|
2023-04-28
| null | null | null | null |
['automl']
|
['methodology']
|
[-3.83540392e-02 3.61325949e-01 8.71125534e-02 -2.53313929e-01
-3.05410951e-01 -6.25784338e-01 4.83141243e-01 4.56034660e-01
-3.94324780e-01 5.01009285e-01 -4.91223961e-01 -4.80787456e-01
-5.47772348e-01 -1.03969014e+00 -7.03159690e-01 -1.03880860e-01
6.08580075e-02 8.00627887e-01 8.22463706e-02 -2.73909450e-01
6.58853233e-01 7.50482738e-01 -1.98689961e+00 6.57129467e-01
1.00318432e+00 1.31868708e+00 5.86618125e-01 2.94742733e-01
-5.70884109e-01 7.55818367e-01 -5.94044268e-01 -4.23525274e-01
4.06888276e-01 -1.16209157e-01 -5.20860076e-01 -6.91740513e-02
-1.30206212e-01 -7.78811052e-02 8.66221309e-01 7.96644449e-01
1.28950223e-01 -3.35013777e-01 6.66138411e-01 -1.48496246e+00
-5.87952770e-02 7.59947598e-01 -7.43116587e-02 -3.20718586e-01
3.07316452e-01 8.86797905e-03 9.85437214e-01 -7.35441864e-01
6.13491654e-01 7.86986411e-01 7.57980168e-01 -8.00906792e-02
-1.26685250e+00 -4.24581885e-01 2.27649406e-01 3.63202125e-01
-1.06685770e+00 1.93096697e-01 9.74940896e-01 -1.02362001e+00
9.95056510e-01 2.02557445e-01 8.77580822e-01 6.53494000e-01
2.02095404e-01 5.19713461e-01 1.31510973e+00 -8.28569293e-01
6.42305136e-01 7.99222410e-01 -2.47263506e-01 4.79629114e-02
3.53982240e-01 -1.07916571e-01 -1.75398484e-01 9.28278342e-02
4.43584263e-01 6.23398721e-02 1.68778554e-01 -3.35934371e-01
-8.48857224e-01 9.21141863e-01 -7.20545128e-02 7.78725088e-01
-7.64806569e-01 -2.09015757e-01 2.56937474e-01 7.07421780e-01
4.47036564e-01 8.55697930e-01 -9.69909072e-01 -3.80687892e-01
-1.21511173e+00 3.07304770e-01 1.24337959e+00 9.10222769e-01
8.83202910e-01 -2.61934847e-01 1.57359883e-01 5.17489910e-01
2.83344418e-01 2.72806644e-01 3.76542002e-01 -6.93204999e-01
2.01721296e-01 1.13882411e+00 3.43823850e-01 -8.73848379e-01
-5.46659350e-01 -5.02989113e-01 -3.55155677e-01 6.92743957e-01
3.89874905e-01 6.96428567e-02 -5.58802605e-01 1.06427836e+00
-1.81260221e-02 -2.14855939e-01 -1.39172941e-01 5.27454436e-01
2.31226280e-01 4.91382241e-01 1.45774949e-02 -5.84066510e-01
1.04870081e+00 -6.57260537e-01 -7.47442842e-01 1.90374658e-01
8.08493614e-01 -8.71460557e-01 1.11427784e+00 1.06936336e+00
-1.11204398e+00 -7.29784489e-01 -8.99370313e-01 4.14341062e-01
-1.00187778e+00 3.89851570e-01 5.01337111e-01 6.44393265e-01
-7.15644062e-01 1.20563829e+00 -5.26272416e-01 -9.63492319e-02
1.65753320e-01 5.07715702e-01 3.96674238e-02 1.44598037e-01
-1.14698553e+00 1.36590159e+00 4.92725074e-01 2.10967615e-01
-1.27212461e-02 -6.60331666e-01 -4.03511137e-01 2.75128394e-01
5.33184111e-01 -3.47892553e-01 1.29802239e+00 -1.08845174e+00
-1.52797556e+00 4.67215389e-01 4.18996006e-01 -4.76497501e-01
8.62268627e-01 -5.05170859e-02 -5.72280109e-01 -2.29767919e-01
-1.08326085e-01 1.50649875e-01 6.65104628e-01 -1.24790180e+00
-1.12791264e+00 -1.85572982e-01 -5.05137593e-02 -4.74833310e-01
-3.75435442e-01 -2.58150131e-01 1.32629846e-03 -4.31759059e-01
-1.51309267e-01 -7.76098430e-01 -2.26123795e-01 -6.48012102e-01
2.13325858e-01 -3.29248101e-01 5.28976977e-01 -8.20094764e-01
1.37034976e+00 -1.86967766e+00 8.10578186e-03 7.80974030e-01
-2.51649380e-01 1.22017525e-01 3.12445343e-01 6.02732182e-01
-3.39813232e-02 -1.56319719e-02 -4.34461832e-02 1.61675543e-01
4.32716519e-01 1.87551051e-01 4.96709868e-02 -8.64934400e-02
3.70180577e-01 5.65993786e-01 -8.17094326e-01 -2.70152360e-01
5.72321713e-01 1.74888358e-01 -4.19484437e-01 7.07866773e-02
-6.98996007e-01 4.15704697e-01 -2.94618785e-01 5.79881907e-01
6.54240787e-01 5.37218861e-02 2.58005470e-01 -3.36360067e-01
-5.10875225e-01 -6.10784926e-02 -1.62134385e+00 1.49968767e+00
-1.16928208e+00 1.46065146e-01 -1.83114305e-01 -1.06133366e+00
1.21073115e+00 2.74349004e-01 6.66817486e-01 -8.45645308e-01
1.88623160e-01 8.33899677e-01 -5.83591834e-02 -6.44841254e-01
3.01449984e-01 -1.50291011e-01 -7.66523853e-02 2.54815042e-01
8.88391882e-02 -2.48508841e-01 5.45195401e-01 -5.46280324e-01
9.18286026e-01 6.21155143e-01 3.13397318e-01 -1.67389736e-01
7.02771664e-01 2.54016221e-01 4.13659960e-01 3.99566203e-01
3.44414353e-01 2.36673042e-01 6.31192684e-01 -7.07290351e-01
-1.15885544e+00 -5.12746751e-01 1.41693830e-01 9.60308731e-01
-4.89057720e-01 -3.71289402e-01 -8.07536244e-01 -7.06726253e-01
2.30644882e-01 1.04256320e+00 -3.66599679e-01 1.95144370e-01
-2.93683380e-01 -2.40817443e-01 -3.00088972e-01 3.14761192e-01
3.35147791e-02 -1.12908685e+00 -1.04871583e+00 7.19634533e-01
2.73383230e-01 -9.84972358e-01 4.50250179e-01 2.90791154e-01
-6.79204166e-01 -8.94631445e-01 -4.77400690e-01 -3.00695270e-01
4.02210623e-01 -5.93725920e-01 1.31192756e+00 -6.84504360e-02
-3.33156854e-01 2.78597862e-01 -5.74239850e-01 -1.08209503e+00
-8.99388731e-01 4.59808916e-01 -3.10920209e-01 -6.59439787e-02
4.53153551e-01 -6.37931943e-01 -2.65245020e-01 2.68361956e-01
-9.29452240e-01 -5.99957556e-02 1.00922704e+00 4.32133973e-01
4.98757124e-01 5.63944936e-01 6.85566783e-01 -1.14553750e+00
7.68810928e-01 -2.51412570e-01 -1.28931773e+00 3.47464919e-01
-1.22241139e+00 1.84255719e-01 5.88065803e-01 -3.84084344e-01
-9.01878893e-01 3.76405686e-01 -2.94123948e-01 -2.79947579e-01
-3.88186365e-01 9.30591762e-01 -2.97503144e-01 1.67154297e-01
3.77500802e-01 -2.75975525e-01 9.02717859e-02 -6.88839555e-01
2.91275710e-01 6.95946097e-01 1.40240341e-01 -3.62656116e-01
9.22636569e-01 -7.00162798e-02 3.01150322e-01 -4.26857948e-01
-3.26321751e-01 -3.20342332e-01 -9.70439196e-01 -5.14812529e-01
7.88601577e-01 -4.82735634e-01 -8.48564804e-01 -8.08923021e-02
-1.19548380e+00 -3.00137341e-01 -7.22784638e-01 5.32251954e-01
-7.41930783e-01 -1.66004464e-01 1.03975656e-02 -1.16717839e+00
-6.87845051e-02 -1.07723761e+00 8.24901998e-01 -1.77693844e-01
-4.85437870e-01 -9.48831141e-01 -1.57580122e-01 1.32049814e-01
6.00701988e-01 5.97203910e-01 1.02807200e+00 -8.45651984e-01
-4.44558382e-01 -5.67564905e-01 1.09985843e-01 8.70507598e-01
1.16454601e-01 1.19828932e-01 -8.10435414e-01 2.16598332e-01
1.99055880e-01 2.29989976e-01 3.23259234e-01 1.95426941e-01
9.51181233e-01 -5.78238396e-04 -1.32131740e-01 -2.36515656e-01
1.62817585e+00 2.62301773e-01 4.65832680e-01 8.92986894e-01
2.84673631e-01 1.23569536e+00 1.14135098e+00 6.73044920e-01
1.25814170e-01 7.73630023e-01 3.60982180e-01 -1.22598276e-01
4.09630775e-01 1.35784999e-01 1.80802435e-01 5.99489152e-01
-6.24389768e-01 3.28528494e-01 -1.02544188e+00 3.40245724e-01
-2.09823704e+00 -8.68276000e-01 -4.10353899e-01 2.28460383e+00
4.69232589e-01 7.89868355e-01 4.05456036e-01 6.53268993e-01
3.83943349e-01 -6.27382457e-01 -1.49430055e-02 -7.60252655e-01
2.84508228e-01 3.97225410e-01 6.91158772e-01 1.26269430e-01
-8.62558424e-01 3.06598693e-01 4.73453140e+00 6.34519517e-01
-8.46794605e-01 6.65633529e-02 2.88570702e-01 1.27246603e-01
-3.68521512e-01 1.15472093e-01 -6.47594750e-01 6.06487155e-01
1.17238045e+00 -6.16630949e-02 5.43874681e-01 1.40687394e+00
4.86089647e-01 -5.10149598e-02 -1.38668787e+00 8.30412269e-01
-3.14627767e-01 -1.51413739e+00 -3.15065414e-01 2.80325800e-01
5.14578462e-01 -2.86136955e-01 -1.80944383e-01 3.70400399e-01
-1.64716858e-02 -8.51643205e-01 9.92852211e-01 8.37853670e-01
2.25375429e-01 -7.92548656e-01 1.33870816e+00 5.50772965e-01
-1.04239690e+00 -5.53509295e-01 -1.46009177e-01 -2.70449936e-01
1.91873565e-01 9.57901061e-01 -9.35335040e-01 9.04627800e-01
6.20106578e-01 1.78371817e-01 -4.36352581e-01 7.84421563e-01
4.52227741e-02 8.66342857e-02 -3.02791446e-01 -5.05033787e-03
7.83414394e-02 -3.10825199e-01 -1.83513939e-01 1.31638300e+00
6.88616037e-01 -4.87864316e-01 -4.64374647e-02 1.18587315e+00
4.38605636e-01 5.02905309e-01 -5.31649590e-01 -4.34021913e-02
1.37379780e-01 1.27280784e+00 -9.66302872e-01 -2.16146275e-01
-5.64372063e-01 6.69828594e-01 -1.78370818e-01 -3.42753455e-02
-5.47321260e-01 -4.98663396e-01 7.20727369e-02 6.08591139e-01
6.07655108e-01 -1.18959524e-01 -5.77144980e-01 -5.77029705e-01
3.26283395e-01 -8.95925820e-01 1.02684282e-01 -5.66197693e-01
-1.24275398e+00 5.08485138e-01 6.36244416e-02 -1.49818873e+00
-6.51312351e-01 -9.60939944e-01 -2.49939308e-01 9.36591744e-01
-1.54381979e+00 -1.31330371e+00 -2.75031567e-01 1.26985729e-01
4.85764205e-01 -2.48302788e-01 7.05357671e-01 2.00744659e-01
-1.00631960e-01 -4.38121669e-02 -6.35970831e-02 -4.22472894e-01
2.98765004e-01 -1.46180153e+00 1.17115229e-01 4.43111330e-01
-1.81395514e-03 2.61342317e-01 9.15451765e-01 -7.00349331e-01
-1.16675019e+00 -8.32117021e-01 1.20254290e+00 -5.10086179e-01
7.85936534e-01 -5.44534683e-01 -8.57045889e-01 2.32645169e-01
2.52255350e-02 -3.66751313e-01 5.37600636e-01 3.20583843e-02
1.01417154e-01 -4.03086811e-01 -1.27816415e+00 -4.45022881e-02
5.49230754e-01 -2.23114818e-01 -6.77340031e-01 3.40328068e-01
2.60519058e-01 9.70632955e-02 -1.56749701e+00 5.50092399e-01
5.45621574e-01 -1.02992940e+00 5.29394865e-01 -1.25895843e-01
3.73744428e-01 -2.35984653e-01 7.27467760e-02 -1.08392131e+00
-2.34096069e-02 -4.18568552e-01 1.10241314e-02 1.44244182e+00
8.04380715e-01 -6.00985110e-01 6.81447923e-01 9.80189383e-01
1.39398770e-02 -6.86620772e-01 -8.12519789e-01 -9.38214004e-01
-1.50680035e-01 -7.61688411e-01 8.12357664e-01 9.09967542e-01
7.47067854e-02 5.12124449e-02 5.44581525e-02 -6.84913481e-03
2.45019287e-01 3.40667546e-01 1.00982630e+00 -2.04568815e+00
-4.80464488e-01 -5.66305161e-01 -5.27935088e-01 -1.13186687e-01
1.34159222e-01 -4.83106285e-01 -1.26620799e-01 -1.85064125e+00
-4.01729196e-01 -4.51108217e-01 -5.68465292e-01 2.86929488e-01
4.56947535e-01 -1.21249042e-01 3.06567252e-01 1.90667942e-01
-1.19164772e-01 -1.19006127e-01 8.50897551e-01 8.81218836e-02
-4.84686166e-01 3.85696918e-01 -3.56816530e-01 9.15082335e-01
7.36530781e-01 -4.09419268e-01 -3.80366713e-01 1.22419477e-01
8.74639928e-01 -3.39003444e-01 7.83421472e-02 -9.97877836e-01
4.82990220e-02 -2.30849490e-01 3.33582699e-01 -6.55417740e-01
-1.75022200e-01 -1.54037809e+00 6.18252635e-01 6.58097744e-01
2.48637516e-02 8.65523294e-02 4.21471484e-02 8.13984126e-02
-3.55474234e-01 -6.65196836e-01 2.38226622e-01 -2.97023565e-01
-7.63800085e-01 -3.55380625e-01 -3.53957385e-01 -6.86647117e-01
1.12508631e+00 -4.08182919e-01 2.00777695e-01 4.49923857e-04
-9.24099386e-01 -9.10649151e-02 4.91314113e-01 2.98914760e-01
1.61016732e-01 -9.18903410e-01 -4.82983947e-01 2.54261225e-01
2.69699991e-01 -4.43737470e-02 -3.36984247e-02 8.27262402e-01
-6.18294001e-01 3.54978293e-01 -4.16069537e-01 -4.42785174e-01
-8.94433379e-01 1.00618327e+00 -6.66209534e-02 -5.79174101e-01
-2.39541516e-01 1.81659594e-01 -5.79545081e-01 -3.32759589e-01
6.79591373e-02 -9.29755092e-01 -4.95634317e-01 4.69463795e-01
5.79464324e-02 5.43924034e-01 6.97232008e-01 -2.18380198e-01
-2.07644254e-01 5.81647515e-01 2.47751638e-01 -1.77112803e-01
1.72830725e+00 1.12898707e-01 -7.53866807e-02 6.26995087e-01
7.41158485e-01 1.05443113e-01 -9.56958950e-01 2.86560208e-01
1.01033318e+00 -3.42644244e-01 1.03570327e-01 -8.64297748e-01
-8.37123990e-01 7.10880637e-01 8.33486736e-01 8.33108246e-01
1.42332256e+00 -1.51210487e-01 4.04425085e-01 3.82548928e-01
6.16773427e-01 -1.85565257e+00 -3.14898789e-01 -1.11291714e-01
9.05687928e-01 -1.14144981e+00 5.48541844e-02 -7.32214272e-01
-4.68781114e-01 1.47315264e+00 1.13842092e-01 1.72460988e-01
7.87887931e-01 4.51261759e-01 -2.22095922e-02 -1.29923895e-01
-4.57066089e-01 -3.46029669e-01 1.36247203e-01 5.54132283e-01
4.88075227e-01 4.00403440e-02 -6.85911357e-01 1.13122058e+00
-3.45213503e-01 7.45725691e-01 5.46108007e-01 1.20285964e+00
-4.40101415e-01 -1.68465972e+00 -3.49173695e-01 5.09520173e-01
-3.69794846e-01 2.66319484e-01 -2.88463652e-01 1.17694807e+00
9.08580959e-01 8.96625042e-01 -9.01188552e-02 -5.38431108e-01
1.10192227e+00 2.35263065e-01 5.02743065e-01 -6.25573456e-01
-9.08609688e-01 2.19245255e-01 3.16790998e-01 -4.61181313e-01
-5.35382867e-01 -7.45198965e-01 -9.62734342e-01 2.05392819e-02
-4.92707402e-01 2.11897045e-01 1.46426666e+00 1.01573706e+00
2.28102654e-01 7.56482065e-01 6.89804316e-01 -9.26506877e-01
-7.12589383e-01 -9.01934206e-01 -7.02953041e-01 7.21738517e-01
-2.82497734e-01 -8.84975433e-01 -3.54942113e-01 3.25106531e-01]
|
[9.053727149963379, 6.103189945220947]
|
eb7eddec-de2f-4d2c-88e7-fa26979f8f2d
|
visual-and-semantic-knowledge-transfer-for
|
1801.03145
| null |
http://arxiv.org/abs/1801.03145v2
|
http://arxiv.org/pdf/1801.03145v2.pdf
|
Visual and Semantic Knowledge Transfer for Large Scale Semi-supervised Object Detection
|
Deep CNN-based object detection systems have achieved remarkable success on
several large-scale object detection benchmarks. However, training such
detectors requires a large number of labeled bounding boxes, which are more
difficult to obtain than image-level annotations. Previous work addresses this
issue by transforming image-level classifiers into object detectors. This is
done by modeling the differences between the two on categories with both
image-level and bounding box annotations, and transferring this information to
convert classifiers to detectors for categories without bounding box
annotations. We improve this previous work by incorporating knowledge about
object similarities from visual and semantic domains during the transfer
process. The intuition behind our proposed method is that visually and
semantically similar categories should exhibit more common transferable
properties than dissimilar categories, e.g. a better detector would result by
transforming the differences between a dog classifier and a dog detector onto
the cat class, than would by transforming from the violin class. Experimental
results on the challenging ILSVRC2013 detection dataset demonstrate that each
of our proposed object similarity based knowledge transfer methods outperforms
the baseline methods. We found strong evidence that visual similarity and
semantic relatedness are complementary for the task, and when combined notably
improve detection, achieving state-of-the-art detection performance in a
semi-supervised setting.
|
['Yu-Xing Tang', 'Emmanuel Dellandrea', 'Boyang Gao', 'Xiaofang Wang', 'Robert Gaizauskas', 'Liming Chen', 'Josiah Wang']
|
2018-01-09
| null | null | null | null |
['semi-supervised-object-detection']
|
['computer-vision']
|
[ 1.57281026e-01 5.44776693e-02 -1.64198324e-01 -5.81944168e-01
-6.14283204e-01 -8.03283215e-01 6.59618139e-01 5.38849115e-01
-7.74635971e-01 2.14025378e-01 4.90191802e-02 -8.76934677e-02
3.43673348e-01 -8.08090866e-01 -1.02793717e+00 -4.33445841e-01
1.95768341e-01 3.70523036e-01 8.37927103e-01 -8.58863965e-02
2.54573971e-01 3.56710255e-01 -1.71359730e+00 6.62911713e-01
4.65773880e-01 1.06435966e+00 2.70114422e-01 4.59740937e-01
1.21515132e-01 6.71702266e-01 -3.79436553e-01 -4.91665125e-01
5.46406269e-01 -3.06285709e-01 -8.75706553e-01 -4.72969487e-02
1.06435537e+00 -4.17210817e-01 -1.49320319e-01 1.28485215e+00
1.37101710e-01 8.62354264e-02 9.75773752e-01 -1.48672688e+00
-9.72684026e-01 4.18805152e-01 -6.69747233e-01 3.63934249e-01
1.74573362e-01 2.30225831e-01 1.15111458e+00 -9.64382887e-01
5.34818828e-01 1.47581899e+00 8.01805198e-01 5.35455942e-01
-1.49185479e+00 -8.53895605e-01 1.87738344e-01 3.88673872e-01
-1.50211585e+00 -1.45954028e-01 5.43587029e-01 -7.52371132e-01
1.05377996e+00 -8.28551948e-02 5.05515873e-01 9.62141931e-01
-2.09095433e-01 9.20714200e-01 1.05327487e+00 -4.07137215e-01
5.10136560e-02 6.14276767e-01 2.27168769e-01 6.11728966e-01
5.72503328e-01 3.59108418e-01 -2.69938737e-01 1.24551468e-01
7.09465384e-01 7.88745359e-02 1.71482768e-02 -8.24875653e-01
-1.17025352e+00 1.02038527e+00 1.29801464e+00 2.02783629e-01
-1.13077171e-01 2.86266387e-01 4.82655853e-01 1.82557821e-01
1.74610198e-01 4.58404869e-01 -2.88312614e-01 5.12992442e-01
-7.51266122e-01 2.28127584e-01 4.97141302e-01 1.22228158e+00
8.69195402e-01 -3.53708833e-01 -3.79603535e-01 5.59705436e-01
4.30521309e-01 3.72764140e-01 2.77347982e-01 -7.63583601e-01
3.57695490e-01 6.80880129e-01 2.12211564e-01 -9.36639071e-01
-3.51110369e-01 -3.33790034e-01 -4.11617815e-01 3.61697912e-01
7.14712918e-01 2.95615703e-01 -9.81737971e-01 1.88816810e+00
3.75270516e-01 -3.53283845e-02 -1.19495429e-02 1.18491030e+00
1.02828634e+00 3.50600898e-01 5.67904651e-01 4.30553168e-01
1.83954298e+00 -9.18888032e-01 -1.50441289e-01 -5.94948292e-01
8.26669157e-01 -7.98067093e-01 9.94680345e-01 -1.68182060e-01
-8.37105811e-01 -9.24177349e-01 -1.15255582e+00 -2.78838992e-01
-4.83298093e-01 3.78232151e-01 5.19643247e-01 5.28198957e-01
-8.22000623e-01 4.47803169e-01 -7.86929667e-01 -8.78715277e-01
7.65282631e-01 4.07241238e-03 -3.80718857e-01 -2.10016575e-02
-8.68652701e-01 1.29606879e+00 8.04739177e-01 -3.98808748e-01
-8.69439602e-01 -6.60200119e-01 -9.28596020e-01 1.33828536e-01
1.85197011e-01 -6.18323207e-01 1.27456629e+00 -1.16421616e+00
-6.44503415e-01 1.50915468e+00 1.36219293e-01 -7.07585216e-01
5.53586721e-01 -1.15489379e-01 -8.33700076e-02 1.16492026e-02
3.34674329e-01 1.47677171e+00 8.62737179e-01 -1.26878345e+00
-1.13226509e+00 -3.86114240e-01 6.41759560e-02 1.18013255e-01
-3.39165986e-01 1.55659169e-01 -1.76472783e-01 -4.14704531e-01
5.16911149e-02 -1.01991951e+00 1.02249160e-01 6.80970907e-01
-2.75623083e-01 -6.91343486e-01 9.39339936e-01 -4.44351673e-01
5.07006228e-01 -2.21619439e+00 -1.30269229e-01 -1.37810752e-01
2.64172375e-01 4.45586622e-01 -3.60201001e-01 6.26466647e-02
2.09271777e-02 -9.85352248e-02 1.73045266e-02 -4.96193543e-02
4.37978283e-02 -7.01523647e-02 -4.37098563e-01 6.07801557e-01
5.30544877e-01 9.95580792e-01 -9.84684110e-01 -5.59098899e-01
3.02304983e-01 2.19537079e-01 -6.44767702e-01 1.13551959e-01
-1.82255860e-02 2.02056766e-01 -1.60941258e-01 5.95076740e-01
5.93511581e-01 -1.91324696e-01 -1.19067982e-01 -4.83912855e-01
1.68223530e-01 2.92600662e-01 -9.89297092e-01 1.44097543e+00
-2.08596528e-01 9.78618741e-01 -1.58229455e-01 -1.20146024e+00
1.00795197e+00 -1.23401592e-02 -5.11969887e-02 -7.13394821e-01
2.91803896e-01 1.32894099e-01 4.18792993e-01 -2.41274267e-01
3.67750734e-01 -1.74783111e-01 -3.02330442e-02 3.37648183e-01
3.58207494e-01 -3.34269196e-01 1.83267877e-01 2.50280321e-01
8.04162323e-01 4.69337776e-02 4.62589025e-01 -2.33276039e-01
2.10231915e-01 3.01286310e-01 4.74934667e-01 1.06206715e+00
-7.63576329e-01 4.61785257e-01 2.24810570e-01 -3.79048795e-01
-1.29061568e+00 -1.26311183e+00 -3.60276490e-01 1.36760497e+00
4.74063903e-01 -1.15100011e-01 -7.50396013e-01 -9.35316026e-01
4.27374005e-01 6.84977233e-01 -8.91597450e-01 -4.99750823e-01
-3.47370625e-01 -3.24018836e-01 6.28033638e-01 1.14313519e+00
5.29406667e-01 -1.09541750e+00 -7.98651576e-01 3.93917263e-02
-3.41209248e-02 -1.32900548e+00 -4.82942909e-01 3.36100906e-01
-6.79967523e-01 -1.25830519e+00 -4.50606376e-01 -1.12510705e+00
6.35679007e-01 6.51160300e-01 1.14841175e+00 8.99508968e-02
-7.71674037e-01 4.14694995e-01 -3.33430737e-01 -7.91941881e-01
-4.64475572e-01 -2.26851180e-01 7.22824559e-02 -1.16719075e-01
9.34760392e-01 -2.74151261e-03 -7.31829286e-01 5.24431825e-01
-5.87789953e-01 -2.12921109e-02 7.10559249e-01 7.78535128e-01
3.62749219e-01 -4.79421139e-01 4.46369857e-01 -3.51213366e-01
1.94746658e-01 -2.19846502e-01 -6.53540552e-01 2.07965046e-01
-3.33102733e-01 1.40107321e-02 2.33002841e-01 -6.70139074e-01
-8.71634066e-01 4.78963166e-01 3.68119389e-01 -3.89089465e-01
-4.18833852e-01 -1.82275206e-01 7.13661760e-02 -1.87892437e-01
1.08991575e+00 2.43744329e-02 -2.15154096e-01 -2.08403796e-01
7.11959839e-01 6.47030890e-01 7.77098894e-01 -3.83769661e-01
1.02353573e+00 7.38273680e-01 -2.60832280e-01 -4.71061587e-01
-1.14656699e+00 -9.56142604e-01 -9.75671291e-01 -2.36171484e-02
1.25179374e+00 -1.10286832e+00 -5.79902172e-01 1.84845403e-01
-1.37564564e+00 -6.89873174e-02 -4.11399513e-01 5.84773362e-01
-4.71809894e-01 7.92484879e-02 -5.15395284e-01 -4.68988180e-01
-1.00500219e-01 -9.41745162e-01 1.20622110e+00 2.19024122e-01
-3.76204997e-01 -7.37931311e-01 -1.29555315e-01 3.58979732e-01
2.09316343e-01 -1.23843830e-02 8.53170216e-01 -9.41178024e-01
-5.84642947e-01 -4.75226462e-01 -9.62078094e-01 4.32456613e-01
-7.34003708e-02 -2.44159818e-01 -1.03579772e+00 -4.22167569e-01
-2.83024997e-01 -6.56738043e-01 1.25589168e+00 1.06279597e-01
8.60018790e-01 7.62980804e-02 -7.26847053e-01 4.45087731e-01
1.33094847e+00 -6.67397976e-02 3.00877005e-01 3.74593884e-01
8.22373152e-01 7.29615867e-01 6.86922133e-01 1.56304743e-02
2.80471623e-01 9.78575349e-01 4.11600858e-01 -3.30786407e-01
-5.17438531e-01 -3.65433693e-01 2.23106816e-01 -1.96200665e-02
1.73351869e-01 1.08141406e-02 -1.08345866e+00 7.81085968e-01
-1.90628767e+00 -9.86971557e-01 -2.66076922e-01 2.11927938e+00
8.48615050e-01 1.86883420e-01 3.75965625e-01 -5.77398352e-02
8.95236194e-01 -3.97180319e-01 -5.99717855e-01 -1.93515673e-01
1.80096850e-02 -4.81487103e-02 6.45118535e-01 -8.24041665e-02
-1.57966721e+00 1.17097187e+00 6.04753351e+00 5.73284149e-01
-8.14297616e-01 3.48263144e-01 2.27452457e-01 8.05012360e-02
4.04567599e-01 7.48737231e-02 -1.06442976e+00 1.55079693e-01
4.28051770e-01 -4.04986441e-02 -7.72775561e-02 1.24661100e+00
-2.22409710e-01 -3.13791037e-02 -1.79884541e+00 9.54766989e-01
2.23985970e-01 -1.11767519e+00 5.10105416e-02 -1.54344842e-01
7.40712583e-01 1.21986948e-01 1.38563598e-02 4.81856018e-01
6.91620350e-01 -9.10994232e-01 1.05655348e+00 2.11609341e-02
3.93300205e-01 -3.21994722e-01 6.68217063e-01 2.36308053e-01
-1.40885055e+00 -2.71182090e-01 -7.56890535e-01 -1.06662869e-01
-2.20316768e-01 4.80888225e-02 -1.11849523e+00 -5.47620989e-02
1.02030206e+00 7.15489626e-01 -8.60688269e-01 1.35472143e+00
-4.82699215e-01 3.89534026e-01 -1.93204135e-01 -1.10048773e-02
2.31125578e-01 2.17074141e-01 3.61175895e-01 1.48401630e+00
-3.10403369e-02 -8.65108967e-02 3.89371037e-01 1.41414034e+00
-2.16423005e-01 -9.37134624e-02 -5.24047256e-01 2.45682776e-01
5.63843727e-01 1.31764984e+00 -9.34343457e-01 -5.64332962e-01
-6.04559779e-01 9.23180878e-01 5.46150565e-01 9.32370052e-02
-1.01024270e+00 -4.62133467e-01 7.33683944e-01 6.80771917e-02
6.13296926e-01 2.79604457e-02 -4.25594568e-01 -8.89856100e-01
-1.80667508e-02 -4.51718390e-01 5.76866388e-01 -8.08924139e-01
-1.52551854e+00 2.33684376e-01 1.14236008e-02 -1.37344134e+00
9.10837725e-02 -9.50922012e-01 -5.37156641e-01 6.86932385e-01
-1.49248385e+00 -1.47476673e+00 -6.46709561e-01 5.64376891e-01
3.92173797e-01 4.40032780e-02 5.53407967e-01 2.13394448e-01
-1.63231432e-01 7.38223016e-01 -2.17349008e-01 6.04470789e-01
9.90388870e-01 -1.19738424e+00 4.87422556e-01 8.46406758e-01
5.34302831e-01 4.46706444e-01 4.44104820e-01 -4.88253176e-01
-9.33086038e-01 -1.37660491e+00 5.36770165e-01 -9.17694569e-01
7.33280122e-01 -5.41585147e-01 -1.14270484e+00 6.79458439e-01
-2.75222920e-02 3.48292440e-01 2.97696978e-01 4.65447530e-02
-1.16410661e+00 1.21679716e-01 -1.07559979e+00 2.22752824e-01
1.28187644e+00 -7.95029163e-01 -1.01010048e+00 5.13339877e-01
5.89978993e-01 -2.02573806e-01 -4.23902541e-01 4.16748673e-01
4.53425080e-01 -7.33325422e-01 1.24762750e+00 -9.29436207e-01
3.05238754e-01 -5.84123671e-01 -2.91858464e-01 -1.14607286e+00
-5.17440081e-01 3.64738941e-01 1.91722333e-01 1.06392312e+00
2.47516498e-01 -3.50608796e-01 5.60157418e-01 3.84537935e-01
-3.58581282e-02 -1.00190833e-01 -6.81352735e-01 -1.36945283e+00
2.40073249e-01 -2.54222542e-01 1.99722067e-01 9.48657632e-01
-2.01257467e-01 3.48848373e-01 7.47264549e-02 2.88143009e-01
7.20739424e-01 2.02161118e-01 8.49529266e-01 -1.37579250e+00
-2.01524287e-01 -7.02181518e-01 -1.07058370e+00 -8.13924015e-01
1.55871063e-01 -1.26751423e+00 3.66788447e-01 -1.47998405e+00
6.61598265e-01 -4.17948812e-01 -3.72018754e-01 8.23588312e-01
-2.81251520e-01 7.38565207e-01 5.44444323e-01 4.33954746e-01
-7.32626379e-01 2.69983292e-01 9.17171478e-01 -4.74409491e-01
2.89712287e-02 -3.25413018e-01 -5.71201861e-01 8.33977580e-01
4.29909497e-01 -6.94727063e-01 -9.70144570e-02 -3.34844142e-01
-1.17962986e-01 -5.95688999e-01 1.05305707e+00 -1.13492680e+00
2.65592307e-01 9.21145529e-02 6.65337324e-01 -4.08060253e-01
1.72653198e-01 -7.97536552e-01 -4.31155443e-01 7.87165046e-01
-4.88043308e-01 -4.76666033e-01 4.77522969e-01 7.86671162e-01
-1.48972377e-01 -1.16303146e-01 1.17791688e+00 2.91607901e-02
-1.17890429e+00 3.61845903e-02 -7.61871785e-02 1.60934612e-01
1.31237650e+00 -5.05237341e-01 -4.55538452e-01 -1.12152316e-01
-6.84044719e-01 7.37514868e-02 4.19770777e-01 7.98493624e-01
4.78939354e-01 -1.34016073e+00 -8.57799053e-01 1.51594486e-02
7.41388798e-01 -3.43715400e-01 -1.93799376e-01 7.35244334e-01
-1.24352872e-01 4.22684520e-01 -4.99596715e-01 -1.05537224e+00
-1.45135522e+00 8.91360819e-01 4.72397506e-01 2.69587547e-01
-4.33224052e-01 1.13502097e+00 8.41759086e-01 -5.06763577e-01
3.42024416e-01 -4.34241623e-01 3.59062217e-02 1.31407619e-01
5.10945439e-01 1.05432712e-01 -1.48939947e-02 -6.88045740e-01
-5.70931852e-01 7.67722487e-01 -3.13864648e-01 2.14965999e-01
9.44994688e-01 1.54678836e-01 2.82094508e-01 1.17347598e-01
1.26424468e+00 -5.91708481e-01 -1.32155144e+00 -6.25045121e-01
7.34431744e-02 -5.23952067e-01 5.37784994e-02 -8.00862610e-01
-7.78214157e-01 1.11861634e+00 9.92281914e-01 -7.48171844e-03
7.39769816e-01 6.63362920e-01 4.87259418e-01 6.15463972e-01
4.19696003e-01 -9.83836412e-01 3.96610796e-01 3.88435215e-01
8.36831748e-01 -1.89560616e+00 -4.52745296e-02 -5.88637650e-01
-6.78584754e-01 8.51694524e-01 1.06415808e+00 -3.46758932e-01
3.13450545e-01 -1.12359568e-01 -1.76638126e-01 -1.89402163e-01
-3.36550266e-01 -6.90236747e-01 6.69571817e-01 9.13080215e-01
3.83414596e-01 2.08366573e-01 -9.67090391e-03 5.29980361e-01
2.17870586e-02 -3.50406289e-01 1.87144488e-01 8.47304642e-01
-7.65536666e-01 -7.20970213e-01 -4.57535475e-01 3.58218551e-01
6.92371801e-02 -5.81336617e-02 -7.02362299e-01 1.00338221e+00
3.13713104e-01 8.68428409e-01 5.44241607e-01 -1.81974217e-01
3.63480031e-01 -1.68933291e-02 6.93949521e-01 -9.64410126e-01
-5.37132740e-01 -3.21890771e-01 -1.95419341e-01 -5.74668705e-01
-3.34274679e-01 -6.89078033e-01 -1.31334496e+00 3.09264716e-02
-6.44611359e-01 -3.18293750e-01 7.34754384e-01 8.60936582e-01
2.73286194e-01 3.54723126e-01 1.66703939e-01 -9.46888149e-01
-7.67781913e-01 -9.71619666e-01 -3.32085907e-01 1.04270446e+00
1.98662356e-01 -9.95571494e-01 -2.19495922e-01 1.76944107e-01]
|
[9.511161804199219, 1.4875057935714722]
|
bf73dcf2-3987-4904-a060-3da67afda7fc
|
text-categorization-for-conflict-event
| null | null |
https://aclanthology.org/2020.aespen-1.5
|
https://aclanthology.org/2020.aespen-1.5.pdf
|
Text Categorization for Conflict Event Annotation
|
We cast the problem of event annotation as one of text categorization, and compare state of the art text categorization techniques on event data produced within the Uppsala Conflict Data Program (UCDP). Annotating a single text involves assigning the labels pertaining to at least 17 distinct categorization tasks, e.g., who were the attacking organization, who was attacked, and where did the event take place. The text categorization techniques under scrutiny are a classical Bag-of-Words approach; character-based contextualized embeddings produced by ELMo; embeddings produced by the BERT base model, and a version of BERT base fine-tuned on UCDP data; and a pre-trained and fine-tuned classifier based on ULMFiT. The categorization tasks are very diverse in terms of the number of classes to predict as well as the skeweness of the distribution of classes. The categorization results exhibit a large variability across tasks, ranging from 30.3{\%} to 99.8{\%} F-score.
|
['Fehmi ben Abdesslem', 'Magnus Sahlgren', 'Ariel Ekgren', 'Kristine Eck', 'Fredrik Olsson']
|
2020-05-01
| null | null | null |
lrec-2020-5
|
['text-categorization']
|
['natural-language-processing']
|
[-5.47663011e-02 -1.17557831e-02 -1.57727212e-01 -4.20838803e-01
-6.24234915e-01 -8.80444646e-01 1.24861443e+00 1.20053864e+00
-9.88332987e-01 4.99097824e-01 9.05667245e-01 -5.18351495e-01
-4.71135646e-01 -7.46113122e-01 -5.96357090e-03 -6.64348722e-01
-1.80805504e-01 9.27966952e-01 1.28006577e-01 -2.94064462e-01
6.12978280e-01 3.07662994e-01 -1.14223766e+00 4.46576864e-01
6.58582002e-02 1.02868450e+00 -3.16325247e-01 7.00386405e-01
-1.77657187e-01 8.52210760e-01 -8.51079047e-01 -3.95575225e-01
-4.09664102e-02 4.92571406e-02 -1.14973593e+00 -3.73920083e-01
-6.03016652e-03 1.44260805e-02 -2.97601968e-01 5.79919219e-01
4.50122088e-01 4.29597080e-01 1.31092381e+00 -1.34061432e+00
-6.51334405e-01 8.40009034e-01 -4.51519310e-01 9.09920037e-01
3.14115793e-01 -4.06800777e-01 1.29996407e+00 -7.78491378e-01
8.90357852e-01 1.31328440e+00 5.64118326e-01 1.02773458e-01
-1.40575218e+00 -5.02658784e-01 1.20796412e-01 4.31205869e-01
-1.21393728e+00 -1.43156201e-01 5.42891204e-01 -1.05083287e+00
9.20951426e-01 -6.16903715e-02 1.62381947e-01 1.33899283e+00
4.84785676e-01 1.18534803e-01 9.52655375e-01 -5.45284271e-01
5.11670351e-01 9.24679339e-02 5.11812270e-01 1.49210036e-01
2.48046711e-01 -5.00955656e-02 -5.59527457e-01 -8.58900666e-01
-2.90415995e-02 3.93156558e-02 1.25975549e-01 -1.85165778e-01
-1.22416389e+00 1.28030038e+00 1.57404646e-01 4.76921827e-01
-2.87960649e-01 1.75280944e-02 1.12347150e+00 1.22198306e-01
6.84460402e-01 4.80531991e-01 -6.70758128e-01 -3.08434129e-01
-7.05304801e-01 4.80383813e-01 7.09570825e-01 2.85077393e-01
4.17852432e-01 -2.78489709e-01 -2.82900017e-02 8.74564111e-01
1.20366611e-01 -1.48137599e-01 7.35786498e-01 -5.34392297e-01
7.02427506e-01 5.53898811e-01 2.08072692e-01 -1.06396282e+00
-7.52984345e-01 7.92869255e-02 -2.45328829e-01 9.22332928e-02
6.32584989e-01 -3.26545358e-01 -5.96462369e-01 1.72716546e+00
3.14037800e-01 -4.38683510e-01 2.69626379e-02 5.63826680e-01
4.87617671e-01 6.80207372e-01 5.76738119e-01 5.23714796e-02
1.60361457e+00 -4.88374054e-01 -5.74819028e-01 -3.38637978e-01
8.26647162e-01 -5.68092108e-01 6.02904558e-01 1.72912702e-01
-4.26791161e-01 -4.11950618e-01 -8.92256975e-01 4.85541038e-02
-9.26918566e-01 -3.48013729e-01 4.37955976e-01 4.77342248e-01
-5.47007203e-01 4.41568106e-01 -5.69155216e-01 -5.39649129e-01
1.25041589e-01 3.10448799e-02 -6.16854131e-01 2.58552194e-01
-1.39512503e+00 1.16658998e+00 9.93002534e-01 -4.74241287e-01
-6.02451980e-01 -5.17978013e-01 -7.42757142e-01 1.32500142e-01
1.91761002e-01 2.08296120e-01 9.37475085e-01 -5.50766051e-01
-5.73456347e-01 1.17977405e+00 2.58559287e-01 -5.20769358e-01
2.67047882e-01 6.19842187e-02 -7.77130961e-01 2.57028699e-01
5.57138085e-01 3.69895190e-01 4.59575385e-01 -8.91249537e-01
-9.18681800e-01 -6.98469162e-01 2.27712886e-03 7.21798763e-02
-5.18278360e-01 6.61585569e-01 5.93159795e-01 -8.66314232e-01
-1.97019592e-01 -8.25258791e-01 -1.56443976e-02 -3.62334549e-01
-1.52660340e-01 -8.33412290e-01 6.55855179e-01 -6.98489785e-01
1.40616691e+00 -2.14036798e+00 1.08487085e-01 -2.00668480e-02
1.70291275e-01 -2.00775251e-01 2.18406111e-01 1.04651594e+00
-5.31454027e-01 4.19512421e-01 1.02117009e-01 3.89317819e-03
4.15038645e-01 8.41810554e-02 -6.96757317e-01 5.44923186e-01
4.38375399e-02 1.68287382e-01 -8.69384468e-01 -6.01562560e-01
4.55573983e-02 -7.36494809e-02 -2.59414375e-01 -6.19943962e-02
2.91356206e-01 -1.21268630e-01 -4.94733006e-01 3.99593800e-01
6.33973554e-02 3.58731449e-01 4.04630691e-01 4.02201749e-02
-4.82586712e-01 3.54026794e-01 -1.02246726e+00 1.23187304e+00
-1.31178930e-01 8.34270418e-01 -1.74762547e-01 -1.21247995e+00
9.67467546e-01 3.83703589e-01 5.25622129e-01 -6.67584166e-02
4.36844885e-01 7.69958794e-02 1.55604348e-01 -2.32751265e-01
7.45639443e-01 -2.41925418e-01 -1.06147492e+00 8.90552521e-01
4.55669105e-01 1.30830705e-01 4.19519156e-01 2.84492999e-01
1.30467844e+00 -2.44342908e-01 5.30849278e-01 -6.05199695e-01
5.59387393e-02 4.64183480e-01 4.87695754e-01 7.13970184e-01
-5.35594940e-01 1.73755780e-01 9.22212481e-01 -7.96820343e-01
-1.06883764e+00 -9.81489837e-01 -7.08098650e-01 1.76449251e+00
-1.48156404e-01 -6.26260161e-01 -4.07399476e-01 -7.55362749e-01
6.39675334e-02 9.89846468e-01 -1.17829084e+00 -2.81518191e-01
-1.49015412e-01 -8.15619528e-01 7.86974251e-01 6.69484675e-01
-5.45061044e-02 -1.04851413e+00 -7.28068292e-01 3.31385285e-01
-2.76143610e-01 -9.01081204e-01 -2.34922513e-01 8.27879965e-01
-2.22106695e-01 -1.28510559e+00 -5.46423420e-02 -6.03068054e-01
1.67588174e-01 -3.87175351e-01 9.63337660e-01 -4.66437757e-01
-3.19788009e-01 4.17236015e-02 -7.73848593e-01 -7.20092654e-01
-3.62462491e-01 3.71222571e-02 4.04227942e-01 -8.48242790e-02
8.99652064e-01 -1.58094153e-01 -1.66227326e-01 2.52051532e-01
-8.95780802e-01 -5.89789987e-01 -4.59353365e-02 8.58320355e-01
1.23247229e-01 3.77548724e-01 6.05620265e-01 -6.52749419e-01
8.68054330e-01 -9.64796484e-01 3.57957445e-02 1.39759079e-01
-3.73231441e-01 -1.83575541e-01 5.67932725e-01 -6.05982721e-01
-8.45278025e-01 -3.95135701e-01 1.16980389e-01 5.41194603e-02
-4.32070374e-01 5.27820230e-01 2.92235821e-01 4.79103833e-01
1.04315305e+00 -2.68817633e-01 -5.45361400e-01 -2.80112624e-01
3.13563704e-01 1.16719544e+00 4.54685628e-01 -7.23610640e-01
5.88580608e-01 4.18636739e-01 -4.45791215e-01 -5.99357545e-01
-8.95983458e-01 -6.28466308e-01 -7.97502458e-01 -1.88777849e-01
1.15179551e+00 -6.21438324e-01 -4.61670607e-01 1.19575709e-01
-8.47029448e-01 -3.70569825e-01 -2.37489387e-01 5.80705643e-01
-4.24894303e-01 2.17187002e-01 -4.24497813e-01 -6.60845816e-01
-2.42047489e-01 -6.74364507e-01 8.17318261e-01 -1.14626937e-01
-8.31142366e-01 -1.11246800e+00 4.14709151e-01 2.92217344e-01
1.32480472e-01 5.73529780e-01 1.39605093e+00 -1.51512325e+00
8.97969604e-01 -4.70586181e-01 -7.62066841e-02 -6.90245479e-02
1.05182633e-01 -3.90765034e-02 -6.49485588e-01 -2.83579648e-01
-1.03304267e-01 -5.51114500e-01 7.07956314e-01 2.88882982e-02
9.48545218e-01 -1.66964214e-02 -4.57450926e-01 -9.73663181e-02
1.18712819e+00 5.51606596e-01 1.39502198e-01 8.17631483e-01
1.58782586e-01 8.25849950e-01 5.85499108e-01 6.68490112e-01
1.83263928e-01 7.03996062e-01 1.80221304e-01 5.67932725e-01
4.57801878e-01 -2.66415954e-01 3.14542800e-01 2.06146061e-01
2.19540596e-01 -4.09449875e-01 -1.48962712e+00 7.85056353e-01
-1.74635911e+00 -1.11063075e+00 -5.49782440e-02 1.85288262e+00
6.06112897e-01 3.82718027e-01 2.01065212e-01 5.38321078e-01
8.47607076e-01 4.44839329e-01 -2.28762940e-01 -9.46379364e-01
1.50750861e-01 2.52334952e-01 4.53725070e-01 1.43998116e-01
-1.43205237e+00 8.06149542e-01 6.32822943e+00 9.11432624e-01
-6.85692728e-01 3.21090430e-01 3.84174168e-01 -1.28554245e-02
3.06457669e-01 2.52975792e-01 -7.01728821e-01 6.84138834e-01
1.33868468e+00 -4.22890514e-01 1.85551066e-02 7.73861647e-01
-6.41868636e-02 -2.51071334e-01 -9.53594327e-01 5.36575079e-01
3.73265624e-01 -1.20677173e+00 -1.46798402e-01 1.88983232e-01
3.55880320e-01 -5.03379339e-03 -4.62248355e-01 5.15386999e-01
8.42215896e-01 -6.37706339e-01 1.10931003e+00 6.33105189e-02
7.40779757e-01 -7.35153794e-01 8.75190198e-01 3.18120301e-01
-9.95670438e-01 -5.85014880e-01 -2.26586819e-01 -1.82133898e-01
1.85597897e-01 3.11620772e-01 -6.95797086e-01 4.88436103e-01
1.04174244e+00 2.30491772e-01 -5.60850263e-01 3.70602399e-01
-1.72097057e-01 7.32663214e-01 -1.41972631e-01 -8.18593130e-02
3.56637448e-01 7.86606520e-02 4.83480304e-01 1.30422449e+00
-5.22655994e-02 5.49290106e-02 6.08308196e-01 2.53956109e-01
8.75886381e-02 2.10375283e-02 -4.60588753e-01 -3.60440731e-01
6.03874207e-01 1.21785545e+00 -9.69257951e-01 -2.35669032e-01
-3.51087227e-02 4.88160729e-01 4.47770417e-01 -4.75078560e-02
-7.75747001e-01 -9.13261533e-01 5.37364185e-01 1.64407551e-01
1.77804455e-01 -2.40891799e-01 -1.15302496e-01 -8.50515366e-01
-4.56971049e-01 -3.34549785e-01 1.01186633e+00 -6.73548460e-01
-1.58249640e+00 7.28534937e-01 4.87235874e-01 -7.85246789e-01
-3.27179074e-01 -7.88052917e-01 -6.32888794e-01 5.71875155e-01
-6.46490037e-01 -9.36787665e-01 -5.62076569e-02 3.21708739e-01
5.20631492e-01 -3.99001598e-01 1.02135086e+00 1.37439787e-01
-5.89663923e-01 2.92734027e-01 3.01731825e-01 7.29811728e-01
9.88977313e-01 -1.27723289e+00 8.22552070e-02 3.65551621e-01
-3.21757980e-02 2.25252926e-01 5.95316708e-01 -5.53887904e-01
-7.70372093e-01 -1.02156556e+00 1.14522910e+00 -6.30976617e-01
1.45378494e+00 -6.19886100e-01 -5.23069382e-01 8.37149203e-01
1.32175595e-01 -1.92397833e-01 1.09240842e+00 4.58574742e-01
-5.71838260e-01 3.00610751e-01 -1.18327737e+00 4.43804741e-01
7.62128949e-01 -6.12748027e-01 -1.39023840e+00 6.01781666e-01
3.17580223e-01 8.14111233e-02 -1.07194054e+00 -1.99769869e-01
3.51389378e-01 -4.89204228e-01 7.68754661e-01 -1.30433571e+00
5.50996542e-01 -7.23625591e-04 -6.13863647e-01 -1.39018989e+00
-7.43729532e-01 -2.11294089e-02 3.77270162e-01 1.53363049e+00
1.61562592e-01 -7.72743523e-01 2.13382021e-01 4.02181625e-01
-1.81942284e-01 -3.79697025e-01 -1.17827487e+00 -6.03338897e-01
5.00266135e-01 -4.09075588e-01 2.99625039e-01 1.31766689e+00
4.88663971e-01 3.36252540e-01 1.43509179e-01 -1.22483142e-01
4.25751239e-01 -5.12990728e-02 4.72233355e-01 -1.58204257e+00
8.44774097e-02 -5.73152900e-01 -7.77212024e-01 7.93634206e-02
5.38601935e-01 -1.02058876e+00 -8.60718116e-02 -1.30013800e+00
2.43721887e-01 -5.22958338e-01 -3.59559000e-01 7.21054137e-01
1.20196387e-01 2.72460617e-02 -2.99943401e-03 3.45738441e-01
-2.94000417e-01 2.19107598e-01 1.07770171e-02 -9.47264507e-02
6.89532086e-02 -4.50822145e-01 -7.86534548e-01 9.40887868e-01
8.39027405e-01 -8.85263681e-01 3.59419994e-02 -2.43302181e-01
2.28704244e-01 3.22702304e-02 1.72451839e-01 -8.03178370e-01
1.29116967e-01 -3.95911545e-01 4.67966080e-01 -3.57169986e-01
1.10782824e-01 -5.62250495e-01 1.23841323e-01 2.90167302e-01
-7.94120193e-01 1.70497343e-01 1.35341257e-01 7.72706032e-01
-1.59701586e-01 -5.97599626e-01 6.39124513e-01 1.59910336e-01
-6.65300667e-01 -6.17922470e-02 -8.55544806e-01 4.21472251e-01
1.22468877e+00 -6.26748949e-02 -5.02530217e-01 -7.50611499e-02
-9.79786932e-01 1.53437644e-01 1.72033712e-01 6.72564864e-01
-1.07250199e-01 -1.24496889e+00 -7.53635049e-01 -2.91393816e-01
4.86998528e-01 -6.53687239e-01 7.93265477e-02 3.46883893e-01
-2.85355449e-01 3.34434509e-01 -3.96978587e-01 -6.60636788e-03
-9.68107879e-01 5.25791407e-01 1.18245117e-01 -5.06663322e-01
-4.30928528e-01 4.29520875e-01 -2.33955942e-02 -5.30227780e-01
7.13602547e-03 1.04662344e-01 -5.38211882e-01 9.82883513e-01
6.67272806e-01 4.45474595e-01 1.06657349e-01 -8.37297857e-01
-6.87843919e-01 1.74550429e-01 -1.33154720e-01 -2.76445508e-01
1.30551779e+00 2.14099884e-01 4.70420681e-02 7.86607623e-01
1.02656806e+00 -1.98963404e-01 -7.84850180e-01 -1.28949791e-01
3.43789518e-01 -2.01641932e-01 1.20319642e-01 -8.68477106e-01
-3.03909659e-01 7.42697954e-01 3.99118721e-01 6.60778165e-01
5.27126431e-01 4.37226295e-02 2.45578423e-01 2.52263963e-01
4.46217209e-01 -1.54805613e+00 2.27544340e-03 6.33286238e-01
9.46058214e-01 -8.06316912e-01 5.13413399e-02 7.11257309e-02
-7.06129074e-01 1.13702893e+00 2.37570211e-01 -2.77365178e-01
8.32053423e-01 5.36818504e-02 -2.69604087e-01 -3.88648629e-01
-7.30792403e-01 -5.71719594e-02 -1.87729537e-01 4.08792645e-01
2.06940025e-01 2.32271045e-01 -5.63029587e-01 7.98202097e-01
-3.89157087e-01 -4.52376306e-01 4.95195329e-01 1.07061076e+00
-4.89371359e-01 -8.16949964e-01 -4.67789292e-01 8.14424813e-01
-7.68905342e-01 1.46455660e-01 -7.19793856e-01 9.09452200e-01
2.89460450e-01 1.11456168e+00 5.80476642e-01 -4.66817409e-01
3.66554111e-01 4.72144186e-01 -8.52949247e-02 -7.48112082e-01
-9.09669995e-01 -5.14113963e-01 4.61837918e-01 2.92291753e-02
-5.80411553e-02 -9.94028986e-01 -1.28216243e+00 -3.74697000e-01
-3.53478640e-01 6.00181401e-01 7.37497449e-01 1.06272662e+00
3.88647132e-02 2.22474098e-01 6.34748161e-01 -9.22599375e-01
-5.59104085e-01 -1.36307347e+00 -7.73143411e-01 6.85547233e-01
-2.57478952e-01 -1.09217453e+00 -7.53925264e-01 -1.64966565e-02]
|
[10.212337493896484, 8.850110054016113]
|
fbcb6291-1cd4-4735-844e-513d1972bbad
|
a-convolutional-neural-network-for-language
|
1904.00805
| null |
http://arxiv.org/abs/1904.00805v1
|
http://arxiv.org/pdf/1904.00805v1.pdf
|
A Convolutional Neural Network for Language-Agnostic Source Code Summarization
|
Descriptive comments play a crucial role in the software engineering process.
They decrease development time, enable better bug detection, and facilitate the
reuse of previously written code. However, comments are commonly the last of a
software developer's priorities and are thus either insufficient or missing
entirely. Automatic source code summarization may therefore have the ability to
significantly improve the software development process. We introduce a novel
encoder-decoder model that summarizes source code, effectively writing a
comment to describe the code's functionality. We make two primary innovations
beyond current source code summarization models. First, our encoder is fully
language-agnostic and requires no complex input preprocessing. Second, our
decoder has an open vocabulary, enabling it to predict any word, even ones not
seen in training. We demonstrate results comparable to state-of-the-art methods
on a single-language data set and provide the first results on a data set
consisting of multiple programming languages.
|
['David Slater', 'Jessica Moore', 'Ben Gelman']
|
2019-03-29
| null | null | null | null |
['code-summarization']
|
['computer-code']
|
[ 1.35941803e-01 2.00482458e-01 -5.54850221e-01 -2.49217555e-01
-9.32292998e-01 -5.87675273e-01 2.00063333e-01 8.55140209e-01
-1.41832635e-01 4.87917662e-01 5.56741655e-01 -5.25701046e-01
2.55882710e-01 -4.37742054e-01 -7.19950974e-01 1.83675557e-01
1.56863138e-01 -4.44617048e-02 2.50931889e-01 -2.88925976e-01
5.87670863e-01 -3.15874785e-01 -1.53295374e+00 4.98753726e-01
9.44110155e-01 1.62879318e-01 4.41670477e-01 8.03846419e-01
-5.40084541e-01 1.26372397e+00 -8.27800632e-01 -5.84148943e-01
-3.97079796e-01 -4.83866394e-01 -8.51636291e-01 -2.34536547e-02
4.02493477e-01 -2.10177064e-01 1.17505774e-01 1.13258505e+00
2.54454017e-01 -4.07340139e-01 2.95117319e-01 -8.51412356e-01
-9.02467787e-01 9.62205827e-01 -6.92081690e-01 1.16907403e-01
6.18260741e-01 -2.11286411e-01 1.29716420e+00 -7.40998268e-01
8.40383589e-01 5.92030525e-01 7.48740196e-01 7.33717322e-01
-1.32656705e+00 -2.81029761e-01 3.98517884e-02 -1.43890128e-01
-9.47020650e-01 -5.90126574e-01 6.15865111e-01 -9.16055799e-01
1.56845343e+00 2.49425173e-02 6.06224120e-01 8.09773922e-01
5.64678073e-01 6.21513844e-01 2.97761858e-01 -6.72587156e-01
3.10820267e-02 1.66285649e-01 4.45521235e-01 9.56328869e-01
5.51168442e-01 -6.50233805e-01 -3.76612782e-01 -3.52554858e-01
-3.17456259e-04 1.54857049e-02 -4.00530964e-01 -1.54252604e-01
-9.98454690e-01 8.17543745e-01 -1.66178197e-01 3.80982339e-01
-4.95486856e-02 5.99326968e-01 6.83584929e-01 5.88657081e-01
5.64926386e-01 7.79786408e-01 -5.47395408e-01 -8.57748032e-01
-1.13450277e+00 2.53986210e-01 1.15556800e+00 1.10746348e+00
8.40110183e-01 -3.87390424e-03 8.41835961e-02 8.79127383e-01
3.93512756e-01 2.26305872e-01 5.12901187e-01 -6.65033877e-01
6.46245480e-01 1.04871368e+00 -3.39139327e-02 -7.24109173e-01
-8.17346349e-02 -4.99753892e-01 -3.04675728e-01 2.94413596e-01
-1.22966662e-01 -6.20109774e-02 -4.05693978e-01 1.37470353e+00
-3.11706692e-01 -6.07714653e-01 -9.93833691e-02 9.86788869e-02
7.95681715e-01 5.87002575e-01 -2.96931922e-01 -1.98292777e-01
1.08952498e+00 -9.13335621e-01 -5.36598802e-01 -6.25680745e-01
9.40232813e-01 -1.05629265e+00 9.89218175e-01 4.28258032e-01
-1.02931893e+00 -3.40190291e-01 -1.21339452e+00 -2.47383043e-01
-6.52399957e-02 3.63224387e-01 6.36592746e-01 7.13035524e-01
-1.21175551e+00 5.29747486e-01 -8.24421346e-01 -2.46429294e-01
3.83416235e-01 1.45831620e-02 -4.79039073e-01 -4.70853448e-02
-4.21410352e-01 7.41798997e-01 1.47822410e-01 -4.59970891e-01
-6.46975219e-01 -8.99073303e-01 -1.09244072e+00 2.33268723e-01
4.33398575e-01 -5.14833748e-01 1.88265371e+00 -9.88398194e-01
-1.09137249e+00 6.35725379e-01 -3.56594026e-01 -2.28644952e-01
-1.06745549e-01 -3.45798016e-01 -3.03001851e-01 -3.95311475e-01
4.51736987e-01 3.35914135e-01 6.82098091e-01 -9.60419118e-01
-6.18894637e-01 1.70036808e-01 2.55387038e-01 -4.52872694e-01
-5.30221224e-01 3.51914078e-01 -4.94052798e-01 -7.00289011e-01
-5.04584968e-01 -6.75400734e-01 -2.41640463e-01 -1.91101193e-01
-3.26302826e-01 -5.79790771e-02 3.98292899e-01 -9.59458470e-01
2.08917880e+00 -2.21925879e+00 2.44925722e-01 -2.38881543e-01
5.21403074e-01 1.75851628e-01 -3.57391089e-02 8.99618208e-01
-1.63030639e-01 5.16378582e-01 -4.63407964e-01 -3.75367522e-01
1.78369731e-02 -2.04942137e-01 -2.74853587e-01 1.57588556e-01
3.69177252e-01 8.58706236e-01 -9.23482716e-01 -3.08516681e-01
-3.94568712e-01 9.54987258e-02 -1.02090311e+00 2.33911108e-02
-6.02960348e-01 3.48968431e-03 -3.24535280e-01 5.55971682e-01
1.03610791e-01 -3.41614962e-01 6.90587936e-03 4.59263712e-01
-4.49712425e-01 7.79686570e-01 -6.59905434e-01 1.87907064e+00
-8.84219110e-01 1.05077910e+00 -1.60207331e-01 -4.90945548e-01
8.30290258e-01 2.88497597e-01 1.15292393e-01 -4.56495315e-01
-1.70769140e-01 4.42372382e-01 2.75098205e-01 -6.64224744e-01
8.05308282e-01 1.61732674e-01 -3.76794219e-01 9.56530988e-01
-9.32309180e-02 -3.02535385e-01 7.07529068e-01 3.47433805e-01
1.52435529e+00 8.67900327e-02 8.26952755e-01 -7.84115195e-02
2.96735048e-01 2.07821518e-01 4.75707769e-01 5.10770738e-01
4.28886622e-01 4.74712431e-01 1.05993271e+00 -2.18226433e-01
-1.21378624e+00 -5.64552128e-01 8.76015574e-02 1.04947305e+00
-3.58422548e-01 -1.25291169e+00 -6.04716897e-01 -7.63743699e-01
-6.47134110e-02 9.36550140e-01 -4.87098247e-01 -2.04029873e-01
-5.58395326e-01 -1.85207486e-01 6.01837099e-01 5.76040328e-01
-1.62190095e-01 -7.70595968e-01 -8.17726433e-01 4.70521688e-01
1.34720383e-02 -5.82280576e-01 -6.62234068e-01 2.57090211e-01
-6.61030591e-01 -1.12608993e+00 -4.11992133e-01 -7.57266462e-01
7.82181501e-01 3.82212773e-02 1.50013638e+00 5.17066300e-01
-3.21938962e-01 2.55322933e-01 -6.43592894e-01 -4.11924034e-01
-1.11004066e+00 2.79526919e-01 -5.25646448e-01 -7.64522135e-01
3.81978720e-01 -4.28238600e-01 -4.33413461e-02 -3.18437904e-01
-8.25236857e-01 1.03240617e-01 6.70991838e-01 8.21121693e-01
3.56606208e-02 -2.28312276e-02 5.74851394e-01 -1.24147713e+00
9.43235159e-01 -6.22489989e-01 -5.10855079e-01 2.56594658e-01
-4.82038438e-01 3.31320256e-01 7.54342318e-01 -1.12578988e-01
-9.63317394e-01 -8.11500549e-02 -4.06084925e-01 3.39962423e-01
1.10130541e-01 1.12376893e+00 2.84430057e-01 2.95702890e-02
6.43527746e-01 1.40817121e-01 -8.58729482e-02 -6.74893737e-01
1.63552880e-01 8.74888718e-01 1.40687138e-01 -4.17890429e-01
5.36432087e-01 -3.81903723e-02 -5.80356836e-01 -7.90995359e-01
-3.85560721e-01 -5.48359334e-01 -4.18190151e-01 6.58210600e-04
5.68089187e-01 -7.88726568e-01 -1.39641687e-01 1.62156403e-01
-1.61639595e+00 -1.47806630e-01 -1.83606565e-01 1.50823161e-01
-1.32994682e-01 5.14961958e-01 -5.21289229e-01 -4.03352857e-01
-2.76521295e-01 -1.32700026e+00 9.75598693e-01 -2.51648128e-02
-8.87817085e-01 -9.53397691e-01 3.94777685e-01 -8.46124999e-03
3.70797545e-01 3.75077166e-02 1.36607075e+00 -4.51411664e-01
-4.69448894e-01 -2.93192774e-01 3.99492756e-02 5.04138172e-01
3.69485468e-01 4.75453347e-01 -5.75509846e-01 -2.67752498e-01
-9.89823490e-02 -2.07424775e-01 8.04492474e-01 -5.88484257e-02
9.25517380e-01 -2.53249019e-01 -2.79258251e-01 1.83327734e-01
1.39634287e+00 1.33963868e-01 4.23621595e-01 2.55258471e-01
6.18254781e-01 4.50954348e-01 2.79370546e-01 6.35927081e-01
4.97644365e-01 4.00386214e-01 4.27585542e-01 2.65532583e-01
-1.76897287e-01 -2.35789910e-01 7.43047357e-01 1.42212057e+00
3.31946105e-01 -2.87336856e-01 -1.20442033e+00 9.69836414e-01
-1.54211748e+00 -7.62203097e-01 -5.57720304e-01 2.04664707e+00
1.18515909e+00 2.96490580e-01 -1.86645001e-01 8.45605358e-02
2.57249117e-01 9.73343700e-02 -2.64810175e-01 -7.82651961e-01
4.85556126e-01 3.71513486e-01 4.27577376e-01 4.07648087e-01
-8.07144582e-01 5.53013742e-01 6.47649908e+00 3.95815551e-01
-9.85446393e-01 2.46848941e-01 -1.02962002e-01 -2.81220432e-02
-8.91779661e-01 2.72284210e-01 -7.35391557e-01 4.84625340e-01
9.91237223e-01 -7.11276948e-01 2.51602560e-01 1.24486029e+00
-1.27174780e-01 -3.74925673e-01 -1.64437079e+00 6.24019325e-01
3.40271533e-01 -1.58994639e+00 -1.63684726e-01 -1.44413367e-01
7.48730361e-01 5.36858775e-02 -2.83102542e-01 2.11888671e-01
2.91082084e-01 -9.01433945e-01 9.12945271e-01 3.33573967e-01
9.71080840e-01 -6.16401851e-01 7.03766167e-01 5.17387688e-01
-9.46291745e-01 -6.99386513e-03 -4.70314801e-01 -2.61722535e-01
8.53245184e-02 7.25193620e-01 -8.21366608e-01 3.30869079e-01
4.35316682e-01 1.23474932e+00 -9.92534280e-01 1.07141161e+00
-2.90733874e-01 6.62938356e-01 2.37504363e-01 -4.93266553e-01
-6.91976398e-02 4.08803552e-01 5.00324667e-01 1.48888731e+00
4.31581795e-01 -4.78757888e-01 3.52821872e-02 9.34881270e-01
-1.48051336e-01 1.34287253e-01 -7.17159748e-01 -6.32920146e-01
3.30141187e-01 7.87521303e-01 -5.49331367e-01 -1.18957900e-01
-9.79811668e-01 8.12883615e-01 3.16814035e-01 2.29300428e-02
-6.28741264e-01 -8.75982940e-01 7.80295789e-01 1.44807860e-01
4.73092139e-01 -3.05949599e-01 -3.82850230e-01 -1.29570830e+00
4.52345341e-01 -1.12136936e+00 -2.30237871e-01 -5.12399077e-01
-7.04246163e-01 5.53425074e-01 -3.07959199e-01 -1.29234684e+00
-4.13588732e-01 -5.41939378e-01 -5.72348654e-01 7.01771438e-01
-1.38328803e+00 -7.15495527e-01 -1.93338804e-02 -2.97508836e-01
8.15359652e-01 -3.77655059e-01 8.50944698e-01 5.57300687e-01
-3.21725637e-01 4.92998332e-01 1.07762277e-01 2.13672016e-02
6.31260753e-01 -1.41513765e+00 9.45936739e-01 1.27576888e+00
1.51101679e-01 1.38568604e+00 7.01982141e-01 -8.73884082e-01
-1.56613243e+00 -1.01992667e+00 1.26485395e+00 -6.94611490e-01
8.61709654e-01 -4.29923981e-01 -9.07120168e-01 6.95210278e-01
4.52379674e-01 -6.02498829e-01 1.05151117e+00 2.70002574e-01
-5.36503255e-01 1.77213624e-01 -4.66587752e-01 4.84385431e-01
6.75406933e-01 -7.74378181e-01 -8.26768100e-01 1.17925346e-01
8.68460834e-01 -4.29085463e-01 -7.45796978e-01 -1.65703148e-01
4.37691033e-01 -9.34477687e-01 3.63943160e-01 -3.42499584e-01
1.26536596e+00 -2.67777175e-01 1.37629375e-01 -1.46220720e+00
-2.52494186e-01 -7.05073953e-01 -1.52171165e-01 1.43148398e+00
7.45370328e-01 -4.57364649e-01 4.29385424e-01 2.57804483e-01
-4.96351838e-01 -6.27117217e-01 -4.61661279e-01 -5.16632557e-01
-5.59842177e-02 -7.43853748e-01 4.36819434e-01 7.17269480e-01
6.08923912e-01 4.34731603e-01 -2.19539702e-01 -2.53095359e-01
2.40455627e-01 6.12122230e-02 7.61625946e-01 -1.25918174e+00
-7.24303603e-01 -5.87746263e-01 -2.86079526e-01 -1.09347332e+00
3.17429721e-01 -1.11403060e+00 1.56376451e-01 -1.79569685e+00
5.14243484e-01 -1.47509575e-01 2.57845044e-01 7.87418962e-01
-7.76937529e-02 -1.47480458e-01 -1.05098113e-01 1.31548464e-01
-4.54785645e-01 1.50022537e-01 6.41186237e-01 -2.44379550e-01
2.17119567e-02 7.93282017e-02 -1.21616268e+00 6.42409801e-01
5.90488970e-01 -8.39644790e-01 -5.26986539e-01 -8.36923778e-01
1.02918291e+00 5.93859516e-02 2.98761297e-02 -7.76143670e-01
3.71242821e-01 4.19164486e-02 -3.01319033e-01 -2.65897334e-01
-3.59997183e-01 -5.23622334e-01 2.66955830e-02 6.29810095e-01
-3.86656731e-01 2.25420982e-01 3.78355920e-01 3.50814968e-01
-3.84297222e-01 -9.42404509e-01 4.62315351e-01 -9.42823961e-02
-8.73964012e-01 -1.96762860e-01 -7.25806892e-01 2.91666448e-01
8.30751896e-01 -1.87802583e-01 -5.03196299e-01 -1.13246284e-01
1.42644132e-02 6.00553714e-02 1.03051615e+00 7.27564752e-01
5.77756524e-01 -8.37606668e-01 -7.27741122e-01 2.56370008e-01
6.08971477e-01 -1.33990169e-01 -1.98951349e-01 5.21597862e-01
-8.98047686e-01 4.57494318e-01 -1.98252201e-02 -2.64606714e-01
-1.48002207e+00 3.14051837e-01 -5.55340983e-02 -1.25638276e-01
-6.97055757e-01 8.12026560e-01 -1.65672973e-01 -2.74272829e-01
1.20044136e-02 -7.53299832e-01 -9.48963463e-02 8.65387451e-03
7.83651352e-01 1.86996520e-01 3.25877726e-01 -1.93453580e-01
-5.03135622e-01 3.84040505e-01 -4.88190919e-01 8.10942799e-02
1.59096813e+00 1.81037694e-01 -7.05703080e-01 7.83932388e-01
1.20729184e+00 5.79212189e-01 -8.52260172e-01 5.31358570e-02
5.51099181e-01 -5.00818431e-01 -1.66967288e-02 -6.98435724e-01
-5.51901698e-01 8.76589835e-01 -1.89339176e-01 2.89913684e-01
8.63139749e-01 2.42285013e-01 6.77540958e-01 5.39773881e-01
4.40660596e-01 -8.62806380e-01 1.94602415e-01 7.47723401e-01
6.93961918e-01 -1.11625040e+00 8.54487792e-02 -2.85171807e-01
-4.63578850e-01 1.43561590e+00 5.02241015e-01 1.93342939e-01
3.18470865e-01 8.49640727e-01 -8.37805942e-02 -2.43011758e-01
-1.23168778e+00 1.13543056e-01 1.64650008e-01 3.34522605e-01
1.25657535e+00 -2.57946163e-01 -3.94455522e-01 7.56808698e-01
-1.23103596e-01 1.10061534e-01 1.33493042e+00 1.30927026e+00
-5.94810963e-01 -1.51697075e+00 6.48687854e-02 7.21261978e-01
-7.27322042e-01 -5.52888572e-01 -4.56076384e-01 5.18398166e-01
-5.71371103e-03 9.18219984e-01 -2.09958732e-01 -2.92715579e-01
3.23750168e-01 1.20191544e-01 4.03517604e-01 -1.64813006e+00
-7.93488801e-01 -2.65948564e-01 2.58913755e-01 -4.33295012e-01
-9.47972462e-02 -7.58810043e-01 -1.13506377e+00 -3.22490454e-01
-3.42785805e-01 2.77929246e-01 7.90721893e-01 7.98799753e-01
5.97634971e-01 9.84915495e-01 2.33533159e-01 -2.94251114e-01
-4.61210161e-01 -7.11074114e-01 -1.51338458e-01 -1.03116289e-01
6.67891383e-01 -2.84786135e-01 -2.09395662e-02 5.80361128e-01]
|
[7.630922317504883, 7.928878307342529]
|
0203c7ba-961f-44c1-8aa0-9396d2875e5f
|
fuzzy-attention-neural-network-to-tackle
|
2209.02048
| null |
https://arxiv.org/abs/2209.02048v2
|
https://arxiv.org/pdf/2209.02048v2.pdf
|
Fuzzy Attention Neural Network to Tackle Discontinuity in Airway Segmentation
|
Airway segmentation is crucial for the examination, diagnosis, and prognosis of lung diseases, while its manual delineation is unduly burdensome. To alleviate this time-consuming and potentially subjective manual procedure, researchers have proposed methods to automatically segment airways from computerized tomography (CT) images. However, some small-sized airway branches (e.g., bronchus and terminal bronchioles) significantly aggravate the difficulty of automatic segmentation by machine learning models. In particular, the variance of voxel values and the severe data imbalance in airway branches make the computational module prone to discontinuous and false-negative predictions. especially for cohorts with different lung diseases. Attention mechanism has shown the capacity to segment complex structures, while fuzzy logic can reduce the uncertainty in feature representations. Therefore, the integration of deep attention networks and fuzzy theory, given by the fuzzy attention layer, should be an escalated solution for better generalization and robustness. This paper presents an efficient method for airway segmentation, comprising a novel fuzzy attention neural network and a comprehensive loss function to enhance the spatial continuity of airway segmentation. The deep fuzzy set is formulated by a set of voxels in the feature map and a learnable Gaussian membership function. Different from the existing attention mechanism, the proposed channel-specific fuzzy attention addresses the issue of heterogeneous features in different channels. Furthermore, a novel evaluation metric is proposed to assess both the continuity and completeness of airway structures. The efficiency, generalization and robustness of the proposed method have been proved by training on normal lung disease while testing on datasets of lung cancer, COVID-19 and pulmonary fibrosis.
|
['Guang Yang', 'Simon Walsh', 'Witold Pedrycz', 'Francisco Herrera', 'Yingying Fang', 'Xiaodan Xing', 'Peng Tang', 'Zeyu Tang', 'Javier Del Ser', 'Yang Nan']
|
2022-09-05
| null | null | null | null |
['deep-attention', 'deep-attention']
|
['computer-vision', 'natural-language-processing']
|
[ 5.42085506e-02 -9.05105099e-02 -1.94197576e-02 -2.53381312e-01
-4.13021743e-01 -4.02800381e-01 -1.04280189e-01 -3.79656442e-02
-4.69267875e-01 5.73564172e-01 -1.35305926e-01 -1.96764573e-01
-4.92521942e-01 -8.15388262e-01 -3.27777982e-01 -8.68167937e-01
2.20859602e-01 6.63163960e-01 3.47485721e-01 7.27807656e-02
2.09976420e-01 5.92462063e-01 -1.39475811e+00 2.94694364e-01
1.30652988e+00 1.18017864e+00 4.86458987e-01 3.80328387e-01
-6.04898810e-01 1.88594371e-01 -4.52230752e-01 -2.13135928e-01
1.11913100e-01 -1.36124343e-01 -8.13088357e-01 1.01913743e-01
2.54005283e-01 -1.79571718e-01 7.36670792e-02 1.25313807e+00
5.34850180e-01 1.47169590e-01 9.20462370e-01 -9.37792182e-01
-6.66736364e-01 3.28608602e-01 -3.71886432e-01 6.71016634e-01
-3.10592383e-01 3.21274489e-01 8.44724715e-01 -7.39316761e-01
1.39290206e-02 9.35637832e-01 7.95943320e-01 5.37158728e-01
-7.71698415e-01 -5.62921405e-01 6.03788234e-02 1.38887107e-01
-1.41542339e+00 2.60363579e-01 5.43368578e-01 -6.21607184e-01
6.22833550e-01 4.00771171e-01 6.72863185e-01 5.09887040e-01
1.54602408e-01 2.83815563e-01 8.87163877e-01 -3.79854143e-02
-6.44057915e-02 4.08978313e-01 2.87655536e-02 7.17492461e-01
4.93526965e-01 -3.85560514e-03 6.83435082e-01 1.19082727e-01
7.73032784e-01 3.72733235e-01 -4.81105208e-01 3.65941189e-02
-9.46997762e-01 7.72211790e-01 9.51241553e-01 6.75803244e-01
-4.26026553e-01 -4.49285842e-02 3.16357911e-01 -3.53478789e-01
1.13116346e-01 4.96398866e-01 -2.93883353e-01 3.04059684e-01
-8.77783597e-01 -8.97220597e-02 3.03517103e-01 5.16758740e-01
3.68985593e-01 -7.53792096e-03 -3.65996450e-01 7.28650212e-01
5.22740841e-01 3.81054401e-01 8.89494181e-01 -7.17213333e-01
4.89407003e-01 8.69160473e-01 -3.22868675e-01 -1.02225602e+00
-5.53242385e-01 -4.46778119e-01 -1.16250694e+00 1.07130352e-02
2.05156535e-01 -3.51885892e-02 -1.11335623e+00 1.36070275e+00
5.56004345e-01 1.46554396e-01 -4.11234617e-01 1.13421762e+00
9.25154269e-01 4.18150157e-01 2.91696042e-01 -1.80655450e-01
1.33110023e+00 -8.11844110e-01 -7.22223520e-01 2.03839436e-01
2.52538443e-01 -3.86339009e-01 1.17268133e+00 8.69177952e-02
-1.09207058e+00 -7.94689775e-01 -8.26145411e-01 -5.22087105e-02
-4.24746752e-01 -8.38204548e-02 4.42194372e-01 6.81796253e-01
-7.48696208e-01 5.86861074e-01 -7.91373551e-01 -8.12925622e-02
9.67046082e-01 7.59784698e-01 1.24647692e-02 2.40854755e-01
-1.32213831e+00 7.30643094e-01 4.33963209e-01 3.04485768e-01
-2.05163166e-01 -6.45495772e-01 -4.50265378e-01 3.94853860e-01
1.03853941e-01 -9.57197726e-01 9.13313925e-01 -8.54737878e-01
-1.11071241e+00 5.30697286e-01 1.37214854e-01 -3.51251632e-01
5.07832348e-01 4.54322286e-02 -2.16482192e-01 5.13400137e-01
-1.23909093e-01 7.31044829e-01 7.24493682e-01 -1.11044157e+00
-8.24052751e-01 -4.87394243e-01 -2.22433716e-01 2.78789192e-01
-5.38435936e-01 -3.47745508e-01 -1.31921202e-01 -6.61996424e-01
1.34826258e-01 -6.77790999e-01 -4.21274304e-01 9.18354690e-02
-2.77559549e-01 -3.59908998e-01 7.82851458e-01 -6.62328124e-01
1.50426948e+00 -2.00699306e+00 -2.04094984e-02 4.09899026e-01
2.91837722e-01 5.02986133e-01 3.72034311e-01 -4.63473767e-01
6.13979474e-02 8.23955536e-01 -6.97802126e-01 2.48537958e-01
-2.03510493e-01 2.98833311e-01 1.70690045e-01 2.58070290e-01
3.36496532e-01 8.16039264e-01 -6.31425977e-01 -1.02806354e+00
4.17506814e-01 6.92385256e-01 -6.53515637e-01 1.19032972e-01
-1.93125799e-01 6.40332699e-01 -7.12155461e-01 6.25090122e-01
6.89808905e-01 -3.95346224e-01 -3.45327139e-01 -3.21584612e-01
-1.05523333e-01 4.97740731e-02 -9.90706980e-01 1.15866852e+00
-5.04764140e-01 9.83420089e-02 2.12918282e-01 -8.58070135e-01
5.93966782e-01 6.13272130e-01 7.48004973e-01 -3.99384081e-01
5.07513285e-01 2.88683563e-01 4.02745157e-01 -1.04008925e+00
-9.59144980e-02 -5.42840660e-01 3.29100549e-01 2.06817746e-01
-2.88059831e-01 -4.44075167e-01 -1.61480755e-04 -3.69986236e-01
5.73336482e-01 -3.28264266e-01 7.93580487e-02 -2.13888884e-01
9.11603808e-01 -1.92452576e-02 3.89724284e-01 2.27361843e-01
-5.21749139e-01 8.19612563e-01 7.30429962e-02 -1.91311270e-01
-7.75563538e-01 -1.05338812e+00 -6.58164322e-01 4.84987497e-01
8.10550898e-02 3.42456400e-01 -8.92759383e-01 -9.90048468e-01
8.39798227e-02 5.08298516e-01 -4.65506375e-01 -1.69943064e-01
-6.76118493e-01 -7.70471573e-01 4.20672417e-01 6.57595336e-01
6.36696517e-01 -1.21175277e+00 -5.26157379e-01 9.72093269e-02
-3.24516982e-01 -6.95559740e-01 -6.40519083e-01 6.51363358e-02
-1.14069426e+00 -1.12582147e+00 -6.73309743e-01 -9.97884154e-01
8.67645681e-01 1.48456469e-01 7.42424011e-01 5.42963386e-01
-5.70335329e-01 1.69616520e-01 -2.89379116e-02 -3.44712853e-01
-1.98144808e-01 1.22258134e-01 -2.58144587e-01 -1.01866640e-01
1.29874960e-01 -4.13374454e-01 -8.52743328e-01 3.16691697e-01
-1.01959872e+00 -3.76463056e-01 6.20608985e-01 1.07129455e+00
9.82692480e-01 6.12073779e-01 7.32627928e-01 -7.58472145e-01
6.56898618e-01 -6.87793434e-01 -3.97974551e-01 7.68046007e-02
-8.17524731e-01 -1.13111205e-01 8.77947211e-01 -2.74235308e-01
-1.11418605e+00 -1.69235840e-01 -3.97066683e-01 -6.42054856e-01
-2.37488225e-01 2.70744175e-01 -1.08717047e-01 -1.55613758e-02
4.72991645e-01 3.13451933e-03 1.96592391e-01 -2.09445685e-01
1.68334201e-01 9.18929458e-01 3.30744565e-01 -1.85339853e-01
6.76118314e-01 3.84017408e-01 1.05921045e-01 -5.29970884e-01
-5.63246906e-01 -5.09482205e-01 -7.69912720e-01 -1.38763580e-02
1.21918964e+00 -3.67508352e-01 -6.33324742e-01 1.03013568e-01
-8.71279180e-01 9.67882946e-02 -4.99947131e-01 6.38508677e-01
-3.03886533e-01 5.52595317e-01 -5.35527885e-01 -5.61002910e-01
-8.29035878e-01 -1.28966498e+00 6.04000449e-01 4.58400309e-01
-1.05214417e-01 -1.10665035e+00 -1.22186184e-01 6.30573452e-01
4.11694765e-01 1.01182148e-01 1.35307920e+00 -7.56466031e-01
-4.78530556e-01 -1.39569148e-01 -4.23964381e-01 8.10309231e-01
3.53037924e-01 2.61337578e-01 -7.88015544e-01 4.41687778e-02
4.17738318e-01 -8.00936595e-02 9.57511365e-01 6.08905077e-01
1.65696216e+00 -3.35744470e-01 -3.99746269e-01 7.23527074e-01
1.29410553e+00 5.64489543e-01 4.04068589e-01 -1.32926665e-02
8.67857456e-01 7.31911480e-01 5.72618842e-01 1.52489170e-01
1.95239618e-01 1.12149745e-01 7.42714822e-01 -2.59201288e-01
-2.17796221e-01 2.85282046e-01 -2.99288064e-01 8.98904324e-01
-2.45698482e-01 -3.05604488e-01 -8.03956807e-01 5.74195147e-01
-1.26884389e+00 -1.02392232e+00 -3.42243105e-01 2.11781740e+00
7.77500451e-01 1.03047289e-01 -1.55285478e-01 2.18026683e-01
1.02365589e+00 -2.50907272e-01 -5.47103226e-01 -4.39474553e-01
1.74663574e-01 5.37173331e-01 2.07888469e-01 2.85099417e-01
-1.15154409e+00 3.95322919e-01 5.57523441e+00 8.55231583e-01
-1.44970584e+00 9.84050408e-02 6.60857975e-01 -7.31067657e-02
-4.75788474e-01 -5.52705824e-01 -6.68696404e-01 9.31344211e-01
5.63573241e-01 2.88923025e-01 3.05399269e-01 5.50183058e-01
1.85610503e-01 2.45252863e-01 -5.74143827e-01 7.05275118e-01
-2.18894333e-01 -9.84913647e-01 1.01757899e-01 -3.34643945e-02
6.60325706e-01 1.09402671e-01 3.63369197e-01 2.54507512e-01
-3.47897619e-01 -1.27186513e+00 1.48993671e-01 4.72507656e-01
9.05182302e-01 -6.12311900e-01 9.88375127e-01 3.09401274e-01
-1.14667773e+00 -2.82982081e-01 -3.54551136e-01 4.99316841e-01
8.92121717e-03 3.84525746e-01 -1.15832031e+00 4.43149507e-01
7.53174305e-01 2.00940877e-01 -5.22586107e-01 1.19872892e+00
1.57431047e-02 5.85593760e-01 -5.81890702e-01 -9.43699479e-02
2.82750309e-01 -3.58624071e-01 4.73185688e-01 9.67706680e-01
5.41373074e-01 1.98267549e-01 5.85751012e-02 1.13292301e+00
-9.20010209e-02 4.91959482e-01 -3.91018569e-01 6.42730370e-02
3.78051311e-01 1.48240721e+00 -8.37349296e-01 -2.61463493e-01
-4.61988896e-01 4.55067396e-01 1.66100696e-01 8.77117813e-02
-1.09854686e+00 -2.14486971e-01 2.28929996e-01 5.63070774e-01
3.77202660e-01 3.66842270e-01 -6.60687864e-01 -6.43536031e-01
-1.37458891e-01 -4.78161424e-01 5.58696210e-01 -3.93110186e-01
-1.29266131e+00 7.95745552e-01 -1.63825423e-01 -1.17387760e+00
1.26437828e-01 -5.20851254e-01 -8.77553582e-01 8.69779766e-01
-1.53253198e+00 -8.35902035e-01 -5.16896486e-01 5.65146267e-01
3.80221456e-01 1.78872511e-01 4.35707003e-01 6.85405076e-01
-7.75686204e-01 4.31824714e-01 -7.11674392e-02 -9.48843509e-02
4.32731867e-01 -1.46694398e+00 -1.98183388e-01 3.45117509e-01
-2.28435665e-01 5.09864688e-01 1.72945708e-01 -6.73621356e-01
-5.62070489e-01 -1.40080953e+00 5.63827634e-01 -1.53137699e-01
2.27252111e-01 3.60548526e-01 -1.28431940e+00 3.44267249e-01
-1.31691871e-02 2.90027767e-01 6.44634664e-01 -4.95106250e-01
2.93433607e-01 -6.13112599e-02 -1.52474058e+00 4.75035787e-01
7.61040568e-01 -5.14622107e-02 -6.86644077e-01 3.88894796e-01
6.17208123e-01 -2.34635279e-01 -1.16084433e+00 8.65954578e-01
3.80472183e-01 -8.52576375e-01 1.06731248e+00 -4.26266760e-01
1.59318328e-01 -4.74359512e-01 1.97671667e-01 -8.69719088e-01
-6.32195175e-01 1.21855915e-01 1.45471931e-01 1.09545839e+00
3.28195304e-01 -6.86287940e-01 7.70897746e-01 5.25967240e-01
-5.04316866e-01 -1.17820716e+00 -9.87966716e-01 -3.80202174e-01
3.71590227e-01 -1.16854981e-01 5.65661967e-01 6.64639711e-01
-4.69126314e-01 1.22374482e-01 4.09079671e-01 3.30496840e-02
2.37584814e-01 6.76749423e-02 5.22166528e-02 -1.43132961e+00
-9.34829190e-02 -1.03186798e+00 -1.81212872e-01 -5.71082115e-01
-7.08859637e-02 -1.25050855e+00 3.43307257e-02 -1.79038084e+00
3.33395563e-02 -6.62661254e-01 -5.87657392e-01 1.94151878e-01
-5.06376207e-01 7.65572488e-02 -1.10393465e-01 2.69028157e-01
-2.12486297e-01 4.94189411e-01 2.03941131e+00 -3.11831683e-01
-2.46218830e-01 5.18914282e-01 -6.00543737e-01 8.55983853e-01
9.56886232e-01 -4.61422622e-01 -5.94424307e-01 -9.90473777e-02
-1.19169876e-01 -9.43143815e-02 3.16676289e-01 -9.57766175e-01
3.06935370e-01 -3.57768834e-02 5.12974322e-01 -7.26234853e-01
9.48231667e-04 -1.29935837e+00 6.21282086e-02 7.46147037e-01
-1.14436835e-01 -1.16966195e-01 1.14661857e-01 4.77395236e-01
-4.04914409e-01 -5.84970653e-01 1.02707553e+00 -4.23622191e-01
-2.43147492e-01 6.08465970e-01 -2.28363529e-01 2.83359855e-01
1.13247812e+00 -4.77576077e-01 5.12093827e-02 1.83949471e-01
-7.06248403e-01 3.92692655e-01 1.17318720e-01 9.10182744e-02
6.17248356e-01 -1.08904350e+00 -4.25689489e-01 3.45049322e-01
-2.83241719e-01 7.84818232e-01 5.85712969e-01 1.24110246e+00
-8.15523565e-01 4.77819949e-01 -1.05269365e-01 -7.76388347e-01
-1.10218799e+00 7.32744753e-01 7.09752858e-01 -4.19147164e-01
-3.30032676e-01 9.42065954e-01 4.30272281e-01 -3.17000866e-01
2.02046588e-01 -9.09339845e-01 -6.32515550e-01 3.68230939e-02
3.21420915e-02 5.78742087e-01 5.85886613e-02 -5.57639718e-01
-2.58617073e-01 9.62717354e-01 2.45990884e-02 3.79393548e-01
7.35911846e-01 -2.21254438e-01 -2.28194535e-01 1.10568395e-02
1.02467668e+00 -3.47018056e-02 -9.76768374e-01 1.61511049e-01
-2.60607034e-01 -8.67884755e-02 2.05978617e-01 -9.38380539e-01
-1.33994031e+00 1.30154514e+00 8.05748343e-01 4.01625991e-01
1.19850063e+00 -2.31224224e-01 1.20495856e+00 1.26011688e-02
-9.13782418e-02 -1.00446105e+00 -6.25806898e-02 1.73529848e-01
7.88890779e-01 -1.43056095e+00 -1.23380773e-01 -5.60816288e-01
-5.37967086e-01 1.15587199e+00 9.20763731e-01 1.05157062e-01
8.90252948e-01 1.05589680e-01 9.41002741e-02 -1.92112848e-01
-3.84787649e-01 -2.03713730e-01 5.16265452e-01 5.25886059e-01
3.31142038e-01 -3.37468833e-02 -3.51031393e-01 1.00340223e+00
3.80168296e-02 -6.11448362e-02 -1.54470969e-02 5.01975596e-01
-7.34807372e-01 -6.62494361e-01 -5.27018666e-01 7.66950727e-01
-8.97443235e-01 -1.60002887e-01 -3.35011892e-02 9.81908500e-01
7.16017365e-01 6.87282920e-01 3.63925070e-01 -1.77374020e-01
1.81562856e-01 -1.14700673e-02 3.07939470e-01 -6.83763444e-01
-1.06718290e+00 1.18178383e-01 -4.59595203e-01 -2.90553849e-02
-2.22040474e-01 -4.59372163e-01 -2.06997132e+00 1.32481515e-01
-6.85798287e-01 2.25580350e-01 4.13888931e-01 7.33769357e-01
2.82469004e-01 9.22690690e-01 7.41616428e-01 -2.74878323e-01
-8.10333669e-01 -9.79292572e-01 -4.27905321e-01 6.15266979e-01
2.46055409e-01 -7.80490637e-01 -6.88984215e-01 -2.15624422e-01]
|
[15.072725296020508, -2.1668169498443604]
|
590712e9-379f-473d-8908-9083c9b7ccd9
|
a-meta-learning-based-generalizable-indoor
|
2305.13453
| null |
https://arxiv.org/abs/2305.13453v2
|
https://arxiv.org/pdf/2305.13453v2.pdf
|
A Meta-learning based Generalizable Indoor Localization Model using Channel State Information
|
Indoor localization has gained significant attention in recent years due to its various applications in smart homes, industrial automation, and healthcare, especially since more people rely on their wireless devices for location-based services. Deep learning-based solutions have shown promising results in accurately estimating the position of wireless devices in indoor environments using wireless parameters such as Channel State Information (CSI) and Received Signal Strength Indicator (RSSI). However, despite the success of deep learning-based approaches in achieving high localization accuracy, these models suffer from a lack of generalizability and can not be readily-deployed to new environments or operate in dynamic environments without retraining. In this paper, we propose meta-learning-based localization models to address the lack of generalizability that persists in conventionally trained DL-based localization models. Furthermore, since meta-learning algorithms require diverse datasets from several different scenarios, which can be hard to collect in the context of localization, we design and propose a new meta-learning algorithm, TB-MAML (Task Biased Model Agnostic Meta Learning), intended to further improve generalizability when the dataset is limited. Lastly, we evaluate the performance of TB-MAML-based localization against conventionally trained localization models and localization done using other meta-learning algorithms.
|
['Kurt Turck', 'Jonathan Ashdown', 'Fatemeh Afghah', 'Linke Guo', 'ChunChih Lin', 'Ali Owfi']
|
2023-05-22
| null | null | null | null |
['indoor-localization']
|
['computer-vision']
|
[-1.45303935e-01 -3.68571132e-01 -2.48650566e-01 -5.62020063e-01
-1.11064041e+00 -1.78741708e-01 3.20890397e-01 1.37061223e-01
-4.41410571e-01 9.78361785e-01 1.08053423e-01 -4.66406763e-01
-4.96716470e-01 -7.21480608e-01 -8.27057719e-01 -7.01503694e-01
-2.64767408e-01 2.84630954e-01 -1.86999410e-01 2.33840883e-01
-1.49067938e-01 4.10601765e-01 -1.00591087e+00 -3.36635523e-02
7.58077919e-01 1.01729381e+00 3.38153332e-01 5.90762556e-01
8.00023973e-02 5.39104879e-01 -1.03701806e+00 2.10361734e-01
-8.46098438e-02 -1.31434798e-01 -4.68269408e-01 -5.21899939e-01
3.28721702e-01 -2.89814025e-01 -5.73130071e-01 6.18496954e-01
1.03576601e+00 1.23320684e-01 2.82201707e-01 -1.41115928e+00
-4.87213165e-01 5.85262716e-01 -1.99227110e-01 1.82285532e-01
6.60328388e-01 -2.92241067e-01 3.92351478e-01 -4.56290960e-01
-6.41156211e-02 7.96396792e-01 1.45989990e+00 4.11878467e-01
-9.48056161e-01 -7.99964845e-01 1.91169381e-01 1.47181168e-01
-1.48560441e+00 -5.11101604e-01 5.85671544e-01 2.33673155e-01
8.04400444e-01 2.84730494e-02 4.47425663e-01 1.81056345e+00
5.54629147e-01 6.07645929e-01 7.78038919e-01 -3.91851187e-01
7.00122833e-01 -8.78619030e-02 -3.41083169e-01 4.20954913e-01
4.57336068e-01 -1.31814599e-01 -5.27358532e-01 -2.86876291e-01
5.59456885e-01 6.50780678e-01 -1.45538270e-01 -4.70206380e-01
-1.50456917e+00 3.19294721e-01 9.52547908e-01 6.36272490e-01
-5.46486795e-01 7.73465693e-01 1.88376099e-01 4.74907532e-02
4.80031818e-01 5.35839200e-01 -9.75310922e-01 -4.77101475e-01
-9.55222189e-01 3.69634144e-02 8.52064431e-01 1.20523393e+00
6.01375282e-01 -9.16055441e-02 -1.04071535e-02 6.63547873e-01
3.98161381e-01 7.41492629e-01 7.47925937e-01 -9.00448501e-01
7.07138240e-01 1.29023552e-01 3.64144713e-01 -9.06362951e-01
-9.73372281e-01 -9.49167192e-01 -1.00012016e+00 -5.39023578e-01
1.04692206e-01 -5.59140027e-01 -6.54039741e-01 1.87123215e+00
8.65964592e-02 6.51053965e-01 5.66173568e-02 2.74858207e-01
6.40214145e-01 5.54975152e-01 1.44778579e-01 1.44085214e-01
6.89145505e-01 -9.63296354e-01 -5.89712203e-01 -4.65834886e-01
8.71312976e-01 -3.98829669e-01 7.00687230e-01 1.09619789e-01
-6.32025540e-01 -6.67527020e-01 -1.23503101e+00 4.14665610e-01
-7.30569899e-01 -3.61906402e-02 9.01214242e-01 1.06495643e+00
-1.14230537e+00 3.04616421e-01 -1.12094319e+00 -7.96393335e-01
4.79498982e-01 7.02384055e-01 -2.29237586e-01 -3.50356251e-01
-1.02725351e+00 7.81299472e-01 1.34924814e-01 3.97239834e-01
-5.37238419e-01 -5.73604763e-01 -8.73748183e-01 -4.15168963e-02
-8.74894261e-02 -7.82854199e-01 1.32290018e+00 -2.74853408e-01
-1.44021308e+00 -1.64757743e-01 -4.52466160e-01 -6.23628914e-01
3.85005832e-01 -1.20182589e-01 -7.75934398e-01 -5.50813437e-01
3.96589011e-01 3.48936230e-01 4.33635771e-01 -1.27939224e+00
-1.02498353e+00 -3.52637291e-01 -7.89480191e-03 -3.09785038e-01
-3.47553194e-01 -8.52421761e-01 -1.80930540e-01 -4.10623819e-01
5.11125684e-01 -8.62371147e-01 -5.01310349e-01 -3.62635076e-01
-2.12060839e-01 1.96799655e-02 9.80046690e-01 -2.35963076e-01
1.13604498e+00 -1.98906076e+00 -5.66622496e-01 3.37205559e-01
1.01728283e-01 7.94678703e-02 -9.69713405e-02 4.64350879e-01
4.88368779e-01 -1.71160862e-01 2.46475413e-01 -7.28065670e-01
1.21504612e-01 3.29540730e-01 2.18771137e-02 5.70652306e-01
-6.92014039e-01 1.13394248e+00 -1.36437213e+00 -1.34144396e-01
6.90411210e-01 7.30682373e-01 -2.76627392e-01 -8.41982663e-02
2.60729969e-01 7.63401389e-01 -6.19179487e-01 7.89957285e-01
7.02352464e-01 -4.13107276e-01 1.99153855e-01 -2.64193267e-01
9.37306359e-02 3.72531265e-01 -9.07553971e-01 2.05584979e+00
-1.41500354e+00 7.76992381e-01 -1.80645242e-01 -9.01677072e-01
6.03905559e-01 3.49532396e-01 8.95249128e-01 -9.16826546e-01
1.33721828e-01 3.37178707e-01 -1.71341345e-01 -3.89886916e-01
3.50933105e-01 4.69982624e-01 -3.51908773e-01 3.46108466e-01
1.61251545e-01 3.04949492e-01 -3.50204378e-01 -2.26943851e-01
1.63559747e+00 -7.52316490e-02 1.82455629e-01 1.21205956e-01
3.27907979e-01 -3.85241359e-01 3.01743388e-01 1.37025285e+00
-2.37204254e-01 4.20749307e-01 -6.27030194e-01 -4.60766256e-01
-5.35682321e-01 -1.23289585e+00 -2.78031260e-01 1.25675046e+00
3.42554450e-01 -4.04133677e-01 -5.40683389e-01 -7.43433416e-01
3.23409855e-01 5.98796189e-01 -4.35577750e-01 -2.65495867e-01
-5.29271364e-01 -8.98472130e-01 8.45612466e-01 8.33122015e-01
1.01061606e+00 -6.56244755e-01 -5.45113385e-01 6.27650917e-01
-3.01347286e-01 -1.24067736e+00 5.04288599e-02 7.51652420e-01
-7.00481474e-01 -7.43104339e-01 -7.96688616e-01 -7.94908702e-01
6.21816576e-01 5.60921848e-01 7.57674396e-01 -1.46556631e-01
1.13558352e-01 9.23850417e-01 -2.83292264e-01 -4.48125660e-01
-1.65547077e-02 8.13362539e-01 4.27423507e-01 -2.50991881e-01
3.42207044e-01 -7.64469564e-01 -6.67155623e-01 3.67181540e-01
-1.51411071e-01 -5.84179699e-01 7.36060977e-01 7.24561930e-01
2.52980679e-01 4.06699181e-01 1.05199933e+00 -5.03271759e-01
6.61123157e-01 -7.74367094e-01 -3.62343282e-01 2.80204684e-01
-5.34863174e-01 -2.41722986e-01 5.95750272e-01 -4.93137836e-01
-7.23378837e-01 -1.49865687e-01 -3.66169721e-01 2.52630919e-01
-2.32922733e-01 4.03985262e-01 -3.57714176e-01 -5.72376251e-01
5.03417850e-01 1.40656784e-01 -5.68022847e-01 -4.94102865e-01
1.14016771e-01 9.54450846e-01 5.67162395e-01 -5.33499777e-01
6.65952623e-01 3.89871955e-01 1.62906364e-01 -6.81418657e-01
-1.08212662e+00 -5.60509026e-01 -5.85596681e-01 6.45594671e-02
3.45707715e-01 -1.07280171e+00 -8.73565257e-01 3.47710609e-01
-9.98765290e-01 -6.19710803e-01 -2.55065411e-03 6.27878070e-01
-4.28674877e-01 -5.52299395e-02 -1.78209126e-01 -6.66674376e-01
-1.91806778e-01 -1.20738602e+00 1.16173494e+00 2.39755452e-01
-2.95877755e-01 -1.58192885e+00 4.19746749e-02 4.49417159e-02
1.06562161e+00 2.41378412e-01 4.75798398e-01 -5.31690836e-01
-5.05230069e-01 -7.77620792e-01 -4.29129321e-03 -1.22213475e-02
7.58701324e-01 -8.35811555e-01 -1.14756823e+00 -6.19372010e-01
-3.39153022e-01 -2.14964952e-02 3.97887439e-01 7.16488302e-01
1.39824927e+00 -2.04497367e-01 -1.03452623e+00 9.09187198e-01
1.09083593e+00 1.31834507e-01 1.99412704e-01 8.00242245e-01
5.80012023e-01 -2.82540113e-01 2.57288933e-01 3.46560478e-01
8.39284360e-01 8.06176722e-01 4.52220201e-01 -2.17709973e-01
-1.68270215e-01 -4.13984656e-01 4.55878414e-02 4.63104784e-01
1.73611060e-01 -5.02737045e-01 -7.83469498e-01 4.23709303e-01
-2.09376073e+00 -7.76688755e-01 1.91316485e-01 2.14052439e+00
2.47863516e-01 9.86655131e-02 -2.10152268e-01 8.85286778e-02
3.68735075e-01 1.54219046e-01 -6.10828459e-01 1.13589451e-01
2.27319911e-01 2.16165647e-01 9.61697578e-01 3.65140438e-01
-1.31691718e+00 7.29753792e-01 6.56018305e+00 4.69398230e-01
-1.32381570e+00 5.80179274e-01 2.87695557e-01 7.64950439e-02
1.99657947e-01 -5.50905883e-01 -6.81530237e-01 7.63888896e-01
1.09610558e+00 4.50347543e-01 3.48545969e-01 1.27899504e+00
8.72971341e-02 -1.99628681e-01 -1.36129224e+00 1.40886962e+00
-8.81090984e-02 -1.25278926e+00 -5.40902793e-01 2.56840527e-01
9.91377115e-01 5.20218194e-01 4.80012149e-01 6.97684824e-01
6.78063333e-02 -8.94586682e-01 6.53668940e-01 2.83787400e-01
4.66224164e-01 -7.38009036e-01 1.25802076e+00 3.51636827e-01
-1.26583493e+00 -4.27389294e-01 -3.86436254e-01 -9.93136838e-02
2.65730917e-01 6.84314728e-01 -1.10436761e+00 4.21805948e-01
9.00230765e-01 5.59214652e-01 -6.79811776e-01 1.43413687e+00
-1.82167754e-01 5.73308170e-01 -5.19332528e-01 -2.49988303e-01
3.01212341e-01 5.19610822e-01 5.27994297e-02 1.11403656e+00
7.80407548e-01 -5.94443381e-01 2.12878421e-01 5.16583502e-01
-1.46440387e-01 -6.05294764e-01 -7.63503253e-01 6.39163971e-01
1.08228719e+00 1.01536655e+00 -4.58362520e-01 9.74591151e-02
-2.70677060e-01 1.06822729e+00 1.80179882e-03 7.32809722e-01
-8.98366988e-01 -3.09928328e-01 5.84683537e-01 1.20590581e-02
4.78287749e-02 -6.51437283e-01 -2.37830341e-01 -8.54877472e-01
-1.93031967e-01 -3.26123685e-01 -1.79444790e-01 -5.42752445e-01
-1.09039330e+00 3.01307946e-01 -1.00178294e-01 -9.94697213e-01
-5.58311284e-01 -4.99132007e-01 -5.06927788e-01 4.68966335e-01
-1.39007199e+00 -1.49801517e+00 -6.57327473e-01 5.38903773e-01
3.43989640e-01 -1.36512145e-01 1.05113971e+00 7.60503531e-01
-5.10943294e-01 1.00369644e+00 8.13204587e-01 7.75173157e-02
8.67118239e-01 -1.16819036e+00 6.14575148e-01 4.93407160e-01
2.90783644e-01 1.06281567e+00 5.56434095e-01 -2.69390523e-01
-1.55398619e+00 -1.52705824e+00 7.92186975e-01 -8.80522490e-01
2.88824260e-01 -5.43012202e-01 -1.60247311e-01 9.97020721e-01
-3.54891896e-01 3.61669838e-01 6.92773342e-01 5.76724589e-01
8.68148059e-02 -6.91406608e-01 -1.47455490e+00 4.70943928e-01
1.14960074e+00 -4.44506586e-01 -3.06375120e-02 3.71503979e-01
4.58068728e-01 -4.72554982e-01 -7.04044700e-01 5.36875963e-01
8.13628495e-01 -7.32257009e-01 1.26882362e+00 6.49745017e-02
-6.68444037e-01 -2.92417884e-01 -5.87736845e-01 -1.45028019e+00
-4.03903186e-01 -3.56674641e-01 -5.62955976e-01 1.29029107e+00
3.09204698e-01 -1.04959595e+00 9.81380761e-01 2.59668112e-01
-3.62870954e-02 -7.20912516e-01 -1.36611652e+00 -9.23815072e-01
-3.23124379e-01 -7.97474146e-01 1.22577119e+00 6.53094292e-01
-2.39128485e-01 2.37230863e-02 -3.75847459e-01 5.70197642e-01
5.46759248e-01 -4.50246543e-01 1.02888894e+00 -1.12855411e+00
-8.92465636e-02 4.21883771e-03 -5.62000990e-01 -1.38557339e+00
2.25345194e-01 -5.43267906e-01 2.56761760e-01 -2.01957512e+00
-2.92873025e-01 -1.16126418e+00 -8.75463605e-01 6.04059458e-01
3.28393877e-01 3.13136011e-01 -3.49708110e-01 -3.47783603e-03
-1.29901493e+00 5.43681979e-01 4.76133913e-01 -5.14473259e-01
-4.02703047e-01 7.75769591e-01 -5.03291428e-01 7.36470520e-01
9.20723617e-01 -3.50161612e-01 -3.69934261e-01 -7.29219556e-01
1.67359158e-01 -2.69315183e-01 5.47770321e-01 -1.86080527e+00
6.32387161e-01 2.82266557e-01 1.10093915e+00 -6.28908336e-01
5.18227339e-01 -1.34339392e+00 5.51102534e-02 4.78304565e-01
4.64357622e-02 1.77664131e-01 9.99373645e-02 6.73492074e-01
2.66522348e-01 1.03990301e-01 7.73332939e-02 1.12019271e-01
-8.13736081e-01 2.92380095e-01 -4.88368511e-01 -5.32907486e-01
7.08134294e-01 -3.37327272e-01 -1.97322786e-01 -6.22518718e-01
-3.56938601e-01 9.31651667e-02 3.51929694e-01 5.05240917e-01
4.55166876e-01 -1.57505310e+00 1.24124579e-01 2.02276468e-01
2.52901495e-01 6.74743876e-02 3.92793156e-02 8.32210600e-01
-3.37979853e-01 6.90734744e-01 2.22348332e-01 -7.65704632e-01
-5.63988745e-01 2.08677843e-01 5.30579686e-01 -3.77017632e-02
-4.85964358e-01 7.20502853e-01 -4.17631060e-01 -8.98921847e-01
7.91118443e-01 -7.92269051e-01 5.24215281e-01 -5.88795960e-01
3.42530578e-01 7.22691238e-01 4.10512745e-01 -4.60203379e-01
-7.11801469e-01 4.74666536e-01 2.74378687e-01 1.66173831e-01
1.12753010e+00 -4.23737466e-01 2.15279713e-01 4.77794766e-01
1.28313816e+00 1.07542746e-01 -1.08990240e+00 -1.71555907e-01
2.69593209e-01 -1.89158469e-01 2.03303888e-01 -9.86734569e-01
-7.71026909e-01 5.03551781e-01 1.14777470e+00 6.15572147e-02
6.89570427e-01 -4.34257984e-02 1.18733871e+00 1.05594718e+00
1.42156005e+00 -7.76002169e-01 -1.07237853e-01 4.71604615e-01
5.82693405e-02 -1.39164519e+00 -1.68599989e-02 4.03618872e-01
3.36326867e-01 8.58221412e-01 3.81614923e-01 2.13045374e-01
1.04286516e+00 3.48461717e-01 1.82789117e-01 2.65082777e-01
-5.67345731e-02 5.22718020e-02 3.29507305e-03 1.34051538e+00
4.39636409e-01 2.32649129e-02 5.55102527e-01 6.85435236e-01
-2.51111746e-01 -6.69053346e-02 -1.99809402e-01 1.14977598e+00
-4.67973411e-01 -1.19612706e+00 -4.40094441e-01 3.80220592e-01
-2.77991414e-01 2.28096738e-01 1.96366578e-01 7.21304715e-01
5.93964398e-01 1.41709638e+00 -1.83051631e-01 -5.00717819e-01
2.49782234e-01 -3.06940168e-01 6.34426534e-01 -4.58554059e-01
-2.61201918e-01 -4.44719225e-01 -3.26684952e-01 -6.21650755e-01
-3.24108660e-01 -2.51641452e-01 -1.15799582e+00 -3.69089901e-01
-3.60506356e-01 2.66107798e-01 1.12974870e+00 1.22186673e+00
6.06517732e-01 1.01141787e+00 5.36262274e-01 -1.40088487e+00
-3.49375218e-01 -9.95014310e-01 -3.89424592e-01 -2.30981931e-01
7.59648204e-01 -7.40376234e-01 -7.99933672e-02 -6.95247054e-01]
|
[6.398256778717041, 0.9231898784637451]
|
9914c214-6afb-4489-86dd-165d89d49588
|
incremental-learning-for-heterogeneous
|
2305.19404
| null |
https://arxiv.org/abs/2305.19404v1
|
https://arxiv.org/pdf/2305.19404v1.pdf
|
Incremental Learning for Heterogeneous Structure Segmentation in Brain Tumor MRI
|
Deep learning (DL) models for segmenting various anatomical structures have achieved great success via a static DL model that is trained in a single source domain. Yet, the static DL model is likely to perform poorly in a continually evolving environment, requiring appropriate model updates. In an incremental learning setting, we would expect that well-trained static models are updated, following continually evolving target domain data -- e.g., additional lesions or structures of interest -- collected from different sites, without catastrophic forgetting. This, however, poses challenges, due to distribution shifts, additional structures not seen during the initial model training, and the absence of training data in a source domain. To address these challenges, in this work, we seek to progressively evolve an ``off-the-shelf" trained segmentation model to diverse datasets with additional anatomical categories in a unified manner. Specifically, we first propose a divergence-aware dual-flow module with balanced rigidity and plasticity branches to decouple old and new tasks, which is guided by continuous batch renormalization. Then, a complementary pseudo-label training scheme with self-entropy regularized momentum MixUp decay is developed for adaptive network optimization. We evaluated our framework on a brain tumor segmentation task with continually changing target domains -- i.e., new MRI scanners/modalities with incremental structures. Our framework was able to well retain the discriminability of previously learned structures, hence enabling the realistic life-long segmentation model extension along with the widespread accumulation of big medical data.
|
['Jonghye Woo', 'Georges El Fakhri', 'Emiliano Santarnecchi', 'Fangxu Xing', 'Helen A. Shih', 'Xiaofeng Liu']
|
2023-05-30
| null | null | null | null |
['tumor-segmentation', 'brain-tumor-segmentation', 'incremental-learning', 'pseudo-label']
|
['computer-vision', 'medical', 'methodology', 'miscellaneous']
|
[ 5.95809042e-01 4.83594894e-01 5.88568486e-03 -2.61260986e-01
-7.09388018e-01 -5.68725705e-01 4.33667481e-01 1.22585043e-01
-5.98677456e-01 9.05355513e-01 1.17985748e-01 -6.32266849e-02
-1.57215521e-01 -5.48883021e-01 -6.21242702e-01 -9.72754180e-01
-5.74726388e-02 7.91532278e-01 5.25747061e-01 -3.19570124e-01
-3.58240008e-02 5.10600388e-01 -9.82561946e-01 6.36583939e-02
1.01899266e+00 7.69695938e-01 5.71328521e-01 4.07159865e-01
-4.48079184e-02 4.38843787e-01 -7.28776082e-02 -3.56408864e-01
2.62075126e-01 -3.96709472e-01 -9.10150766e-01 5.05714230e-02
3.37096125e-01 -4.92527410e-02 -3.56361479e-01 9.50338125e-01
5.98441064e-01 7.37961605e-02 5.61517358e-01 -4.66457337e-01
-6.90265000e-01 6.37712002e-01 -4.37160164e-01 3.57279152e-01
-4.08214390e-01 5.29811740e-01 6.42568231e-01 -6.03767276e-01
1.07355499e+00 5.61653972e-01 8.90211821e-01 1.05749607e+00
-1.56972241e+00 -2.59321213e-01 4.19330925e-01 -2.16738448e-01
-9.46231067e-01 -2.86776245e-01 9.19538498e-01 -5.48428357e-01
5.45749903e-01 7.20461085e-02 8.78036618e-01 1.55588496e+00
4.58563179e-01 5.56155145e-01 1.03997731e+00 3.81262414e-02
3.27880889e-01 1.14364497e-01 1.91455185e-01 5.72544515e-01
1.69170350e-01 -3.44344117e-02 -2.39063561e-01 -1.46758826e-02
7.11054802e-01 -3.06955371e-02 -3.50861073e-01 -5.95059931e-01
-1.34188688e+00 5.28639674e-01 6.52456462e-01 6.48700833e-01
-4.20152605e-01 -1.92555366e-03 5.89840829e-01 2.40681529e-01
6.31759942e-01 5.12811244e-01 -6.88310146e-01 6.74223453e-02
-1.19257033e+00 -4.50854190e-02 5.21706760e-01 5.32102823e-01
4.55963433e-01 1.40178041e-03 -2.32938603e-01 8.31121206e-01
-1.95653029e-02 1.08396709e-01 9.27824497e-01 -6.77529335e-01
1.98117927e-01 5.50841987e-01 -3.08684736e-01 -5.25033295e-01
-8.05257440e-01 -1.16154051e+00 -1.23833179e+00 9.99794602e-02
4.06247824e-01 -2.36106619e-01 -1.35464644e+00 2.38767147e+00
4.81253922e-01 1.96384549e-01 -5.36278859e-02 6.88981235e-01
4.44201022e-01 1.44743592e-01 5.48922867e-02 -3.86524469e-01
9.95431364e-01 -1.05852747e+00 -4.16393489e-01 -3.86078328e-01
6.59621835e-01 -1.87448218e-01 1.12572348e+00 2.99675047e-01
-1.16253138e+00 -3.97978127e-01 -8.73917341e-01 7.68373311e-02
-2.45586067e-01 -3.43721300e-01 5.74382782e-01 4.74752367e-01
-1.29114747e+00 8.24548542e-01 -1.19350278e+00 -3.30024481e-01
7.31501997e-01 5.31351149e-01 -4.11006749e-01 9.67474580e-02
-1.16570640e+00 9.45491672e-01 4.68397945e-01 1.87707245e-01
-1.11375237e+00 -1.09999776e+00 -4.82032210e-01 -1.48154348e-01
2.90423334e-01 -1.02182913e+00 9.81668055e-01 -1.28618014e+00
-1.67003965e+00 9.55439270e-01 1.05997855e-02 -6.19849920e-01
9.37126338e-01 1.13789313e-01 -1.48943067e-01 3.98847982e-02
-1.52406439e-01 7.31339514e-01 1.02768064e+00 -1.30991781e+00
-7.64167011e-02 -5.52182257e-01 -1.28379777e-01 4.23116475e-01
-4.73593205e-01 -5.08016348e-01 -2.02221930e-01 -7.38226712e-01
1.95722431e-01 -1.23002911e+00 -6.32926464e-01 8.02826583e-02
-7.51278624e-02 4.68226045e-01 5.30818522e-01 -7.71920443e-01
1.11872292e+00 -2.12574697e+00 7.76610494e-01 1.11324511e-01
6.04743481e-01 2.13188201e-01 -1.56850216e-03 -1.77313492e-01
-2.66197801e-01 7.99452141e-03 -8.93855393e-01 -5.13877451e-01
-4.47635978e-01 2.79365778e-01 1.35453761e-01 3.58113497e-01
2.10631043e-01 1.12007046e+00 -9.50292528e-01 -4.20267940e-01
-1.19495556e-01 3.86837751e-01 -7.54798651e-01 -6.61421865e-02
-2.67307609e-01 1.16002953e+00 -2.69766539e-01 4.27866608e-01
6.74280882e-01 -5.83045900e-01 1.52003199e-01 -7.37106651e-02
1.48829088e-01 -1.98311135e-01 -7.56486356e-01 2.31056595e+00
-4.92795169e-01 1.95404515e-01 4.20818746e-01 -1.25374961e+00
5.59385359e-01 2.51242578e-01 9.70506370e-01 -5.27326643e-01
1.24024928e-01 4.63633597e-01 3.79134029e-01 -2.70772606e-01
1.67701468e-01 -5.28816164e-01 -1.58468429e-02 2.69942164e-01
3.72955978e-01 -1.32061273e-01 -6.40269229e-03 5.51382452e-02
1.35197532e+00 -7.91324824e-02 -9.16282982e-02 -4.15351838e-01
5.00415325e-01 7.03346059e-02 5.71812391e-01 7.15432584e-01
-4.63573933e-01 8.46601248e-01 3.42542678e-01 -2.92530686e-01
-9.94494498e-01 -1.23131490e+00 -5.24308562e-01 8.17771673e-01
-2.49414183e-02 3.65205780e-02 -8.54671240e-01 -9.75385368e-01
-1.74918368e-01 3.33419204e-01 -9.14065719e-01 -6.83489382e-01
-8.61936033e-01 -1.35751784e+00 3.93046558e-01 2.05174044e-01
2.98856378e-01 -9.89489853e-01 -7.23918021e-01 6.65833712e-01
6.20534532e-02 -6.98237062e-01 -4.34789896e-01 7.13181436e-01
-1.25898445e+00 -7.56094337e-01 -1.25707304e+00 -8.30307245e-01
8.67022574e-01 -2.32686982e-01 9.83324349e-01 -2.64010355e-02
-2.71898180e-01 3.03506583e-01 -3.63177955e-02 1.58156484e-01
-6.12476707e-01 6.87331736e-01 3.53200696e-02 2.98969317e-02
-4.07721579e-01 -8.75352263e-01 -8.93830776e-01 1.70152694e-01
-1.10729718e+00 3.45645159e-01 7.37237513e-01 1.29329336e+00
6.80564106e-01 -3.01216781e-01 8.05943429e-01 -9.74190235e-01
5.67844152e-01 -7.16726422e-01 -2.09973127e-01 3.59039068e-01
-6.93931043e-01 2.46935815e-01 8.14124882e-01 -8.44616950e-01
-1.20663738e+00 1.70161687e-02 -4.05328751e-01 -2.43968889e-01
3.10237575e-02 6.08128369e-01 1.28168389e-01 -2.94215977e-01
8.46573353e-01 3.51212412e-01 3.36032838e-01 -4.07716244e-01
4.56242144e-01 1.33910328e-01 5.70561469e-01 -7.25017190e-01
6.66823328e-01 6.30458891e-01 -4.18503862e-03 -3.71717572e-01
-8.07363212e-01 -1.26955733e-01 -1.03827274e+00 -2.73995042e-01
7.73270547e-01 -5.28944790e-01 -7.05093071e-02 8.96630585e-01
-7.49237299e-01 -7.13921249e-01 -8.12325895e-01 2.93986529e-01
-5.01856446e-01 3.16051990e-01 -7.84729481e-01 -1.05124228e-01
-5.34418285e-01 -1.32581949e+00 7.45459795e-01 2.62374789e-01
-2.83153728e-02 -1.29958475e+00 4.41741347e-01 1.15105025e-01
8.92348528e-01 4.16248232e-01 9.82177913e-01 -6.23014092e-01
-2.85577089e-01 3.35101672e-02 6.54859692e-02 4.79256988e-01
3.93097281e-01 -3.30402613e-01 -7.37530053e-01 -6.19272709e-01
2.76900560e-01 -3.26523513e-01 1.07911777e+00 4.23668057e-01
1.24643087e+00 -2.26499680e-02 -4.62189883e-01 8.63156557e-01
1.27831304e+00 4.89539839e-02 2.46324345e-01 5.18249646e-02
6.71200275e-01 4.68169093e-01 -5.21619543e-02 8.95745754e-02
2.38554016e-01 4.92340088e-01 3.50236803e-01 -3.18274975e-01
-4.36701238e-01 4.37793098e-02 1.77956834e-01 1.20549738e+00
-6.16431832e-02 5.73727749e-02 -1.13901782e+00 5.75161994e-01
-1.69824564e+00 -6.46470547e-01 1.46071047e-01 2.12716866e+00
1.19132161e+00 5.26878119e-01 -1.73737943e-01 -4.25368249e-01
5.43328881e-01 1.34721681e-01 -1.26346791e+00 9.70307440e-02
-2.12503061e-01 3.45125675e-01 3.94369304e-01 3.69453996e-01
-9.27782059e-01 8.54878783e-01 5.85534668e+00 7.01437116e-01
-1.66123664e+00 5.99789917e-01 9.03607070e-01 -3.13831329e-01
-3.57251108e-01 -2.62406860e-02 -5.21815777e-01 6.36820436e-01
1.07562912e+00 -2.15769187e-01 3.66247028e-01 6.30121589e-01
-8.04788470e-02 4.93035018e-02 -9.05142665e-01 5.75665534e-01
-6.86821789e-02 -1.50096059e+00 -1.73248991e-01 8.44532698e-02
8.13582897e-01 5.67178667e-01 3.28484327e-01 3.13562989e-01
2.47126147e-01 -6.82362556e-01 6.25267506e-01 7.02238858e-01
9.30850208e-01 -1.97099715e-01 4.89538074e-01 4.27818090e-01
-7.87506938e-01 -6.72624484e-02 -4.19250056e-02 6.06958449e-01
3.67649943e-01 6.90361202e-01 -6.00036204e-01 6.36697650e-01
4.69260871e-01 7.33311057e-01 -8.14792156e-01 9.82761204e-01
3.24531198e-01 5.23550391e-01 -5.16680837e-01 4.59976494e-01
2.47736216e-01 -1.91705897e-01 6.10296905e-01 1.02237773e+00
2.44616807e-01 -1.07998438e-01 -5.23923477e-03 7.66457438e-01
-1.36560306e-01 -2.18618847e-02 -2.88248420e-01 4.34375674e-01
-1.02548257e-01 1.34954643e+00 -1.02215004e+00 -2.58016109e-01
-1.16809204e-01 1.10698903e+00 5.71789384e-01 2.46650696e-01
-1.01385677e+00 2.88187027e-01 3.09604079e-01 2.68961161e-01
8.88506472e-02 -2.43800133e-01 -4.44079161e-01 -1.50811160e+00
4.16716980e-03 -5.57798445e-01 4.15082365e-01 -2.40904823e-01
-1.42870641e+00 8.74849439e-01 -6.92578256e-02 -8.52910221e-01
4.67620157e-02 -3.84492129e-01 -5.40371418e-01 6.49656236e-01
-1.56428695e+00 -1.07693756e+00 -2.21104950e-01 5.38655341e-01
5.07710159e-01 -1.36417588e-02 6.51179850e-01 5.46753407e-01
-6.57927990e-01 5.41796029e-01 5.02796888e-01 -2.19672292e-01
7.30804741e-01 -1.20725453e+00 2.48107806e-01 6.99692488e-01
-1.33642256e-01 4.96026844e-01 3.95368755e-01 -8.10764432e-01
-9.21650469e-01 -1.04527259e+00 1.60071373e-01 -4.27910149e-01
8.42652142e-01 -5.10485590e-01 -1.41382921e+00 5.28020918e-01
-2.02591829e-02 3.14956427e-01 4.87267464e-01 -1.30669579e-01
1.80029035e-01 -1.22078545e-01 -1.29946601e+00 5.16309440e-01
1.41348970e+00 -2.78255403e-01 -5.46641648e-01 4.77953017e-01
8.21177304e-01 -6.52921736e-01 -1.02024245e+00 6.51426315e-01
2.91258514e-01 -5.96224189e-01 9.60753977e-01 -7.24541366e-01
1.56181723e-01 4.56453627e-03 3.51067007e-01 -1.47287869e+00
-4.42355990e-01 -5.93220770e-01 -1.80825353e-01 8.99518251e-01
6.32880628e-01 -7.95570433e-01 8.08977246e-01 7.38298714e-01
-5.32864451e-01 -1.07001686e+00 -1.33047068e+00 -5.58154941e-01
6.69556201e-01 -1.85432270e-01 2.64628410e-01 9.29400921e-01
-1.91256538e-01 1.78399011e-01 -2.50294805e-01 -2.32346892e-01
4.94667351e-01 -2.22676113e-01 1.75334498e-01 -1.29763234e+00
-4.50109124e-01 -7.33727276e-01 -3.74523737e-02 -8.22440743e-01
2.06007123e-01 -1.35774946e+00 -1.16943792e-01 -1.27761340e+00
2.54944772e-01 -7.71017671e-01 -6.36659086e-01 3.77844065e-01
-1.91524506e-01 -3.45652252e-02 2.25891941e-03 4.44802642e-01
-4.12629187e-01 7.72211611e-01 1.61193967e+00 -9.82328728e-02
-3.13277781e-01 1.96346045e-02 -6.92861974e-01 5.93993306e-01
8.06787372e-01 -5.89230418e-01 -5.08389592e-01 -4.51736540e-01
1.53904194e-02 -3.41662914e-02 2.90137351e-01 -1.14228189e+00
2.56676286e-01 9.61267427e-02 2.74928153e-01 -8.01088065e-02
1.20370097e-01 -8.27104986e-01 4.14756685e-01 6.21836185e-01
-3.48674536e-01 -1.14964671e-01 2.57833064e-01 6.28622830e-01
4.42040339e-02 -2.81023353e-01 1.15905857e+00 -4.98149842e-01
-3.27544153e-01 6.55638039e-01 -1.55568033e-01 2.89692670e-01
1.08047092e+00 -2.28839248e-01 -1.35902286e-01 2.95824051e-01
-1.40268290e+00 1.14521258e-01 6.92483068e-01 3.53357792e-01
2.74366617e-01 -1.04859352e+00 -6.42748773e-01 1.60543755e-01
-2.18162134e-01 4.33086574e-01 6.77682996e-01 1.20467305e+00
-3.91169846e-01 -1.41375978e-02 -3.27378422e-01 -6.79306209e-01
-6.48479044e-01 4.33975220e-01 7.88838148e-01 -9.57940817e-01
-9.12532687e-01 8.48071396e-01 3.13986063e-01 -6.76767290e-01
-2.02619471e-04 -5.18816233e-01 1.47752734e-02 1.50880471e-01
1.24807982e-02 -2.14115515e-01 3.84393036e-01 -4.81450200e-01
-2.53880948e-01 5.30424416e-01 -4.52132940e-01 -7.87057355e-02
1.55254900e+00 -1.14036888e-01 -9.87378042e-03 4.81382966e-01
1.05822647e+00 -3.23594809e-01 -1.69031811e+00 -3.76287192e-01
9.39643234e-02 4.20342013e-02 1.15688220e-01 -9.54822183e-01
-1.36637461e+00 7.22684741e-01 8.85565162e-01 -6.86543211e-02
1.08389807e+00 1.26107842e-01 9.55688655e-01 2.35545058e-02
4.81323034e-01 -1.02819538e+00 2.08144605e-01 4.82895464e-01
7.98022449e-01 -1.24181569e+00 -2.85680115e-01 2.24636778e-01
-6.59406185e-01 1.02117288e+00 5.64854920e-01 3.02686729e-02
8.27013254e-01 3.05074304e-01 -2.84618348e-01 -2.78613657e-01
-7.76478410e-01 2.40401570e-02 1.30846843e-01 4.44178462e-01
1.25080660e-01 -1.52219355e-01 -1.56395614e-01 5.86687446e-01
7.68758953e-02 1.77832067e-01 4.11550850e-01 9.12771285e-01
-5.84298335e-02 -1.11360526e+00 1.42217092e-02 6.78714573e-01
-3.76326889e-01 -1.00940831e-01 -5.25599569e-02 6.29483461e-01
1.83242768e-01 -7.84008801e-02 -1.63192600e-01 -9.81974751e-02
2.38950193e-01 2.19498590e-01 5.34141183e-01 -6.72434747e-01
-8.23011816e-01 -3.15075219e-02 -4.32481021e-01 -2.47303054e-01
-3.03215712e-01 -9.94031906e-01 -1.18454301e+00 9.01795179e-02
-2.57034361e-01 -1.60724074e-01 4.75225329e-01 8.04191113e-01
4.72330570e-01 8.83741140e-01 4.77881074e-01 -8.82472157e-01
-6.43638790e-01 -7.40741193e-01 -2.65169322e-01 5.27562737e-01
4.51179564e-01 -6.84726834e-01 -4.09534544e-01 1.10484753e-02]
|
[14.588056564331055, -2.1214759349823]
|
9a23ad37-30a3-40a7-8008-0108ac23f09c
|
source-free-domain-adaptation-with-image
|
2008.07514
| null |
https://arxiv.org/abs/2008.07514v2
|
https://arxiv.org/pdf/2008.07514v2.pdf
|
Source Free Domain Adaptation with Image Translation
|
Effort in releasing large-scale datasets may be compromised by privacy and intellectual property considerations. A feasible alternative is to release pre-trained models instead. While these models are strong on their original task (source domain), their performance might degrade significantly when deployed directly in a new environment (target domain), which might not contain labels for training under realistic settings. Domain adaptation (DA) is a known solution to the domain gap problem, but usually requires labeled source data. In this paper, we study the problem of source free domain adaptation (SFDA), whose distinctive feature is that the source domain only provides a pre-trained model, but no source data. Being source free adds significant challenges to DA, especially when considering that the target dataset is unlabeled. To solve the SFDA problem, we propose an image translation approach that transfers the style of target images to that of unseen source images. To this end, we align the batch-wise feature statistics of generated images to that stored in batch normalization layers of the pre-trained model. Compared with directly classifying target images, higher accuracy is obtained with these style transferred images using the pre-trained model. On several image classification datasets, we show that the above-mentioned improvements are consistent and statistically significant.
|
['Yunzhong Hou', 'Liang Zheng']
|
2020-08-17
| null | null | null | null |
['source-free-domain-adaptation']
|
['computer-vision']
|
[ 6.55046046e-01 9.58799571e-02 -2.45938078e-01 -5.85740566e-01
-8.36928129e-01 -9.26100433e-01 6.92371845e-01 -1.32496059e-01
-5.37403047e-01 9.60396588e-01 -1.28609493e-01 -3.92691651e-03
3.05845290e-01 -6.65402353e-01 -1.03593218e+00 -8.63577962e-01
5.46019495e-01 3.75515044e-01 -5.30699193e-02 -3.50750722e-02
3.76260914e-02 5.73524177e-01 -1.22545695e+00 3.62990260e-01
7.61309743e-01 1.03239787e+00 2.09354814e-02 3.53169560e-01
-1.00444354e-01 4.57394332e-01 -1.04465258e+00 -6.73705637e-01
6.40592754e-01 -6.54708862e-01 -8.77386630e-01 3.60148191e-01
6.03976786e-01 -3.46524090e-01 6.28142012e-03 1.28287983e+00
3.58494729e-01 -2.68917382e-01 5.51321447e-01 -1.69645131e+00
-1.07082248e+00 2.41211474e-01 -3.36665213e-01 -1.51724279e-01
-2.07653828e-02 -8.76654983e-02 4.40211117e-01 -7.19575167e-01
1.06506634e+00 7.49411762e-01 2.43342668e-01 9.67051744e-01
-1.50480068e+00 -7.93174982e-01 -8.24262723e-02 -1.53867215e-01
-1.39753866e+00 -5.51496923e-01 7.44433761e-01 -5.52109122e-01
3.92915696e-01 1.31145373e-01 1.15622925e-02 1.55931842e+00
1.61021426e-02 4.11505729e-01 1.28611493e+00 -5.98837078e-01
3.81653070e-01 1.03788769e+00 -1.59736171e-01 -3.15595381e-02
5.41371644e-01 -2.65971273e-02 -5.00378132e-01 -2.71228850e-01
5.85927248e-01 -9.58636589e-03 -2.58429348e-01 -8.69437337e-01
-1.21787775e+00 7.12861717e-01 2.49802142e-01 3.10442239e-01
-2.83402115e-01 -3.55415195e-01 2.87253439e-01 6.29873633e-01
6.47786975e-01 4.93571490e-01 -5.82713068e-01 3.91130388e-01
-9.05940354e-01 3.56932431e-02 5.50060391e-01 1.30916321e+00
9.34687972e-01 -1.40046418e-01 7.09226495e-03 6.97413146e-01
-2.60580946e-02 6.43124640e-01 7.53999233e-01 -6.50580943e-01
4.98768896e-01 4.00196880e-01 2.70858645e-01 -7.45699704e-01
1.12810999e-01 -4.96255755e-01 -7.09796786e-01 1.82163090e-01
6.08402908e-01 -7.06934854e-02 -9.08077836e-01 2.03711724e+00
4.91324753e-01 -4.15336341e-02 5.30471861e-01 7.28598237e-01
4.07091290e-01 6.40453696e-01 1.91512164e-02 -1.56420067e-01
1.05607152e+00 -9.45105612e-01 -5.46354473e-01 -4.12574202e-01
4.90530372e-01 -7.62925923e-01 1.14978480e+00 2.21744210e-01
-6.29268944e-01 -5.65636933e-01 -1.18535411e+00 5.65514378e-02
-5.11412740e-01 2.06552878e-01 -1.26379654e-02 8.99083138e-01
-1.07108593e+00 2.72559494e-01 -2.59571582e-01 -7.49693155e-01
5.54238200e-01 3.09086502e-01 -7.77967870e-01 -2.38906935e-01
-1.03186786e+00 7.28925943e-01 3.18937212e-01 -4.44632173e-01
-9.61042464e-01 -6.33319736e-01 -7.05477536e-01 -1.69205233e-01
2.31484920e-01 -4.26609933e-01 1.16616786e+00 -1.51970124e+00
-1.43166256e+00 1.19966435e+00 -1.14385948e-01 -4.21415716e-01
6.60278857e-01 4.60601505e-03 -6.59823835e-01 6.51021227e-02
3.52436900e-01 8.10450733e-01 1.35492420e+00 -1.59066224e+00
-6.26069665e-01 -3.08001965e-01 -7.85591826e-02 -6.72028363e-02
-8.29579473e-01 1.61707401e-01 -2.57646412e-01 -5.79267502e-01
-2.42606908e-01 -1.09438276e+00 -1.04263037e-01 9.23177525e-02
-2.45371327e-01 3.15034568e-01 9.28196728e-01 -6.26165152e-01
8.14458668e-01 -2.51202512e+00 -1.20954327e-01 2.35867426e-01
-2.85238754e-02 4.78574753e-01 -3.99595946e-01 1.60008430e-01
-3.09587806e-01 1.70744613e-01 -3.87832463e-01 -3.45791906e-01
-2.18370035e-01 1.48883656e-01 -6.54475272e-01 3.81788194e-01
4.10280049e-01 6.40772641e-01 -6.73586249e-01 -3.10296237e-01
-2.53539264e-01 3.17914218e-01 -3.33154261e-01 3.77498358e-01
-7.52745792e-02 7.94642806e-01 -1.82523921e-01 3.39578032e-01
1.13008285e+00 -3.53391230e-01 2.34094545e-01 6.29981160e-02
2.60942996e-01 -7.64865652e-02 -8.59823585e-01 1.70413184e+00
-3.43816996e-01 5.38453460e-01 2.97836959e-02 -9.11467433e-01
1.17872691e+00 3.26050520e-01 2.91964412e-01 -7.72012949e-01
-1.78628473e-03 4.00855869e-01 -2.08171517e-01 -2.77613223e-01
5.30731320e-01 -3.37834537e-01 -2.56227255e-01 7.26588368e-01
1.89016163e-01 1.35046579e-02 -1.02725230e-01 1.89962640e-01
8.40982378e-01 1.04734369e-01 7.65876621e-02 -3.21644209e-02
4.83057618e-01 1.87185228e-01 5.92327118e-01 4.70238060e-01
-2.72106618e-01 8.49251032e-01 4.49375421e-01 -2.29592159e-01
-1.36087143e+00 -1.02726364e+00 -2.00167358e-01 9.33998346e-01
4.16146144e-02 -3.81258242e-02 -1.05878866e+00 -1.15057290e+00
-1.06241614e-01 6.50362134e-01 -6.38279259e-01 -3.55193526e-01
-3.53552103e-01 -4.15797710e-01 6.45574868e-01 3.14657569e-01
5.61459363e-01 -7.71308482e-01 -4.13849652e-01 9.87583026e-02
-3.02419901e-01 -1.27175379e+00 -5.49164534e-01 -9.58006270e-03
-7.24205911e-01 -7.72581220e-01 -1.00732648e+00 -7.85681188e-01
1.20717478e+00 3.27397674e-01 7.44846165e-01 -3.00869256e-01
1.49289891e-01 2.97680408e-01 -4.35057491e-01 -5.74274361e-01
-8.41377974e-01 1.28736958e-01 1.80498332e-01 5.19895673e-01
4.75501835e-01 -3.06863904e-01 -2.71955699e-01 5.59241533e-01
-1.31592953e+00 -1.33874878e-01 6.46293759e-01 8.85966063e-01
5.63834906e-01 -2.97831260e-02 7.52772868e-01 -1.22442865e+00
4.52027529e-01 -4.35726374e-01 -6.32521093e-01 4.15040404e-01
-8.00290585e-01 -1.43325672e-01 7.79806376e-01 -6.64700449e-01
-1.20319009e+00 2.71076769e-01 2.88600564e-01 -5.88906288e-01
-3.72887135e-01 1.71101138e-01 -5.51093519e-01 -8.19533244e-02
9.79087114e-01 2.19716132e-01 1.95200503e-01 -5.45306325e-01
2.34264016e-01 1.06545866e+00 5.68682492e-01 -5.36103785e-01
1.05178726e+00 5.60593903e-01 -3.37974042e-01 -4.69111532e-01
-7.38418758e-01 -2.81915724e-01 -8.06184411e-01 1.33027613e-01
6.16180301e-01 -9.02000666e-01 2.68688679e-01 6.49583340e-01
-1.21261239e+00 -2.41774946e-01 -4.75026906e-01 2.86496878e-01
-4.41661179e-01 2.13909820e-01 -9.91155431e-02 -3.78162861e-01
-9.43415686e-02 -1.01447737e+00 8.79909098e-01 1.37483582e-01
-6.81652874e-02 -8.77098858e-01 2.72589792e-02 4.17356551e-01
6.04291201e-01 3.87840897e-01 7.67144859e-01 -9.16669607e-01
-4.46918041e-01 -4.25851673e-01 -2.19933778e-01 9.51855481e-01
4.01163310e-01 -2.40621820e-01 -1.17361093e+00 -6.32328331e-01
4.67301816e-01 -2.32306615e-01 3.55983824e-01 -1.18279561e-01
9.10840690e-01 -5.17116129e-01 -2.50108153e-01 4.61810499e-01
1.44121897e+00 2.34231070e-01 7.14689732e-01 5.06162286e-01
4.41215664e-01 8.13311219e-01 8.38036299e-01 3.05931062e-01
1.50916994e-01 7.37809002e-01 1.40000969e-01 -3.21380228e-01
-2.00580910e-01 -4.54789609e-01 5.48470557e-01 4.15454388e-01
4.49692130e-01 -3.35856616e-01 -7.65328050e-01 6.76853299e-01
-1.59816778e+00 -7.48435795e-01 1.42450541e-01 2.55931950e+00
1.00771773e+00 -9.71015394e-02 8.40420574e-02 -1.99237451e-01
1.07077479e+00 -1.78964332e-01 -7.32742488e-01 -3.89798969e-01
-3.64029706e-01 1.66403994e-01 7.05989838e-01 6.50458634e-02
-1.07165647e+00 8.00952315e-01 6.22542334e+00 6.93840981e-01
-1.52089214e+00 5.62787473e-01 6.78500295e-01 -1.11631401e-01
-2.38552764e-01 1.49548039e-01 -7.32262313e-01 5.12008965e-01
1.21241748e+00 -5.09198070e-01 2.98665047e-01 1.10346067e+00
-1.10591464e-01 1.95278063e-01 -1.42651927e+00 9.23045158e-01
3.42860520e-01 -1.22182643e+00 1.91174358e-01 3.77774954e-01
9.17236626e-01 -1.42004564e-01 3.98874611e-01 8.66424069e-02
6.26001656e-02 -7.27163672e-01 9.44344282e-01 1.25423402e-01
1.30156851e+00 -5.42516112e-01 7.50404179e-01 3.48800540e-01
-4.95261937e-01 1.00349495e-02 -6.23647988e-01 2.05928147e-01
-1.31754309e-01 5.14189482e-01 -1.01468849e+00 5.49226224e-01
7.35713065e-01 4.87277657e-01 -7.64303207e-01 8.09492528e-01
-2.06161052e-01 4.88731027e-01 -1.83656752e-01 4.84253168e-01
-1.68245181e-01 -5.55458851e-02 3.24722350e-01 8.06433856e-01
5.50236583e-01 -2.95637339e-01 -2.15684310e-01 9.55537438e-01
-4.77915943e-01 1.91257864e-01 -9.40210640e-01 -2.42138013e-01
4.41234529e-01 1.18631184e+00 -4.81953204e-01 -4.35879439e-01
-4.16483879e-01 1.32908010e+00 7.00844154e-02 4.89399672e-01
-8.49975288e-01 -3.72594416e-01 6.54983103e-01 2.69530416e-01
2.56801933e-01 3.93232927e-02 -3.49189132e-01 -1.32637823e+00
2.96320677e-01 -1.00991166e+00 2.94165850e-01 -6.91337109e-01
-1.38918817e+00 8.67485166e-01 -7.00014904e-02 -1.72963834e+00
-1.38653204e-01 -4.35696214e-01 -1.13751419e-01 8.66870642e-01
-1.60426617e+00 -1.33777475e+00 -2.32815117e-01 8.49856615e-01
2.48549566e-01 -3.57941359e-01 9.31843996e-01 4.80176479e-01
-3.72717261e-01 1.16139948e+00 5.00103891e-01 2.41638809e-01
1.38787627e+00 -7.36664295e-01 2.37268656e-01 1.16788387e+00
1.19543791e-01 6.53801203e-01 5.41467011e-01 -7.19164729e-01
-1.01839268e+00 -1.52403319e+00 1.02245235e+00 -6.76344812e-01
3.57828498e-01 -6.33274138e-01 -9.88110542e-01 8.17281961e-01
1.55295253e-01 1.60443798e-01 9.96012330e-01 -4.26436961e-01
-6.95170403e-01 -2.68782109e-01 -1.69836736e+00 2.87892163e-01
7.44468749e-01 -5.91877699e-01 -4.29182291e-01 4.01450247e-01
8.67610753e-01 -2.99282640e-01 -7.86118746e-01 -4.99969758e-02
2.43750498e-01 -6.57988489e-01 7.20411777e-01 -6.39654517e-01
4.16321456e-01 -4.35657680e-01 -2.44076267e-01 -1.20902503e+00
-5.26466332e-02 -4.33964550e-01 4.68179286e-01 1.74169219e+00
5.71773887e-01 -9.36700284e-01 8.53571415e-01 1.08779871e+00
2.35764444e-01 1.21577434e-01 -9.86734450e-01 -1.32105136e+00
3.35591316e-01 -5.31087145e-02 1.01368320e+00 1.25530183e+00
-3.35787326e-01 2.13035911e-01 -5.91221392e-01 2.73324579e-01
5.94480753e-01 4.53802571e-02 1.08716786e+00 -1.04634261e+00
-9.09963250e-02 1.16169818e-01 -2.91796595e-01 -5.73934376e-01
2.06031963e-01 -1.02906704e+00 -4.17175405e-02 -1.04043257e+00
2.03562722e-01 -6.72672212e-01 -3.82123560e-01 6.78924859e-01
2.27795407e-01 4.28783983e-01 2.74484336e-01 6.24428153e-01
-1.73908651e-01 2.97525108e-01 9.36254621e-01 -2.25262433e-01
-1.12603813e-01 -3.75283770e-02 -9.21987891e-01 3.04453522e-01
9.23088372e-01 -1.05756807e+00 -5.02474427e-01 -6.11396730e-01
-2.09181830e-01 -4.20107543e-01 3.12323213e-01 -9.19082642e-01
1.46423250e-01 -3.27090353e-01 2.89728522e-01 -2.31320225e-02
1.09503172e-01 -1.13699460e+00 5.08844435e-01 1.49134696e-01
-4.29164290e-01 -2.97300011e-01 7.07191676e-02 5.22044420e-01
-3.38640779e-01 -3.58929753e-01 9.85850930e-01 1.37616917e-01
-6.23394847e-01 1.84769437e-01 -1.16275035e-01 -2.21134961e-01
1.31198716e+00 -4.08246815e-01 -5.18785834e-01 -2.74878114e-01
-5.39799571e-01 -3.12418103e-01 1.11429906e+00 5.86176693e-01
4.59642053e-01 -1.53607607e+00 -6.59138680e-01 4.93953526e-01
5.39070308e-01 -3.10447235e-02 1.52642742e-01 5.25164545e-01
-1.54768988e-01 2.22776696e-01 -5.22554040e-01 -4.64308470e-01
-1.17912245e+00 1.00932813e+00 -6.31373236e-03 -7.78513029e-02
-2.14515537e-01 7.08853841e-01 4.96799260e-01 -5.27527869e-01
-4.39534783e-02 1.76079012e-02 1.88754886e-01 7.47442022e-02
5.59981883e-01 -8.51272866e-02 2.05567166e-01 -8.71603727e-01
-4.16683406e-01 2.82334566e-01 -2.94384331e-01 -2.14759782e-01
1.29946375e+00 -2.76337087e-01 -2.26878121e-01 9.75207761e-02
1.41407931e+00 1.36073723e-01 -1.22449505e+00 -5.50843120e-01
-5.94107360e-02 -8.75784516e-01 -3.91278416e-01 -8.71002078e-01
-1.10030663e+00 8.53863478e-01 7.82105088e-01 -2.19848510e-02
1.34113300e+00 2.53148526e-02 7.80030429e-01 1.96430460e-01
6.33953929e-01 -1.20483851e+00 1.04647495e-01 1.43610343e-01
8.96728694e-01 -1.19482493e+00 -2.80655831e-01 -3.46351355e-01
-9.32641566e-01 8.29261541e-01 6.41477942e-01 1.95710257e-01
3.74097198e-01 -2.57261395e-02 3.54654729e-01 3.19316208e-01
-3.97975683e-01 1.89912662e-01 -6.25523850e-02 1.02466202e+00
-8.45596194e-02 -8.06611404e-02 -5.92964627e-02 7.48210788e-01
-9.35955718e-02 2.27288231e-01 8.60813856e-01 1.08638906e+00
4.13283259e-02 -1.75353503e+00 -7.08210349e-01 2.14972422e-01
-4.23916012e-01 1.08218223e-01 -5.90900540e-01 5.12733102e-01
2.57744312e-01 9.86136079e-01 -4.21217196e-02 -4.70403016e-01
3.52348894e-01 2.11360589e-01 2.06117138e-01 -7.31535316e-01
-4.25026208e-01 -6.36925176e-02 -1.71477422e-01 -2.95429707e-01
-4.96650845e-01 -6.92732155e-01 -8.34412456e-01 -1.84875429e-01
-2.67580003e-01 1.79106653e-01 9.93222654e-01 5.73221266e-01
7.42160916e-01 -9.26470309e-02 9.03053641e-01 -5.30595005e-01
-6.52425706e-01 -5.71992815e-01 -5.56055605e-01 7.65030980e-01
2.42804244e-01 -4.46247160e-01 -3.27836454e-01 5.68025947e-01]
|
[10.337627410888672, 3.1173887252807617]
|
928e6a2b-5795-498d-adcd-e4929825a34a
|
affordance-transfer-learning-for-human-object
|
2104.02867
| null |
https://arxiv.org/abs/2104.02867v2
|
https://arxiv.org/pdf/2104.02867v2.pdf
|
Affordance Transfer Learning for Human-Object Interaction Detection
|
Reasoning the human-object interactions (HOI) is essential for deeper scene understanding, while object affordances (or functionalities) are of great importance for human to discover unseen HOIs with novel objects. Inspired by this, we introduce an affordance transfer learning approach to jointly detect HOIs with novel objects and recognize affordances. Specifically, HOI representations can be decoupled into a combination of affordance and object representations, making it possible to compose novel interactions by combining affordance representations and novel object representations from additional images, i.e. transferring the affordance to novel objects. With the proposed affordance transfer learning, the model is also capable of inferring the affordances of novel objects from known affordance representations. The proposed method can thus be used to 1) improve the performance of HOI detection, especially for the HOIs with unseen objects; and 2) infer the affordances of novel objects. Experimental results on two datasets, HICO-DET and HOI-COCO (from V-COCO), demonstrate significant improvements over recent state-of-the-art methods for HOI detection and object affordance detection. Code is available at https://github.com/zhihou7/HOI-CL
|
['DaCheng Tao', 'Xiaojiang Peng', 'Yu Qiao', 'Baosheng Yu', 'Zhi Hou']
|
2021-04-07
| null |
http://openaccess.thecvf.com//content/CVPR2021/html/Hou_Affordance_Transfer_Learning_for_Human-Object_Interaction_Detection_CVPR_2021_paper.html
|
http://openaccess.thecvf.com//content/CVPR2021/papers/Hou_Affordance_Transfer_Learning_for_Human-Object_Interaction_Detection_CVPR_2021_paper.pdf
|
cvpr-2021-1
|
['affordance-recognition', 'human-object-interaction-concept-discovery', 'affordance-detection']
|
['computer-vision', 'computer-vision', 'computer-vision']
|
[ 1.77654088e-01 -1.70744415e-02 4.35477346e-02 -2.09537849e-01
-1.62445262e-01 -4.94331717e-01 5.14413178e-01 2.25310400e-01
-9.08178017e-02 3.75940502e-01 1.17653228e-01 7.88914561e-02
-2.54806995e-01 -7.16082633e-01 -9.30424631e-01 -4.91919577e-01
-2.66257912e-01 4.28860396e-01 4.47148353e-01 -1.52863517e-01
2.75695920e-01 8.96968544e-01 -2.03283310e+00 2.94410169e-01
7.69813895e-01 9.05748665e-01 5.81673443e-01 7.14954197e-01
2.28697583e-01 4.70264226e-01 -3.57247889e-01 2.44804919e-01
2.52556115e-01 -6.79690167e-02 -6.73086703e-01 3.16216111e-01
6.48049831e-01 -6.71715319e-01 -4.19013083e-01 9.02465284e-01
1.57257199e-01 5.21904290e-01 7.62115002e-01 -1.41813350e+00
-1.20683873e+00 3.62840801e-01 -4.18536365e-01 4.02463108e-01
7.31642187e-01 3.78241897e-01 1.15168393e+00 -1.13492167e+00
7.18105018e-01 1.57519722e+00 -1.07123055e-01 5.93910158e-01
-1.15370405e+00 -6.65606618e-01 2.46782169e-01 4.65754300e-01
-1.39985502e+00 -3.58164698e-01 7.78642774e-01 -4.95583892e-01
1.15170383e+00 4.46224749e-01 8.72472763e-01 8.67900789e-01
-3.12640488e-01 1.58418477e+00 6.76223159e-01 -5.13878286e-01
-3.33876573e-02 -4.32157591e-02 1.14304036e-01 8.06630373e-01
3.50008100e-01 2.30451569e-01 -4.05782372e-01 4.51570749e-02
9.50318933e-01 6.54665053e-01 -4.40209955e-01 -4.48140413e-01
-1.85084116e+00 5.15155315e-01 1.23834312e+00 4.13973033e-01
-2.26923868e-01 1.69617176e-01 3.12685687e-03 1.55068681e-01
2.13104449e-02 6.33133411e-01 -2.97331929e-01 3.69883627e-01
-4.76376573e-03 7.48442054e-01 3.58617991e-01 1.52773130e+00
9.60346997e-01 -3.23549151e-01 -3.05497646e-01 4.24258083e-01
2.83660144e-01 4.97372746e-01 3.81900072e-01 -6.04463518e-01
3.55908722e-01 9.00344193e-01 3.45430225e-01 -7.89170861e-01
-4.66772646e-01 4.54191230e-02 -3.48304629e-01 1.98484614e-01
3.00979376e-01 4.14356738e-01 -9.91403162e-01 1.46413243e+00
7.01340318e-01 4.28530499e-02 -9.59700793e-02 1.19181859e+00
1.12485480e+00 6.05258107e-01 7.81373214e-03 3.45574021e-01
1.65341771e+00 -1.05189478e+00 -5.11401772e-01 -8.10795352e-02
6.54933691e-01 -5.41259348e-01 1.43590176e+00 1.42848089e-01
-5.55691361e-01 -8.19763303e-01 -1.17400515e+00 -4.48403239e-01
-5.98531902e-01 3.82294476e-01 1.05509543e+00 2.14950051e-02
-3.95045757e-01 5.32756746e-01 -7.95198560e-01 -5.22871971e-01
8.35036039e-01 2.37039804e-01 -4.38906401e-01 -1.71503261e-01
-7.30027080e-01 6.54731870e-01 1.03291452e+00 1.65979207e-01
-1.15853274e+00 -5.18522143e-01 -1.20368254e+00 2.01379672e-01
9.28196549e-01 -7.72091627e-01 9.98556197e-01 -5.74654281e-01
-8.77650380e-01 6.10687733e-01 -7.73827583e-02 -1.93130281e-02
3.04348916e-01 -4.84542727e-01 -4.10522521e-01 6.38109818e-02
3.84417415e-01 8.41311574e-01 1.10855711e+00 -1.40317273e+00
-8.30954015e-01 -5.27749896e-01 4.14051324e-01 2.87938088e-01
-9.10184458e-02 -4.46612462e-02 3.66376787e-02 -5.56176543e-01
5.44882417e-01 -1.01433206e+00 1.79036066e-01 5.51922083e-01
-6.42615020e-01 -9.12559688e-01 9.93206978e-01 -3.13412577e-01
7.05212355e-01 -2.24559808e+00 2.90743440e-01 -2.43454188e-01
5.17238081e-01 1.91328317e-01 -2.11432412e-01 3.99020463e-01
-1.12607114e-01 -8.25890079e-02 5.29704392e-02 1.15198754e-01
4.72628362e-02 3.51311952e-01 -3.18525612e-01 3.77896279e-01
5.90498924e-01 1.27963161e+00 -1.24015439e+00 -1.32439077e-01
4.96788412e-01 1.77893847e-01 -5.01658976e-01 5.74186921e-01
-4.24579471e-01 8.07567954e-01 -6.09526634e-01 9.39308465e-01
4.31875378e-01 -1.37835220e-01 -1.99444950e-01 -2.50595808e-01
-4.24836352e-02 2.21968994e-01 -1.27244318e+00 1.68461096e+00
-1.54103652e-01 4.83350486e-01 -5.05980015e-01 -6.59022927e-01
7.61397839e-01 1.84590474e-01 6.04636595e-03 -2.92746127e-01
1.19038582e-01 3.02377284e-01 2.46116787e-01 -9.48098004e-01
1.62267581e-01 2.46537719e-02 -3.53724742e-03 2.92529851e-01
3.67808670e-01 -2.40673512e-01 3.31415802e-01 2.23727673e-02
6.26563132e-01 3.77072901e-01 7.99883306e-01 -1.00748017e-01
4.48797911e-01 -3.41238946e-01 3.48005772e-01 7.44381368e-01
-2.57397830e-01 4.25231636e-01 1.50939198e-02 -8.61595511e-01
-8.50108325e-01 -1.46455956e+00 -1.96082458e-01 1.34763753e+00
3.74403358e-01 -5.96451871e-02 -3.48602206e-01 -5.58154821e-01
3.98474813e-01 7.37047017e-01 -8.66655529e-01 -9.13656205e-02
-4.89611715e-01 -1.50225386e-01 -9.21881050e-02 7.85703242e-01
2.24506557e-01 -1.72904027e+00 -1.18381846e+00 -1.16019659e-02
-2.21301075e-02 -1.11920691e+00 -5.31331122e-01 -3.62956040e-02
-7.30573654e-01 -1.07205784e+00 -7.10647583e-01 -8.19068491e-01
8.16036463e-01 7.28139937e-01 6.30007386e-01 4.83520091e-01
-9.68740642e-01 4.12618190e-01 -5.64164460e-01 -6.90133691e-01
-1.53729111e-01 -1.40074221e-02 3.18486899e-01 2.52701908e-01
4.56454396e-01 -6.31800532e-01 -5.69090366e-01 1.75948024e-01
-1.11550212e+00 2.96963066e-01 5.27239978e-01 8.42063785e-01
5.56013703e-01 -2.91303813e-01 3.73289883e-01 -5.35195589e-01
1.01966031e-01 -2.88961709e-01 -4.81763780e-01 2.79668957e-01
-8.94216895e-02 1.88279673e-01 2.15492278e-01 -8.55152845e-01
-1.09739184e+00 5.60521595e-02 3.16884279e-01 -4.62991357e-01
-7.36692131e-01 1.02087274e-01 -3.78576487e-01 2.96083489e-03
7.67954826e-01 -3.30777429e-02 -4.68798071e-01 -7.34245658e-01
7.51711607e-01 6.26493692e-01 4.31851506e-01 -7.60969162e-01
8.62784624e-01 5.92861474e-01 -2.21451148e-01 -1.07534552e+00
-1.01130629e+00 -8.49850178e-01 -1.34269869e+00 -1.51292875e-01
1.02681339e+00 -7.70016730e-01 -8.75298619e-01 2.00439215e-01
-1.21887100e+00 3.88013572e-02 -4.85683560e-01 5.47484756e-01
-5.73962510e-01 2.74312109e-01 -4.11467314e-01 -7.67218888e-01
-8.36732704e-03 -1.10737598e+00 1.35751104e+00 5.16627371e-01
-2.68730402e-01 -4.94888425e-01 -4.14612949e-01 3.76358837e-01
-2.27317408e-01 5.00262141e-01 1.03001416e+00 -5.48087597e-01
-1.11084533e+00 -2.87215769e-01 -5.99624276e-01 -1.75372213e-02
4.05773401e-01 -3.07509542e-01 -1.12926018e+00 -3.80540937e-01
-1.98460311e-01 -5.14108360e-01 7.23367989e-01 -7.59938434e-02
9.30570960e-01 -2.51347542e-01 -5.45512676e-01 5.14545858e-01
1.14985025e+00 1.67707369e-01 4.31466848e-01 1.78323463e-01
1.16787934e+00 6.32159591e-01 8.08151245e-01 5.12435377e-01
3.04855496e-01 6.01489484e-01 5.65801561e-01 2.48940632e-01
-3.29895973e-01 -4.91668016e-01 2.97493488e-02 4.44209367e-01
-4.76898611e-01 2.55624592e-01 -8.14132094e-01 5.70833087e-01
-1.97050107e+00 -5.74789107e-01 -1.58373982e-01 1.95850503e+00
3.45509291e-01 -6.91149607e-02 6.08420260e-02 7.26941153e-02
4.45750654e-01 -9.65475142e-02 -9.90330279e-01 8.23017657e-02
1.12426877e-01 -1.09891206e-01 -6.10884838e-02 6.94925487e-02
-1.12775624e+00 1.02932596e+00 4.90639973e+00 3.56296659e-01
-6.18485093e-01 1.96444377e-01 -3.01371396e-01 6.10695630e-02
-1.33253694e-01 8.45240504e-02 -7.14471579e-01 -2.22832188e-01
-1.24113075e-02 -7.91597553e-03 7.94416070e-01 1.05412567e+00
-3.80202234e-01 -1.04313992e-01 -1.66965830e+00 1.09939790e+00
2.70346314e-01 -9.68173265e-01 3.59812677e-01 -3.63965109e-02
6.29659355e-01 -1.83396250e-01 -1.21884327e-02 4.31189299e-01
-1.70324117e-01 -7.46173620e-01 8.84565353e-01 5.07990241e-01
5.98421216e-01 -2.20558524e-01 2.67254025e-01 5.41914344e-01
-1.43999326e+00 -4.00047898e-01 -7.07274854e-01 -4.91795480e-01
-1.57138988e-01 -1.37684145e-03 -9.48687792e-01 5.11557281e-01
9.79692519e-01 1.01767838e+00 -7.99749196e-01 1.00490284e+00
-6.71534956e-01 -1.53059706e-01 -3.59448582e-01 -2.81832159e-01
4.01931331e-02 1.65504031e-02 8.55285645e-01 6.87607288e-01
1.95177808e-01 2.03371823e-01 3.27114075e-01 1.47395337e+00
3.99479531e-02 1.20573826e-01 -7.13461399e-01 -1.81172132e-01
2.13287830e-01 1.16035533e+00 -7.92170227e-01 -4.77973908e-01
-4.20648783e-01 1.13575578e+00 7.39514112e-01 5.62581182e-01
-5.36566257e-01 -3.79011393e-01 7.62398839e-01 6.59891292e-02
5.53596437e-01 -3.49400192e-01 2.62157947e-01 -1.49680865e+00
1.90646946e-01 -5.40647626e-01 5.53508520e-01 -1.11140585e+00
-1.31378412e+00 3.14730763e-01 3.89580876e-01 -1.38710463e+00
1.24525458e-01 -1.10647213e+00 -2.52702981e-01 5.79418957e-01
-1.56535304e+00 -1.63259768e+00 -5.14053285e-01 5.05640447e-01
8.27881932e-01 9.94671434e-02 7.78219342e-01 -5.56267276e-02
-1.50531530e-01 2.44795665e-01 -4.04221982e-01 4.43439791e-03
3.47306341e-01 -1.15927124e+00 4.20963049e-01 5.44839323e-01
6.14142120e-01 7.67873764e-01 4.61628199e-01 -5.87512910e-01
-1.51609814e+00 -8.60289395e-01 3.94305587e-01 -1.08174217e+00
5.80747366e-01 -7.57236898e-01 -1.08620250e+00 1.00526452e+00
-2.99199820e-01 4.39322531e-01 5.45499682e-01 1.04016855e-01
-6.52230680e-01 1.97547674e-01 -8.27736676e-01 8.30031514e-01
1.55231452e+00 -6.73476458e-01 -1.04260778e+00 5.10016024e-01
1.04536784e+00 -4.44739282e-01 -8.12231243e-01 4.99434292e-01
8.44637752e-01 -8.46566260e-01 1.13470185e+00 -9.01558399e-01
-8.25020149e-02 -6.44553483e-01 -2.14752167e-01 -8.81576657e-01
-5.47135115e-01 -2.11425424e-01 -5.14369309e-01 6.61435604e-01
1.18678816e-01 -5.12447119e-01 1.07769512e-01 3.67191106e-01
-8.75960067e-02 -4.90239620e-01 -7.27654874e-01 -8.74957919e-01
-4.41110760e-01 -4.39330071e-01 9.77981806e-01 7.70727158e-01
7.00405315e-02 4.94293153e-01 -2.03260675e-01 3.55506361e-01
6.61041737e-01 6.15405560e-01 1.02795911e+00 -1.35224032e+00
-4.62845445e-01 -3.61836255e-01 -7.32253969e-01 -1.26871550e+00
7.49084428e-02 -1.04158139e+00 1.82247877e-01 -1.39297366e+00
3.51315826e-01 -2.93780208e-01 -4.36842918e-01 8.46491873e-01
-4.88974482e-01 1.28827184e-01 6.16460145e-01 4.69278038e-01
-6.10401213e-01 7.68146574e-01 1.91070879e+00 -3.82833451e-01
-4.02493864e-01 -9.84348655e-02 -3.94839793e-01 7.23435283e-01
4.48006392e-01 -4.45968866e-01 -2.69618720e-01 -4.24088299e-01
2.84368508e-02 -2.05537573e-01 7.34442651e-01 -1.08934665e+00
-2.10031033e-01 -7.63060823e-02 5.15370071e-01 -6.06266677e-01
2.23067969e-01 -8.83099616e-01 -2.48123735e-01 6.72112703e-01
-1.90763250e-01 -4.49930608e-01 2.06101075e-01 7.64165342e-01
1.52998909e-01 -2.75940508e-01 2.94040680e-01 -2.54750192e-01
-1.26050627e+00 4.56271201e-01 -1.23640716e-01 -3.49463612e-01
1.04648423e+00 -1.20675251e-01 -1.92556888e-01 7.01291636e-02
-1.01671398e+00 1.74186260e-01 2.56599218e-01 1.10204768e+00
9.86169517e-01 -1.41015553e+00 -5.35433054e-01 6.48762822e-01
9.91992772e-01 2.95629174e-01 2.71028250e-01 5.20714164e-01
-3.90493900e-01 2.56316066e-01 -3.58865678e-01 -8.69665325e-01
-1.08278537e+00 1.28162861e+00 -8.32356662e-02 4.21675742e-01
-1.09300053e+00 8.20012689e-01 8.22399616e-01 -7.91003466e-01
1.66499645e-01 -9.87368107e-01 -1.86651900e-01 -3.51551265e-01
6.53503954e-01 2.91939318e-01 -5.50245941e-01 -7.31644511e-01
-1.92617282e-01 4.96587723e-01 -1.63954776e-02 4.12314594e-01
1.11935031e+00 -1.70369998e-01 -1.61317840e-01 9.45730627e-01
1.11458898e+00 -7.05415010e-01 -1.34251451e+00 -5.14989734e-01
1.13279350e-01 -7.58847773e-01 -4.17468101e-01 -5.19759715e-01
-3.73009086e-01 1.20855975e+00 6.44961476e-01 -1.17020532e-01
8.05933595e-01 6.43800318e-01 6.19284928e-01 9.31976080e-01
6.47753417e-01 -4.52276647e-01 5.33806145e-01 1.79543257e-01
1.38022542e+00 -1.56135428e+00 5.69110794e-04 -6.96152091e-01
-4.21144664e-01 1.17421913e+00 6.76337063e-01 -1.70731738e-01
6.19722366e-01 -3.34594727e-01 -2.96355635e-01 -4.92097169e-01
-1.97866902e-01 -5.36816716e-01 8.61385942e-01 7.30357111e-01
1.04696602e-01 5.27346909e-01 2.52229661e-01 4.63898182e-01
4.29206230e-02 -1.66318163e-01 1.97948426e-01 9.16625857e-01
-5.20571947e-01 -5.25828421e-01 -3.01723778e-01 4.28099871e-01
3.88673872e-01 9.91136730e-02 -1.09195501e-01 7.98807323e-01
3.88059646e-01 6.03896856e-01 -1.07916668e-02 -7.38427192e-02
6.45158350e-01 4.31023687e-02 9.34419453e-01 -1.14826500e+00
1.16149768e-01 -1.86479259e-02 -3.34787071e-01 -6.38949633e-01
-4.09556359e-01 -7.30339527e-01 -1.39665711e+00 4.41877604e-01
-6.28994107e-01 -3.23394120e-01 3.68987858e-01 1.08794355e+00
1.71161920e-01 7.08056450e-01 2.28429228e-01 -1.44705427e+00
-3.87382448e-01 -1.17046952e+00 -7.68615007e-01 8.44837070e-01
5.49852431e-01 -1.19472551e+00 -1.73210919e-01 7.48145804e-02]
|
[5.167304039001465, -0.11266295611858368]
|
f7f8d9ac-4fd1-4e4f-8e4f-ebf0f3c82ca3
|
stochastic-parallelizable-eigengap-dilation
|
2207.14589
| null |
https://arxiv.org/abs/2207.14589v1
|
https://arxiv.org/pdf/2207.14589v1.pdf
|
Stochastic Parallelizable Eigengap Dilation for Large Graph Clustering
|
Large graphs commonly appear in social networks, knowledge graphs, recommender systems, life sciences, and decision making problems. Summarizing large graphs by their high level properties is helpful in solving problems in these settings. In spectral clustering, we aim to identify clusters of nodes where most edges fall within clusters and only few edges fall between clusters. This task is important for many downstream applications and exploratory analysis. A core step of spectral clustering is performing an eigendecomposition of the corresponding graph Laplacian matrix (or equivalently, a singular value decomposition, SVD, of the incidence matrix). The convergence of iterative singular value decomposition approaches depends on the eigengaps of the spectrum of the given matrix, i.e., the difference between consecutive eigenvalues. For a graph Laplacian corresponding to a well-clustered graph, the eigenvalues will be non-negative but very small (much less than $1$) slowing convergence. This paper introduces a parallelizable approach to dilating the spectrum in order to accelerate SVD solvers and in turn, spectral clustering. This is accomplished via polynomial approximations to matrix operations that favorably transform the spectrum of a matrix without changing its eigenvectors. Experiments demonstrate that this approach significantly accelerates convergence, and we explain how this transformation can be parallelized and stochastically approximated to scale with available compute.
|
['Richard Everett', 'Yoram Bachrach', 'Ian Gemp', 'Elise van der Pol']
|
2022-07-29
| null | null | null | null |
['graph-clustering']
|
['graphs']
|
[-3.46047916e-02 1.45096213e-01 1.93594489e-02 2.27151021e-01
-1.63414583e-01 -9.32748854e-01 3.80028784e-02 3.89283776e-01
-6.13796599e-02 2.59725362e-01 1.30942971e-01 -5.89887738e-01
-4.31864887e-01 -7.61311650e-01 -3.72875184e-01 -7.24828720e-01
-5.45630991e-01 4.59321141e-01 7.80993775e-02 -1.62655145e-01
-1.85506772e-02 5.40152431e-01 -9.58836019e-01 -3.89868878e-02
9.21846628e-01 4.06981230e-01 -1.17193855e-01 6.44141614e-01
-2.48627707e-01 3.57225299e-01 -1.85066342e-01 -2.41391167e-01
4.18915570e-01 -5.49079239e-01 -8.67383718e-01 3.87134969e-01
-1.00542426e-01 4.11401629e-01 -1.37289420e-01 1.38505733e+00
1.63267657e-01 3.11799973e-01 7.31036961e-01 -1.39648116e+00
-2.80834228e-01 7.80614197e-01 -1.09480202e+00 4.84397262e-02
3.94937128e-01 -2.53519088e-01 1.20742297e+00 -7.40138888e-01
6.43775523e-01 1.14292455e+00 9.85764921e-01 3.86027992e-02
-1.78436935e+00 -2.72008300e-01 1.96462292e-02 -6.81284145e-02
-1.74084747e+00 -1.96986318e-01 7.24046171e-01 -5.56388319e-01
6.73500061e-01 5.09848297e-01 8.93262029e-01 3.17956984e-01
-1.45357803e-01 1.55169204e-01 5.84107578e-01 -3.77203584e-01
3.92779291e-01 3.62312011e-02 3.22307914e-01 6.66598678e-01
5.13642430e-01 -5.92213213e-01 -1.42825559e-01 -6.40422165e-01
5.00904262e-01 4.78844084e-02 -2.24864334e-01 -7.24999011e-01
-1.30806315e+00 9.00335371e-01 4.52713519e-01 3.64205569e-01
-5.14699280e-01 1.27810404e-01 2.77836829e-01 4.56374943e-01
2.23801658e-01 3.29549223e-01 3.78118269e-02 9.64914039e-02
-8.33870053e-01 -1.12483434e-01 1.20512354e+00 7.43672490e-01
1.07024336e+00 -2.36150011e-01 3.30357641e-01 8.07667792e-01
6.73883334e-02 4.86104429e-01 1.25219628e-01 -9.87495005e-01
1.69376075e-01 9.57090616e-01 -1.20684437e-01 -1.43001151e+00
-7.29930401e-01 -3.03036720e-01 -1.24791229e+00 -3.13666493e-01
6.46370232e-01 -4.03251737e-01 -4.49694067e-01 1.61683953e+00
6.04178667e-01 2.98676878e-01 -1.42330438e-01 9.29585159e-01
3.47361267e-01 7.06189394e-01 -1.69399917e-01 -6.28950834e-01
1.17917860e+00 -4.79777664e-01 -4.06437516e-01 -2.10601445e-02
7.23263264e-01 -7.99584091e-01 9.48755920e-01 1.40953839e-01
-8.42884302e-01 1.00795664e-01 -7.93418467e-01 3.53154749e-01
4.65237312e-02 2.31133401e-03 5.24939477e-01 6.50475681e-01
-1.32060063e+00 5.73852777e-01 -8.55775952e-01 -4.66422111e-01
-6.35806993e-02 5.07659018e-01 -2.67058700e-01 1.20651580e-01
-8.34343076e-01 3.37994009e-01 2.91746229e-01 -3.98563774e-04
7.87703395e-02 -7.15246975e-01 -5.15277088e-01 2.78128386e-01
5.31873941e-01 -5.36817193e-01 5.46630144e-01 -1.26570427e+00
-1.02106786e+00 6.54064119e-01 -2.04167783e-01 -2.38331795e-01
2.88628638e-01 5.22266686e-01 -2.68335283e-01 1.84719265e-01
1.18782990e-01 -6.56761527e-02 9.32141185e-01 -1.04076242e+00
-1.10210761e-01 -4.26293790e-01 -2.03367561e-01 2.32900769e-01
-6.98756754e-01 -1.13087304e-01 -6.07193708e-01 -2.70115584e-01
5.96418321e-01 -1.29312658e+00 -7.68456101e-01 -4.20622736e-01
-4.72733498e-01 -1.33814812e-01 5.37173748e-01 -4.97256964e-01
1.42044401e+00 -2.30656266e+00 4.29103822e-01 1.29041374e+00
6.99984312e-01 -2.08713144e-01 -8.38966221e-02 9.28847551e-01
-2.89674133e-01 1.11062698e-01 -1.93652585e-01 2.56505877e-01
-1.83287755e-01 -1.26824930e-01 -4.42659184e-02 8.35799336e-01
-2.20966876e-01 4.13662881e-01 -1.04971099e+00 -2.55517036e-01
-6.86503053e-02 2.64531821e-01 -6.44000828e-01 -2.13368371e-01
1.47793934e-01 -4.85443138e-02 -2.96643227e-01 1.15713105e-01
6.44642055e-01 -6.24691665e-01 9.76448655e-01 -4.31180000e-01
-1.35445699e-01 -1.53668046e-01 -1.84534621e+00 1.11632109e+00
-1.82929173e-01 5.58628082e-01 5.50833702e-01 -1.19795048e+00
6.71473622e-01 1.01616636e-01 1.07078362e+00 1.07254900e-01
4.35667783e-02 6.76832795e-02 2.99931377e-01 5.77228442e-02
3.58703464e-01 8.91837031e-02 2.28443921e-01 7.19762266e-01
-4.38501179e-01 9.77508724e-02 6.71189368e-01 8.90662491e-01
1.32627797e+00 -7.59515703e-01 2.44645372e-01 -7.14452982e-01
4.85579103e-01 1.20537162e-01 3.96194756e-01 3.24524462e-01
7.14068189e-02 3.54910314e-01 1.12421942e+00 -1.78369120e-01
-1.05571592e+00 -8.59643757e-01 1.09115794e-01 9.25676942e-01
1.70562744e-01 -9.32780445e-01 -1.11077917e+00 -2.34996825e-01
1.18999019e-01 1.63522020e-01 -3.33673924e-01 -1.55149981e-01
-3.15750539e-01 -9.86495078e-01 -6.99035218e-03 2.05925092e-01
-4.62160110e-02 -3.71313900e-01 -1.29571170e-01 1.14012375e-01
-1.01554498e-01 -8.52135897e-01 -7.80183613e-01 6.91466630e-02
-8.79327595e-01 -1.34039664e+00 -4.31492358e-01 -8.60127985e-01
1.10762501e+00 7.19330430e-01 8.51544738e-01 1.99197248e-01
-2.15192407e-01 4.85030711e-01 -1.69924483e-01 1.05992109e-01
-3.62826437e-01 1.47307277e-01 3.21388781e-01 4.57261890e-01
1.52748436e-01 -8.00063610e-01 -4.77433622e-01 3.76657635e-01
-7.53310442e-01 -9.72817000e-03 3.17232281e-01 6.29632652e-01
6.20828211e-01 3.38221908e-01 2.38526285e-01 -9.76843894e-01
1.17894304e+00 -6.83422446e-01 -6.95069015e-01 3.37464094e-01
-6.14340723e-01 1.81646034e-01 8.15106392e-01 -4.64540124e-01
-3.00846398e-01 6.11874223e-01 3.20815802e-01 -5.54821730e-01
5.74494243e-01 8.41739178e-01 1.80489391e-01 -2.56385267e-01
8.74370396e-01 -3.70628610e-02 3.01432550e-01 -2.25784525e-01
6.47994876e-01 4.94316667e-01 1.88638359e-01 -3.62704605e-01
8.63198102e-01 4.83832866e-01 3.11802626e-01 -1.21529138e+00
-3.26677710e-01 -1.01472127e+00 -6.24085784e-01 -4.32174593e-01
3.65622759e-01 -6.63581371e-01 -1.15519571e+00 -1.99079081e-01
-7.02032804e-01 -1.06568903e-01 -1.85602203e-01 3.29722106e-01
-2.66315371e-01 6.26866877e-01 -5.48675120e-01 -7.62392402e-01
-2.71849364e-01 -7.14021683e-01 4.24414217e-01 -2.51648948e-02
-3.61409247e-01 -1.05990195e+00 2.24415913e-01 2.17025261e-02
1.32673547e-01 2.79725760e-01 9.74229336e-01 -4.62082267e-01
-2.09714651e-01 -2.60481417e-01 -2.22650781e-01 -7.09858835e-02
-2.25214660e-03 2.87504494e-01 -1.30692676e-01 -6.04170084e-01
-3.96736383e-01 3.20570827e-01 5.69828689e-01 4.18385476e-01
6.38661683e-01 -3.32977057e-01 -5.72609007e-01 4.45799738e-01
1.24380136e+00 -1.78950116e-01 7.39865005e-02 -1.64029136e-01
8.64393950e-01 4.85279322e-01 1.11576699e-01 5.74602842e-01
1.50685921e-01 3.97434950e-01 -3.27866785e-02 -1.31672338e-01
1.44745260e-01 -2.50664894e-02 3.36019546e-01 1.13479948e+00
-1.22348323e-01 2.23818198e-01 -1.34230423e+00 4.97825325e-01
-2.16503024e+00 -9.38315690e-01 -7.81617343e-01 2.31082916e+00
6.69832349e-01 -1.23922676e-01 6.02846920e-01 1.88350365e-01
1.07932508e+00 -1.45068049e-01 -6.13183916e-01 -2.74561316e-01
8.68046358e-02 -5.37457205e-02 7.58180022e-01 7.52866864e-01
-7.62290418e-01 7.25588679e-01 6.64325190e+00 5.17296255e-01
-8.90668094e-01 -1.02726467e-01 3.48501533e-01 -2.29845066e-02
-3.34159136e-01 2.00904891e-01 -1.48084328e-01 3.21095437e-01
1.02410734e+00 -6.99688256e-01 9.41343904e-01 8.73945475e-01
3.93767834e-01 -5.35178408e-02 -8.24896336e-01 1.13589919e+00
-3.02514076e-01 -1.24576199e+00 -2.74187297e-01 4.10657197e-01
8.96369576e-01 -2.25919619e-04 -2.06704408e-01 -2.06086710e-01
6.34173274e-01 -6.90562189e-01 1.24510273e-01 1.42528325e-01
5.67559421e-01 -1.11154354e+00 3.64325106e-01 2.17922583e-01
-1.52817535e+00 -9.03965831e-02 -5.38258612e-01 -8.23696237e-03
7.70978769e-03 1.11975873e+00 -9.64562595e-01 2.99579412e-01
6.01520061e-01 7.08797038e-01 -2.50559449e-01 8.77306163e-01
2.15354025e-01 6.77269876e-01 -7.34568894e-01 -1.19111799e-01
2.10042402e-01 -1.13944507e+00 6.02743447e-01 1.00203836e+00
4.00986373e-01 2.99487501e-01 4.43530113e-01 7.51407444e-01
-1.58009157e-01 6.62908256e-01 -5.30107379e-01 -3.48969638e-01
4.97331887e-01 1.60417235e+00 -1.30087984e+00 -1.55416563e-01
-4.83302981e-01 8.41287255e-01 4.57146347e-01 5.20774424e-01
-4.54193354e-01 -5.71987033e-01 8.48699689e-01 4.63463038e-01
9.33882520e-02 -4.39911366e-01 -2.26597637e-01 -1.16372037e+00
-1.00255907e-01 -8.28242898e-01 5.21076500e-01 -2.54266202e-01
-1.18850636e+00 3.46027941e-01 -3.39894146e-01 -9.67318892e-01
-3.26187730e-01 -2.78969646e-01 -5.69125116e-01 7.62893200e-01
-5.89783132e-01 -4.53796089e-01 -1.89238906e-01 9.76871848e-01
-2.37573296e-01 1.75643250e-01 5.59364617e-01 2.56640553e-01
-8.27681422e-01 3.26225370e-01 6.69906497e-01 1.12397127e-01
5.26680946e-01 -1.26685619e+00 2.96651840e-01 9.30407584e-01
2.40658745e-01 8.82969737e-01 8.37332308e-01 -7.59257972e-01
-1.83886218e+00 -8.75167966e-01 5.59394300e-01 7.75064528e-03
1.13869035e+00 -3.13426822e-01 -7.24512339e-01 5.47650695e-01
-6.57069385e-02 5.40816672e-02 8.26473713e-01 5.66693962e-01
-1.53482810e-01 -1.81545228e-01 -8.29408348e-01 8.19563866e-01
1.02066207e+00 -4.07946557e-01 2.72903442e-01 6.93299592e-01
2.35345244e-01 -6.27228320e-02 -1.00741565e+00 -1.66090474e-01
3.38817626e-01 -8.83071065e-01 9.09385741e-01 -6.20097339e-01
-1.63138628e-01 -5.17781675e-01 1.19004168e-01 -1.49196744e+00
-6.53290391e-01 -1.13302922e+00 -1.77663982e-01 8.63923669e-01
4.48965698e-01 -5.79921305e-01 9.86507595e-01 5.66077888e-01
3.49736422e-01 -5.17617404e-01 -5.97132385e-01 -7.52510846e-01
-2.41255417e-01 -8.10277313e-02 2.02038452e-01 1.30065882e+00
7.20419466e-01 7.59251595e-01 -1.43045262e-01 3.27473968e-01
8.22566152e-01 3.70693743e-01 6.69193208e-01 -1.50013137e+00
-2.23317608e-01 -4.94621992e-01 -4.61926758e-01 -7.54306853e-01
1.97082926e-02 -1.13991249e+00 -3.07507396e-01 -1.44111192e+00
3.18799317e-01 -5.35356641e-01 1.11816712e-01 2.85514265e-01
-1.56652287e-01 1.52738884e-01 1.83637410e-01 4.21038151e-01
-5.91224730e-01 -2.43700482e-02 7.81539202e-01 4.13043648e-02
-7.52101600e-01 -6.09235605e-03 -5.51706016e-01 6.28083229e-01
7.09161699e-01 -3.84072840e-01 -5.56538582e-01 1.07934199e-01
7.00458169e-01 1.63978398e-01 -1.04231872e-01 -7.31161892e-01
4.15013015e-01 -1.74090475e-01 -3.80566269e-02 -2.77944803e-01
-4.31212299e-02 -8.46095860e-01 6.94057524e-01 6.44926727e-01
-1.65332839e-01 1.51168525e-01 -1.49028480e-01 7.09103823e-01
-7.98374340e-02 -1.71404272e-01 7.01421618e-01 5.41053228e-02
-2.79135942e-01 1.96445420e-01 -5.86681724e-01 1.43919364e-01
1.16394567e+00 -1.60187334e-01 9.40247774e-02 -6.73057854e-01
-1.01515985e+00 3.77992451e-01 6.24323130e-01 -6.04539178e-02
3.30079615e-01 -1.31358206e+00 -5.66331923e-01 3.41370888e-02
-2.46956468e-01 -3.57075185e-01 4.82192449e-02 1.29671443e+00
-6.77233636e-01 -6.15677051e-02 1.36603802e-01 -6.40451550e-01
-1.57109487e+00 5.59194505e-01 3.53241324e-01 -1.28750563e-01
-5.05934358e-01 5.27416229e-01 -7.76690021e-02 -2.50617802e-01
8.56945142e-02 -5.75139001e-02 7.27180615e-02 3.02001059e-01
3.81245941e-01 6.11397684e-01 4.91183475e-02 -5.24429739e-01
-4.92101640e-01 5.27567089e-01 1.03398181e-01 7.67980097e-03
1.30911231e+00 -1.71747312e-01 -7.77244270e-01 1.61679387e-01
1.27147424e+00 4.03206348e-01 -8.42224836e-01 -1.59580618e-01
1.08718932e-01 -1.35666341e-01 6.99659958e-02 -9.40842461e-03
-1.27199841e+00 3.35298896e-01 9.04586166e-02 9.33865786e-01
1.08728433e+00 9.37176272e-02 4.53903437e-01 7.25172043e-01
4.46062796e-02 -1.15711093e+00 -1.35209382e-01 3.29421997e-01
5.08019447e-01 -8.13903093e-01 1.85602173e-01 -9.32165861e-01
-5.37582874e-01 1.21594524e+00 8.35203230e-02 -2.35577285e-01
9.54924822e-01 1.97525427e-01 -2.13580608e-01 -2.27115408e-01
-4.62028086e-01 -2.45112255e-01 2.14039251e-01 3.65261406e-01
4.40975010e-01 2.64758438e-01 -4.57654715e-01 2.78178602e-01
-2.07336649e-01 -6.56103849e-01 7.05706894e-01 4.97541845e-01
-4.95835483e-01 -9.74363685e-01 -5.54589212e-01 6.72916412e-01
-1.83578748e-02 -1.66156560e-01 -8.44511271e-01 3.08246940e-01
-4.84088749e-01 1.06666160e+00 -2.64700997e-04 -4.20959353e-01
1.54249534e-01 -1.06605664e-01 1.46614343e-01 -5.17479181e-01
-4.20161158e-01 2.32526913e-01 5.27531877e-02 -5.85697353e-01
-1.57831416e-01 -8.83197665e-01 -1.48812091e+00 -9.15559530e-01
-3.04925919e-01 6.26660109e-01 5.33125281e-01 4.33436036e-01
4.90308851e-01 1.91273734e-01 8.80735934e-01 -5.29292285e-01
-4.26410407e-01 -4.43298668e-01 -9.10752535e-01 4.67142791e-01
4.04596748e-03 -2.08193809e-01 -4.06680167e-01 6.49326816e-02]
|
[7.031836986541748, 5.122690677642822]
|
f9e6460a-47ad-437a-a768-32791afba232
|
l-ac-learning-latent-decision-aware-models
|
2306.17366
| null |
https://arxiv.org/abs/2306.17366v1
|
https://arxiv.org/pdf/2306.17366v1.pdf
|
$λ$-AC: Learning latent decision-aware models for reinforcement learning in continuous state-spaces
|
The idea of decision-aware model learning, that models should be accurate where it matters for decision-making, has gained prominence in model-based reinforcement learning. While promising theoretical results have been established, the empirical performance of algorithms leveraging a decision-aware loss has been lacking, especially in continuous control problems. In this paper, we present a study on the necessary components for decision-aware reinforcement learning models and we showcase design choices that enable well-performing algorithms. To this end, we provide a theoretical and empirical investigation into prominent algorithmic ideas in the field. We highlight that empirical design decisions established in the MuZero line of works are vital to achieving good performance for related algorithms, and we showcase differences in behavior between different instantiations of value-aware algorithms in stochastic environments. Using these insights, we propose the Latent Model-Based Decision-Aware Actor-Critic framework ($\lambda$-AC) for decision-aware model-based reinforcement learning in continuous state-spaces and highlight important design choices in different environments.
|
['Amir-Massoud Farahmand', 'Igor Gilitschenski', 'Romina Abachi', 'Arash Ahmadian', 'Claas A Voelcker']
|
2023-06-30
| null | null | null | null |
['continuous-control', 'model-based-reinforcement-learning', 'decision-making']
|
['playing-games', 'reasoning', 'reasoning']
|
[-9.64928884e-03 1.27538204e-01 -8.05197835e-01 -2.75259078e-01
-9.42583144e-01 -2.22847402e-01 3.94806355e-01 1.15557365e-01
-6.07725978e-01 1.08877051e+00 -6.66845590e-02 -4.99670506e-01
-7.05509186e-01 -5.62278509e-01 -5.11102438e-01 -8.19945991e-01
-4.89926964e-01 5.20853341e-01 -6.52860105e-02 -4.82749015e-01
2.81200230e-01 1.93878263e-01 -1.31867707e+00 -2.36239314e-01
7.05071628e-01 8.49720061e-01 -8.70901793e-02 9.33498979e-01
3.13452393e-01 1.02241838e+00 -5.33041835e-01 -1.39379025e-01
4.07119155e-01 -6.03831768e-01 -5.93695402e-01 1.36486679e-01
-7.55683854e-02 -5.55788994e-01 -1.28759161e-01 6.69135332e-01
8.53769541e-01 3.75695348e-01 6.30844414e-01 -1.46794152e+00
-1.36787355e-01 8.00618947e-01 -4.82488781e-01 2.16302291e-01
-1.22489408e-01 7.67891705e-01 1.10841835e+00 -2.53852189e-01
3.81091982e-01 1.39532733e+00 5.44824719e-01 9.17367756e-01
-1.51620412e+00 -4.12880301e-01 6.63698196e-01 3.65187585e-01
-6.72950387e-01 -3.10314953e-01 8.41737270e-01 -4.16307837e-01
1.05903673e+00 -1.16023846e-01 1.04987478e+00 1.21366882e+00
6.49436116e-01 1.15551507e+00 1.43942463e+00 -4.90799278e-01
8.91319156e-01 -1.08681105e-01 -1.16098989e-02 4.36069697e-01
1.86058804e-01 9.07843769e-01 -5.02458870e-01 -6.35016561e-02
8.57826352e-01 -8.57410505e-02 1.93909436e-01 -1.08970869e+00
-9.03215885e-01 9.69922781e-01 6.54915869e-02 4.47019339e-02
-5.65813661e-01 9.10836399e-01 4.71420109e-01 7.95224845e-01
3.65522623e-01 8.06106985e-01 -5.03210068e-01 -5.51899493e-01
-9.93080854e-01 7.08360136e-01 6.86839104e-01 7.62094080e-01
6.62893116e-01 7.23405898e-01 -3.71743590e-01 5.03412068e-01
3.85615706e-01 4.07509685e-01 3.97832245e-01 -1.44037724e+00
2.81348884e-01 4.11737300e-02 6.09270394e-01 -5.94545484e-01
-5.34977078e-01 -7.74110079e-01 -5.16185284e-01 8.94869864e-01
3.33724320e-01 -6.65390790e-01 -4.86655891e-01 1.80020034e+00
2.24696249e-01 -1.18654943e-03 1.19528227e-01 7.53401458e-01
-4.06079292e-01 4.07707840e-01 1.95309505e-01 -6.05860174e-01
6.97219908e-01 -9.69231129e-01 -7.81840861e-01 -1.49579138e-01
5.63561618e-01 -2.85452574e-01 1.08742905e+00 4.79666829e-01
-1.36911023e+00 -4.94093031e-01 -1.20945358e+00 7.94958353e-01
9.54559967e-02 -3.50671381e-01 7.45808125e-01 6.68550313e-01
-1.14448917e+00 9.53820944e-01 -1.16596639e+00 -2.72554547e-01
1.76543489e-01 4.23746109e-01 3.84813994e-01 3.13348770e-01
-1.14221084e+00 1.27402890e+00 2.88646787e-01 -1.74568683e-01
-1.65947449e+00 -5.41207433e-01 -4.51492637e-01 -1.55112594e-01
9.41649675e-01 -7.86062121e-01 1.96121824e+00 -1.10222208e+00
-2.00331759e+00 2.02437803e-01 2.43249774e-01 -1.12164056e+00
8.26642513e-01 -1.23535797e-01 -1.87572971e-01 -5.29610831e-03
-2.62192994e-01 3.05289984e-01 1.20758700e+00 -1.33503067e+00
-9.85266268e-01 -1.13009863e-01 4.12126601e-01 2.79197484e-01
-4.78986092e-02 -2.72046298e-01 4.22651410e-01 -5.80802560e-01
-5.54200590e-01 -7.84182608e-01 -9.06665444e-01 -2.58200597e-02
-3.81994434e-02 -2.63553917e-01 5.64612567e-01 -6.59875153e-03
1.49111784e+00 -1.63001585e+00 -4.36203144e-02 4.84207571e-02
1.05214655e-01 2.43525565e-01 -3.88892591e-02 8.95032227e-01
1.72977671e-01 6.45084009e-02 -2.34529883e-01 -2.73830861e-01
5.13977945e-01 4.49494064e-01 -4.09087002e-01 5.29107928e-01
2.15086684e-01 9.68151331e-01 -1.05848014e+00 -3.95262063e-01
5.47984898e-01 -3.40583995e-02 -7.75339663e-01 2.74638295e-01
-6.62244976e-01 2.46131718e-01 -8.54601920e-01 4.33013409e-01
5.82152382e-02 3.60945500e-02 2.03606859e-01 5.06887317e-01
-2.32578427e-01 1.50114253e-01 -1.18642879e+00 1.57158458e+00
-7.37322927e-01 4.36584145e-01 4.11551833e-01 -1.32536530e+00
7.05930293e-01 2.93378979e-01 7.83540010e-01 -7.41253555e-01
1.54183716e-01 1.09276034e-01 5.55892847e-02 -3.13269138e-01
4.66316164e-01 -4.37164098e-01 -7.31533393e-02 6.62819088e-01
-5.73741496e-02 -3.82700324e-01 3.84114206e-01 -1.32463843e-01
1.11909890e+00 5.81034362e-01 4.38001692e-01 -4.67043191e-01
8.19441527e-02 1.27403572e-01 6.45228505e-01 1.11425793e+00
-7.33995795e-01 -1.11824475e-01 6.32146180e-01 -4.36224401e-01
-9.68752384e-01 -7.50755191e-01 9.43227112e-03 1.27499080e+00
1.26578668e-02 -3.18973809e-01 -5.38931429e-01 -4.67471898e-01
2.73955822e-01 8.56649160e-01 -9.12874460e-01 -4.11615998e-01
-5.88462770e-01 -8.26958835e-01 1.84790581e-01 4.70392972e-01
2.78970569e-01 -9.04540002e-01 -1.24944198e+00 8.62391114e-01
5.00559032e-01 -2.99565703e-01 -5.11543043e-02 6.30890012e-01
-1.10455632e+00 -9.84738290e-01 -7.56504774e-01 -2.52207130e-01
1.77631229e-01 -2.45511889e-01 1.20337331e+00 -2.64191508e-01
-1.56925973e-02 9.46241617e-01 -1.96698278e-01 -4.64187831e-01
-6.22341096e-01 -1.45695154e-02 2.88509458e-01 -1.69398129e-01
-1.00289755e-01 -6.50996625e-01 -6.73715115e-01 2.62495100e-01
-6.04753137e-01 -1.02730706e-01 5.60324669e-01 9.92806852e-01
5.61732709e-01 -1.10361658e-01 7.47002482e-01 -5.47196686e-01
1.04043341e+00 -2.90093869e-01 -8.28545988e-01 1.76611587e-01
-1.23099744e+00 5.55815339e-01 6.65389776e-01 -4.38333005e-01
-8.69255662e-01 -2.22592697e-01 -1.15350485e-01 -3.78471732e-01
1.02646679e-01 2.53640145e-01 1.94953442e-01 1.42294690e-01
7.04174578e-01 1.40224591e-01 1.68130353e-01 -3.07880163e-01
3.57688069e-01 3.15458924e-01 -8.76867697e-02 -1.12884021e+00
4.81157631e-01 2.27751404e-01 2.50833124e-01 -4.05120045e-01
-6.35442197e-01 -1.58710614e-01 -1.24503084e-01 -3.76525939e-01
5.25761187e-01 -7.79110253e-01 -9.80500996e-01 2.14567721e-01
-4.86964732e-01 -9.61314082e-01 -1.01129544e+00 3.83835196e-01
-1.57279599e+00 -9.08251032e-02 -4.57970500e-01 -1.45484006e+00
-2.24144720e-02 -1.34867036e+00 5.15163422e-01 2.20988765e-01
5.89643717e-02 -1.32555592e+00 6.06890380e-01 -2.55979151e-01
7.29455233e-01 3.36751699e-01 8.01227927e-01 -4.75949585e-01
-4.59824622e-01 2.13802919e-01 4.85928923e-01 2.49590367e-01
-2.35824406e-01 9.99474525e-02 -8.74346673e-01 -6.06301725e-01
-1.06741019e-01 -5.33660293e-01 8.13804507e-01 8.46627176e-01
8.55924129e-01 -3.53963822e-01 -1.42529622e-01 2.52774447e-01
1.51815283e+00 4.09289628e-01 2.12193523e-02 6.73004448e-01
3.19021717e-02 4.97295260e-01 9.44052935e-01 1.00272667e+00
2.69842625e-01 7.29753256e-01 8.04487586e-01 6.23918101e-02
1.39524013e-01 -4.32842791e-01 6.66361451e-01 3.98388267e-01
2.85508968e-02 -1.87346786e-01 -6.37952805e-01 6.06674731e-01
-2.19377232e+00 -1.06725371e+00 4.11555022e-01 2.25264573e+00
8.80526602e-01 3.54571611e-01 6.10158324e-01 1.04456976e-01
5.72940946e-01 3.00196379e-01 -1.11131787e+00 -8.09396565e-01
6.70914724e-02 2.73537319e-02 5.11269033e-01 6.29944086e-01
-9.87005234e-01 7.99356401e-01 7.40347576e+00 1.00048327e+00
-9.84985411e-01 6.62326291e-02 6.78030074e-01 -4.14793134e-01
-1.43575817e-01 3.17137599e-01 -7.11221337e-01 2.90196627e-01
1.26211071e+00 -5.40708125e-01 4.36649889e-01 1.28966784e+00
8.19908738e-01 -2.20811218e-01 -1.10337973e+00 7.29592144e-01
-4.95196462e-01 -1.09968948e+00 -3.44630420e-01 1.98812053e-01
9.14271057e-01 -1.33340031e-01 2.89213091e-01 7.32940376e-01
1.03135240e+00 -8.24234664e-01 9.91074562e-01 4.42021787e-01
4.00527924e-01 -1.08064508e+00 3.96565199e-01 4.38470662e-01
-8.55234921e-01 -5.67903399e-01 -3.67215604e-01 -3.61579567e-01
2.26060078e-01 4.62726504e-01 -5.31420290e-01 2.42002755e-01
1.57435179e-01 9.59656715e-01 -3.81008908e-02 1.09756255e+00
-3.43055874e-01 9.31975305e-01 -7.54556805e-02 -3.59049588e-01
5.68626463e-01 4.22919840e-02 6.05889618e-01 1.03414571e+00
-3.13364365e-03 -3.40901047e-01 4.74294484e-01 7.20923007e-01
3.73838514e-01 1.18522253e-02 -4.83502388e-01 8.52045417e-02
1.21114634e-01 8.98136258e-01 -4.14728910e-01 -2.62260914e-01
7.10690320e-02 3.71996790e-01 3.11147600e-01 3.99498105e-01
-8.83279800e-01 -1.47980407e-01 1.00338602e+00 5.27612232e-02
1.49501801e-01 -4.02694494e-01 -1.19773827e-01 -9.40744042e-01
-2.39984900e-01 -9.73513246e-01 2.76636600e-01 -2.81192333e-01
-1.16315961e+00 7.96134323e-02 2.19758362e-01 -1.42227769e+00
-8.54609132e-01 -4.40736145e-01 -3.89432430e-01 3.79699916e-01
-1.61426961e+00 -6.75538659e-01 5.17596424e-01 4.31549788e-01
8.77590835e-01 -1.07507147e-01 5.20847201e-01 -1.43728405e-01
-5.96962154e-01 5.26046991e-01 7.31224120e-01 -3.82923901e-01
5.71325719e-01 -1.48467040e+00 1.52668923e-01 6.14916801e-01
-1.22967042e-01 3.83043110e-01 1.06768620e+00 -3.33725274e-01
-1.69299078e+00 -7.60215819e-01 1.46655217e-01 -3.02175194e-01
8.26620877e-01 3.76813114e-02 -2.67526120e-01 4.08986002e-01
2.08214894e-01 -3.30462068e-01 2.87863523e-01 1.06491461e-01
7.93966427e-02 -3.36215883e-01 -1.04873347e+00 7.46546447e-01
9.71645057e-01 -1.37316540e-01 -5.12402654e-01 -8.86442547e-04
4.93290722e-01 -2.00212017e-01 -8.61599624e-01 1.80407286e-01
6.33233011e-01 -1.15999091e+00 8.76726389e-01 -9.22578216e-01
2.60067552e-01 -2.30858959e-02 -4.16028053e-02 -1.79713023e+00
-2.36857504e-01 -1.14990079e+00 -5.34523845e-01 6.80125237e-01
2.77003407e-01 -7.26484656e-01 7.44609118e-01 6.48506105e-01
-2.31809556e-01 -1.17760396e+00 -1.15604246e+00 -1.16780114e+00
6.14625335e-01 -6.64237559e-01 3.23507726e-01 5.77922523e-01
1.20390028e-01 -9.88497362e-02 -6.76913381e-01 -4.55507666e-01
6.85977936e-01 1.90029532e-01 6.14906132e-01 -7.64741898e-01
-7.52432883e-01 -7.94039786e-01 -1.79403752e-01 -1.20899594e+00
6.52287975e-02 -3.85551989e-01 5.39303944e-03 -1.56412280e+00
-8.01285356e-03 -6.22562826e-01 -6.93679750e-01 4.72271681e-01
-6.16353229e-02 -4.34593648e-01 4.59004670e-01 3.85943875e-02
-1.03677750e+00 9.45834577e-01 1.41077650e+00 -1.40947521e-01
-2.97718823e-01 3.64054263e-01 -5.55395424e-01 5.19601107e-01
1.05505884e+00 -5.25635540e-01 -6.38025939e-01 -1.98412985e-01
3.02592218e-01 3.70736301e-01 2.20960960e-01 -1.31971586e+00
-1.53998574e-02 -8.46310258e-01 1.49291670e-02 -2.71526784e-01
2.96969682e-01 -8.55782270e-01 -2.84491241e-01 1.05457723e+00
-9.84634161e-01 3.27017725e-01 4.88927998e-02 8.89306545e-01
6.92591667e-02 -1.25924438e-01 8.75677347e-01 -1.75221711e-01
-7.41773367e-01 3.99790466e-01 -8.06818604e-01 4.94649738e-01
1.24623477e+00 -6.61169142e-02 1.47534415e-01 -6.64405048e-01
-6.95109189e-01 4.76616114e-01 4.38689321e-01 1.57964244e-01
3.22709084e-01 -9.47074473e-01 -5.46897292e-01 -2.02035695e-01
-1.28494635e-01 -5.27696073e-01 9.00911391e-02 6.64860070e-01
-1.34937480e-01 4.75343406e-01 -3.67677271e-01 -3.37950259e-01
-6.75248086e-01 6.00349844e-01 7.35611618e-01 -7.98222005e-01
-2.16157034e-01 5.00194430e-01 -3.30263793e-01 -1.10904440e-01
3.98787051e-01 -5.33060670e-01 2.04364449e-01 1.88168064e-02
3.37984592e-01 6.15750253e-01 -2.23217070e-01 7.21543357e-02
-7.89295211e-02 3.70547950e-01 7.22322315e-02 -6.31318390e-01
1.34236133e+00 -2.59350836e-01 7.09001601e-01 7.65351534e-01
6.16937160e-01 -4.95037973e-01 -2.23847008e+00 -1.81663647e-01
1.08972088e-01 -1.83420241e-01 2.09120095e-01 -8.69066536e-01
-8.05950105e-01 1.14648426e+00 8.13514650e-01 2.96864748e-01
9.34980690e-01 -6.36903048e-01 3.44455034e-01 5.05345225e-01
9.44637179e-01 -1.62426329e+00 1.93771914e-01 5.58814764e-01
6.38303339e-01 -1.09702265e+00 7.59515241e-02 5.87538004e-01
-7.83086181e-01 9.85665798e-01 6.02360904e-01 -4.84656453e-01
5.49291313e-01 2.74381697e-01 -9.40459743e-02 2.01433867e-01
-1.37094855e+00 -4.95253086e-01 -3.89018595e-01 8.66483390e-01
1.08437143e-01 1.67870924e-01 -3.29056442e-01 3.55005324e-01
1.09233804e-01 3.33798043e-02 4.58311200e-01 1.41662586e+00
-7.17862964e-01 -1.52322054e+00 -1.68101728e-01 2.92001337e-01
-4.62769270e-01 1.50886357e-01 4.01052311e-02 9.58437383e-01
-3.77844423e-01 1.18286288e+00 -2.44305983e-01 -1.38701856e-01
3.47609550e-01 4.73899096e-02 5.70515454e-01 -1.87011316e-01
-8.33266675e-01 -4.00840342e-02 1.52247414e-01 -7.94116378e-01
-4.41975892e-01 -5.11837482e-01 -1.06091416e+00 -3.05626601e-01
-6.94928169e-02 2.65824437e-01 6.65681243e-01 8.35919380e-01
3.54999453e-01 6.48141384e-01 8.76540720e-01 -7.91017652e-01
-1.39161170e+00 -5.83376527e-01 -7.87807286e-01 -9.28642228e-02
5.89513600e-01 -9.08337474e-01 -2.62769967e-01 -3.04715693e-01]
|
[4.130195140838623, 2.191617965698242]
|
85eba30c-8156-4467-8e0e-34a2fe82ed7a
|
es-un-platano-exploring-the-application-of-a
| null | null |
https://aclanthology.org/W19-1602
|
https://aclanthology.org/W19-1602.pdf
|
?`Es un pl\'atano? Exploring the Application of a Physically Grounded Language Acquisition System to Spanish
|
In this paper we describe a multilingual grounded language learning system adapted from an English-only system. This system learns the meaning of words used in crowd-sourced descriptions by grounding them in the physical representations of the objects they are describing. Our work presents a framework to compare the performance of the system when applied to a new language and to identify modifications necessary to attain equal performance, with the goal of enhancing the ability of robots to learn language from a more diverse range of people. We then demonstrate this system with Spanish, through first analyzing the performance of translated Spanish, and then extending this analysis to a new corpus of crowd-sourced Spanish language data. We find that with small modifications, the system is able to learn color, object, and shape words with comparable performance between languages.
|
['Francis Ferraro', 'Cynthia Matuszek', 'Caroline Kery']
|
2019-06-01
| null | null | null |
ws-2019-6
|
['grounded-language-learning']
|
['natural-language-processing']
|
[-3.08473766e-01 6.00456819e-02 2.59742647e-01 -3.03562492e-01
-8.04441452e-01 -7.20137715e-01 8.02553773e-01 3.39338541e-01
-8.07738662e-01 9.79754448e-01 2.65065283e-01 3.98024842e-02
3.68751466e-01 -8.67563069e-01 -7.34859109e-01 -4.31652635e-01
-3.74120958e-02 9.40999210e-01 4.08815056e-01 -5.92443645e-01
1.89644158e-01 6.07826233e-01 -1.63626003e+00 3.85498703e-01
5.00132382e-01 3.50358099e-01 8.03569615e-01 4.89830673e-01
-3.97640109e-01 1.14344811e+00 -7.81873584e-01 -1.76694825e-01
2.37364501e-01 -3.00374568e-01 -9.80944455e-01 5.90389781e-02
4.57936764e-01 -1.87597126e-01 -1.71035007e-01 7.09907055e-01
5.07644176e-01 2.04439878e-01 6.19133115e-01 -1.06737649e+00
-9.26037431e-01 4.71314669e-01 2.34125331e-01 -2.04946369e-01
1.04853058e+00 1.68413833e-01 6.54870331e-01 -7.75924504e-01
9.89200592e-01 1.63985038e+00 5.82072258e-01 7.70277202e-01
-9.95654285e-01 -4.21759129e-01 1.34497639e-02 -2.54240334e-02
-1.60634160e+00 -5.35074592e-01 4.60110337e-01 -4.87798035e-01
1.18089712e+00 -5.26304722e-01 6.74916387e-01 1.08403254e+00
8.42918754e-02 5.39279819e-01 1.11848688e+00 -8.29561234e-01
5.21584332e-01 5.46605110e-01 -3.18561316e-01 1.06385350e+00
4.43756878e-01 -7.91776925e-02 -8.39569569e-01 -5.03217205e-02
5.31851649e-01 -5.40687680e-01 2.43245363e-02 -5.85476339e-01
-1.33443439e+00 8.70845795e-01 5.49043834e-01 5.97083569e-01
-1.40595630e-01 2.97672600e-01 2.66545653e-01 1.38599873e-01
3.35307658e-01 5.07625222e-01 -3.40404838e-01 -4.15870957e-02
-4.38491493e-01 5.47924042e-01 1.01596522e+00 1.34885001e+00
1.01699293e+00 4.61944332e-03 2.73588747e-01 5.86651683e-01
3.67852449e-01 1.08967996e+00 6.98391557e-01 -1.03656459e+00
3.92680347e-01 4.85867202e-01 3.43056381e-01 -8.49654496e-01
-4.41396296e-01 1.69511974e-01 5.12520783e-02 3.93758595e-01
4.74628389e-01 -2.26135164e-01 -5.42080522e-01 1.61914945e+00
2.29965337e-02 -3.87364298e-01 7.35508859e-01 5.88108718e-01
1.05177498e+00 5.78779578e-01 2.88310140e-01 2.75565535e-01
1.17032599e+00 -8.08253586e-01 -2.82338798e-01 -4.17097658e-01
9.76998508e-01 -3.55035186e-01 1.35270619e+00 2.50388563e-01
-6.85434639e-01 -6.19571567e-01 -1.08441603e+00 -4.05208021e-01
-9.20436084e-01 6.61995858e-02 5.55637956e-01 5.12077510e-01
-1.31798291e+00 2.89197773e-01 -5.77133298e-01 -1.09051442e+00
2.78487597e-02 2.86875844e-01 -6.53990269e-01 -1.48069352e-01
-9.24570858e-01 1.36579537e+00 6.80092990e-01 -6.14055276e-01
-1.12858891e+00 -2.11608298e-02 -1.26769519e+00 -5.26025295e-01
1.65180415e-01 -4.76792723e-01 1.32284141e+00 -9.57514405e-01
-1.49113679e+00 1.16058469e+00 -1.99980289e-01 -3.81407559e-01
3.58238429e-01 1.54744480e-02 -3.42592955e-01 2.18231440e-01
5.81610382e-01 1.22847402e+00 2.54824251e-01 -1.69186616e+00
-8.08115363e-01 -2.39822671e-01 4.60696876e-01 3.41866344e-01
-2.42138267e-01 -2.07386926e-01 -3.05359304e-01 -2.45322153e-01
-2.53571808e-01 -1.06982398e+00 -5.99594414e-03 1.24915078e-01
1.74456656e-01 -3.07686508e-01 3.45397681e-01 -5.50892174e-01
1.72387064e-01 -2.12510848e+00 1.27970174e-01 -5.69520667e-02
-8.86593312e-02 -1.31537750e-01 -4.94357765e-01 5.68049371e-01
6.05923355e-01 -5.48440330e-02 -5.12923598e-02 -3.92132938e-01
-3.14979218e-02 5.71862340e-01 4.20439914e-02 4.45593894e-01
1.38324007e-01 6.88330710e-01 -1.28524518e+00 -6.27880573e-01
8.47907215e-02 4.89077628e-01 -5.40577233e-01 2.37217382e-01
-2.27119878e-01 4.56614673e-01 -4.75537688e-01 2.68001139e-01
7.24079534e-02 2.00215906e-01 2.14243457e-01 3.07536542e-01
-2.12296963e-01 9.10499319e-02 -9.73038673e-01 2.07503843e+00
-8.96475196e-01 8.87551665e-01 -1.02039218e-01 -5.29194474e-01
1.23863578e+00 2.91783750e-01 1.83741629e-01 -7.92189717e-01
1.84431568e-01 4.30856705e-01 -2.93701082e-01 -8.78587544e-01
7.02625871e-01 -3.94738138e-01 -3.76468092e-01 3.71644288e-01
3.81184489e-01 -5.50373018e-01 4.14226353e-01 7.68523067e-02
6.54401004e-01 4.83602434e-01 4.93288696e-01 -6.51976585e-01
5.72169781e-01 5.56177318e-01 -9.87211689e-02 5.67570567e-01
-6.91805035e-02 3.58483821e-01 -6.04921952e-02 -5.34127712e-01
-9.82435107e-01 -1.06192088e+00 2.44811699e-01 1.35863090e+00
2.79364258e-01 -3.35777700e-01 -7.38977253e-01 -5.21924317e-01
-8.12833458e-02 1.04202652e+00 -6.20529652e-01 1.01787433e-01
-4.66487497e-01 -2.71932669e-02 7.77836919e-01 4.56352711e-01
5.35831153e-01 -1.44369173e+00 -1.13088226e+00 5.52386157e-02
-1.83656603e-01 -1.32673419e+00 1.02907859e-01 1.90164372e-01
-4.38521475e-01 -9.39637005e-01 -6.93219781e-01 -1.31646705e+00
6.65364861e-01 2.50787616e-01 1.32569551e+00 6.70851544e-02
9.61411670e-02 9.70042586e-01 -6.67856991e-01 -8.50794435e-01
-8.08967471e-01 2.86626726e-01 2.89997369e-01 -5.64168870e-01
5.60137928e-01 -5.69680631e-02 2.66465008e-01 -5.74570298e-02
-1.01231384e+00 -1.12212822e-01 2.55363584e-01 3.30624074e-01
1.13067254e-01 -3.48260820e-01 1.90398127e-01 -3.81528884e-01
6.76365495e-01 -3.59234720e-01 -4.14822638e-01 3.90150458e-01
-1.06025077e-01 4.91989434e-01 5.85916400e-01 -4.11058992e-01
-9.83401477e-01 2.05251247e-01 -1.27213355e-02 2.32385069e-01
-5.89515150e-01 2.45236591e-01 -1.78693399e-01 -1.96933821e-01
9.81858611e-01 2.26904109e-01 1.22353975e-02 -4.47428763e-01
5.61877370e-01 8.22226822e-01 5.26194572e-01 -9.64143932e-01
5.63316345e-01 5.98859787e-01 -1.54251322e-01 -1.18834591e+00
-4.54798758e-01 -3.82563472e-01 -1.02742171e+00 -1.34478688e-01
1.03190780e+00 -1.17899728e+00 -3.51563126e-01 2.61479825e-01
-1.31013799e+00 -6.88046098e-01 -4.26965237e-01 5.13017297e-01
-8.08514416e-01 2.73613222e-02 -2.90505588e-01 -5.94572663e-01
1.34748131e-01 -1.06661415e+00 1.34443152e+00 1.06623434e-01
-2.54401445e-01 -1.26858449e+00 4.96301830e-01 -4.11056094e-02
1.59714520e-01 6.29228950e-02 7.56666899e-01 -8.64309132e-01
-1.83141995e-02 4.53876927e-02 1.23704769e-01 5.09409085e-02
1.62896797e-01 -5.13739049e-01 -7.91085899e-01 -3.63228053e-01
-2.43306682e-01 -7.91245461e-01 3.65665346e-01 -3.66896421e-01
7.89394751e-02 -6.85636327e-02 -2.37402290e-01 2.64456689e-01
1.56239843e+00 1.53305024e-01 3.52939039e-01 7.21344590e-01
5.46964705e-01 8.83454680e-01 4.23180580e-01 6.49515465e-02
7.79086828e-01 7.08712459e-01 7.30812401e-02 7.52166137e-02
-1.78506806e-01 -4.88689899e-01 6.57068491e-01 5.56250274e-01
-1.20461859e-01 -3.19041640e-01 -1.25698578e+00 7.21154511e-01
-1.61135554e+00 -6.79798484e-01 3.02513450e-01 1.86613655e+00
6.21033370e-01 -2.22571090e-01 3.04461271e-01 -4.16421592e-01
3.77897799e-01 -2.49451563e-01 -2.74370134e-01 -3.94254029e-01
-3.36846471e-01 8.73643756e-02 3.68519425e-01 6.42074406e-01
-6.68788314e-01 1.35359597e+00 7.36740494e+00 3.34179282e-01
-1.18346810e+00 1.34034261e-01 -6.50156289e-03 2.34339923e-01
-2.00064018e-01 -1.92277431e-01 -7.06174374e-01 -1.00388385e-01
8.69705021e-01 -1.18873455e-01 6.92872345e-01 8.36147308e-01
1.03748061e-01 -2.49867678e-01 -1.38653231e+00 9.25622761e-01
7.96745896e-01 -9.81058717e-01 3.16441625e-01 -2.01903343e-01
5.17806411e-01 2.31759936e-01 -4.74224865e-01 3.41844559e-01
5.31151950e-01 -1.09750342e+00 1.39272904e+00 5.40987313e-01
6.65222526e-01 -4.45474714e-01 7.56608725e-01 4.23833311e-01
-1.01339602e+00 2.97876783e-02 -4.23061043e-01 -4.19682115e-01
-1.17573775e-01 -3.92875284e-01 -1.06916118e+00 3.04209352e-01
7.33578086e-01 5.04192531e-01 -8.72226417e-01 6.28420174e-01
-2.93705046e-01 -6.32341430e-02 -2.44893834e-01 -7.04250991e-01
3.14642340e-01 1.27757415e-01 3.81044716e-01 1.41560495e+00
4.46327984e-01 -8.49529803e-02 5.48799157e-01 6.29621208e-01
2.85982192e-01 4.84629214e-01 -1.06113410e+00 1.56585142e-01
3.67027640e-01 7.97279954e-01 -7.30427265e-01 -4.62871611e-01
-4.22745466e-01 8.96668434e-01 6.23215675e-01 2.91822881e-01
-4.65812922e-01 -3.48412812e-01 1.90490305e-01 8.88253674e-02
7.56903142e-02 -7.23574817e-01 7.46451095e-02 -1.09038544e+00
-1.12769879e-01 -9.19719696e-01 -7.23747462e-02 -1.23133910e+00
-1.18163645e+00 8.56344819e-01 4.51022953e-01 -9.75619316e-01
-5.30877233e-01 -1.11452234e+00 4.95420806e-02 7.11800754e-01
-1.24198890e+00 -1.54874706e+00 -4.24070835e-01 6.86660111e-01
5.63391447e-01 -6.85765088e-01 1.12155128e+00 -1.17443278e-01
2.62919307e-01 1.29773587e-01 -1.29530221e-01 2.91827798e-01
6.90091610e-01 -1.08662701e+00 2.28756711e-01 3.80846262e-01
5.88249981e-01 4.67051685e-01 8.89704525e-01 -6.29040182e-01
-1.16918421e+00 -8.01603138e-01 9.53040719e-01 -5.94713926e-01
5.19347012e-01 -5.59634030e-01 -5.41656673e-01 6.91316128e-01
4.13703352e-01 -2.61807472e-01 4.63698238e-01 -1.32090718e-01
-5.03860176e-01 1.35197639e-01 -1.39481151e+00 5.38301289e-01
1.08182454e+00 -7.10680485e-01 -9.52679992e-01 4.26545858e-01
6.63301766e-01 -1.92999557e-01 -4.33280528e-01 -1.26151130e-01
4.44188774e-01 -8.28898549e-01 7.13487625e-01 -5.52956700e-01
1.74080148e-01 -3.85903299e-01 -6.57999635e-01 -1.48109365e+00
-3.68187845e-01 -1.46990076e-01 5.70585072e-01 1.06627202e+00
4.64720100e-01 -6.51532769e-01 3.32498878e-01 2.08348840e-01
6.69734627e-02 -1.93672962e-02 -8.93573582e-01 -8.47599983e-01
5.62298059e-01 -3.32527727e-01 6.22839928e-01 6.65460587e-01
1.44860908e-01 5.06238163e-01 8.23918264e-03 5.72859459e-02
3.59743118e-01 -2.53100812e-01 9.73832130e-01 -1.14266574e+00
-5.63524738e-02 -2.27075934e-01 -5.40573895e-01 -7.32845426e-01
4.94272292e-01 -1.00073504e+00 5.41531324e-01 -1.78254533e+00
2.04841420e-01 -3.77378851e-01 2.69254833e-01 5.60234427e-01
2.93472201e-01 5.07102668e-01 5.97493231e-01 1.82800740e-01
-7.80294240e-01 4.41906303e-01 8.81666660e-01 -2.78897136e-01
-1.45414546e-01 -6.38050675e-01 -5.69064260e-01 9.78416264e-01
8.67242813e-01 -5.10362148e-01 -2.57010609e-01 -8.84725690e-01
3.86901438e-01 -6.56120598e-01 3.92666340e-01 -1.49179411e+00
2.22603634e-01 -8.71695951e-02 3.68511200e-01 -1.14197144e-02
2.32007504e-01 -9.13055062e-01 -1.01810716e-01 4.96785164e-01
-4.26179022e-01 3.82111400e-01 4.16167676e-01 3.11169386e-01
5.79755791e-02 -5.00354409e-01 4.79971081e-01 -6.29393697e-01
-1.28607583e+00 -4.19029236e-01 -7.05051780e-01 1.37341097e-01
9.82365608e-01 -1.36727616e-01 -2.71034352e-02 -5.24787545e-01
-5.69819927e-01 -2.22236849e-02 1.07722414e+00 7.26936936e-01
3.82529914e-01 -1.30032921e+00 -8.33544135e-01 1.15903556e-01
5.49448311e-01 -2.30198711e-01 -4.20214266e-01 -8.13933164e-02
-1.05557370e+00 5.57609379e-01 -4.82197136e-01 -5.37271619e-01
-9.94543254e-01 5.94154835e-01 4.68507916e-01 1.65711924e-01
-4.55365181e-01 4.50949013e-01 -9.23862308e-02 -8.72959793e-01
2.35203698e-01 -4.00738388e-01 -3.44744444e-01 -1.54409483e-01
4.30285066e-01 2.56969392e-01 -1.07369699e-01 -1.31714892e+00
-4.52061087e-01 8.21566224e-01 7.02015281e-01 -5.97066820e-01
1.19643521e+00 -2.15819716e-01 -1.23535834e-01 9.73008513e-01
1.07858372e+00 4.36571062e-01 -1.05236483e+00 -2.04671666e-01
4.29931693e-02 -5.49336374e-02 -4.13612068e-01 -6.32541776e-01
-5.77876210e-01 6.25566959e-01 7.53857255e-01 -6.09207377e-02
6.64020181e-01 4.37339783e-01 4.21399206e-01 1.15287745e+00
1.22378290e+00 -1.06226826e+00 2.78354585e-01 8.43571842e-01
9.42373097e-01 -1.18078768e+00 7.27109537e-02 5.04964478e-02
-7.72291183e-01 1.17070460e+00 4.25604165e-01 -1.68621272e-01
3.73477697e-01 3.23423356e-01 4.17708695e-01 -2.56875724e-01
-3.17316711e-01 -4.54228878e-01 8.68169218e-02 1.27616823e+00
3.95446002e-01 1.14576504e-01 2.00762805e-02 2.29069322e-01
-4.73497480e-01 1.35505740e-02 6.29195392e-01 1.10103858e+00
-6.45348489e-01 -1.00590861e+00 -5.35884142e-01 -3.91801983e-01
-2.02736445e-02 4.68870103e-02 -8.07647824e-01 1.17965102e+00
6.84271216e-01 1.23171616e+00 2.26707056e-01 -1.73479006e-01
5.33125043e-01 1.61813349e-01 7.42686391e-01 -8.83069158e-01
-4.07236516e-01 -4.49208647e-01 2.75415510e-01 -4.42924321e-01
-9.11421776e-01 -7.16371655e-01 -1.60735631e+00 2.63909735e-02
3.05399388e-01 2.55142450e-01 7.15023994e-01 1.18294120e+00
5.20714894e-02 1.53747395e-01 -6.40139449e-03 -1.14133489e+00
-6.32515401e-02 -9.21542287e-01 -4.57492530e-01 5.67128837e-01
2.20784426e-01 -7.07853913e-01 -6.98043928e-02 2.98861504e-01]
|
[4.357269287109375, 0.7647721767425537]
|
2e875711-0ae7-4659-99e5-1e8f6d380a96
|
research-on-smoking-behavior-detection-system
| null | null |
https://kns.cnki.net/kcms/detail/detail.aspx?dbcode=CMFD&dbname=CMFDTEMP&filename=1022545108.nh&uniplatform=NZKPT&v=EO9SeGGV79D5fqruvEuEUASMUu2YRCcwEnhyFqQY-l52O5A4MxkYJvMC3lR2N9KV
|
https://kns.cnki.net/kcms/detail/detail.aspx?dbcode=CMFD&dbname=CMFDTEMP&filename=1022545108.nh&uniplatform=NZKPT&v=EO9SeGGV79D5fqruvEuEUASMUu2YRCcwEnhyFqQY-l52O5A4MxkYJvMC3lR2N9KV
|
Research on Smoking Behavior Detection System Based on Deep Learning
|
As we all know,smoking endangers the health of smokers,and the harm of second-hand smoke to the health of people around us can not be ignored;in addition, improper smoking can sometimes cause many safety accidents,such as fire or explosion, and bring huge property losses to the society.Therefore,strengthening the supervision of smoking prohibition in public places has become a growing concern of all sectors of society.Intelligent smoking behavior detection system has become a very urgent demand,and the core of this kind of system is how to design an algorithm to detect smoking behavior quickly and accurately.According to the existing research status of smoking behavior detection,this dissertation proposes a smoking behavior detection method combined with face detection.The method is tested on the self-made smoking behavior data set,and its accuracy is 94.78%,and its detection speed is 26.76FPS, which basically meets the requirements of accuracy and real-time of smoking behavior detection.The main research contents of this dissertation are as follows:
(1)In view of the misjudgment of smoking behavior that may occur in the direct detection of cigarettes,a smoking behavior detection method combining face detection and cigarette detection is proposed.The face detection model and target detection model are used to detect faces and cigarettes respectively,and whether there is smoking behavior is judged according to the position relationship between face detection frame and cigarette detection frame.
(2)Pyramid feature fusion is carried out on the existing face detection model FaceBoxes,which enhances the ability of the model to detect multi-scale faces,and adjusts the structure of rapid digestion convolution layer to retain more face feature information.
(3)Several powerful deep learning target detection algorithms are studied,and their performance on the self-made smoking behavior data set is tested.According to the experimental results,YOLOv4 with excellent performance is selected for subsequent improvement to adapt to the detection of targets such as cigarettes.
(4)In order to improve the detection ability of YOLOv4 for the smaller cigarette target in the image,this dissertation improves the neck of YOLOv4,improves the accuracy of the model,and uses PP-LCNet to replace the original backbone network CSPDarknet53,which speeds up the reasoning speed.
(5)According to the existing web application development technology and the smoking behavior detection method in this dissertation,a simple online smoking behavior detection system is designed,which can realize the interaction between users and the system,and provide the possibility for better applying the smoking behavior detection algorithm to the actual society.
|
['万里波']
|
2022-06-01
| null | null | null |
2022-2022-6
|
['face-detection']
|
['computer-vision']
|
[-1.45978760e-02 -5.27638137e-01 -2.82146633e-01 -2.04383969e-01
2.93610126e-01 -1.48381546e-01 -5.91624975e-02 -8.15822124e-01
-4.04633939e-01 -2.54122801e-02 -2.40700334e-01 -4.83579457e-01
1.25094965e-01 -1.35652351e+00 1.62122503e-01 -9.73656356e-01
5.87700129e-01 -1.26797095e-01 4.61090505e-01 -1.58237889e-01
1.70956478e-01 7.79287755e-01 -1.68706071e+00 3.83510500e-01
3.51080745e-01 9.92360234e-01 5.74386597e-01 3.63034904e-01
-5.81419587e-01 4.12819952e-01 -5.26849389e-01 -2.59502083e-01
2.19568670e-01 -4.65032071e-01 -2.24508762e-01 2.58520931e-01
2.15448171e-01 -6.71310127e-01 -5.66211939e-01 1.26662016e+00
6.25213027e-01 -1.55511498e-01 3.96381974e-01 -1.29584265e+00
-7.21447825e-01 2.61158109e-01 -7.46216655e-01 3.73928845e-01
-1.02890603e-01 2.82198012e-01 2.47933432e-01 -7.42185295e-01
-9.30987969e-02 1.66494620e+00 6.37067616e-01 8.90777946e-01
-4.60347325e-01 -1.46266401e+00 7.88454413e-02 5.26378334e-01
-1.46111679e+00 -3.35346103e-01 6.58084333e-01 -1.33438230e-01
5.40366828e-01 2.89516896e-01 7.02588022e-01 7.94789255e-01
1.21918134e-01 3.83945882e-01 6.64995253e-01 -3.25096250e-01
-1.72022983e-01 3.88782084e-01 9.10139456e-02 1.07292318e+00
4.48226392e-01 4.09739047e-01 2.92406768e-01 1.82641938e-01
8.90652180e-01 5.55275142e-01 1.22140041e-02 5.19359946e-01
-3.12252343e-01 7.86529362e-01 5.67424476e-01 8.95178914e-01
4.05436568e-03 7.58551881e-02 1.54451773e-01 3.30423787e-02
2.62490451e-01 -6.73538089e-01 -1.51129961e-01 1.76057264e-01
-6.31186068e-01 -1.63624659e-01 5.92639148e-01 5.77866852e-01
8.04908335e-01 4.94885862e-01 -5.19851781e-02 8.60095739e-01
7.85349667e-01 1.15751696e+00 6.26670122e-01 -8.79003167e-01
4.59297858e-02 8.57159734e-01 -4.47915703e-01 -1.18685937e+00
-2.90523797e-01 -1.58823907e-01 -8.54658902e-01 5.48922241e-01
1.98986292e-01 -1.65939391e-01 -8.97239506e-01 1.41349816e+00
4.91777390e-01 7.02579096e-02 -2.69654781e-01 6.23501956e-01
1.31327617e+00 8.70153904e-01 2.31555238e-01 -6.23240829e-01
1.65310967e+00 -5.76772094e-01 -9.10983920e-01 -1.24030381e-01
1.93056986e-01 -8.70235205e-01 1.01516902e+00 1.53098941e-01
-6.45060062e-01 -9.63523269e-01 -6.87072635e-01 1.34315327e-01
-5.45016527e-01 3.15010309e-01 4.14073259e-01 1.50960982e+00
-7.61221826e-01 2.63480872e-01 -4.77067232e-01 -5.27507544e-01
6.97465301e-01 4.34628248e-01 1.40754700e-01 -3.07963789e-01
-1.13655102e+00 8.75839233e-01 1.98485106e-02 4.74738240e-01
-5.31534731e-01 -4.61485803e-01 -4.26109314e-01 2.31648579e-01
2.89774984e-01 -3.00605237e-01 1.22356713e+00 -1.16309428e+00
-1.16302574e+00 7.20420122e-01 -3.31907898e-01 3.78493607e-01
3.22011113e-01 5.09688795e-01 -9.68054891e-01 2.73846030e-01
-3.03311623e-03 2.57913083e-01 8.81422997e-01 -1.01639640e+00
-9.71149623e-01 -7.69922733e-01 -1.45109624e-01 -1.10758012e-02
-7.99400568e-01 3.66779894e-01 -4.80994433e-01 -1.42398968e-01
-1.70749173e-01 -6.69847012e-01 -4.69836667e-02 3.76760662e-01
1.60594150e-01 -6.51678741e-01 1.86179256e+00 -6.20550513e-01
1.34055817e+00 -2.45338631e+00 -8.03391218e-01 4.31251585e-01
7.25379586e-02 6.42250299e-01 -4.89003137e-02 -1.61453605e-01
-1.07182413e-01 1.42668277e-01 -1.29092678e-01 2.26898029e-01
-1.75912946e-01 4.29065645e-01 2.35244066e-01 4.94146079e-01
-3.25612694e-01 5.53289056e-01 -4.04502422e-01 -8.51560533e-01
3.97993803e-01 6.03300452e-01 -1.93400234e-01 1.80106312e-01
1.94548637e-01 -1.17423914e-01 -6.95265353e-01 7.82540143e-01
1.12806678e+00 2.85558134e-01 -2.08331831e-02 -4.39163297e-01
-1.44728765e-01 -3.61197621e-01 -1.28306019e+00 6.48161411e-01
-4.65486825e-01 1.85327038e-01 4.89857912e-01 -8.46545875e-01
1.06693995e+00 6.32928371e-01 5.65201163e-01 -8.34895134e-01
5.66276014e-01 -6.03034832e-02 5.78550622e-02 -1.19803393e+00
-3.60420346e-01 -5.23730814e-01 4.51012641e-01 3.13803822e-01
-3.51937085e-01 6.78936066e-03 3.04303646e-01 -6.62774965e-02
7.67204523e-01 -2.26464719e-01 6.30334243e-02 -1.15130417e-01
9.97574627e-01 -2.21470594e-01 5.53863287e-01 3.55120391e-01
-4.93529588e-01 1.03113782e-02 -5.13032340e-02 -3.18589628e-01
-4.79988754e-01 -1.04591084e+00 -3.10510874e-01 1.17210841e+00
1.82097122e-01 1.49114683e-01 -8.37094545e-01 -9.47292447e-01
2.45847367e-02 6.38105452e-01 -2.81299382e-01 -4.57444727e-01
-5.71339309e-01 -7.94997275e-01 4.27139133e-01 5.66508114e-01
1.19005704e+00 -1.68663073e+00 -1.53765842e-01 5.29251173e-02
-1.84893638e-01 -8.74036491e-01 -4.84374493e-01 -5.34796357e-01
-7.13109076e-01 -1.41238570e+00 -3.87680173e-01 -1.23652625e+00
6.04854226e-01 8.96853447e-01 5.80857515e-01 9.01477754e-01
-5.96607864e-01 5.16228318e-01 -7.99386576e-02 -7.34004796e-01
-6.47152066e-01 -3.12155992e-01 1.30074588e-03 -1.33044466e-01
8.72144997e-01 -4.75348741e-01 -6.00411832e-01 4.47922766e-01
-9.23596442e-01 -6.06849372e-01 6.31860733e-01 1.16872288e-01
-1.83360264e-01 6.42987669e-01 7.88327515e-01 -8.09008598e-01
4.19689476e-01 -2.16910839e-01 -7.42379308e-01 1.11924700e-01
-8.83720696e-01 -7.72393584e-01 6.69053257e-01 -3.22360218e-01
-1.46242297e+00 1.42778540e-02 -5.82521975e-01 -3.87120664e-01
-3.55876833e-01 -2.90190071e-01 -5.99512100e-01 -2.46699780e-01
4.65184480e-01 3.59029382e-01 4.45194155e-01 -4.64149058e-01
1.02049880e-01 1.20444143e+00 2.83299565e-01 1.56872928e-01
9.86623764e-01 4.60234225e-01 6.45200163e-02 -9.76224184e-01
-5.56173265e-01 -5.73905826e-01 -3.61730337e-01 -5.88675022e-01
1.08152044e+00 -6.29284441e-01 -1.44718194e+00 5.41852534e-01
-1.07600331e+00 3.43396246e-01 1.85187727e-01 4.75271791e-01
1.34641141e-01 6.61870003e-01 -8.03901196e-01 -1.11740410e+00
-5.82453609e-01 -9.31821823e-01 5.61072588e-01 6.47295177e-01
3.02681595e-01 -8.43837559e-01 -3.95300984e-01 4.68686521e-01
6.03387952e-01 -4.05369550e-01 9.25157487e-01 -8.67458135e-02
-4.26368892e-01 -2.54911721e-01 -4.08304602e-01 6.55062914e-01
4.83977973e-01 3.13649654e-01 -1.19541013e+00 -8.91461670e-02
5.34714282e-01 2.78860182e-01 8.39542627e-01 3.52792174e-01
1.44262767e+00 -2.02005252e-01 -4.78589237e-01 1.90712675e-01
1.38247263e+00 9.87507522e-01 8.91289711e-01 3.18561122e-02
5.67606986e-01 7.01286018e-01 3.59887630e-01 -3.33412290e-02
1.17817177e-02 2.20017597e-01 7.19321668e-01 -4.41777796e-01
-4.22806114e-01 1.15659684e-01 6.02306485e-01 5.72383106e-01
-2.59257168e-01 -4.95082997e-02 -1.51083872e-01 6.44926727e-02
-1.33973098e+00 -1.59763849e+00 -4.77731436e-01 1.81372499e+00
3.66673976e-01 -1.62211061e-01 3.38171870e-01 2.95415968e-01
1.20196474e+00 -1.13314558e-02 -4.26988572e-01 -3.86666626e-01
4.69556034e-01 1.77375585e-01 1.84143335e-01 2.13241935e-01
-9.32125151e-01 5.39753616e-01 6.26556873e+00 1.12062204e+00
-1.28144062e+00 3.54425550e-01 5.25171697e-01 4.80366014e-02
1.24165505e-01 -4.18040246e-01 -1.12559044e+00 7.31834769e-01
4.78986084e-01 -1.41861469e-01 6.22626781e-01 1.10431385e+00
5.69531500e-01 1.38116717e-01 -4.43135411e-01 1.10949194e+00
9.67670977e-02 -8.05023193e-01 -1.42113358e-01 3.84670086e-02
8.55633765e-02 -3.86919439e-01 -2.26242412e-02 5.88503957e-01
-7.03443065e-02 -8.82833540e-01 1.77416101e-01 7.65420124e-02
6.43941641e-01 -7.75250554e-01 8.65326643e-01 5.22335172e-01
-1.52902436e+00 -5.90543151e-01 -6.48269951e-01 -1.19496025e-02
-9.05835778e-02 5.46667576e-01 -5.92381418e-01 2.62202770e-01
8.17264736e-01 5.25421977e-01 -4.75715727e-01 7.44286537e-01
-1.10458106e-01 7.36969948e-01 -4.53320444e-01 -3.55118930e-01
-7.74725229e-02 -4.74761307e-01 2.01542079e-02 1.15068328e+00
5.02191067e-01 3.61476123e-01 1.52326643e-01 8.95919144e-01
-5.80633655e-02 1.79017514e-01 -5.42650461e-01 2.10329905e-01
2.39140898e-01 1.68222249e+00 -6.09097123e-01 -3.62158865e-01
-7.55480886e-01 4.38980192e-01 -3.36541265e-01 7.07887486e-02
-1.00513363e+00 -2.71561593e-01 3.63356531e-01 4.31221843e-01
3.16242158e-01 2.51272827e-01 -5.29302843e-02 -5.14519215e-01
-2.16885075e-01 -6.06340289e-01 6.00095332e-01 -7.11166918e-01
-1.13995349e+00 1.62484497e-01 -1.99812770e-01 -9.48727250e-01
4.11385149e-01 -7.85664976e-01 -1.16562212e+00 7.98587024e-01
-1.31574869e+00 -1.07136083e+00 -4.95930731e-01 8.23686898e-01
6.14924669e-01 -2.44035318e-01 6.84061766e-01 8.80200565e-01
-9.33749437e-01 4.84856308e-01 -1.75277308e-01 3.17226380e-01
3.26258481e-01 -4.79493946e-01 -2.24255353e-01 8.00071955e-01
-1.18430376e-01 3.99762630e-01 9.95878056e-02 -5.46463132e-01
-1.24113631e+00 -1.25946236e+00 7.55836129e-01 -3.60294059e-02
3.51900421e-02 -9.80708227e-02 -8.71092141e-01 3.36565405e-01
-8.79683706e-04 -1.49180040e-01 3.51695746e-01 -2.69888848e-01
-1.62823386e-02 -7.42093801e-01 -1.69413149e+00 5.29925525e-01
9.71707225e-01 -2.41305456e-01 -3.34061503e-01 4.84604478e-01
4.44271684e-01 2.50019252e-01 -3.02138448e-01 2.49576315e-01
6.85633898e-01 -1.05570376e+00 1.04912710e+00 -1.27526164e-01
8.52259621e-02 -4.46117759e-01 1.49655208e-01 -4.93351370e-01
-9.79898095e-01 1.61377802e-01 2.49112964e-01 1.52145576e+00
-3.89399603e-02 -6.37189865e-01 1.05518579e+00 4.13223863e-01
1.29679322e-01 -3.84295076e-01 -9.54632401e-01 -6.32465184e-01
-1.65579438e-01 -2.99739063e-01 5.72296143e-01 6.97214901e-01
-2.49371901e-01 3.57096225e-01 5.22562228e-02 1.96924090e-01
4.64377522e-01 3.31302024e-02 4.18548912e-01 -9.25284445e-01
1.17484055e-01 -5.80158651e-01 -2.06734866e-01 -8.32000196e-01
-8.09558257e-02 -7.27069795e-01 -7.82063678e-02 -1.71254575e+00
2.78049052e-01 1.56153496e-02 -3.44805121e-01 7.50608027e-01
-2.29753360e-01 4.74211633e-01 2.29014978e-01 6.40115933e-03
1.06645338e-02 2.97220111e-01 1.59369946e+00 -3.23816925e-01
1.56482413e-01 6.10781550e-01 -8.88926625e-01 9.37691033e-01
7.68799901e-01 -4.90515292e-01 -3.49934816e-01 -9.78747681e-02
-4.63191211e-01 -9.35688540e-02 2.14194015e-01 -9.36754346e-01
1.75357237e-01 -4.07176733e-01 7.48061836e-01 -5.18264294e-01
2.62061179e-01 -1.49853730e+00 -1.40798330e-01 1.17223096e+00
3.60218287e-01 -1.43182263e-01 1.01777859e-01 3.93352836e-01
3.32372725e-01 -5.54706037e-01 1.18570125e+00 -3.89041543e-01
-9.19773638e-01 4.42393094e-01 -8.00932646e-01 -5.14937699e-01
1.18528068e+00 -4.30563569e-01 -2.60114670e-01 -3.50111187e-01
-5.51413119e-01 6.41858280e-02 -2.03475356e-01 4.89161670e-01
5.71935773e-01 -1.43331182e+00 -4.13211793e-01 5.09155869e-01
-3.38909000e-01 -4.91731048e-01 4.62996006e-01 5.37801445e-01
-3.42878997e-01 -2.96193026e-02 6.52497709e-02 -3.88170958e-01
-1.89981663e+00 7.99632251e-01 5.85476041e-01 2.47788966e-01
-4.23745304e-01 5.57298958e-01 3.02263439e-01 -3.51104379e-01
1.44820675e-01 -8.68596584e-02 -6.97230875e-01 -2.60929346e-01
8.43799889e-01 6.63393855e-01 -2.85059530e-02 -6.96844876e-01
-4.56800938e-01 7.42241204e-01 -1.42632686e-02 4.85498250e-01
8.64758313e-01 -8.83920938e-02 -2.89438069e-01 -1.47641167e-01
1.23675454e+00 9.97001082e-02 -6.55099571e-01 1.75402552e-01
-6.77481234e-01 -7.39677489e-01 1.72486082e-01 -7.15848863e-01
-1.69371343e+00 9.17624176e-01 1.32564151e+00 5.61852813e-01
1.28517509e+00 -2.90807277e-01 1.01087725e+00 3.23422790e-01
-2.79860268e-03 -9.72597837e-01 5.69384210e-02 1.03452079e-01
5.72702944e-01 -1.08869243e+00 -4.15819436e-02 -8.28585505e-01
-1.94623932e-01 1.42508614e+00 1.03805792e+00 5.81534505e-02
9.01453257e-01 4.41943705e-01 3.96337986e-01 -2.04655409e-01
-9.20665637e-02 -2.65166968e-01 -5.56502230e-02 6.94340229e-01
1.63509145e-01 -2.04712301e-01 -4.98965949e-01 5.94374120e-01
1.08931810e-01 2.35871166e-01 7.70793483e-02 6.73169136e-01
-1.05587828e+00 -9.89927590e-01 -7.80772150e-01 6.03964746e-01
-5.31192243e-01 1.44386068e-01 -7.50233531e-02 7.08464026e-01
8.46655071e-01 1.54765463e+00 1.13736548e-01 -2.63502002e-01
5.08268356e-01 -9.78028253e-02 2.75003731e-01 -5.34329057e-01
-6.44349337e-01 1.85176030e-01 -2.16980979e-01 -2.46234044e-01
-4.81558889e-01 -1.86486542e-01 -1.44610667e+00 -7.93253481e-01
-4.87196416e-01 2.66681053e-02 6.20803356e-01 1.08295465e+00
-1.13911822e-01 5.74482143e-01 1.00406611e+00 -3.25723678e-01
-5.92034519e-01 -9.14566219e-01 -8.54243994e-01 4.74024743e-01
-2.71329403e-01 -4.66509491e-01 -5.32731175e-01 -2.47570276e-02]
|
[13.26362419128418, 0.8344709873199463]
|
7ce650b3-9cfe-4172-8551-d080417cef7b
|
semisupervised-regression-in-latent-structure
|
2305.02473
| null |
https://arxiv.org/abs/2305.02473v1
|
https://arxiv.org/pdf/2305.02473v1.pdf
|
Semisupervised regression in latent structure networks on unknown manifolds
|
Random graphs are increasingly becoming objects of interest for modeling networks in a wide range of applications. Latent position random graph models posit that each node is associated with a latent position vector, and that these vectors follow some geometric structure in the latent space. In this paper, we consider random dot product graphs, in which an edge is formed between two nodes with probability given by the inner product of their respective latent positions. We assume that the latent position vectors lie on an unknown one-dimensional curve and are coupled with a response covariate via a regression model. Using the geometry of the underlying latent position vectors, we propose a manifold learning and graph embedding technique to predict the response variable on out-of-sample nodes, and we establish convergence guarantees for these responses. Our theoretical results are supported by simulations and an application to Drosophila brain data.
|
['Carey E. Priebe', 'Youngser Park', 'Michael W. Trosset', 'Joshua Agterberg', 'Aranyak Acharyya']
|
2023-05-04
| null | null | null | null |
['graph-embedding']
|
['graphs']
|
[ 1.61753029e-01 5.72936356e-01 -3.52953583e-01 -2.44402602e-01
2.68828779e-01 -5.22162974e-01 6.78113997e-01 2.36478597e-01
-9.60218608e-02 2.56459624e-01 1.39771447e-01 -8.69830921e-02
-4.41657573e-01 -7.62670219e-01 -7.06524134e-01 -8.20286930e-01
-6.22130036e-01 7.12714076e-01 -1.42955974e-01 1.07217155e-01
9.38185602e-02 2.46277034e-01 -5.43465495e-01 -5.65614700e-01
6.86114132e-01 2.16602921e-01 -6.51919171e-02 6.12710595e-01
2.38932729e-01 1.53494805e-01 -2.15598837e-01 -3.46409738e-01
1.92337692e-01 -3.01655471e-01 -1.88964725e-01 2.89433718e-01
9.89418030e-02 2.70660818e-01 -7.50378549e-01 1.23919582e+00
8.83781239e-02 -8.84369090e-02 1.29905629e+00 -1.76351309e+00
-9.55148458e-01 5.22984326e-01 -9.76494491e-01 -2.39726678e-01
-2.99022458e-02 -3.24356347e-01 1.23810434e+00 -7.82166243e-01
6.66557789e-01 1.21513402e+00 5.15571892e-01 4.24761772e-01
-2.14605093e+00 -3.52812588e-01 1.00010760e-01 -2.76320875e-01
-1.46662700e+00 2.53324048e-03 1.04494381e+00 -9.16647971e-01
3.27994615e-01 -1.13315478e-01 7.17209697e-01 7.86403775e-01
6.56227648e-01 3.94765198e-01 4.76824015e-01 -3.33745331e-01
3.62842113e-01 -2.04332769e-01 2.00934768e-01 9.06496346e-01
4.19042796e-01 -9.26752314e-02 -1.60695747e-01 -5.25605381e-01
1.04670417e+00 4.26963240e-01 -2.33258903e-01 -1.32495856e+00
-1.15545380e+00 1.37009430e+00 4.17270392e-01 2.11957544e-01
-3.84574413e-01 3.88033301e-01 -3.67814690e-01 1.23656169e-01
5.76041698e-01 -6.23097010e-02 -1.46435812e-01 4.89246368e-01
-6.88720703e-01 7.50039369e-02 7.72111714e-01 1.03329396e+00
9.23362553e-01 -2.68288165e-01 2.15145066e-01 5.15985429e-01
9.88399267e-01 7.10427046e-01 8.34298655e-02 -8.92167449e-01
4.11595076e-01 7.42947698e-01 -9.67967287e-02 -1.61318707e+00
-4.81259346e-01 -3.98598403e-01 -1.20854247e+00 1.19909421e-02
5.54459333e-01 -1.03431828e-01 -6.83850050e-01 2.12178946e+00
2.18374044e-01 4.19471383e-01 -2.30429769e-01 5.50355732e-01
-1.75091222e-01 8.30123186e-01 8.39784145e-02 -5.55569291e-01
8.11692774e-01 -5.80317378e-01 -7.30472684e-01 3.80568323e-03
5.98796725e-01 -4.34543192e-01 6.55903995e-01 4.49769050e-02
-8.81218255e-01 -1.94855794e-01 -8.95269573e-01 3.25124502e-01
-2.48518400e-02 -7.06954002e-02 3.16773593e-01 4.12185580e-01
-1.14903843e+00 7.44452894e-01 -1.08540881e+00 -2.25442305e-01
4.51586917e-02 4.18175101e-01 -4.14531291e-01 7.32473657e-02
-8.42538238e-01 2.20465049e-01 -2.94799030e-01 5.10270111e-02
-6.93951845e-01 -6.42408788e-01 -8.60089779e-01 -8.52509737e-02
6.01204624e-03 -7.00552702e-01 5.06600380e-01 -3.60334575e-01
-1.10981011e+00 6.51506484e-01 -1.10078730e-01 -2.83154279e-01
4.03215289e-01 9.64436382e-02 -3.65456551e-01 1.52354270e-01
3.44039679e-01 4.94026124e-01 1.29862118e+00 -9.68893409e-01
4.38562669e-02 -6.90703809e-01 -3.46346945e-01 4.19890089e-03
-5.17597079e-01 -4.63743567e-01 -1.49088427e-01 -4.21500474e-01
7.20492899e-01 -1.28374422e+00 -5.87062001e-01 3.04361492e-01
-5.67350984e-01 -3.09297085e-01 8.75111043e-01 -2.74880379e-01
1.16713464e+00 -2.26900697e+00 7.10499704e-01 5.06154597e-01
8.89953434e-01 -3.41683358e-01 -2.63786763e-01 5.74685693e-01
-3.95217150e-01 3.52332026e-01 -3.99712116e-01 -4.05916005e-01
9.19455942e-03 4.06080857e-02 -2.09377065e-01 1.10763490e+00
7.03607947e-02 8.50835502e-01 -1.04037237e+00 -2.55344480e-01
3.02202906e-02 6.64184868e-01 -5.52872360e-01 1.29646182e-01
-3.60087678e-02 1.72233045e-01 -6.40078068e-01 -1.32279620e-01
5.96707404e-01 -6.24758065e-01 3.22067529e-01 2.11644210e-02
1.80862755e-01 -1.87804416e-01 -1.25188184e+00 1.32829678e+00
6.57430962e-02 5.16644299e-01 -9.41621512e-02 -9.71077442e-01
9.50721622e-01 3.32123756e-01 6.79680645e-01 2.01069519e-01
9.43327497e-04 -1.68955669e-01 1.09325454e-01 -1.32877082e-02
1.15446344e-01 3.66262235e-02 -1.30673423e-01 5.97868979e-01
-1.31866857e-02 3.03949565e-02 1.47009119e-01 7.40454912e-01
1.17506576e+00 -3.29806089e-01 8.75423700e-02 -5.58092654e-01
1.86729819e-01 -4.18722719e-01 6.46854937e-01 2.27769181e-01
-1.85521170e-01 4.56964254e-01 1.11463773e+00 -2.96856500e-02
-1.28889692e+00 -1.41174662e+00 -1.66138425e-01 4.84707117e-01
4.17805165e-02 -4.18305725e-01 -8.33704412e-01 -4.12130147e-01
2.00146273e-01 4.39782560e-01 -8.64370167e-01 -3.22522759e-01
-2.35493645e-01 -5.77782810e-01 1.05434014e-02 1.36451170e-01
-8.19252953e-02 -4.24524426e-01 -5.81806265e-02 2.63547927e-01
8.80873129e-02 -6.23824358e-01 -9.06401098e-01 7.39845680e-03
-1.19381845e+00 -1.04106438e+00 -8.91830087e-01 -8.03993106e-01
1.18645740e+00 4.06352907e-01 6.44270122e-01 -2.95791388e-01
-1.89931795e-01 6.23551130e-01 1.04037665e-01 1.44562557e-01
-3.36216778e-01 5.36139915e-03 3.41742128e-01 3.88611257e-01
1.86758712e-01 -8.20059240e-01 -6.16227090e-01 5.23012936e-01
-9.91434753e-01 7.46453926e-02 2.59343654e-01 8.07472348e-01
6.13706052e-01 8.17255527e-02 3.11353326e-01 -7.16656208e-01
5.69753826e-01 -9.20607507e-01 -9.36860442e-01 1.51285052e-01
-6.25491917e-01 3.63719076e-01 2.37388283e-01 -8.40808392e-01
-1.95254400e-01 1.52078822e-01 7.11386502e-01 -6.49028122e-01
2.40215182e-01 6.73760593e-01 -2.36558273e-01 1.93175390e-01
4.72461849e-01 -2.69648194e-01 3.25000256e-01 -2.24634454e-01
7.07114697e-01 1.20980769e-01 -9.87498984e-02 -1.80935085e-01
1.10654831e+00 3.88896525e-01 6.29108787e-01 -1.03886390e+00
-3.09879839e-01 -2.89139390e-01 -9.54274297e-01 -1.68218315e-01
9.69501734e-01 -5.41513622e-01 -6.98948205e-01 2.14848578e-01
-1.02089274e+00 -3.25142741e-01 -1.67603493e-01 6.16639316e-01
-7.58075893e-01 2.61189610e-01 -6.12386703e-01 -8.08705866e-01
2.10614651e-01 -8.83673310e-01 8.25376391e-01 8.12648833e-02
-3.33793551e-01 -1.57826912e+00 5.01560509e-01 -3.26234668e-01
1.24741919e-01 1.44450545e-01 1.28158832e+00 -3.96114081e-01
-7.56591320e-01 -5.95107317e-01 6.17249906e-02 -1.88512027e-01
2.39459217e-01 1.97789475e-01 -1.69255763e-01 -3.43932420e-01
2.99079828e-02 3.86752844e-01 5.04581988e-01 8.75135124e-01
5.95490038e-01 -2.05585018e-01 -8.19835424e-01 4.91404235e-01
1.24252880e+00 -1.73656121e-01 2.44768620e-01 -4.31585103e-01
8.86299431e-01 7.58447945e-01 1.88073590e-01 2.92396903e-01
1.40466064e-01 4.49317783e-01 3.59392107e-01 1.13458648e-01
3.98885936e-01 -6.32162690e-01 2.63503343e-01 9.62288797e-01
2.39509761e-01 -4.28194523e-01 -8.96793425e-01 5.80090284e-01
-2.00877333e+00 -8.27976108e-01 -5.67553997e-01 2.42936206e+00
4.61798072e-01 -3.92941758e-02 1.93191230e-01 7.78903998e-03
1.20603275e+00 1.52370006e-01 -6.25077605e-01 2.52649337e-01
-1.25348747e-01 -2.19197378e-01 6.72553301e-01 7.55810916e-01
-7.45745540e-01 6.94747388e-01 6.86214876e+00 2.37581775e-01
-7.83476889e-01 -6.68419972e-02 5.80433130e-01 3.51914972e-01
-4.67928678e-01 2.86468953e-01 -7.14452624e-01 2.33405650e-01
8.70810628e-01 -5.41287601e-01 2.98896700e-01 5.93857348e-01
2.02799514e-01 2.87647605e-01 -1.21945918e+00 5.37944496e-01
-1.93615377e-01 -1.11703253e+00 -3.65162045e-02 8.33551049e-01
9.40091252e-01 -2.56272405e-01 5.91778040e-01 -3.11086625e-01
6.57379925e-01 -9.75255370e-01 1.42044231e-01 6.54753983e-01
7.17036486e-01 -5.89913607e-01 6.78010881e-02 3.83861154e-01
-1.06546640e+00 1.73422173e-01 -6.31019473e-01 -1.46394712e-03
3.51572096e-01 7.74488330e-01 -9.41398621e-01 -1.31390123e-02
-7.79425353e-02 9.21817303e-01 -2.84193784e-01 8.60701025e-01
-4.88658488e-01 7.98639715e-01 -3.82668018e-01 -7.41859823e-02
-2.50129849e-01 -9.40399528e-01 7.65819073e-01 4.56696004e-01
3.89873743e-01 2.92565301e-02 8.04421380e-02 1.23799837e+00
-2.29589045e-02 2.36081600e-01 -9.62487817e-01 -4.77569997e-01
4.50482756e-01 1.36183810e+00 -1.01908791e+00 -1.75099969e-02
-3.83510321e-01 9.45813000e-01 3.53313029e-01 6.43195570e-01
-6.47335589e-01 -2.61163503e-01 8.43838990e-01 4.88743156e-01
2.18548030e-01 -6.77570760e-01 5.89636446e-04 -1.13848019e+00
-1.57317132e-01 -1.63197878e-03 1.08736612e-01 -5.35499275e-01
-1.41976786e+00 2.65695721e-01 -1.03870947e-02 -1.05367386e+00
-3.15566450e-01 -4.50950205e-01 -5.48561394e-01 9.88369465e-01
-7.43128777e-01 -8.63619506e-01 1.23771176e-01 6.30084217e-01
-7.37013221e-02 -2.01429669e-02 8.26359510e-01 -1.24083944e-01
-5.00441194e-01 2.11666659e-01 3.48849207e-01 1.95939213e-01
2.30592251e-01 -1.41203928e+00 6.80533648e-01 5.16550660e-01
4.81196523e-01 1.05461943e+00 8.98763061e-01 -1.11488140e+00
-1.46417725e+00 -1.07813728e+00 9.24543083e-01 -3.81573886e-01
1.19995236e+00 -6.04338586e-01 -9.08487260e-01 1.07400095e+00
-4.90506701e-02 3.36097360e-01 8.02669466e-01 1.66543826e-01
-1.57087952e-01 2.81792611e-01 -9.52975929e-01 8.84318233e-01
1.06824088e+00 -3.98289055e-01 -2.28630528e-01 4.20922518e-01
8.03662598e-01 2.82951593e-01 -8.65338266e-01 2.36162350e-01
2.73705095e-01 -3.36945623e-01 8.38890076e-01 -7.02082336e-01
2.44021431e-01 -3.17169338e-01 -1.66149274e-01 -1.65841269e+00
-7.63699591e-01 -7.37076342e-01 -1.59792960e-01 1.11319542e+00
4.92283314e-01 -6.29785597e-01 1.18217492e+00 7.48862624e-01
4.41832930e-01 -4.85590011e-01 -7.79577255e-01 -5.90590596e-01
1.96186349e-01 -6.19660243e-02 2.25436434e-01 8.14277947e-01
2.38264054e-01 8.29634905e-01 -3.83258164e-01 3.47692728e-01
1.26884711e+00 -1.55301943e-01 7.21676171e-01 -1.65556753e+00
-3.00867021e-01 -2.63673574e-01 -8.97279561e-01 -1.31798685e+00
3.07547688e-01 -1.03318012e+00 -1.53762057e-01 -1.36262429e+00
2.62801081e-01 -4.68853921e-01 -8.14081654e-02 -1.58048853e-01
-5.36590330e-02 -3.63313854e-02 -1.51034117e-01 2.92208284e-01
-3.08636457e-01 8.70377302e-01 1.02013457e+00 6.34834170e-02
-3.95814747e-01 3.69926453e-01 -2.50115007e-01 7.53104866e-01
6.57489359e-01 -6.82354867e-01 -6.33312702e-01 -1.52140856e-01
3.32351774e-01 4.93021876e-01 1.49644166e-01 -5.58677077e-01
2.71422684e-01 -1.37753904e-01 2.16147825e-01 -6.02856755e-01
4.15739954e-01 -8.99749279e-01 4.74873632e-01 5.35158575e-01
-4.69574481e-01 1.56633496e-01 -4.97407615e-01 1.11979747e+00
2.66228318e-01 -2.30776682e-01 6.76279843e-01 5.72493315e-01
2.14441344e-01 7.68518865e-01 -4.25003290e-01 -1.19240418e-01
1.31733978e+00 5.75013049e-02 -1.23657443e-01 -7.23924100e-01
-9.91336048e-01 2.44003206e-01 5.72919548e-01 1.94141597e-01
7.75290608e-01 -1.64224350e+00 -4.85589683e-01 4.67806906e-01
2.88327001e-02 -2.77551502e-01 9.06438008e-02 7.54135311e-01
-4.80125956e-02 2.66769826e-01 2.01900601e-02 -9.28530395e-01
-1.12117279e+00 8.86622965e-01 9.66132283e-02 -6.86497763e-02
-3.60886186e-01 6.95583403e-01 5.81838369e-01 -5.05371630e-01
-1.15358122e-02 8.53328183e-02 -1.40226737e-01 1.12590678e-02
2.15300635e-01 3.83200675e-01 -7.44799316e-01 -9.29337680e-01
9.08592269e-02 7.40124822e-01 2.83615999e-02 -5.11954784e-01
1.38447714e+00 -1.64970934e-01 -3.46332848e-01 9.64293897e-01
1.60775185e+00 1.19142555e-01 -1.41414487e+00 -4.78450358e-01
1.81149915e-01 -4.94118571e-01 -2.23194450e-01 1.42609090e-01
-1.09809089e+00 8.92266512e-01 3.65272075e-01 6.10227585e-01
4.18739319e-01 4.20823604e-01 5.29472530e-02 -1.98933885e-01
2.12671310e-01 -6.33115470e-01 3.99511099e-01 2.21655414e-01
6.38377726e-01 -7.60647595e-01 -1.22235410e-01 -7.49157250e-01
-1.74248219e-01 9.59807813e-01 8.45842063e-02 -5.63190579e-01
1.17731893e+00 -4.06888239e-02 -2.29371175e-01 -3.00846249e-01
-7.44136870e-01 2.99640983e-01 2.63556570e-01 7.35558093e-01
5.59714794e-01 4.64492738e-01 -3.23058426e-01 7.06428066e-02
-1.26578480e-01 -5.10870099e-01 6.42206967e-01 4.94819552e-01
-5.04319251e-01 -1.37924492e+00 -2.26393595e-01 4.55528587e-01
-1.44121096e-01 6.34186417e-02 -1.08200029e-01 4.35376465e-01
-4.41775948e-01 8.21160078e-01 3.07098299e-01 -2.69922167e-01
3.14528011e-02 -2.88440753e-02 3.83818358e-01 -7.10844576e-01
7.10227013e-01 3.77213001e-01 -6.83737636e-01 -1.47423670e-01
-2.07906198e-02 -1.07418954e+00 -1.12150073e+00 -3.65768880e-01
-4.41706270e-01 3.09920102e-01 8.25139403e-01 4.60349172e-01
4.15981978e-01 2.46469155e-01 9.45354521e-01 -5.13511121e-01
-6.47069931e-01 -9.65428174e-01 -9.53834236e-01 8.68066326e-02
1.69199169e-01 -7.14323580e-01 -6.52769804e-01 1.43139020e-01]
|
[7.093810558319092, 5.223530292510986]
|
9b40103d-4c5a-44b3-a1f6-892e20d9b2bc
|
applying-a-generic-sequence-to-sequence-model
|
2201.05302
| null |
https://arxiv.org/abs/2201.05302v1
|
https://arxiv.org/pdf/2201.05302v1.pdf
|
Applying a Generic Sequence-to-Sequence Model for Simple and Effective Keyphrase Generation
|
In recent years, a number of keyphrase generation (KPG) approaches were proposed consisting of complex model architectures, dedicated training paradigms and decoding strategies. In this work, we opt for simplicity and show how a commonly used seq2seq language model, BART, can be easily adapted to generate keyphrases from the text in a single batch computation using a simple training procedure. Empirical results on five benchmarks show that our approach is as good as the existing state-of-the-art KPG systems, but using a much simpler and easy to deploy framework.
|
['Alfio Gliozzo', 'Nandana Mihindukulasooriya', 'Michael Glass', 'Gaetano Rossiello', 'Md Faisal Mahbub Chowdhury']
|
2022-01-14
| null | null | null | null |
['keyphrase-generation']
|
['natural-language-processing']
|
[ 8.30154568e-02 -3.21922712e-02 -6.56929836e-02 -3.75785120e-02
-9.33841407e-01 -6.45050168e-01 1.13370311e+00 1.36108130e-01
-7.12360084e-01 1.10699987e+00 3.35830390e-01 -6.50202215e-01
1.54366091e-01 -7.84527898e-01 -6.82155848e-01 -3.79745066e-01
9.99145284e-02 4.15880054e-01 2.84925878e-01 -8.47090006e-01
5.06584585e-01 3.00448298e-01 -1.35799754e+00 4.20899034e-01
4.72232491e-01 8.84831309e-01 6.50207028e-02 1.17371440e+00
-5.02007067e-01 1.12213886e+00 -8.91268432e-01 -5.96610725e-01
2.97430515e-01 -4.27734911e-01 -8.89830470e-01 -5.65689027e-01
4.45651859e-01 -5.60894191e-01 -4.62030381e-01 6.85918152e-01
8.94936144e-01 -2.29077339e-02 2.49310166e-01 -8.89949739e-01
-5.12752652e-01 1.23559880e+00 1.20733410e-01 2.40637049e-01
5.83179295e-01 -7.31394514e-02 1.04072404e+00 -9.82799172e-01
5.60129464e-01 1.06723356e+00 5.11125565e-01 5.51350117e-01
-1.10646105e+00 -6.70923889e-01 6.55716285e-03 1.01812199e-01
-1.37278628e+00 -4.07480121e-01 4.14539337e-01 1.95907637e-01
1.38890636e+00 5.33574820e-01 7.57278681e-01 1.30991602e+00
6.16084754e-01 9.91315722e-01 1.26705420e+00 -6.46296322e-01
1.89233571e-01 6.28523752e-02 2.30881721e-02 4.91506904e-01
1.54851198e-01 -8.32109363e-04 -8.77101123e-01 -4.56596106e-01
4.16437924e-01 -1.75501168e-01 -1.90920562e-01 1.81965545e-01
-1.68201315e+00 7.49390721e-01 -4.98088785e-02 3.39020222e-01
-4.29751456e-01 5.39713621e-01 4.99358088e-01 7.15070426e-01
5.92724800e-01 6.40012026e-01 -5.24663866e-01 -6.44285023e-01
-1.21217930e+00 8.31521571e-01 1.43998826e+00 1.09207630e+00
5.82286716e-01 1.61436960e-01 -6.44160450e-01 3.86319727e-01
-1.41467392e-01 3.58732551e-01 6.80541754e-01 -4.87832993e-01
5.59443772e-01 1.32884562e-01 8.61954167e-02 -7.17081308e-01
-2.39632756e-01 -9.20083225e-02 -9.31667209e-01 -3.70147586e-01
5.62718138e-02 -3.90876293e-01 -1.04862678e+00 1.21162295e+00
-1.06361024e-01 3.09759527e-01 1.82439566e-01 3.15298736e-01
8.90352011e-01 9.97367203e-01 -7.93445632e-02 -4.61144894e-02
1.25244164e+00 -9.41883743e-01 -7.07435906e-01 -2.55940318e-01
5.55036366e-01 -1.02531385e+00 9.45149720e-01 6.62271738e-01
-1.31810105e+00 -3.80291730e-01 -8.64845216e-01 -3.56310606e-03
-8.20828140e-01 1.61025777e-01 1.09081233e+00 6.94254935e-01
-1.35177088e+00 7.09394038e-01 -4.92395997e-01 -4.58016455e-01
-4.56209257e-02 2.43746772e-01 -1.94517016e-01 2.05303580e-01
-1.54421258e+00 7.19887614e-01 9.85935628e-01 -2.51622684e-03
-9.08707380e-01 -7.92835057e-01 -5.78473330e-01 2.65496466e-02
7.09250808e-01 -9.58012164e-01 1.66274142e+00 -3.76909852e-01
-2.24406815e+00 5.12647748e-01 2.71623060e-02 -8.25130165e-01
4.39251512e-01 -6.33506656e-01 -3.72409493e-01 1.42911553e-01
-3.26504886e-01 6.25854492e-01 9.90074754e-01 -5.55926323e-01
-5.19262195e-01 3.87528807e-01 2.62594074e-01 5.96859120e-02
-1.63328767e-01 6.49776101e-01 -4.11916405e-01 -1.23451114e+00
-5.51857054e-01 -8.33675504e-01 -4.46255773e-01 -5.20099938e-01
-5.28809786e-01 -2.25977674e-01 2.31019378e-01 -5.59314668e-01
1.69936407e+00 -1.61125958e+00 -6.14132285e-02 3.12388502e-02
-3.30944434e-02 6.33565128e-01 -2.21916422e-01 1.28374612e+00
2.04673246e-01 3.27225655e-01 -8.67033750e-02 -1.54527426e-01
2.01797217e-01 6.84441030e-02 -9.80602503e-01 -1.03154741e-01
1.65992752e-01 1.31605101e+00 -1.08539414e+00 -2.45044068e-01
-1.17796533e-01 1.97600037e-01 -5.53201437e-01 3.83324832e-01
-6.13598883e-01 -9.60813686e-02 -2.37312630e-01 4.12400872e-01
3.90209615e-01 -1.43776655e-01 8.50169212e-02 1.50725380e-01
-2.81750023e-01 9.03079152e-01 -1.28079200e+00 1.84572232e+00
-3.85494351e-01 5.07761300e-01 -8.11937898e-02 -8.12246561e-01
8.40993464e-01 5.44202924e-01 6.73502833e-02 -3.99869472e-01
1.38868108e-01 3.90746683e-01 -1.82853147e-01 -1.39535330e-02
1.28504002e+00 2.28504047e-01 -2.21299484e-01 9.32742298e-01
2.59837836e-01 -6.11875355e-01 6.04197562e-01 6.32031024e-01
1.08561754e+00 -2.44419016e-02 5.28532147e-01 -2.18891159e-01
5.69375336e-01 -1.06510393e-01 -4.13282309e-03 1.29192269e+00
3.39740396e-01 4.15909529e-01 3.94961506e-01 -4.92994130e-01
-1.01582372e+00 -6.58833683e-01 4.96324778e-01 1.10292506e+00
-3.78710747e-01 -1.33519566e+00 -6.40133798e-01 -6.40064061e-01
-2.58193702e-01 7.04099953e-01 -2.94789106e-01 -3.34887654e-02
-8.17785144e-01 -7.39637733e-01 1.13076782e+00 4.53448296e-01
4.40452099e-01 -1.32479835e+00 -4.91224617e-01 6.06756747e-01
1.36688933e-01 -1.10460079e+00 -4.07848597e-01 2.67229706e-01
-5.79166830e-01 -5.00701368e-01 -1.02211237e+00 -4.53011543e-01
3.05974245e-01 3.58960092e-01 1.45691335e+00 1.00589842e-02
-1.93764254e-01 3.91451538e-01 -7.07516730e-01 -6.99036479e-01
-7.23336101e-01 5.71367979e-01 -3.16415168e-02 -1.53310254e-01
1.66713715e-01 -3.96055371e-01 -4.37720358e-01 -3.38746101e-01
-1.24063694e+00 3.48934084e-01 9.10123527e-01 8.70727777e-01
3.50502253e-01 -1.16668150e-01 5.01101375e-01 -1.08218133e+00
1.23760819e+00 -3.36784065e-01 -5.66277266e-01 3.14112157e-01
-9.43658829e-01 2.74573714e-01 1.01381052e+00 -3.99352968e-01
-8.34551752e-01 -4.06517804e-01 -3.13437462e-01 9.64577794e-02
1.37608042e-02 6.98796570e-01 3.88822675e-01 -5.30439653e-02
5.03550828e-01 9.13556933e-01 -2.41708413e-01 -5.18389106e-01
8.08541417e-01 7.15378284e-01 3.28160912e-01 -6.66553736e-01
1.03163147e+00 2.34744713e-01 -1.67105779e-01 -5.93524039e-01
-7.01158047e-01 -3.49975556e-01 -4.36341852e-01 1.23760812e-01
1.65737212e-01 -1.03295922e+00 -3.60426366e-01 7.42488027e-01
-1.36306655e+00 -5.34421623e-01 -2.81287819e-01 2.42294207e-01
-5.12096465e-01 4.91445750e-01 -6.66544795e-01 -5.35876632e-01
-1.12195420e+00 -7.72411764e-01 1.19121194e+00 1.64043039e-01
-3.42646211e-01 -9.05298948e-01 1.44736901e-01 -1.77006304e-01
9.38841701e-01 -3.30532104e-01 6.90109015e-01 -7.99211025e-01
-6.42927229e-01 -3.61295670e-01 -1.19221143e-01 4.52389359e-01
-1.05441339e-01 -2.95553301e-02 -6.31640196e-01 -2.88879126e-01
-4.49683875e-01 -4.47892040e-01 9.61410999e-01 -1.61969736e-01
1.10914314e+00 -6.53815746e-01 -3.64104621e-02 6.09286189e-01
1.27826941e+00 -1.25459686e-01 6.57820761e-01 3.55147332e-01
3.87283653e-01 2.78698001e-02 5.55550098e-01 4.54886854e-01
6.17164671e-01 4.90745664e-01 -1.49235681e-01 1.22769125e-01
4.41922173e-02 -5.95856130e-01 8.33525419e-01 1.10845006e+00
2.49865726e-02 -3.29575181e-01 -7.13009536e-01 4.19744372e-01
-1.88334501e+00 -1.01003706e+00 1.17067486e-01 1.86042273e+00
1.40617812e+00 1.78319409e-01 9.29978192e-02 1.23468302e-01
-5.63491210e-02 5.75675488e-01 1.41460314e-01 -7.84426689e-01
-1.62381440e-01 9.15429771e-01 7.99932361e-01 2.96454847e-01
-9.26318467e-01 1.39549839e+00 8.25481033e+00 1.07924831e+00
-1.25365734e+00 -2.10006058e-01 2.47939363e-01 -1.56674534e-01
-4.89397943e-01 4.57520448e-02 -1.15740943e+00 4.48267430e-01
1.38169718e+00 -5.59636295e-01 6.09628141e-01 6.27739012e-01
9.16649699e-02 9.44442023e-03 -9.46999788e-01 9.70744193e-01
1.56887382e-01 -1.66871417e+00 4.45928484e-01 -2.93386787e-01
5.59600353e-01 1.76099926e-01 -2.80720085e-01 4.97932553e-01
5.85972607e-01 -8.53524804e-01 8.39275241e-01 7.21241534e-01
7.51546979e-01 -6.42750978e-01 5.84694743e-01 4.99105930e-01
-1.16095436e+00 6.55215755e-02 -3.91378105e-01 -2.59181172e-01
2.19069287e-01 4.83085841e-01 -9.15042996e-01 9.57178235e-01
6.94260895e-01 3.88467968e-01 -7.85747290e-01 8.61918628e-01
-5.90124130e-01 9.09760833e-01 -3.09915960e-01 -5.49927413e-01
5.30805945e-01 8.93431082e-02 5.15715480e-01 1.68402815e+00
3.29165220e-01 -1.37977200e-02 2.58486181e-01 5.95007598e-01
-5.87662756e-02 3.40307087e-01 -4.99987483e-01 -4.19115961e-01
3.35794717e-01 1.39730394e+00 -5.13116837e-01 -8.03174376e-01
-2.90566266e-01 1.14487898e+00 1.90474138e-01 2.35344768e-01
-6.63505256e-01 -8.03773522e-01 3.89587045e-01 -1.58692837e-01
4.06670064e-01 -4.51789260e-01 1.77188903e-01 -1.45435798e+00
4.06530164e-02 -1.35062373e+00 3.08271170e-01 -7.24304080e-01
-1.02985418e+00 7.67220676e-01 2.33347073e-01 -9.72312748e-01
-6.95173204e-01 -6.11381292e-01 -5.27547121e-01 8.89804780e-01
-1.83060551e+00 -1.11958778e+00 -1.48923635e-01 4.86296624e-01
5.15759528e-01 -3.68453294e-01 1.09592795e+00 9.28175598e-02
-2.74037600e-01 6.96724534e-01 1.13088049e-01 8.74660164e-02
7.73610830e-01 -1.42039025e+00 1.24219835e+00 9.21109080e-01
4.62380499e-01 8.87491703e-01 5.15363634e-01 -5.18833458e-01
-1.90293074e+00 -9.16702867e-01 1.33891940e+00 -1.77902415e-01
1.01877260e+00 -7.12018669e-01 -5.52353024e-01 5.36534309e-01
6.63411856e-01 -3.25951099e-01 5.29630244e-01 -3.51170003e-01
-2.98972696e-01 -1.16474688e-01 -4.15011734e-01 1.00286639e+00
8.48265171e-01 -6.60965204e-01 -5.98223805e-01 2.23932728e-01
8.36112916e-01 -6.38547540e-01 -6.62908554e-01 3.02974224e-01
5.91513097e-01 -6.95162535e-01 8.37370932e-01 -6.78299069e-01
2.46934921e-01 -1.82856277e-01 3.98380272e-02 -1.47220480e+00
-2.06740513e-01 -1.73505592e+00 -7.08796620e-01 1.08293033e+00
2.77779013e-01 -7.28713572e-01 5.25646389e-01 2.03852683e-01
6.95193708e-02 -8.32431674e-01 -6.74844086e-01 -7.65271723e-01
-1.25729358e-02 -7.44625747e-01 7.81903803e-01 4.29642051e-01
-3.41516025e-02 6.25336468e-01 -6.89849615e-01 -2.91834503e-01
1.20795473e-01 -6.09094910e-02 1.22307563e+00 -9.18999851e-01
-4.57032233e-01 -4.82450098e-01 4.59776968e-02 -1.45446181e+00
4.09263372e-02 -9.04009104e-01 -2.18389332e-01 -1.57520759e+00
-8.35196078e-02 -1.82945281e-01 -2.90687203e-01 5.55328548e-01
-4.83046114e-01 2.19922155e-01 4.47800100e-01 5.70954457e-02
-5.56970358e-01 5.29247761e-01 9.29351628e-01 -2.58307159e-01
-1.54772103e-01 2.87925685e-03 -9.60754097e-01 2.49761403e-01
8.16147506e-01 -5.10350883e-01 -3.40947688e-01 -2.39970863e-01
9.19474661e-01 -2.85163969e-01 3.05702150e-01 -8.34986508e-01
5.13909459e-01 -1.12980017e-02 -1.64378762e-01 -8.12730014e-01
1.97806984e-01 -4.73297060e-01 5.78748398e-02 4.84344900e-01
-3.85797799e-01 2.36919165e-01 4.16414827e-01 3.58142167e-01
-2.78156251e-01 -2.26972476e-01 2.29487911e-01 -3.42562139e-01
-6.80186093e-01 2.58984596e-01 -7.61699200e-01 -1.77249443e-02
7.15023637e-01 1.54137284e-01 -2.27998272e-01 -6.80837274e-01
2.46120505e-02 -7.91020766e-02 -1.67120993e-02 5.96347630e-01
7.15527534e-01 -1.05300939e+00 -1.09130275e+00 2.81040132e-01
1.16571918e-01 -2.71667659e-01 -3.11445653e-01 6.56453192e-01
-6.86384261e-01 1.05954444e+00 1.81739286e-01 2.21190089e-03
-9.74761963e-01 4.36550856e-01 1.18422844e-01 -8.82698357e-01
-6.48711622e-01 7.22579479e-01 -3.94338429e-01 -4.21156675e-01
1.37371823e-01 -6.78437829e-01 5.41074872e-02 1.11241937e-01
9.40976322e-01 1.96939081e-01 3.17522883e-01 -1.31552279e-01
-4.67089042e-02 1.85530424e-01 -4.72337127e-01 -1.75232261e-01
1.05966568e+00 2.57090330e-01 -2.88403124e-01 4.34370190e-01
6.61329627e-01 1.16303779e-01 -5.39388299e-01 -4.84441012e-01
3.73865594e-03 -9.44245011e-02 2.78303362e-02 -9.73153412e-01
-5.87006748e-01 6.24531388e-01 -3.45750712e-02 4.55533803e-01
1.15945733e+00 -3.76424670e-01 1.16837978e+00 8.25804293e-01
6.41419530e-01 -1.18792450e+00 -1.21789880e-01 9.60067153e-01
8.27364564e-01 -7.80858517e-01 1.94542125e-01 -1.57548279e-01
-4.60220724e-01 1.29327154e+00 4.48485650e-02 -1.51139468e-01
7.47627556e-01 5.33175886e-01 1.07785314e-01 2.42429785e-02
-1.36627054e+00 -1.28834650e-01 3.72314155e-01 2.15400755e-01
6.28387392e-01 -5.32441586e-02 -6.80855274e-01 5.64037859e-01
-6.29195035e-01 2.54673183e-01 6.88848615e-01 1.14383042e+00
-2.30073169e-01 -1.48872781e+00 -1.75126269e-01 4.83092248e-01
-1.08218694e+00 -8.35203528e-01 -6.29372418e-01 7.12262690e-01
-4.42253441e-01 7.16783643e-01 -2.43397385e-01 -3.96854311e-01
1.74303964e-01 2.16047496e-01 4.35013473e-01 -7.70479262e-01
-1.13054895e+00 -1.78358123e-01 3.05949390e-01 -6.82465971e-01
-2.97664940e-01 -5.48823290e-02 -8.15452516e-01 -5.30331373e-01
-2.36677036e-01 2.77345717e-01 4.81983215e-01 7.97223866e-01
7.67955661e-01 3.95738453e-01 5.39133191e-01 -8.60981524e-01
-9.29595113e-01 -1.23780811e+00 -4.90836531e-01 -1.82802930e-01
1.54369473e-01 1.89019606e-01 -2.30633914e-02 7.56403208e-02]
|
[12.263518333435059, 8.90402603149414]
|
d852ece1-a362-415d-9629-27245b75a070
|
generative-adversarial-network-applications
|
2201.09152
| null |
https://arxiv.org/abs/2201.09152v1
|
https://arxiv.org/pdf/2201.09152v1.pdf
|
Generative Adversarial Network Applications in Creating a Meta-Universe
|
Generative Adversarial Networks (GANs) are machine learning methods that are used in many important and novel applications. For example, in imaging science, GANs are effectively utilized in generating image datasets, photographs of human faces, image and video captioning, image-to-image translation, text-to-image translation, video prediction, and 3D object generation to name a few. In this paper, we discuss how GANs can be used to create an artificial world. More specifically, we discuss how GANs help to describe an image utilizing image/video captioning methods and how to translate the image to a new image using image-to-image translation frameworks in a theme we desire. We articulate how GANs impact creating a customized world.
|
['Hamid R. Arabnia', 'Khaled Rasheed', 'Thiab R. Taha', 'Soheyla Amirian']
|
2022-01-23
| null | null | null | null |
['video-prediction']
|
['computer-vision']
|
[ 1.01074851e+00 5.05593359e-01 1.05783194e-01 -2.82874823e-01
-6.10749960e-01 -7.09818482e-01 6.97562933e-01 -8.54750395e-01
1.55095577e-01 9.42191064e-01 2.23741561e-01 -1.78606972e-01
4.74024028e-01 -9.28216577e-01 -1.15940070e+00 -5.88640451e-01
6.50519967e-01 2.99008071e-01 -3.88560474e-01 -1.46733195e-01
1.05520800e-01 6.97986066e-01 -1.16095054e+00 4.66519356e-01
4.39547986e-01 9.57529366e-01 2.31163949e-02 8.96103203e-01
-2.27133721e-01 9.25237834e-01 -8.08567166e-01 -7.71465600e-01
3.00377756e-01 -9.70691502e-01 -5.86096883e-01 1.93685070e-01
5.01188517e-01 -5.60629547e-01 -3.81338745e-01 9.85692680e-01
4.32993144e-01 -3.67655158e-01 9.28835154e-01 -1.54291427e+00
-1.27690196e+00 2.09005937e-01 -5.17874718e-01 -3.91369730e-01
6.24991536e-01 3.79843742e-01 4.39444184e-02 -7.55677521e-01
1.04876292e+00 1.31843197e+00 4.20399725e-01 1.18347323e+00
-1.10571659e+00 -5.93922079e-01 -5.94960868e-01 -1.40182376e-01
-9.14225519e-01 -4.84911621e-01 9.84743476e-01 -5.22223115e-01
4.62150425e-01 3.79180670e-01 7.41919219e-01 1.49347341e+00
5.36219358e-01 8.20056081e-01 1.11739624e+00 -5.72127283e-01
5.73741347e-02 1.22797295e-01 -1.03986263e+00 5.42548180e-01
-1.10240541e-01 2.58272171e-01 -5.65748334e-01 1.86893046e-01
1.29538667e+00 -9.25091002e-03 -1.36649441e-02 4.23376774e-03
-1.28198624e+00 8.02482545e-01 3.80529046e-01 1.73648596e-02
-4.70432609e-01 5.57671726e-01 1.52311876e-01 2.46366411e-01
3.79145086e-01 5.83506405e-01 8.99321586e-02 -1.79930050e-02
-7.79420674e-01 3.01326126e-01 4.41239625e-01 1.17766225e+00
5.70636332e-01 7.86413372e-01 -3.41219157e-01 6.48990631e-01
8.04634243e-02 9.15282190e-01 6.18202329e-01 -1.21316540e+00
2.05620036e-01 1.94435135e-01 5.94536774e-03 -7.54270375e-01
4.79110181e-02 1.10735886e-01 -9.52959120e-01 4.06253010e-01
-1.72829345e-01 -2.38443345e-01 -1.31563461e+00 1.49020171e+00
1.87901586e-01 1.62517890e-01 2.03362375e-01 6.87239885e-01
1.10596609e+00 9.09504771e-01 3.22541073e-02 -2.35419758e-02
1.22267294e+00 -1.01478779e+00 -8.10584784e-01 -5.06959140e-01
-1.17397927e-01 -1.08503592e+00 8.22067261e-01 1.47517510e-02
-1.55000436e+00 -7.34294772e-01 -8.51165771e-01 -1.81630805e-01
-3.87565881e-01 -8.08273256e-02 4.48778957e-01 7.77473211e-01
-1.29009032e+00 1.78733543e-01 -4.54293132e-01 -2.84616828e-01
7.60569394e-01 5.42709492e-02 -5.47532678e-01 -1.74357012e-01
-8.99045467e-01 9.09325480e-01 3.62100035e-01 -2.14034885e-01
-1.21565259e+00 -7.62523592e-01 -9.57311988e-01 -2.92238653e-01
-5.66572044e-03 -1.44764447e+00 1.18723190e+00 -1.36422193e+00
-1.76573312e+00 1.35837162e+00 -8.09459835e-02 -5.95067024e-01
5.68862796e-01 1.67035963e-02 -4.16919351e-01 2.40931109e-01
1.44441975e-02 1.42916715e+00 1.50958824e+00 -1.27176750e+00
-8.71004537e-02 -6.88314214e-02 -9.70476046e-02 3.56571488e-02
2.25285932e-01 3.42473537e-02 -2.21056044e-01 -1.13354731e+00
-1.67385355e-01 -9.27914500e-01 -3.89785469e-02 2.54557788e-01
-5.39152682e-01 4.87262756e-01 1.28294325e+00 -7.09558547e-01
2.64806390e-01 -2.00323915e+00 1.46960780e-01 -2.05721945e-01
1.38154253e-01 4.58304077e-01 -4.25191224e-01 5.98001778e-01
-1.69506192e-01 3.06555867e-01 -2.94477791e-01 -2.79522330e-01
-2.52371252e-01 1.16330303e-01 -5.11314392e-01 -3.46277952e-02
6.11808658e-01 1.82483518e+00 -6.93483293e-01 -5.00810444e-01
5.51424861e-01 8.49565089e-01 -5.19137323e-01 5.72320044e-01
-2.27102265e-01 1.24166477e+00 -3.00931871e-01 8.21004629e-01
6.70571625e-01 5.61515205e-02 -2.35759705e-01 -2.40735754e-01
3.63711357e-01 -4.66759324e-01 -2.08169907e-01 1.58341396e+00
-5.36316812e-01 9.38709676e-01 -5.16086034e-02 -8.41246367e-01
9.58604753e-01 3.15064609e-01 4.30761278e-01 -6.25152528e-01
4.38811183e-02 1.35922670e-01 -2.29603037e-01 -6.67229712e-01
3.50150734e-01 -4.21685487e-01 1.55325666e-01 4.73945260e-01
5.47748767e-02 -1.02264369e+00 2.54518911e-02 6.68168068e-02
7.61981010e-01 4.36055839e-01 1.00674719e-01 3.19419771e-01
4.10122424e-01 4.79123220e-02 -6.78595603e-02 5.36912143e-01
1.23881213e-01 1.32466495e+00 2.94180244e-01 -6.04475379e-01
-1.74186540e+00 -1.18948603e+00 3.89103055e-01 4.06793833e-01
-5.96079119e-02 1.72025368e-01 -1.16176748e+00 -4.59599644e-01
-3.52874607e-01 7.72510648e-01 -7.06795990e-01 -3.51908088e-01
-5.12537479e-01 -3.11832994e-01 5.37887812e-01 3.13791096e-01
6.15477502e-01 -1.70114386e+00 -3.93818349e-01 1.70838013e-01
-1.24618284e-01 -1.44838607e+00 -7.00507998e-01 -4.32808161e-01
-7.76898623e-01 -7.52715766e-01 -1.45266843e+00 -1.06046212e+00
1.09806752e+00 2.20739678e-01 1.34261346e+00 -1.22206770e-01
-5.57315350e-01 7.37698138e-01 -4.53703076e-01 -7.92373657e-01
-1.25472200e+00 -3.13527912e-01 -5.30032068e-02 -4.94353510e-02
-1.66876182e-01 -6.52166545e-01 -7.88677752e-01 2.11387634e-01
-1.52566314e+00 7.79686749e-01 7.75388062e-01 8.49940121e-01
6.20055020e-01 -4.63614792e-01 5.29783368e-01 -1.05490088e+00
6.83320403e-01 -1.92289561e-01 -3.66096020e-01 2.82481641e-01
-1.28132299e-01 -2.48963639e-01 6.92840993e-01 -5.34879148e-01
-9.59024072e-01 1.49222016e-01 -3.17533284e-01 -5.91768146e-01
-2.25464880e-01 1.75815448e-01 -1.19397722e-01 -3.71515006e-01
6.81108713e-01 6.83549643e-01 5.32225549e-01 2.31016293e-01
6.52292252e-01 5.10365963e-01 7.58230865e-01 -3.57720643e-01
1.20164537e+00 6.40556097e-01 7.64571130e-02 -6.52844250e-01
-3.99200857e-01 4.30526495e-01 -6.05132043e-01 -3.82288784e-01
1.24189758e+00 -7.54148901e-01 -5.05192935e-01 6.12267554e-01
-1.41190755e+00 -4.69502151e-01 -5.95166802e-01 5.06510586e-02
-1.11198568e+00 -2.45156914e-01 -4.11653727e-01 -4.98261780e-01
-6.89156890e-01 -1.32902241e+00 1.42033958e+00 6.38774574e-01
-1.05951272e-01 -9.51500952e-01 -2.93745041e-01 6.15012348e-01
8.46863866e-01 7.89016604e-01 7.74887562e-01 1.41649634e-01
-6.88232601e-01 -2.23244160e-01 -2.75386035e-01 5.56115210e-01
3.29354614e-01 -5.32391444e-02 -8.47612441e-01 -3.74399237e-02
-8.46881419e-02 -3.75218511e-01 3.19222093e-01 6.61973834e-01
1.34036195e+00 -5.54307580e-01 -3.18768799e-01 8.58897150e-01
1.20086908e+00 6.65366054e-01 1.37655270e+00 8.88985544e-02
9.35729086e-01 3.19953799e-01 1.07010327e-01 -8.36392790e-02
-6.76714331e-02 7.32662916e-01 4.05302972e-01 -5.56680441e-01
-7.17381895e-01 -7.15027273e-01 4.14263219e-01 4.50341791e-01
4.03784998e-02 -3.88801992e-01 -5.42787731e-01 1.94402516e-01
-1.26762342e+00 -9.75710392e-01 1.36134237e-01 1.71176171e+00
6.82855904e-01 -2.54279017e-01 -1.64871186e-01 -3.21009278e-01
8.72826159e-01 -1.24305308e-01 -8.31739962e-01 -6.07955575e-01
-2.47157976e-01 4.14942920e-01 4.69469815e-01 2.10895166e-01
-7.31783032e-01 1.21906018e+00 7.34835100e+00 7.56280720e-01
-1.38292909e+00 1.33738324e-01 1.20788825e+00 2.81179368e-01
-4.71894294e-01 -2.33395711e-01 -2.10042149e-01 4.27173883e-01
7.45740116e-01 -2.94797271e-01 6.06965780e-01 7.05411434e-01
2.51558870e-01 1.45343676e-01 -1.03759432e+00 1.36745012e+00
4.14623916e-01 -1.77240109e+00 7.53944159e-01 1.32095382e-01
1.21039176e+00 -5.10086834e-01 4.72011238e-01 -1.55489892e-01
3.07972226e-02 -1.50553823e+00 8.31012249e-01 6.63659990e-01
1.38415742e+00 -4.55683947e-01 5.34979165e-01 -1.07590191e-01
-5.21827817e-01 3.24161321e-01 -1.14796869e-01 4.10037369e-01
6.54183865e-01 4.38357145e-01 -1.13550591e+00 3.22950006e-01
4.51593906e-01 3.57523382e-01 -3.64798039e-01 7.44813800e-01
-3.06500733e-01 3.89510661e-01 1.52886957e-01 1.49429277e-01
1.50099263e-01 -4.57898438e-01 5.49236476e-01 6.85104370e-01
8.11092556e-01 8.43719020e-02 -5.11640251e-01 1.24733365e+00
-4.39698339e-01 -1.75271109e-01 -1.19653821e+00 -4.12328690e-01
1.41105995e-01 1.03704929e+00 -7.07754493e-01 -4.74353433e-01
-8.34535509e-02 1.51000273e+00 -3.19432735e-01 5.51169395e-01
-1.06520724e+00 -2.98526049e-01 4.81964022e-01 4.74041253e-01
1.16224840e-01 -8.39636475e-02 -2.12116018e-01 -9.32166100e-01
-1.52295470e-01 -1.10747182e+00 -3.65568399e-01 -1.68645644e+00
-1.17493010e+00 8.19531202e-01 -9.43636894e-02 -1.30599999e+00
-6.43923759e-01 -6.59365952e-01 -8.16800416e-01 7.39778697e-01
-1.09381783e+00 -1.79825830e+00 -5.50237775e-01 4.78974372e-01
6.85114384e-01 -6.17604494e-01 7.32338071e-01 1.15839519e-01
3.52007337e-04 5.11764407e-01 8.61599669e-03 2.98887968e-01
6.75492346e-01 -8.09541643e-01 9.15003240e-01 6.64075375e-01
2.75857538e-01 3.08229476e-01 7.07263172e-01 -5.76723397e-01
-1.59522188e+00 -1.16987169e+00 1.87701881e-01 -3.71826917e-01
5.41323870e-02 -2.62683600e-01 -3.37361366e-01 7.13101566e-01
6.22059166e-01 -8.08151811e-03 4.12962496e-01 -1.08361816e+00
-1.66745558e-02 3.29796784e-02 -1.59023106e+00 8.43985379e-01
9.54951942e-01 -5.27450860e-01 -4.32687066e-02 5.75357735e-01
8.33991647e-01 -9.27186131e-01 -7.33132184e-01 2.10356653e-01
6.67767406e-01 -9.16710019e-01 1.34514606e+00 -5.29034972e-01
1.20863175e+00 -2.87111729e-01 1.08643658e-01 -1.45986819e+00
-6.52159052e-03 -9.64310169e-01 1.14007369e-01 1.18813503e+00
2.65631258e-01 -6.35110676e-01 8.71850908e-01 6.31334484e-01
-1.41446903e-01 -4.02490020e-01 -7.17200279e-01 -5.68969846e-01
1.21405795e-01 -1.08276069e-01 9.12014782e-01 7.29613781e-01
-6.90794408e-01 6.22303374e-02 -8.61232758e-01 -4.85160202e-01
4.79880780e-01 -1.74238026e-01 1.17140079e+00 -4.70994323e-01
-5.11056976e-03 -4.41048026e-01 -6.63973153e-01 -8.35773289e-01
5.37054017e-02 -8.24868023e-01 -1.93817541e-01 -1.53717160e+00
1.00633390e-01 -1.28270835e-01 4.94315356e-01 2.69012094e-01
7.91955218e-02 9.84041452e-01 5.33080161e-01 2.15631008e-01
9.63979289e-02 4.73780781e-01 1.80495059e+00 -4.38940287e-01
1.12653323e-01 -1.39426246e-01 -8.66360486e-01 3.34099829e-01
9.00668561e-01 -4.03425932e-01 -3.12263876e-01 -6.15920722e-01
9.15713161e-02 1.79588377e-01 5.42533994e-01 -8.76376629e-01
-1.77821234e-01 -3.00470531e-01 8.62036109e-01 -6.20625205e-02
6.01697385e-01 -7.00234771e-01 8.88525009e-01 5.40370226e-01
-2.48678491e-01 1.29708588e-01 2.88676292e-01 2.24677593e-01
-3.67080480e-01 -3.57367694e-01 8.94058347e-01 -5.25089681e-01
-4.55967724e-01 4.01468545e-01 -2.93435723e-01 -1.59160122e-01
1.44943726e+00 -3.40454161e-01 -4.19879556e-01 -1.15848112e+00
-4.91322726e-01 -3.91775817e-01 6.50322676e-01 6.39494002e-01
1.00377619e+00 -1.81100321e+00 -9.90917683e-01 3.75422120e-01
-7.60019943e-02 -1.37534901e-01 3.40375781e-01 3.28318447e-01
-1.15202975e+00 3.72191876e-01 -6.74024642e-01 -6.47698581e-01
-1.13425756e+00 5.16169667e-01 3.02680761e-01 2.12523386e-01
-3.67349654e-01 6.41383708e-01 5.21983027e-01 -2.80459315e-01
-4.32626247e-01 1.87078699e-01 2.35716790e-01 -6.08570695e-01
5.83207190e-01 -1.20004244e-01 -2.63236582e-01 -7.51708210e-01
7.26558715e-02 8.26896787e-01 9.30059254e-02 -2.15135738e-01
1.29144895e+00 -2.50786170e-02 -1.96318448e-01 -8.69587585e-02
1.22394538e+00 -2.17653736e-01 -1.23188007e+00 3.26499194e-01
-7.88107991e-01 -5.78406930e-01 -3.17647010e-01 -8.17473590e-01
-1.47642922e+00 8.85613203e-01 7.23304868e-01 2.57206827e-01
1.27947032e+00 2.22809650e-02 1.08457696e+00 -1.81318536e-01
3.55319470e-01 -6.60517514e-01 3.68590981e-01 7.12272525e-02
1.38787425e+00 -1.29830241e+00 -1.38698205e-01 -3.23869675e-01
-8.98143470e-01 1.20210075e+00 5.32696724e-01 7.25879893e-02
3.45822155e-01 3.72066617e-01 2.77438253e-01 1.18079312e-01
-3.79927546e-01 1.86713263e-01 3.11870426e-01 1.33676648e+00
4.34842914e-01 -1.47358896e-02 5.89138530e-02 -1.27657041e-01
-4.26240981e-01 2.73036450e-01 6.98691308e-01 6.64945960e-01
2.55741030e-01 -1.37918603e+00 -6.79079056e-01 4.21769172e-01
-5.27575374e-01 -1.76796749e-01 -5.55877805e-01 4.30426419e-01
5.59885502e-02 6.35809839e-01 -6.58373674e-03 -3.09782177e-01
-6.02865145e-02 -8.57547298e-02 7.34287262e-01 -4.51816082e-01
-2.41497502e-01 -2.36178115e-01 -3.40752572e-01 -4.04076487e-01
-6.11786962e-01 -3.05591911e-01 -7.77796090e-01 -3.94752145e-01
1.87573358e-01 -1.75406665e-01 1.05728781e+00 6.82156265e-01
6.12877369e-01 5.55457830e-01 5.74667335e-01 -1.20032883e+00
2.72354484e-01 -8.64210665e-01 -1.76447481e-01 5.85947096e-01
9.57129970e-02 -2.26146519e-01 -8.62656757e-02 8.74158084e-01]
|
[11.819926261901855, -0.38448426127433777]
|
8a39ce85-7f88-44a6-8d01-29e593edc6b5
|
ynu_dyx-at-semeval-2019-task-9-a-stacked
| null | null |
https://aclanthology.org/S19-2223
|
https://aclanthology.org/S19-2223.pdf
|
YNU\_DYX at SemEval-2019 Task 9: A Stacked BiLSTM for Suggestion Mining Classification
|
In this paper we describe a deep-learning system that competed as SemEval 2019 Task 9-SubTask A: Suggestion Mining from Online Reviews and Forums. We use Word2Vec to learn the distributed representations from sentences. This system is composed of a Stacked Bidirectional Long-Short Memory Network (SBiLSTM) for enriching word representations before and after the sequence relationship with context. We perform an ensemble to improve the effectiveness of our model. Our official submission results achieve an F1-score 0.5659.
|
['Xue-jie Zhang', 'Yunxia Ding', 'Xiaobing Zhou']
|
2019-06-01
| null | null | null |
semeval-2019-6
|
['suggestion-mining']
|
['natural-language-processing']
|
[-2.60335475e-01 2.98643202e-01 -3.70497882e-01 -7.34390974e-01
-3.95911545e-01 -5.56935705e-02 7.97511339e-01 6.59794882e-02
-8.23561788e-01 8.76736581e-01 6.63035691e-01 -9.06116366e-01
2.38714561e-01 -4.94077027e-01 -6.23101950e-01 -2.92725652e-01
-5.12665398e-02 3.47146004e-01 1.36673272e-01 -6.52107716e-01
4.51245815e-01 -6.66140914e-02 -1.20859826e+00 8.70682716e-01
4.85395253e-01 8.03311944e-01 1.39403507e-01 8.47253084e-01
-7.22250402e-01 7.49341726e-01 -7.79091239e-01 -7.70059347e-01
7.69863501e-02 -1.23321921e-01 -1.00428236e+00 -4.49538231e-01
3.21489811e-01 -4.05550987e-01 -5.03290951e-01 8.63144457e-01
4.90363061e-01 5.89550674e-01 7.40233243e-01 -5.67781806e-01
-1.13929892e+00 1.37141597e+00 -8.39777410e-01 4.32119608e-01
-7.15326704e-03 -2.10106239e-01 1.59530294e+00 -1.58232534e+00
6.23743474e-01 1.24144363e+00 3.82160753e-01 7.08580494e-01
-9.65222657e-01 -5.68463743e-01 6.19740605e-01 4.23922837e-01
-1.07081437e+00 -1.05070375e-01 5.90455472e-01 -1.49762824e-01
1.86805499e+00 9.13793817e-02 2.61547267e-01 1.48631716e+00
5.25982559e-01 1.08895183e+00 5.89048386e-01 -4.52635735e-01
1.47617370e-01 2.62914091e-01 1.05625665e+00 4.29347932e-01
2.33660355e-01 -3.12751770e-01 -9.20842767e-01 -2.45382473e-01
8.59335139e-02 9.23662633e-02 1.73999220e-01 4.54929233e-01
-6.09245777e-01 1.24532378e+00 5.12892306e-01 3.17130119e-01
-4.27728802e-01 8.23377259e-03 7.09187031e-01 4.89975452e-01
9.80954587e-01 6.91174030e-01 -1.02352250e+00 3.26694474e-02
-6.46931529e-01 3.76932979e-01 8.64749193e-01 5.07953167e-01
6.65178537e-01 8.61860961e-02 -4.26078022e-01 1.36512685e+00
4.91702765e-01 -1.09415483e-02 1.16501558e+00 -2.17547983e-01
4.73507673e-01 3.37988824e-01 -8.22516605e-02 -9.15607333e-01
-4.09240574e-01 -5.88927746e-01 -8.21523428e-01 -2.66522139e-01
-4.40589130e-01 -6.71562970e-01 -8.54215264e-01 1.21609735e+00
-5.17810769e-02 2.62787104e-01 -1.97040532e-02 6.99852645e-01
1.43293488e+00 1.02396560e+00 2.47014880e-01 -5.27785867e-02
9.81276929e-01 -1.77649271e+00 -1.06972575e+00 -3.13067526e-01
8.75504732e-01 -6.15959048e-01 1.02133238e+00 4.64573771e-01
-9.03932035e-01 -7.96092451e-01 -1.16115022e+00 -2.25307971e-01
-8.67774725e-01 3.06404997e-02 5.69396615e-01 2.25514844e-01
-1.02579927e+00 5.51854432e-01 -3.67657036e-01 -8.50640759e-02
1.24303952e-01 4.90615696e-01 -3.83141160e-01 2.57105529e-01
-1.61521053e+00 1.03405786e+00 2.59061754e-01 1.54295459e-01
-5.24003565e-01 -5.68916142e-01 -8.73257935e-01 1.11357003e-01
-1.56927984e-02 -4.18067902e-01 1.50546014e+00 -8.62771749e-01
-1.65960419e+00 7.61514425e-01 -3.27460229e-01 -1.00827098e+00
-1.78770036e-01 -7.37848818e-01 -7.97520220e-01 -5.37440360e-01
-9.15332735e-02 5.91245592e-01 5.18890858e-01 -7.04054117e-01
-5.53383052e-01 -1.87190369e-01 -1.50981769e-01 1.57863542e-01
-8.09625864e-01 2.63492793e-01 -4.72234100e-01 -7.79427648e-01
-6.64818645e-01 -6.35617554e-01 -6.85715020e-01 -9.67285991e-01
-6.95716381e-01 -9.62854981e-01 3.93897027e-01 -6.84335172e-01
1.73677528e+00 -1.84897184e+00 -1.58883542e-01 1.32964611e-01
1.06604479e-01 6.24733448e-01 -7.25069642e-01 6.75281346e-01
-2.76024435e-02 3.83200347e-01 2.19881490e-01 -8.04203570e-01
1.31275188e-02 1.06238931e-01 -6.43284142e-01 6.15984537e-02
3.52095485e-01 1.08339560e+00 -7.52471626e-01 -1.18529506e-01
-9.34236497e-02 2.98421383e-01 -3.28254759e-01 4.73529071e-01
-1.85425922e-01 -3.41363609e-01 -1.63095310e-01 6.55029193e-02
5.18352628e-01 -3.24443847e-01 2.65604198e-01 3.58706981e-01
1.37834758e-01 9.83614802e-01 -6.65954471e-01 1.80318117e+00
-5.34906328e-01 5.39540052e-01 -1.69994697e-01 -6.58639848e-01
1.34668279e+00 -6.23280481e-02 1.24156915e-01 -7.19920278e-01
8.42693746e-02 4.32437584e-02 6.56579435e-02 -3.81770760e-01
1.03308284e+00 1.96587518e-01 1.07880846e-01 4.94107842e-01
6.36231422e-01 2.42128313e-01 9.60225761e-02 5.98377407e-01
1.26846802e+00 -2.53692210e-01 3.69916081e-01 -3.14807534e-01
4.68490928e-01 -5.95469892e-01 1.85611963e-01 8.87197673e-01
1.72775328e-01 4.42024082e-01 2.95437098e-01 -8.29510987e-01
-7.37458766e-01 -9.07564163e-01 -4.52059731e-02 1.65512276e+00
-1.96752116e-01 -1.10399592e+00 -1.48453622e-03 -1.23019648e+00
2.10374445e-01 1.12347722e+00 -1.05928731e+00 -3.01978827e-01
-3.83560032e-01 -6.56020343e-01 2.03074202e-01 6.99652493e-01
-1.38320446e-01 -1.40190208e+00 1.83691338e-01 2.90314436e-01
4.82780725e-01 -7.80065835e-01 -5.14217913e-01 6.51169240e-01
-5.67322612e-01 -6.90985441e-01 -5.25837183e-01 -8.44965219e-01
2.01319426e-01 1.89548954e-01 1.60913277e+00 1.13436937e-01
8.15752242e-03 -1.15236141e-01 -5.71479380e-01 -7.07320452e-01
1.10811375e-01 3.88961554e-01 1.13299735e-01 -1.84780598e-01
9.81787682e-01 -2.14927837e-01 -5.57537377e-01 9.20378137e-03
-5.84610641e-01 -1.25324383e-01 4.40694273e-01 1.29579175e+00
4.22741622e-01 -6.66796505e-01 9.84908998e-01 -1.35230589e+00
1.29640210e+00 -9.25788522e-01 -1.72400832e-01 7.79572204e-02
-1.02461636e+00 -8.48768428e-02 4.19152290e-01 -2.61748254e-01
-9.48618889e-01 -3.26889902e-01 -5.83912373e-01 9.03734118e-02
1.00383513e-01 9.57664847e-01 4.17061597e-01 4.89794672e-01
6.82166576e-01 -2.22202688e-02 -2.86720634e-01 -7.06094921e-01
8.19500029e-01 9.24099922e-01 1.00183174e-01 -8.99929553e-02
3.59713808e-02 -2.34166414e-01 -6.34362102e-01 -9.71737862e-01
-1.14471197e+00 -6.99912250e-01 -2.26586312e-01 1.34278715e-01
5.36142886e-01 -8.46721530e-01 -3.36552113e-01 4.88663055e-02
-1.46467352e+00 -1.10250615e-01 -2.36073673e-01 3.82651299e-01
2.29164451e-01 1.16245441e-01 -1.03679049e+00 -7.98229694e-01
-1.09091961e+00 -7.33749509e-01 7.18021154e-01 2.83465981e-01
-4.75766242e-01 -1.10592985e+00 6.82528198e-01 6.98761418e-02
7.51596153e-01 -6.69361353e-01 4.04485077e-01 -1.66301870e+00
4.33381468e-01 -3.45079333e-01 -1.59233809e-01 6.90569401e-01
-1.45793632e-01 -3.49823721e-02 -1.02746129e+00 4.37936969e-02
-4.86286610e-01 -7.01211214e-01 1.67060578e+00 4.78627324e-01
1.80167425e+00 -4.10023272e-01 -2.02952415e-01 1.90618619e-01
8.12974155e-01 -1.16843477e-01 2.87304401e-01 3.13651085e-01
4.96558875e-01 6.90210879e-01 4.35521930e-01 5.46803772e-01
3.63064617e-01 4.39883113e-01 3.55720401e-01 1.51823470e-02
5.40502444e-02 -3.45904231e-01 5.30532837e-01 1.29141176e+00
4.37588006e-01 -5.37700355e-01 -7.83700526e-01 5.95865250e-01
-1.97967589e+00 -6.29468143e-01 -2.05090240e-01 1.49283338e+00
9.55407917e-01 2.71939307e-01 -1.08803183e-01 -4.80414420e-01
5.29719055e-01 5.52631736e-01 -4.48814541e-01 -1.32769763e+00
-3.38377774e-01 6.29720449e-01 2.24275813e-01 7.80553699e-01
-1.19102216e+00 1.03888416e+00 6.80992317e+00 1.05746865e+00
-7.73155272e-01 7.94906542e-02 1.06948626e+00 -3.06300759e-01
-7.22859085e-01 -6.23310387e-01 -1.19595921e+00 4.13433313e-01
1.61368489e+00 -2.01289400e-01 -2.44037971e-01 9.74603355e-01
-1.58887357e-01 2.39742041e-01 -7.06795514e-01 7.40911543e-01
4.34716821e-01 -1.74095166e+00 2.64563084e-01 -7.99220875e-02
9.97989357e-01 7.31016397e-01 2.60940254e-01 1.00849509e+00
7.46944070e-01 -1.36476362e+00 -1.35429382e-01 4.00898099e-01
2.96741724e-01 -7.78408885e-01 1.31851590e+00 1.81494400e-01
-6.45696104e-01 7.69252107e-02 -6.69249654e-01 -2.65188128e-01
1.71839446e-01 8.00051928e-01 -1.14688921e+00 3.68036360e-01
5.89119315e-01 1.09285712e+00 -5.84195733e-01 6.41375184e-01
-5.95814049e-01 1.10140097e+00 9.08118114e-02 -7.58884311e-01
5.47911465e-01 -5.48891276e-02 1.71646997e-01 1.77222788e+00
-1.05759509e-01 -2.15940610e-01 1.42244682e-01 3.52338612e-01
-6.75714731e-01 6.07905984e-01 -7.15096056e-01 6.77597430e-03
3.00911516e-01 1.55330324e+00 -4.60761823e-02 -6.16263270e-01
-7.66497254e-01 8.85269940e-01 7.74026930e-01 3.83373708e-01
-5.30053616e-01 -4.46004093e-01 7.25906074e-01 -4.94611174e-01
4.50970352e-01 9.43222200e-04 -4.52997267e-01 -1.09510612e+00
-3.34930092e-01 -6.19176030e-01 4.27587658e-01 -4.46360171e-01
-2.04059291e+00 1.03813398e+00 -4.92896765e-01 -5.86032689e-01
-3.25889349e-01 -9.06663358e-01 -9.36372817e-01 1.06041837e+00
-1.50551999e+00 -1.05836320e+00 4.40658420e-01 2.84262389e-01
1.05577588e+00 -6.64506733e-01 1.11626256e+00 1.93678103e-02
-6.98998153e-01 7.91793108e-01 2.18392208e-01 -1.16854263e-02
1.02079928e+00 -1.58926570e+00 9.49427605e-01 4.08385605e-01
6.24640644e-01 1.01531148e+00 4.40258086e-01 -6.66260123e-01
-9.75212753e-01 -1.29411697e+00 1.39252758e+00 -3.07662487e-01
1.16956353e+00 -4.10400748e-01 -8.86888266e-01 6.21319115e-01
1.10285354e+00 2.93664169e-02 1.34139132e+00 9.37404931e-01
-4.71757919e-01 -9.34684277e-02 -3.01186830e-01 5.31740427e-01
5.94655633e-01 -6.33373320e-01 -9.97832716e-01 4.21040803e-01
1.25917470e+00 9.12301913e-02 -5.06467044e-01 4.19063181e-01
4.36597079e-01 -4.83083874e-01 7.61044264e-01 -1.41101336e+00
9.78621602e-01 2.48709559e-01 -1.71446040e-01 -1.52205265e+00
-3.89369845e-01 -4.68525112e-01 -7.50777781e-01 1.00164306e+00
1.16595340e+00 -2.95071661e-01 6.23319387e-01 3.96181464e-01
-3.15120697e-01 -1.42142677e+00 -6.81884050e-01 -4.91109371e-01
4.34181452e-01 -6.19414985e-01 3.67241889e-01 8.13684464e-01
4.20387328e-01 1.10261667e+00 -5.14736652e-01 -5.62671304e-01
-1.46942541e-01 7.21647544e-03 6.33731127e-01 -8.34533811e-01
-4.03346062e-01 -6.59387589e-01 1.43159240e-01 -1.44015396e+00
7.36412942e-01 -9.70089018e-01 3.23097818e-02 -1.45670676e+00
2.88911909e-01 3.79023939e-01 -1.20592630e+00 2.37354621e-01
-6.03334248e-01 5.88908270e-02 -2.40833625e-01 -3.33122730e-01
-1.15491867e+00 1.11546183e+00 7.14445651e-01 -4.20530349e-01
-3.97255784e-03 6.35168748e-03 -6.92255437e-01 5.73797524e-01
9.76817846e-01 -1.18149191e-01 -2.14335218e-01 -6.79581761e-01
3.85683000e-01 -5.97904444e-01 -3.27691972e-01 -8.12848359e-02
3.20020229e-01 1.75592825e-01 3.62548679e-01 -9.38832402e-01
3.37056220e-01 -3.06124002e-01 -9.04159129e-01 1.86464652e-01
-1.20390248e+00 2.25050911e-01 2.37949714e-01 7.50715256e-01
-4.05764997e-01 -2.39389718e-01 1.74899381e-02 -4.50939462e-02
-7.99059510e-01 2.88295358e-01 -7.03823209e-01 -3.21180284e-01
5.31470001e-01 6.95468783e-01 -6.45685196e-01 -3.55706275e-01
-6.40647650e-01 5.22409916e-01 -3.54124725e-01 1.01167607e+00
1.03601480e+00 -1.19940674e+00 -8.50042343e-01 8.22417215e-02
3.31228495e-01 -4.81800050e-01 2.40301847e-01 4.70589429e-01
9.94322821e-02 6.28613412e-01 2.23523468e-01 2.20927298e-02
-1.21344626e+00 4.67689544e-01 4.34120037e-02 -6.93088055e-01
-4.41135585e-01 1.71645081e+00 3.64326537e-02 -7.03785479e-01
3.47384751e-01 -1.07262551e-03 -9.21704769e-01 1.26638412e-01
9.99103308e-01 3.03962290e-01 3.64818037e-01 -6.17978163e-02
-3.73169482e-01 -1.70779437e-01 -7.54157901e-01 -1.71474174e-01
1.61041915e+00 -8.95720050e-02 -3.52318883e-01 7.37270236e-01
1.34140265e+00 -5.64772263e-02 -5.07521749e-01 -4.75398690e-01
3.50983202e-01 3.25323753e-02 1.62465036e-01 -9.78654206e-01
-8.44866991e-01 9.56647217e-01 1.49599567e-01 4.08804804e-01
5.16113281e-01 -1.04145244e-01 8.30739915e-01 8.22452068e-01
-2.43207321e-01 -1.30057943e+00 6.72582313e-02 1.20264041e+00
1.13752198e+00 -1.53331220e+00 -1.64873630e-01 3.35112065e-01
-8.93675447e-01 1.28686392e+00 1.13841522e+00 -5.70628881e-01
1.11648083e+00 1.82090089e-01 5.50063439e-02 -3.57800603e-01
-1.51014161e+00 4.92896251e-02 7.67434895e-01 1.27992332e-01
1.17640555e+00 1.31475225e-01 -8.05714667e-01 1.42410433e+00
-6.11328669e-02 -5.41530311e-01 3.28817666e-01 4.83347833e-01
-4.35938239e-01 -1.26972485e+00 4.76798534e-01 8.35911095e-01
-5.81112027e-01 -5.90707541e-01 -8.66187572e-01 1.99023291e-01
-1.99110329e-01 1.22993708e+00 1.25299945e-01 -8.90303075e-01
9.74697694e-02 1.42029315e-01 -3.42867851e-01 -1.14869106e+00
-1.11764872e+00 -1.68417171e-02 6.39435768e-01 -6.67929947e-01
-7.34492913e-02 -1.67985454e-01 -9.44724202e-01 -7.65892342e-02
-5.16344547e-01 6.54890060e-01 7.35603392e-01 8.84103537e-01
6.76444232e-01 8.05255294e-01 7.69860208e-01 -4.34973270e-01
-8.01400840e-01 -1.78095937e+00 -6.50550902e-01 2.59246200e-01
2.46580452e-01 -1.47788122e-01 -1.50323808e-01 -4.52793121e-01]
|
[10.917896270751953, 7.5656609535217285]
|
48582cea-417e-4889-8a4e-75670826e3a3
|
tedigan-text-guided-diverse-image-generation
|
2012.03308
| null |
https://arxiv.org/abs/2012.03308v3
|
https://arxiv.org/pdf/2012.03308v3.pdf
|
TediGAN: Text-Guided Diverse Face Image Generation and Manipulation
|
In this work, we propose TediGAN, a novel framework for multi-modal image generation and manipulation with textual descriptions. The proposed method consists of three components: StyleGAN inversion module, visual-linguistic similarity learning, and instance-level optimization. The inversion module maps real images to the latent space of a well-trained StyleGAN. The visual-linguistic similarity learns the text-image matching by mapping the image and text into a common embedding space. The instance-level optimization is for identity preservation in manipulation. Our model can produce diverse and high-quality images with an unprecedented resolution at 1024. Using a control mechanism based on style-mixing, our TediGAN inherently supports image synthesis with multi-modal inputs, such as sketches or semantic labels, with or without instance guidance. To facilitate text-guided multi-modal synthesis, we propose the Multi-Modal CelebA-HQ, a large-scale dataset consisting of real face images and corresponding semantic segmentation map, sketch, and textual descriptions. Extensive experiments on the introduced dataset demonstrate the superior performance of our proposed method. Code and data are available at https://github.com/weihaox/TediGAN.
|
['Baoyuan Wu', 'Jing-Hao Xue', 'Yujiu Yang', 'Weihao Xia']
|
2020-12-06
| null |
http://openaccess.thecvf.com//content/CVPR2021/html/Xia_TediGAN_Text-Guided_Diverse_Face_Image_Generation_and_Manipulation_CVPR_2021_paper.html
|
http://openaccess.thecvf.com//content/CVPR2021/papers/Xia_TediGAN_Text-Guided_Diverse_Face_Image_Generation_and_Manipulation_CVPR_2021_paper.pdf
|
cvpr-2021-1
|
['face-sketch-synthesis']
|
['computer-vision']
|
[ 4.78933901e-01 -2.43331753e-02 -1.43230423e-01 -3.89333308e-01
-8.51930678e-01 -7.79254198e-01 7.98294365e-01 -7.70079553e-01
8.88727978e-02 4.68684137e-01 -1.34498952e-02 2.05930248e-01
-5.54583361e-03 -8.74869287e-01 -9.16347802e-01 -6.94378614e-01
6.86469257e-01 5.66473901e-01 -2.52534002e-01 -1.08205482e-01
3.57787818e-01 5.23621917e-01 -1.52867556e+00 3.74272108e-01
7.35956669e-01 1.03916347e+00 2.68013477e-01 6.57145858e-01
-3.68604869e-01 3.80742937e-01 -4.35031593e-01 -7.41808057e-01
4.30329561e-01 -6.47769034e-01 -4.93845731e-01 5.35522819e-01
8.33274543e-01 -4.11669493e-01 -3.12887907e-01 1.14013529e+00
5.75799942e-01 -1.32508934e-01 7.34204769e-01 -1.77355957e+00
-1.21214223e+00 3.80342245e-01 -6.63163245e-01 -6.94984019e-01
3.82976711e-01 4.90914613e-01 8.33058953e-01 -1.01360917e+00
8.83742094e-01 1.62285995e+00 2.85913229e-01 8.74073982e-01
-1.26912439e+00 -1.09054482e+00 -2.24558577e-01 -8.56718570e-02
-1.45619464e+00 -6.92980289e-01 1.09992135e+00 -3.61686617e-01
6.34357408e-02 2.62886852e-01 5.92653036e-01 1.25104344e+00
-1.11977406e-01 9.88045812e-01 1.15081894e+00 -2.94533670e-01
-1.68811411e-01 4.39956874e-01 -6.35905206e-01 1.10908830e+00
-1.36017084e-01 6.84670582e-02 -7.08543479e-01 -1.24376342e-01
1.20764637e+00 3.69888134e-02 -8.19408968e-02 -3.09417874e-01
-1.34926105e+00 8.43725264e-01 2.21300706e-01 9.25981253e-02
-1.47969514e-01 3.04964542e-01 1.97714463e-01 2.65219539e-01
2.94144958e-01 2.58309811e-01 9.80404299e-03 2.66460568e-01
-1.10630763e+00 2.98104823e-01 6.16254508e-01 1.59074473e+00
8.32036138e-01 2.92469233e-01 -3.60923886e-01 9.72563922e-01
4.93379921e-01 1.03963804e+00 3.71644914e-01 -1.18519759e+00
4.84653950e-01 4.78473604e-01 -1.45272374e-01 -1.18548846e+00
2.94865370e-01 8.22031721e-02 -9.75499630e-01 1.28944129e-01
2.31309786e-01 -2.57864445e-02 -8.89498830e-01 1.74701989e+00
3.17611724e-01 2.80523062e-01 -3.16844769e-02 9.29321706e-01
1.00299001e+00 7.86638200e-01 -9.48754847e-02 3.73470224e-02
1.42377412e+00 -1.13578594e+00 -9.09201384e-01 -6.84426501e-02
-1.24742135e-01 -1.12160635e+00 1.44027567e+00 1.34500086e-01
-1.31714976e+00 -5.67664087e-01 -1.03567457e+00 -3.78277391e-01
-3.36648047e-01 6.07189953e-01 4.55774933e-01 5.63559711e-01
-1.01767182e+00 3.58788460e-01 -2.67114103e-01 -2.97216445e-01
6.43738329e-01 1.89870507e-01 -5.31541407e-01 -1.61293000e-01
-9.45895135e-01 3.45289290e-01 3.67076665e-01 5.46243694e-03
-1.03532732e+00 -6.23422563e-01 -9.98456001e-01 -2.00279742e-01
2.98724174e-01 -7.82718062e-01 8.54150832e-01 -1.18272519e+00
-1.88045609e+00 1.24742711e+00 -1.26189902e-01 2.50637770e-01
7.31619596e-01 1.14124641e-01 -3.03683192e-01 3.57147396e-01
3.29522222e-01 1.08294988e+00 1.44239295e+00 -1.57206798e+00
-4.28970963e-01 -3.83964181e-01 -4.29595895e-02 7.91097954e-02
-3.19052666e-01 -9.49096754e-02 -9.26330745e-01 -1.02831173e+00
-4.47081774e-02 -9.62356031e-01 1.89134419e-01 3.97193611e-01
-7.05602288e-01 2.40999851e-02 9.68747437e-01 -5.80279112e-01
7.44710505e-01 -2.27962518e+00 4.48878288e-01 2.08764017e-01
1.57073900e-01 -9.02114029e-04 -5.07359743e-01 3.53906691e-01
1.03847668e-01 2.30524242e-01 -3.93960059e-01 -5.93966722e-01
2.60148108e-01 6.23340718e-02 -4.57369059e-01 4.82942581e-01
3.80124599e-01 1.14887166e+00 -6.92384720e-01 -8.24616730e-01
3.07705730e-01 4.72601831e-01 -4.59571421e-01 5.15658140e-01
-5.06613851e-01 6.85092151e-01 -4.41266626e-01 9.33886051e-01
8.06233943e-01 -1.84951752e-01 -1.53012518e-02 -5.34741223e-01
3.76799032e-02 -5.00608742e-01 -1.30469263e+00 2.08726192e+00
-4.42613155e-01 5.16021609e-01 1.70138583e-01 -4.90723938e-01
9.75038528e-01 1.98177770e-01 2.12197483e-01 -6.30568445e-01
2.82379597e-01 2.51667976e-01 -6.37613118e-01 -5.87318957e-01
5.22189260e-01 -1.57348931e-01 -1.86880648e-01 7.70620108e-01
2.28014290e-01 -7.27729440e-01 2.59008497e-01 1.88363522e-01
2.68440545e-01 3.53636414e-01 -1.72835976e-01 -1.39325604e-01
6.83812499e-01 -1.99491590e-01 3.27318341e-01 3.87704253e-01
1.65637195e-01 7.74326444e-01 4.46273029e-01 -1.39162540e-01
-1.37870979e+00 -1.11067891e+00 5.93988821e-02 1.01703262e+00
4.13401037e-01 -2.08833367e-01 -9.89453554e-01 -4.51809943e-01
3.89066413e-02 4.67439026e-01 -6.17860317e-01 -5.27166650e-02
-4.08264130e-01 -1.35274768e-01 8.32073987e-01 4.44481475e-03
8.66811275e-01 -1.23569429e+00 -2.63662562e-02 -2.13588268e-01
-2.00831205e-01 -1.14081931e+00 -1.10537529e+00 -6.60055697e-01
-5.58720291e-01 -9.28128183e-01 -9.47005391e-01 -1.10654747e+00
8.62937152e-01 3.62072363e-02 7.76638567e-01 -1.62445065e-02
-3.46786916e-01 5.30582249e-01 -5.52017242e-03 -7.98843876e-02
-4.06357139e-01 -8.67904425e-02 -1.50757864e-01 5.03675818e-01
-1.99042127e-01 -6.03740990e-01 -5.77882767e-01 4.13293749e-01
-1.28358078e+00 4.97671664e-01 6.52793586e-01 1.04968584e+00
6.80058300e-01 -3.03651273e-01 3.18942696e-01 -9.14983034e-01
5.04558861e-01 -2.51937449e-01 -5.44394791e-01 5.57177961e-01
-2.37978414e-01 2.03138381e-01 5.04682839e-01 -6.36321902e-01
-1.08881819e+00 1.69172928e-01 6.61295503e-02 -7.49558926e-01
-1.56429470e-01 -8.80312175e-02 -6.56525552e-01 -3.04006219e-01
2.61059642e-01 6.23625994e-01 2.16390729e-01 -2.45807931e-01
9.56359446e-01 7.75095820e-01 7.20718384e-01 -9.31437016e-01
1.07391870e+00 5.59552252e-01 -1.74128771e-01 -8.78544509e-01
-3.83637428e-01 7.65405521e-02 -6.13310099e-01 -3.18150789e-01
9.46212769e-01 -8.96681070e-01 -7.34849334e-01 7.28738725e-01
-1.31102872e+00 -2.83344835e-01 -1.98282033e-01 3.35642695e-02
-9.12353158e-01 1.58628374e-01 -5.66160917e-01 -4.93452132e-01
-4.36419964e-01 -1.36630380e+00 1.57899880e+00 4.45178926e-01
1.74428165e-01 -8.27190816e-01 -3.59094143e-01 7.12967515e-01
3.56147915e-01 3.87917429e-01 7.67485321e-01 -1.55879915e-01
-1.00196421e+00 2.43341953e-01 -5.25605917e-01 1.60180733e-01
2.10108161e-01 2.45935962e-01 -9.57945943e-01 -2.47060433e-01
-2.92764753e-01 -5.85438132e-01 5.22487998e-01 2.51695001e-03
1.24415445e+00 -4.50206280e-01 -1.08407281e-01 1.07767785e+00
1.46529984e+00 7.45225921e-02 5.45161664e-01 2.16519814e-02
1.03876054e+00 6.28960669e-01 6.26791835e-01 4.67811316e-01
4.53446478e-01 6.39499962e-01 3.27023178e-01 -2.19398350e-01
-3.71414274e-01 -6.25146449e-01 2.45739922e-01 7.07550824e-01
2.31401607e-01 -2.02512667e-01 -4.66528535e-01 4.59607035e-01
-1.59675479e+00 -9.30404603e-01 1.99698865e-01 1.75783777e+00
1.04846632e+00 -3.76904666e-01 -5.27621321e-02 -1.86677888e-01
1.01596582e+00 2.66779304e-01 -6.83720946e-01 -1.08626343e-01
-2.35407233e-01 6.76580742e-02 3.35886300e-01 5.05611598e-01
-9.66822386e-01 1.30828714e+00 5.21815348e+00 1.30948913e+00
-1.30548477e+00 1.90949157e-01 6.03167117e-01 -1.31684467e-01
-7.16234267e-01 -3.17996114e-01 -6.67879760e-01 5.45512140e-01
3.67317528e-01 -1.90954402e-01 7.68121958e-01 6.22263849e-01
9.48880799e-03 3.44360501e-01 -9.52526987e-01 1.41360998e+00
4.03983146e-01 -1.62307060e+00 5.96158564e-01 -1.06274515e-01
9.20671880e-01 -6.37081146e-01 4.91308630e-01 -4.25897390e-02
1.62601531e-01 -9.74382579e-01 1.18254912e+00 6.23408258e-01
1.60439718e+00 -7.78098404e-01 5.43868355e-02 -2.45021861e-02
-1.24709463e+00 9.67104137e-02 -1.24300666e-01 6.09534383e-01
1.09448582e-01 1.59059584e-01 -4.48090434e-01 5.05401313e-01
3.54113311e-01 8.61457646e-01 -4.54618305e-01 3.55835229e-01
-3.39411318e-01 1.05332702e-01 -4.87433001e-02 1.40360892e-01
1.66183695e-01 -5.96552372e-01 4.71757978e-01 1.00790656e+00
3.83378744e-01 5.00059053e-02 1.16588563e-01 1.34049010e+00
-3.58590156e-01 1.80727154e-01 -8.52025807e-01 -5.25034189e-01
6.22703493e-01 1.43351555e+00 -5.64265430e-01 -3.69873405e-01
-1.30058397e-02 1.35113728e+00 8.10438115e-03 3.63545656e-01
-1.09074998e+00 -5.40110469e-01 5.43941200e-01 -8.10271651e-02
1.32120743e-01 -8.18633065e-02 -3.85524005e-01 -1.20727360e+00
3.72388177e-02 -1.07649839e+00 1.25697274e-02 -9.91545379e-01
-1.20074129e+00 6.27232909e-01 -1.19093612e-01 -1.07385457e+00
3.59369926e-02 -5.11810064e-01 -4.84189749e-01 8.54892254e-01
-1.31786036e+00 -1.96238160e+00 -5.07799208e-01 9.87718344e-01
6.83207870e-01 -6.06793880e-01 6.64625108e-01 4.86644924e-01
-4.92773294e-01 9.27522540e-01 -2.30174270e-02 2.43257761e-01
9.36436892e-01 -1.02713323e+00 3.14848483e-01 5.42844892e-01
7.96358436e-02 5.27258277e-01 3.76428187e-01 -6.32850707e-01
-1.71195948e+00 -1.26387441e+00 3.93004984e-01 -1.65854856e-01
5.18433273e-01 -5.47115207e-01 -4.58306760e-01 7.24081576e-01
3.92557859e-01 -2.53159910e-01 5.06944358e-01 -7.80494511e-01
-3.27933341e-01 -2.16468334e-01 -1.41276491e+00 8.10414076e-01
1.20702326e+00 -8.86526465e-01 -2.57724673e-01 5.30453622e-01
8.26164603e-01 -5.19092917e-01 -8.83940697e-01 1.11617558e-01
8.09132338e-01 -7.26330519e-01 1.01516187e+00 -2.57479101e-01
8.00466299e-01 -3.79578739e-01 -3.42454404e-01 -1.03370869e+00
-4.43840120e-03 -7.60222316e-01 3.81287515e-01 1.69119751e+00
1.88879952e-01 -4.22949314e-01 6.97682440e-01 4.10170734e-01
1.98545769e-01 -5.01029193e-01 -5.64147711e-01 -4.29515511e-01
-1.31904662e-01 -1.70938864e-01 1.04560709e+00 1.07995403e+00
-6.16441429e-01 1.98832810e-01 -7.74518132e-01 8.96412060e-02
9.96918857e-01 4.45966065e-01 1.00325406e+00 -7.18431652e-01
-1.56861767e-01 -4.38925922e-01 -2.18422607e-01 -1.05503726e+00
5.36370754e-01 -1.05143964e+00 -1.40839517e-01 -1.09410894e+00
2.22488150e-01 -4.26026732e-01 3.29764575e-01 4.07231539e-01
1.57651603e-01 7.13914096e-01 4.71148938e-01 1.95352569e-01
-4.34916049e-01 7.76793301e-01 1.85528541e+00 -5.48189938e-01
7.19887167e-02 -3.68164450e-01 -6.96706295e-01 4.17690754e-01
7.98812091e-01 -3.76102656e-01 -4.59853977e-01 -5.71826339e-01
-1.56057566e-01 3.41703236e-01 5.07173061e-01 -5.79866350e-01
2.43413359e-01 -3.90503019e-01 4.20713872e-01 -4.06012684e-01
5.83188832e-01 -8.19074810e-01 5.74307680e-01 3.80961299e-01
-6.21137619e-01 -1.01330571e-01 7.59944543e-02 4.00054306e-01
-1.75634503e-01 1.39202652e-02 1.00829720e+00 -2.50530452e-01
-6.18442178e-01 6.62258923e-01 2.15772748e-01 -3.38596404e-02
1.11457849e+00 -1.04850389e-01 -3.65129620e-01 -3.92259926e-01
-6.45228982e-01 1.78147614e-01 7.14637697e-01 6.42656803e-01
9.57147300e-01 -1.90379095e+00 -7.96231449e-01 6.65403545e-01
2.32045516e-01 -6.64086267e-02 2.85095394e-01 4.43486035e-01
-6.69229686e-01 2.02157900e-01 -6.15756571e-01 -6.69417143e-01
-1.23223209e+00 2.65684843e-01 2.43106455e-01 2.41281852e-01
-2.31004298e-01 8.14192951e-01 3.54915440e-01 -7.76965261e-01
1.34970117e-02 1.67987928e-01 1.29814208e-01 1.96452755e-02
4.70183164e-01 2.28069097e-01 -5.67707896e-01 -8.55984449e-01
-1.82650574e-02 9.98986959e-01 1.75340965e-01 -4.57314581e-01
1.15906525e+00 -2.00468719e-01 -5.46368599e-01 3.09332341e-01
1.33502305e+00 -3.87451276e-02 -1.22328293e+00 -2.45984375e-01
-4.51414198e-01 -8.16823542e-01 -2.82019705e-01 -5.35819709e-01
-1.45292187e+00 8.33930671e-01 5.58088303e-01 -2.69293904e-01
1.13840961e+00 1.35663180e-02 9.82745409e-01 8.75808522e-02
3.02971095e-01 -8.28482270e-01 5.72951972e-01 9.57818627e-02
1.29110765e+00 -1.26389682e+00 -3.05530816e-01 -2.81373471e-01
-8.04222226e-01 1.06161082e+00 6.35999918e-01 -1.71638057e-01
4.21247602e-01 1.08990327e-01 1.59996569e-01 -1.80727795e-01
-4.35509145e-01 5.74352928e-02 3.07153463e-01 6.03169799e-01
2.49189530e-02 1.05618052e-01 7.42098689e-03 4.33980256e-01
-3.15442622e-01 -2.11509764e-02 2.90538847e-01 5.57985008e-01
-1.72573194e-01 -1.14641964e+00 -5.32646120e-01 1.38309434e-01
-2.16543078e-01 -8.20751116e-02 -3.72645497e-01 6.32996321e-01
2.83463418e-01 7.76389778e-01 5.25388084e-02 -4.89664912e-01
2.93966800e-01 -3.48597914e-02 6.32275045e-01 -3.97504151e-01
-4.11820233e-01 1.28654718e-01 -2.31368154e-01 -6.86513662e-01
-3.38141918e-01 -5.09424210e-01 -9.87966478e-01 -6.75591111e-01
-3.93495820e-02 -5.52587546e-02 7.93658853e-01 6.87775791e-01
3.74575764e-01 2.73537278e-01 8.51414740e-01 -1.09918702e+00
-2.50782073e-01 -7.08725870e-01 -7.40599632e-01 6.12297773e-01
2.10388795e-01 -4.88435745e-01 -2.29407281e-01 5.11257708e-01]
|
[12.281638145446777, -0.22180306911468506]
|
78b92365-5bdc-4dfb-b459-80a14b6a1446
|
multi-dataset-co-training-with-sharpness
|
2305.19953
| null |
https://arxiv.org/abs/2305.19953v2
|
https://arxiv.org/pdf/2305.19953v2.pdf
|
Multi-Dataset Co-Training with Sharpness-Aware Optimization for Audio Anti-spoofing
|
Audio anti-spoofing for automatic speaker verification aims to safeguard users' identities from spoofing attacks. Although state-of-the-art spoofing countermeasure(CM) models perform well on specific datasets, they lack generalization when evaluated with different datasets. To address this limitation, previous studies have explored large pre-trained models, which require significant resources and time. We aim to develop a compact but well-generalizing CM model that can compete with large pre-trained models. Our approach involves multi-dataset co-training and sharpness-aware minimization, which has not been investigated in this domain. Extensive experiments reveal that proposed method yield competitive results across various datasets while utilizing 4,000 times less parameters than the large pre-trained models.
|
['Tomi Kinnunen', 'Jee-weon Jung', 'Hye-jin Shim']
|
2023-05-31
| null | null | null | null |
['speaker-verification']
|
['speech']
|
[ 5.23070693e-01 -2.91325569e-01 -1.82081386e-01 -2.99692333e-01
-1.15462601e+00 -5.89923680e-01 5.19360602e-01 -1.65717036e-01
-3.06590945e-01 4.15533632e-01 2.63781041e-01 -5.78544199e-01
1.73562676e-01 -1.73711017e-01 -5.93966126e-01 -5.22386014e-01
-2.41064727e-01 1.51245236e-01 1.58721015e-01 -2.80826956e-01
2.58319676e-01 3.65088820e-01 -1.23274875e+00 4.29467082e-01
7.78419912e-01 6.64216280e-01 -3.89534354e-01 8.68118763e-01
2.02048942e-01 2.66009420e-01 -8.72201622e-01 -7.32578039e-01
1.58159718e-01 -3.13522905e-01 -8.56856585e-01 -1.65002376e-01
7.21690536e-01 -6.97870776e-02 -2.38633528e-01 1.35590875e+00
8.52215171e-01 -2.48992160e-01 2.26658478e-01 -1.36391056e+00
-4.60654467e-01 7.71969557e-01 -6.96008861e-01 4.43664193e-01
4.99571234e-01 -1.57915682e-01 7.93105185e-01 -6.00673616e-01
1.47892937e-01 1.53407776e+00 9.63140845e-01 8.43585789e-01
-1.09201837e+00 -1.20979500e+00 3.33359174e-04 2.50272691e-01
-1.61779940e+00 -9.78554428e-01 1.13436556e+00 1.30293056e-01
7.96742678e-01 4.92206097e-01 5.26431436e-03 1.39933670e+00
-2.61265457e-01 7.89908171e-01 1.16381860e+00 -4.47070301e-01
-2.36248165e-01 6.24966502e-01 1.07871801e-01 5.01483917e-01
4.86253738e-01 1.54665917e-01 -5.12850642e-01 -6.64196610e-01
1.69511780e-01 -3.77488911e-01 -3.89000595e-01 -2.08875071e-02
-9.18730617e-01 1.01236987e+00 -6.62676990e-02 5.92317104e-01
-1.60081089e-01 -1.60855860e-01 7.16495633e-01 4.67206806e-01
4.41312432e-01 3.29689652e-01 -5.51699460e-01 -5.40808775e-02
-1.25871837e+00 9.34219658e-02 7.75626600e-01 6.24894679e-01
4.33502108e-01 3.36412191e-01 4.25675035e-01 8.36291909e-01
4.30096447e-01 6.73083901e-01 4.66073424e-01 -4.55902189e-01
8.03375959e-01 4.57687825e-02 -1.41233683e-01 -1.30915058e+00
-2.68992811e-01 -6.05570436e-01 -8.18427026e-01 -4.61623102e-01
1.59876287e-01 -2.17885002e-01 -6.10980153e-01 1.85831928e+00
3.63011867e-01 8.14883947e-01 8.36068168e-02 3.65581304e-01
4.85632420e-01 4.01008666e-01 -6.44194633e-02 -3.97805929e-01
1.30386138e+00 -7.99127400e-01 -7.55259931e-01 -1.73139453e-01
7.30143011e-01 -1.12552691e+00 7.58910477e-01 5.55208206e-01
-7.70208240e-01 -4.10020560e-01 -1.09901941e+00 3.95776451e-01
-3.98731291e-01 -1.80124938e-01 7.28596568e-01 1.92240977e+00
-1.03772521e+00 1.18215188e-01 -3.64422202e-01 -2.19786584e-01
5.81125915e-01 5.72294950e-01 -4.66356128e-01 1.93650499e-01
-1.51360154e+00 7.00579464e-01 1.56970367e-01 -5.57263643e-02
-1.09736550e+00 -6.45665765e-01 -7.44429171e-01 6.69432012e-03
1.27702624e-01 -2.26859033e-01 1.03623641e+00 -6.94783568e-01
-1.59130812e+00 9.12520289e-01 -2.91357696e-01 -8.43229651e-01
2.53742158e-01 -2.88236797e-01 -1.32779133e+00 2.72532076e-01
-9.16754007e-02 3.80448222e-01 1.45128059e+00 -1.38749862e+00
-3.84941936e-01 -3.65104347e-01 -2.61577696e-01 -3.62504929e-01
-8.57440174e-01 6.97672188e-01 -8.46450999e-02 -8.26327443e-01
8.69452860e-03 -1.06224859e+00 -2.42872033e-02 -7.21178293e-01
-5.09626806e-01 2.02083305e-01 1.52235663e+00 -9.50239539e-01
1.52377236e+00 -2.13973284e+00 -4.25949186e-01 3.96011651e-01
-1.44930273e-01 1.09306991e+00 -2.09617570e-01 4.21760321e-01
-2.01771811e-01 4.07193810e-01 -2.24067643e-01 -7.29411244e-01
-1.01267971e-01 -2.93159932e-01 -5.04072249e-01 9.24906850e-01
-4.05309498e-02 3.06620091e-01 -6.39484107e-01 -7.33158827e-01
1.35231838e-01 8.72780204e-01 -8.66700590e-01 -1.04844123e-01
2.27476403e-01 3.49076211e-01 -2.85193205e-01 7.90552080e-01
1.39670920e+00 -4.95479966e-04 2.38916487e-01 -2.46672314e-02
3.40920329e-01 6.95275307e-01 -1.42594504e+00 1.42180467e+00
-3.73150289e-01 5.94913244e-01 5.60298085e-01 -1.20356488e+00
7.78722346e-01 6.79543376e-01 3.76811355e-01 -2.40041271e-01
3.27799439e-01 4.34042215e-01 -5.88990748e-02 -4.84489799e-01
3.37695479e-01 -1.96274191e-01 5.56900585e-03 5.61920881e-01
2.74660978e-02 1.33714318e-01 -3.32648993e-01 1.38831884e-01
7.56062627e-01 -5.43817759e-01 3.38764563e-02 -1.59614027e-01
1.18265986e+00 -6.10928297e-01 4.27741110e-01 7.89177656e-01
-7.79780149e-01 3.90657842e-01 -1.21750250e-01 1.16754072e-02
-5.96987844e-01 -5.72272480e-01 -3.11546832e-01 1.05259645e+00
7.97448829e-02 -4.71629888e-01 -9.82152224e-01 -9.57798719e-01
-6.37340546e-02 4.55734193e-01 -3.35150093e-01 -1.01562701e-02
-9.58586276e-01 -9.37025726e-01 1.46463072e+00 -1.66387169e-03
6.52390659e-01 -7.11162746e-01 -8.95139128e-02 1.75003365e-01
-4.92820948e-01 -1.39170015e+00 -6.77972198e-01 -3.47173184e-01
-6.83329463e-01 -9.69011843e-01 -5.57619989e-01 -9.76433754e-01
3.18590790e-01 7.43878663e-01 6.50434911e-01 1.65876314e-01
-7.05386028e-02 1.72665402e-01 -2.40791172e-01 -5.61167240e-01
-7.49532700e-01 2.02574581e-01 5.80974400e-01 5.75699031e-01
5.84947169e-01 -6.57589495e-01 -2.12294579e-01 5.15341818e-01
-8.50190401e-01 -6.24175668e-01 3.46780896e-01 7.64156044e-01
-2.19151989e-01 3.64853829e-01 8.60382080e-01 -7.50931442e-01
6.63962305e-01 -3.45643908e-01 -2.79022425e-01 2.59513319e-01
-6.35114074e-01 8.72334093e-03 3.40253025e-01 -8.40223730e-01
-9.91251111e-01 -2.38950029e-01 -3.82937938e-01 -3.18576276e-01
-1.11933902e-01 1.70052290e-01 -2.86944926e-01 -7.68875062e-01
4.90730494e-01 4.07335371e-01 1.87286343e-02 -7.31608987e-01
1.41162336e-01 1.15338230e+00 4.01010275e-01 -3.79614294e-01
1.17139304e+00 6.17292583e-01 -2.95756221e-01 -1.20796704e+00
-3.93273026e-01 -9.92201388e-01 -3.04685682e-01 2.92587906e-01
4.78717446e-01 -8.61476600e-01 -8.89836490e-01 7.26778567e-01
-1.24637961e+00 5.71837127e-01 6.11356854e-01 5.67338288e-01
-8.22691098e-02 1.02300751e+00 -6.21871173e-01 -1.15517211e+00
-5.92015922e-01 -9.58606958e-01 9.74147499e-01 -1.71850801e-01
-1.20224684e-01 -1.00944042e+00 3.59403998e-01 8.00148427e-01
6.98981881e-01 -1.01746969e-01 3.13203961e-01 -1.02795660e+00
-7.41181374e-02 -5.94312429e-01 -1.87815875e-01 6.20191276e-01
2.63149977e-01 -1.80439457e-01 -1.43800378e+00 -6.95991874e-01
4.20254707e-01 -2.57170856e-01 1.07114244e+00 2.85649121e-01
1.08283794e+00 -7.08595932e-01 -5.02635598e-01 6.86306298e-01
1.00311065e+00 -2.92232055e-02 6.54125988e-01 3.23671639e-01
6.97927415e-01 6.75327837e-01 2.70618528e-01 3.48292023e-01
3.28795373e-01 8.17560256e-01 3.44935417e-01 4.18186486e-02
1.31465241e-01 -4.22370166e-01 6.26574218e-01 9.04295981e-01
2.23521918e-01 -2.30679542e-01 -6.81011856e-01 6.18906677e-01
-1.23150945e+00 -1.39345658e+00 3.00032701e-02 2.17206860e+00
7.32025206e-01 2.21300140e-01 4.79042977e-01 8.14594924e-01
1.01190186e+00 5.91582179e-01 2.47297902e-03 -4.56461251e-01
-4.01355356e-01 1.88253239e-01 5.52513957e-01 8.66164863e-01
-1.53880513e+00 1.05659819e+00 6.83650446e+00 1.14138138e+00
-1.45635116e+00 3.96444350e-01 2.92225152e-01 4.23593856e-02
-2.19351664e-01 -2.36006342e-02 -1.05409670e+00 4.01180387e-01
1.22988796e+00 1.40104577e-01 3.45878184e-01 7.25009918e-01
-1.15000233e-02 7.19334662e-01 -5.40956438e-01 1.21326184e+00
4.93303716e-01 -1.34289753e+00 -4.26466726e-02 2.87636578e-01
6.11855865e-01 -2.56161690e-02 2.88509876e-01 2.14371219e-01
1.23779297e-01 -9.41622138e-01 4.29462075e-01 -3.58997375e-01
5.80076575e-01 -7.36037791e-01 6.49619639e-01 2.43997484e-01
-1.27433670e+00 -3.24069768e-01 -2.31032804e-01 2.61405557e-01
3.76483560e-01 3.58941406e-01 -1.10451484e+00 5.43922544e-01
5.28195500e-01 2.59070218e-01 -5.41691005e-01 7.83339202e-01
9.48812068e-02 1.08697522e+00 -5.17342448e-01 -5.31788357e-03
1.96053267e-01 4.42245096e-01 7.06850708e-01 1.68268585e+00
1.41086683e-01 -3.14580202e-01 -1.27977831e-02 2.80280679e-01
-1.12209357e-01 1.57569006e-01 -5.42949855e-01 -2.24787235e-01
7.01251745e-01 8.76648068e-01 -9.39305872e-02 -3.15686613e-01
-1.99432403e-01 9.05595958e-01 -6.35673404e-02 2.14453518e-01
-9.88640189e-01 -3.33844572e-01 8.22586298e-01 1.74812093e-01
3.47149253e-01 -2.23052934e-01 -2.97859937e-01 -1.31229854e+00
-1.24277025e-01 -1.46943974e+00 6.24531507e-01 1.20292947e-01
-1.08815825e+00 7.42224932e-01 -2.95444518e-01 -1.25470054e+00
-1.43891200e-01 -2.57257074e-01 -5.40959597e-01 5.80554247e-01
-1.73645413e+00 -1.36752915e+00 2.69218832e-01 9.11243856e-01
2.84409940e-01 -4.65990961e-01 1.02438462e+00 8.05293381e-01
-8.18888009e-01 1.27674496e+00 -5.70479371e-02 1.80572867e-01
7.02757239e-01 -6.05764627e-01 5.89161038e-01 1.02356279e+00
3.86879325e-01 1.10580337e+00 9.18888211e-01 -4.64476943e-01
-1.33537948e+00 -6.83322430e-01 1.28351426e+00 -4.13400173e-01
6.13858700e-01 -3.96954089e-01 -9.90414500e-01 4.05804366e-01
2.30184689e-01 8.19782838e-02 9.83495116e-01 2.28263184e-01
-1.03644526e+00 -1.40094263e-02 -1.56661224e+00 1.40063301e-01
9.11372602e-01 -1.14915717e+00 -3.06026995e-01 1.21102624e-01
6.23733699e-01 -1.52774593e-02 -5.43918848e-01 4.16936219e-01
6.28513873e-01 -9.95866835e-01 1.40959966e+00 -7.30928183e-01
-3.70322973e-01 -1.31892487e-01 -2.22972274e-01 -8.16422105e-01
4.61164974e-02 -1.37462139e+00 -2.90540248e-01 1.62219191e+00
4.17060465e-01 -8.49557877e-01 8.84554982e-01 3.24734114e-02
1.84420452e-01 -2.54260331e-01 -1.06280088e+00 -1.10267687e+00
-2.16620471e-02 -6.91430688e-01 9.70387399e-01 1.38906229e+00
1.61020294e-01 9.06349048e-02 -1.08858383e+00 8.06648135e-01
9.00707603e-01 -2.80246258e-01 8.96862388e-01 -9.61641192e-01
-4.18321431e-01 -4.79458332e-01 -4.42986459e-01 -1.08307135e+00
3.79044056e-01 -6.12141192e-01 -3.61956924e-01 -6.06906354e-01
1.21742837e-01 -6.31513953e-01 -4.15317357e-01 1.58325195e-01
-3.57085645e-01 4.25991923e-01 4.82887775e-02 9.01126638e-02
-3.95488948e-01 3.69390517e-01 6.65827096e-01 -4.08291817e-01
-1.98646232e-01 4.43742067e-01 -8.02018762e-01 5.68135381e-01
9.90432262e-01 -8.32748294e-01 -3.50546330e-01 -1.77764073e-01
-3.59300733e-01 -6.71609566e-02 1.93555579e-01 -1.05130005e+00
8.71105716e-02 -1.59882922e-02 -1.16644658e-01 -4.37507480e-01
5.71550846e-01 -7.39840865e-01 -6.95641786e-02 6.45605385e-01
-3.05339098e-01 2.96247704e-03 2.18228251e-01 5.05941093e-01
-3.23058307e-01 -1.68134302e-01 1.03087604e+00 1.71821773e-01
-5.69284372e-02 2.15476915e-01 -9.61208567e-02 1.30857620e-02
6.25607073e-01 -2.58728653e-01 -3.83516312e-01 -5.78565836e-01
-1.79209381e-01 -3.51981491e-01 1.66095003e-01 5.62163711e-01
6.66900575e-01 -1.10404742e+00 -8.97319555e-01 5.14038682e-01
-1.39869362e-01 -7.34315991e-01 3.30591232e-01 6.94660664e-01
-3.55514079e-01 8.94758642e-01 -2.10926076e-03 -4.15118515e-01
-1.83984125e+00 9.33348060e-01 1.51241615e-01 -4.80957657e-01
-2.56308675e-01 1.17446327e+00 -2.28395671e-01 -5.96164525e-01
4.36275065e-01 1.97322518e-01 -1.34520352e-01 -1.10161908e-01
9.84044790e-01 4.69135076e-01 1.36000827e-01 -1.18567479e+00
-8.07863712e-01 4.92934138e-01 -3.20360482e-01 -3.21443915e-01
1.00900996e+00 -2.13620335e-01 1.76215515e-01 -2.67600030e-01
1.75927734e+00 6.05634630e-01 -6.29313529e-01 -4.11588073e-01
2.48034224e-01 -8.17603111e-01 2.14459077e-02 -2.46032029e-01
-9.33103919e-01 1.03634310e+00 7.23733425e-01 4.37243760e-01
1.04327714e+00 -3.81834358e-01 1.21040285e+00 1.05665565e-01
5.24913430e-01 -8.71464789e-01 1.83363616e-01 2.88753629e-01
6.12720788e-01 -1.48204207e+00 -1.71946302e-01 -7.81239092e-01
-5.72285235e-01 6.12440944e-01 2.35305250e-01 2.36313805e-01
9.46247995e-01 1.29754245e-01 1.99533492e-01 1.32852169e-02
-3.24081153e-01 9.90018100e-02 1.85255170e-01 9.30205822e-01
2.17051446e-01 2.38850847e-01 -1.15416475e-01 3.70434314e-01
-3.53139281e-01 -2.92063296e-01 2.48410136e-01 9.46284354e-01
-2.49244243e-01 -1.46399140e+00 -7.15699136e-01 -1.16004042e-01
-1.12407970e+00 -2.45233566e-01 -3.99552554e-01 2.76876122e-01
-2.20710132e-02 1.49178755e+00 -6.83916032e-01 -8.61017644e-01
-1.28777757e-01 -8.45870897e-02 2.71148741e-01 -2.79424697e-01
-7.60684490e-01 2.46603154e-02 4.13690388e-01 -2.23628312e-01
-6.30067766e-01 -6.72183037e-01 -6.93455279e-01 -7.34898031e-01
-8.94593358e-01 3.29612970e-01 9.33173180e-01 8.40279162e-01
3.95808995e-01 -1.02022052e-01 1.13825238e+00 -7.17596352e-01
-8.67361665e-01 -1.08153605e+00 -3.67759347e-01 4.67048883e-01
9.00759280e-01 -4.67132568e-01 -6.50301158e-01 -1.46237597e-01]
|
[14.061816215515137, 5.863542079925537]
|
6103eda8-9ff0-43ed-aa26-258efea78c04
|
opening-the-black-box-of-deep-neural-networks
|
1703.00810
| null |
http://arxiv.org/abs/1703.00810v3
|
http://arxiv.org/pdf/1703.00810v3.pdf
|
Opening the Black Box of Deep Neural Networks via Information
|
Despite their great success, there is still no comprehensive theoretical
understanding of learning with Deep Neural Networks (DNNs) or their inner
organization. Previous work proposed to analyze DNNs in the \textit{Information
Plane}; i.e., the plane of the Mutual Information values that each layer
preserves on the input and output variables. They suggested that the goal of
the network is to optimize the Information Bottleneck (IB) tradeoff between
compression and prediction, successively, for each layer.
In this work we follow up on this idea and demonstrate the effectiveness of
the Information-Plane visualization of DNNs. Our main results are: (i) most of
the training epochs in standard DL are spent on {\emph compression} of the
input to efficient representation and not on fitting the training labels. (ii)
The representation compression phase begins when the training errors becomes
small and the Stochastic Gradient Decent (SGD) epochs change from a fast drift
to smaller training error into a stochastic relaxation, or random diffusion,
constrained by the training error value. (iii) The converged layers lie on or
very close to the Information Bottleneck (IB) theoretical bound, and the maps
from the input to any hidden layer and from this hidden layer to the output
satisfy the IB self-consistent equations. This generalization through noise
mechanism is unique to Deep Neural Networks and absent in one layer networks.
(iv) The training time is dramatically reduced when adding more hidden layers.
Thus the main advantage of the hidden layers is computational. This can be
explained by the reduced relaxation time, as this it scales super-linearly
(exponentially for simple diffusion) with the information compression from the
previous layer.
|
['Ravid Shwartz-Ziv', 'Naftali Tishby']
|
2017-03-02
| null | null | null | null |
['information-plane']
|
['methodology']
|
[ 3.41940373e-01 5.69295943e-01 -2.92852782e-02 -2.46261179e-01
-1.14903590e-02 -3.25858563e-01 4.02284563e-01 1.01978600e-01
-6.61463499e-01 5.75053275e-01 1.11297496e-01 -3.30703795e-01
-5.22975445e-01 -6.18813217e-01 -7.42390215e-01 -9.55684781e-01
-1.74135119e-01 5.82789123e-01 2.70464510e-01 2.75902543e-02
2.18250945e-01 6.39908433e-01 -1.45139611e+00 2.38766819e-01
7.45777726e-01 1.17375350e+00 3.03522825e-01 7.75137365e-01
-3.33121926e-01 9.20092106e-01 -4.63533968e-01 -4.18766350e-01
4.48326379e-01 -6.00089371e-01 -9.18124020e-01 -3.28568280e-01
1.55396447e-01 -2.84691632e-01 -4.57965761e-01 1.45280635e+00
3.44638258e-01 7.11919442e-02 7.32491195e-01 -9.13699985e-01
-4.28829432e-01 1.03181684e+00 -4.35544580e-01 2.17813358e-01
-4.78793055e-01 -1.97747499e-01 8.97509456e-01 -6.12267852e-01
8.10225129e-01 9.81728613e-01 7.16204643e-01 7.03160226e-01
-1.23055518e+00 -5.31628728e-01 8.73082057e-02 8.42002034e-03
-1.23602772e+00 -3.32010806e-01 6.01983249e-01 -3.72089535e-01
9.73953962e-01 3.43124658e-01 7.77884603e-01 5.45321465e-01
7.40588754e-02 6.22253001e-01 8.53180647e-01 -6.08903289e-01
2.68817574e-01 7.25492001e-01 4.54672843e-01 6.69717491e-01
5.06242812e-01 -4.93707657e-02 -5.08356929e-01 4.56446737e-01
7.91307747e-01 -1.97389238e-02 -3.51791471e-01 -2.53293663e-01
-5.38335562e-01 7.67830014e-01 5.91250241e-01 5.52435100e-01
-1.03431255e-01 9.88568217e-02 3.59100252e-01 6.66767180e-01
4.21280950e-01 3.58263463e-01 -6.39916837e-01 -1.72017366e-01
-1.29349148e+00 4.36256342e-02 9.32844400e-01 8.23032975e-01
8.85508537e-01 -5.06520690e-03 2.48088539e-01 7.81543851e-01
3.41373622e-01 2.29674831e-01 6.90770984e-01 -1.13376307e+00
6.57027066e-01 7.33481348e-01 -5.52894294e-01 -9.65304554e-01
-7.02840984e-01 -9.00609553e-01 -1.21196187e+00 4.22593266e-01
7.36621559e-01 -3.82357985e-01 -6.32913351e-01 1.90464473e+00
-1.50582761e-01 -5.31729817e-01 2.73183614e-01 6.53106093e-01
4.99268591e-01 6.43897295e-01 -2.40154609e-01 -3.97469342e-01
1.01099706e+00 -6.98998690e-01 -8.33681405e-01 -2.45043576e-01
9.69380736e-01 -3.75769764e-01 8.25923979e-01 3.34086001e-01
-1.50074530e+00 -3.67012382e-01 -1.31856799e+00 -2.06075147e-01
-4.95572180e-01 -4.16296814e-03 3.48214090e-01 5.31614721e-01
-1.30927014e+00 1.25193608e+00 -9.01311100e-01 -1.22815128e-02
6.04748726e-01 6.75817370e-01 -2.73864865e-01 1.52912989e-01
-1.03952849e+00 8.80164444e-01 7.29397595e-01 1.73442364e-01
-3.33962351e-01 -8.87335002e-01 -4.13732231e-01 4.14557308e-01
-1.17980942e-01 -4.74405855e-01 8.24838102e-01 -1.50558114e+00
-1.44568133e+00 7.16300666e-01 -4.20260206e-02 -7.15739489e-01
7.45756030e-01 -1.79957166e-01 2.37395048e-01 -6.18749633e-02
-5.17002702e-01 9.05856669e-01 4.81724292e-01 -1.34039199e+00
-3.09795320e-01 -6.55987918e-01 -3.41758460e-01 3.03322971e-01
-6.79617465e-01 -3.58934194e-01 -4.01646107e-01 -4.09036756e-01
5.80660462e-01 -8.55825007e-01 -2.47483142e-02 6.39148355e-02
-4.37867641e-01 -4.65227813e-02 7.84133375e-01 -7.20609546e-01
1.38677764e+00 -2.16266012e+00 2.88869828e-01 3.52060586e-01
4.96958435e-01 3.31764221e-01 1.55428767e-01 2.13718891e-01
-2.62498945e-01 3.85318667e-01 -3.81580800e-01 -4.69284385e-01
-4.24142219e-02 2.75841177e-01 -1.78126708e-01 5.06855786e-01
-6.22879378e-02 5.71591020e-01 -4.60365802e-01 -3.61156702e-01
-1.96069717e-01 5.63366890e-01 -6.65164411e-01 -7.67255947e-02
7.06681982e-02 3.91479656e-02 4.84316982e-03 -1.56146407e-01
9.28301513e-01 -1.68536618e-01 2.57022291e-01 -1.38069928e-01
-4.05631602e-01 3.00789833e-01 -1.14858460e+00 1.40617907e+00
-1.26329109e-01 1.21834183e+00 -4.48388755e-02 -1.21024942e+00
9.23040211e-01 5.38363755e-02 3.92637789e-01 -7.29625821e-01
1.25111699e-01 1.79144695e-01 4.13832515e-01 -1.76126257e-01
5.25920652e-02 -3.34043317e-02 6.21427894e-01 6.45878792e-01
3.51363212e-01 3.47293407e-01 2.80688733e-01 1.70123279e-01
8.34423363e-01 -2.14025021e-01 -4.95996140e-02 -5.06218970e-01
3.00817728e-01 -3.72007221e-01 4.25333887e-01 6.24956608e-01
3.08076963e-02 4.84379500e-01 1.10631907e+00 -2.50842065e-01
-1.35763490e+00 -7.37010658e-01 -3.64191920e-01 9.09849882e-01
-1.67269081e-01 -3.09597552e-01 -1.30675137e+00 -3.09493601e-01
-2.34967455e-01 7.00589240e-01 -6.66489184e-01 -4.10413533e-01
-6.13407969e-01 -6.58034980e-01 4.87246692e-01 3.44013900e-01
9.07210648e-01 -9.51938689e-01 -6.18402243e-01 -1.72319844e-01
1.81034297e-01 -6.52166784e-01 -1.71897858e-02 6.88284993e-01
-1.50535619e+00 -6.22793615e-01 -7.47348428e-01 -6.29468501e-01
7.61325538e-01 -2.83902586e-01 8.01100850e-01 -2.47235480e-03
-4.17320877e-02 -1.65629104e-01 1.87498748e-01 -4.96440738e-01
-3.97286028e-01 3.58378470e-01 -5.84045313e-02 -2.94526488e-01
1.64532229e-01 -7.80990303e-01 -6.46407187e-01 1.94307849e-01
-8.10283422e-01 3.53581071e-01 5.95561266e-01 4.56938684e-01
5.57060540e-01 3.04564267e-01 1.15394376e-01 -8.22242618e-01
5.43559372e-01 -2.50208288e-01 -6.84063077e-01 9.11181793e-02
-9.86989558e-01 6.38625085e-01 5.55077314e-01 -3.73387158e-01
-9.38273132e-01 -2.13778228e-03 -1.52159974e-01 -2.97238976e-01
1.69061482e-01 3.25204134e-01 -7.54398778e-02 8.37356201e-04
6.08682215e-01 1.21920690e-01 6.86286390e-02 -6.33116245e-01
2.35964715e-01 4.18367952e-01 1.78397894e-01 -4.19380479e-02
4.61455852e-01 2.34815121e-01 1.69799984e-01 -7.74408281e-01
-7.84541309e-01 3.51709276e-02 -9.68078077e-01 -2.97419012e-01
8.70006680e-01 -3.10942531e-01 -9.47027624e-01 3.16796392e-01
-1.25418901e+00 -4.97953385e-01 -7.82069862e-01 6.26214206e-01
-3.84927362e-01 6.48361146e-02 -7.51047313e-01 -9.73741710e-01
-3.49652022e-01 -9.70147848e-01 1.96624577e-01 3.19693714e-01
-3.64399957e-03 -1.28480136e+00 -1.64131913e-02 -4.15310673e-02
4.74070758e-01 -1.38128445e-01 1.27724838e+00 -8.51278126e-01
-3.72655481e-01 -6.93012401e-02 -4.86923575e-01 7.72263288e-01
-3.70812178e-01 4.31233132e-03 -1.10183060e+00 -1.35571554e-01
4.49271739e-01 -6.83001801e-02 1.22614348e+00 7.01270461e-01
1.16684461e+00 -6.87979639e-01 -1.99766323e-01 9.88237023e-01
1.49261236e+00 2.15038076e-01 5.20922303e-01 3.09732378e-01
5.28334081e-01 8.39077771e-01 -3.02424461e-01 2.04829574e-01
-6.22282028e-02 3.25795054e-01 3.09952617e-01 -2.19206735e-01
-3.81029367e-01 -2.43195802e-01 4.00277644e-01 1.28113616e+00
-2.32193425e-01 -1.48421764e-01 -8.42137337e-01 4.83032465e-02
-1.69790232e+00 -9.04374301e-01 -9.14060846e-02 2.49867725e+00
1.03334582e+00 5.38911045e-01 -2.02197701e-01 4.35457438e-01
4.04972017e-01 -1.67649072e-02 -8.62956882e-01 -7.22160876e-01
-2.39662588e-01 -1.68696374e-01 9.31177914e-01 7.28361368e-01
-7.09741890e-01 6.17259264e-01 6.60386992e+00 6.97119474e-01
-1.10833502e+00 5.68780899e-02 8.79393995e-01 -1.78619072e-01
-2.51768738e-01 -2.53256291e-01 -1.01171076e+00 2.88379163e-01
1.18302870e+00 -3.77343968e-02 6.03695810e-01 7.88354099e-01
1.19082844e-02 -1.10867796e-02 -1.16644025e+00 9.67043400e-01
-2.63564378e-01 -1.39719784e+00 1.10612042e-01 3.55595142e-01
6.63138926e-01 1.84150398e-01 3.12434137e-01 7.66766816e-02
2.33153328e-01 -1.00585186e+00 7.04605460e-01 6.38760567e-01
7.64400840e-01 -9.49535251e-01 7.12883115e-01 5.98417461e-01
-7.35785246e-01 -3.32357347e-01 -8.36639404e-01 3.80911492e-02
-2.36480534e-01 6.10961616e-01 -6.06862009e-01 -4.56120074e-02
5.48371673e-01 5.07936001e-01 -5.29336751e-01 7.16470540e-01
-5.05399145e-03 5.84395885e-01 -4.12635475e-01 5.20996973e-02
2.25026324e-01 -5.30170679e-01 4.30914462e-01 1.26989853e+00
2.02601418e-01 -8.81390125e-02 -6.33311808e-01 1.11215484e+00
-2.75577098e-01 -5.52254822e-03 -5.80893636e-01 -9.93032083e-02
1.98387519e-01 9.20836687e-01 -8.57717216e-01 -1.70832545e-01
-7.27070346e-02 8.38678360e-01 5.91667652e-01 4.54362065e-01
-4.58638102e-01 -5.43899059e-01 5.07126749e-01 1.69479653e-01
2.18731388e-01 -8.66009444e-02 -8.41755688e-01 -6.34335160e-01
3.42574390e-03 -3.31321299e-01 1.08932495e-01 -3.44849110e-01
-6.70154691e-01 6.65658176e-01 -1.02276333e-01 -6.99658990e-01
-2.03010604e-01 -6.41742110e-01 -2.68305749e-01 8.43958199e-01
-1.19136107e+00 -3.56673330e-01 5.12397476e-02 5.45297205e-01
2.93690145e-01 -1.26879409e-01 6.92133367e-01 4.59600240e-01
-7.05050826e-01 9.20781493e-01 5.12088895e-01 1.95171028e-01
1.26506966e-02 -1.28369057e+00 1.93306934e-02 6.08487666e-01
1.99913353e-01 6.41915500e-01 7.06397116e-01 -3.65014017e-01
-1.13794589e+00 -7.33412802e-01 1.04281032e+00 -6.48402749e-03
5.11148095e-01 -5.45911968e-01 -1.13157701e+00 5.26995718e-01
1.02535844e-01 -3.20130348e-01 5.19446969e-01 1.16560906e-02
-2.53627062e-01 -4.33737189e-01 -9.15241718e-01 5.08415341e-01
9.60659146e-01 -4.01539713e-01 -2.31256649e-01 4.05342281e-01
8.44834745e-01 -1.14349805e-01 -7.10370898e-01 7.70363957e-02
6.31477535e-01 -1.20153189e+00 6.12286389e-01 -5.18912435e-01
4.92731452e-01 1.97513804e-01 -3.80849652e-02 -1.04013383e+00
-1.99299827e-01 -6.26510918e-01 -9.22622681e-02 1.14181077e+00
8.74571443e-01 -5.31887650e-01 8.23778093e-01 7.73878455e-01
-9.96925607e-02 -8.39002490e-01 -1.03915930e+00 -5.19033492e-01
2.74155408e-01 -3.89491796e-01 1.72334369e-02 7.24071264e-01
4.82158847e-02 3.34971368e-01 -3.19253318e-02 -2.65041322e-01
5.60912907e-01 -5.10856092e-01 8.52788910e-02 -1.27189493e+00
-3.13473642e-01 -9.31502163e-01 -2.85160929e-01 -1.23900807e+00
-2.50717759e-01 -1.06995785e+00 -1.42642573e-01 -1.45789528e+00
2.01023176e-01 -5.26780486e-01 -4.00889486e-01 2.33950466e-01
3.62283379e-01 -2.10482955e-01 4.18396264e-01 5.39464772e-01
-1.89506248e-01 1.88296229e-01 1.19096732e+00 1.87818453e-01
-4.35123205e-01 2.06105456e-01 -4.89219606e-01 9.54608321e-01
8.23166192e-01 -6.30177617e-01 -6.44611657e-01 -6.74006701e-01
6.17777348e-01 4.78623137e-02 -7.67568946e-02 -1.08854496e+00
5.56105435e-01 5.54750264e-01 5.60809374e-01 -4.69085515e-01
2.64409304e-01 -8.93038630e-01 -3.15168686e-02 8.95658731e-01
-8.50299060e-01 9.23248529e-02 1.60985261e-01 3.02201778e-01
-2.07286160e-02 -7.19511807e-01 1.12756217e+00 -1.40919104e-01
1.34546682e-02 1.22740388e-01 -2.80732155e-01 8.64951387e-02
6.41055167e-01 -3.53942931e-01 -2.80681431e-01 -3.17988187e-01
-8.76926661e-01 -2.23463953e-01 1.50420532e-01 -2.03554034e-01
4.66580391e-01 -1.09705293e+00 -4.22057509e-01 4.20522541e-01
-7.59160340e-01 1.57350674e-01 5.27763776e-02 9.17249918e-01
-6.53103352e-01 4.92325157e-01 -2.12530166e-01 -5.89524627e-01
-1.12151527e+00 2.69369662e-01 7.06036329e-01 -4.15349662e-01
-7.29357362e-01 1.13436592e+00 1.92203432e-01 -1.79102629e-01
7.45266378e-01 -4.66089308e-01 -3.61099005e-01 3.61897290e-01
6.19000554e-01 5.71625412e-01 2.25553475e-02 -2.88472176e-01
2.80623650e-03 6.39085233e-01 -2.28037432e-01 -2.78322071e-01
1.48049951e+00 -3.09270490e-02 -3.05692971e-01 7.18771756e-01
1.73066771e+00 -3.68038505e-01 -1.52067101e+00 -1.71076298e-01
1.57616973e-01 7.76316077e-02 2.98397869e-01 -6.39205217e-01
-1.37622905e+00 1.21122468e+00 9.98603523e-01 4.81136084e-01
1.02327681e+00 -7.97090083e-02 4.58367765e-01 5.60647964e-01
-4.08910185e-01 -1.30949748e+00 -5.50513975e-02 6.59932911e-01
7.82726526e-01 -7.56274581e-01 -6.99699074e-02 5.37305623e-02
-5.78229129e-01 1.42100596e+00 2.98487663e-01 -1.56734496e-01
9.17861640e-01 4.98884648e-01 -2.41086707e-01 -1.70275122e-01
-7.93283105e-01 2.40909457e-01 1.62386790e-01 3.63475084e-01
4.80367720e-01 -2.14718506e-01 -1.38488874e-01 4.48102742e-01
-5.30027747e-01 -2.28894293e-01 6.86251149e-02 4.24547046e-01
-8.24706376e-01 -7.05513120e-01 1.52287737e-01 3.80090594e-01
-4.23021168e-01 -2.25250065e-01 -3.31794828e-01 8.47475052e-01
2.08674952e-01 5.65606892e-01 4.95416582e-01 -3.00540268e-01
1.56971142e-01 1.75429523e-01 3.30461472e-01 -7.97621310e-02
-4.66463447e-01 -1.24766015e-01 -3.73327106e-01 -4.70204443e-01
-3.61339301e-02 -6.62929595e-01 -1.46802104e+00 -5.69349706e-01
-3.44629139e-01 9.26478878e-02 1.12029827e+00 9.10510123e-01
5.83379716e-02 5.75648963e-01 4.78022844e-01 -7.13540494e-01
-5.73747218e-01 -9.07958508e-01 -5.80175102e-01 1.07312664e-01
3.66378427e-01 -1.63496092e-01 -7.80708015e-01 -2.17826720e-02]
|
[8.021390914916992, 3.538625717163086]
|
c1ef34e9-ee6c-4ed2-a62f-cd3026d20391
|
deep-adaptation-of-adult-child-facial
|
2209.08614
| null |
https://arxiv.org/abs/2209.08614v1
|
https://arxiv.org/pdf/2209.08614v1.pdf
|
Deep Adaptation of Adult-Child Facial Expressions by Fusing Landmark Features
|
Imaging of facial affects may be used to measure psychophysiological attributes of children through their adulthood, especially for monitoring lifelong conditions like Autism Spectrum Disorder. Deep convolutional neural networks have shown promising results in classifying facial expressions of adults. However, classifier models trained with adult benchmark data are unsuitable for learning child expressions due to discrepancies in psychophysical development. Similarly, models trained with child data perform poorly in adult expression classification. We propose domain adaptation to concurrently align distributions of adult and child expressions in a shared latent space to ensure robust classification of either domain. Furthermore, age variations in facial images are studied in age-invariant face recognition yet remain unleveraged in adult-child expression classification. We take inspiration from multiple fields and propose deep adaptive FACial Expressions fusing BEtaMix SElected Landmark Features (FACE-BE-SELF) for adult-child facial expression classification. For the first time in the literature, a mixture of Beta distributions is used to decompose and select facial features based on correlations with expression, domain, and identity factors. We evaluate FACE-BE-SELF on two pairs of adult-child data sets. Our proposed FACE-BE-SELF approach outperforms adult-child transfer learning and other baseline domain adaptation methods in aligning latent representations of adult and child expressions.
|
['Khan M. Iftekharuddin', 'Haim Y. Bar', 'Norou Diawara', 'Manar D. Samad', 'Megan A. Witherow']
|
2022-09-18
| null | null | null | null |
['age-invariant-face-recognition']
|
['computer-vision']
|
[ 2.42362708e-01 8.57977271e-02 -3.51328403e-01 -1.12180221e+00
-3.11823517e-01 -4.62317020e-01 4.66244489e-01 9.68619362e-02
-4.22819704e-01 4.92434978e-01 1.21430710e-01 7.06020176e-01
2.48293709e-02 -4.67525035e-01 -2.88157731e-01 -8.72598827e-01
-1.35473996e-01 3.94784629e-01 -7.41954923e-01 7.79219344e-02
-2.48451054e-01 7.04257607e-01 -1.74770200e+00 3.32811117e-01
6.05564833e-01 1.17851329e+00 -5.31775594e-01 3.10809851e-01
-1.26655608e-01 3.98399651e-01 -4.99046654e-01 -5.51402926e-01
7.62039199e-02 -4.88636643e-01 -5.10847270e-01 2.73474455e-01
9.74695146e-01 -4.51540947e-01 5.22632040e-02 1.13072991e+00
6.88758194e-01 -1.57963932e-01 1.22808278e+00 -1.71667159e+00
-6.21209383e-01 1.89026505e-01 -9.31952596e-01 9.33870003e-02
4.30427760e-01 -2.30626434e-01 5.15912235e-01 -7.99764693e-01
4.51588899e-01 1.30095136e+00 5.24593830e-01 1.07702053e+00
-1.38451982e+00 -1.44280124e+00 1.01314195e-01 2.62594353e-02
-1.25027764e+00 -7.46607542e-01 8.54946017e-01 -7.98284471e-01
7.35407412e-01 -3.47896993e-01 8.38767409e-01 1.53476155e+00
3.76177914e-02 5.66569507e-01 1.37253821e+00 -2.50346601e-01
2.65725683e-02 -1.58958882e-02 -2.05596626e-01 7.45374084e-01
-2.95830220e-01 1.24987431e-01 -9.83956516e-01 -1.09049410e-01
7.38930285e-01 -3.10301811e-01 1.13726325e-01 -1.95679337e-01
-6.46005630e-01 7.21693099e-01 7.47857541e-02 3.36065173e-01
-5.11905372e-01 1.11593474e-02 4.96783555e-01 5.70420563e-01
1.08122051e+00 1.35130703e-01 -5.76954126e-01 -2.59751439e-01
-8.32109272e-01 3.33294868e-01 2.00379908e-01 6.68136537e-01
8.11661005e-01 2.37143606e-01 -7.87356645e-02 1.41157019e+00
5.84339201e-02 4.36990589e-01 5.41506410e-01 -1.04314995e+00
-1.55052453e-01 5.13492584e-01 -4.26716536e-01 -6.95954382e-01
-7.36319542e-01 -1.46499068e-01 -6.95153236e-01 5.11714280e-01
3.35905075e-01 -2.05002740e-01 -7.62400568e-01 2.43000150e+00
4.72491056e-01 3.12298071e-02 3.68216783e-02 5.62925518e-01
8.60522568e-01 1.42337263e-01 6.34647489e-01 -4.05825287e-01
1.22301483e+00 -5.14418125e-01 -5.35267532e-01 -1.69921070e-01
4.92448837e-01 -3.61607581e-01 5.05203009e-01 5.34497559e-01
-1.15349805e+00 -4.78047490e-01 -6.91844463e-01 2.46472195e-01
-1.31809145e-01 1.55995116e-01 9.82629418e-01 8.91317129e-01
-1.06910658e+00 4.33882594e-01 -7.47537076e-01 -4.69091326e-01
9.56143379e-01 6.58852577e-01 -1.07851315e+00 3.66279215e-01
-8.27110231e-01 8.13988268e-01 -9.13252905e-02 -1.70768768e-01
-7.51001358e-01 -9.48542833e-01 -9.88967478e-01 -1.02704540e-01
-5.47197044e-01 -3.57354760e-01 1.17290807e+00 -1.97259748e+00
-1.72721779e+00 1.82446003e+00 3.27412109e-03 -1.90703366e-02
-7.28033110e-02 5.53906113e-02 -5.75863957e-01 4.40823257e-01
1.33790255e-01 1.16292739e+00 1.22586751e+00 -6.36233449e-01
-2.29383007e-01 -1.12534559e+00 -4.84033614e-01 5.34827076e-02
-6.19821846e-01 6.02137327e-01 1.69156983e-01 -5.36491334e-01
1.19304739e-01 -8.08054686e-01 2.56661415e-01 5.68990767e-01
5.16477108e-01 -4.60538864e-01 5.22319376e-01 -4.97265965e-01
6.25733972e-01 -2.18432593e+00 2.03236356e-01 3.70606072e-02
2.05741033e-01 -2.15102434e-02 -3.24561924e-01 -1.83045473e-02
-6.96406841e-01 -3.34120721e-01 2.08046455e-02 -6.02011383e-01
-7.92167783e-02 1.20204300e-01 7.82323331e-02 9.43717182e-01
6.45306945e-01 7.34572709e-01 -7.32198179e-01 -5.47727466e-01
-1.34099498e-01 4.26804185e-01 -5.50155699e-01 5.59475362e-01
-1.50342276e-02 8.38112772e-01 -2.19503939e-01 9.17393923e-01
7.53094852e-01 3.92947614e-01 7.25586480e-03 4.09141332e-02
2.64349818e-01 -1.47436097e-01 -2.18639299e-01 1.74183798e+00
-5.92325389e-01 5.79255044e-01 5.76873183e-01 -1.16475475e+00
1.34621668e+00 6.00286424e-01 8.72552991e-01 -7.01851070e-01
4.02317524e-01 1.56275019e-01 1.48591876e-01 -4.84327614e-01
-3.88307273e-01 -6.55755162e-01 9.08514708e-02 3.62322867e-01
8.07394922e-01 -1.78523183e-01 -2.10159749e-01 -3.64144653e-01
8.41689646e-01 3.25633407e-01 2.68796742e-01 -1.59029856e-01
3.85544628e-01 -7.82284617e-01 6.13373339e-01 -9.46630016e-02
-6.52305186e-01 4.31813270e-01 7.13063240e-01 -4.90165979e-01
-7.78232813e-01 -1.31174684e+00 -5.01732349e-01 1.64001060e+00
-6.90441191e-01 7.48804733e-02 -7.93803394e-01 -6.08771920e-01
2.24835170e-03 4.01780367e-01 -1.00165081e+00 -3.47369671e-01
-1.23180971e-01 -6.60110295e-01 9.50123966e-01 7.27320850e-01
7.58857876e-02 -1.11685598e+00 -3.19899797e-01 -2.60518640e-02
4.33453649e-01 -1.16996348e+00 -1.03537455e-01 1.89488843e-01
-4.76385742e-01 -8.95876944e-01 -7.90238678e-01 -7.65090406e-01
8.07010770e-01 -6.10355198e-01 1.02163184e+00 -1.73455328e-01
-3.25198174e-01 9.29765046e-01 -2.96812236e-01 -7.27698088e-01
-5.26143193e-01 -3.77974808e-01 6.31589532e-01 2.95705229e-01
8.38068902e-01 -1.18975616e+00 -5.74948847e-01 8.08359757e-02
-8.04600656e-01 -2.40002528e-01 3.18006217e-01 7.05016196e-01
7.97783509e-02 -5.59175849e-01 8.18331003e-01 -4.37285483e-01
3.30766559e-01 -7.13666975e-01 -2.41523772e-01 5.23692742e-02
-5.37993073e-01 -2.51620293e-01 2.71403551e-01 -8.18039536e-01
-1.07352793e+00 1.20425820e-01 -2.54503101e-01 -7.31819689e-01
-6.67932034e-01 2.05643680e-02 -1.20844759e-01 -1.04751125e-01
5.34091353e-01 -2.57586062e-01 4.84732807e-01 -1.51517227e-01
2.87734699e-02 7.56941497e-01 7.29723334e-01 -1.06767964e+00
3.67740959e-01 3.20129275e-01 2.99338400e-01 -7.19921350e-01
-7.56354034e-01 -2.32558727e-01 -1.02641821e+00 -4.16967005e-01
1.13425159e+00 -1.12402987e+00 -7.10768998e-01 1.02804351e+00
-1.08873606e+00 -2.61802942e-01 4.04318050e-02 4.46296334e-01
-8.28984141e-01 -7.11501762e-02 -4.65790957e-01 -8.13089132e-01
-4.43569541e-01 -9.42315817e-01 1.35609984e+00 2.36803502e-01
-6.12888575e-01 -9.51268971e-01 3.76916349e-01 1.01364896e-01
9.93903652e-02 6.79347754e-01 8.93091917e-01 -6.98725462e-01
5.17104268e-01 -1.48852810e-01 -6.72900826e-02 5.95950246e-01
4.15495485e-01 2.91208297e-01 -1.36556184e+00 -2.11211979e-01
-9.16543156e-02 -1.02522993e+00 5.39207160e-01 2.95908719e-01
1.23881280e+00 -1.35970891e-01 3.83314081e-02 9.64889765e-01
8.24490070e-01 1.83607936e-01 2.97126025e-01 -1.91822842e-01
3.16602230e-01 1.38236499e+00 4.13670063e-01 6.57319069e-01
1.89780697e-01 7.60920942e-01 1.79430649e-01 -2.22223774e-02
1.37857795e-01 -1.79025784e-01 5.50193131e-01 2.89075196e-01
-3.58997136e-02 2.99018979e-01 -8.89668286e-01 5.96073389e-01
-1.18577886e+00 -8.08390439e-01 4.21029985e-01 1.85401607e+00
1.05037415e+00 -3.29781085e-01 5.01977265e-01 -1.97848901e-01
5.89624822e-01 2.20247302e-02 -6.34204805e-01 -8.33691657e-01
-1.69798613e-01 9.30045784e-01 -1.42521217e-01 -5.06603569e-02
-1.07437503e+00 8.90670955e-01 5.95601177e+00 5.34756601e-01
-1.47117603e+00 2.02528536e-01 9.18346703e-01 -2.99017757e-01
-1.81786660e-02 -4.78279203e-01 -4.71153766e-01 1.20636947e-01
9.45140779e-01 -1.00761428e-01 1.53171852e-01 9.06428933e-01
-1.43416122e-01 2.45673478e-01 -1.52191043e+00 1.28251255e+00
1.83247417e-01 -4.75429654e-01 -3.07122350e-01 -2.68667787e-01
7.29090929e-01 -2.76836127e-01 5.87998807e-01 3.54155421e-01
-6.71559423e-02 -1.54775727e+00 5.07913530e-01 4.65929717e-01
1.42252076e+00 -7.61078835e-01 1.61689475e-01 -3.77835818e-02
-7.50336051e-01 -5.50027639e-02 -1.24497958e-01 -3.00459921e-01
-3.19563329e-01 -2.54846495e-02 -7.93463171e-01 -1.30017832e-01
8.36333334e-01 7.09176064e-01 -2.95178980e-01 1.22637682e-01
-8.63680542e-02 4.58008587e-01 -1.61750808e-01 2.91094124e-01
4.21904959e-02 -2.70463467e-01 2.11136311e-01 1.23322284e+00
3.26898962e-01 1.89521939e-01 -2.68242568e-01 9.52669144e-01
-8.83586332e-03 4.08126056e-01 -6.69754207e-01 -3.36562455e-01
1.86189279e-01 1.42325699e+00 -1.83442578e-01 1.00877561e-01
-5.09955943e-01 1.01535022e+00 5.10789692e-01 2.18408376e-01
-5.16922176e-01 3.78567517e-01 1.27869809e+00 8.35450888e-02
-1.60723910e-01 4.39764634e-02 -2.93679107e-02 -9.29111600e-01
-3.18732262e-01 -9.74065959e-01 3.13034475e-01 -8.14403772e-01
-1.56242919e+00 6.13806188e-01 4.21279818e-01 -9.59991693e-01
-6.99825287e-01 -8.42433572e-01 -7.83072829e-01 7.29028821e-01
-9.81967330e-01 -1.71175396e+00 -3.10633332e-01 6.21793866e-01
2.05682188e-01 -4.96171087e-01 1.22979915e+00 1.50429249e-01
-5.26855588e-01 9.84535813e-01 -3.74037594e-01 2.77474463e-01
9.98812020e-01 -1.06131589e+00 -3.46684977e-02 1.78082332e-01
1.70205981e-01 5.21951854e-01 5.13853908e-01 -3.97895992e-01
-6.81106150e-01 -7.80644417e-01 3.42396647e-01 -2.85218596e-01
5.91550231e-01 -4.27614361e-01 -7.81544745e-01 6.47942841e-01
1.54526591e-01 2.26087451e-01 1.16018498e+00 3.47011179e-01
-9.61412251e-01 -3.75896275e-01 -1.44715834e+00 2.29863301e-01
9.69512701e-01 -8.02012861e-01 -2.72003412e-01 9.33639109e-02
8.53056237e-02 -1.18154548e-01 -1.24085879e+00 7.86809981e-01
1.07999527e+00 -1.14982092e+00 5.78330100e-01 -9.78352904e-01
5.78380883e-01 5.03147781e-01 -1.07128054e-01 -1.30445921e+00
-5.50571568e-02 -4.03958827e-01 2.73527384e-01 1.56712949e+00
-9.71233472e-02 -5.43891609e-01 9.97076273e-01 6.60803676e-01
3.32458079e-01 -9.74766970e-01 -1.15510368e+00 -4.45506901e-01
6.09807611e-01 -3.36668462e-01 6.62402630e-01 1.06819367e+00
-1.84643921e-02 2.40496516e-01 4.28467765e-02 -2.09161342e-04
4.82539564e-01 -8.50131810e-02 6.77714407e-01 -1.23544657e+00
-1.29199162e-01 -7.15650320e-01 -9.53526616e-01 -2.44832009e-01
1.20907748e+00 -1.03525901e+00 -3.48743886e-01 -6.22826815e-01
3.40921313e-01 -1.25853226e-01 -4.66604859e-01 6.89239860e-01
1.27951384e-01 3.01443547e-01 -2.92576283e-01 -4.58828390e-01
1.59592424e-02 7.41298914e-01 8.68569911e-01 -1.28854990e-01
1.16081789e-01 1.10768355e-01 -3.50382507e-01 9.65339482e-01
7.09461749e-01 -2.79467434e-01 -3.89524013e-01 2.37416718e-02
-8.89340416e-02 1.41254976e-01 1.96733698e-01 -8.02907944e-01
-3.59466374e-01 -1.60186991e-01 8.30068290e-01 -1.15659915e-01
6.21219277e-01 -6.95842564e-01 -1.05941072e-01 4.53544408e-02
-2.98308909e-01 -7.15171769e-02 2.70279527e-01 -4.06376794e-02
-2.69747645e-01 -1.81091532e-01 1.16535902e+00 1.42922133e-01
-4.74969774e-01 7.90375292e-01 -3.70106071e-01 -1.01657532e-01
9.29145098e-01 -1.99704662e-01 -3.26441303e-02 -6.18314385e-01
-8.67563784e-01 -1.17346318e-02 5.69640219e-01 4.88783121e-01
4.61456478e-01 -1.39711857e+00 -8.60525489e-01 5.13198733e-01
4.07675058e-01 -2.87474215e-01 3.11035216e-01 8.78012002e-01
-2.03804523e-01 -1.40622124e-01 -9.64113951e-01 -6.92634106e-01
-1.74727702e+00 3.31021428e-01 4.12815243e-01 1.01412021e-01
2.04089656e-01 1.31830120e+00 7.29291916e-01 -5.64590931e-01
1.23824961e-01 1.05160885e-01 -3.01241100e-01 5.17253697e-01
4.91365254e-01 1.07067168e-01 -2.62929231e-01 -1.17326593e+00
-3.49916339e-01 7.13256896e-01 -5.08838817e-02 -1.23801567e-02
1.54245651e+00 1.90661520e-01 -5.12916505e-01 3.65128696e-01
1.52855706e+00 -1.56619415e-01 -1.16076040e+00 -9.76573378e-02
-1.32683411e-01 -2.53974766e-01 -2.51978546e-01 -4.69680607e-01
-1.17397237e+00 1.13550758e+00 9.59249437e-01 -4.86151665e-01
1.71864867e+00 2.36554906e-01 2.44370893e-01 -1.55210778e-01
3.55184466e-01 -9.22831118e-01 3.40966016e-01 4.43214923e-02
1.13784802e+00 -1.15725338e+00 7.43881194e-03 -1.48153365e-01
-6.47605002e-01 1.29285264e+00 1.13795948e+00 2.22097561e-02
6.42546892e-01 3.57795149e-01 3.30765903e-01 -2.53070384e-01
-9.06831205e-01 -6.66931495e-02 4.21123445e-01 9.84399498e-01
7.17496872e-01 4.54185866e-02 -1.03442907e-01 7.78018475e-01
-3.55312914e-01 -1.97508767e-01 -1.35570005e-01 4.83172238e-01
-1.18340440e-02 -1.40512896e+00 -2.81009167e-01 4.44291592e-01
-8.66163909e-01 2.88006485e-01 -4.92113590e-01 5.88698506e-01
5.21217942e-01 4.52969223e-01 5.21851778e-01 -1.71228036e-01
-6.43300191e-02 5.47866762e-01 1.24683070e+00 -8.30045879e-01
-4.42936927e-01 -3.67501006e-02 2.86773797e-02 -4.79052007e-01
-7.92561948e-01 -9.80056643e-01 -1.10630548e+00 2.13055938e-01
1.75581902e-01 -3.74484420e-01 7.37473071e-01 7.65001714e-01
-3.34353708e-02 -1.58522934e-01 7.57068634e-01 -8.48433673e-01
-4.33436781e-01 -1.16146219e+00 -9.35867906e-01 6.98591769e-01
3.68011206e-01 -1.03282964e+00 -1.99889004e-01 6.98982179e-02]
|
[13.54330062866211, 1.7002654075622559]
|
7554ab7f-c20b-4265-927c-b6850a9d8324
|
compressed-predictive-information-coding
|
2203.02051
| null |
https://arxiv.org/abs/2203.02051v1
|
https://arxiv.org/pdf/2203.02051v1.pdf
|
Compressed Predictive Information Coding
|
Unsupervised learning plays an important role in many fields, such as artificial intelligence, machine learning, and neuroscience. Compared to static data, methods for extracting low-dimensional structure for dynamic data are lagging. We developed a novel information-theoretic framework, Compressed Predictive Information Coding (CPIC), to extract useful representations from dynamic data. CPIC selectively projects the past (input) into a linear subspace that is predictive about the compressed data projected from the future (output). The key insight of our framework is to learn representations by minimizing the compression complexity and maximizing the predictive information in latent space. We derive variational bounds of the CPIC loss which induces the latent space to capture information that is maximally predictive. Our variational bounds are tractable by leveraging bounds of mutual information. We find that introducing stochasticity in the encoder robustly contributes to better representation. Furthermore, variational approaches perform better in mutual information estimation compared with estimates under a Gaussian assumption. We demonstrate that CPIC is able to recover the latent space of noisy dynamical systems with low signal-to-noise ratios, and extracts features predictive of exogenous variables in neuroscience data.
|
['Kristofer Bouchard', 'Tianyi Luo', 'Rui Meng']
|
2022-03-03
| null | null | null | null |
['mutual-information-estimation']
|
['methodology']
|
[ 5.06651819e-01 3.87870461e-01 -4.00145531e-01 -1.75069854e-01
-7.52309144e-01 -5.33030391e-01 7.15400338e-01 -9.23231766e-02
-1.15065098e-01 6.10058844e-01 6.29159927e-01 1.14670090e-01
-5.88952243e-01 -3.49794120e-01 -7.81643748e-01 -9.14828598e-01
-4.94172692e-01 2.39727005e-01 -2.98410982e-01 4.19825077e-01
2.12395817e-01 4.81599450e-01 -1.40393269e+00 1.50810108e-01
5.53988159e-01 8.99747968e-01 5.34040391e-01 6.42702341e-01
2.39021167e-01 7.55225837e-01 -1.53211504e-01 -9.50726569e-02
2.33418882e-01 -5.41726232e-01 -7.26390183e-01 2.87144836e-02
-3.23593616e-01 -6.56546652e-03 -9.68941689e-01 1.25905609e+00
1.64915189e-01 1.98817268e-01 1.02365017e+00 -1.15081155e+00
-5.00966012e-01 5.96782029e-01 -1.31605208e-01 3.89064461e-01
7.73746595e-02 -1.65721208e-01 1.20613384e+00 -6.51114106e-01
9.60111141e-01 1.17261267e+00 3.60143125e-01 5.62066197e-01
-1.67982829e+00 -4.11667198e-01 -7.63731077e-02 1.23519674e-01
-1.25560308e+00 -8.74397516e-01 9.53073323e-01 -7.84490824e-01
6.11015975e-01 1.63930938e-01 4.93690461e-01 1.24455893e+00
5.25718093e-01 1.16122854e+00 8.08925569e-01 -2.88119465e-01
5.18982828e-01 3.24910157e-03 5.52972592e-02 4.48877424e-01
1.23254061e-01 2.49101132e-01 -8.62291932e-01 -2.18142226e-01
7.44085014e-01 4.11586881e-01 -4.52567816e-01 -5.60108840e-01
-1.23315322e+00 9.37995255e-01 -4.22538556e-02 3.59751970e-01
-4.56227601e-01 2.32488364e-01 3.36474210e-01 3.81121904e-01
5.33825099e-01 4.84011889e-01 -4.25062627e-01 -5.45712411e-01
-1.00022328e+00 9.59901884e-03 8.50161552e-01 1.10518122e+00
4.45988655e-01 4.54912819e-02 -1.60497978e-01 5.58044970e-01
2.49516249e-01 6.99531376e-01 7.40786552e-01 -1.43523002e+00
4.61640328e-01 1.57475486e-01 -8.60205218e-02 -1.11053443e+00
-2.04271078e-01 -5.05913913e-01 -1.06332862e+00 -4.93287653e-01
-3.34784016e-02 -1.00488417e-01 -5.68397939e-01 2.12332082e+00
-2.02337652e-01 1.53959289e-01 3.40666920e-01 6.22605860e-01
6.53750077e-02 8.91510010e-01 -4.71390605e-01 -9.66079652e-01
7.19340622e-01 -2.91118562e-01 -1.04037857e+00 -2.92385370e-01
4.26281005e-01 -2.31569469e-01 4.84667361e-01 2.91925788e-01
-1.08326018e+00 -1.85464099e-01 -8.82353663e-01 1.18258119e-01
1.89087495e-01 -8.69203359e-02 5.12090802e-01 3.33012342e-01
-9.25595045e-01 1.02609968e+00 -1.39720762e+00 -1.14094138e-01
4.57074821e-01 2.24140316e-01 -5.34600854e-01 7.81230070e-03
-8.67332041e-01 5.60838163e-01 3.65244746e-01 -2.76192874e-01
-1.04826069e+00 -5.78726649e-01 -8.98133397e-01 2.40166396e-01
1.97076380e-01 -4.76823181e-01 8.64602089e-01 -6.72179341e-01
-1.35560524e+00 3.11637640e-01 -4.87345040e-01 -8.78768027e-01
2.15910241e-01 -1.82469353e-01 -1.15422651e-01 5.84523141e-01
1.02432117e-01 2.12725371e-01 1.07220328e+00 -9.16543603e-01
6.45642504e-02 -5.31642079e-01 -5.03573835e-01 -2.21602079e-02
-4.39543247e-01 -3.61831456e-01 -4.68173951e-01 -7.35995829e-01
7.05743253e-01 -1.05985367e+00 -3.35934043e-01 3.20359878e-02
-2.66057551e-01 8.04162771e-02 7.41940260e-01 -6.38755202e-01
1.28707850e+00 -2.20718050e+00 7.42444932e-01 2.01614037e-01
4.84600216e-01 -2.62059439e-02 1.06795602e-01 6.44228220e-01
-6.70119375e-02 5.50619848e-02 -2.66374052e-01 -5.64197958e-01
-1.16894312e-01 6.03943944e-01 -8.33678126e-01 7.29678214e-01
-5.61360363e-03 8.59365344e-01 -9.28553760e-01 -3.22141647e-01
-3.18216234e-02 4.04418558e-01 -7.07921743e-01 2.39113614e-01
-1.07250623e-01 5.95212579e-01 -6.28009558e-01 2.33703777e-01
3.75282109e-01 -5.94897747e-01 2.18158394e-01 -4.50941138e-02
2.38226235e-01 1.99274227e-01 -7.89542377e-01 1.87442875e+00
-3.52513313e-01 1.00557327e+00 1.91764627e-02 -1.51825917e+00
6.68863773e-01 3.95083606e-01 8.61818552e-01 -2.65223384e-01
-6.37852922e-02 -1.25851884e-01 -3.11783522e-01 -5.84013283e-01
1.49147883e-01 -1.75852165e-01 -1.11957073e-01 4.17566448e-01
3.80240738e-01 1.91172853e-01 -1.19243123e-01 5.75610876e-01
1.19798315e+00 -9.07732546e-02 4.86371696e-01 -2.00514480e-01
2.41164863e-02 -4.90959555e-01 8.54183733e-01 8.81911755e-01
-2.43528664e-01 5.35703003e-01 7.23555505e-01 7.82993436e-02
-1.16684020e+00 -1.13853168e+00 -3.25245947e-01 6.05385184e-01
-1.46875888e-01 -6.30257308e-01 -5.06769240e-01 -2.32654378e-01
-1.39644623e-01 7.34242082e-01 -6.18490696e-01 -4.89876181e-01
-1.51766360e-01 -5.00777423e-01 2.72961229e-01 3.15691382e-01
3.61431725e-02 -6.34904921e-01 -5.75654984e-01 1.53569251e-01
-3.89435887e-01 -1.18435657e+00 -4.41252589e-01 4.49774116e-01
-1.25472653e+00 -7.07102716e-01 -6.41402841e-01 -2.53325373e-01
5.64601600e-01 8.55074897e-02 6.39243305e-01 -7.05812573e-01
-1.57958418e-01 7.16553211e-01 -2.95413107e-01 -1.46634832e-01
-4.44852501e-01 -2.26427510e-01 5.03622234e-01 1.22209080e-01
1.25077292e-01 -9.15355742e-01 -4.09238011e-01 -9.69620887e-03
-9.50986207e-01 2.64963686e-01 5.86516500e-01 1.13660240e+00
7.32872605e-01 1.52629584e-01 3.94384116e-01 -7.26168573e-01
3.64595681e-01 -8.32463086e-01 -4.58930194e-01 -6.21608738e-03
-7.48232901e-01 7.91192472e-01 7.00686991e-01 -4.16414917e-01
-7.41315782e-01 2.24590465e-01 2.04133064e-01 -1.00221717e+00
1.74264356e-01 6.12451971e-01 1.82108989e-03 4.17211175e-01
4.15840745e-01 7.36722291e-01 3.07342678e-01 -6.12617850e-01
2.99659699e-01 5.76159537e-01 4.75589722e-01 -4.29614723e-01
4.37467754e-01 6.02445662e-01 2.69696206e-01 -9.96957898e-01
-9.33940411e-01 -6.11532032e-01 -7.91242480e-01 -4.41650040e-02
4.89107400e-01 -8.95621479e-01 -6.39365494e-01 -1.08561823e-02
-9.48285103e-01 2.02822506e-01 -6.04828000e-01 8.15591931e-01
-1.07398915e+00 3.71654332e-01 -5.36225259e-01 -1.17732608e+00
-1.53972328e-01 -1.05373430e+00 1.04397607e+00 -2.16003790e-01
-1.18416734e-01 -1.08024049e+00 3.45573902e-01 2.89337561e-02
1.03001438e-01 1.61843792e-01 6.43644571e-01 -7.04497218e-01
-5.69711089e-01 -3.95033956e-01 1.43950164e-01 3.76402736e-01
-2.91031599e-02 -3.15854430e-01 -9.19452965e-01 -3.46157372e-01
4.73810941e-01 -1.05035990e-01 1.26534855e+00 6.23062074e-01
1.19377363e+00 -8.94918442e-01 -6.10515177e-01 8.82507503e-01
1.21717966e+00 1.38916612e-01 3.91969502e-01 -2.79121846e-01
4.47772950e-01 4.99305278e-01 1.69014916e-01 8.29401791e-01
7.77129689e-03 5.42314947e-01 2.47899577e-01 8.22615802e-01
2.68879026e-01 -4.64421779e-01 6.09584570e-01 1.40435493e+00
1.62018463e-01 -2.57168319e-02 -6.38850093e-01 6.76672995e-01
-1.94262075e+00 -1.27193654e+00 4.71818775e-01 2.17656326e+00
9.27294433e-01 -1.69092827e-02 -2.66306907e-01 2.01729804e-01
4.92301345e-01 1.41124904e-01 -9.09471035e-01 5.73269278e-02
-2.17667133e-01 -1.98980093e-01 4.54421580e-01 3.32709998e-01
-1.06403899e+00 5.05401671e-01 6.94034958e+00 8.43735754e-01
-7.41903543e-01 1.50428280e-01 5.19155800e-01 -3.21329743e-01
-3.62113446e-01 2.45388448e-02 -7.05637336e-01 6.92373037e-01
1.44159913e+00 -6.52035058e-01 5.97972155e-01 9.14411306e-01
3.40156019e-01 1.68289825e-01 -1.22988093e+00 1.11494374e+00
-2.21305657e-02 -1.50312436e+00 3.36366855e-02 3.52875978e-01
9.10948098e-01 1.22389190e-01 3.06760043e-01 -7.56712630e-02
1.74403384e-01 -7.93328285e-01 5.83265245e-01 1.15569496e+00
7.48140872e-01 -6.30625010e-01 3.63651931e-01 7.86745191e-01
-8.73011231e-01 -3.36942196e-01 -7.94345498e-01 -6.67125955e-02
7.87743703e-02 9.27464604e-01 -7.02081442e-01 2.52768308e-01
3.60909969e-01 1.31967044e+00 -1.19467638e-01 7.26988912e-01
1.50483400e-01 8.00628006e-01 -3.75832289e-01 2.18650311e-01
-1.04685530e-01 -3.59309763e-01 9.53911126e-01 9.11257446e-01
4.24472749e-01 1.31205782e-01 8.40676054e-02 9.83536780e-01
-8.15658048e-02 -1.84684768e-01 -1.01701987e+00 -6.78872228e-01
5.33030987e-01 6.12686515e-01 -4.79267389e-01 -2.14871541e-01
-1.74639925e-01 1.15656507e+00 1.95366427e-01 4.62895453e-01
-2.12760583e-01 -1.42219573e-01 7.11679637e-01 -1.38825402e-01
4.17257965e-01 -5.21263123e-01 -2.53363758e-01 -1.45951498e+00
-6.18300214e-02 -4.27789658e-01 3.08857620e-01 -4.77758437e-01
-1.16238594e+00 2.46137649e-01 2.88806945e-01 -1.38912070e+00
-8.43906045e-01 -2.91121274e-01 -1.82251751e-01 6.80581987e-01
-9.11104739e-01 -7.09245384e-01 4.36589420e-01 5.53586841e-01
5.34975648e-01 -3.92441690e-01 8.58130634e-01 -1.74286604e-01
-5.27773678e-01 4.13100719e-01 1.05866706e+00 -7.24116862e-02
1.37398183e-01 -1.08753538e+00 3.63124460e-02 9.36754405e-01
4.13024157e-01 9.00943756e-01 8.49430501e-01 -6.34957433e-01
-1.70957899e+00 -1.00802588e+00 8.37463021e-01 -3.37086648e-01
8.89801621e-01 -3.26277137e-01 -7.36812592e-01 8.64507020e-01
-3.55418593e-01 5.66493906e-02 7.86282122e-01 -3.75348106e-02
-3.31516623e-01 5.52575625e-02 -8.53999078e-01 4.20022547e-01
9.94534016e-01 -1.03311145e+00 -5.70499599e-01 4.90947127e-01
9.69376087e-01 -9.62053612e-02 -1.06968904e+00 4.59852293e-02
6.01809263e-01 -5.24000227e-01 1.04647899e+00 -7.76281893e-01
5.69804966e-01 2.36692101e-01 -4.95823443e-01 -1.22708118e+00
-3.55583221e-01 -1.06120622e+00 -7.97575116e-01 8.94517839e-01
1.75529107e-01 -3.87821287e-01 6.50610745e-01 6.76484048e-01
1.18134182e-03 -7.14083254e-01 -1.11433935e+00 -9.15730000e-01
1.19046479e-01 -6.61461532e-01 -6.09832257e-02 5.91490388e-01
5.58863521e-01 8.51919800e-02 -7.57655978e-01 8.12553763e-02
9.19676542e-01 2.15435490e-01 2.61742055e-01 -1.26791561e+00
-5.33303618e-01 -1.46073818e-01 -7.72408068e-01 -1.36341333e+00
3.09804559e-01 -1.21875811e+00 -1.14302719e-02 -1.05941260e+00
5.43738604e-01 -5.93393706e-02 -3.71429861e-01 2.03082949e-01
1.52201056e-01 -2.76626676e-01 3.15827787e-01 8.43982637e-01
-5.53795576e-01 9.90899920e-01 1.09490025e+00 -2.01916378e-02
-1.87466666e-01 2.10010916e-01 -6.34098649e-01 7.38737404e-01
5.71227014e-01 -6.53712094e-01 -5.74455917e-01 -7.09457025e-02
1.43835127e-01 5.83247900e-01 7.79379383e-02 -9.45134163e-01
4.38250333e-01 -2.11207077e-01 3.43236953e-01 -6.14264607e-01
6.11863613e-01 -6.36082470e-01 1.14216365e-01 3.43017578e-01
-9.15312767e-01 -4.46039706e-01 -2.46667817e-01 1.17096496e+00
-1.54654980e-01 -3.19665313e-01 6.97488904e-01 -5.09536006e-02
-3.16099018e-01 4.68550652e-01 -4.85669017e-01 2.33412296e-01
7.12437868e-01 1.57575518e-01 4.00191769e-02 -8.19518745e-01
-1.05230045e+00 -1.83896683e-02 2.10097387e-01 1.40864968e-01
8.52689862e-01 -1.30317414e+00 -4.31196362e-01 4.90592778e-01
8.10663402e-03 -4.18236315e-01 2.43263230e-01 8.38214695e-01
5.92501499e-02 8.42502058e-01 6.02367520e-02 -6.75656736e-01
-8.21341753e-01 6.49229467e-01 -6.17583804e-02 -2.44103760e-01
-9.98037696e-01 5.80260575e-01 2.34870553e-01 1.05605550e-01
2.55074233e-01 -1.73180580e-01 -1.37798026e-01 1.06362611e-01
6.45874023e-01 3.24001908e-01 -3.71384889e-01 -7.56049275e-01
-1.18211739e-01 2.45271027e-01 -1.67124733e-01 -4.65367764e-01
1.50306094e+00 -4.20118481e-01 5.18016741e-02 9.09578025e-01
1.79881513e+00 -2.90541202e-01 -1.53223550e+00 -6.84938669e-01
1.35980338e-01 -4.36073005e-01 3.25517654e-01 -1.13393389e-01
-8.87949705e-01 9.48503673e-01 3.39653790e-01 1.84139177e-01
1.04421723e+00 3.35869968e-01 4.67536956e-01 7.00890541e-01
4.58982825e-01 -1.01223040e+00 1.08757220e-01 5.57748616e-01
8.44566107e-01 -9.15704727e-01 2.71100532e-02 4.24035639e-02
-5.95382273e-01 1.06636679e+00 -3.14736247e-01 -5.89022562e-02
1.05257654e+00 2.28485644e-01 -8.03306997e-01 -1.65469170e-01
-1.17519343e+00 1.41823282e-02 4.18950796e-01 6.15896523e-01
1.15819998e-01 1.28537327e-01 1.07262373e-01 8.64369988e-01
-2.94801205e-01 -8.34333897e-02 5.23692608e-01 8.58998477e-01
-4.77279872e-01 -6.63652658e-01 -1.77015007e-01 7.56479263e-01
-5.67305982e-01 -1.50381744e-01 -4.28609960e-02 5.65081909e-02
-4.84959513e-01 8.08266878e-01 1.71615154e-01 -4.80739474e-01
-3.15922499e-01 3.02482426e-01 2.71790802e-01 -4.37758893e-01
5.14938235e-01 3.25694829e-01 -3.30910444e-01 -8.04750681e-01
-4.62796211e-01 -1.12120485e+00 -9.42214847e-01 -8.83722231e-02
-1.64408326e-01 2.45099619e-01 6.46511257e-01 1.10807574e+00
5.59627414e-01 3.78931344e-01 8.51450324e-01 -8.30479860e-01
-6.94869399e-01 -8.34074318e-01 -8.28956008e-01 3.56314480e-01
6.64786994e-01 -6.96493208e-01 -6.63946211e-01 4.04182255e-01]
|
[7.971439838409424, 3.838175058364868]
|
37193254-345a-4ce0-a7e5-646881968e2d
|
global-and-local-collaborative-learning-for
|
2204.08917
| null |
https://arxiv.org/abs/2204.08917v1
|
https://arxiv.org/pdf/2204.08917v1.pdf
|
Global-and-Local Collaborative Learning for Co-Salient Object Detection
|
The goal of co-salient object detection (CoSOD) is to discover salient objects that commonly appear in a query group containing two or more relevant images. Therefore, how to effectively extract inter-image correspondence is crucial for the CoSOD task. In this paper, we propose a global-and-local collaborative learning architecture, which includes a global correspondence modeling (GCM) and a local correspondence modeling (LCM) to capture comprehensive inter-image corresponding relationship among different images from the global and local perspectives. Firstly, we treat different images as different time slices and use 3D convolution to integrate all intra features intuitively, which can more fully extract the global group semantics. Secondly, we design a pairwise correlation transformation (PCT) to explore similarity correspondence between pairwise images and combine the multiple local pairwise correspondences to generate the local inter-image relationship. Thirdly, the inter-image relationships of the GCM and LCM are integrated through a global-and-local correspondence aggregation (GLA) module to explore more comprehensive inter-image collaboration cues. Finally, the intra- and inter-features are adaptively integrated by an intra-and-inter weighting fusion (AEWF) module to learn co-saliency features and predict the co-saliency map. The proposed GLNet is evaluated on three prevailing CoSOD benchmark datasets, demonstrating that our model trained on a small dataset (about 3k images) still outperforms eleven state-of-the-art competitors trained on some large datasets (about 8k-200k images).
|
['Sam Kwong', 'Qingming Huang', 'Yao Zhao', 'Huazhu Fu', 'Chongyi Li', 'Ning Yang', 'Runmin Cong']
|
2022-04-19
| null | null | null | null |
['co-saliency-detection']
|
['computer-vision']
|
[ 2.38647208e-01 -4.22164977e-01 -2.01351851e-01 -3.77318531e-01
-6.43142462e-01 -2.79845685e-01 7.36013472e-01 3.04510385e-01
-3.68581206e-01 2.45847613e-01 4.38529849e-01 3.58596265e-01
-4.35645878e-01 -4.95667219e-01 -5.23659348e-01 -7.60212064e-01
5.86377122e-02 -2.14777932e-01 7.57888615e-01 -3.26817393e-01
4.35010403e-01 4.30801034e-01 -2.02168679e+00 3.72804075e-01
9.32871759e-01 1.13676441e+00 7.90343642e-01 1.14362098e-01
-1.18936583e-01 6.62571430e-01 -2.94764638e-01 1.57132760e-01
2.28098541e-01 -3.32110405e-01 -6.91998720e-01 2.22130448e-01
6.27977908e-01 3.02804392e-02 5.52030839e-02 1.28825116e+00
5.74399292e-01 2.63077617e-01 3.54762465e-01 -1.30399144e+00
-6.80235803e-01 4.08957869e-01 -7.82598019e-01 6.85610235e-01
6.94157109e-02 4.88924608e-02 1.22120595e+00 -1.27503371e+00
6.72892928e-01 1.25649929e+00 7.10838288e-03 -1.91689238e-01
-1.09880686e+00 -6.67749822e-01 3.64105493e-01 6.96313679e-01
-1.35075688e+00 -1.50405779e-01 1.14136362e+00 -2.20833570e-01
7.07567394e-01 4.24963981e-01 8.11218083e-01 5.56297243e-01
2.08798587e-01 1.06300390e+00 1.17108333e+00 -2.74447888e-01
-2.37180442e-02 7.76913436e-03 7.57905021e-02 3.57756972e-01
3.74103375e-02 -1.20681720e-02 -9.17643964e-01 1.70663327e-01
6.65528119e-01 3.70384485e-01 -3.29121083e-01 -6.77449167e-01
-1.59927297e+00 6.82180464e-01 1.01500154e+00 5.52348435e-01
-4.87380534e-01 -2.78054595e-01 2.52308011e-01 1.31892025e-01
3.14285576e-01 5.67712426e-01 -3.34432095e-01 3.90906602e-01
-8.87092948e-01 1.95434809e-01 4.75829579e-02 9.84855592e-01
1.22601473e+00 -4.97487545e-01 -4.71245140e-01 9.46299493e-01
1.99668631e-01 3.75893652e-01 6.46794319e-01 -7.10338652e-01
4.61077094e-01 9.78142023e-01 3.55206728e-02 -1.56185579e+00
-3.54700476e-01 -6.82965279e-01 -8.02257121e-01 -8.15442726e-02
-2.55443063e-03 4.24095064e-01 -7.25360692e-01 1.59815037e+00
5.87082088e-01 3.12389791e-01 -1.32174492e-01 1.31863236e+00
1.07211649e+00 5.48786879e-01 9.20772776e-02 -2.29241088e-01
1.39933968e+00 -1.32442999e+00 -5.49934804e-01 -4.63153481e-01
3.99682105e-01 -1.20372772e+00 9.41875041e-01 -1.33781105e-01
-1.07544541e+00 -9.91125464e-01 -1.03437614e+00 -3.36135477e-01
-5.97095668e-01 2.88342774e-01 5.37081122e-01 -3.86890858e-01
-8.80819380e-01 8.58449116e-02 -3.23725015e-01 -3.34222376e-01
4.01917696e-01 3.28463763e-01 -3.55719656e-01 -2.74152100e-01
-1.21262383e+00 7.37982333e-01 6.03433192e-01 2.75370050e-02
-7.41456270e-01 -6.92437351e-01 -9.16288078e-01 2.77384967e-02
4.63067323e-01 -5.68712950e-01 6.87098980e-01 -1.02616930e+00
-7.73648679e-01 8.07716370e-01 -3.41399878e-01 -8.87730494e-02
-6.14181533e-02 -9.60084349e-02 -5.56161821e-01 2.93836743e-01
5.06744087e-01 9.97903883e-01 9.67059851e-01 -1.33102643e+00
-1.09900057e+00 -3.46064895e-01 -4.00804430e-02 8.93485069e-01
-2.69480914e-01 2.26036400e-01 -7.17325151e-01 -7.61661708e-01
5.70140183e-01 -6.38633490e-01 -8.72841850e-02 -1.25950083e-01
-3.06214541e-01 -2.51243472e-01 1.07379222e+00 -4.70146596e-01
1.16301775e+00 -2.29921627e+00 3.79688591e-01 1.03482842e-01
3.43232661e-01 1.68337002e-01 -3.32780331e-01 1.97615221e-01
-2.03135431e-01 -4.62418318e-01 2.11116765e-02 -3.11623186e-01
-3.20290923e-01 -7.73487315e-02 -8.71405900e-02 1.29232869e-01
3.33000064e-01 1.24214566e+00 -1.11867177e+00 -7.09554136e-01
4.90995198e-01 2.23648027e-01 -1.25352874e-01 1.50557935e-01
1.62165925e-01 3.54499787e-01 -5.97156227e-01 6.69417202e-01
8.52259517e-01 -5.37937462e-01 -2.12421760e-01 -6.32175922e-01
-2.09254608e-01 -2.75341719e-02 -1.17018199e+00 1.70463991e+00
-2.14057818e-01 3.36427301e-01 -2.18724042e-01 -9.69725370e-01
1.14218628e+00 -3.46928269e-01 5.25933325e-01 -9.70704198e-01
3.51066478e-02 3.59282702e-01 -6.22543879e-03 -2.96760798e-01
5.01003504e-01 3.83653164e-01 9.17720571e-02 3.39076072e-01
2.99535096e-01 2.40486622e-01 8.27469975e-02 3.21257532e-01
5.65372467e-01 -1.88369080e-01 4.57507938e-01 -6.50383234e-01
1.07183719e+00 -1.63577691e-01 6.25564396e-01 4.34015840e-01
-2.79538453e-01 8.32833767e-01 1.21535853e-01 -5.19943297e-01
-7.39751101e-01 -1.03307545e+00 -1.29532786e-02 1.09364760e+00
1.29270244e+00 -3.47073466e-01 -4.08447921e-01 -6.41485751e-01
-2.31387187e-02 3.37513924e-01 -7.47231185e-01 -3.39657336e-01
-3.77030522e-01 -6.87821925e-01 -2.65605986e-01 5.38588703e-01
8.42107058e-01 -1.31295824e+00 -5.92026830e-01 2.24464238e-01
-3.48716408e-01 -1.05372930e+00 -1.05896032e+00 5.45819066e-02
-5.15359342e-01 -9.00452435e-01 -7.27596819e-01 -1.20282698e+00
6.88527763e-01 1.05612826e+00 9.68718708e-01 1.07802548e-01
-3.33374828e-01 1.93028972e-02 -4.08428162e-01 -2.03173250e-01
3.48763674e-01 -1.45477444e-01 -1.52356863e-01 4.63765234e-01
6.09067321e-01 -5.73551059e-01 -8.37790489e-01 7.32200027e-01
-7.89521039e-01 5.60999513e-01 7.56289899e-01 9.45147634e-01
9.88721609e-01 -2.00026575e-02 5.09316802e-01 -2.23392829e-01
2.76289642e-01 -3.36319625e-01 -5.23584664e-01 4.56263214e-01
-4.74182338e-01 -9.61851627e-02 3.05200726e-01 -4.01729643e-01
-9.86550152e-01 -5.39907813e-02 5.64632118e-01 -6.79360747e-01
5.84111437e-02 4.78973985e-01 -4.92374003e-01 -1.23434171e-01
2.65560985e-01 6.46381140e-01 -2.12224990e-01 -3.68362814e-01
3.08570355e-01 3.93458188e-01 6.88747644e-01 -3.01592201e-01
8.07222188e-01 4.49145555e-01 -2.38364432e-02 -4.73552257e-01
-1.15638375e+00 -1.00239563e+00 -9.12431180e-01 -1.71573192e-01
9.46878731e-01 -1.26950574e+00 -1.24862395e-01 6.82437897e-01
-9.14393067e-01 1.28805444e-01 -2.62469530e-01 4.62815791e-01
-2.13850930e-01 2.94726610e-01 -1.73816770e-01 -1.83544844e-01
-3.90674382e-01 -1.36229694e+00 1.44413877e+00 6.64853096e-01
1.88675374e-01 -7.66854823e-01 -3.31587523e-01 3.73719066e-01
3.14273566e-01 -1.78278852e-02 5.79917967e-01 -5.21135390e-01
-8.58014226e-01 1.99409440e-01 -8.35959435e-01 2.69434929e-01
5.31936169e-01 -2.54090786e-01 -7.04217553e-01 -3.44400138e-01
-1.88706249e-01 -5.34828939e-02 8.88108730e-01 4.88650024e-01
1.13926578e+00 -6.27701208e-02 -4.50451672e-01 4.47796583e-01
1.30167150e+00 -2.02964898e-02 3.99705470e-01 4.22794729e-01
9.50350881e-01 7.12269127e-01 1.24707413e+00 2.32376948e-01
5.96477509e-01 8.85190547e-01 3.49565923e-01 -3.46477062e-01
-3.34648401e-01 -3.07384670e-01 1.80095509e-01 9.00000095e-01
3.26570794e-02 1.88462362e-01 -4.85584557e-01 8.06242466e-01
-2.21041846e+00 -8.43371391e-01 -1.38437092e-01 2.11746931e+00
7.06288218e-01 2.97213010e-02 7.40396008e-02 -1.93483815e-01
1.02941883e+00 3.98224264e-01 -4.58557338e-01 2.01526135e-01
-6.20786726e-01 -1.13199927e-01 2.78541148e-01 2.42557809e-01
-1.36774898e+00 9.55021918e-01 4.75682116e+00 1.35448599e+00
-1.14001131e+00 1.25684410e-01 7.64745712e-01 2.38651671e-02
-3.86742949e-01 1.63419351e-01 -7.47375548e-01 6.33545697e-01
-1.88005846e-02 -2.33055487e-01 2.73277253e-01 7.56452739e-01
8.32175240e-02 -3.73231977e-01 -6.71209335e-01 1.20925236e+00
2.87697107e-01 -1.39188588e+00 1.68875843e-01 2.33539995e-02
1.12160802e+00 9.35778469e-02 1.55793265e-01 -4.97374609e-02
-1.47804499e-01 -7.67338574e-01 7.31721520e-01 5.56483805e-01
3.89156103e-01 -7.44689047e-01 7.83302903e-01 2.66542256e-01
-1.67068136e+00 -1.36179090e-01 -4.54270363e-01 3.06465447e-01
-2.85407212e-02 7.65740752e-01 -3.99739951e-01 9.48450029e-01
1.15720415e+00 1.28400850e+00 -1.06796718e+00 1.10119498e+00
-9.02017280e-02 -1.37318689e-02 -3.10845464e-01 3.87269929e-02
4.93589103e-01 -2.22489357e-01 6.29201531e-01 8.97335351e-01
1.96609214e-01 -7.69282579e-02 2.41746053e-01 8.75699818e-01
1.77346632e-01 2.94507951e-01 -6.84398115e-02 3.93203735e-01
5.27963400e-01 1.66884863e+00 -7.93586016e-01 -3.73271257e-01
-3.78733486e-01 1.01048589e+00 1.57351330e-01 3.63175690e-01
-6.81759417e-01 -2.37890422e-01 6.06925666e-01 -2.16493428e-01
5.29849768e-01 -9.64096282e-03 -2.33311310e-01 -1.18552601e+00
2.41488174e-01 -6.93020761e-01 4.99149024e-01 -1.07536662e+00
-1.25420582e+00 3.92735600e-01 8.14879239e-02 -1.60082424e+00
2.67058164e-01 -7.66303614e-02 -8.08924139e-01 1.11281574e+00
-1.81264210e+00 -1.44303799e+00 -6.79681420e-01 7.80591547e-01
6.35425210e-01 -1.47710204e-01 2.53842652e-01 1.09001532e-01
-2.91130126e-01 4.27544862e-01 -1.48687422e-01 -1.84269443e-01
7.57739544e-01 -1.09759963e+00 1.38516188e-01 9.81775522e-01
3.70208830e-01 5.55657566e-01 3.13700646e-01 -5.80695987e-01
-9.85742509e-01 -1.22024477e+00 1.03044999e+00 -2.68723369e-01
4.56328750e-01 -1.49601623e-01 -1.07300842e+00 2.62160152e-01
1.26015712e-02 5.22648811e-01 2.27242589e-01 -1.18222311e-01
-2.29009315e-01 -3.37590605e-01 -8.64745498e-01 5.56392193e-01
1.17596591e+00 -6.70496047e-01 -6.39699399e-01 9.20957178e-02
1.00159597e+00 -1.61482796e-01 -7.88641214e-01 6.73870027e-01
2.74498820e-01 -1.09231877e+00 1.16290724e+00 -1.62141368e-01
4.70144361e-01 -1.07015216e+00 -3.31503183e-01 -1.13169062e+00
-6.66938484e-01 -3.63540798e-01 7.47234523e-02 1.37525260e+00
8.28914121e-02 -4.96971250e-01 2.63764262e-01 -3.80431712e-02
-1.88075140e-01 -9.57078278e-01 -7.77450442e-01 -4.75753486e-01
-4.86465812e-01 5.11175627e-03 8.69913578e-01 1.01776099e+00
-1.56609520e-01 3.26779634e-01 -2.60700524e-01 3.21901947e-01
5.22799909e-01 7.43991852e-01 5.15195966e-01 -1.04336655e+00
-8.19461793e-02 -5.15629888e-01 -7.01785266e-01 -1.09481966e+00
-1.34136200e-01 -8.55887115e-01 -1.28494218e-01 -1.14675939e+00
6.96937621e-01 -3.59611183e-01 -9.33773339e-01 3.70734155e-01
-7.47775495e-01 3.58804911e-01 3.24946225e-01 5.37758350e-01
-1.06581414e+00 8.91267598e-01 1.61940444e+00 -2.32930213e-01
-1.84364468e-01 -4.01162177e-01 -8.27805340e-01 4.60375577e-01
4.65467900e-01 -1.76964760e-01 -5.02053261e-01 -1.34826332e-01
-1.52131185e-01 -3.73030365e-01 7.33979523e-01 -1.09814131e+00
5.24086595e-01 -3.36240917e-01 6.01249397e-01 -9.58034635e-01
2.19734773e-01 -6.32128298e-01 8.08712747e-03 -2.85551674e-03
-3.42310011e-01 -1.29881561e-01 -8.16093534e-02 5.06892800e-01
-7.50155926e-01 2.35404938e-01 7.84585118e-01 8.06433856e-02
-1.33282292e+00 4.89616215e-01 4.28636104e-01 -5.42740636e-02
1.17271352e+00 -1.97625801e-01 -3.55229020e-01 -1.26243517e-01
-4.81944114e-01 5.20042419e-01 5.38644433e-01 9.09172773e-01
8.31506133e-01 -1.66373765e+00 -5.29768229e-01 4.72980946e-01
6.67197287e-01 1.62622035e-01 5.91935933e-01 1.13976705e+00
-3.62510653e-03 2.52435148e-01 -3.91879380e-01 -9.52762306e-01
-1.38675618e+00 6.10325515e-01 2.04224616e-01 -3.23470592e-01
-4.62279677e-01 1.08289504e+00 8.31913173e-01 -3.35478932e-01
-2.00969398e-01 -1.62072852e-01 -5.25332808e-01 2.28599101e-01
6.59189403e-01 3.63885723e-02 -8.45252478e-04 -1.12141228e+00
-5.07031262e-01 9.81918156e-01 -2.00918049e-01 3.18289399e-01
1.20139384e+00 -5.60026765e-01 -3.16636622e-01 2.64110357e-01
1.42727828e+00 -3.18910480e-01 -1.30060267e+00 -8.30964506e-01
-4.45006080e-02 -7.68378496e-01 2.13843629e-01 -5.87541819e-01
-1.19018543e+00 7.76210189e-01 7.20661342e-01 -2.24268079e-01
1.50214303e+00 3.71470928e-01 7.06033766e-01 -7.02106357e-02
3.57754290e-01 -1.14541030e+00 4.55860823e-01 2.50210047e-01
1.10469615e+00 -1.38006794e+00 1.64226621e-01 -6.38991117e-01
-8.51913273e-01 8.49720359e-01 7.60223567e-01 -1.44129872e-01
7.60257781e-01 -3.28277975e-01 -1.38999104e-01 -4.01880503e-01
-5.32943785e-01 -5.19930065e-01 9.77793038e-01 3.15942287e-01
-5.50044402e-02 1.03954978e-01 -1.92350417e-01 5.29327393e-01
-2.22909842e-02 -2.31677040e-01 2.60913689e-02 6.86839044e-01
-4.96375203e-01 -9.24622655e-01 -2.48927176e-01 4.12237138e-01
8.74125212e-02 -2.91417032e-01 -2.43877843e-01 5.06872773e-01
5.26435256e-01 9.48275685e-01 2.26754457e-01 -6.23412311e-01
-4.03700918e-02 -4.24712092e-01 1.27777904e-01 -5.11030078e-01
-3.95983666e-01 4.54523265e-01 -3.98699701e-01 -7.02703178e-01
-8.45922232e-01 -8.69804382e-01 -1.11481547e+00 1.95347860e-01
-4.10642803e-01 -5.00734411e-02 2.36233771e-01 9.09890294e-01
5.82241476e-01 5.25930464e-01 9.51394379e-01 -1.12703276e+00
-4.23952267e-02 -7.94636011e-01 -7.81997859e-01 7.13242352e-01
3.56520355e-01 -9.45851982e-01 -4.03145969e-01 -6.49627820e-02]
|
[9.803727149963379, -0.31251269578933716]
|
012887cd-80d9-4c7f-b8b6-96b4cb6480ef
|
ecnu-at-semeval-2017-task-1-leverage-kernel
| null | null |
https://aclanthology.org/S17-2028
|
https://aclanthology.org/S17-2028.pdf
|
ECNU at SemEval-2017 Task 1: Leverage Kernel-based Traditional NLP features and Neural Networks to Build a Universal Model for Multilingual and Cross-lingual Semantic Textual Similarity
|
To address semantic similarity on multilingual and cross-lingual sentences, we firstly translate other foreign languages into English, and then feed our monolingual English system with various interactive features. Our system is further supported by combining with deep learning semantic similarity and our best run achieves the mean Pearson correlation 73.16{\%} in primary track.
|
['Zhiheng Zhou', 'Man Lan', 'Yuanbin Wu', 'Junfeng Tian']
|
2017-08-01
| null | null | null |
semeval-2017-8
|
['cross-lingual-semantic-textual-similarity']
|
['natural-language-processing']
|
[-7.61632204e-01 -2.84935057e-01 -3.73350471e-01 -5.83076358e-01
-9.79849696e-01 -8.69880021e-01 6.97509527e-01 2.37020224e-01
-9.10062671e-01 1.07230616e+00 4.88079339e-01 -5.35425007e-01
-2.15268484e-03 -6.64889514e-01 -4.88520712e-01 1.09876044e-01
6.91861510e-02 6.32253289e-01 3.12056988e-02 -7.41129935e-01
2.01413244e-01 9.92200598e-02 -9.43851888e-01 4.27535802e-01
1.14039695e+00 6.52457893e-01 5.07335663e-01 3.33092064e-01
-6.02474988e-01 6.15692139e-01 -5.62129617e-01 -9.70345199e-01
1.96655929e-01 -3.51338536e-01 -1.28762150e+00 -6.89053595e-01
4.83580947e-01 2.32799754e-01 -2.43903458e-01 1.44414997e+00
7.92096496e-01 2.60042608e-01 3.64565998e-01 -8.95326436e-01
-1.30562782e+00 1.14520431e+00 -4.81674708e-02 6.06404662e-01
1.00399840e+00 -2.68798023e-01 1.20005989e+00 -1.34272575e+00
9.53412533e-01 1.27049792e+00 9.56727386e-01 2.00160101e-01
-6.27718627e-01 -8.02219331e-01 -1.46042719e-01 5.98233044e-01
-1.71446168e+00 -6.01303399e-01 2.11414263e-01 7.43021145e-02
1.72550356e+00 3.93565856e-02 1.95625380e-01 1.11832690e+00
1.99623078e-01 6.09361827e-01 1.29928541e+00 -5.14821172e-01
-3.44428957e-01 4.78130788e-01 1.75780013e-01 6.51044965e-01
-3.10164653e-02 -2.12465167e-01 -4.43634510e-01 1.06124237e-01
4.35277037e-02 -2.34713525e-01 3.99440229e-02 5.35450876e-01
-1.43664467e+00 7.17016757e-01 2.52565175e-01 8.03713322e-01
-4.25761100e-03 -4.20679390e-01 8.23458314e-01 9.84466672e-01
3.33990425e-01 7.96927631e-01 -7.71682739e-01 -1.27108097e-01
-7.15242863e-01 1.38268128e-01 8.87618005e-01 1.38314474e+00
9.05748844e-01 -3.83926660e-01 1.99056968e-01 1.36078453e+00
1.12154759e-01 8.72827888e-01 1.04319835e+00 -9.18061376e-01
4.93062586e-01 2.80571073e-01 -3.73025596e-01 -7.81325340e-01
-4.73888606e-01 -3.72240394e-01 -5.46109140e-01 -8.11272562e-01
9.12374184e-02 -1.26938298e-01 -1.77573428e-01 1.48321247e+00
-1.11214951e-01 -3.35662901e-01 7.54893482e-01 6.05983973e-01
1.30318701e+00 5.26759148e-01 2.99682617e-01 -1.03788093e-01
1.31207931e+00 -1.09797120e+00 -8.55192423e-01 -2.72764057e-01
1.03159153e+00 -1.38503361e+00 1.30967677e+00 -1.14398845e-01
-1.04262197e+00 -7.84308076e-01 -7.92277932e-01 -4.88005489e-01
-8.03640068e-01 -3.16655152e-02 4.92615283e-01 2.67385811e-01
-1.46683097e+00 6.17154479e-01 -3.59311521e-01 -9.37690735e-01
-1.73154861e-01 1.21991999e-01 -7.47937024e-01 -1.91843823e-01
-2.10880876e+00 1.39767277e+00 6.70758843e-01 -4.18370515e-01
-1.46119997e-01 -6.89879179e-01 -1.08867180e+00 -3.54926586e-01
-1.41616389e-01 -6.23647153e-01 9.51676309e-01 -6.33496583e-01
-1.26171470e+00 1.43948591e+00 -2.04761937e-01 -3.65762830e-01
3.32629919e-01 -2.45992064e-01 -1.10751843e+00 -1.53742269e-01
8.60967994e-01 5.67451775e-01 -1.79858729e-01 -7.24557281e-01
-7.64336824e-01 -2.37784281e-01 -3.66247706e-02 6.29293501e-01
-4.94748592e-01 7.94815183e-01 -6.34790957e-01 -6.29704297e-01
-4.82106619e-02 -8.19585860e-01 -1.98447630e-02 -6.63087130e-01
-2.12813094e-01 -5.54317713e-01 3.00659776e-01 -1.24905288e+00
1.17315090e+00 -2.06088376e+00 -1.60795644e-01 7.39767104e-02
-3.58742744e-01 1.39499009e-01 -3.11759472e-01 6.07951641e-01
2.68389396e-02 5.85896783e-02 1.28587916e-01 -2.62782395e-01
3.88742983e-02 2.98257023e-01 2.75718067e-02 2.21704513e-01
-2.20985174e-01 1.01044321e+00 -1.27366138e+00 -8.66718054e-01
1.03192896e-01 1.33476660e-01 -4.36996669e-01 4.16459776e-02
3.76632363e-01 1.52072713e-01 -9.62625071e-02 6.79392576e-01
5.12134254e-01 8.19263831e-02 4.44312900e-01 -1.08361207e-01
-1.44855246e-01 8.39639604e-01 -9.10962939e-01 2.42742491e+00
-9.29465294e-01 6.37858689e-01 -2.31288284e-01 -9.35200036e-01
1.18867111e+00 1.45320952e-01 2.44103998e-01 -1.15987694e+00
3.98848914e-02 7.63189018e-01 -9.61317495e-02 -4.30528581e-01
8.55887949e-01 -8.44685659e-02 -4.52536076e-01 3.38670880e-01
4.17657971e-01 -2.85892546e-01 3.21101844e-01 2.05003634e-01
7.77575135e-01 -2.11933851e-02 5.34791112e-01 -9.65328991e-01
9.06342506e-01 1.28434345e-01 2.66385704e-01 5.39916277e-01
-4.14414406e-01 1.43163487e-01 -9.27128941e-02 -4.02052313e-01
-9.16957855e-01 -1.41090071e+00 -4.16148663e-01 1.36527562e+00
2.14028269e-01 -6.25385523e-01 -6.10786796e-01 -8.35435271e-01
-5.49911596e-02 7.88502038e-01 -1.94221839e-01 -1.46747068e-01
-3.22152704e-01 -4.22806919e-01 8.34500492e-01 5.54624081e-01
4.87149417e-01 -1.14423287e+00 6.49386287e-01 2.08185270e-01
-5.92502892e-01 -1.44781554e+00 -5.60323477e-01 7.60783628e-02
-4.02943760e-01 -8.97352397e-01 -3.58701378e-01 -1.38887274e+00
3.37484898e-03 7.71539360e-02 1.50049007e+00 -7.60268047e-02
3.43713686e-02 2.60545593e-02 -4.97698635e-01 -5.73183829e-03
-5.51166296e-01 4.57960665e-01 6.72673464e-01 -8.32468033e-01
1.12544036e+00 -3.91487241e-01 -1.84696585e-01 9.81781259e-02
-1.54808089e-01 -3.64486098e-01 2.72089779e-01 7.04114318e-01
6.14081383e-01 -2.24589854e-01 6.44702017e-01 -6.03582084e-01
8.44257653e-01 -7.70065546e-01 -3.58567499e-02 5.57823062e-01
-5.61129212e-01 7.32759088e-02 8.33748400e-01 4.00857106e-02
-9.25886750e-01 -5.73364019e-01 -5.36483645e-01 -2.50062775e-02
-1.77267432e-01 5.43421090e-01 -6.29454628e-02 1.99636117e-01
7.02679574e-01 2.59020627e-01 -3.09522361e-01 -5.15798092e-01
7.07539320e-01 1.16259873e+00 9.16612267e-01 -6.71585619e-01
4.27332550e-01 5.51018864e-02 -6.54175699e-01 -6.42797112e-01
-9.55355644e-01 -4.57883358e-01 -8.16451073e-01 2.87387043e-01
8.40490043e-01 -1.31400621e+00 -3.88486922e-01 2.52943695e-01
-1.26913691e+00 2.41992418e-02 -3.92377377e-02 8.61102879e-01
-3.48056853e-01 2.28223503e-01 -8.93207788e-01 1.68768808e-01
-6.81270838e-01 -8.55550468e-01 8.41413796e-01 -6.96710423e-02
-7.14937270e-01 -1.48168612e+00 3.50576341e-01 4.14070636e-01
4.45359051e-01 -6.91052318e-01 7.19440937e-01 -1.32983720e+00
4.00945127e-01 -1.16777748e-01 -4.22697663e-01 5.43556035e-01
4.25018638e-01 -1.24564767e-01 -4.16393012e-01 -3.12955409e-01
-3.28381181e-01 -4.36490655e-01 3.13929409e-01 2.69742105e-02
8.59391630e-01 1.30151346e-01 -1.80467665e-01 8.10399592e-01
1.30962288e+00 1.35076448e-01 1.12086669e-01 7.46227622e-01
5.83690405e-01 6.08311594e-01 6.17425680e-01 7.09294975e-02
1.05457151e+00 5.26218295e-01 -5.73033333e-01 -7.94404224e-02
-2.66804278e-01 -3.07305634e-01 5.77551663e-01 1.71857285e+00
1.41314149e-01 2.32341111e-01 -1.23504508e+00 5.25257885e-01
-1.56410837e+00 -1.04972458e+00 -1.71423107e-01 1.67537761e+00
1.12772238e+00 -2.99955569e-02 -2.81757176e-01 -4.42859441e-01
8.03711057e-01 -9.02706683e-02 -1.47634044e-01 -8.20255935e-01
-7.05488801e-01 5.73550522e-01 5.81667364e-01 8.21793497e-01
-9.99519348e-01 1.67370868e+00 7.54786205e+00 9.47464645e-01
-1.03482401e+00 3.80639315e-01 2.01834753e-01 1.45176724e-01
-4.85883981e-01 -5.04433997e-02 -9.73387361e-01 4.63485867e-01
1.06475675e+00 -9.35511112e-01 4.49456424e-01 8.58878076e-01
-4.26929258e-02 1.14190340e-01 -1.05930376e+00 9.57856297e-01
3.07540655e-01 -1.15397751e+00 7.99254999e-02 -4.85226661e-01
9.80009794e-01 5.93624413e-01 -2.46035650e-01 8.23518038e-01
7.43999481e-01 -8.49319160e-01 5.14415085e-01 4.11790907e-01
8.85777295e-01 -1.00845313e+00 1.04534161e+00 1.27991736e-02
-1.38407922e+00 4.04691219e-01 -4.79131967e-01 -2.94068106e-03
3.79326530e-02 2.72546500e-01 -6.53188884e-01 8.82453680e-01
1.11288953e+00 1.22495329e+00 -7.31686294e-01 3.18860531e-01
-2.92640805e-01 -6.09713700e-03 -1.23254433e-01 -1.08823679e-01
2.93889344e-01 -2.32106701e-01 3.01295131e-01 1.66302800e+00
4.77750719e-01 -4.49456513e-01 3.74567270e-01 3.15035909e-01
-2.42816567e-01 9.46977913e-01 -8.18634689e-01 1.62613139e-01
9.02794957e-01 8.90952110e-01 -3.87497932e-01 -6.95868552e-01
-7.66962826e-01 1.34667766e+00 6.86709464e-01 1.31928369e-01
-6.60252273e-01 -8.27089489e-01 9.05831337e-01 -4.79964346e-01
-4.30558681e-01 -1.29622951e-01 -1.19624250e-01 -1.35506260e+00
4.29782756e-02 -9.77086306e-01 6.05692208e-01 -7.50675619e-01
-1.85949922e+00 9.15341198e-01 -3.48220557e-01 -8.67042124e-01
-4.06851292e-01 -6.16401017e-01 -3.32974762e-01 1.26823866e+00
-1.40265083e+00 -9.07811642e-01 1.83615386e-01 1.14913988e+00
6.13674939e-01 -7.64460266e-01 1.11538506e+00 7.48991311e-01
-2.19063133e-01 1.02138042e+00 3.18342119e-01 3.07638496e-01
1.32024705e+00 -1.22761667e+00 7.46351182e-01 6.62509024e-01
2.04741284e-01 8.72438610e-01 5.14057934e-01 -5.54137588e-01
-1.15349638e+00 -1.08244002e+00 1.91963851e+00 -5.72953999e-01
1.34643054e+00 -3.36701833e-02 -7.69465864e-01 6.78695679e-01
8.12735319e-01 -1.51317671e-01 8.88454735e-01 5.65547824e-01
-3.32180649e-01 -5.48812076e-02 -9.62008893e-01 7.64140010e-01
1.47693741e+00 -1.26973546e+00 -1.11618745e+00 5.08707166e-01
1.02977169e+00 -2.00430930e-01 -1.25612724e+00 4.22534525e-01
1.13280930e-01 -5.50976753e-01 1.08462071e+00 -9.24737334e-01
3.31520349e-01 -1.29245386e-01 -6.62718952e-01 -1.67542982e+00
-3.79366845e-01 -2.96044439e-01 6.57098353e-01 1.19863832e+00
7.45507896e-01 -6.54320955e-01 1.56760126e-01 -8.14446900e-03
-4.52943921e-01 -3.30265522e-01 -8.03533614e-01 -1.18308926e+00
5.46888411e-01 -5.83436608e-01 5.93924105e-01 1.70859957e+00
4.87836063e-01 3.68150860e-01 -4.29411270e-02 -4.90455478e-02
2.80239075e-01 -1.08719859e-02 3.01407546e-01 -7.32256770e-01
-1.81693211e-02 -6.41465962e-01 -4.09699023e-01 -5.23795784e-01
9.78661656e-01 -1.83300960e+00 -2.67470658e-01 -1.44043636e+00
3.10362518e-01 -4.93996471e-01 -5.00974298e-01 2.95808464e-01
-3.69599491e-01 2.13212192e-01 -1.18776917e-01 8.06898922e-02
-8.19854200e-01 4.52406257e-01 9.38758850e-01 -2.75815520e-02
1.00833632e-01 -5.24936855e-01 -7.82948077e-01 6.15468264e-01
7.47909486e-01 -3.43402207e-01 -2.88683679e-02 -8.76727283e-01
3.70348305e-01 -2.37920702e-01 -2.88381457e-01 -8.39852929e-01
1.15947522e-01 -1.36241198e-01 2.12028950e-01 -3.57991368e-01
-3.63430709e-01 -5.39321959e-01 -6.47617877e-02 3.26024979e-01
-4.64328438e-01 9.58392859e-01 1.29791051e-01 -2.15255573e-01
-6.80459261e-01 -1.60211861e-01 6.26799345e-01 -2.99217075e-01
-1.14817357e+00 1.18631348e-01 -2.79355139e-01 6.41790867e-01
8.07637751e-01 2.79902369e-01 -3.52804482e-01 -1.52122071e-02
-9.66902792e-01 4.29987103e-01 5.30834079e-01 1.19075215e+00
1.16074614e-01 -1.82343054e+00 -8.99343014e-01 7.19138384e-02
4.42270517e-01 -8.14742386e-01 -8.55712891e-02 7.08133399e-01
-6.94327772e-01 5.99675119e-01 -3.46483082e-01 -2.40108758e-01
-1.09622836e+00 4.69817966e-01 4.92968142e-01 -6.96467236e-02
-3.04752231e-01 9.06391025e-01 -5.27629733e-01 -1.24767780e+00
-3.02137379e-02 2.10284725e-01 -2.14541540e-01 1.01828068e-01
4.66803312e-01 3.92200291e-01 3.58905286e-01 -1.05385983e+00
-8.89446676e-01 6.37010872e-01 1.24793574e-02 -2.28178695e-01
9.34844553e-01 -5.04470706e-01 -6.38370216e-01 4.86251444e-01
1.84937787e+00 3.50011885e-01 1.35442376e-01 -7.51960337e-01
4.21401948e-01 -2.08674997e-01 -3.13995719e-01 -7.39854991e-01
-7.98188627e-01 5.22622764e-01 4.86367047e-01 -2.55836904e-01
6.78488374e-01 4.30109978e-01 9.04391825e-01 9.41620946e-01
5.75913668e-01 -1.69282055e+00 -4.20650333e-01 1.28524077e+00
6.47293270e-01 -1.57318568e+00 -1.38982609e-01 -1.09732978e-01
-6.16479814e-01 9.26673472e-01 6.51128590e-01 -1.05058752e-01
8.48692536e-01 2.76050925e-01 6.94379985e-01 -5.70020825e-02
-6.31903291e-01 -3.33203167e-01 3.14716369e-01 5.28915763e-01
1.13067913e+00 2.53760040e-01 -8.90750647e-01 7.99355865e-01
-9.53936994e-01 -3.42016846e-01 -1.41151890e-01 5.41036248e-01
-2.82606453e-01 -1.43757606e+00 -1.47863273e-02 1.97724134e-01
-6.76625609e-01 -8.27390730e-01 -2.62654990e-01 7.37714291e-01
3.37666363e-01 9.60536659e-01 3.59576821e-01 -4.74010319e-01
4.70638812e-01 2.41842628e-01 3.37550521e-01 -6.58409357e-01
-8.28988731e-01 -4.17954981e-01 4.35046881e-01 -7.09523201e-01
-2.86297321e-01 -8.74228895e-01 -1.41365635e+00 -6.29075885e-01
2.69717306e-01 4.44356978e-01 7.09435105e-01 1.26544797e+00
1.58374757e-01 6.32559061e-02 7.36958504e-01 -9.52714905e-02
-2.16540515e-01 -1.11326039e+00 -3.01261514e-01 7.80316114e-01
-2.91365772e-01 -2.38681227e-01 -4.34158117e-01 -1.58307195e-01]
|
[10.952589988708496, 9.779767036437988]
|
c0eb036f-fd18-4c72-a4c4-70216669e654
|
analysis-of-temporal-expressions-annotated-in
| null | null |
https://aclanthology.org/W15-0211
|
https://aclanthology.org/W15-0211.pdf
|
Analysis of Temporal Expressions Annotated in Clinical Notes
| null |
['Marcus Didonet Del Fabro', 'Genevieve Gorrell', 'Angus Roberts', 'Leon Derczynski', 'Hegler Tissot']
|
2015-04-01
| null | null | null |
ws-2015-4
|
['temporal-information-extraction']
|
['natural-language-processing']
|
[-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01]
|
[-7.335432529449463, 3.614875078201294]
|
d9715cb0-3440-40c1-9aa2-db2a04136527
|
a-note-on-monotone-submodular-maximization
|
2006.09327
| null |
https://arxiv.org/abs/2006.09327v5
|
https://arxiv.org/pdf/2006.09327v5.pdf
|
Submodular Maximization in Clean Linear Time
|
In this paper, we provide the first deterministic algorithm that achieves the tight $1-1/e$ approximation guarantee for submodular maximization under a cardinality (size) constraint while making a number of queries that scales only linearly with the size of the ground set $n$. To complement our result, we also show strong information-theoretic lower bounds. More specifically, we show that when the maximum cardinality allowed for a solution is constant, no algorithm making a sub-linear number of function evaluations can guarantee any constant approximation ratio. Furthermore, when the constraint allows the selection of a constant fraction of the ground set, we show that any algorithm making fewer than $\Omega(n/\log(n))$ function evaluations cannot perform better than an algorithm that simply outputs a uniformly random subset of the ground set of the right size. We then provide a variant of our deterministic algorithm for the more general knapsack constraint, which is the first linear-time algorithm that achieves $1/2$-approximation guarantee for this constraint. Finally, we extend our results to the general case of maximizing a monotone submodular function subject to the intersection of a $p$-set system and multiple knapsack constraints. We extensively evaluate the performance of our algorithms on multiple real-life machine learning applications, including movie recommendation, location summarization, twitter text summarization and video summarization.
|
['Amin Karbasi', 'Ehsan Kazemi', 'Moran Feldman', 'Wenxin Li']
|
2020-06-16
| null | null | null | null |
['movie-recommendation']
|
['miscellaneous']
|
[ 2.17060894e-01 4.11047369e-01 -5.10716558e-01 -1.03336066e-01
-1.05915570e+00 -1.04842687e+00 -5.21794379e-01 6.50589466e-01
-6.08736217e-01 8.18287492e-01 -8.52477998e-02 -3.36617559e-01
-5.68451345e-01 -1.31370199e+00 -1.14172113e+00 -6.81602538e-01
-4.34080988e-01 9.49885249e-01 2.46045396e-01 -2.37757504e-01
4.31344867e-01 2.86610246e-01 -1.28265500e+00 -8.27427953e-02
7.99356997e-01 1.22464907e+00 1.68767482e-01 8.38659763e-01
-3.65474559e-02 1.24554485e-01 -6.91429019e-01 -2.40104482e-01
7.38465667e-01 -4.07831907e-01 -8.68705690e-01 2.07284868e-01
6.26127064e-01 -3.97230536e-01 -4.93199080e-01 1.24021196e+00
2.73831069e-01 1.66796505e-01 2.26510599e-01 -1.42763615e+00
-3.31738442e-01 1.22958732e+00 -7.53991485e-01 2.75377780e-01
5.76531351e-01 -3.20309639e-01 1.49502873e+00 -1.04421049e-01
6.22591376e-01 8.87319863e-01 1.68620169e-01 2.07968205e-01
-1.06207609e+00 -7.41437137e-01 3.09514016e-01 -7.03953728e-02
-1.46153235e+00 -1.71213299e-01 4.02019918e-01 1.79071635e-01
9.81597006e-01 9.07475054e-01 4.20865625e-01 -4.70409840e-01
-2.38009959e-01 7.94476688e-01 5.13213992e-01 -4.08098072e-01
2.67343760e-01 6.69790059e-02 3.84544522e-01 7.61673510e-01
9.72811341e-01 -6.62160397e-01 -5.71979225e-01 -6.23598158e-01
1.21341363e-01 -2.72809342e-02 -6.39267325e-01 -1.64835021e-01
-7.92602301e-01 8.48866284e-01 1.37915075e-01 7.44540393e-02
-1.73478395e-01 6.33720756e-01 -7.63121918e-02 6.04170203e-01
2.70838499e-01 5.36666334e-01 -5.49529850e-01 4.38945740e-02
-1.29389691e+00 5.31562686e-01 1.34879756e+00 1.37551773e+00
7.16571808e-01 -2.84424812e-01 -1.00015283e-01 3.19826096e-01
-1.27760097e-01 9.14502859e-01 -2.00208351e-01 -1.40686560e+00
7.91805923e-01 6.40739501e-01 4.00848538e-01 -1.03439379e+00
-2.52551377e-01 -2.34755650e-01 -4.03699189e-01 -3.64263833e-01
4.94074315e-01 -2.80830055e-01 -2.20646128e-01 2.07916594e+00
3.27963322e-01 -6.22169197e-01 -1.60906836e-01 6.13319576e-01
3.11730951e-01 1.03015971e+00 -7.81033337e-01 -1.00805855e+00
1.27666557e+00 -6.55824900e-01 -5.27225792e-01 1.49895614e-02
8.49992394e-01 -3.46989036e-01 7.15104699e-01 5.95200956e-01
-1.76157284e+00 4.04724270e-01 -1.14151871e+00 8.72324482e-02
-5.19592315e-02 -4.62217331e-01 5.54946125e-01 8.70134354e-01
-1.15200949e+00 2.23737285e-01 -3.47488850e-01 -2.07806483e-01
1.60374835e-01 5.96814990e-01 -9.45471078e-02 -3.55992138e-01
-5.31220138e-01 2.68341333e-01 2.94239849e-01 -4.33522195e-01
-3.68869781e-01 -6.52146697e-01 -5.11073649e-01 4.14033026e-01
1.07331407e+00 -6.68119729e-01 1.00215733e+00 -4.84095395e-01
-7.77710319e-01 6.72554016e-01 -3.29152286e-01 -5.84813118e-01
3.50832403e-01 1.42708123e-01 5.22096872e-01 5.12895525e-01
-6.92325607e-02 1.06538273e-01 9.65092927e-02 -1.01951265e+00
-8.22793663e-01 -7.40449727e-01 7.79636919e-01 4.87920403e-01
-7.74791837e-01 -4.63510156e-02 -5.54053247e-01 -1.06691696e-01
2.89346993e-01 -9.30836737e-01 -5.08857489e-01 -7.47687593e-02
-5.73576689e-01 -2.45645523e-01 3.94427866e-01 -1.93756729e-01
1.74901342e+00 -2.01860857e+00 1.78008616e-01 6.00902736e-01
3.48845303e-01 -3.01804155e-01 -8.96346048e-02 7.57460833e-01
5.28799474e-01 5.08977056e-01 -3.89652729e-01 -3.09786230e-01
2.34220862e-01 2.24114254e-01 -9.89463627e-02 7.40572572e-01
-9.75038588e-01 5.69274127e-01 -5.39274395e-01 -2.22262099e-01
-5.67890763e-01 -5.71150243e-01 -9.68548656e-01 -2.94739306e-01
-7.12128699e-01 -6.43598199e-01 -3.54050308e-01 4.66768205e-01
1.02570879e+00 -4.49250966e-01 5.03325164e-01 2.92316079e-01
1.56080410e-01 -7.54005387e-02 -1.69451010e+00 1.28127193e+00
-2.12175876e-01 3.86861950e-01 6.92450881e-01 -9.94648576e-01
2.77336895e-01 -7.64201805e-02 7.94435382e-01 -3.63790542e-01
1.57028243e-01 4.36430693e-01 -3.86414587e-01 -1.83969170e-01
8.32491934e-01 -8.28319043e-02 -4.90269452e-01 1.16090393e+00
-5.11575580e-01 -3.04137804e-02 4.14926201e-01 7.17737913e-01
1.31111467e+00 -1.08303297e+00 1.04679205e-01 -2.37351656e-01
1.82728752e-01 3.88348959e-02 5.31061530e-01 1.14588809e+00
1.86729297e-01 4.56243426e-01 1.01490819e+00 -1.00887880e-01
-9.15718079e-01 -7.63014734e-01 9.83757526e-02 1.28243768e+00
4.95218188e-01 -5.73759377e-01 -8.77619743e-01 -4.15349275e-01
3.82366955e-01 6.33515954e-01 -4.86456126e-01 3.03093225e-01
-3.60762954e-01 -8.64771128e-01 3.15052211e-01 1.37722045e-01
3.85982424e-01 -4.49325711e-01 -3.84204298e-01 1.84803456e-02
-3.05228323e-01 -1.27678633e+00 -1.18900335e+00 -5.50996214e-02
-8.91520083e-01 -1.20634127e+00 -3.09084594e-01 -7.07564056e-01
9.21254218e-01 7.00284362e-01 6.25598311e-01 4.08016741e-01
1.76907301e-01 7.20612109e-01 -2.35341564e-01 -4.09727633e-01
-5.86424433e-02 2.19786733e-01 -2.57849004e-02 -4.17719126e-01
1.23907678e-01 -3.23940605e-01 -7.21662939e-01 2.40090732e-02
-1.25686395e+00 -4.90278959e-01 8.57548118e-02 2.74854928e-01
8.12195957e-01 5.02644062e-01 4.30331886e-01 -1.03768587e+00
7.99311280e-01 -6.41178906e-01 -9.89133000e-01 3.14981550e-01
-7.82583952e-01 1.42090529e-01 6.94920123e-01 -2.19680011e-01
-2.14279592e-01 3.53250168e-02 1.88627198e-01 -3.45861958e-03
7.03300953e-01 7.95671344e-01 -1.81810349e-01 -2.92692147e-03
5.54233074e-01 5.36452115e-01 1.19450808e-01 -1.54229999e-01
3.38977963e-01 6.26599431e-01 4.09261465e-01 -5.87437451e-01
6.28092229e-01 6.92643106e-01 4.54504669e-01 -8.96271050e-01
-7.04209089e-01 -6.91663504e-01 1.40645474e-01 1.61016077e-01
2.27020949e-01 -5.42751670e-01 -1.20524859e+00 1.02844894e-01
-8.44895005e-01 -2.20937729e-01 -4.56420839e-01 -4.32162285e-02
-6.81715906e-01 7.65954733e-01 -2.50454515e-01 -1.10715055e+00
-6.39926791e-01 -7.91756928e-01 5.22548795e-01 5.30728176e-02
2.54225910e-01 -4.37898457e-01 -1.07550718e-01 5.38772225e-01
3.40091527e-01 2.92313606e-01 9.02812302e-01 -8.26042771e-01
-1.08090889e+00 -5.51239967e-01 -5.04879728e-02 7.18780607e-02
-3.36343467e-01 -3.76480281e-01 -1.09388918e-01 -7.13752031e-01
-1.37664497e-01 -1.35809332e-01 8.37487876e-01 5.49872160e-01
1.33461642e+00 -1.16448975e+00 -4.32748795e-01 6.49674654e-01
1.66200960e+00 8.02419893e-03 4.13950145e-01 2.04152748e-01
6.62181973e-02 2.38394529e-01 5.90811074e-01 1.00998402e+00
6.38010561e-01 4.12999004e-01 5.88455200e-01 4.72427428e-01
5.93715429e-01 -2.27124803e-02 1.52841106e-01 2.59816825e-01
1.17464192e-01 -9.47245181e-01 -2.89632469e-01 8.62888753e-01
-1.91266525e+00 -9.06816542e-01 7.29035065e-02 2.69578433e+00
9.07478750e-01 1.42842487e-01 4.69790965e-01 2.06225038e-01
5.15436292e-01 1.08812883e-01 -6.72358274e-01 -6.90964937e-01
-3.06175977e-01 1.85305908e-01 1.33305967e+00 8.40202391e-01
-5.67371070e-01 6.14534557e-01 6.34429646e+00 8.33228350e-01
-5.26787698e-01 4.24455889e-02 5.17980516e-01 -9.54439223e-01
-9.68864918e-01 -1.52821332e-01 -8.18995893e-01 5.50106823e-01
9.45961058e-01 -9.07555938e-01 8.61599445e-01 8.54912996e-01
-1.12081449e-02 -5.67827404e-01 -1.08978271e+00 9.52916026e-01
1.94314837e-01 -1.52869773e+00 -5.45274653e-03 6.45789981e-01
9.72013950e-01 -2.49015406e-01 4.89941426e-02 -2.07457572e-01
2.38947481e-01 -8.15501630e-01 5.56547940e-01 -5.82325198e-02
8.27077746e-01 -1.11609173e+00 5.07615030e-01 8.06565166e-01
-1.00442362e+00 -5.08967698e-01 -4.56430465e-01 9.26215574e-02
3.04935664e-01 5.96791446e-01 -5.97870767e-01 3.93596709e-01
6.57344818e-01 -3.99543941e-01 1.70843318e-01 1.19659770e+00
3.81003529e-01 3.18928391e-01 -1.36746311e+00 -3.80874008e-01
3.68023291e-02 -6.52303398e-02 8.17935050e-01 1.07945836e+00
4.75973964e-01 8.52052331e-01 2.65187502e-01 3.23712170e-01
-6.73860610e-01 3.42382044e-01 -4.43281561e-01 -9.76858139e-02
9.56362247e-01 8.08131337e-01 -5.44267535e-01 -2.01077655e-01
-1.30095705e-01 7.89169192e-01 2.40327284e-01 1.14582643e-01
-7.36073911e-01 -6.40117407e-01 6.55313432e-01 5.06147802e-01
6.45450473e-01 -4.71703261e-01 -4.73486692e-01 -7.46661067e-01
5.24649441e-01 -5.19525588e-01 1.01649988e+00 -1.39413834e-01
-7.86757946e-01 1.30009472e-01 2.82448173e-01 -5.12771368e-01
-1.31473662e-02 -1.68282181e-01 -3.51076096e-01 5.18398702e-01
-1.27582550e+00 -3.11180741e-01 -1.92589185e-03 4.55983728e-01
3.13251722e-03 1.97651803e-01 3.88245165e-01 8.06335583e-02
-2.17871979e-01 9.75413918e-01 3.86583567e-01 -4.01248127e-01
8.73457342e-02 -1.21389377e+00 -4.08460528e-01 9.79116321e-01
-1.64318889e-01 3.96679580e-01 9.32279885e-01 -2.63295680e-01
-1.98697841e+00 -6.19262457e-01 1.05534220e+00 -4.32144292e-02
3.09707761e-01 -2.30257407e-01 -3.85569900e-01 7.28142679e-01
3.35964188e-02 -7.23809078e-02 7.79823720e-01 -5.65785654e-02
-1.51058480e-01 -5.10341287e-01 -1.62485743e+00 2.93354630e-01
1.21331656e+00 -2.52752677e-02 -1.69055238e-02 5.81720531e-01
1.03217447e+00 -6.28291965e-01 -7.87070096e-01 3.51097524e-01
4.67808962e-01 -4.95942235e-01 7.71124244e-01 -5.25377572e-01
-2.29334203e-03 -1.65510342e-01 -6.16716564e-01 -7.07434177e-01
-6.11578971e-02 -1.00698364e+00 -5.62848449e-01 5.94641447e-01
6.80621922e-01 -6.80068970e-01 1.26931512e+00 8.57821405e-01
2.26689294e-01 -1.13124156e+00 -1.20946610e+00 -7.59891689e-01
2.68599361e-01 -3.20706218e-01 6.57564044e-01 5.90329289e-01
3.60990703e-01 -1.32390082e-01 -2.28318155e-01 4.46695507e-01
6.43272161e-01 6.19996369e-01 9.22744274e-01 -7.11618543e-01
-7.13464141e-01 -4.06932235e-01 -2.70310417e-03 -1.66791725e+00
-2.73003399e-01 -8.93019259e-01 -4.17772867e-02 -1.70273352e+00
5.91010153e-01 -5.58950901e-01 -2.71626003e-02 3.09934527e-01
1.16740555e-01 -2.03260779e-02 4.67419505e-01 1.86395664e-02
-1.11346698e+00 1.32802173e-01 1.10711694e+00 1.39689874e-02
-3.99853408e-01 2.17705235e-01 -1.37695861e+00 2.40212545e-01
6.07302606e-01 -4.96642202e-01 -4.17294145e-01 -4.50252861e-01
7.97197580e-01 6.52995884e-01 -4.42201704e-01 -5.09668410e-01
6.63293362e-01 -5.17541289e-01 -5.21559775e-01 -7.75103688e-01
2.35591516e-01 -7.62738645e-01 -5.42874932e-02 5.17072558e-01
-2.91742027e-01 1.64699644e-01 1.13493159e-01 7.28222847e-01
2.54951447e-01 -5.09352982e-01 6.21292472e-01 -5.79243526e-02
1.23235859e-01 7.11406887e-01 -2.56733865e-01 3.09723198e-01
1.36178923e+00 -4.12934244e-01 -6.95674479e-01 -9.46944058e-01
-4.48839664e-01 8.83608341e-01 5.64558148e-01 -4.69039261e-01
5.18302500e-01 -7.49285340e-01 -7.54616082e-01 -3.70093405e-01
-1.46443442e-01 2.89548188e-01 3.07706207e-01 7.10400403e-01
-6.22471035e-01 5.97486854e-01 4.64411288e-01 -1.56740665e-01
-1.54597461e+00 7.15777934e-01 2.12050661e-01 -4.41345096e-01
-1.24797128e-01 8.81489575e-01 -1.62888408e-01 7.32410774e-02
4.48732018e-01 -2.26607755e-01 4.91009474e-01 -8.13659802e-02
5.77585161e-01 6.68160200e-01 -1.24077797e-01 -1.46012306e-01
-2.52871275e-01 2.30436191e-01 -1.03340372e-01 -4.62059796e-01
1.50323069e+00 -2.71280795e-01 -4.56939310e-01 -2.79878646e-01
1.19345367e+00 4.40177292e-01 -7.09086955e-01 -4.81594861e-01
-4.14224982e-01 -6.58677518e-01 -9.21214223e-02 -4.81263071e-01
-1.24478114e+00 8.85878354e-02 -8.42156187e-02 7.12321222e-01
1.46696413e+00 3.91301721e-01 1.12393546e+00 9.17475700e-01
9.01320159e-01 -1.19921136e+00 -1.61593050e-01 2.11096883e-01
6.88948095e-01 -7.73310721e-01 4.38094765e-01 -5.72787166e-01
-1.77609622e-01 9.74849522e-01 3.71397585e-01 -1.13619931e-01
7.28969276e-01 4.83450025e-01 -8.66239846e-01 -2.45332554e-01
-9.74083602e-01 -1.19745478e-01 -1.71109453e-01 -5.26153594e-02
-1.32252738e-01 2.42357895e-01 -9.99181569e-01 9.51391697e-01
-5.07900655e-01 -8.61640722e-02 8.94912124e-01 1.07691896e+00
-1.26794839e+00 -1.04444933e+00 -4.15973663e-01 7.56943583e-01
-7.75572419e-01 -1.30628020e-01 -3.47764552e-01 4.57509607e-01
-2.90551186e-01 1.05852115e+00 2.24929124e-01 -1.54036939e-01
2.39995286e-01 -2.11587146e-01 7.38323808e-01 -6.23618484e-01
-2.58174688e-01 -1.75739035e-01 8.54558647e-02 -3.45345795e-01
9.48237106e-02 -6.94691300e-01 -1.61645055e+00 -9.65898931e-01
-4.64827955e-01 3.31863195e-01 5.41184723e-01 7.10544705e-01
2.57784575e-01 -2.78544039e-01 8.41340959e-01 7.66884312e-02
-9.04711604e-01 -5.78459144e-01 -8.59837532e-01 -5.61245531e-02
1.93470344e-01 9.68729854e-02 -5.88335454e-01 -4.14701074e-01]
|
[6.569189548492432, 4.8917083740234375]
|
bb72cc20-b93f-4d4d-b458-18b55805c8ee
|
distinctive-image-features-from-scale
| null | null |
https://www.cs.ubc.ca/~lowe/papers/ijcv04.pdf
|
https://www.cs.ubc.ca/~lowe/papers/ijcv04.pdf
|
Distinctive Image Features from Scale-Invariant Keypoints
|
This paper presents a method for extracting distinctive invariant features from images that can be used to perform reliable matching between different views of an object or scene. The features are invariant to image scale and rotation, and are shown to provide robust matching across a substantial range of affine distortion, change in 3D viewpoint, addition of noise, and change in illumination. The features are highly distinctive, in the sense that a single feature can be correctly matched with high probability against a large database of features from many images. This paper also describes an approach to using these features for object recognition. The recognition proceeds by matching individual features to a database of features from known objects using a fast nearest-neighbor algorithm, followed by a Hough transform to identify clusters belonging to a single object, and finally performing verification through least-squares solution for consistent pose parameters. This approach to recognition can robustly identify objects among clutter and occlusion while achieving near real-time performance.
|
['David Lowe']
|
2004-01-05
| null | null | null | null |
['patch-matching']
|
['computer-vision']
|
[ 5.51107526e-01 -7.92337000e-01 5.93746789e-02 -6.48565471e-01
-1.12150609e+00 -7.28301048e-01 7.02639043e-01 -8.16918015e-02
-5.03092334e-02 1.70931518e-01 -1.02204971e-01 -8.71359557e-03
-2.81387091e-01 -3.54211271e-01 -4.06541824e-01 -6.58491015e-01
-7.58215338e-02 5.45969844e-01 5.09621978e-01 1.77174911e-01
6.98307753e-01 1.54864907e+00 -1.86923802e+00 4.07839753e-02
1.32760918e-03 1.08655989e+00 8.23631808e-02 7.84321070e-01
2.62763679e-01 2.27236822e-01 -6.41124785e-01 3.54779209e-03
6.50389910e-01 -2.09310204e-01 -6.22238219e-01 8.19489717e-01
1.09493065e+00 -2.14771435e-01 -2.14062512e-01 8.75404656e-01
5.47464311e-01 1.95965841e-01 8.21121812e-01 -1.24305427e+00
-2.71271229e-01 -5.63180864e-01 -6.12138808e-01 1.57911226e-01
9.91244853e-01 -1.48398550e-02 6.65725887e-01 -1.26816237e+00
6.66647255e-01 1.19689023e+00 4.92457241e-01 1.80511534e-01
-1.37356114e+00 -3.60026091e-01 -3.50943595e-01 8.19465369e-02
-1.72026610e+00 -7.31376410e-01 6.45177543e-01 -4.31351453e-01
9.22297239e-01 7.14246631e-01 4.88285840e-01 2.12724522e-01
2.53544778e-01 1.29018411e-01 1.02765846e+00 -6.55834198e-01
-2.15169396e-02 1.81113854e-01 -1.02838585e-02 8.29818308e-01
3.78507406e-01 2.38293439e-01 -3.81009161e-01 -5.28504968e-01
6.83544397e-01 1.43032372e-01 -1.67778581e-01 -1.00167918e+00
-1.25523460e+00 3.65557164e-01 3.06514621e-01 1.64775327e-01
-1.52573809e-01 -2.30360717e-01 1.21470191e-01 3.30925852e-01
-1.71317428e-01 4.50130135e-01 -1.96226507e-01 2.59944618e-01
-6.24914706e-01 2.75320530e-01 7.75437117e-01 1.06829870e+00
7.23995447e-01 2.09137965e-02 3.93939167e-01 6.77001119e-01
4.39642161e-01 9.37762499e-01 3.04837883e-01 -8.71614158e-01
1.28959656e-01 5.24632990e-01 1.74701005e-01 -1.35505259e+00
-3.98325980e-01 8.42900500e-02 -2.95894474e-01 8.10078263e-01
3.45885754e-01 4.98384804e-01 -7.43315816e-01 1.05233645e+00
6.78584814e-01 -3.01870912e-01 -7.14343935e-02 7.59034276e-01
5.84017038e-01 4.01418090e-01 -6.54856503e-01 -1.57688797e-01
1.35458875e+00 -3.39438349e-01 -2.04041213e-01 -2.11432114e-01
1.15133956e-01 -1.37638366e+00 3.56962413e-01 3.62091988e-01
-8.37632596e-01 -7.92006314e-01 -1.19832206e+00 1.89386413e-01
-2.97826052e-01 -1.83278635e-01 1.95947424e-01 5.96078038e-01
-8.22348714e-01 4.37366694e-01 -5.07155418e-01 -5.42651534e-01
-4.80234884e-02 7.07660437e-01 -9.35304463e-01 -2.72147685e-01
-2.82007873e-01 1.27411509e+00 2.20708475e-01 -7.01384023e-02
-4.48960155e-01 -4.00860608e-02 -9.62761641e-01 -3.17799628e-01
-3.16094607e-02 -3.65658700e-01 9.52625573e-01 -8.11038554e-01
-9.39496279e-01 1.16838276e+00 -6.82144463e-01 1.31327853e-01
1.82280764e-01 4.75166775e-02 -7.98446655e-01 3.63688678e-01
1.26016378e-01 1.82753056e-01 1.25149953e+00 -1.34907997e+00
-5.75125337e-01 -7.57834733e-01 -4.97294247e-01 4.09415036e-01
1.86603546e-01 5.70227325e-01 -4.80366319e-01 -2.78367519e-01
8.82407963e-01 -8.49421561e-01 2.92527732e-02 1.85668737e-01
1.76995136e-02 6.10592924e-02 1.36255097e+00 -5.53477049e-01
6.82102621e-01 -2.16459179e+00 -1.82331413e-01 7.31398582e-01
-5.34761623e-02 1.10574476e-01 -1.85183302e-01 4.62148130e-01
-3.13808590e-01 -2.59641051e-01 3.15455496e-01 2.84250915e-01
-2.43759677e-01 9.93466303e-02 -8.18401799e-02 1.08024359e+00
1.81261957e-01 5.47370613e-01 -5.45670927e-01 -5.72853088e-01
5.26012719e-01 3.46161813e-01 5.97016849e-02 1.78319216e-01
6.46511912e-01 -6.71256632e-02 -3.18649411e-01 9.51007366e-01
6.54719472e-01 1.12440757e-01 -6.47620633e-02 -3.68255198e-01
-2.97027845e-02 3.05539486e-03 -1.87023735e+00 7.34932899e-01
-7.79439183e-03 6.79737568e-01 1.09920859e-01 -7.75353968e-01
1.39854920e+00 2.76470751e-01 5.17264009e-01 -4.48356271e-01
1.69791877e-01 3.56296331e-01 -8.10147151e-02 -3.86545628e-01
4.00301307e-01 -4.10121940e-02 1.21301167e-01 5.78800917e-01
-1.62870973e-01 -6.22954488e-01 -3.95894349e-02 -1.01284534e-01
7.53044605e-01 -3.69623780e-01 7.20282376e-01 -1.57300949e-01
6.33131862e-01 -8.51681605e-02 2.81555712e-01 6.51805580e-01
-2.09172830e-01 8.39936256e-01 -3.01606983e-01 -7.55142868e-01
-1.21715331e+00 -1.04579401e+00 -4.81764197e-01 4.49624151e-01
5.51377535e-01 2.24461049e-01 -2.19537392e-01 -2.84063190e-01
3.28549176e-01 -6.28979970e-03 -3.04058224e-01 -1.05234653e-01
-4.34504330e-01 -3.74051601e-01 1.26219451e-01 4.83755410e-01
2.60149479e-01 -7.25897789e-01 -7.92368948e-01 6.72052801e-02
1.85023069e-01 -9.98449147e-01 -5.55987537e-01 2.68228859e-01
-8.44219625e-01 -1.43644297e+00 -3.82844836e-01 -1.03948891e+00
1.07438195e+00 9.26909566e-01 8.57628345e-01 1.96477816e-01
-8.73438060e-01 7.36956835e-01 -2.42730871e-01 -1.67640388e-01
-3.60452116e-01 -6.57904208e-01 5.28545201e-01 5.12197576e-02
5.26713848e-01 -1.52998418e-01 -1.57726720e-01 1.09264898e+00
-4.94671404e-01 -5.68262637e-01 3.67403954e-01 8.25398147e-01
6.72176778e-01 1.82141930e-01 3.28460112e-02 -1.43542707e-01
1.53251320e-01 1.81951478e-01 -8.27578127e-01 3.55650246e-01
-4.58193243e-01 -2.43319690e-01 2.71692753e-01 -3.72807443e-01
-7.31552780e-01 4.66976166e-01 3.94634157e-01 -2.50921100e-01
-4.71476525e-01 -9.36268456e-03 -3.54354829e-01 -9.75830138e-01
9.63255644e-01 4.06330973e-01 3.12479943e-01 -2.02057227e-01
1.93839863e-01 7.77681708e-01 8.48274052e-01 -2.85656273e-01
1.33217824e+00 5.63671470e-01 3.09419781e-01 -1.01882076e+00
-4.01517749e-01 -1.12037647e+00 -1.16992629e+00 -2.68406659e-01
4.31155920e-01 -7.04281926e-01 -4.42416936e-01 5.44921041e-01
-9.60871100e-01 5.40487468e-01 -1.07256487e-01 7.19908714e-01
-6.56763017e-01 6.20007515e-01 -1.87821016e-02 -7.52284050e-01
6.18344992e-02 -1.03934288e+00 1.01685882e+00 2.46406853e-01
-3.00502479e-01 -7.56164253e-01 -1.29318744e-01 3.25679570e-01
2.14072376e-01 2.15403110e-01 5.03969014e-01 -7.55626976e-01
-6.68980062e-01 -1.09043503e+00 -2.77032405e-01 1.92165777e-01
6.36820257e-01 4.20210987e-01 -8.76021922e-01 -7.19735503e-01
1.73583403e-01 -2.04495370e-01 2.51495421e-01 1.80241078e-01
5.07172167e-01 -1.12826481e-01 -5.07837594e-01 6.00247085e-01
1.46089137e+00 6.68386579e-01 5.00922382e-01 3.66084486e-01
3.73810381e-01 5.29596627e-01 7.15319693e-01 1.53313607e-01
-1.77629545e-01 8.45442832e-01 2.01147079e-01 -2.67791212e-01
1.19099356e-01 2.89966613e-01 4.99115027e-02 4.06283289e-01
2.98602358e-02 3.00964504e-01 -8.92445266e-01 4.02911812e-01
-1.28505766e+00 -1.29170048e+00 -9.70916227e-02 2.54216671e+00
1.47548631e-01 2.38878299e-02 2.30232239e-01 2.78727174e-01
1.11153662e+00 9.45351087e-03 -5.95053554e-01 -3.44875753e-01
-2.30296478e-01 -1.42633170e-01 6.69582307e-01 4.74954724e-01
-1.22268629e+00 3.81784737e-01 7.92528582e+00 3.60455006e-01
-1.06375241e+00 -6.69611573e-01 1.75093934e-01 3.76678169e-01
1.01635166e-01 7.24359974e-02 -8.96385670e-01 -9.24757496e-02
3.17717254e-01 -3.90092105e-01 1.89478725e-01 7.96330869e-01
-1.64761886e-01 -3.21632862e-01 -1.02790129e+00 1.06275821e+00
6.79359972e-01 -9.21002090e-01 -3.65717784e-02 6.30227998e-02
5.67569017e-01 -1.92285046e-01 3.13353054e-02 -4.90502149e-01
1.21489331e-01 -7.55211592e-01 4.09040749e-01 4.32210058e-01
6.73133016e-01 -5.53733289e-01 5.29891253e-01 2.14709833e-01
-1.37647378e+00 6.64981008e-02 -5.77446520e-01 1.14397049e-01
-5.79051562e-02 1.30984411e-01 -1.02659106e+00 4.26888645e-01
5.65751910e-01 3.33193123e-01 -6.64772570e-01 1.51434290e+00
3.39560360e-01 -2.50634164e-01 -4.60487783e-01 -4.19594953e-03
-1.31351680e-01 -7.23770112e-02 5.76292992e-01 1.00629926e+00
3.55324239e-01 1.57803014e-01 4.35822994e-01 2.44315580e-01
5.55485785e-01 6.54085875e-02 -1.04789495e+00 3.35024655e-01
6.66175544e-01 1.23676980e+00 -8.90546858e-01 -2.58350670e-01
-4.89712000e-01 9.53560472e-01 -1.63618714e-01 1.96063474e-01
-2.65050739e-01 -7.60370433e-01 4.03543800e-01 5.35675399e-02
4.87528533e-01 -3.65100652e-01 1.08284421e-01 -8.78281772e-01
2.58249938e-01 -1.04281521e+00 2.93330550e-01 -8.57959986e-01
-1.22787929e+00 5.94143629e-01 1.01603873e-01 -1.65821028e+00
-4.11406428e-01 -8.25984359e-01 -5.19436598e-01 1.09540319e+00
-9.07676697e-01 -1.05184138e+00 -2.33375534e-01 8.19484472e-01
2.17375636e-01 -5.42634666e-01 9.45086837e-01 -1.68525763e-02
6.99914154e-03 3.63901615e-01 3.50695133e-01 1.72888324e-01
8.31883788e-01 -9.56707180e-01 4.33876693e-01 7.87409544e-01
3.18078846e-01 6.58149064e-01 6.62777007e-01 -4.86530870e-01
-1.61497152e+00 -6.93279862e-01 1.02557969e+00 -6.25762105e-01
3.83376420e-01 -8.51345658e-02 -7.96386719e-01 4.76393342e-01
-3.40122581e-01 4.67754781e-01 5.73820114e-01 -5.93033768e-02
-6.85232222e-01 -2.77297914e-01 -1.40096557e+00 1.54111519e-01
7.07160652e-01 -8.20532918e-01 -7.21711159e-01 2.56832987e-01
-1.97729487e-02 -5.56874871e-01 -8.47950459e-01 1.98284522e-01
1.07562578e+00 -9.22032475e-01 1.24815464e+00 -5.05564153e-01
-4.63956714e-01 -7.81670928e-01 -5.82954168e-01 -9.06755030e-01
-6.50218904e-01 -4.26920623e-01 4.98578012e-01 1.03055203e+00
7.98759088e-02 -6.33056819e-01 5.15052617e-01 7.52780437e-01
2.58565545e-01 -2.33071849e-01 -9.34185266e-01 -1.24796593e+00
-7.02045918e-01 6.21168613e-02 4.86711830e-01 6.89997315e-01
-1.26462290e-03 -1.21959980e-04 -2.06984669e-01 4.25510883e-01
8.84244382e-01 6.09917879e-01 1.17185843e+00 -1.36601138e+00
2.43591275e-02 -1.47116125e-01 -1.39911377e+00 -6.34583592e-01
-9.93374363e-02 -5.79738438e-01 8.59143361e-02 -1.13864565e+00
2.97885597e-01 -3.80564600e-01 1.25786692e-01 2.63955623e-01
4.97455709e-02 4.84824836e-01 1.62559345e-01 4.56210285e-01
-3.34191144e-01 -7.78370649e-02 7.85801411e-01 -8.08510706e-02
1.52437583e-01 2.23683134e-01 -4.62551415e-01 7.83219278e-01
6.16886318e-01 -4.38787758e-01 4.59858142e-02 -1.57834023e-01
-4.71350193e-01 1.63885295e-01 3.98559928e-01 -1.21658111e+00
2.58548379e-01 -3.32434326e-01 9.60151017e-01 -6.95944130e-01
4.53094333e-01 -1.27920890e+00 3.76076758e-01 5.55791259e-01
1.32373802e-03 5.46258748e-01 1.27562523e-01 4.38504040e-01
-3.46419036e-01 -3.35102737e-01 1.16006994e+00 -4.17850055e-02
-1.03683639e+00 5.40287867e-02 -3.12933803e-01 -3.46405685e-01
1.29534721e+00 -8.64003301e-01 -1.25866681e-01 -4.45403188e-01
-6.08195841e-01 -1.50572672e-01 8.40690434e-01 2.90653795e-01
8.99868250e-01 -1.58657801e+00 -6.45532250e-01 8.72902632e-01
3.91605794e-01 -5.12758553e-01 -1.67275533e-01 4.71684247e-01
-6.74907744e-01 3.45287085e-01 -3.60799134e-01 -9.61017370e-01
-2.11991787e+00 6.49702668e-01 4.66353148e-01 3.97793382e-01
-3.90396029e-01 6.71159565e-01 -2.21979916e-01 -2.88826078e-01
1.82491750e-01 2.05617502e-01 1.78489894e-01 -1.94505319e-01
7.57813096e-01 2.82854378e-01 2.42486104e-01 -1.29319072e+00
-6.53312504e-01 1.31643224e+00 -1.22211397e-01 -1.33485585e-01
8.49859178e-01 -2.07546055e-01 -1.45162027e-02 3.58258784e-01
1.59510052e+00 5.80008850e-02 -1.00494707e+00 -5.32391787e-01
-2.70273630e-02 -1.10674107e+00 -3.82855624e-01 -5.08883417e-01
-7.48871505e-01 5.67255378e-01 8.89446199e-01 1.58306107e-01
1.00569129e+00 1.45678118e-01 1.14924513e-01 7.13178098e-01
5.57544112e-01 -8.51814568e-01 6.19337298e-02 2.97950417e-01
9.39250290e-01 -1.29893279e+00 5.14919341e-01 -3.52250397e-01
-3.83279353e-01 1.39058685e+00 4.57786322e-01 -2.33336300e-01
5.85765004e-01 3.33431333e-01 3.56235057e-01 -2.25624979e-01
-4.40245211e-01 -1.37745231e-01 7.22062767e-01 8.83430302e-01
1.83719754e-01 -1.55924186e-01 1.41231358e-01 -6.55220747e-01
4.00655642e-02 -4.63030279e-01 2.41826221e-01 1.17172945e+00
-9.50405717e-01 -9.42862511e-01 -1.03406942e+00 3.78382832e-01
-1.85968399e-01 3.95367712e-01 -6.10024452e-01 8.51046443e-01
-2.62995183e-01 7.63282299e-01 1.73114344e-01 -3.13620239e-01
6.16921246e-01 -8.87918472e-03 7.97705293e-01 -4.05105263e-01
-2.54919469e-01 4.49828446e-01 -6.10206425e-02 -6.77831531e-01
-5.06798744e-01 -1.25275767e+00 -8.10658693e-01 -1.25391796e-01
-7.34843612e-01 2.17844434e-02 6.38633907e-01 6.85393631e-01
2.89934855e-02 -2.87332505e-01 1.18459558e+00 -1.13236046e+00
-7.33713508e-01 -3.38402539e-01 -9.27163124e-01 6.26504838e-01
4.14859921e-01 -7.24762559e-01 -4.84643817e-01 3.22088093e-01]
|
[8.159666061401367, -2.286419630050659]
|
3eaea165-1eeb-42a3-9688-1ada69a01eb8
|
developing-an-icu-scoring-system-with
|
1604.06730
| null |
http://arxiv.org/abs/1604.06730v1
|
http://arxiv.org/pdf/1604.06730v1.pdf
|
Developing an ICU scoring system with interaction terms using a genetic algorithm
|
ICU mortality scoring systems attempt to predict patient mortality using
predictive models with various clinical predictors. Examples of such systems
are APACHE, SAPS and MPM. However, most such scoring systems do not actively
look for and include interaction terms, despite physicians intuitively taking
such interactions into account when making a diagnosis. One barrier to
including such terms in predictive models is the difficulty of using most
variable selection methods in high-dimensional datasets. A genetic algorithm
framework for variable selection with logistic regression models is used to
search for two-way interaction terms in a clinical dataset of adult ICU
patients, with separate models being built for each category of diagnosis upon
admittance to the ICU. The models had good discrimination across all
categories, with a weighted average AUC of 0.84 (>0.90 for several categories)
and the genetic algorithm was able to find several significant interaction
terms, which may be able to provide greater insight into mortality prediction
for health practitioners. The GA selected models had improved performance
against stepwise selection and random forest models, and provides greater
flexibility in terms of variable selection by being able to optimize over any
modeler-defined model performance metric instead of a specific variable
importance metric.
|
['Chee Chun Gan', 'Gerard Learmonth']
|
2016-04-22
| null | null | null | null |
['icu-mortality']
|
['medical']
|
[ 3.84319127e-02 -2.12685481e-01 -3.12427044e-01 -4.84740108e-01
-4.50357556e-01 -2.69303083e-01 -3.11738923e-02 6.70347691e-01
-2.95058668e-01 9.57597852e-01 2.92847931e-01 -8.99627745e-01
-9.01382446e-01 -6.69624507e-01 2.76068658e-01 -6.09847188e-01
-5.25025189e-01 1.07472157e+00 -1.87726200e-01 -1.86123829e-02
2.99313188e-01 3.49940956e-01 -1.07074130e+00 1.93575054e-01
7.70818889e-01 4.66503412e-01 -3.57656293e-02 8.50271225e-01
2.57193539e-02 9.04755533e-01 -3.75405610e-01 -4.26030755e-02
1.69816732e-01 -9.00686383e-01 -5.69188237e-01 -4.60623085e-01
-3.94414753e-01 -2.68926650e-01 1.85586527e-01 7.53121600e-02
8.50457966e-01 -1.38021186e-01 9.68898892e-01 -1.19355059e+00
-1.16608068e-02 5.32787323e-01 6.74871579e-02 1.65270463e-01
3.76204789e-01 3.90011877e-01 9.42298055e-01 -7.11352289e-01
3.15415323e-01 1.03333569e+00 9.52773988e-01 4.34994221e-01
-1.58114266e+00 -5.87223291e-01 -2.15836063e-01 2.71240950e-01
-1.19233692e+00 -3.57970633e-02 2.11754605e-01 -8.37374210e-01
1.40941012e+00 5.81540108e-01 9.72437382e-01 3.38454872e-01
5.06488025e-01 -4.45386283e-02 9.73010421e-01 -5.66383302e-01
1.88071430e-01 1.16667673e-01 2.86023766e-01 6.85844362e-01
4.64758635e-01 4.04114544e-01 -3.94518197e-01 -1.01278007e+00
7.54560530e-01 6.15253627e-01 -2.68996924e-01 -5.29937327e-01
-1.34857047e+00 1.07309544e+00 1.89921662e-01 -7.49620274e-02
-6.01430476e-01 -2.14089915e-01 5.40784657e-01 3.50587726e-01
-1.00741751e-01 9.68533099e-01 -1.12345898e+00 -2.61570513e-01
-8.47632647e-01 5.98743148e-02 9.71628070e-01 3.93935084e-01
1.95354149e-01 -2.03962311e-01 -2.78256655e-01 9.77489114e-01
2.85563320e-01 1.02657549e-01 6.17213845e-01 -9.04324651e-01
1.83971047e-01 8.82669747e-01 -1.74113009e-02 -5.03468335e-01
-9.99447942e-01 -1.94858938e-01 -7.71919072e-01 4.35217619e-01
3.58724385e-01 -3.43006670e-01 -7.15604782e-01 1.38930929e+00
-1.27458870e-01 -2.64663547e-01 -4.70779128e-02 5.65397680e-01
5.01886547e-01 -4.91579697e-02 6.01340175e-01 -6.17654681e-01
1.30579495e+00 -7.54545093e-01 -3.00944537e-01 -1.25905141e-01
9.68631089e-01 -2.74079442e-01 6.21702075e-01 5.28302670e-01
-9.51635659e-01 8.06205645e-02 -6.23102307e-01 5.09497404e-01
1.09644845e-01 -2.37909541e-01 7.30103850e-01 3.64212245e-01
-1.01812398e+00 7.49878764e-01 -7.90082574e-01 -6.68657601e-01
1.67015418e-01 7.40860045e-01 -2.83554673e-01 -8.04326683e-02
-1.19924045e+00 1.35372674e+00 2.98397750e-01 -3.97682607e-01
-4.32435453e-01 -5.87286294e-01 -5.95037818e-01 2.44768068e-01
-8.81161019e-02 -1.73104715e+00 7.34148145e-01 -7.16561139e-01
-9.34772015e-01 6.61120653e-01 -3.93219709e-01 -3.92054737e-01
5.35508811e-01 8.04265309e-03 -1.48583636e-01 2.66301602e-01
-1.50583088e-01 1.41982138e-01 5.48062861e-01 -8.16965938e-01
-5.10339677e-01 -4.62998807e-01 -3.52689981e-01 3.69482547e-01
3.21181193e-02 3.69614601e-01 4.52022761e-01 -4.34640855e-01
2.27508545e-01 -7.61885047e-01 -8.91822517e-01 -1.81199789e-01
1.77086536e-02 1.32387087e-01 2.80542791e-01 -6.08649015e-01
1.69867444e+00 -1.42383516e+00 1.78307429e-01 2.54232734e-01
5.24465978e-01 -5.99539764e-02 2.94291615e-01 5.93842804e-01
-4.21128631e-01 5.64150631e-01 -2.53710061e-01 1.62131935e-01
-5.70442259e-01 8.65188166e-02 2.87285537e-01 1.12450927e-01
2.91366369e-01 7.93722153e-01 -7.92719901e-01 -7.10607231e-01
4.58490014e-01 4.29906815e-01 -8.47800612e-01 4.82991278e-01
3.76614034e-01 4.85971153e-01 -4.28285211e-01 7.04052985e-01
3.66394557e-02 -3.99791449e-01 5.65762341e-01 6.82917655e-01
1.59272641e-01 2.06641510e-01 -6.38532341e-01 7.60641873e-01
-2.01874986e-01 1.12506941e-01 -1.55532077e-01 -8.98497462e-01
1.09767497e+00 6.37384653e-01 9.25653815e-01 1.80986851e-01
8.81082937e-02 3.46218050e-01 4.75705504e-01 -6.19320571e-01
-4.72613811e-01 -9.20393109e-01 3.03289115e-01 2.94180930e-01
-3.04068506e-01 -3.16868663e-01 -1.44791633e-01 -1.04007483e-01
1.48763943e+00 -2.64080614e-01 1.01920724e+00 -2.21790954e-01
3.58373016e-01 3.43029231e-01 6.38569295e-01 7.50382423e-01
-2.80872434e-01 9.53415155e-01 5.71339488e-01 -7.67057776e-01
-7.23529458e-01 -6.30146027e-01 -5.74371874e-01 8.46036732e-01
-5.25784910e-01 -2.96537429e-01 -4.53993045e-02 -5.46479046e-01
2.97976285e-01 9.24199939e-01 -6.71772182e-01 -3.29347968e-01
-3.12948495e-01 -1.29472399e+00 3.31975102e-01 4.76417512e-01
-3.26420724e-01 -1.03164792e+00 -1.17738521e+00 7.18946636e-01
-8.07719901e-02 -4.03095828e-03 1.83375418e-01 1.09963870e+00
-1.59791946e+00 -1.39504957e+00 -5.68622530e-01 -5.59808671e-01
5.00708103e-01 -8.31667855e-02 1.35308683e+00 6.78036034e-01
-3.93533587e-01 1.74439773e-01 -2.99668491e-01 -5.84911585e-01
-3.82238656e-01 -1.96610838e-01 1.92831308e-01 -8.39557707e-01
7.07060039e-01 -5.64707100e-01 -8.71478319e-01 4.84148473e-01
-4.04793799e-01 -1.04284152e-01 4.71776545e-01 1.49491966e+00
3.50870073e-01 -1.67927340e-01 7.27117717e-01 -7.55071282e-01
6.02075815e-01 -9.66964900e-01 1.22993439e-01 1.12862892e-01
-1.46986198e+00 -1.45840392e-01 5.02917349e-01 -3.25092494e-01
-3.52698326e-01 8.64078403e-02 1.25609219e-01 -1.50328591e-01
-1.74926132e-01 6.45452619e-01 2.76151210e-01 1.77072793e-01
6.81271374e-01 -1.10246435e-01 3.36543590e-01 -4.22579259e-01
-5.12907982e-01 3.50125432e-01 -2.69360870e-01 -3.17523152e-01
3.56207103e-01 -2.44695470e-01 3.58885288e-01 -2.73319602e-01
-8.09150338e-02 -6.34502232e-01 -9.21114504e-01 9.70021859e-02
8.01737487e-01 -6.43167138e-01 -7.17029452e-01 -1.88013334e-02
-4.25227433e-01 -3.62997442e-01 -1.05802692e-01 9.25708711e-01
-6.96052849e-01 1.17530122e-01 -3.65734190e-01 -8.06030929e-01
-4.11629349e-01 -1.26473045e+00 4.58797067e-01 -8.74522179e-02
-1.05082369e+00 -1.36233366e+00 2.63174593e-01 2.31751084e-01
6.26444161e-01 4.54474419e-01 1.71085262e+00 -1.17244327e+00
4.86980751e-02 -3.99396151e-01 1.28747880e-01 1.93377547e-02
3.38669419e-01 1.58961773e-01 -3.76583576e-01 -1.75501287e-01
1.43551631e-02 -1.16939478e-01 6.21875882e-01 7.54719436e-01
9.67792928e-01 -3.06422085e-01 -5.53962171e-01 7.34994650e-01
1.38724685e+00 7.51965046e-01 3.64441186e-01 7.68686593e-01
3.75195682e-01 5.88454902e-01 4.75337416e-01 6.22505963e-01
2.86775768e-01 5.16713560e-01 3.70241493e-01 -2.60687888e-01
4.89303917e-01 1.67414293e-01 -9.46690142e-02 4.04431313e-01
-2.41119340e-01 4.97066006e-02 -1.58009768e+00 3.41918677e-01
-1.71306503e+00 -7.71901131e-01 -5.92739761e-01 2.32724476e+00
7.54057407e-01 6.76926151e-02 3.16705018e-01 1.57231182e-01
4.16707784e-01 -5.01928926e-01 -6.96894944e-01 -8.42014849e-01
-1.91216692e-02 1.65075347e-01 5.02633929e-01 3.50019932e-01
-7.28690863e-01 2.53567517e-01 8.42505741e+00 -2.95000404e-01
-7.80234337e-01 -4.20183510e-01 8.81527483e-01 -3.13147515e-01
1.94543451e-01 1.94793433e-01 -3.01190376e-01 3.92795354e-01
1.15352571e+00 -4.47483897e-01 2.86866426e-01 8.31182539e-01
6.83522344e-01 -1.35568067e-01 -1.12174010e+00 5.61944246e-01
-3.19746941e-01 -8.87251079e-01 -8.82535800e-02 5.78622110e-02
3.10759544e-01 -2.05015719e-01 -4.92350608e-01 2.81829894e-01
6.04560256e-01 -1.51627386e+00 1.25571266e-01 7.05159962e-01
9.13841128e-01 -5.56979477e-01 1.11398697e+00 2.28994280e-01
-5.83990216e-01 -5.14927447e-01 -1.33531541e-01 -6.39074743e-01
2.67125607e-01 4.13281083e-01 -1.44221091e+00 2.42208168e-01
7.20622957e-01 3.52287024e-01 -5.73996305e-01 1.33632481e+00
6.01142310e-02 1.01565981e+00 -1.29953206e-01 -2.20380872e-02
-1.67291626e-01 -9.89711881e-02 5.11765063e-01 1.17263985e+00
4.37207669e-01 4.66403723e-01 2.01052357e-03 1.93145648e-01
6.86845779e-01 4.67772782e-01 -6.12456381e-01 2.39572719e-01
3.64195079e-01 7.34762967e-01 -5.21758258e-01 -3.82098705e-01
-3.20801973e-01 5.35092711e-01 -4.39772233e-02 1.85685024e-01
-3.92684817e-01 -1.76150903e-01 7.69424677e-01 4.69022423e-01
-1.51910320e-01 8.27378556e-02 -1.18042386e+00 -8.89821053e-01
-7.02733755e-01 -1.08390951e+00 1.02828288e+00 -7.24026263e-01
-1.22591054e+00 5.19594133e-01 -4.01197039e-02 -1.28087735e+00
-5.88349521e-01 -5.02998054e-01 -6.27274632e-01 1.41995490e+00
-8.91939104e-01 -6.01777554e-01 -1.90930560e-01 3.36671293e-01
3.41742128e-01 -2.71728396e-01 1.44152284e+00 -1.62986234e-01
-4.37377930e-01 2.95574725e-01 2.12510630e-01 -2.50475943e-01
1.07997417e+00 -1.32960761e+00 -3.43414694e-01 1.38751283e-01
-8.77361655e-01 1.03003252e+00 6.94204986e-01 -8.60151589e-01
-8.48222077e-01 -8.72352302e-01 1.11799586e+00 -8.10597539e-01
3.01523656e-01 6.06676340e-01 -1.04130507e+00 5.43768346e-01
-2.07824752e-01 -5.87940276e-01 1.40179813e+00 3.54243994e-01
8.77999067e-02 4.36458945e-01 -1.36357880e+00 3.45968425e-01
7.81809032e-01 1.89447388e-01 -7.55168080e-01 2.63508230e-01
2.86987573e-01 2.09740847e-02 -1.24561918e+00 6.36056125e-01
9.05192673e-01 -9.79910433e-01 1.18407464e+00 -1.27496815e+00
6.03990138e-01 -3.57301943e-02 1.24290407e-01 -1.23248029e+00
-1.10945475e+00 -2.68291444e-01 -3.85583267e-02 6.90073192e-01
7.29076385e-01 -1.00363851e+00 4.34726894e-01 1.36651850e+00
3.52484286e-01 -1.19895816e+00 -6.92929208e-01 -4.04288262e-01
3.36133182e-01 -1.00639395e-01 6.73633873e-01 1.05839205e+00
4.58661854e-01 3.83178830e-01 -3.58033627e-01 -2.41241992e-01
4.03733581e-01 -1.78132862e-01 4.90591168e-01 -1.53323865e+00
-4.41781759e-01 -8.45044792e-01 -5.15727639e-01 -2.15081014e-02
-5.00770807e-01 -7.87506521e-01 -3.40174526e-01 -1.98227644e+00
6.10694826e-01 -8.47265124e-01 -7.09110916e-01 6.82978034e-01
-8.49307477e-01 -1.78113520e-01 -3.02234516e-02 6.25902295e-01
6.98615685e-02 -2.40442306e-02 7.03356385e-01 2.83996284e-01
-6.28875613e-01 2.43193135e-01 -1.01604867e+00 6.35610580e-01
1.15505481e+00 -8.09956610e-01 -3.66854906e-01 2.65939124e-02
-2.21865579e-01 6.34436190e-01 1.61163732e-01 -6.70684397e-01
-2.26556271e-01 -8.04697156e-01 7.77455628e-01 -2.11209506e-01
6.43015951e-02 -9.39872444e-01 8.34706008e-01 1.05194259e+00
-3.39543670e-01 4.35042739e-01 1.15259118e-01 1.79229230e-01
4.46953960e-02 -2.64224052e-01 8.23937416e-01 -2.28193477e-01
-3.02230597e-01 1.01529388e-02 -6.08138084e-01 -1.55940533e-01
1.07246304e+00 -5.13175905e-01 4.82225455e-02 -4.23208922e-01
-1.06902611e+00 4.61156160e-01 7.52371728e-01 1.12899110e-01
5.75012982e-01 -8.67428362e-01 -9.12205040e-01 9.05354396e-02
2.23739982e-01 -9.57705304e-02 -6.21228963e-02 1.21575582e+00
-9.48021531e-01 5.42998850e-01 -2.99804360e-01 -4.44610059e-01
-1.69117796e+00 6.60789132e-01 6.54164732e-01 -5.57078600e-01
-6.96437478e-01 7.65056431e-01 1.24173395e-01 -2.94262379e-01
2.33270675e-02 -3.87258455e-02 -3.72710317e-01 9.27540138e-02
5.00912428e-01 6.27429485e-01 -7.96962976e-02 -3.51413667e-01
-8.29648674e-01 4.03773546e-01 3.44021052e-01 1.43600494e-01
1.61146438e+00 -1.16229095e-01 -3.19645733e-01 4.97071087e-01
8.97369981e-01 -3.65650713e-01 -7.52921581e-01 2.13735133e-01
1.29224852e-01 -4.18259323e-01 -1.33592203e-01 -1.37647164e+00
-5.39713919e-01 7.95782149e-01 5.51354229e-01 7.52739161e-02
1.28345335e+00 -2.58645564e-01 2.20256001e-01 2.37393692e-01
3.42886746e-01 -4.95507836e-01 -4.40971643e-01 3.66962343e-01
7.79125273e-01 -1.31153440e+00 2.24988505e-01 1.22339129e-02
-1.07818878e+00 1.43074024e+00 2.78829902e-01 1.84588358e-01
8.35096896e-01 5.51794395e-02 3.91126573e-01 -1.78847641e-01
-1.28506529e+00 2.42254972e-01 5.18821403e-02 7.70607352e-01
8.63862574e-01 4.45484549e-01 -9.39808011e-01 6.79480135e-01
-1.84689566e-01 2.49976948e-01 3.19665402e-01 9.83452559e-01
-5.23106158e-01 -1.29390311e+00 -5.16994178e-01 1.35583496e+00
-6.99290395e-01 -3.76411796e-01 -5.41599572e-01 5.46890497e-01
-1.63611978e-01 1.04086339e+00 -4.12809223e-01 -4.01708990e-01
3.57434183e-01 8.08247328e-01 -1.21398456e-01 -6.03301466e-01
-1.02626371e+00 -6.28138334e-02 2.30980933e-01 -2.56055504e-01
8.41868967e-02 -1.07066178e+00 -1.22863722e+00 -1.65943176e-01
-2.90888906e-01 4.63122129e-01 3.93951148e-01 4.76759315e-01
2.36177996e-01 6.17841601e-01 6.33133650e-01 -5.02365708e-01
-5.48902035e-01 -9.89963531e-01 -5.65315962e-01 2.14660838e-01
3.40875804e-01 -7.57945657e-01 -6.52197421e-01 1.32603735e-01]
|
[8.06434440612793, 6.001269340515137]
|
d12f8fb3-d7d9-480f-a2f1-71cf0935c999
|
simplebert-a-pre-trained-model-that-learns-to
|
2204.07779
| null |
https://arxiv.org/abs/2204.07779v1
|
https://arxiv.org/pdf/2204.07779v1.pdf
|
SimpleBERT: A Pre-trained Model That Learns to Generate Simple Words
|
Pre-trained models are widely used in the tasks of natural language processing nowadays. However, in the specific field of text simplification, the research on improving pre-trained models is still blank. In this work, we propose a continued pre-training method for text simplification. Specifically, we propose a new masked language modeling (MLM) mechanism, which does not randomly mask words but only masks simple words. The new mechanism can make the model learn to generate simple words. We use a small-scale simple text dataset for continued pre-training and employ two methods to identify simple words from the texts. We choose BERT, a representative pre-trained model, and continue pre-training it using our proposed method. Finally, we obtain SimpleBERT, which surpasses BERT in both lexical simplification and sentence simplification tasks and has achieved state-of-the-art results on multiple datasets. What's more, SimpleBERT can replace BERT in existing simplification models without modification.
|
['Xiaojun Wan', 'Renliang Sun']
|
2022-04-16
| null | null | null | null |
['lexical-simplification']
|
['natural-language-processing']
|
[ 3.81777793e-01 1.49154738e-01 3.15700583e-02 -4.04277086e-01
-6.05245233e-01 -4.79700826e-02 5.11923850e-01 2.74722338e-01
-8.09844494e-01 7.73819566e-01 3.65801454e-01 -4.12968397e-01
3.03644776e-01 -6.67132795e-01 -4.85574484e-01 -3.00142467e-01
5.71256578e-01 4.28138435e-01 1.60108097e-02 -5.79267979e-01
3.04378301e-01 3.61628324e-01 -1.31179392e+00 5.18862963e-01
1.48994195e+00 -7.80336261e-02 6.46894753e-01 6.01165235e-01
-4.84540403e-01 7.86350250e-01 -9.34266388e-01 -7.09096313e-01
7.96561241e-02 -4.52161193e-01 -8.04613471e-01 -1.26066625e-01
3.28756750e-01 -2.14320630e-01 -3.96412134e-01 1.00712359e+00
5.42751014e-01 3.22080582e-01 6.40450597e-01 -5.69126189e-01
-7.31020868e-01 1.15853059e+00 -3.83488089e-01 1.78791672e-01
2.08405182e-01 -1.36566743e-01 7.47087836e-01 -9.16283429e-01
6.11030400e-01 1.72938383e+00 6.82723701e-01 8.91116560e-01
-9.72723424e-01 -7.12760389e-01 5.19349158e-01 1.32848039e-01
-1.52833366e+00 -6.04345620e-01 4.96207565e-01 -1.10153250e-01
1.18146706e+00 4.36091661e-01 2.98835039e-01 5.58980107e-01
4.85449702e-01 9.85093415e-01 8.31295729e-01 -9.75890040e-01
-2.91097730e-01 1.66407414e-03 4.12590027e-01 6.10335648e-01
3.89206797e-01 -4.53906536e-01 -2.14505792e-01 2.24508017e-01
2.32049584e-01 1.05598852e-01 -4.75422330e-02 4.42168921e-01
-8.99526775e-01 7.29001701e-01 -1.44604249e-02 4.75104392e-01
-8.14753175e-02 -2.22821534e-01 4.34617132e-01 4.54555750e-01
6.88661993e-01 7.02857375e-01 -4.68617111e-01 8.92532393e-02
-1.13348782e+00 3.92012388e-01 7.10043550e-01 1.16250491e+00
8.68637085e-01 2.56745279e-01 -7.31225014e-01 1.03647614e+00
-1.35018118e-02 6.15215838e-01 5.47785282e-01 -4.55064416e-01
7.04564154e-01 6.44550502e-01 -2.40932569e-01 -7.17047095e-01
-4.40745026e-01 -2.33974501e-01 -1.20786262e+00 -3.21519971e-01
-1.75757587e-01 -2.34488502e-01 -1.15804183e+00 1.61216128e+00
-1.53091177e-02 -1.75941423e-01 -4.10177857e-02 2.63436377e-01
1.05441701e+00 8.26592922e-01 2.08848104e-01 -4.47122216e-01
1.15117204e+00 -1.33825850e+00 -1.20799601e+00 -3.41599286e-01
1.10693884e+00 -1.14899826e+00 1.38417804e+00 5.96810222e-01
-1.07991266e+00 -5.86575627e-01 -1.08078647e+00 -7.21007347e-01
-5.61760366e-01 3.09768140e-01 5.69431901e-01 5.59758961e-01
-7.07412481e-01 8.54566514e-01 -5.00045836e-01 -2.62171865e-01
3.34217548e-01 2.90739477e-01 -2.38216847e-01 -4.97237235e-01
-1.38901258e+00 1.32638967e+00 4.08957303e-01 -1.44769132e-01
-5.45635283e-01 -6.56672180e-01 -1.01090765e+00 1.01251528e-01
3.71847183e-01 -9.77156818e-01 1.46478629e+00 -6.31708860e-01
-1.64763248e+00 7.55353630e-01 -8.27243745e-01 -4.01969373e-01
3.20212901e-01 -6.90893829e-01 -5.02729893e-01 -3.71205539e-01
2.07184795e-02 5.16590834e-01 1.04172421e+00 -1.08618379e+00
-4.51253295e-01 9.14318636e-02 1.27507508e-01 2.13967174e-01
-3.79380375e-01 2.96272367e-01 -5.22509038e-01 -1.06025732e+00
-2.88237423e-01 -6.91080689e-01 -4.35210973e-01 -6.35403872e-01
-6.66556299e-01 -4.37331855e-01 5.38098454e-01 -1.03780019e+00
2.10521340e+00 -1.99316084e+00 2.80875266e-01 -1.96416810e-01
3.13201904e-01 1.01934338e+00 -4.06422436e-01 5.78484654e-01
-2.57415548e-02 6.80907905e-01 -4.90636021e-01 -1.28990543e+00
-5.94354160e-02 4.14833069e-01 -4.14310843e-01 -2.17751656e-02
1.75597757e-01 9.83850121e-01 -6.59270704e-01 -7.93546975e-01
4.94478196e-01 2.10717753e-01 -6.67919219e-01 1.17498241e-01
-1.94466770e-01 1.63556397e-01 -7.26460516e-02 2.76857585e-01
9.94781494e-01 4.98559207e-01 -8.25554952e-02 5.78468367e-02
-3.32545310e-01 5.95136225e-01 -9.96496856e-01 1.71083939e+00
-6.81299567e-01 5.83705485e-01 -1.83974758e-01 -6.78017974e-01
7.89996266e-01 3.31927673e-03 -1.38659492e-01 -5.61077595e-01
3.67051810e-01 1.13688204e-02 1.22661397e-01 -4.89862859e-01
9.19277370e-01 -1.66920364e-01 -2.61848345e-02 3.49750370e-01
3.92923318e-02 -6.21800184e-01 7.85080016e-01 3.58670086e-01
1.01772141e+00 -1.19143002e-01 5.65410018e-01 -2.11877525e-01
8.43864441e-01 -1.72568381e-01 5.71718156e-01 9.24019933e-01
1.56981796e-01 7.57129908e-01 2.48454228e-01 -2.10756153e-01
-1.11665380e+00 -5.90951085e-01 1.04541779e-01 1.25797439e+00
2.95187719e-02 -1.19967937e+00 -1.00689387e+00 -5.43979466e-01
-1.29555151e-01 1.30390251e+00 -3.34188193e-01 -3.94343019e-01
-1.07986510e+00 -7.05382526e-01 7.61251390e-01 -2.12597623e-02
2.80992478e-01 -1.27100313e+00 2.44901150e-01 1.64980769e-01
-3.38688046e-01 -1.04197133e+00 -7.04134703e-01 -6.12974949e-02
-9.02892590e-01 -7.09526598e-01 -4.75419521e-01 -9.59769905e-01
7.54049778e-01 6.62146926e-01 1.25321209e+00 4.75356400e-01
-1.20989777e-01 -4.94292408e-01 -4.72628623e-01 -9.68283534e-01
-6.06269300e-01 6.60648763e-01 -5.61300851e-03 -3.27385366e-01
4.30168778e-01 -3.90406370e-01 -3.64553183e-02 -1.75279021e-01
-1.29662895e+00 3.45588386e-01 8.31819355e-01 6.60461128e-01
6.25743747e-01 -1.90080144e-02 3.80847454e-01 -1.42020297e+00
1.06088245e+00 -1.08889183e-02 -8.54121894e-03 4.12074178e-01
-5.73122323e-01 2.48846486e-01 1.21022856e+00 -4.92828220e-01
-1.22164857e+00 -2.53604144e-01 -5.55622280e-01 -1.19857546e-02
-7.32193291e-02 6.16822124e-01 -4.23247546e-01 7.47105181e-02
4.77387637e-01 -1.20808491e-02 -3.72169107e-01 -1.09273756e+00
6.02027416e-01 7.55240977e-01 2.21866176e-01 -3.66959631e-01
1.04138064e+00 3.30120385e-01 -2.10824385e-01 -1.05291271e+00
-1.16048038e+00 -2.42236197e-01 -8.59094560e-01 4.81420398e-01
4.51408088e-01 -7.57802784e-01 -4.57664765e-02 5.44109821e-01
-1.73301196e+00 -2.33097434e-01 -2.93425858e-01 3.12250733e-01
3.76906283e-02 8.32880437e-01 -6.07393444e-01 -5.31886816e-01
-8.35710585e-01 -9.10699785e-01 1.02767837e+00 2.65977502e-01
-4.90945548e-01 -8.47316682e-01 6.45732656e-02 1.03315897e-01
4.25509781e-01 -3.07123721e-01 1.22675002e+00 -6.17247164e-01
-2.21934021e-01 -3.05868179e-01 5.86422859e-03 6.81212962e-01
1.97238669e-01 2.50117898e-01 -7.59146810e-01 -1.03296697e-01
1.68374300e-01 1.40945107e-01 1.14143431e+00 3.51913869e-01
1.34834087e+00 -2.68659681e-01 -2.64311075e-01 8.16148818e-01
8.62857759e-01 -6.23162612e-02 1.02676058e+00 3.49498272e-01
1.06065226e+00 3.83150905e-01 6.19162977e-01 3.44861239e-01
4.51190084e-01 4.10601795e-01 -1.71226233e-01 -4.89450872e-01
-4.10198569e-01 -3.86975139e-01 3.87565702e-01 1.59512925e+00
1.41973376e-01 -3.57947767e-01 -6.59385860e-01 3.86757135e-01
-1.67887115e+00 -6.94813311e-01 -3.71070325e-01 1.94126070e+00
1.24837959e+00 3.47298265e-01 -4.55375582e-01 2.31187001e-01
8.09723377e-01 2.52108186e-01 -2.23042533e-01 -8.93717289e-01
-5.17889142e-01 6.99341357e-01 3.64626914e-01 1.00500095e+00
-1.18716216e+00 1.66777861e+00 6.17747593e+00 1.49922585e+00
-9.36956823e-01 1.64773107e-01 2.88238555e-01 -1.55873328e-01
-4.50197428e-01 7.61549622e-02 -1.16907954e+00 2.13209420e-01
6.92657173e-01 -6.54387355e-01 4.96587217e-01 5.55504084e-01
6.16904914e-01 -2.35861465e-02 -1.02412152e+00 1.02743447e+00
4.75600958e-01 -9.98305202e-01 8.09184849e-01 -5.69922686e-01
8.88779879e-01 -3.16173404e-01 -2.98379779e-01 8.97390723e-01
7.11387396e-02 -1.29179752e+00 4.45864916e-01 6.55601740e-01
6.33215308e-01 -9.69806433e-01 7.64774323e-01 7.51866579e-01
-9.77365017e-01 2.86873072e-01 -6.22196317e-01 -4.99830872e-01
3.29169244e-01 9.32425022e-01 -5.64370155e-01 6.59626842e-01
1.51468068e-01 7.62198150e-01 -8.60291541e-01 9.95505869e-01
-6.64123058e-01 5.94038963e-01 -1.70449942e-01 -3.55369672e-02
-1.12122484e-01 -2.20571265e-01 5.61284840e-01 1.65244305e+00
3.64274420e-02 8.38203803e-02 1.67702615e-01 4.28630054e-01
-3.03057998e-01 5.30673504e-01 -5.08247316e-01 1.18901934e-02
4.16851521e-01 1.18931103e+00 -2.48437136e-01 -7.64393687e-01
-2.56899744e-01 1.00022137e+00 4.56753373e-01 2.65297681e-01
-6.93008602e-01 -1.04153335e+00 5.87561309e-01 -1.78493381e-01
4.69125099e-02 -4.06606436e-01 -8.84245396e-01 -1.59649944e+00
5.94990999e-02 -1.18482614e+00 1.47357036e-03 -5.59586108e-01
-1.01746559e+00 6.07408941e-01 7.04412684e-02 -8.42709064e-01
4.24743630e-02 -3.63141030e-01 -6.85391307e-01 1.14068818e+00
-1.75252366e+00 -1.04761660e+00 -5.49184158e-02 2.98606455e-01
1.06091094e+00 -8.29979486e-04 9.01758075e-01 5.70496380e-01
-9.90989447e-01 9.23103034e-01 2.46077150e-01 4.01843758e-03
9.89425659e-01 -9.58528221e-01 9.76684391e-01 1.23159325e+00
-2.90584899e-02 1.09072542e+00 7.22761631e-01 -8.59290600e-01
-1.08658540e+00 -1.29904211e+00 1.70488417e+00 -3.92026484e-01
2.61522204e-01 -5.93637586e-01 -1.12512910e+00 6.23693645e-01
3.65212470e-01 -8.36557448e-01 4.70717132e-01 2.57018562e-02
-1.10496918e-03 -5.44585250e-02 -9.51138377e-01 1.18421483e+00
1.13444865e+00 -3.50995749e-01 -1.10272872e+00 4.87799466e-01
1.15423572e+00 -5.45863271e-01 -2.72418499e-01 4.30964828e-01
1.08157247e-01 -4.46155667e-01 6.78757966e-01 -7.35372126e-01
4.74929959e-01 -1.33507907e-01 1.14964232e-01 -1.56824660e+00
-5.14527082e-01 -1.12947989e+00 -1.14385895e-01 1.58333945e+00
4.26091880e-01 -4.40166473e-01 3.47846061e-01 4.10482377e-01
-6.19500518e-01 -6.99221075e-01 -5.63751161e-01 -7.02563524e-01
4.61591721e-01 -4.18158442e-01 8.10607672e-01 6.90706074e-01
-2.83698112e-01 7.01327741e-01 -6.21111810e-01 -2.73707062e-01
2.78011411e-01 -2.01614127e-01 9.70663965e-01 -1.03955960e+00
6.01961948e-02 -4.11217272e-01 2.43128270e-01 -1.42694378e+00
4.28368926e-01 -1.03058887e+00 7.16509148e-02 -1.70235968e+00
1.92654312e-01 -1.69169500e-01 3.36356163e-02 4.43041056e-01
-7.74282634e-01 -6.66268915e-02 4.79587644e-01 1.46838427e-01
-6.34865224e-01 8.99246514e-01 1.45534110e+00 -2.45341197e-01
-3.77480954e-01 -1.23689115e-01 -1.00714982e+00 8.84512365e-01
9.08765316e-01 -7.56343126e-01 -1.25170141e-01 -1.03945780e+00
3.08616702e-02 -7.12734938e-01 -3.74822587e-01 -8.09182525e-01
2.85116509e-02 -2.70298153e-01 -6.07310503e-05 -7.80550718e-01
1.44887909e-01 -3.25846076e-01 -2.80297250e-01 4.87331092e-01
-3.34144115e-01 2.81837761e-01 3.94414783e-01 -5.03242500e-02
-1.67340264e-01 -7.36989081e-01 7.78897524e-01 -2.42065832e-01
-4.91656989e-01 1.81902766e-01 -5.96199214e-01 2.79473662e-01
5.01827180e-01 1.56008989e-01 -3.10593486e-01 -1.92826346e-01
-3.40424895e-01 2.74583459e-01 4.18351144e-01 4.74701464e-01
5.00086248e-01 -9.77437556e-01 -8.93069625e-01 1.68413088e-01
-2.16066599e-01 2.48912498e-01 2.10727334e-01 4.75178748e-01
-7.31795967e-01 6.25863910e-01 2.82109141e-01 -8.02954957e-02
-1.40606034e+00 6.64274335e-01 6.57256767e-02 -6.14852309e-01
-4.59386230e-01 7.25070238e-01 1.46171257e-01 -7.06179321e-01
1.64257869e-01 -6.62373900e-01 -3.74457568e-01 -1.23178758e-01
8.30543518e-01 4.36609864e-01 3.20958376e-01 -4.72379804e-01
-9.38584059e-02 5.20023704e-01 -5.83730221e-01 1.74111634e-01
1.21832418e+00 -3.52134615e-01 -6.26004517e-01 4.17958826e-01
1.03600299e+00 3.99762839e-01 -3.20184320e-01 -3.69462878e-01
1.98918451e-02 -3.07603121e-01 -2.45830998e-01 -4.77614552e-01
-6.11433089e-01 1.17105961e+00 -3.74921970e-02 -9.73727927e-02
1.12174892e+00 -5.06848991e-01 1.22896004e+00 9.03187037e-01
-1.63609278e-03 -1.21760130e+00 -3.24460298e-01 1.38545287e+00
9.07364130e-01 -1.05192268e+00 3.72462094e-01 -8.18230629e-01
-5.97944736e-01 9.33833241e-01 5.66172957e-01 -2.23700702e-01
4.80106741e-01 1.74034923e-01 1.24886453e-01 3.03344935e-01
-8.08854580e-01 -3.33495229e-01 2.64492303e-01 4.43167478e-01
6.91414475e-01 1.11606345e-01 -1.00138247e+00 9.38187182e-01
-6.22547328e-01 8.97647142e-02 6.20504022e-01 8.41793478e-01
-6.02785230e-01 -1.54879248e+00 -3.14323634e-01 6.33056819e-01
-4.67678845e-01 -8.00859094e-01 -6.11567974e-01 7.22400725e-01
3.08874726e-01 1.11483824e+00 -4.36462939e-01 -4.54735041e-01
6.97783351e-01 1.11509643e-01 3.52592587e-01 -1.13755655e+00
-7.88424492e-01 -1.54931709e-01 1.44394308e-01 -5.05404808e-02
-3.36148255e-02 -3.40066105e-01 -1.19724619e+00 -7.82936573e-01
-4.02594030e-01 1.36822954e-01 4.26631391e-01 1.29487288e+00
4.62469965e-01 7.87596405e-01 4.51239496e-01 -1.06792605e+00
-4.52549368e-01 -1.38067269e+00 -2.76090533e-01 5.88888407e-01
2.12322831e-01 -4.40466970e-01 -2.28854373e-01 2.03505144e-01]
|
[11.02657413482666, 10.311452865600586]
|
5fe88cbf-ac0c-4967-bd21-676646882206
|
missing-values-and-imputation-in-healthcare
|
2304.11749
| null |
https://arxiv.org/abs/2304.11749v1
|
https://arxiv.org/pdf/2304.11749v1.pdf
|
Missing Values and Imputation in Healthcare Data: Can Interpretable Machine Learning Help?
|
Missing values are a fundamental problem in data science. Many datasets have missing values that must be properly handled because the way missing values are treated can have large impact on the resulting machine learning model. In medical applications, the consequences may affect healthcare decisions. There are many methods in the literature for dealing with missing values, including state-of-the-art methods which often depend on black-box models for imputation. In this work, we show how recent advances in interpretable machine learning provide a new perspective for understanding and tackling the missing value problem. We propose methods based on high-accuracy glass-box Explainable Boosting Machines (EBMs) that can help users (1) gain new insights on missingness mechanisms and better understand the causes of missingness, and (2) detect -- or even alleviate -- potential risks introduced by imputation algorithms. Experiments on real-world medical datasets illustrate the effectiveness of the proposed methods.
|
['Rich Caruana', 'Cynthia Rudin', 'Urszula Chajewska', 'Sarah Tan', 'Zhi Chen']
|
2023-04-23
| null | null | null | null |
['interpretable-machine-learning']
|
['methodology']
|
[ 5.38957119e-01 1.93995014e-01 -7.40795612e-01 -8.39988470e-01
-7.78887630e-01 3.93228084e-02 -1.68314055e-01 4.17062551e-01
1.18505679e-01 1.34975040e+00 5.01937211e-01 -4.56401080e-01
-3.65345150e-01 -8.26079488e-01 -8.81695032e-01 -8.41495275e-01
1.39824077e-01 4.49615598e-01 -5.65425158e-01 -7.43239969e-02
2.15526983e-01 -3.40798274e-02 -1.38078988e+00 8.94599915e-01
1.25527918e+00 5.73381841e-01 -1.75398752e-01 1.35218576e-01
-3.80269140e-01 9.47441995e-01 -4.37854648e-01 -4.34400469e-01
-3.13147493e-02 -5.64465582e-01 -4.21785712e-01 -3.63578707e-01
-2.58672208e-01 -2.27315784e-01 3.91424149e-01 7.77755558e-01
2.68845260e-01 -4.44405258e-01 7.49936044e-01 -1.35989523e+00
-5.78521311e-01 8.00059319e-01 -8.33459556e-01 -2.81955749e-01
2.34483376e-01 -1.48883060e-01 4.37011063e-01 -7.77540743e-01
3.68773758e-01 1.17524743e+00 9.50375974e-01 8.19822252e-01
-1.37388015e+00 -7.25042760e-01 -1.17526345e-01 2.46793523e-01
-8.93754840e-01 -4.44963366e-01 7.15153039e-01 -7.15636253e-01
4.20366168e-01 7.42720962e-01 3.26680392e-01 9.93918777e-01
5.28688014e-01 5.86758375e-01 1.27934802e+00 -5.02365053e-01
2.47060165e-01 7.25115897e-05 3.45994145e-01 2.08192378e-01
6.73176527e-01 1.19682193e-01 -4.61631656e-01 -8.13807726e-01
4.97974187e-01 6.83919191e-01 -7.28715211e-02 -2.07834482e-01
-1.43868375e+00 9.10720527e-01 1.20272271e-01 -3.52472253e-02
-4.97449994e-01 -6.48740167e-03 2.15764940e-01 2.04192176e-01
5.03434300e-01 5.65289520e-02 -9.56383049e-01 5.13695367e-02
-8.05300832e-01 2.88339913e-01 4.07430768e-01 5.72446287e-01
5.47717750e-01 -2.88501143e-01 -3.55399102e-01 7.94623375e-01
1.61962837e-01 3.49663675e-01 3.22077364e-01 -5.85377157e-01
5.14737189e-01 9.53133106e-01 4.96825606e-01 -8.48654866e-01
-6.14418030e-01 -2.51817316e-01 -1.46044970e+00 1.12299696e-01
6.94655240e-01 -1.37742698e-01 -9.29487407e-01 1.75150919e+00
5.26917696e-01 2.21721351e-01 -1.40662091e-02 6.77068651e-01
7.81668305e-01 2.20149666e-01 3.22570741e-01 -3.56170177e-01
1.38349199e+00 -4.45484698e-01 -1.09724128e+00 -4.66523618e-02
8.48173559e-01 -6.74599767e-01 7.54701734e-01 6.05886340e-01
-1.10154486e+00 -3.57082814e-01 -6.06743753e-01 -2.24240959e-01
-4.52437848e-02 2.28546169e-02 7.45921910e-01 7.43864655e-01
-2.74535686e-01 5.36877751e-01 -9.06273782e-01 3.96549888e-02
5.05126119e-01 4.39758450e-01 -2.37496600e-01 -2.08618060e-01
-1.12978041e+00 7.58046567e-01 1.05421662e-01 3.56599927e-01
-4.73422080e-01 -1.08709753e+00 -5.84359586e-01 -2.81232391e-02
3.21463376e-01 -1.17761815e+00 8.20956886e-01 -9.51483607e-01
-5.05136013e-01 6.74574196e-01 -6.99186683e-01 -3.40566546e-01
7.18318760e-01 -1.81896135e-01 -3.34335774e-01 -6.71315789e-01
7.53777474e-02 1.96983114e-01 5.80666184e-01 -1.16631818e+00
-4.49556559e-01 -8.84518087e-01 -6.63717926e-01 -3.56439918e-01
2.05927908e-01 -1.67000324e-01 4.71004069e-01 -1.11655128e+00
3.99816155e-01 -6.98305011e-01 -5.54387987e-01 -1.52010068e-01
-7.76033461e-01 1.75551921e-01 1.31788239e-01 -9.79547083e-01
1.27262795e+00 -1.75419223e+00 1.35450035e-01 1.07327133e-01
2.41524637e-01 -1.59406319e-01 2.27390975e-01 6.09176219e-01
-3.05667192e-01 1.68374062e-01 -6.37635112e-01 -1.82477668e-01
-4.51379001e-01 3.95779222e-01 -4.56728399e-01 3.37554961e-01
-4.71591577e-03 7.72135377e-01 -3.54209423e-01 -1.90078527e-01
3.57710451e-01 5.60758948e-01 -4.84950095e-01 2.70561516e-01
-9.31193084e-02 1.01118922e+00 -3.85152847e-01 7.91621864e-01
1.09286153e+00 -1.03397287e-01 2.50680685e-01 7.48935863e-02
1.54879600e-01 7.03791529e-02 -1.07821405e+00 1.27298737e+00
-2.75514901e-01 -9.15064216e-02 -1.02940938e-02 -1.19882512e+00
9.87272084e-01 1.37111545e-01 4.60939944e-01 -4.00591701e-01
-6.51464462e-02 1.40108198e-01 1.24206662e-01 -8.15834045e-01
-1.06027331e-02 -3.83132845e-01 9.88897458e-02 2.02997655e-01
-7.43894398e-01 7.04490244e-01 -3.80812377e-01 -3.08637410e-01
9.81806874e-01 -1.54277280e-01 6.46280885e-01 -2.17315063e-01
4.65715170e-01 1.17500201e-01 1.34740388e+00 9.20940638e-01
2.12910354e-01 8.37101221e-01 6.37622237e-01 -1.04935157e+00
-8.47224355e-01 -8.24164689e-01 -5.70112944e-01 7.22586274e-01
-2.06831634e-01 1.19885921e-01 -6.94263458e-01 -4.25034136e-01
5.40439010e-01 8.32365870e-01 -9.33296859e-01 -1.03762344e-01
-4.47130978e-01 -1.46091330e+00 2.52251029e-01 5.74117064e-01
-3.13798860e-02 -9.84439909e-01 -5.01988471e-01 4.89758849e-01
-7.05675960e-01 -6.45992041e-01 1.24567762e-01 9.71893743e-02
-1.58370674e+00 -1.20574939e+00 -5.20644009e-01 -2.36779600e-01
1.09166944e+00 1.06442317e-01 1.33294594e+00 5.51765442e-01
-3.83893102e-01 -4.40037012e-01 -3.74244988e-01 -7.71224499e-01
-6.76869154e-01 -9.86828282e-02 -1.32770479e-01 6.57706559e-02
6.04490697e-01 -3.36108476e-01 -6.49436772e-01 3.71129006e-01
-8.81101012e-01 5.06557763e-01 6.52873993e-01 1.21481955e+00
6.19199336e-01 -1.32175356e-01 1.24812376e+00 -1.55563271e+00
3.16850007e-01 -9.46918309e-01 -3.95910501e-01 4.34516013e-01
-8.15225482e-01 2.44327873e-01 3.59493703e-01 -7.75054842e-02
-1.17206681e+00 1.43724814e-01 -2.00603217e-01 2.69572645e-01
-3.48927200e-01 6.48991108e-01 -4.05205667e-01 3.71449322e-01
3.78672868e-01 9.52275023e-02 4.90698032e-03 -8.55678141e-01
1.58806294e-01 8.62368464e-01 2.82362431e-01 -4.11859602e-01
2.87505776e-01 7.51557887e-01 1.99716076e-01 -4.37189549e-01
-8.07385564e-01 -3.18426877e-01 -5.58095813e-01 3.19587916e-01
7.05576658e-01 -9.81470942e-01 -8.37834477e-01 3.21483821e-01
-1.02596152e+00 1.10336356e-01 4.53297012e-02 5.07869065e-01
-3.92522246e-01 1.06967621e-01 -3.54285449e-01 -1.12187457e+00
-3.78713816e-01 -1.05618501e+00 8.10464263e-01 -6.88592121e-02
-5.48396587e-01 -8.33179951e-01 1.21597372e-01 9.91247118e-01
3.39590192e-01 7.16321886e-01 1.59531307e+00 -3.96694422e-01
-4.06656623e-01 -3.16597521e-02 -1.32489458e-01 -1.05759256e-01
3.76990169e-01 -3.78279328e-01 -1.00121784e+00 -6.76731532e-03
1.28835976e-01 2.07229465e-01 9.74052906e-01 1.09481978e+00
1.55361450e+00 -6.08408272e-01 -4.52769011e-01 5.38038969e-01
1.26616836e+00 5.76892197e-02 9.81908619e-01 1.38219088e-01
5.32547653e-01 9.24857438e-01 6.17577136e-01 7.57395804e-01
5.67696035e-01 6.02060199e-01 6.47292137e-01 -5.18888295e-01
3.01394552e-01 -3.68199855e-01 -2.26821095e-01 4.80534136e-01
-6.93663657e-02 2.10761562e-01 -8.63009512e-01 7.11146176e-01
-2.46264362e+00 -9.30592954e-01 -9.19905007e-01 2.58995152e+00
1.07152057e+00 -1.69825003e-01 9.94957387e-02 3.70638937e-01
8.31547439e-01 -4.35830772e-01 -6.27006233e-01 -4.66443151e-01
-5.19624054e-02 -2.16862231e-01 3.74198109e-01 2.58752704e-01
-8.74101043e-01 1.18240722e-01 6.29475451e+00 4.45821434e-01
-5.49574018e-01 1.04662694e-01 1.15720725e+00 1.37799019e-02
-5.15296996e-01 7.61663392e-02 -6.14223421e-01 7.71633625e-01
7.01919794e-01 1.46553114e-01 1.71235576e-01 7.89547563e-01
7.81105876e-01 -3.66810650e-01 -1.15401316e+00 8.73796642e-01
-2.39122614e-01 -1.42983425e+00 -6.47154972e-02 2.44778574e-01
8.81747127e-01 -3.40253115e-01 -2.16437891e-01 -7.01242909e-02
1.11821055e-01 -1.29183137e+00 2.26888835e-01 9.15099800e-01
5.54521203e-01 -8.54937673e-01 1.11973107e+00 7.33496666e-01
-4.45277274e-01 -1.96321249e-01 -6.63173735e-01 -5.40072143e-01
8.87652636e-02 1.47627079e+00 -7.47593582e-01 6.09324455e-01
6.53669298e-01 4.10598755e-01 -2.96025246e-01 1.15520751e+00
-2.44575381e-01 8.59682262e-01 9.48587954e-02 3.71014535e-01
-5.31182587e-01 -2.73612767e-01 1.01885460e-01 7.38302767e-01
5.76148391e-01 2.02809274e-01 -2.67512292e-01 1.00699973e+00
-7.28591159e-02 1.19489562e-02 -5.65739810e-01 4.16753411e-01
2.48179376e-01 8.73457551e-01 -4.52217966e-01 -2.38420859e-01
-5.48370838e-01 4.30974513e-01 -9.23411995e-02 4.49993238e-02
-7.28988171e-01 1.34112179e-01 8.04300666e-01 4.65065867e-01
-1.94018200e-01 2.71656245e-01 -1.18905747e+00 -1.35597932e+00
1.39373958e-01 -1.08707261e+00 7.45468855e-01 -6.11294866e-01
-1.56352651e+00 8.50368515e-02 -1.50258243e-01 -1.10740721e+00
-2.13305518e-01 -3.98661457e-02 -3.47170562e-01 8.55862319e-01
-1.41625583e+00 -1.11087334e+00 -3.20226103e-01 5.10771811e-01
3.34152550e-01 1.07531406e-01 1.05137515e+00 2.66968876e-01
-4.98662412e-01 3.54517311e-01 6.09300792e-01 2.18077246e-02
7.44051039e-01 -9.58056509e-01 1.26571849e-01 4.75090951e-01
-2.99959332e-02 6.71073616e-01 8.22353303e-01 -9.61563110e-01
-1.26592386e+00 -1.04146135e+00 1.26034236e+00 -8.07218015e-01
-1.72236592e-01 -3.35534066e-01 -1.15622842e+00 5.68520546e-01
-2.91930974e-01 -1.60248905e-01 1.21544945e+00 6.95709229e-01
-1.54526949e-01 -2.39187971e-01 -1.70697284e+00 1.61043435e-01
6.88206911e-01 2.33274609e-01 -7.00221300e-01 1.80780977e-01
4.15243894e-01 -2.33766779e-01 -5.48925936e-01 6.67571127e-01
8.54327857e-01 -1.21647000e+00 9.93818223e-01 -1.21207666e+00
6.96794689e-01 -1.17695406e-01 -4.73310277e-02 -1.08913410e+00
-1.87834799e-01 -1.84806824e-01 -4.42924313e-02 1.09892690e+00
4.26771611e-01 -5.46701789e-01 9.70079780e-01 1.29700291e+00
1.84348315e-01 -8.37348104e-01 -1.06381011e+00 -4.46291208e-01
2.28764310e-01 -5.32064438e-01 1.14317536e+00 1.19555259e+00
-2.40919329e-02 -2.83747733e-01 -1.10833299e+00 8.89946520e-02
1.05201900e+00 2.95004487e-01 7.16476142e-01 -1.63250244e+00
-1.85894415e-01 3.52431655e-01 -1.42509416e-01 -8.62950087e-02
-1.77627996e-01 -6.19045198e-01 -4.67112511e-01 -1.60237241e+00
7.82838643e-01 -6.57517135e-01 -4.17799145e-01 8.30641210e-01
-4.58283842e-01 1.34113029e-01 -1.90660581e-01 3.19733888e-01
1.20244838e-01 3.39507014e-01 8.25780034e-01 -7.85703436e-02
-3.10772926e-01 4.35339570e-01 -1.07431078e+00 9.15595710e-01
8.69977355e-01 -1.16093826e+00 -1.35784253e-01 -6.67232811e-01
6.15739405e-01 6.07379735e-01 5.85754633e-01 -4.46545392e-01
-1.90589070e-01 -6.72699690e-01 7.91624010e-01 -7.21902728e-01
-1.21671008e-02 -1.07473159e+00 7.92911947e-01 8.58791113e-01
-4.38982815e-01 1.06761448e-01 -1.71542093e-02 5.59444666e-01
-1.01462439e-01 -2.18802884e-01 5.85574269e-01 -1.06277406e-01
3.81067507e-02 -1.50387838e-01 -1.15231618e-01 -5.17901853e-02
8.78166676e-01 3.35950553e-02 -4.07314152e-01 -4.37216938e-01
-9.66335595e-01 2.03621909e-01 2.18892246e-01 3.99119079e-01
6.48487747e-01 -1.36968780e+00 -1.09775400e+00 4.94165778e-01
2.30568305e-01 -2.12702259e-01 6.94971800e-01 9.38239813e-01
-1.59462988e-01 2.38017350e-01 -1.70253918e-01 -4.06488866e-01
-1.38070333e+00 6.19629920e-01 -9.77171659e-02 -3.10688943e-01
-4.18204278e-01 3.29716295e-01 3.39200318e-01 -5.65802813e-01
1.09188363e-01 -4.19533789e-01 1.36017129e-01 -5.92896193e-02
7.36452222e-01 7.57058799e-01 2.91933089e-01 -2.81493068e-01
-5.42641759e-01 4.45265144e-01 4.90884930e-02 4.10560071e-01
1.61469221e+00 -1.86313227e-01 -3.94914329e-01 4.91567731e-01
5.07818282e-01 -1.40560523e-01 -7.81196356e-01 5.10100387e-02
1.13945007e-01 -7.42879748e-01 -2.40631104e-01 -1.25797391e+00
-8.58256578e-01 1.07456827e+00 5.75471520e-01 1.18168825e-02
1.16528821e+00 -3.00351351e-01 6.97712362e-01 1.54947191e-02
4.37411487e-01 -7.23635554e-01 -7.83758283e-01 -3.50184023e-01
9.08311963e-01 -1.72762561e+00 1.73475429e-01 -4.56072181e-01
-6.36745870e-01 1.00180364e+00 1.21394441e-01 3.19897473e-01
5.73609352e-01 2.67890155e-01 2.22657934e-01 2.35909641e-01
-8.56671512e-01 2.41731599e-01 1.06124148e-01 7.10060120e-01
6.43630087e-01 3.58766556e-01 -7.13820696e-01 1.27242076e+00
6.65062815e-02 5.99562287e-01 6.35160804e-01 5.48349202e-01
-1.75443158e-01 -1.57944429e+00 -9.87489283e-01 8.31332207e-01
-8.22585106e-01 -1.94936067e-01 -2.95435846e-01 4.62092608e-01
3.92374396e-01 1.25686872e+00 -3.95616740e-01 -5.32754324e-02
1.10234156e-01 2.31852606e-01 2.00688645e-01 -3.18239272e-01
-5.25363207e-01 3.23203765e-02 -2.59619523e-02 -3.21094215e-01
-3.33173662e-01 -6.34933770e-01 -1.14933586e+00 -3.85491401e-01
-2.34819517e-01 1.60968974e-01 6.30952597e-01 9.33296025e-01
7.20281899e-01 4.42968667e-01 4.18487906e-01 -1.13706142e-01
-6.29397392e-01 -8.17870677e-01 -5.47743857e-01 6.64698422e-01
5.06478548e-01 -5.59950888e-01 -2.99013048e-01 1.53531596e-01]
|
[7.726368427276611, 4.930809497833252]
|
36509940-6547-420b-9826-11da1737f097
|
adversarial-open-domain-adaption-for-sketch
|
2104.05703
| null |
https://arxiv.org/abs/2104.05703v2
|
https://arxiv.org/pdf/2104.05703v2.pdf
|
Adversarial Open Domain Adaptation for Sketch-to-Photo Synthesis
|
In this paper, we explore open-domain sketch-to-photo translation, which aims to synthesize a realistic photo from a freehand sketch with its class label, even if the sketches of that class are missing in the training data. It is challenging due to the lack of training supervision and the large geometric distortion between the freehand sketch and photo domains. To synthesize the absent freehand sketches from photos, we propose a framework that jointly learns sketch-to-photo and photo-to-sketch generation. However, the generator trained from fake sketches might lead to unsatisfying results when dealing with sketches of missing classes, due to the domain gap between synthesized sketches and real ones. To alleviate this issue, we further propose a simple yet effective open-domain sampling and optimization strategy to "fool" the generator into treating fake sketches as real ones. Our method takes advantage of the learned sketch-to-photo and photo-to-sketch mapping of in-domain data and generalizes it to the open-domain classes. We validate our method on the Scribble and SketchyCOCO datasets. Compared with the recent competing methods, our approach shows impressive results in synthesizing realistic color, texture, and maintaining the geometric composition for various categories of open-domain sketches. Our code is available at https://github.com/Mukosame/AODA
|
['Jan P. Allebach', 'Xiaohui Shen', 'Yiheng Zhu', 'Xiao Yang', 'Ding Liu', 'Xiaoyu Xiang']
|
2021-04-12
| null | null | null | null |
['sketch-to-image-translation']
|
['computer-vision']
|
[ 3.44330490e-01 3.33319634e-01 5.77485599e-02 -1.76826611e-01
-9.93128955e-01 -1.11892688e+00 7.37327874e-01 -9.01592374e-01
4.38121200e-01 8.45871985e-01 -7.77219236e-02 -2.24251323e-03
3.46603543e-01 -9.27859783e-01 -9.41583693e-01 -6.42156899e-01
6.21927202e-01 6.15002692e-01 -1.05188653e-01 -1.93773359e-01
1.43479362e-01 5.60650587e-01 -1.47795391e+00 5.09797513e-01
9.39733446e-01 7.98900068e-01 4.89528738e-02 5.58943450e-01
-3.20346564e-01 2.52005011e-01 -7.29593992e-01 -7.69721866e-01
8.09555113e-01 -6.11878514e-01 -4.70258653e-01 3.82029474e-01
1.15652156e+00 -7.26352751e-01 -4.70500410e-01 1.00719273e+00
4.76015747e-01 -9.90773886e-02 8.89448643e-01 -1.67424643e+00
-1.14191246e+00 3.30764428e-02 -3.52203488e-01 -8.21597099e-01
3.92229348e-01 3.22603405e-01 7.36742556e-01 -1.11492050e+00
1.09984636e+00 1.30152476e+00 4.06733453e-01 1.01671362e+00
-1.29061735e+00 -1.08436275e+00 1.59151349e-02 -3.35958689e-01
-1.62979925e+00 -5.60927212e-01 1.11105418e+00 -4.31301624e-01
1.65000245e-01 2.14725912e-01 5.36372125e-01 1.64438224e+00
-3.93281043e-01 7.54313767e-01 1.04097283e+00 -3.11953485e-01
1.23039834e-01 3.28216821e-01 -7.65361130e-01 7.17614889e-01
6.86516538e-02 4.75984752e-01 -2.52414107e-01 -3.21681082e-01
1.32454586e+00 4.82527055e-02 -1.51681840e-01 -6.15280449e-01
-1.09080672e+00 5.99404633e-01 3.95828098e-01 -1.09617906e-02
-1.48062050e-01 1.29618570e-01 -1.25124365e-01 3.98486048e-01
4.97847170e-01 4.85147893e-01 -1.17325723e-01 1.32711917e-01
-1.06500733e+00 4.87441421e-01 8.23568225e-01 1.39309621e+00
9.68859017e-01 1.30489439e-01 1.34547437e-02 9.67104852e-01
-1.15477070e-01 8.29282641e-01 7.40897432e-02 -1.17105949e+00
6.16919696e-01 4.14708167e-01 3.87847245e-01 -9.67977762e-01
5.28646290e-01 2.32407629e-01 -7.92889237e-01 5.32428741e-01
6.61990047e-01 1.60735473e-02 -1.07347929e+00 1.53118503e+00
3.07040989e-01 1.51094422e-01 3.05439439e-02 1.10469818e+00
8.29837382e-01 7.29418516e-01 -3.52302760e-01 2.38525316e-01
1.02796185e+00 -1.08176172e+00 -5.35184503e-01 -2.34599516e-01
6.00569136e-02 -9.63702619e-01 1.36951625e+00 2.81581610e-01
-9.67328846e-01 -6.81209624e-01 -9.94071662e-01 -3.69010687e-01
-3.90861034e-01 5.15557110e-01 2.01468676e-01 5.49844921e-01
-7.34176397e-01 7.78282523e-01 -1.27328783e-01 -2.60644734e-01
8.14644396e-01 -5.13778068e-02 -5.26725769e-01 -4.93280232e-01
-9.38407302e-01 5.00674307e-01 -4.14699540e-02 -1.83043167e-01
-1.02554286e+00 -8.54220927e-01 -5.95911801e-01 -4.17527817e-02
5.36594212e-01 -7.60515392e-01 1.03076112e+00 -1.28852069e+00
-1.74996579e+00 9.34326172e-01 6.27465472e-02 2.98615038e-01
9.42953944e-01 1.22489370e-01 -2.22626641e-01 6.14864565e-02
-1.34665472e-02 1.11854291e+00 1.56864715e+00 -1.82840765e+00
-1.64730400e-01 -2.02930227e-01 8.69013369e-02 3.10021956e-02
-1.28837317e-01 -6.00314796e-01 -3.72255713e-01 -9.12233353e-01
-4.32318933e-02 -1.01717353e+00 6.54093847e-02 7.65092492e-01
-4.83124167e-01 4.04793732e-02 1.04478419e+00 -6.30563915e-01
6.19793534e-01 -2.31578422e+00 1.21520676e-01 1.39540002e-01
9.31581184e-02 4.25940990e-01 -5.72673023e-01 7.13946342e-01
-2.84696147e-02 1.05341315e-01 -3.52046013e-01 -5.00584543e-01
2.62462020e-01 4.68697965e-01 -9.54331338e-01 1.87685251e-01
4.03853625e-01 9.45306599e-01 -1.00977600e+00 -5.25807977e-01
1.49439126e-01 4.56895590e-01 -4.11090314e-01 5.23307323e-01
-5.66451252e-01 6.91708326e-01 -3.50811571e-01 8.57497096e-01
1.19115198e+00 -3.93619835e-02 1.90824360e-01 -1.52728036e-01
2.22930685e-01 -8.32497329e-02 -1.19987726e+00 1.99161911e+00
-4.43511099e-01 4.75019604e-01 1.64610848e-01 -4.70120132e-01
1.29147816e+00 2.69334018e-01 1.16681628e-01 -5.16409516e-01
-4.16085087e-02 4.28198665e-01 -4.44518656e-01 -3.76169741e-01
4.72221941e-01 -4.91222382e-01 -7.41349459e-02 6.26378000e-01
1.42255295e-02 -8.16038251e-01 -7.81716630e-02 2.07642496e-01
7.31211364e-01 6.35151505e-01 -1.60338497e-03 3.36919613e-02
2.83209682e-01 -9.51642767e-02 2.74623871e-01 4.51146930e-01
8.98897424e-02 1.17467475e+00 4.52042639e-01 -4.37201560e-01
-1.65087724e+00 -1.29708123e+00 1.68338239e-01 5.76636493e-01
1.72720879e-01 -2.35993773e-01 -7.77222931e-01 -9.04369891e-01
2.82807410e-01 4.93110627e-01 -4.94520634e-01 -5.07418700e-02
-4.03991818e-01 2.33007595e-01 8.30428779e-01 1.98128089e-01
2.66848892e-01 -1.05707324e+00 -2.43306905e-03 -1.66833892e-01
-2.03944236e-01 -1.15575588e+00 -8.54252636e-01 -7.04541445e-01
-6.51716948e-01 -1.08220387e+00 -1.17137957e+00 -8.47174942e-01
1.18937540e+00 3.07910144e-01 1.01939487e+00 2.45097116e-01
-3.23205441e-01 2.53331989e-01 -2.47881651e-01 -1.78918034e-01
-8.45230043e-01 -1.30940914e-01 -2.25850284e-01 2.85785884e-01
-2.33460844e-01 -6.44907117e-01 -6.54326081e-01 5.34322441e-01
-1.10949266e+00 3.46025437e-01 5.67857802e-01 9.22080398e-01
4.11181867e-01 -2.42328584e-01 4.72173810e-01 -8.53229046e-01
4.27064598e-01 -2.93690890e-01 -7.01573610e-01 3.28379333e-01
-3.62599164e-01 6.94601536e-02 9.37925935e-01 -7.99483716e-01
-1.14763427e+00 2.95514792e-01 1.09947570e-01 -1.01842260e+00
-2.79775321e-01 -4.22456592e-01 -4.89809483e-01 -2.03933373e-01
7.33971119e-01 1.72044367e-01 1.91045821e-01 -5.57319939e-01
8.01651597e-01 6.56677127e-01 5.35726309e-01 -9.88105297e-01
1.37830234e+00 8.30601931e-01 -9.43572149e-02 -7.36236572e-01
-4.29857671e-01 4.61183861e-02 -4.92898017e-01 -3.86049263e-02
3.66959631e-01 -8.36265564e-01 -2.70164460e-01 4.94107097e-01
-1.38705504e+00 -4.42558318e-01 -5.02914190e-01 -2.43361726e-01
-7.02446163e-01 5.40468872e-01 -3.41228485e-01 -7.19191134e-01
-1.58364937e-01 -1.07378471e+00 1.59914172e+00 -1.86633840e-02
-1.01392064e-02 -6.51048005e-01 -1.26122190e-02 1.72314376e-01
1.51760191e-01 2.87210524e-01 6.99364543e-01 -1.52223753e-02
-1.05917728e+00 -2.31231317e-01 -5.16183496e-01 5.92541158e-01
2.05967933e-01 2.26035997e-01 -1.03430331e+00 -2.27104396e-01
-5.00806212e-01 -6.40258491e-01 5.07877290e-01 -2.56798059e-01
1.14900970e+00 -5.11479616e-01 -2.47631386e-01 7.04123378e-01
1.45124173e+00 -7.14761466e-02 9.22014415e-01 -4.35013205e-01
8.73775840e-01 7.16127872e-01 5.64718008e-01 4.20225024e-01
1.15404256e-01 7.78550327e-01 2.65149057e-01 -1.82391450e-01
-6.29128098e-01 -1.13374436e+00 3.27351362e-01 2.71939993e-01
1.92065299e-01 -3.05522084e-01 -4.19890493e-01 6.45474195e-01
-1.55533147e+00 -9.40643430e-01 2.38193572e-02 2.23419762e+00
8.41997445e-01 -3.87260735e-01 6.57597184e-02 -1.09261602e-01
8.72597277e-01 1.56561568e-01 -5.43874025e-01 -2.66622603e-01
-1.55559242e-01 2.74526507e-01 2.21492827e-01 5.28431535e-01
-7.12787747e-01 1.18835521e+00 5.43667364e+00 1.17515445e+00
-1.08250654e+00 9.94865447e-02 3.87176365e-01 -3.44834402e-02
-6.54024661e-01 2.96874344e-01 -4.56939906e-01 5.81925035e-01
3.01451474e-01 7.45684728e-02 9.29258347e-01 7.36850560e-01
-1.71998039e-01 1.76077604e-01 -1.37516999e+00 1.19838262e+00
1.43909439e-01 -1.54791617e+00 3.98283988e-01 1.21312208e-01
1.10017002e+00 -6.42541707e-01 6.57484457e-02 1.17059916e-01
2.48030201e-01 -9.80026603e-01 8.28252137e-01 4.56770092e-01
1.58697307e+00 -3.99029762e-01 6.59945980e-02 4.51320708e-01
-9.32082891e-01 1.55300558e-01 -4.69020069e-01 -3.57321203e-02
2.51415018e-02 5.02698362e-01 -8.44878793e-01 5.90828717e-01
1.56218633e-01 6.27812326e-01 -2.18081459e-01 5.67936480e-01
-4.93421793e-01 8.49718507e-03 -2.28350699e-01 3.46110195e-01
-1.07065819e-01 -5.36280394e-01 4.75470722e-01 7.07991838e-01
6.77313745e-01 1.19611427e-01 6.04046546e-02 1.62410009e+00
-3.55247825e-01 -2.40282565e-01 -1.17547596e+00 -3.31578702e-01
6.98612034e-01 1.16118801e+00 -3.92710477e-01 -5.81050813e-01
-9.93811786e-02 1.60410440e+00 2.71724135e-01 5.32488644e-01
-8.60596001e-01 -4.72604364e-01 8.07735503e-01 2.98097521e-01
2.68656164e-01 -3.89104411e-02 -3.02763194e-01 -1.40398848e+00
3.46822798e-01 -7.64123201e-01 -1.96579546e-02 -1.07629192e+00
-1.62610114e+00 5.66526294e-01 -2.12915719e-01 -1.63118684e+00
-1.93804741e-01 -4.84435290e-01 -5.87482095e-01 9.87753153e-01
-1.37500370e+00 -1.55041873e+00 -4.44365978e-01 5.71027577e-01
5.11415839e-01 -8.10214430e-02 8.61246943e-01 4.37689096e-01
-6.74880370e-02 8.34448218e-01 1.21352658e-01 1.19766332e-01
1.02987874e+00 -9.48354602e-01 7.17990994e-01 5.03283918e-01
2.48202249e-01 2.72812486e-01 3.67755562e-01 -6.75856769e-01
-1.41253006e+00 -1.00542533e+00 7.00226605e-01 -6.74762607e-01
4.15586054e-01 -7.97562897e-01 -7.78068006e-01 4.90098715e-01
5.55885583e-02 2.83257335e-01 1.59644410e-01 -5.07837296e-01
-7.18157232e-01 -5.42525984e-02 -1.43757963e+00 8.14706981e-01
1.42408085e+00 -7.23496914e-01 -3.76886249e-01 2.99275935e-01
4.86038685e-01 -6.30664945e-01 -7.11027741e-01 -5.89101613e-02
9.32149231e-01 -8.73786032e-01 1.22242200e+00 -4.54679906e-01
9.25924420e-01 -3.02419305e-01 -4.50602286e-02 -1.51581836e+00
3.28866988e-02 -8.95581126e-01 1.25156149e-01 1.38058734e+00
1.75804272e-01 -4.49778140e-01 1.00102758e+00 5.31575978e-01
3.37506793e-02 -3.85940760e-01 -7.65359819e-01 -1.03908777e+00
3.52676988e-01 -3.10022086e-02 1.02918589e+00 1.03517020e+00
-3.38074654e-01 1.40068859e-01 -8.50747287e-01 2.56155785e-02
7.21572936e-01 6.16092980e-01 1.34559202e+00 -1.04867864e+00
-2.34999642e-01 -2.25477517e-01 -8.22543353e-03 -1.17998779e+00
3.19433123e-01 -8.76307130e-01 -2.34081876e-04 -1.26099741e+00
-5.33538498e-02 -8.56964529e-01 4.94630456e-01 4.29551125e-01
9.01838765e-02 6.09932005e-01 5.92655718e-01 3.07779104e-01
-7.85916001e-02 7.17642128e-01 1.92265534e+00 -2.38169834e-01
3.13875964e-04 -2.11066604e-01 -5.66444993e-01 4.70457494e-01
5.61235309e-01 -5.64833462e-01 -4.03150976e-01 -4.40794170e-01
-3.69212851e-02 3.59942675e-01 7.98135817e-01 -7.33180523e-01
-2.67421246e-01 -4.33054060e-01 2.79608905e-01 -3.40124816e-01
7.86827862e-01 -1.00233459e+00 5.35533786e-01 1.81026101e-01
-2.03252882e-01 -4.39989239e-01 -1.83339924e-01 4.93773997e-01
-3.84606887e-03 -4.03395332e-02 8.76545668e-01 -3.01849365e-01
-3.41615170e-01 5.13637424e-01 2.66625434e-01 7.57241622e-02
8.97223055e-01 -4.06935275e-01 -6.81528032e-01 -5.77010632e-01
-5.35218060e-01 -2.09718972e-01 1.13380229e+00 6.70272589e-01
8.34980667e-01 -1.84436512e+00 -6.50246561e-01 6.62268698e-01
3.55099976e-01 1.53542995e-01 3.63660842e-01 1.26565501e-01
-5.11255503e-01 1.59644008e-01 -4.04922336e-01 -2.65210927e-01
-1.01771820e+00 7.96692669e-01 1.21701024e-01 1.24844015e-01
-6.30492806e-01 5.57268381e-01 6.46802425e-01 -8.25197697e-01
6.72118217e-02 -1.92407042e-01 8.18209469e-01 -4.01466399e-01
3.30700397e-01 3.47744226e-01 -2.31938258e-01 -4.71184671e-01
1.21444566e-02 6.84687018e-01 3.48969638e-01 -1.05982862e-01
9.89099324e-01 1.79025054e-01 7.26774111e-02 4.64631729e-02
1.31874943e+00 2.73085892e-01 -1.77589023e+00 -1.29166961e-01
-6.68250561e-01 -1.18591344e+00 -5.29126585e-01 -9.19737458e-01
-1.05693913e+00 1.07295275e+00 2.33568266e-01 -2.03871235e-01
7.87703753e-01 1.74093321e-02 1.08133733e+00 1.70427069e-01
5.40209830e-01 -9.27072167e-01 4.19925958e-01 1.09463580e-01
1.34878230e+00 -1.18854213e+00 -4.72425558e-02 -6.47764087e-01
-7.42494822e-01 1.11662567e+00 6.72498047e-01 -4.81303394e-01
3.21686447e-01 5.39839007e-02 1.34450078e-01 -1.68363936e-02
-4.86822039e-01 1.16258480e-01 2.52998978e-01 8.02858293e-01
-1.82009503e-01 2.85289288e-01 3.29924375e-02 4.34258401e-01
-1.73450142e-01 2.87636191e-01 5.86012423e-01 7.07080960e-01
2.99003031e-02 -1.60951602e+00 -6.14047170e-01 4.62262556e-02
2.91103363e-01 1.61377564e-02 -8.50256741e-01 7.67971277e-01
3.91374558e-01 6.12586915e-01 3.15913372e-02 -3.31824243e-01
2.83553004e-01 9.72429141e-02 8.16455305e-01 -5.36614478e-01
-1.18247956e-01 -1.32897571e-01 5.94572686e-02 -5.45939863e-01
-4.92901541e-02 -3.61402899e-01 -8.40643764e-01 -5.08625150e-01
-5.10046370e-02 -1.29046828e-01 5.56393504e-01 4.61271048e-01
6.64636850e-01 -7.70265162e-02 8.17224443e-01 -1.22911692e+00
-7.20051885e-01 -6.34025931e-01 -7.23092079e-01 9.04177487e-01
3.45946878e-01 -7.30152786e-01 -2.66778678e-01 1.31635189e-01]
|
[11.968707084655762, 0.007284647785127163]
|
90b40aee-a8ba-45de-94bb-e92f26a61531
|
shortest-path-networks-for-graph-property
|
2206.01003
| null |
https://arxiv.org/abs/2206.01003v4
|
https://arxiv.org/pdf/2206.01003v4.pdf
|
Shortest Path Networks for Graph Property Prediction
|
Most graph neural network models rely on a particular message passing paradigm, where the idea is to iteratively propagate node representations of a graph to each node in the direct neighborhood. While very prominent, this paradigm leads to information propagation bottlenecks, as information is repeatedly compressed at intermediary node representations, which causes loss of information, making it practically impossible to gather meaningful signals from distant nodes. To address this, we propose shortest path message passing neural networks, where the node representations of a graph are propagated to each node in the shortest path neighborhoods. In this setting, nodes can directly communicate between each other even if they are not neighbors, breaking the information bottleneck and hence leading to more adequately learned representations. Our framework generalizes message passing neural networks, resulting in a class of more expressive models, including some recent state-of-the-art models. We verify the capacity of a basic model of this framework on dedicated synthetic experiments, and on real-world graph classification and regression benchmarks, and obtain state-of-the art results.
|
['İsmail İlkan Ceylan', 'Radoslav Dimitrov', 'Ralph Abboud']
|
2022-06-02
| null | null | null | null |
['graph-property-prediction']
|
['graphs']
|
[ 3.31778765e-01 6.69537187e-01 -4.39004064e-01 -1.78197652e-01
-3.48088816e-02 -3.52463067e-01 7.61926949e-01 1.00266933e+00
-3.29300702e-01 6.20993316e-01 -5.04730083e-02 -4.31254804e-01
-2.17356831e-01 -1.48111463e+00 -9.67520356e-01 -5.38795531e-01
-8.58781099e-01 4.23283190e-01 4.80040431e-01 -3.83797318e-01
4.71527576e-02 5.28495431e-01 -1.11337888e+00 8.47684368e-02
3.22346002e-01 6.01644158e-01 -1.07227124e-01 8.14873397e-01
-3.39123249e-01 9.60035205e-01 -4.76194948e-01 -5.35262167e-01
5.95642924e-02 -2.86294192e-01 -1.08895004e+00 -2.55538434e-01
3.01311076e-01 -3.46660793e-01 -9.59016085e-01 9.90924060e-01
-1.45641521e-01 1.25052631e-01 5.28701842e-01 -1.40856850e+00
-9.19783235e-01 1.35888124e+00 -6.01678729e-01 9.33745205e-02
1.59203321e-01 -1.59744129e-01 1.42531097e+00 -3.49420011e-01
7.95387328e-01 1.15852594e+00 8.41197848e-01 3.59077126e-01
-1.54495430e+00 -4.25254226e-01 7.69954622e-01 8.05708170e-02
-1.19789350e+00 -2.14154363e-01 9.20076668e-01 -2.09025275e-02
1.03807056e+00 1.05268426e-01 7.85203755e-01 9.90825951e-01
2.38649487e-01 6.89183354e-01 3.78472120e-01 -1.96527377e-01
1.99105233e-01 -7.56399259e-02 6.33274138e-01 1.01030684e+00
4.92259413e-01 -4.42087501e-02 -4.85246539e-01 -1.35644001e-03
6.93882883e-01 3.92063707e-01 -4.69278783e-01 -3.84547681e-01
-1.12873375e+00 1.08420432e+00 1.49477899e+00 5.23648798e-01
-5.06576180e-01 7.15585470e-01 3.59050542e-01 7.29566276e-01
5.45302927e-01 1.54410213e-01 -1.46051064e-01 3.74785721e-01
-6.15194559e-01 2.47580688e-02 1.15194678e+00 8.19567025e-01
1.14459431e+00 -2.24403903e-01 -1.31505683e-01 4.23470736e-01
5.80906630e-01 -8.23206678e-02 9.96502191e-02 -5.79036653e-01
5.22284091e-01 9.69137073e-01 -4.52295095e-01 -1.59801745e+00
-6.36992395e-01 -7.28803158e-01 -1.29419923e+00 4.52985913e-02
3.87979597e-01 3.88555601e-02 -8.32725585e-01 1.97268367e+00
2.37962622e-02 3.18744034e-01 -1.82840563e-02 6.74959958e-01
8.11035991e-01 9.27011490e-01 1.36313969e-02 -3.71140912e-02
8.15937161e-01 -9.71683264e-01 -1.62626013e-01 -3.40231001e-01
7.83709049e-01 6.82439730e-02 5.47057211e-01 1.00937180e-01
-1.12424743e+00 -3.81054759e-01 -1.06657851e+00 1.67251397e-02
-6.16754472e-01 -6.19951904e-01 1.00968134e+00 3.00244123e-01
-1.54439175e+00 1.05603945e+00 -9.84313369e-01 -4.83825386e-01
5.43121696e-01 3.64026099e-01 -5.42670488e-01 -1.19830348e-01
-1.16760457e+00 3.19146276e-01 4.54964101e-01 1.68065175e-01
-1.19060040e+00 -6.68105960e-01 -1.00715482e+00 3.75613540e-01
2.50305593e-01 -8.53414416e-01 1.01680183e+00 -1.01098561e+00
-1.09391367e+00 5.61746061e-01 -1.23465106e-01 -8.05073559e-01
3.63282621e-01 1.60523310e-01 -3.60162593e-02 2.12813437e-01
-3.60309064e-01 7.43982852e-01 7.31897771e-01 -1.18228161e+00
-2.73367882e-01 -3.44910562e-01 5.41777253e-01 -8.26399177e-02
-5.95380783e-01 -3.07871461e-01 -4.44709152e-01 -3.76220971e-01
2.63555169e-01 -7.44597077e-01 -5.64515769e-01 3.18916649e-01
-7.57462680e-01 -3.95949155e-01 6.08808398e-01 -1.51316896e-01
1.03727520e+00 -1.84556293e+00 3.37817729e-01 4.05120105e-01
1.04116178e+00 -1.34496108e-01 -4.72708315e-01 8.11643124e-01
-4.25230861e-02 4.25346315e-01 -2.51753002e-01 -6.17667913e-01
2.15084217e-02 4.32458162e-01 -2.78732717e-01 6.38096929e-01
3.06732595e-01 1.05065775e+00 -1.09420764e+00 -1.98712245e-01
-3.87639366e-02 6.24217629e-01 -6.45145595e-01 7.90614262e-02
-2.90126890e-01 2.25776490e-02 -4.94307727e-01 1.05971061e-01
4.94085848e-01 -8.81993949e-01 8.51879269e-02 1.68907821e-01
4.32997018e-01 5.42790353e-01 -9.51326072e-01 1.64633036e+00
-3.71860176e-01 7.31652081e-01 8.83055255e-02 -1.43993247e+00
8.41618359e-01 4.93590208e-03 3.84458005e-01 -5.16928554e-01
-2.71798521e-02 -2.48116836e-01 -1.81263998e-01 1.23960905e-01
5.13881087e-01 2.33548567e-01 2.33325571e-01 6.12138867e-01
1.03143997e-01 2.71026969e-01 3.58039260e-01 8.50478470e-01
1.47236836e+00 -4.28759009e-01 -1.48355088e-03 -7.45019391e-02
2.57387370e-01 -3.64592254e-01 1.80365294e-01 1.25934851e+00
-1.71690322e-02 5.64720869e-01 9.64526534e-01 -5.48710465e-01
-7.06877470e-01 -1.23179257e+00 3.08869600e-01 1.32304692e+00
9.87057537e-02 -6.92629755e-01 -5.63574612e-01 -7.72840500e-01
1.99631732e-02 1.79494873e-01 -9.70963657e-01 -4.53158110e-01
-5.36086500e-01 -5.67912638e-01 5.03466189e-01 4.48475122e-01
1.54479176e-01 -8.44023466e-01 -2.37827733e-01 5.43610156e-01
1.98669583e-01 -8.00233722e-01 -4.86881658e-03 2.17466518e-01
-1.10664952e+00 -1.11807704e+00 -7.73999393e-01 -8.32725286e-01
8.70518684e-01 2.95304358e-01 1.50544345e+00 8.16798091e-01
-3.14657055e-02 9.22490731e-02 -2.73790836e-01 4.48144972e-02
-7.12415755e-01 5.78212619e-01 -3.76989365e-01 1.10452242e-01
-3.23310718e-02 -8.42873633e-01 -6.55812263e-01 -4.28650752e-02
-9.26060617e-01 1.27294123e-01 5.75593531e-01 7.21131265e-01
2.61149704e-01 -3.68209003e-04 6.15035892e-01 -1.23322093e+00
5.69929540e-01 -9.69934225e-01 -6.19456887e-01 1.50869980e-01
-5.44301271e-01 2.14162514e-01 1.06140506e+00 -2.20377028e-01
-3.22654307e-01 -2.28196755e-01 -6.61968663e-02 -9.34062153e-02
-7.93526247e-02 9.17189121e-01 3.01804900e-01 -2.27386370e-01
7.32809603e-01 2.82591850e-01 -1.23691261e-01 -3.51768315e-01
6.15048766e-01 1.08980969e-01 4.16996717e-01 -2.83051550e-01
7.52385139e-01 5.37824929e-01 2.99310684e-01 -9.45339143e-01
-4.50738013e-01 -5.15190735e-02 -5.27966261e-01 1.74729712e-02
3.15081626e-01 -5.76177537e-01 -8.36729467e-01 4.64506716e-01
-1.37783074e+00 -3.37455124e-01 -2.13689476e-01 2.41881728e-01
-8.59089717e-02 3.20670336e-01 -1.13290918e+00 -5.12432218e-01
1.88872740e-02 -9.10635054e-01 5.08865118e-01 1.94474488e-01
3.80648933e-02 -1.37217271e+00 7.90469125e-02 -4.45274830e-01
5.98142385e-01 1.57007262e-01 1.25887680e+00 -8.09445202e-01
-8.38836372e-01 -3.52043599e-01 -5.25353193e-01 -3.22145708e-02
-1.58010840e-01 4.82616387e-02 -7.16569781e-01 -5.09349227e-01
-5.32947361e-01 -2.02613503e-01 1.48949182e+00 1.59214526e-01
9.91730630e-01 -5.32202363e-01 -7.22845554e-01 6.82285964e-01
1.35213757e+00 -5.21719277e-01 4.71068889e-01 -5.86162582e-02
7.15413094e-01 4.56440479e-01 -5.18889666e-01 1.59644485e-01
6.07848763e-01 2.24351749e-01 8.75810862e-01 -2.13017538e-01
-2.05651209e-01 -6.17717326e-01 2.08378285e-01 9.80222166e-01
1.58021823e-02 -7.99890161e-01 -7.46313453e-01 6.63094640e-01
-1.71580541e+00 -5.38371801e-01 -2.19505519e-01 2.01796293e+00
5.71299791e-01 2.25677013e-01 -8.57003685e-03 2.37352654e-01
5.25304854e-01 7.64974833e-01 -4.16829467e-01 -3.40267301e-01
1.00842901e-01 1.25096470e-01 5.17757416e-01 5.52082241e-01
-9.00597990e-01 7.73642421e-01 6.83374119e+00 2.37476677e-01
-1.08290195e+00 -9.34906956e-03 7.06826329e-01 1.60179794e-01
-6.06855989e-01 1.02490865e-01 -5.09394705e-01 1.61641940e-01
1.08695757e+00 -3.17120075e-01 7.04371452e-01 6.27581775e-01
-1.88954875e-01 1.05712131e-01 -1.56233561e+00 5.04841447e-01
-1.41674668e-01 -1.62581861e+00 4.10100132e-01 -1.24010602e-02
5.40879250e-01 5.71520329e-01 -2.20333427e-01 2.84407616e-01
7.62414992e-01 -1.20546234e+00 2.70367235e-01 3.70726377e-01
3.07323933e-01 -6.66847765e-01 3.50982845e-01 3.74734491e-01
-1.31135345e+00 -4.93849963e-02 -6.46704972e-01 -3.34743291e-01
-1.60857085e-02 8.03500235e-01 -8.11466098e-01 7.54404604e-01
3.53563040e-01 1.01198769e+00 -6.59746468e-01 9.44334328e-01
-3.30873489e-01 7.56339550e-01 -5.21138549e-01 -2.87151068e-01
5.18814206e-01 1.62074223e-01 6.29302979e-01 1.22195208e+00
1.00722581e-01 -2.73545623e-01 2.54947394e-01 1.32786930e+00
-8.35722387e-01 -1.09408110e-01 -9.66914058e-01 -2.29064494e-01
5.22544503e-01 1.05487120e+00 -9.46877658e-01 -2.16303572e-01
-5.73585927e-01 9.15893853e-01 9.81274724e-01 6.95507884e-01
-4.71821994e-01 -6.24652267e-01 4.78829980e-01 5.12634702e-02
1.84311360e-01 -3.59783262e-01 6.05440363e-02 -8.49690080e-01
-6.62001371e-02 -4.51434463e-01 5.38661957e-01 -2.38781676e-01
-1.28389239e+00 6.16863489e-01 -2.89076447e-01 -6.22506142e-01
-1.11248769e-01 -4.92110163e-01 -9.10415590e-01 6.16775453e-01
-1.71406686e+00 -1.04845691e+00 -3.05955052e-01 6.42944038e-01
8.54660943e-02 1.55720159e-01 8.75592649e-01 1.96667477e-01
-4.08707023e-01 6.40636086e-01 -6.36710674e-02 4.82482642e-01
-1.95007096e-03 -1.22550070e+00 9.87010658e-01 7.53587544e-01
6.85819924e-01 7.79093027e-01 3.68905425e-01 -4.89227802e-01
-1.74448180e+00 -1.18112373e+00 7.85391331e-01 -8.94922614e-02
8.55984449e-01 -5.64490020e-01 -1.19343019e+00 1.01411152e+00
8.41798261e-02 2.21443459e-01 1.30252466e-01 4.10907269e-01
-6.34419501e-01 -1.60159677e-01 -6.93814754e-01 5.56908071e-01
1.25838578e+00 -5.72382808e-01 -2.14230865e-01 3.63578379e-01
1.03410554e+00 -1.09716114e-02 -8.56212199e-01 5.60448244e-02
2.57461399e-01 -9.26227272e-01 9.91434038e-01 -8.93008351e-01
6.38641477e-01 -1.21283576e-01 2.03016371e-01 -1.76203585e+00
-4.03972834e-01 -7.44486094e-01 -3.73677015e-01 9.28202927e-01
8.66870880e-01 -8.50302935e-01 9.28938627e-01 2.28030980e-01
1.23743759e-02 -6.48848951e-01 -8.57189894e-01 -5.56336939e-01
3.08720171e-01 -4.78546649e-01 5.96149206e-01 9.25244749e-01
1.47401258e-01 4.69765306e-01 -1.12430900e-01 3.34837317e-01
6.83848321e-01 5.78642152e-02 5.94385803e-01 -1.46216834e+00
-3.97922575e-01 -7.90435076e-01 -6.20254874e-01 -1.56172693e+00
2.71203905e-01 -1.40867722e+00 -5.90752773e-02 -1.83610439e+00
2.05152705e-01 -6.46677077e-01 -6.57242477e-01 5.92493594e-01
6.13595806e-02 2.39701614e-01 2.27674425e-01 -3.23269982e-03
-5.76849759e-01 4.39748466e-01 1.13414550e+00 -5.74633062e-01
-1.75460786e-01 -7.00390637e-02 -8.95965040e-01 6.55555844e-01
5.33920586e-01 -7.63624310e-01 -6.04332924e-01 -8.75872195e-01
7.41500318e-01 1.12371042e-01 7.11257696e-01 -9.48486865e-01
6.87485695e-01 1.58132061e-01 2.88979530e-01 -3.12542647e-01
2.56761253e-01 -6.97074711e-01 -4.76220846e-02 7.10310340e-01
-8.53735447e-01 3.44108487e-03 -1.37466818e-01 8.78605545e-01
-3.90289128e-02 2.77880579e-02 5.84608734e-01 -1.57538027e-01
-4.48275298e-01 7.14451015e-01 -2.18577608e-01 -3.94693241e-02
7.53721237e-01 7.73228100e-03 -6.44664764e-01 -7.07813025e-01
-7.01400340e-01 3.11663181e-01 4.50818926e-01 4.21983123e-01
6.50335193e-01 -1.01793873e+00 -1.06538999e+00 4.81693655e-01
2.57922858e-01 1.81358352e-01 1.75059438e-02 7.24393845e-01
-4.77005929e-01 2.51989603e-01 1.61362424e-01 -6.39787495e-01
-6.80557907e-01 6.93474889e-01 4.39304560e-01 -2.64568716e-01
-1.11827552e+00 8.81897390e-01 3.21261704e-01 -3.22099000e-01
4.77300137e-01 -5.96494019e-01 -1.51112661e-01 -3.05880845e-01
6.76587820e-01 2.09458813e-01 3.66118997e-02 -3.36262733e-01
-1.97887033e-01 2.68220365e-01 -4.16827738e-01 2.15916291e-01
1.16804659e+00 -8.75576511e-02 -4.01300132e-01 4.42897290e-01
1.51220095e+00 -4.11799848e-01 -9.28826153e-01 -4.71793801e-01
1.26782641e-01 -3.61563474e-01 1.12268776e-01 -3.64203840e-01
-1.42454302e+00 1.01601505e+00 -4.29209461e-03 8.54008675e-01
8.37844312e-01 2.88240790e-01 7.95834780e-01 9.20394242e-01
4.35281038e-01 -5.12560785e-01 -1.00913703e-01 6.59970462e-01
6.57131135e-01 -8.17779243e-01 -2.48032883e-01 -3.85471255e-01
9.72103104e-02 1.11058533e+00 2.08311945e-01 -4.53867912e-01
8.59971821e-01 1.61942646e-01 -3.95420283e-01 -3.37202489e-01
-1.11656106e+00 2.14793772e-01 -1.34750426e-01 6.88572943e-01
3.87510657e-01 6.22926764e-02 1.93720870e-02 3.55232447e-01
8.19270834e-02 -8.05163607e-02 6.36076570e-01 6.41629159e-01
-4.29264337e-01 -8.87321353e-01 1.97749168e-01 4.73121375e-01
-2.74077982e-01 -2.43904293e-01 -5.92220604e-01 7.64224410e-01
-5.41565716e-01 9.58898127e-01 3.26290816e-01 -2.66554415e-01
3.13642668e-03 -2.93397337e-01 3.21629375e-01 -5.14312387e-01
-5.28003156e-01 -5.83758116e-01 3.51159126e-02 -6.86365128e-01
-1.12359591e-01 4.59287092e-02 -1.35459852e+00 -7.48799682e-01
-1.77472651e-01 2.10730255e-01 6.15920007e-01 6.73011422e-01
4.35304254e-01 7.39627302e-01 7.74602234e-01 -8.51899207e-01
-4.33996797e-01 -8.37954640e-01 -5.51185012e-01 4.31707323e-01
8.22291613e-01 -1.96282536e-01 -3.47530901e-01 -4.00416464e-01]
|
[6.959756851196289, 6.200713634490967]
|
8d347d91-4038-47ac-a705-636fa4e8f4b3
|
udapter-typology-based-language-adapters-for
| null | null |
https://aclanthology.org/2022.cl-3.3
|
https://aclanthology.org/2022.cl-3.3.pdf
|
UDapter: Typology-based Language Adapters for Multilingual Dependency Parsing and Sequence Labeling
|
Recent advances in multilingual language modeling have brought the idea of a truly universal parser closer to reality. However, such models are still not immune to the “curse of multilinguality”: Cross-language interference and restrained model capacity remain major obstacles. To address this, we propose a novel language adaptation approach by introducing contextual language adapters to a multilingual parser. Contextual language adapters make it possible to learn adapters via language embeddings while sharing model parameters across languages based on contextual parameter generation. Moreover, our method allows for an easy but effective integration of existing linguistic typology features into the parsing model. Because not all typological features are available for every language, we further combine typological feature prediction with parsing in a multi-task model that achieves very competitive parsing performance without the need for an external prediction system for missing features. The resulting parser, UDapter, can be used for dependency parsing as well as sequence labeling tasks such as POS tagging, morphological tagging, and NER. In dependency parsing, it outperforms strong monolingual and multilingual baselines on the majority of both high-resource and low-resource (zero-shot) languages, showing the success of the proposed adaptation approach. In sequence labeling tasks, our parser surpasses the baseline on high resource languages, and performs very competitively in a zero-shot setting. Our in-depth analyses show that adapter generation via typological features of languages is key to this success.1
|
['Gertjan van Noord', 'Gosse Bouma', 'Arianna Bisazza', 'Ahmet Üstün']
| null | null | null | null |
cl-acl-2022-9
|
['morphological-tagging']
|
['natural-language-processing']
|
[-2.23625153e-01 -2.85195392e-02 -4.54560041e-01 -3.27517331e-01
-1.23846245e+00 -9.76700425e-01 4.06227022e-01 2.05008626e-01
-8.73351872e-01 6.11081898e-01 4.68514889e-01 -7.03263104e-01
3.20578188e-01 -5.75550020e-01 -8.36335659e-01 -4.21695471e-01
2.69856155e-01 5.06370068e-01 2.03612298e-01 -3.33392084e-01
-3.25230956e-01 2.04407096e-01 -9.87236619e-01 2.53067285e-01
1.02222204e+00 2.20708624e-01 7.23504543e-01 5.55909634e-01
-4.34723496e-01 5.08804619e-01 -2.40912408e-01 -7.63012648e-01
3.62932310e-02 -1.39157370e-01 -8.17368269e-01 -3.64198536e-01
2.78032571e-01 1.07696712e-01 -1.89434097e-04 8.86418164e-01
5.51739037e-01 -1.82518706e-01 3.24459851e-01 -5.63446164e-01
-8.64008486e-01 1.19707966e+00 -3.19194973e-01 1.88652843e-01
2.43264645e-01 9.51256678e-02 1.55722249e+00 -1.04126394e+00
9.67167556e-01 1.37623262e+00 7.59945989e-01 6.36621833e-01
-1.38474250e+00 -4.92556512e-01 4.47848380e-01 -9.81118307e-02
-9.99933541e-01 -4.63120073e-01 4.61035997e-01 -4.04810935e-01
1.34322178e+00 -8.45606551e-02 1.54098049e-01 1.39110422e+00
1.91412061e-01 7.95408309e-01 1.11350620e+00 -7.63954341e-01
-1.01947576e-01 1.95918337e-01 3.40251744e-01 6.52770579e-01
7.62809068e-02 -1.65268127e-02 -2.89969832e-01 6.42043650e-02
4.23324078e-01 -3.89106631e-01 -1.28789004e-02 -3.52314934e-02
-1.29506695e+00 9.72414196e-01 1.07325479e-01 7.04636931e-01
-4.85112071e-02 -2.03945816e-01 7.00550199e-01 3.04752082e-01
4.59190935e-01 4.91260141e-01 -9.64116752e-01 3.16526406e-02
-5.01581669e-01 -1.54766783e-01 6.97737098e-01 9.83482718e-01
9.06289637e-01 -1.22700082e-02 3.94487269e-02 1.34109914e+00
-9.92847956e-04 6.34968579e-01 7.14724720e-01 -5.63907683e-01
6.83888733e-01 4.70400125e-01 -3.04728121e-01 -2.72592783e-01
-7.92764544e-01 -2.89208591e-01 -4.46741313e-01 -2.94624686e-01
5.48032701e-01 -3.71380031e-01 -6.94593906e-01 2.24985600e+00
2.70967782e-01 -2.19707802e-01 2.76001185e-01 4.71806675e-01
6.13710284e-01 6.52706623e-01 6.73919380e-01 -7.08873123e-02
1.69452608e+00 -1.03580642e+00 -5.36929071e-01 -7.14675844e-01
1.12018466e+00 -9.64424908e-01 1.46290994e+00 -9.30074155e-02
-8.87324512e-01 -6.07668698e-01 -8.07585478e-01 -4.08165216e-01
-4.32685345e-01 1.92595080e-01 9.03500319e-01 7.45950043e-01
-9.10885751e-01 2.87495553e-01 -7.60656536e-01 -6.09413624e-01
-2.42475308e-02 2.63802767e-01 -7.74253905e-01 -3.56520504e-01
-1.25684381e+00 1.16934144e+00 5.47235191e-01 -1.65156111e-01
-4.03172493e-01 -6.20872676e-01 -1.08715284e+00 -5.81948496e-02
3.64063084e-01 -5.70469856e-01 1.22565460e+00 -8.97567213e-01
-1.49244678e+00 9.17691171e-01 -1.45113647e-01 -2.07419157e-01
8.32088962e-02 -3.01737428e-01 -5.72728395e-01 -3.91743630e-01
3.46295208e-01 6.45009279e-01 2.01878354e-01 -9.18917716e-01
-9.38348293e-01 -2.61070907e-01 7.13864267e-02 1.55407757e-01
-4.23142582e-01 4.29706782e-01 -5.56601048e-01 -5.12589633e-01
-1.97004795e-01 -1.02283752e+00 -4.04432803e-01 -8.68615508e-01
-2.10462630e-01 -3.22650373e-01 2.59430349e-01 -8.66140842e-01
1.22378123e+00 -2.06935763e+00 1.52259126e-01 -2.78413057e-01
-3.02136004e-01 3.67934078e-01 -6.74671292e-01 5.85169673e-01
-4.18091938e-02 3.36975008e-01 -1.97044089e-01 -5.68722904e-01
4.85698022e-02 5.59943557e-01 1.15896370e-02 1.40510023e-01
4.67162043e-01 1.03035188e+00 -9.12119567e-01 -3.80939186e-01
8.56157392e-02 4.46498424e-01 -6.61028326e-01 1.70558691e-01
-8.45316648e-02 5.80117285e-01 -1.05657138e-01 6.01400375e-01
4.36305553e-01 2.82655805e-01 7.97072232e-01 7.80483261e-02
-3.88281047e-01 7.77122438e-01 -8.54264975e-01 1.86070573e+00
-1.18676078e+00 4.42916201e-03 -1.40717598e-02 -7.95428038e-01
7.14585066e-01 3.86938602e-01 7.62964785e-02 -8.76909435e-01
-6.89993501e-02 6.45589054e-01 4.40653175e-01 -3.05100530e-01
4.33821410e-01 -2.91720986e-01 -7.46763289e-01 4.31647062e-01
6.23854816e-01 2.98901767e-01 3.88153076e-01 9.57042202e-02
1.17866671e+00 3.55900884e-01 5.54398358e-01 -4.93575960e-01
6.28019035e-01 4.62688059e-02 9.96532917e-01 6.04680955e-01
-2.84484196e-02 1.55796647e-01 3.83915454e-01 -3.62793744e-01
-1.18169641e+00 -9.87313926e-01 -2.55918503e-01 1.81111908e+00
-4.30133730e-01 -5.00426531e-01 -8.04365396e-01 -9.74170625e-01
-1.15166984e-01 6.75765276e-01 -3.00641298e-01 1.98254511e-01
-1.22173202e+00 -1.02597213e+00 7.53180325e-01 6.36785686e-01
-2.44449019e-01 -1.26677144e+00 6.07099645e-02 6.46419704e-01
-3.20206642e-01 -1.70547771e+00 -5.47449350e-01 5.32049894e-01
-6.71852708e-01 -8.49331796e-01 -2.84164637e-01 -1.09845257e+00
2.82901257e-01 -5.06659709e-02 1.28336442e+00 -1.46342725e-01
5.78942597e-02 8.60602558e-02 -4.03424412e-01 -1.16708595e-02
-9.51067030e-01 7.16879070e-01 1.73063964e-01 -2.79043645e-01
4.64430094e-01 -5.47966182e-01 -1.19741345e-02 1.08117722e-02
-7.05137491e-01 -1.09705448e-01 7.93627501e-01 1.02160120e+00
5.47911584e-01 -7.07778931e-01 9.01118457e-01 -1.30440986e+00
3.55128288e-01 -5.45361698e-01 -5.54532111e-01 3.30809623e-01
-3.30726713e-01 3.92560571e-01 1.09673917e+00 -4.39843684e-01
-1.24895263e+00 2.09582597e-01 -6.73082829e-01 3.63579094e-01
-2.43065432e-01 5.96361756e-01 -6.57350421e-01 2.44017720e-01
4.31186557e-01 -1.52040914e-01 -4.49872434e-01 -9.90569651e-01
7.81202495e-01 5.62176764e-01 5.78696191e-01 -7.96027660e-01
5.08277476e-01 -1.05088860e-01 -3.09482694e-01 -7.92079866e-01
-8.29037905e-01 -3.54377657e-01 -1.09571910e+00 4.44947332e-01
8.82834911e-01 -1.16313779e+00 -3.17823082e-01 2.07630262e-01
-1.34270799e+00 -3.54431748e-01 -1.20960623e-01 4.58463341e-01
-2.04188988e-01 4.03150499e-01 -1.02097154e+00 -4.37364548e-01
-2.20960215e-01 -1.28715813e+00 9.00031388e-01 -1.66506529e-01
-1.41686738e-01 -1.44540823e+00 3.54689956e-01 1.86241135e-01
2.12369755e-01 -1.99047104e-01 1.37104273e+00 -9.88331079e-01
-4.00359660e-01 1.24224704e-02 -4.31097597e-02 4.46140468e-01
-5.82154514e-03 -1.92637250e-01 -9.59077060e-01 -1.63126484e-01
-3.56238484e-01 -2.41602585e-01 8.28831732e-01 1.45297632e-01
3.75800788e-01 -4.82902750e-02 -2.48926252e-01 6.79818153e-01
1.58751297e+00 -1.18974619e-01 1.39393061e-01 3.91490817e-01
9.94288146e-01 6.84641302e-01 3.45677882e-01 -1.27841011e-01
9.47298825e-01 6.89247668e-01 -3.47590121e-03 -1.80407241e-02
-3.31787199e-01 -4.85121995e-01 7.85257339e-01 1.57466269e+00
2.17010036e-01 -2.41847083e-01 -1.22823346e+00 8.20772886e-01
-1.69567299e+00 -4.51609075e-01 -2.51778007e-01 2.25668836e+00
9.49770987e-01 5.12691066e-02 5.86188063e-02 -3.60986739e-01
6.17421329e-01 3.69751304e-02 -3.03402599e-02 -9.17329848e-01
-4.80503321e-01 4.79184061e-01 6.04251325e-01 8.55242908e-01
-1.23226655e+00 1.71299052e+00 5.91109610e+00 7.29954958e-01
-9.61176455e-01 7.62139082e-01 2.73649514e-01 3.40782166e-01
-4.15310442e-01 3.34796220e-01 -1.36637902e+00 3.33706081e-01
1.35404623e+00 1.05573475e-01 3.67721200e-01 7.27505028e-01
-9.15861651e-02 5.44550642e-03 -1.14936280e+00 4.56776738e-01
-1.18667744e-01 -9.93025899e-01 -6.34641945e-02 -2.31729522e-02
5.24543107e-01 6.50270343e-01 -2.95851946e-01 7.17871666e-01
8.76009583e-01 -8.26535821e-01 5.93491137e-01 -9.06442031e-02
8.88214290e-01 -8.09024811e-01 7.90801525e-01 5.06705582e-01
-1.28388393e+00 -1.59847468e-01 -4.87801701e-01 -1.98997892e-02
4.39537108e-01 4.58323717e-01 -6.52087986e-01 6.13575220e-01
2.32875437e-01 3.28554571e-01 -6.79994047e-01 4.81128603e-01
-5.31217754e-01 7.48838603e-01 -2.71936953e-01 2.38733649e-01
3.52554768e-01 1.05792053e-01 3.55393231e-01 1.82484031e+00
2.36325368e-01 -4.44544226e-01 5.24886072e-01 2.95395494e-01
-2.08315685e-01 8.42030227e-01 -6.31325006e-01 -3.65901738e-02
5.60623467e-01 1.45492697e+00 -5.46569347e-01 -1.98948666e-01
-9.17189181e-01 9.14896846e-01 1.06218231e+00 1.83065161e-02
-5.52986801e-01 9.72575857e-04 6.84161127e-01 -3.16562206e-01
3.31060499e-01 -3.92459959e-01 -1.85827777e-01 -1.36426723e+00
-6.32936507e-02 -9.78515089e-01 5.24183631e-01 -1.04715124e-01
-1.41749609e+00 8.71473432e-01 -3.05635065e-01 -6.86385334e-01
-4.66836005e-01 -9.00435388e-01 -3.45824212e-01 1.11316240e+00
-1.70918620e+00 -1.68710136e+00 5.25727272e-01 3.86597663e-01
6.03291750e-01 -1.88326880e-01 1.26077664e+00 5.61909795e-01
-8.19246590e-01 8.42822731e-01 1.15652315e-01 3.29275936e-01
9.87722993e-01 -1.37476122e+00 8.79545271e-01 1.12766790e+00
4.91477758e-01 5.81568301e-01 2.90338427e-01 -5.33864319e-01
-1.26614881e+00 -1.18028235e+00 1.65659237e+00 -6.72111869e-01
9.95106578e-01 -8.16277325e-01 -1.02317286e+00 9.96501923e-01
2.11493090e-01 3.70268449e-02 9.24347341e-01 8.44571173e-01
-6.92204893e-01 9.29744393e-02 -6.10010684e-01 5.97670078e-01
1.17597306e+00 -5.92729867e-01 -5.75437367e-01 2.51923323e-01
8.61327171e-01 -1.26577929e-01 -1.02780688e+00 1.55085593e-01
5.01395285e-01 -6.09189510e-01 6.59270227e-01 -7.13018179e-01
2.59545624e-01 1.49395421e-01 -2.94674188e-01 -1.64582205e+00
-6.44976258e-01 -5.30644834e-01 3.79273176e-01 1.64818156e+00
8.45459163e-01 -6.65739298e-01 7.62295648e-02 1.74060196e-01
-4.23119426e-01 -3.77401590e-01 -1.00817442e+00 -9.44091678e-01
7.72814989e-01 -6.42866254e-01 5.41525066e-01 1.00628579e+00
1.95102543e-01 8.97998214e-01 -2.87321538e-01 3.39379430e-01
4.04607505e-01 -1.15327202e-01 6.31182134e-01 -1.28917325e+00
-5.35477698e-01 -3.75658363e-01 -5.62040880e-02 -7.81669378e-01
7.61137962e-01 -1.42637587e+00 1.28361300e-01 -1.24605548e+00
2.89075047e-01 -7.84232020e-01 -4.31304306e-01 6.71098769e-01
-5.26250124e-01 1.81726053e-01 4.23844188e-01 2.63359733e-02
-3.97520840e-01 1.65684834e-01 8.43980730e-01 3.46236199e-01
-9.77211669e-02 -3.55509281e-01 -8.35357070e-01 8.91344845e-01
7.37976730e-01 -6.76409423e-01 1.95066497e-01 -9.12774563e-01
3.36196840e-01 1.44070923e-01 -2.72577286e-01 -7.42591441e-01
1.13087974e-03 5.67134209e-02 -5.94336316e-02 2.99566295e-02
-2.23114528e-02 -5.35305917e-01 -1.93112776e-01 1.07712060e-01
-4.26813029e-03 3.65148932e-01 3.07104826e-01 1.63523540e-01
-8.93494859e-02 -4.59608376e-01 8.16727936e-01 -2.43414998e-01
-8.46055031e-01 2.57467568e-01 -3.95602733e-01 4.31360304e-01
5.42275667e-01 2.32315630e-01 -3.43592703e-01 3.37606728e-01
-7.01996028e-01 3.20295580e-02 5.25985003e-01 7.76078522e-01
-2.76980281e-01 -1.10474455e+00 -7.85617828e-01 4.08076972e-01
2.45978221e-01 -4.07248735e-01 1.57388628e-01 7.47872710e-01
-1.34750977e-01 5.38159430e-01 -8.63366127e-02 -3.85368347e-01
-9.91435766e-01 6.45124137e-01 1.45645112e-01 -7.97154546e-01
-5.38292408e-01 7.48662472e-01 3.83452624e-01 -1.10177398e+00
-1.82971269e-01 -3.43918502e-01 -1.85861841e-01 3.55049782e-02
3.35236311e-01 -1.08494177e-01 1.92791224e-01 -1.03860140e+00
-4.33008999e-01 4.26017553e-01 -3.29488307e-01 -1.45962149e-01
1.40526807e+00 -2.38029063e-01 -4.70482707e-02 5.83773434e-01
9.85214233e-01 7.48215258e-01 -1.00982392e+00 -3.96408379e-01
5.72145998e-01 6.73652440e-02 -2.45196342e-01 -1.03746498e+00
-6.88075781e-01 9.96769309e-01 1.41283795e-01 -2.26085842e-01
7.31691539e-01 1.30901635e-01 1.06714451e+00 2.10150629e-01
4.65188593e-01 -1.13269997e+00 -4.62790042e-01 1.15489697e+00
3.56135607e-01 -1.35147214e+00 -5.47959387e-01 -4.04293835e-01
-7.88956642e-01 8.77799094e-01 5.72893679e-01 -6.52734041e-02
4.41398561e-01 5.84443331e-01 3.88285279e-01 3.28198522e-01
-8.79365861e-01 -4.79943931e-01 9.90509391e-02 6.58161461e-01
9.33149278e-01 4.15015101e-01 -7.00193465e-01 1.08279085e+00
-2.54614294e-01 -5.75496972e-01 3.41861069e-01 5.50278425e-01
-3.19962084e-01 -1.95480657e+00 -7.59886429e-02 -1.88901216e-01
-8.96052361e-01 -6.21182263e-01 9.14690420e-02 9.85317111e-01
3.76780033e-01 7.14487612e-01 -4.97556776e-02 -1.25155613e-01
3.77220064e-01 6.65576339e-01 3.86708319e-01 -9.64594960e-01
-8.44279468e-01 1.78414851e-01 3.58585268e-01 -3.89632881e-01
-1.24509200e-01 -7.43171096e-01 -1.09117222e+00 -6.09089248e-03
-2.36195982e-01 -2.03524064e-02 7.13538885e-01 1.09577584e+00
2.78227568e-01 4.11454499e-01 4.45050389e-01 -6.40409231e-01
-4.08305734e-01 -1.07868111e+00 -3.58027279e-01 3.45161229e-01
-8.83219466e-02 -4.10066187e-01 -1.69646591e-01 -1.63301945e-01]
|
[10.549464225769043, 9.87064266204834]
|
67f4323c-9ce2-45b5-9613-1b5a2bbc3e8b
|
clustering-human-mobility-with-multiple
|
2301.08524
| null |
https://arxiv.org/abs/2301.08524v1
|
https://arxiv.org/pdf/2301.08524v1.pdf
|
Clustering Human Mobility with Multiple Spaces
|
Human mobility clustering is an important problem for understanding human mobility behaviors (e.g., work and school commutes). Existing methods typically contain two steps: choosing or learning a mobility representation and applying a clustering algorithm to the representation. However, these methods rely on strict visiting orders in trajectories and cannot take advantage of multiple types of mobility representations. This paper proposes a novel mobility clustering method for mobility behavior detection. First, the proposed method contains a permutation-equivalent operation to handle sub-trajectories that might have different visiting orders but similar impacts on mobility behaviors. Second, the proposed method utilizes a variational autoencoder architecture to simultaneously perform clustering in both latent and original spaces. Also, in order to handle the bias of a single latent space, our clustering assignment prediction considers multiple learned latent spaces at different epochs. This way, the proposed method produces accurate results and can provide reliability estimates of each trajectory's cluster assignment. The experiment shows that the proposed method outperformed state-of-the-art methods in mobility behavior detection from trajectories with better accuracy and more interpretability.
|
['Yao-Yi Chiang', 'Haowen Lin', 'Haoji Hu']
|
2023-01-20
| null | null | null | null |
['deep-clustering', 'deep-clustering']
|
['miscellaneous', 'natural-language-processing']
|
[-2.35654905e-01 -3.48700970e-01 -5.83412051e-01 -2.88910747e-01
-3.67143422e-01 -2.24323481e-01 4.90702868e-01 8.95276070e-02
-3.80605280e-01 5.72579801e-01 4.25404966e-01 -3.29661369e-01
-5.17015219e-01 -9.28790927e-01 -4.99802560e-01 -8.72327745e-01
2.34544277e-02 6.77501142e-01 4.15067554e-01 -7.10200220e-02
1.75788682e-02 3.76675844e-01 -1.72690439e+00 2.30012000e-01
1.26640987e+00 3.70160222e-01 2.54062265e-01 3.06658804e-01
-1.81291759e-01 6.61626518e-01 -5.41522920e-01 -8.74542519e-02
-1.00580655e-01 -3.66497427e-01 -7.73675144e-01 2.12350070e-01
-2.02305883e-01 -2.03843459e-01 -6.31887794e-01 8.88735473e-01
1.30851105e-01 7.78916419e-01 1.02219486e+00 -1.63306832e+00
-5.88465273e-01 7.76423872e-01 -4.34446275e-01 3.96834791e-01
2.55007178e-01 -8.15623105e-02 6.70017540e-01 -3.09392571e-01
3.17596674e-01 1.19885242e+00 6.33742034e-01 3.22692424e-01
-1.07749081e+00 -6.77037895e-01 4.71707225e-01 5.14015675e-01
-1.54087126e+00 -1.35533303e-01 6.54275477e-01 -7.19190836e-01
8.78017128e-01 3.50551665e-01 5.90811491e-01 1.32994902e+00
2.29450203e-02 1.00382710e+00 6.26455963e-01 9.55799595e-02
2.15009123e-01 1.54913694e-01 4.49312478e-01 3.93768698e-01
2.79300332e-01 -3.49076331e-01 1.00316532e-01 -3.20931435e-01
4.33752686e-01 8.29436243e-01 -7.55392164e-02 -2.35037029e-01
-1.24191988e+00 6.00777388e-01 4.91219282e-01 6.15174592e-01
-4.72163945e-01 2.31543463e-02 1.46273777e-01 -8.70380998e-02
9.19190794e-02 -1.73865914e-01 -9.38679352e-02 -4.31179225e-01
-1.19847536e+00 2.94860393e-01 4.78743762e-01 8.95170450e-01
9.18663740e-01 -2.95971949e-02 -2.60539174e-01 5.31989753e-01
3.94759208e-01 5.92934489e-01 9.77546930e-01 -7.47180402e-01
6.96845472e-01 8.19234371e-01 2.41348431e-01 -1.41810739e+00
-6.36639178e-01 -1.49737537e-01 -8.26298296e-01 -3.41781765e-01
3.39720696e-01 -7.94412121e-02 -9.18801367e-01 1.60287750e+00
4.29908484e-01 5.34425616e-01 7.73878247e-02 7.30137348e-01
3.93922865e-01 9.13374245e-01 2.43944615e-01 -1.24135144e-01
9.30853784e-01 -9.82064545e-01 -9.11559224e-01 3.03984672e-01
7.98195183e-01 -2.79205352e-01 9.51769590e-01 2.55831089e-02
-7.32527137e-01 -8.07367027e-01 -7.62266219e-01 1.25648320e-01
-4.88593429e-01 3.68122727e-01 5.13986111e-01 8.04563880e-01
-7.99531937e-01 7.60790884e-01 -1.48246610e+00 -6.61416948e-01
2.78612114e-02 4.99168038e-01 -1.19984835e-01 3.05101573e-01
-1.13679802e+00 4.63521421e-01 5.54103017e-01 1.33796949e-02
-4.29961234e-01 -2.08362773e-01 -7.28352308e-01 2.35920236e-01
1.49716586e-01 -4.16590661e-01 8.90372217e-01 -7.78614163e-01
-1.40315223e+00 3.01187336e-01 -6.64672077e-01 -2.82880515e-01
4.15126681e-01 -8.67952630e-02 -9.22086000e-01 -1.10729992e-01
3.56245160e-01 2.66992152e-01 7.27416158e-01 -1.05391872e+00
-1.04060614e+00 -3.12422425e-01 -1.70616120e-01 1.36870682e-01
-5.32610297e-01 -5.22007763e-01 -6.87218845e-01 -5.20389497e-01
3.07723433e-01 -1.30276060e+00 -1.53452307e-01 -7.22966731e-01
-3.29839855e-01 -5.10259390e-01 8.85363698e-01 -6.54741228e-01
1.83325863e+00 -2.02351165e+00 3.08274239e-01 5.36364973e-01
1.66038230e-01 2.55710751e-01 3.97703618e-01 7.04828382e-01
3.16134065e-01 4.05923352e-02 -3.60855341e-01 -4.31344837e-01
3.07192683e-01 3.73677135e-01 -3.26303065e-01 4.49790895e-01
-3.72639298e-01 7.38585413e-01 -9.70395148e-01 -5.16806722e-01
4.82232690e-01 4.02032495e-01 -5.51333487e-01 -1.43906409e-02
1.91661701e-01 5.53885818e-01 -6.69978678e-01 3.53652388e-01
5.75643778e-01 -3.21325719e-01 5.27784050e-01 1.62538990e-01
-9.77709591e-02 9.33037885e-03 -1.42026055e+00 1.33533311e+00
-2.60994941e-01 5.47300577e-01 -4.16120797e-01 -1.07699859e+00
6.55923605e-01 2.24746093e-01 8.57057333e-01 -3.19568574e-01
7.56134689e-02 3.43554770e-03 -1.91176295e-01 -7.78975189e-01
5.53819716e-01 4.10315335e-01 -1.65187806e-01 6.58096850e-01
-1.75645038e-01 9.85333025e-01 1.94954515e-01 -7.50332922e-02
1.06392503e+00 6.33059144e-02 6.16169460e-02 -2.38854364e-02
6.33874536e-01 4.11208794e-02 8.94450843e-01 8.08123827e-01
-4.45008337e-01 1.54633462e-01 8.82805884e-02 -4.34752017e-01
-6.75624490e-01 -1.30155921e+00 3.49978544e-02 1.12152922e+00
5.65668046e-01 -5.04819989e-01 -8.24035764e-01 -6.60941541e-01
4.37076725e-02 7.07872033e-01 -6.16608381e-01 -1.68843135e-01
-7.43530691e-01 -1.03862846e+00 6.77136600e-01 7.02215970e-01
4.36680526e-01 -9.98385370e-01 -3.55088919e-01 2.49855995e-01
-6.21825039e-01 -7.33768940e-01 -4.63209510e-01 -4.12904322e-01
-9.39837873e-01 -1.03534186e+00 -6.20768666e-01 -8.62978995e-01
7.76180685e-01 5.57628334e-01 4.69843060e-01 1.65383026e-01
3.31990004e-01 3.58187348e-01 -4.32735503e-01 6.95872381e-02
-1.24533750e-01 6.31511152e-01 3.61871302e-01 4.47726816e-01
8.65583181e-01 -8.36026847e-01 -6.47919118e-01 5.60876906e-01
-7.91577518e-01 -2.61234283e-01 4.20197695e-01 4.81612802e-01
4.56997365e-01 6.55014753e-01 2.69412071e-01 -6.31017745e-01
7.83412635e-01 -1.07350111e+00 -2.11335436e-01 2.50467718e-01
-7.12470651e-01 7.11403489e-02 7.12454498e-01 -6.41040444e-01
-1.14593744e+00 4.74105328e-02 -5.99162914e-02 -5.29172778e-01
-6.06043935e-01 3.06662798e-01 -3.09468955e-01 5.21443427e-01
3.98394436e-01 4.29619551e-01 -3.48860770e-01 -5.21364331e-01
2.57322699e-01 9.27151799e-01 2.90560305e-01 -4.39466059e-01
6.95460141e-01 8.02344859e-01 -2.99912810e-01 -9.83478308e-01
-7.87393227e-02 -8.26110899e-01 -9.58871186e-01 -1.25054687e-01
1.01663291e+00 -7.79858649e-01 -9.85286713e-01 2.60023475e-01
-6.45097256e-01 -2.21119210e-01 2.67106324e-01 7.87980258e-01
-3.66245776e-01 5.98862350e-01 -6.46098316e-01 -9.25887525e-01
2.37071007e-01 -1.36881709e+00 9.76758599e-01 2.51845241e-01
-4.66313034e-01 -1.17193532e+00 1.93794742e-01 3.08748603e-01
1.43037423e-01 2.82446295e-01 7.64702559e-01 -5.82491517e-01
-5.12706339e-01 -9.44063738e-02 2.22486109e-01 -3.68078798e-01
4.83398438e-01 2.62548532e-02 -7.22119212e-01 -3.35033774e-01
-2.57845938e-01 3.00199062e-01 8.87224555e-01 5.25355518e-01
1.18143868e+00 -4.03435886e-01 -1.01329482e+00 5.27212381e-01
9.77875769e-01 6.08414948e-01 5.31311929e-01 3.46236587e-01
1.10642421e+00 7.62067735e-01 5.08365273e-01 1.58243999e-01
7.40040898e-01 7.56947696e-01 -4.30139862e-02 1.30529940e-01
4.60456431e-01 -5.02537429e-01 4.29940253e-01 1.02930248e+00
-2.18173057e-01 -4.55042034e-01 -1.00736856e+00 6.96084619e-01
-2.59865379e+00 -1.46580279e+00 -2.66124129e-01 2.01512432e+00
7.75327533e-02 5.37408106e-02 5.49060404e-01 3.53502214e-01
9.50950384e-01 1.27556264e-01 -5.59002638e-01 -1.53865784e-01
3.80191982e-01 -4.67501104e-01 4.85831857e-01 4.96950477e-01
-1.17128122e+00 9.16448236e-01 6.17122078e+00 7.97927737e-01
-8.92916441e-01 2.18992352e-01 1.26380166e-02 1.07561000e-01
-3.61402720e-01 -2.48219624e-01 -7.73112833e-01 1.02290010e+00
9.46813643e-01 7.59652555e-02 4.16190088e-01 8.30947518e-01
4.02991682e-01 1.24981105e-01 -7.23496616e-01 1.06727171e+00
-1.83925122e-01 -9.55528617e-01 2.65957624e-01 3.59596759e-01
7.68147290e-01 -2.54528791e-01 1.43264845e-01 5.46454191e-01
1.58949494e-01 -9.28608656e-01 4.85605836e-01 8.49343061e-01
7.35827461e-02 -1.11291718e+00 6.11605883e-01 7.04730928e-01
-1.51398838e+00 -4.13503230e-01 -3.21522772e-01 -2.00173572e-01
3.25003088e-01 1.98892847e-01 -6.13154709e-01 5.85286677e-01
7.91510820e-01 9.34243679e-01 -4.33853686e-01 7.90999353e-01
-2.14308165e-02 9.38733399e-01 -3.14941972e-01 -1.25784993e-01
3.57695371e-01 -5.23760557e-01 4.90626901e-01 1.17154229e+00
5.84324956e-01 3.10767330e-02 5.43030202e-01 7.34840274e-01
2.67973334e-01 -1.18959516e-01 -5.90428054e-01 4.18053642e-02
7.37733483e-01 7.96674192e-01 -9.75952268e-01 -4.69401121e-01
-2.70034105e-01 1.18803632e+00 2.64359683e-01 6.63447082e-01
-1.13275397e+00 -3.19013000e-01 8.48948002e-01 3.30223083e-01
3.55525583e-01 -4.06015247e-01 2.34592259e-01 -1.28333080e+00
2.17055678e-02 -3.33994329e-01 5.13205647e-01 -2.42844462e-01
-1.08061171e+00 4.34221059e-01 3.28949153e-01 -1.47785699e+00
-4.32810247e-01 -3.03999692e-01 -7.53322840e-01 5.48501670e-01
-8.88860345e-01 -9.40844655e-01 -2.65061766e-01 7.57261932e-01
4.01227146e-01 -1.31415471e-01 6.36057138e-01 6.61167860e-01
-9.31924045e-01 6.66882813e-01 6.62800610e-01 2.41061553e-01
3.40965897e-01 -1.26853478e+00 3.74384195e-01 8.46882403e-01
1.79755852e-01 1.02986860e+00 4.75259036e-01 -9.37938690e-01
-9.06504929e-01 -1.26064157e+00 8.52206171e-01 -6.54779971e-01
3.66545916e-01 4.04860973e-02 -9.53435242e-01 1.12573731e+00
-2.04906285e-01 -6.49270594e-01 9.26264107e-01 3.41864526e-01
3.56778741e-01 -3.30033293e-03 -9.14632678e-01 7.92060554e-01
1.26315022e+00 -4.75039810e-01 -7.95280755e-01 2.53612716e-02
5.39061606e-01 -1.99581459e-01 -8.09487402e-01 1.44827455e-01
5.55572212e-01 -9.57686305e-01 9.26183999e-01 -5.81125319e-01
-1.09552182e-01 -7.92791843e-01 6.91611320e-02 -1.13507450e+00
-7.71670580e-01 -3.59963864e-01 -4.77916867e-01 1.18817818e+00
1.68117344e-01 -5.15156507e-01 8.59138966e-01 6.10408783e-01
-7.62504190e-02 -5.64954698e-01 -7.75033414e-01 -7.15458989e-01
-2.46942148e-01 -3.73002023e-01 1.13592851e+00 1.16382420e+00
-2.09544133e-02 -4.35162075e-02 -5.49297988e-01 4.84929413e-01
6.12926185e-01 1.68045029e-01 9.84278202e-01 -1.23467505e+00
2.70111598e-02 -5.03197253e-01 -5.62914550e-01 -1.15219414e+00
3.47484648e-01 -1.05774796e+00 -1.99937046e-01 -1.61399662e+00
-1.04431398e-01 -2.14907035e-01 -6.14573956e-01 1.75040379e-01
-3.49391311e-01 -3.21669430e-01 -6.24302961e-02 6.83521032e-01
-8.58423293e-01 6.70317590e-01 7.09099352e-01 -2.55661458e-01
-8.71394277e-01 3.91392142e-01 -3.03993225e-01 7.49123514e-01
9.73921955e-01 -6.24853134e-01 -7.81256855e-01 -3.57096702e-01
-2.05411419e-01 -8.07304531e-02 1.64729252e-01 -1.42024255e+00
4.77474272e-01 -2.96997637e-01 5.14419019e-01 -9.17107463e-01
2.92003244e-01 -8.73195350e-01 4.91180032e-01 4.77712303e-01
-2.12407559e-01 2.93815464e-01 -9.43304896e-02 1.05393207e+00
-9.36969817e-02 2.01139346e-01 2.37722769e-01 -1.33158341e-02
-9.08290982e-01 4.14178908e-01 -1.06398094e+00 -3.29019547e-01
1.12632692e+00 -5.83435774e-01 1.84747308e-01 -3.38211685e-01
-9.91628587e-01 6.53239787e-01 5.24319887e-01 7.83037424e-01
4.11511809e-01 -1.71456063e+00 -3.20794657e-02 1.18292488e-01
-1.17808208e-02 -2.90730119e-01 5.24845541e-01 9.26528692e-01
-2.90805250e-01 4.65833783e-01 -3.32263142e-01 -7.80100167e-01
-1.00475943e+00 7.65046179e-01 1.57069206e-01 -3.23907524e-01
-7.61673212e-01 2.08558932e-01 -1.60916239e-01 -7.03166187e-01
3.85389388e-01 -5.49914122e-01 -6.24494076e-01 1.17498688e-01
5.10040581e-01 1.11638474e+00 -3.76058966e-01 -1.18090439e+00
-4.12405938e-01 6.17110372e-01 1.80510491e-01 -1.53908625e-01
8.51394951e-01 -5.33326268e-01 2.51896977e-01 9.13354218e-01
1.13490784e+00 -1.02127202e-01 -1.01467443e+00 -3.42863612e-02
2.51900941e-01 -3.28563273e-01 -2.85107821e-01 -3.07990815e-02
-7.44670451e-01 8.30577075e-01 7.37781644e-01 2.83614039e-01
8.31162632e-01 -3.47872227e-01 1.26933813e+00 6.03261888e-01
3.58832985e-01 -1.38495779e+00 -3.20965409e-01 2.29656219e-01
1.52248591e-01 -1.26952600e+00 -2.88772434e-01 -9.47348028e-02
-7.53344119e-01 9.00318801e-01 6.10667646e-01 6.11229725e-02
8.91530871e-01 -6.32362247e-01 5.11466265e-02 -6.37141541e-02
-2.00568900e-01 -4.03154224e-01 4.43204880e-01 5.93579769e-01
1.48671836e-01 4.56112444e-01 -2.40454853e-01 6.81549191e-01
-1.57163307e-01 -1.80477723e-01 2.05013275e-01 9.04486775e-01
-4.99269158e-01 -1.04169703e+00 -4.95009750e-01 4.21208411e-01
-1.70662314e-01 5.98312080e-01 -7.66950473e-02 7.02060282e-01
4.99611855e-01 1.19254923e+00 1.76246360e-01 -8.52458894e-01
3.27634126e-01 1.58263415e-01 -2.13574860e-02 -3.41581881e-01
-3.21618229e-01 -1.17717728e-01 -3.98232907e-01 -7.41658509e-01
-6.28440976e-01 -7.36938596e-01 -1.57917464e+00 -6.03607476e-01
2.90439967e-02 5.76531470e-01 1.55233949e-01 1.08853209e+00
5.11723638e-01 4.49537098e-01 4.33317631e-01 -8.72550964e-01
-8.69408473e-02 -7.92946339e-01 -7.70581841e-01 8.39867413e-01
3.44032645e-01 -1.02493215e+00 -2.07423776e-01 4.59265783e-02]
|
[6.517541885375977, 1.5581469535827637]
|
c7b9fd4c-3197-40ff-a2b6-315ea1e0661d
|
improving-robustness-via-risk-averse
|
2005.00585
| null |
https://arxiv.org/abs/2005.00585v1
|
https://arxiv.org/pdf/2005.00585v1.pdf
|
Improving Robustness via Risk Averse Distributional Reinforcement Learning
|
One major obstacle that precludes the success of reinforcement learning in real-world applications is the lack of robustness, either to model uncertainties or external disturbances, of the trained policies. Robustness is critical when the policies are trained in simulations instead of real world environment. In this work, we propose a risk-aware algorithm to learn robust policies in order to bridge the gap between simulation training and real-world implementation. Our algorithm is based on recently discovered distributional RL framework. We incorporate CVaR risk measure in sample based distributional policy gradients (SDPG) for learning risk-averse policies to achieve robustness against a range of system disturbances. We validate the robustness of risk-aware SDPG on multiple environments.
|
['Qinsheng Zhang', 'Rahul Singh', 'Yongxin Chen']
|
2020-05-01
| null |
https://openreview.net/forum?id=6_48llFrKdm
|
https://openreview.net/pdf?id=6_48llFrKdm
|
l4dc-2020-6
|
['distributional-reinforcement-learning']
|
['methodology']
|
[-1.95242196e-01 3.13491523e-02 -6.15913868e-02 -9.22307447e-02
-7.36983359e-01 -6.43482268e-01 6.72466636e-01 -2.35818252e-02
-5.84990382e-01 1.27114010e+00 2.48286694e-01 -6.56556368e-01
-4.92186099e-01 -7.18303442e-01 -8.67466807e-01 -7.03095675e-01
-4.80504990e-01 6.21979907e-02 1.04378864e-01 -2.75357097e-01
2.46363208e-01 5.45162082e-01 -1.38861990e+00 -4.50175494e-01
1.05816209e+00 6.42473578e-01 1.54057637e-01 7.85224557e-01
3.76101226e-01 6.89858496e-01 -7.55816281e-01 3.04797888e-01
7.43598640e-01 -1.60353050e-01 -1.57354996e-01 -4.40064996e-01
-1.65983349e-01 -4.16872025e-01 -7.24906772e-02 1.07132411e+00
8.32058430e-01 5.30986547e-01 7.53417671e-01 -1.33778906e+00
-3.17386389e-01 5.66471100e-01 -4.23718363e-01 1.56538174e-01
4.52133752e-02 6.44030154e-01 3.43681246e-01 -2.40207911e-01
2.20110759e-01 1.55062437e+00 3.53473425e-01 5.91122627e-01
-1.06451786e+00 -6.49469554e-01 4.93075758e-01 -2.40755498e-01
-1.15159535e+00 -2.08211794e-01 4.02438760e-01 -3.51715744e-01
1.05640471e+00 -9.89527851e-02 3.16597909e-01 1.41862905e+00
7.00818300e-01 3.07164192e-01 1.54930890e+00 -3.09098482e-01
9.65167999e-01 2.18643472e-01 -5.28170645e-01 2.05822721e-01
6.50868654e-01 1.01055968e+00 8.06588009e-02 -2.94264883e-01
6.46329284e-01 -4.15189922e-01 -9.20229554e-02 -5.81580997e-01
-6.24475598e-01 7.13720381e-01 9.68219936e-02 -1.53663889e-01
-3.74056339e-01 4.80991274e-01 5.54910302e-01 6.38909757e-01
2.33695060e-01 7.64726043e-01 -5.72636545e-01 -3.58943105e-01
-4.60847676e-01 3.51694524e-01 7.08214700e-01 7.24713385e-01
1.70976207e-01 7.59808481e-01 -2.74512589e-01 4.84072745e-01
6.20544553e-01 8.93232644e-01 4.45263773e-01 -9.42646623e-01
5.52557349e-01 1.34127036e-01 6.07925534e-01 -7.60652244e-01
-2.22919554e-01 -5.85633159e-01 -3.68870556e-01 1.03069544e+00
3.94983262e-01 -6.44788563e-01 -7.77193308e-01 1.86228561e+00
4.50513095e-01 3.49446386e-01 4.12302703e-01 6.40641153e-01
-4.90831256e-01 5.80513179e-01 8.94756019e-02 -4.55752164e-01
3.05955440e-01 -3.97734404e-01 -5.02117395e-01 -7.79257938e-02
1.53209090e-01 -3.90256405e-01 1.31139684e+00 5.33353567e-01
-7.77115822e-01 -2.57202148e-01 -1.25560308e+00 8.87987792e-01
-2.63583511e-01 -3.58954638e-01 1.63023137e-02 9.35241461e-01
-9.52091277e-01 9.23313022e-01 -7.08010316e-01 -3.38021308e-01
4.81667966e-02 3.97522837e-01 8.57784450e-02 5.00141203e-01
-1.23951542e+00 1.39671230e+00 5.49118936e-01 -2.14533940e-01
-1.52824676e+00 -7.33173907e-01 -6.42469466e-01 -2.57160842e-01
8.38783324e-01 -1.67919293e-01 1.45967400e+00 -6.69704735e-01
-2.31362247e+00 -1.17683955e-01 5.14916003e-01 -7.84522593e-01
1.01514673e+00 -4.90686297e-01 -4.02385771e-01 -2.15104729e-01
-3.03920448e-01 -1.30018577e-01 1.30393791e+00 -1.53498399e+00
-6.26323938e-01 1.17503330e-02 -9.29890871e-02 3.22519839e-01
7.29248375e-02 -1.87873706e-01 5.60159802e-01 -5.68380654e-01
-7.90240228e-01 -9.00677025e-01 -5.46697855e-01 -5.58226109e-01
-1.72435902e-02 9.53430608e-02 9.60011363e-01 -5.09443402e-01
9.96091068e-01 -1.91353846e+00 -1.48667574e-01 5.23509920e-01
-4.68701094e-01 6.25353992e-01 -1.37967423e-01 7.11741269e-01
9.56962854e-02 1.81920961e-01 -4.36584130e-02 1.89231768e-01
4.16148722e-01 4.83616233e-01 -1.06400275e+00 6.23217523e-01
1.58061475e-01 4.37313914e-01 -9.89206314e-01 -6.00339472e-02
4.04295683e-01 2.07679838e-01 -2.92634070e-01 4.88804758e-01
-4.70335394e-01 7.22666681e-01 -6.99419618e-01 3.30419838e-01
5.77429593e-01 7.75101960e-01 2.68661175e-02 4.31936294e-01
-3.21128696e-01 -5.40242381e-02 -1.39817178e+00 8.83755803e-01
-8.25257838e-01 -3.71875577e-02 1.85237423e-01 -8.50861430e-01
8.62055600e-01 1.97231263e-01 2.26388901e-01 -8.49031031e-01
3.84835809e-01 1.00555018e-01 1.71066336e-02 -3.03713411e-01
2.62659818e-01 -2.32953653e-01 -1.25771150e-01 6.04764283e-01
2.02922169e-02 -5.00188470e-01 -5.92278428e-02 -1.29767358e-01
8.90809000e-01 4.49918747e-01 5.07510781e-01 -6.45781577e-01
3.54393750e-01 -4.70412999e-01 8.38595152e-01 9.86976326e-01
-6.32139981e-01 -6.31954381e-03 5.60263753e-01 -8.94216299e-02
-1.16584957e+00 -1.24540472e+00 6.12270124e-02 8.69167328e-01
-9.34786126e-02 6.18829392e-02 -5.34581363e-01 -8.49646389e-01
4.12190050e-01 1.31011951e+00 -5.10406017e-01 -6.13917351e-01
-5.53932190e-01 -5.27577043e-01 5.73363721e-01 4.48161036e-01
2.04046682e-01 -8.31242025e-01 -9.64199662e-01 4.51970935e-01
7.51220107e-01 -7.67197073e-01 -2.02268645e-01 3.42292517e-01
-6.99498832e-01 -9.81775582e-01 -2.73509204e-01 2.55378950e-02
3.40369403e-01 -2.52819896e-01 7.98505843e-01 -4.18136239e-01
-7.35837594e-02 7.09172070e-01 -1.53206035e-01 -8.99037778e-01
-6.80828333e-01 -4.69982445e-01 8.32305908e-01 -3.21087509e-01
-4.05878365e-01 -7.01090574e-01 -4.47003722e-01 5.08904934e-01
-7.77888417e-01 -7.07344651e-01 4.34450448e-01 8.13865244e-01
4.24422204e-01 4.42136437e-01 9.34356451e-01 -4.33824211e-01
1.33751547e+00 -6.03645265e-01 -1.25093615e+00 4.04785752e-01
-9.13370132e-01 5.08065641e-01 1.10806906e+00 -7.27206826e-01
-1.13641942e+00 -1.76061466e-01 -1.22128315e-02 -3.21473598e-01
-3.32005024e-02 2.56413013e-01 -8.04350600e-02 -9.95672122e-02
7.51719832e-01 3.62962075e-02 4.66868430e-02 -1.50390238e-01
4.87299711e-01 5.01048207e-01 2.27219552e-01 -1.27317166e+00
1.05512643e+00 1.16227176e-02 8.84163007e-02 -6.14653111e-01
-4.80631620e-01 1.37089536e-01 -9.08745080e-02 -3.25735062e-01
3.70968491e-01 -8.35400999e-01 -7.92753637e-01 2.90768176e-01
-4.05313402e-01 -9.15570617e-01 -6.44434929e-01 5.58187842e-01
-1.03219652e+00 7.80537277e-02 -1.49821997e-01 -1.52552295e+00
-2.91562617e-01 -1.01046038e+00 4.24614936e-01 4.50659513e-01
6.78398609e-02 -9.73932803e-01 5.81600666e-01 -4.12081510e-01
7.66939759e-01 5.92186153e-01 7.09243119e-01 -4.21301484e-01
-8.93933922e-02 8.71448964e-02 3.44025582e-01 6.99648499e-01
5.43810800e-02 2.22464815e-01 -8.13748360e-01 -8.46708536e-01
2.80248761e-01 -6.00144982e-01 3.66175443e-01 3.36445481e-01
6.82933092e-01 -5.96993029e-01 7.17958659e-02 3.41301829e-01
1.56263089e+00 5.59460282e-01 3.86582434e-01 5.87026477e-01
1.81773782e-01 3.14352721e-01 8.60974729e-01 7.01093674e-01
-7.72663131e-02 2.80409694e-01 6.26697779e-01 2.99498588e-01
4.97661412e-01 -7.40528047e-01 8.97571743e-01 3.67539823e-01
1.51271015e-01 -1.28806576e-01 -8.11601937e-01 4.02944565e-01
-1.94164109e+00 -9.43104386e-01 6.59265935e-01 2.74351263e+00
6.70332313e-01 3.88785690e-01 3.52880418e-01 -1.50940672e-01
5.85835159e-01 -1.40365541e-01 -1.05457115e+00 -9.55433071e-01
1.22427016e-01 3.04019805e-02 8.46424520e-01 6.74175262e-01
-9.55035806e-01 8.04829836e-01 7.23249054e+00 7.93495536e-01
-1.22586751e+00 -2.31700271e-01 4.32182521e-01 -1.17770664e-01
-6.72548711e-02 -2.06550676e-02 -5.18834531e-01 4.90602434e-01
1.28074956e+00 -7.61329353e-01 5.67498207e-01 1.16643679e+00
7.81558216e-01 -2.50613093e-01 -7.30656207e-01 3.89418960e-01
-4.20458913e-01 -6.38660550e-01 -3.21982026e-01 -5.31814806e-03
7.67818809e-01 -7.57577922e-03 5.31690776e-01 7.43212223e-01
1.19866574e+00 -1.29127944e+00 7.34945297e-01 4.85351026e-01
4.29313123e-01 -1.23851371e+00 6.65854156e-01 5.28758466e-01
-8.12979400e-01 -5.69208860e-01 -4.75226969e-01 -2.66726226e-01
5.79656847e-02 3.35760713e-01 -1.07337964e+00 4.12985146e-01
3.96741390e-01 1.62658051e-01 -1.99313954e-01 1.00286067e+00
-4.32759613e-01 8.34746003e-01 -4.25129294e-01 -2.06103802e-01
5.01572967e-01 -2.11404324e-01 7.42420793e-01 9.67647254e-01
4.45119828e-01 -2.06451073e-01 4.65451062e-01 6.71784461e-01
5.67532361e-01 -3.70572135e-03 -1.02066779e+00 -6.37170821e-02
4.92793530e-01 8.88977408e-01 -4.09091175e-01 9.60813761e-02
9.74034965e-02 3.62958163e-01 3.45952868e-01 4.93336946e-01
-9.19536293e-01 -2.05173433e-01 9.73860085e-01 -2.80838013e-01
1.52556792e-01 -6.12371981e-01 -6.92426264e-02 -6.88496232e-01
-5.73291108e-02 -1.13105369e+00 2.99376369e-01 -2.08729312e-01
-1.51552701e+00 3.62102091e-01 1.82957575e-01 -1.27740669e+00
-5.89674115e-01 -5.71172953e-01 -8.48088801e-01 8.66129577e-01
-1.45543170e+00 -6.57703757e-01 5.27756810e-01 6.61299884e-01
2.96375781e-01 -4.78586406e-01 5.47632337e-01 -2.44092286e-01
-6.73823476e-01 5.53013563e-01 7.10175455e-01 -4.71672654e-01
8.49902689e-01 -1.51457977e+00 1.80268586e-01 1.04590702e+00
-5.22201478e-01 5.57213843e-01 1.33488691e+00 -7.60634780e-01
-1.47447586e+00 -1.25999939e+00 -3.56570750e-01 -3.44568193e-01
8.72709274e-01 -2.65752912e-01 -6.95591211e-01 3.77494663e-01
2.25855589e-01 -1.26815662e-01 1.65618464e-01 -2.58009017e-01
-2.85184085e-01 -1.34673923e-01 -1.55147147e+00 9.40120816e-01
6.13861918e-01 -2.11095586e-01 -5.20072818e-01 -1.29472882e-01
9.14366066e-01 -3.94822925e-01 -8.00590336e-01 5.27363598e-01
3.07642311e-01 -5.68120837e-01 8.48602772e-01 -8.03783059e-01
-2.13799819e-01 -4.85635519e-01 -2.67211467e-01 -2.04551601e+00
2.01155338e-02 -1.10010207e+00 -1.69509560e-01 1.16152978e+00
2.98843116e-01 -9.96259093e-01 2.79582381e-01 6.12152100e-01
6.10043257e-02 -5.56345284e-01 -1.04987109e+00 -1.42253089e+00
7.65217602e-01 -5.24188042e-01 5.99591315e-01 5.79865515e-01
6.38233870e-02 -2.39573270e-01 -6.03863120e-01 4.28880930e-01
7.74230123e-01 -3.34157348e-01 7.34770536e-01 -5.21355391e-01
-5.55619597e-01 -3.39777201e-01 -1.62919849e-01 -8.06277543e-02
3.25688839e-01 -2.47557804e-01 2.56918401e-01 -1.21197367e+00
-4.09972638e-01 -4.63490903e-01 -5.67239642e-01 2.55263269e-01
-1.49869785e-01 -6.57107055e-01 3.14244092e-01 -4.28853780e-01
-4.89740580e-01 1.03169060e+00 9.52833056e-01 8.08993801e-02
-4.49650437e-01 1.87113494e-01 -5.52041888e-01 6.79334819e-01
1.39323628e+00 -5.85105181e-01 -1.09089136e+00 -5.29890470e-02
1.04975224e-01 2.48360797e-03 -5.32572307e-02 -1.30385399e+00
-1.42096639e-01 -8.86334062e-01 1.53230444e-01 -8.96773562e-02
-2.41625607e-01 -8.10570240e-01 -4.67132069e-02 7.52252698e-01
-3.65709931e-01 2.57625461e-01 5.74222088e-01 9.23858404e-01
1.40540242e-01 2.51372419e-02 1.04227948e+00 -4.76644048e-03
-6.01754725e-01 1.74934164e-01 -6.27142608e-01 3.98865551e-01
1.27970588e+00 1.66399956e-01 -3.32656831e-01 -5.13024867e-01
-2.68824786e-01 6.06025219e-01 4.08300459e-01 4.99606669e-01
5.45930445e-01 -9.93254960e-01 -5.05421281e-01 -1.10803377e-02
-2.00697571e-01 -4.30061281e-01 -5.01856357e-02 3.44731629e-01
-2.49301970e-01 2.12232977e-01 -4.61745471e-01 -7.92030841e-02
-7.16706038e-01 8.48727524e-01 6.92377210e-01 -3.07369173e-01
-6.00355506e-01 3.33748132e-01 -1.76951408e-01 -5.65580428e-01
5.77427924e-01 -3.90648395e-01 1.64008185e-01 -4.63953018e-01
6.63511753e-01 4.02501971e-01 -7.36861676e-02 -1.26784533e-01
-1.68559015e-01 3.38442653e-01 1.05307698e-01 -7.12588072e-01
1.11620116e+00 -9.54597667e-02 4.14228022e-01 5.03420591e-01
5.82329869e-01 -8.68324488e-02 -2.02069306e+00 1.76325381e-01
1.63526312e-01 -6.05063796e-01 8.54345262e-02 -1.18276024e+00
-6.69284165e-01 6.27150893e-01 8.95201623e-01 -6.37675449e-02
9.00186121e-01 -8.72208595e-01 2.81247437e-01 4.87040848e-01
7.25363135e-01 -1.82402658e+00 1.56937361e-01 7.71643698e-01
1.04750574e+00 -9.82958436e-01 -8.57427344e-02 3.95396411e-01
-1.12008739e+00 8.38501334e-01 7.95810342e-01 -6.85791314e-01
6.61154211e-01 6.83418393e-01 2.56823629e-01 4.69654173e-01
-9.76474285e-01 -1.24190956e-01 -1.01056702e-01 1.11896491e+00
-1.14877336e-01 2.07957730e-01 -2.48123094e-01 3.00225377e-01
1.05923370e-01 -1.61381960e-01 6.06398821e-01 1.18590546e+00
-5.71784675e-01 -1.08997965e+00 -5.49491465e-01 9.32819843e-02
-3.45586479e-01 3.77784252e-01 -2.11992636e-01 9.08231378e-01
-2.47598574e-01 1.07268548e+00 -5.15897393e-01 -2.81144083e-01
4.69736844e-01 -6.22193217e-02 3.45304817e-01 -2.37709641e-01
-4.65762794e-01 -8.86821449e-02 1.11770965e-01 -7.95667529e-01
1.53196022e-01 -4.95525181e-01 -1.19753540e+00 -9.88011360e-02
2.82004159e-02 1.69789150e-01 6.88422561e-01 8.84048820e-01
3.88889611e-01 5.45966089e-01 9.20878947e-01 -6.54180586e-01
-1.63432932e+00 -6.41434431e-01 -7.51461506e-01 1.60510018e-01
3.72577757e-01 -9.98346210e-01 -4.79948044e-01 -6.92167044e-01]
|
[4.447775363922119, 2.3218441009521484]
|
3d3659a8-d025-463f-b5d6-eb7f093fa1ca
|
mining-duplicate-questions-of-stack-overflow
|
2210.01637
| null |
https://arxiv.org/abs/2210.01637v1
|
https://arxiv.org/pdf/2210.01637v1.pdf
|
Mining Duplicate Questions of Stack Overflow
|
There has a been a significant rise in the use of Community Question Answering sites (CQAs) over the last decade owing primarily to their ability to leverage the wisdom of the crowd. Duplicate questions have a crippling effect on the quality of these sites. Tackling duplicate questions is therefore an important step towards improving quality of CQAs. In this regard, we propose two neural network based architectures for duplicate question detection on Stack Overflow. We also propose explicitly modeling the code present in questions to achieve results that surpass the state of the art.
|
['Pranav Dheram', 'Radhika Parik', 'Anirudha Rayasam', 'Mihir Kale']
|
2022-10-04
| null | null | null | null |
['community-question-answering', 'community-question-answering']
|
['miscellaneous', 'natural-language-processing']
|
[-4.85360026e-01 -2.14863308e-02 4.09224004e-01 -2.30004609e-01
-8.22459638e-01 -5.78150988e-01 5.23096025e-01 5.05162477e-01
-4.36776400e-01 3.81082982e-01 7.26694822e-01 -4.08177406e-01
-1.10044152e-01 -7.87252307e-01 -3.93705398e-01 5.32614142e-02
2.27776811e-01 1.08071484e-01 5.75323641e-01 -5.30243933e-01
6.96394086e-01 1.17240280e-01 -1.53985953e+00 6.24789715e-01
9.28714633e-01 5.65013111e-01 1.95524804e-02 9.54132199e-01
-9.90647197e-01 1.51666319e+00 -9.30130422e-01 -7.36522675e-01
-1.41768068e-01 -3.91612947e-01 -9.51119423e-01 -2.91514575e-01
9.27776933e-01 -4.23915684e-01 -3.10145199e-01 9.94461477e-01
4.04965311e-01 -6.27515242e-02 1.64440852e-02 -1.11571562e+00
-7.69208670e-01 4.39495325e-01 -3.80189449e-01 6.35265768e-01
3.91943842e-01 1.05966635e-01 1.42852032e+00 -6.40131176e-01
5.06828189e-01 1.00674176e+00 7.43281782e-01 3.80734473e-01
-7.07436740e-01 -3.49027127e-01 -9.53356400e-02 3.38551193e-01
-8.69740069e-01 -6.16638184e-01 4.32212025e-01 -5.77433228e-01
1.25862539e+00 2.00845405e-01 4.65113103e-01 4.14804697e-01
8.69609118e-02 6.99184477e-01 7.34710693e-01 -4.88243729e-01
3.66552562e-01 -1.13788031e-01 1.89818591e-01 5.55630863e-01
3.86474639e-01 -6.36548400e-01 -5.70648313e-01 -4.95020211e-01
3.69199097e-01 8.57292861e-02 -1.59489304e-01 -3.89767960e-02
-5.74783206e-01 8.56612623e-01 4.06531602e-01 6.79646432e-01
-4.17492062e-01 4.13632780e-01 4.71457750e-01 3.72973204e-01
2.37687930e-01 1.02564216e+00 -3.00123207e-02 -5.79062223e-01
-9.96828973e-01 6.75219357e-01 1.39279675e+00 5.59432328e-01
5.49109459e-01 -3.13980490e-01 -1.59251913e-01 4.81738657e-01
4.31289107e-01 1.96640208e-01 2.75353253e-01 -1.60088611e+00
6.67980671e-01 1.22394991e+00 4.63327110e-01 -1.34080958e+00
-6.82394505e-02 -4.84537154e-01 -3.82296778e-02 8.86759311e-02
7.18335330e-01 1.35512367e-01 -1.64232954e-01 1.26383114e+00
1.80482030e-01 -3.16948086e-01 -2.77124017e-01 4.05231535e-01
6.39042616e-01 3.96810293e-01 -9.84622762e-02 4.26384419e-01
1.17224550e+00 -9.19347703e-01 -7.95202792e-01 -2.54920781e-01
8.03013623e-01 -8.32845092e-01 1.01803660e+00 2.41917759e-01
-1.13822317e+00 -1.49990112e-01 -9.20077562e-01 -1.39533788e-01
-3.65139812e-01 -3.19429696e-01 4.46910024e-01 9.98010874e-01
-1.17572796e+00 5.17432392e-01 -6.89560413e-01 -4.50214744e-01
5.37577748e-01 4.34141271e-02 -7.01470748e-02 -2.77746111e-01
-8.84950757e-01 6.53910756e-01 -2.85390228e-01 -2.47358885e-02
-3.11065316e-01 -5.39578676e-01 -2.75845289e-01 4.15943980e-01
4.56297368e-01 -5.40308297e-01 1.64058518e+00 -5.65222859e-01
-7.42543578e-01 8.23641896e-01 -2.19820440e-01 -2.12223649e-01
4.57092315e-01 -3.66384894e-01 -1.86412424e-01 2.36933917e-01
3.38252783e-01 7.02444613e-02 2.78376192e-01 -8.00382972e-01
-7.56937623e-01 -5.43799400e-01 5.37679195e-01 -1.95263714e-01
-7.56000578e-01 5.55694222e-01 -3.20831239e-02 -1.11361921e-01
-2.96201050e-01 -4.45228428e-01 -5.81366643e-02 9.71632525e-02
1.30888373e-01 -6.09555602e-01 6.80656970e-01 -9.98560250e-01
1.52692342e+00 -1.85431397e+00 -3.61633360e-01 -3.18191051e-01
7.56873786e-01 5.39555192e-01 -2.82806873e-01 9.43227708e-01
6.39029920e-01 5.25099754e-01 8.56392086e-02 -2.62714565e-01
7.06084892e-02 1.74890861e-01 -1.54893801e-01 4.08736914e-02
2.52587825e-01 7.61831045e-01 -1.01689434e+00 -2.61268944e-01
-4.64995354e-01 2.70609915e-01 -8.78051877e-01 4.30841953e-01
-6.36610329e-01 7.91369081e-02 -4.69589859e-01 4.95084941e-01
4.67351139e-01 -8.16318870e-01 3.82603891e-02 5.95809519e-01
-3.27267677e-01 5.97436249e-01 -8.95648360e-01 1.39027929e+00
-2.80594200e-01 8.51534069e-01 3.93743992e-01 -3.33752394e-01
8.46268535e-01 5.15171945e-01 2.35711351e-01 -9.68606412e-01
-1.28601268e-01 3.64478290e-01 8.56898576e-02 -9.51573014e-01
8.24536562e-01 5.96848764e-02 3.85944098e-01 8.62966776e-01
-3.88133734e-01 1.77314162e-01 3.89603853e-01 5.13908446e-01
1.67020595e+00 -2.28516564e-01 -2.89353505e-02 3.00038736e-02
6.76416993e-01 1.53763846e-01 3.97041380e-01 9.00816977e-01
-7.47383535e-01 4.64628935e-01 7.78343976e-01 -4.31357533e-01
-1.08051181e+00 -7.69187748e-01 3.92422438e-01 1.15691745e+00
-2.64340669e-01 -2.81432331e-01 -7.79905677e-01 -4.38889623e-01
6.16246946e-02 5.34844697e-01 -4.56535965e-01 2.55118549e-01
-8.01313221e-01 -6.57687262e-02 7.41584241e-01 6.27204835e-01
6.57745242e-01 -9.60651875e-01 -6.85297966e-01 3.65887105e-01
-5.47972023e-01 -1.10196590e+00 -3.14176589e-01 -2.31808096e-01
-9.46360171e-01 -1.30044270e+00 -7.37403035e-01 -4.47355360e-01
-2.43330747e-02 5.65454900e-01 1.60805142e+00 6.34437442e-01
-2.91420132e-01 6.36226177e-01 -5.43823719e-01 -3.50310713e-01
-5.59468865e-01 4.16061550e-01 -5.17931163e-01 -3.96461070e-01
9.24659848e-01 -5.99053383e-01 -7.47169554e-01 1.68538764e-01
-9.78264868e-01 -8.24468315e-01 1.04996063e-01 1.21030428e-01
-3.37194622e-01 -2.72354931e-01 9.33733582e-01 -5.63917458e-01
1.00822711e+00 -8.66350651e-01 -5.37873745e-01 1.33971363e-01
-4.93175149e-01 -4.18613218e-02 4.07905579e-01 2.10591644e-01
-1.04607320e+00 -8.05806041e-01 -1.26069501e-01 6.48088455e-02
-1.33070350e-01 5.56278706e-01 3.18226635e-01 -2.27241460e-02
9.15678978e-01 -3.81535798e-01 8.91055688e-02 -7.68099725e-01
9.44864601e-02 8.90645921e-01 3.53381068e-01 -3.47948521e-01
2.79781342e-01 5.14420807e-01 -3.71872008e-01 -9.19742882e-01
-8.20025980e-01 -8.63410413e-01 -1.74287722e-01 -3.00321370e-01
8.63731503e-01 -7.20306516e-01 -7.94705272e-01 3.64556462e-01
-1.46818650e+00 1.35931939e-01 -5.13535142e-02 -1.03136972e-01
-8.60252418e-03 5.73055685e-01 -8.40724349e-01 -9.82271552e-01
-3.32070708e-01 -8.67759824e-01 5.81619978e-01 6.28827035e-01
-3.07553560e-01 -9.47391570e-01 6.07620239e-01 9.99009311e-01
1.09153402e+00 -2.89715175e-02 7.86192060e-01 -9.67176139e-01
-1.02990961e+00 -4.04759496e-01 -5.50485730e-01 3.15925330e-01
-1.33212611e-01 -5.77631779e-02 -9.71410990e-01 -8.33320096e-02
8.97855237e-02 -1.43790469e-01 5.90079010e-01 -7.74618909e-02
3.13099205e-01 -5.33822030e-02 1.04393251e-02 -2.00255483e-01
1.46283698e+00 -1.30759194e-01 5.45966804e-01 6.32950306e-01
3.76823157e-01 8.64981651e-01 -1.69455647e-01 5.58199346e-01
7.50019014e-01 5.90913519e-02 6.84260786e-01 6.61640108e-01
-2.34829947e-01 -1.89600512e-01 8.37370604e-02 9.17389512e-01
2.95938969e-01 -2.55712539e-01 -1.40722716e+00 1.07431436e+00
-1.79752362e+00 -1.04755652e+00 -7.48993754e-01 1.89462399e+00
3.09788316e-01 -1.62688211e-01 1.67079777e-01 -1.85407586e-02
7.39315450e-01 3.26363534e-01 -3.31133664e-01 -6.25913918e-01
2.03117087e-01 8.07214305e-02 -6.15230389e-02 3.40034544e-01
-5.87394178e-01 3.49568516e-01 6.49026585e+00 -2.05397476e-02
-6.47726119e-01 1.49478763e-01 1.42697603e-01 -2.21973332e-03
-5.58041930e-01 6.39765859e-02 -6.25681102e-01 3.95717770e-01
1.27168858e+00 -6.94517419e-02 4.91266072e-01 7.33378649e-01
2.89295763e-01 -4.62987393e-01 -7.93621004e-01 4.98135418e-01
1.54138252e-01 -1.47576058e+00 -3.25581193e-01 1.70546725e-01
7.36302316e-01 5.81264377e-01 -2.15210274e-01 3.62505049e-01
4.13276613e-01 -8.46800625e-01 2.62105227e-01 6.63691878e-01
-1.92848176e-01 -4.84852552e-01 8.64848912e-01 4.99445111e-01
-6.17225945e-01 -3.10925156e-01 -4.00757909e-01 -5.40786624e-01
1.39313251e-01 6.48692071e-01 -6.31382167e-01 2.32370477e-02
9.18864012e-01 9.55944136e-02 -8.91404331e-01 1.66552067e+00
-1.50904879e-01 7.79437125e-01 -3.05050731e-01 -4.89842802e-01
3.60971659e-01 2.16494128e-01 3.75625849e-01 9.67601299e-01
5.36107197e-02 -2.57634401e-01 -1.43843949e-01 1.15942097e+00
-4.05760378e-01 2.35311538e-01 -3.92011136e-01 -5.98306417e-01
5.48311412e-01 9.75756586e-01 -3.10148627e-01 -2.39836797e-01
-9.50345039e-01 7.29358137e-01 4.67794269e-01 1.36393562e-01
-2.87172854e-01 -6.67562604e-01 5.11351764e-01 3.15387934e-01
5.91237724e-01 -2.24324867e-01 -2.49858081e-01 -1.08029068e+00
6.34505272e-01 -1.04290545e+00 3.07062000e-01 -6.90457821e-01
-1.33739269e+00 4.45709884e-01 -6.75806820e-01 -5.01129210e-01
-2.27872819e-01 -2.94256479e-01 -8.09958458e-01 1.07820487e+00
-1.57157207e+00 -7.22048283e-01 -6.17850244e-01 -6.72309706e-03
2.02703215e-02 2.39115015e-01 6.48366213e-01 5.91944218e-01
-2.53168255e-01 1.95586950e-01 3.08757007e-01 2.09154963e-01
6.60337985e-01 -1.11488140e+00 6.87929034e-01 7.16898680e-01
1.70476753e-02 9.35186744e-01 6.75415635e-01 -6.04748547e-01
-1.33365774e+00 -5.46007335e-01 1.57517397e+00 -9.62947249e-01
9.93041396e-01 -1.05202816e-01 -1.26126575e+00 5.25902987e-01
5.01736343e-01 -3.00716698e-01 7.61625230e-01 2.97045439e-01
-7.81602144e-01 7.45281205e-02 -1.03990436e+00 3.27524006e-01
4.70511705e-01 -1.00958097e+00 -7.32039392e-01 2.01112088e-02
7.51842618e-01 2.85375148e-01 -9.36292350e-01 -2.56755352e-01
4.27484334e-01 -1.65951395e+00 4.74132895e-01 -7.48323977e-01
8.22950304e-01 -2.02671722e-01 -2.81255454e-01 -7.03880489e-01
-1.11653827e-01 -5.02195656e-01 -2.54827589e-01 1.33030772e+00
3.56333852e-01 -6.09695554e-01 1.34176409e+00 1.28300738e+00
1.79238126e-01 -4.65970337e-01 -7.17201114e-01 -5.86604893e-01
4.18711931e-01 -7.17014372e-02 6.81774318e-01 8.84700418e-01
1.87450320e-01 3.42501283e-01 1.56850606e-01 -5.52040301e-02
4.68259037e-01 -2.31332034e-02 8.02736163e-01 -1.36193788e+00
-5.08382559e-01 -5.96400857e-01 -2.83377945e-01 -1.03911519e+00
-4.45304722e-01 -7.17228532e-01 -1.89719960e-01 -1.77919340e+00
4.64474082e-01 3.50613426e-03 -3.93934138e-02 -1.00485094e-01
-4.40216631e-01 1.11692129e-02 4.06878799e-01 3.71568292e-01
-1.03632236e+00 2.42947072e-01 9.45821404e-01 1.19940475e-01
7.87228495e-02 -2.21492589e-01 -1.01347959e+00 5.03087282e-01
8.37170184e-01 -5.72713852e-01 1.11590371e-01 -8.89609575e-01
1.11954725e+00 2.07944065e-01 2.23687842e-01 -1.23632705e+00
6.28902018e-01 3.54137987e-01 -2.25667670e-01 -3.79186660e-01
-8.70550498e-02 -4.65145081e-01 -5.31322420e-01 3.96742642e-01
-6.37419760e-01 4.85240102e-01 -6.34382144e-02 7.32496738e-01
-4.31552529e-01 -8.01344275e-01 3.69629741e-01 -6.58317208e-01
-2.46348262e-01 -2.00872347e-01 -6.27514720e-01 6.52690530e-01
2.96950251e-01 -1.06824867e-01 -6.51914895e-01 -5.93670249e-01
-3.73121232e-01 4.28543359e-01 5.69853008e-01 5.07646322e-01
6.31833225e-02 -8.15343201e-01 -6.51641309e-01 -3.15046877e-01
2.56769091e-01 -2.80889720e-01 7.82702565e-02 5.87223470e-01
-6.14993036e-01 6.68208063e-01 1.08179852e-01 -3.61730784e-01
-8.86054039e-01 2.30119213e-01 3.79495025e-01 -3.77400547e-01
-2.93192953e-01 7.10983396e-01 -5.47237456e-01 -6.54365122e-01
3.47698390e-01 -1.31015435e-01 -2.02491164e-01 4.87231137e-03
1.05320513e+00 9.93908644e-01 2.90911168e-01 -1.00431748e-01
-1.43537238e-01 -9.36182514e-02 -2.65823931e-01 -5.42039387e-02
1.39951003e+00 -2.00770050e-01 -5.17160535e-01 3.31217885e-01
1.23180652e+00 1.43091157e-01 -7.90427327e-01 -2.54901290e-01
6.88843668e-01 -6.21930718e-01 5.79197295e-02 -8.80109310e-01
-6.52719021e-01 1.08663201e+00 3.09341580e-01 7.36058652e-01
3.54658514e-01 7.71824345e-02 1.14023316e+00 7.34434187e-01
2.07746163e-01 -1.02728641e+00 4.03635085e-01 6.46857381e-01
5.65849900e-01 -1.24328578e+00 -2.66159624e-01 8.60370696e-02
-1.31093010e-01 1.00731671e+00 5.23340702e-01 -2.09800646e-01
2.92743087e-01 -5.34322765e-03 3.16084772e-01 -6.36338830e-01
-8.50341916e-01 9.55654681e-02 -2.24034160e-01 4.26631093e-01
9.12942767e-01 -2.46853366e-01 -2.26557851e-01 2.51020998e-01
2.34030917e-01 3.30288589e-01 9.31517243e-01 1.19682002e+00
-1.07293344e+00 -9.35177743e-01 -5.30525684e-01 5.71717322e-01
-9.79693770e-01 -2.11449116e-01 -5.60846448e-01 3.97471339e-01
-3.87910753e-01 1.47569001e+00 1.01539735e-02 4.11575399e-02
3.62016171e-01 3.48762035e-01 1.95879608e-01 -5.83734810e-01
-1.25692832e+00 -5.40350676e-01 3.48382443e-01 -6.59790754e-01
-1.95308477e-01 -6.86251283e-01 -1.10690427e+00 -5.92729449e-01
-3.98939371e-01 2.99402416e-01 8.31567824e-01 8.35985005e-01
7.98116744e-01 2.82657176e-01 2.14265287e-01 -1.13672063e-01
-9.04854655e-01 -8.21982980e-01 -2.13194340e-01 3.15254033e-01
2.69162625e-01 4.22845781e-02 -4.85076904e-01 -4.33165908e-01]
|
[11.422355651855469, 8.141997337341309]
|
45099a44-297e-4c6d-b8d4-9c4749e602c3
|
a-method-for-detection-of-atrial-fibrillation
| null | null |
https://doi.org/10.1109/CIC.2000.898539
|
http://www.medicine.mcgill.ca/physio/glasslab/pub_pdf/method_2001.pdf
|
A method for detection of atrial fibrillation using RR intervals
|
This work describes a method for automatic detection of atrial fibrillation (AF) based on RR intervals. We define /spl Delta/RR to be the difference between successive RR intervals. The standard density histograms of RR and /spl Delta/RR intervals are determined from data in the MIT-BIH atrial fibrillation/flutter database. The present method estimates the similarity between the standard density histograms and a best density histogram by the Kolmogorov-Smirnov (KS) test. The algorithm returns significance (p) of difference between given histograms. If the p value is smaller than a value (P/sub c/), the test density histogram is significantly different from the standard density histogram. If the test density histogram is not significantly different from the standard density histogram, we say the data is AF: Using the standard density histogram of /spl Delta/RR with P/sub c/=0.01, the average sensitivity is 93.2% and the average specificity is 96.7%.
|
['L. Glass', 'K. Tateno']
|
2000-09-24
| null | null | null |
computers-in-cardiology-2000-9
|
['atrial-fibrillation-detection']
|
['medical']
|
[-6.72235861e-02 -4.15991396e-01 -2.31445923e-01 -2.27230057e-01
-6.35981560e-01 -6.58938408e-01 -1.50714502e-01 4.55236256e-01
-2.77064770e-01 1.34508121e+00 -7.71309063e-02 -8.79632473e-01
-2.44942978e-01 -7.79775977e-01 -1.33695126e-01 -8.26647401e-01
-5.89124799e-01 5.59792280e-01 1.85023621e-01 3.34141493e-01
3.71410310e-01 5.15901089e-01 -1.19236410e+00 3.31014693e-01
7.55725443e-01 1.23107588e+00 -3.35767537e-01 1.04777563e+00
2.60404907e-02 6.33243695e-02 -1.31659567e+00 4.03577358e-01
2.06641197e-01 -8.85828793e-01 -5.17010272e-01 -6.76309168e-01
-4.00833599e-02 -1.78348973e-01 1.43497065e-01 1.12455225e+00
6.91409111e-01 -2.55009562e-01 9.40168679e-01 -1.10522079e+00
-2.37791300e-01 4.32298988e-01 -5.33429503e-01 9.97620702e-01
5.38220167e-01 -1.79285079e-01 3.80447835e-01 -6.66880310e-01
2.62340456e-01 7.43397832e-01 7.60529160e-01 7.12288469e-02
-1.42338324e+00 -8.47181499e-01 -5.78987479e-01 -2.40057081e-01
-2.08523083e+00 1.70626536e-01 3.44647825e-01 -5.14425218e-01
6.30021930e-01 4.72036302e-01 1.13043272e+00 2.54442483e-01
8.72937262e-01 -1.00675553e-01 1.25088787e+00 -5.77241480e-01
3.86561006e-01 5.75283952e-02 3.70340019e-01 3.08032155e-01
7.02000260e-01 1.86262712e-01 -1.43844925e-03 -6.86895549e-01
9.52883422e-01 -4.92960848e-02 -5.25501132e-01 2.41329059e-01
-1.03982306e+00 7.20997095e-01 -1.44508764e-01 6.86997473e-01
-3.46421838e-01 -1.92155197e-01 2.70481288e-01 5.19453883e-01
-7.53648281e-02 2.35828564e-01 -2.57403314e-01 -3.87005180e-01
-1.13461375e+00 2.70239085e-01 9.46551740e-01 5.01790285e-01
5.64134657e-01 5.34419082e-02 -1.70914397e-01 5.67616045e-01
2.50291169e-01 9.06700015e-01 3.73709500e-01 -7.65570104e-01
-1.58486918e-01 4.48468119e-01 2.94334412e-01 -3.90594572e-01
-1.76359147e-01 -9.95786861e-02 -8.85504603e-01 1.40613750e-01
1.06966174e+00 -3.63306046e-01 -7.97266960e-01 1.27742720e+00
-2.42428854e-01 -8.41071457e-02 1.90900430e-01 6.35013044e-01
7.13577628e-01 6.14855051e-01 1.10620923e-01 -1.09339619e+00
1.35118413e+00 4.20291781e-01 -7.98041224e-01 3.33701700e-01
3.23588818e-01 -6.24375105e-01 8.76595140e-01 3.89798343e-01
-8.17606151e-01 -5.60188234e-01 -1.02033043e+00 6.05816722e-01
-2.00684443e-02 -1.92419454e-01 2.64541898e-02 8.87030482e-01
-9.92723286e-01 4.58235025e-01 -4.02658731e-01 -1.21167652e-01
1.01411380e-02 3.74585569e-01 -1.30912289e-01 4.57336426e-01
-1.32358658e+00 5.58324158e-01 3.01653564e-01 -3.01578879e-01
-1.93957701e-01 -6.16206884e-01 -5.88289142e-01 -2.82949992e-02
-4.49494958e-01 -4.11766380e-01 7.59788692e-01 -4.91090626e-01
-8.93765986e-01 6.38911486e-01 -2.30915263e-01 -3.51487905e-01
1.61718607e-01 3.37420888e-02 -7.19976902e-01 1.81133434e-01
8.36507231e-03 -4.19340543e-02 7.64531374e-01 -9.32250142e-01
-5.07438660e-01 -4.72920835e-01 -6.26184702e-01 -1.46278381e-01
3.94351989e-01 1.23367250e-01 1.61902592e-01 -4.69286174e-01
6.18940294e-02 -6.51594639e-01 -4.98982705e-02 -6.09298229e-01
-2.39656627e-01 -5.44837452e-02 7.10615635e-01 -6.67952120e-01
2.00107551e+00 -2.44100142e+00 -7.54074931e-01 9.71855044e-01
2.92277545e-01 2.36812249e-01 6.46931529e-01 2.53155571e-03
-3.95940036e-01 3.66666049e-01 -5.16078055e-01 1.03659737e+00
-5.69246590e-01 -5.15852384e-02 -2.20417276e-01 6.65001392e-01
-3.70564282e-01 4.26988244e-01 -6.21612906e-01 -4.80457902e-01
-6.36183769e-02 5.86648047e-01 -2.78633386e-02 6.39846697e-02
4.15637642e-01 2.67612815e-01 -1.81118280e-01 4.69664156e-01
7.42304683e-01 1.24661485e-02 2.17585385e-01 -3.69061865e-02
-4.43216771e-01 1.61369946e-02 -1.09851551e+00 6.17034078e-01
3.80051404e-01 7.34588623e-01 -6.59387529e-01 -7.07475245e-01
1.70232701e+00 6.13607168e-01 2.78463095e-01 -5.36890566e-01
1.61041975e-01 4.98899400e-01 5.42030275e-01 -2.48271585e-01
-2.04806492e-01 -5.63366771e-01 4.14077267e-02 4.76395816e-01
-4.68672872e-01 -2.90935010e-01 3.23756784e-01 -2.53423691e-01
9.84439194e-01 -4.91766870e-01 8.99349451e-01 -7.91689157e-01
5.83974302e-01 -3.00813675e-01 5.54045856e-01 1.00002289e+00
-3.07440221e-01 6.59524620e-01 1.12340546e+00 -6.13696277e-01
-7.59524226e-01 -1.70152795e+00 -8.36231947e-01 -1.67924985e-01
-2.30023991e-02 -3.67028087e-01 -5.84579825e-01 -5.67824483e-01
2.10077688e-01 5.40643036e-01 -5.01877129e-01 -7.27026463e-02
-4.49977934e-01 -9.79087889e-01 6.94236159e-01 5.37450790e-01
3.36494207e-01 -1.13952851e+00 -8.37927878e-01 1.44081756e-01
-1.28461635e-02 -4.49988574e-01 -2.99561441e-01 1.85625792e-01
-1.36150718e+00 -1.25283051e+00 -7.52923250e-01 -5.85126400e-01
3.88190210e-01 -3.28032792e-01 1.03340578e+00 6.73966110e-02
-6.64312840e-01 -4.53108400e-02 -2.33972952e-01 -3.98932606e-01
-4.91098583e-01 -6.25350654e-01 1.25932857e-01 -5.00882328e-01
6.07174754e-01 -6.04290366e-01 -8.81443977e-01 3.96029502e-01
-3.16452146e-01 -8.93497288e-01 3.03125381e-01 3.92241597e-01
1.09394240e+00 6.42997473e-02 1.02549267e+00 -6.69409633e-01
8.82869422e-01 -2.27970138e-01 -5.78268409e-01 1.06624953e-01
-9.22595143e-01 2.15219613e-02 5.52106261e-01 -4.20663089e-01
-2.35880330e-01 -2.47875571e-01 6.69417083e-02 -3.53606492e-01
-3.34477216e-01 2.74720013e-01 1.14742793e-01 5.09597123e-01
8.79296422e-01 2.84337968e-01 1.49711639e-01 -5.91172986e-02
-3.93460244e-01 9.58807051e-01 7.98722625e-01 -2.54525572e-01
2.51912653e-01 3.65598164e-02 -1.47988852e-02 -9.83569264e-01
-1.81389540e-01 -4.72022265e-01 -2.55732179e-01 -2.18008101e-01
9.26505148e-01 -4.37032968e-01 -7.23584294e-01 8.49142149e-02
-7.20163226e-01 -1.04220644e-01 -5.35387993e-01 1.00739336e+00
-6.71850085e-01 5.17448127e-01 -3.95601422e-01 -1.21042585e+00
-7.90681601e-01 -9.09948409e-01 3.34604472e-01 5.38508654e-01
-8.12627256e-01 -8.37761700e-01 4.24594790e-01 -4.16909605e-01
3.19332927e-01 4.82301623e-01 1.27184355e+00 -7.78537810e-01
3.86009455e-01 -6.21655822e-01 -5.41696697e-02 3.97555172e-01
4.88800287e-01 1.65179506e-01 -5.78408241e-01 -1.01108439e-01
1.57515809e-01 4.07984644e-01 3.29834372e-01 9.59053457e-01
1.11133265e+00 -1.63668796e-01 -3.82367790e-01 3.26376706e-01
1.21567070e+00 1.25372851e+00 1.13602734e+00 5.48966266e-02
-4.98680845e-02 -1.75847292e-01 6.14257634e-01 5.76455116e-01
-1.23236664e-01 1.96513757e-01 -3.69269460e-01 -7.10324273e-02
1.65877536e-01 -3.76439677e-03 1.29743978e-01 6.73488230e-02
-3.54772419e-01 2.32413501e-01 -1.13513672e+00 2.61091113e-01
-1.21278262e+00 -1.12131798e+00 -5.41216195e-01 2.98092175e+00
9.68591034e-01 4.96735871e-01 4.80834872e-01 4.99280572e-01
9.96440053e-01 -1.65909737e-01 -1.56582698e-01 -8.81682575e-01
3.38198915e-02 7.35422015e-01 8.43674615e-02 7.10473895e-01
-7.71591544e-01 1.60033911e-01 6.70421934e+00 3.93904060e-01
-1.00579596e+00 -5.99030018e-01 6.78741693e-01 4.09874618e-01
-1.02136932e-01 2.32001483e-01 -5.90603232e-01 7.39996314e-01
1.45542061e+00 -6.16965890e-01 -1.99807942e-01 5.12617469e-01
3.17101479e-01 -7.91408420e-01 -9.49084580e-01 1.24677885e+00
-7.34768212e-02 -9.32852089e-01 -1.38655663e-01 2.40812838e-01
-2.91553866e-02 -5.36095202e-01 -1.95264310e-01 1.03920095e-01
-2.92583942e-01 -1.07616472e+00 1.25587106e-01 6.47706449e-01
1.36119092e+00 -8.89242470e-01 1.24584591e+00 1.00968264e-01
-1.38374805e+00 2.50610203e-01 -1.78862184e-01 6.60349354e-02
1.58040881e-01 1.13956928e+00 -1.06679153e+00 6.24832213e-02
9.30018544e-01 2.93293566e-01 -2.98296422e-01 1.46786737e+00
3.25411856e-02 7.46072650e-01 -5.95336437e-01 5.63047417e-02
-5.34381449e-01 -3.76332492e-01 6.40052140e-01 1.25416875e+00
4.34594303e-01 2.45512038e-01 -1.67234167e-01 6.46757543e-01
4.18980151e-01 1.86220124e-01 -6.76225483e-01 -7.11637288e-02
9.61072683e-01 5.92740417e-01 -9.20124888e-01 -5.65837860e-01
-2.11565107e-01 3.49995613e-01 -5.00933230e-01 2.61586010e-01
-5.74690461e-01 -8.97817433e-01 6.52140975e-02 2.88372785e-01
-1.05876297e-01 2.50094593e-01 -6.88236237e-01 -5.11471093e-01
-1.24852277e-01 -6.32679641e-01 8.54242265e-01 -4.34863865e-01
-9.42576170e-01 8.88016284e-01 3.83626610e-01 -1.38710237e+00
-5.18131018e-01 -3.65398109e-01 -7.64343083e-01 1.26516449e+00
-6.67916238e-01 1.98945761e-01 -8.90402347e-02 5.49112439e-01
-3.85602899e-02 -1.46471083e-01 1.27024615e+00 -1.13790687e-02
-1.57224182e-02 4.53019321e-01 -3.66537660e-01 1.85855716e-01
6.63375676e-01 -1.53374815e+00 -1.00161858e-01 4.87563223e-01
-4.26631391e-01 7.57655025e-01 9.04722512e-01 -1.00999916e+00
-6.09981179e-01 -5.41580856e-01 1.19482851e+00 -2.14844093e-01
5.00802770e-02 1.72789052e-01 -1.02153528e+00 2.79379338e-01
-1.96428820e-01 1.99341372e-01 1.09884608e+00 -1.47385180e-01
-2.76133746e-01 -1.30203664e-01 -1.43276107e+00 3.11946094e-01
1.33506238e-01 -3.29368621e-01 -8.45015883e-01 -2.17367142e-01
-1.54826447e-01 -4.64374386e-02 -1.39748478e+00 8.58989477e-01
1.12706351e+00 -1.28989458e+00 7.91826248e-01 -8.01122114e-02
-3.38023245e-01 -5.43599665e-01 -8.86462033e-02 -9.36494768e-01
-1.21601544e-01 -4.77482438e-01 7.73364604e-02 7.10570931e-01
6.91050649e-01 -7.29372144e-01 3.80107522e-01 2.88783312e-01
1.92871317e-01 -6.55189037e-01 -9.51089203e-01 -6.92726731e-01
-6.79356083e-02 -3.44723500e-02 1.97199956e-01 6.35437250e-01
5.08913696e-01 5.88330887e-02 2.34792858e-01 -6.20242879e-02
5.10256171e-01 3.38024080e-01 1.97809175e-01 -1.50660443e+00
-8.95192698e-02 -2.32223213e-01 -5.95938027e-01 -1.37987107e-01
-4.24796045e-01 -5.03023565e-01 1.72967848e-03 -1.54300427e+00
-2.80890726e-02 -2.39033595e-01 -4.72959667e-01 2.37897351e-01
-7.73011222e-02 6.35432154e-02 -3.69537055e-01 2.14174941e-01
2.25432053e-01 -3.49477768e-01 6.80962503e-01 3.31897914e-01
-9.56435919e-01 4.37555641e-01 -3.60183954e-01 5.82360089e-01
1.14167631e+00 -4.13546890e-01 -3.66802096e-01 7.92631030e-01
1.90043338e-02 8.27566564e-01 -8.82762223e-02 -1.08475387e+00
-4.30267632e-01 1.02565117e-01 8.67980421e-01 -6.75144613e-01
-3.02420497e-01 -5.43327570e-01 5.42606711e-01 9.56097364e-01
-3.96051183e-02 6.35719061e-01 2.91291952e-01 2.96247631e-01
-1.01800509e-01 -1.97406366e-01 1.05897164e+00 7.05289096e-02
-9.46617946e-02 -8.45191553e-02 -7.32333362e-01 1.55828640e-01
8.94058883e-01 -6.37603939e-01 -2.98007041e-01 -3.24634612e-01
-1.01461363e+00 -2.97375381e-01 3.35109413e-01 -1.04799092e-01
1.01978493e+00 -1.01572227e+00 -6.04108393e-01 5.97805500e-01
2.15771154e-01 -4.59886461e-01 -3.89643013e-02 1.03735149e+00
-9.11504388e-01 1.79666921e-01 -2.39772961e-01 -7.35847890e-01
-1.41364682e+00 1.75914675e-01 6.37144029e-01 9.88045484e-02
-9.41193223e-01 4.29636866e-01 -8.06124657e-02 6.05864644e-01
2.28595939e-02 -3.66580725e-01 -3.92443597e-01 -1.42156795e-01
1.08563876e+00 4.83509809e-01 -3.93516570e-02 -1.90428138e-01
-6.86733902e-01 6.72366977e-01 -7.92694837e-02 -4.36032563e-01
7.49139607e-01 3.05646479e-01 -3.26207012e-01 8.60773087e-01
9.25080836e-01 2.86668837e-01 -6.28259480e-01 6.48936272e-01
-8.68370384e-02 -5.24619102e-01 -2.70348519e-01 -1.00021005e+00
-5.96202314e-01 4.31122065e-01 1.32036293e+00 5.27712226e-01
1.11804366e+00 1.36078030e-01 2.82031268e-01 -8.52301791e-02
1.74004212e-01 -7.62848258e-01 -3.62738758e-01 2.74808586e-01
6.82471871e-01 -5.24937630e-01 1.27843916e-01 -2.64898181e-01
-7.64704943e-01 1.31210363e+00 2.52510756e-01 -3.82092416e-01
1.18360877e+00 7.20737755e-01 3.94144684e-01 -8.85992125e-02
-3.97434831e-01 2.29903758e-01 1.76133543e-01 6.66898131e-01
8.03524196e-01 2.96733171e-01 -1.11822939e+00 6.15208387e-01
-3.79521221e-01 3.37402582e-01 6.12276316e-01 7.95481205e-01
-9.16820705e-01 -8.90156329e-01 -4.93147284e-01 9.88583982e-01
-5.16379297e-01 -2.99781170e-02 -2.35115960e-01 9.93946612e-01
-6.28284141e-02 9.02847648e-01 6.76658034e-01 -3.00934255e-01
4.08747673e-01 5.81405938e-01 5.14878631e-01 -2.54512757e-01
-3.53639454e-01 5.64865291e-01 -1.06307909e-01 -1.39126599e-01
-5.59502840e-02 -5.73477089e-01 -2.04278922e+00 -2.76072957e-02
-1.00295998e-01 7.60555923e-01 3.37546319e-01 6.59147203e-01
-1.05368033e-01 5.94146065e-02 5.11802435e-01 1.36644796e-01
-1.43986523e-01 -9.11309361e-01 -1.20263171e+00 5.01952656e-02
5.70741475e-01 -5.46742976e-01 -8.55673134e-01 -2.96538603e-02]
|
[14.1600923538208, 3.1771185398101807]
|
9643b313-886b-44d2-a50f-3b114467f07d
|
semantic-line-detection-and-its-applications
| null | null |
http://openaccess.thecvf.com/content_iccv_2017/html/Lee_Semantic_Line_Detection_ICCV_2017_paper.html
|
http://openaccess.thecvf.com/content_ICCV_2017/papers/Lee_Semantic_Line_Detection_ICCV_2017_paper.pdf
|
Semantic Line Detection and Its Applications
|
Semantic lines characterize the layout of an image. Despite their importance in image analysis and scene understanding, there is no reliable research for semantic line detection. In this paper, we propose a semantic line detector using a convolutional neural network with multi-task learning, by regarding the line detection as a combination of classification and regression tasks. We use convolution and max-pooling layers to obtain multi-scale feature maps for an input image. Then, we develop the line pooling layer to extract a feature vector for each candidate line from the feature maps. Next, we feed the feature vector into the parallel classification and regression layers. The classification layer decides whether the line candidate is semantic or not. In case of a semantic line, the regression layer determines the offset for refining the line location. Experimental results show that the proposed detector extracts semantic lines accurately and reliably. Moreover, we demonstrate that the proposed detector can be used successfully in three applications: horizon estimation, composition enhancement, and image simplification.
|
['Chang-Su Kim', 'Han-Ul Kim', 'Jun-Tae Lee', 'Chul Lee']
|
2017-10-01
| null | null | null |
iccv-2017-10
|
['line-detection']
|
['computer-vision']
|
[ 3.91652972e-01 -3.95024538e-01 -3.35635208e-02 -4.33025420e-01
-3.49112481e-01 -2.24899366e-01 1.80154026e-01 4.60462064e-01
-5.05958080e-01 3.60910654e-01 -2.70615101e-01 -1.50392607e-01
-3.30808498e-02 -9.82500911e-01 -8.10691655e-01 -4.97111112e-01
3.46890211e-01 -9.57356021e-02 6.60144687e-01 -1.14762269e-01
6.58970058e-01 8.64935219e-01 -1.38191009e+00 4.40627098e-01
7.35879958e-01 1.28534794e+00 2.58954674e-01 5.90917110e-01
-2.93913275e-01 8.24692070e-01 -7.54187167e-01 -1.25712380e-01
2.15912655e-01 -4.09797281e-01 -5.41388392e-01 4.33035821e-01
5.29179454e-01 -4.62380588e-01 -1.83562294e-01 1.03578067e+00
4.24585283e-01 -5.16744107e-02 4.80184466e-01 -1.33670950e+00
-3.34872544e-01 4.08412755e-01 -1.15911531e+00 5.02922647e-02
4.02348600e-02 -5.20850718e-01 8.37020338e-01 -9.87124383e-01
4.14198041e-01 9.69751418e-01 9.86186981e-01 -1.12747349e-01
-7.08351493e-01 -5.28493822e-01 5.76254278e-02 3.29697430e-01
-1.69031918e+00 -3.03764552e-01 1.11691427e+00 -3.35079253e-01
5.35840452e-01 -1.15498053e-02 4.43406403e-01 2.68161297e-01
1.70210853e-01 1.03624833e+00 9.14835691e-01 -6.50536239e-01
3.51458788e-02 2.69068539e-01 1.01037323e-01 1.02540326e+00
2.49711409e-01 -4.79519814e-01 -4.98573333e-01 2.32113451e-01
1.05088842e+00 -3.26294154e-02 -2.57723153e-01 -4.04652983e-01
-9.17761564e-01 6.69462383e-01 6.24481559e-01 3.59262973e-01
-3.33536834e-01 1.95687860e-01 3.15950006e-01 -8.26793090e-02
2.43282810e-01 2.36589566e-01 -2.97058225e-01 4.01312649e-01
-1.32123208e+00 2.65952826e-01 4.73849326e-01 1.20180345e+00
1.01654112e+00 -2.00415388e-01 -1.79143190e-01 1.10582113e+00
6.00644127e-02 4.75724518e-01 2.30009854e-01 -5.12151480e-01
5.15112817e-01 5.79410255e-01 2.12466240e-01 -1.32985806e+00
-9.28450286e-01 -4.83005822e-01 -6.83711529e-01 3.17400426e-01
4.22406018e-01 -2.99000025e-01 -5.26543736e-01 9.48284984e-01
1.31771013e-01 5.41892797e-02 -5.70604280e-02 8.88928235e-01
7.23788619e-01 6.13994479e-01 -3.14686447e-01 1.04912683e-01
1.53611863e+00 -1.39808726e+00 -6.16596818e-01 -4.71857458e-01
6.79034650e-01 -1.04676497e+00 6.56699002e-01 3.65644962e-01
-8.67960215e-01 -7.03909993e-01 -1.44927192e+00 -2.88682073e-01
-4.73163068e-01 1.11131680e+00 5.90946198e-01 4.07093614e-01
-7.02442825e-01 4.29716647e-01 -4.66865331e-01 -3.28033477e-01
5.49255669e-01 7.64789507e-02 -5.04812486e-02 -1.20345198e-01
-8.58030438e-01 6.94938064e-01 5.74681580e-01 4.66990381e-01
-1.04773983e-01 -4.56555068e-01 -9.98362005e-01 3.14443380e-01
4.26349610e-01 -5.76693356e-01 1.08837199e+00 -1.15415871e+00
-1.31425035e+00 6.76785648e-01 -2.54474193e-01 -4.90708917e-01
6.40226066e-01 -3.85458440e-01 -4.09485996e-01 2.80427486e-01
2.15280622e-01 5.28194249e-01 1.09219420e+00 -1.17478836e+00
-1.42898428e+00 -1.19676247e-01 -5.61616570e-02 3.21879417e-01
-2.72420287e-01 1.06190529e-03 -7.38152385e-01 -5.74247062e-01
3.35804164e-01 -5.14110923e-01 -1.82762817e-01 2.40018368e-01
-7.87373543e-01 -1.18932150e-01 1.14220202e+00 -6.52553916e-01
1.30054581e+00 -2.15569854e+00 -5.13025820e-01 4.39600110e-01
5.29515967e-02 4.61338051e-02 3.47859063e-03 1.33055970e-01
1.01410158e-01 -1.62636504e-01 -1.60454169e-01 -1.55246690e-01
-3.84462357e-01 -5.02445579e-01 -1.00066796e-01 5.03081322e-01
4.90068793e-01 6.12443864e-01 -5.34619689e-01 -5.49164236e-01
5.29342711e-01 3.73412997e-01 -4.03924555e-01 -7.35855103e-02
4.41042706e-03 2.79286169e-02 -5.87997139e-01 4.65901464e-01
1.02472901e+00 -9.93997455e-02 -7.39129707e-02 -7.32183933e-01
-5.22890389e-01 -3.96487266e-01 -1.45567143e+00 1.70355737e+00
-5.24152994e-01 1.04776645e+00 -3.15294087e-01 -8.26090395e-01
1.42400110e+00 -3.48049521e-01 5.28633595e-01 -8.94322336e-01
2.45029628e-01 3.70032430e-01 -1.99363440e-01 -3.81534427e-01
9.74964499e-01 5.05322874e-01 -2.14705348e-01 1.54409362e-02
-9.71435905e-02 -2.82777190e-01 3.25921178e-01 -2.81436712e-01
6.71683609e-01 3.25000077e-01 2.21579373e-01 -1.01133594e-02
8.61617625e-01 1.34139970e-01 5.24839699e-01 6.38039231e-01
2.54437029e-02 6.31065905e-01 3.88422012e-01 -4.81776476e-01
-1.37245452e+00 -7.41814077e-01 -1.16057061e-01 8.24410617e-01
7.78840065e-01 -3.64195824e-01 -9.43895519e-01 -3.34197134e-01
-1.87228210e-02 3.94830465e-01 -2.10852697e-01 7.11623058e-02
-7.04724252e-01 -6.11767650e-01 5.46843588e-01 6.72326207e-01
1.11970174e+00 -7.57208824e-01 -9.59205508e-01 2.89301604e-01
1.29687399e-01 -1.41439855e+00 -4.92162704e-01 1.89596757e-01
-4.61485654e-01 -1.19403517e+00 -7.08368182e-01 -1.15376770e+00
8.48305345e-01 6.34549439e-01 4.14151847e-01 -7.23495558e-02
-5.75996339e-01 -1.04881683e-03 -1.71360075e-01 -4.09842968e-01
5.53724729e-02 5.56925125e-03 -4.84943599e-01 2.60988444e-01
1.70022622e-01 1.13178398e-02 -6.99510336e-01 3.42929870e-01
-5.91339350e-01 5.78071415e-01 6.94384933e-01 6.37018263e-01
6.18986011e-01 4.21617478e-01 3.80667925e-01 -9.02499616e-01
6.87805355e-01 2.05402330e-01 -9.62811410e-01 3.24475259e-01
-2.37013534e-01 1.33378997e-01 6.71233416e-01 1.45643502e-01
-1.15075874e+00 4.25105989e-01 -1.60962746e-01 -2.07543686e-01
-1.85530797e-01 2.72161901e-01 -2.00182255e-02 -2.37161964e-01
4.75698590e-01 6.21742941e-03 -5.08432925e-01 -3.30361247e-01
5.52687228e-01 7.25964963e-01 6.44214094e-01 -4.23475690e-02
8.61720026e-01 6.36832654e-01 1.63767517e-01 -1.24426556e+00
-8.68984759e-01 -7.36274898e-01 -1.04112041e+00 -3.19035292e-01
1.07177401e+00 -7.75627017e-01 -6.00675702e-01 8.76584411e-01
-1.40580809e+00 1.18616614e-02 6.97145090e-02 2.80230314e-01
-5.87365627e-01 2.67650992e-01 -4.50952709e-01 -5.73629200e-01
-3.58164400e-01 -1.19280064e+00 1.04149485e+00 8.28085780e-01
1.04833111e-01 -8.92215967e-01 -4.67440516e-01 1.70450851e-01
2.21832216e-01 6.24881834e-02 8.69391978e-01 -4.28359210e-01
-7.44380772e-01 -3.68732661e-01 -7.63212383e-01 2.17005327e-01
-2.22817976e-02 2.37241149e-01 -1.12974286e+00 8.77233297e-02
-5.71864963e-01 -1.03349462e-01 8.30570877e-01 5.64764977e-01
1.56541228e+00 9.46526378e-02 -4.17206317e-01 8.98744226e-01
1.73209417e+00 3.75655651e-01 7.22503841e-01 7.20562279e-01
9.18138504e-01 4.60249603e-01 7.57175386e-01 6.21768951e-01
2.82835543e-01 5.36200821e-01 -1.64793190e-02 -6.65789962e-01
-2.64743030e-01 -3.81059974e-01 -1.71200931e-01 3.50181431e-01
2.77878016e-01 -2.40276277e-01 -7.76038826e-01 1.54958144e-01
-1.83664358e+00 -9.07157719e-01 -3.88272732e-01 2.04682231e+00
1.86238095e-01 3.91089410e-01 -1.69299901e-01 3.15380633e-01
1.01143897e+00 3.90217490e-02 -4.06030834e-01 -5.41520119e-01
-5.61944127e-01 1.08891025e-01 1.16214681e+00 3.14909369e-01
-1.63564670e+00 1.27739739e+00 5.24327183e+00 1.02318418e+00
-1.33344638e+00 -2.65284449e-01 6.12924516e-01 5.78511178e-01
-2.58573052e-02 -9.57143903e-02 -9.44373727e-01 5.09624043e-03
9.40223411e-02 -1.13876380e-01 1.01836003e-01 6.98617816e-01
3.83995324e-01 -5.72264075e-01 -8.46464813e-01 1.33939314e+00
3.28470916e-01 -1.42770278e+00 -4.97361124e-02 -5.18021405e-01
7.00341165e-01 -3.59762967e-01 -1.04274094e-01 -1.53315693e-01
-3.34879637e-01 -7.80078232e-01 1.07054996e+00 5.41337729e-01
5.32264352e-01 -9.44737077e-01 5.71139932e-01 1.51951313e-01
-1.42372251e+00 -1.82627574e-01 -3.80568117e-01 1.96981847e-01
7.64098689e-02 5.41760385e-01 -9.90738630e-01 6.41376019e-01
3.80009502e-01 6.62550688e-01 -7.72111177e-01 1.67698002e+00
-3.96065027e-01 -8.89501907e-03 -2.98582524e-01 1.18397445e-04
3.21802467e-01 -3.68511051e-01 -8.63308785e-04 1.33758926e+00
4.34607714e-01 -4.45170552e-01 2.52514094e-01 7.82016635e-01
-1.16483949e-01 5.25547385e-01 -3.88973922e-01 1.24753848e-01
4.81609553e-01 1.48730469e+00 -1.34588373e+00 -1.07590720e-01
-6.33982003e-01 1.50351751e+00 3.20951790e-01 4.36969548e-01
-8.12961042e-01 -1.19972467e+00 2.12265059e-01 -1.73154473e-01
4.64784384e-01 -2.73444027e-01 -6.67957485e-01 -7.73588300e-01
4.18906957e-02 -1.33151934e-01 1.12208560e-01 -1.01540601e+00
-5.31838119e-01 3.99152249e-01 -3.65517139e-01 -1.22931576e+00
7.06004575e-02 -7.60531545e-01 -7.48026311e-01 8.53032529e-01
-1.81917787e+00 -1.34336710e+00 -5.87810338e-01 2.91326910e-01
7.02515721e-01 2.07951274e-02 4.89717692e-01 4.30347025e-01
-7.23639011e-01 5.40483832e-01 3.07459623e-01 4.07530338e-01
6.93063259e-01 -9.58350718e-01 2.51462668e-01 9.01090026e-01
-9.28366184e-02 7.60869905e-02 5.32549560e-01 -4.45456415e-01
-9.42871332e-01 -1.22402120e+00 6.85629368e-01 4.73007351e-01
4.77893919e-01 -3.57414126e-01 -6.59912527e-01 4.33266968e-01
-1.34208739e-01 1.21310335e-02 3.28035057e-01 -2.92958260e-01
-9.86208916e-02 -2.88394034e-01 -9.29107487e-01 5.21604896e-01
7.85949647e-01 -4.38781947e-01 -1.64136276e-01 1.81760073e-01
4.63383436e-01 -4.08992290e-01 -7.06770778e-01 2.93368667e-01
6.86846614e-01 -9.56986725e-01 1.01948166e+00 6.66878372e-02
4.51719165e-01 -5.84733605e-01 2.47729316e-01 -1.29473484e+00
-2.79054493e-01 1.05535192e-02 6.65085375e-01 1.18075669e+00
5.83368003e-01 -2.62608051e-01 8.22028995e-01 -2.72909533e-02
-4.18648988e-01 -6.48378909e-01 -5.18922150e-01 -2.80481100e-01
-2.98804104e-01 -3.12789887e-01 5.37134111e-01 7.55345345e-01
-3.55401129e-01 2.48204708e-01 -3.52694124e-01 5.10544538e-01
5.88299513e-01 5.43527007e-01 7.48188972e-01 -1.06125259e+00
7.95620158e-02 -6.67709649e-01 -5.68606913e-01 -1.25289237e+00
8.09572637e-02 -6.28215373e-01 1.68370873e-01 -1.82716405e+00
-7.44141638e-02 -5.17343581e-01 -3.98321822e-02 1.69470727e-01
-1.18715748e-01 1.65540978e-01 2.55124718e-01 -2.43856553e-02
-4.79616910e-01 3.41725945e-01 1.09488738e+00 -3.25823575e-01
-3.98135602e-01 2.94947606e-02 -3.21686000e-01 1.07096505e+00
9.55204427e-01 7.53553361e-02 -1.89056396e-01 -5.66657424e-01
6.15368485e-02 -1.52013421e-01 2.60422885e-01 -1.24545670e+00
6.13561630e-01 6.27922714e-02 8.99530590e-01 -1.05141330e+00
1.83303326e-01 -8.14018428e-01 -3.89074832e-01 4.12568629e-01
-1.72558084e-01 -1.16835102e-01 3.31518859e-01 1.32272094e-01
-2.10016891e-01 -5.12806833e-01 7.97201633e-01 2.40920156e-01
-1.49881291e+00 2.62840390e-01 -1.54932678e-01 -3.85638654e-01
1.26056015e+00 -4.29757267e-01 -4.61432368e-01 -7.74269970e-03
-4.04294729e-01 4.79807258e-01 3.04630995e-01 4.84834641e-01
1.07351410e+00 -1.27956510e+00 -4.69163090e-01 5.05281866e-01
4.16690379e-01 4.23453338e-02 3.54925305e-01 5.63750029e-01
-1.13272178e+00 4.11290497e-01 -4.44210649e-01 -4.99345779e-01
-1.18842351e+00 3.27989936e-01 5.93392551e-01 2.69892037e-01
-5.71216941e-01 6.85972691e-01 1.06031887e-01 9.79810059e-02
9.01782140e-02 -4.59564477e-01 -4.54762310e-01 1.59662604e-01
5.45254052e-01 4.71692115e-01 7.58998916e-02 -6.00771964e-01
-2.41092190e-01 1.02350736e+00 9.73127112e-02 9.66917351e-02
1.16554546e+00 -1.19546838e-01 1.91305578e-02 4.15704101e-01
1.28330135e+00 1.87111661e-01 -1.17708957e+00 -1.79786012e-01
1.52967528e-01 -4.24362153e-01 2.39487052e-01 -2.58230031e-01
-9.69340801e-01 8.79709542e-01 6.03086948e-01 -1.89617660e-03
1.13075960e+00 -3.42513025e-01 7.30375290e-01 2.29870290e-01
2.79275477e-01 -1.49802804e+00 -9.13876295e-02 3.37516248e-01
7.06586003e-01 -9.79185343e-01 -2.14156788e-03 -8.31176043e-01
-3.84462953e-01 1.82421386e+00 7.36992002e-01 -2.30907857e-01
4.40948159e-01 4.78461742e-01 -3.68944705e-02 -1.00023136e-01
-3.65168750e-02 -3.02805185e-01 2.60409534e-01 4.47709352e-01
3.64052951e-01 -1.51939495e-02 -1.97402880e-01 2.45442808e-01
-2.33522654e-01 4.50721309e-02 6.41401470e-01 8.33798230e-01
-1.00745451e+00 -6.80237293e-01 -5.54785967e-01 4.21863765e-01
-2.76586283e-02 -1.11138038e-02 -1.16865344e-01 6.18531108e-01
2.77641594e-01 7.32544661e-01 5.00652254e-01 -2.38157347e-01
6.59827292e-01 -2.75674760e-01 5.78236639e-01 -3.00556839e-01
-2.43278876e-01 1.79096594e-01 6.55380040e-02 -4.01889592e-01
-1.93945035e-01 -3.94387126e-01 -1.55285203e+00 2.02976950e-02
-4.73457992e-01 -2.38892771e-02 1.06617701e+00 7.29658961e-01
-1.19721927e-01 8.29396248e-01 8.69875014e-01 -5.79675972e-01
-1.74631342e-01 -6.40460074e-01 -7.29264081e-01 2.72787988e-01
1.60743311e-01 -4.61161613e-01 4.12966982e-02 -1.41373957e-02]
|
[8.341418266296387, -1.5741368532180786]
|
a2ab69a5-bd80-4dfb-9889-012e61c9ee55
|
gaitmast-motion-aware-spatio-temporal-feature
|
2210.11817
| null |
https://arxiv.org/abs/2210.11817v2
|
https://arxiv.org/pdf/2210.11817v2.pdf
|
Motion Matters: A Novel Motion Modeling For Cross-View Gait Feature Learning
|
As a unique biometric that can be perceived at a distance, gait has broad applications in person authentication, social security, and so on. Existing gait recognition methods suffer from changes in viewpoint and clothing and barely consider extracting diverse motion features, a fundamental characteristic in gaits, from gait sequences. This paper proposes a novel motion modeling method to extract the discriminative and robust representation. Specifically, we first extract the motion features from the encoded motion sequences in the shallow layer. Then we continuously enhance the motion feature in deep layers. This motion modeling approach is independent of mainstream work in building network architectures. As a result, one can apply this motion modeling method to any backbone to improve gait recognition performance. In this paper, we combine motion modeling with one commonly used backbone~(GaitGL) as GaitGL-M to illustrate motion modeling. Extensive experimental results on two commonly-used cross-view gait datasets demonstrate the superior performance of GaitGL-M over existing state-of-the-art methods.
|
['Junping Zhang', 'Hongming Shan', 'Yuzhen Zhang', 'Jiaqi Gao', 'Jingqi Li']
|
2022-10-21
| null | null | null | null |
['gait-recognition']
|
['computer-vision']
|
[-1.30604103e-01 -8.79126906e-01 -3.72994810e-01 -1.51198223e-01
-1.05050400e-01 -3.13917577e-01 2.03541398e-01 -5.15303791e-01
-3.56876224e-01 2.96647012e-01 3.72444272e-01 1.49150774e-01
2.45670930e-01 -6.61427557e-01 5.84210046e-02 -9.03148413e-01
-3.53788048e-01 -2.70820975e-01 2.16161564e-01 -2.32987136e-01
5.22842519e-02 4.33390617e-01 -1.09107423e+00 8.70646015e-02
1.88277394e-01 7.41715014e-01 -8.48479569e-02 7.19964206e-01
9.18247253e-02 3.73843402e-01 -4.80863154e-01 -4.00265515e-01
3.18400934e-02 -3.68456662e-01 -5.64115345e-01 1.44547313e-01
2.66316831e-01 -5.93201816e-01 -9.07935202e-01 8.75730455e-01
9.05253530e-01 2.86037475e-03 3.82346302e-01 -1.38493216e+00
-6.84652150e-01 8.94272998e-02 -1.05836523e+00 2.14399785e-01
5.96922815e-01 4.79092360e-01 6.27769947e-01 -9.39517677e-01
5.40392637e-01 1.25593364e+00 1.02501965e+00 1.02563596e+00
-7.53238678e-01 -5.74624002e-01 1.80743560e-01 5.41643739e-01
-1.32392812e+00 -3.55103463e-01 1.07735848e+00 -2.05014527e-01
8.60494316e-01 8.92786160e-02 8.76962185e-01 1.43561709e+00
2.29543939e-01 1.10176420e+00 4.92635727e-01 -1.06524505e-01
-9.03943479e-02 -8.56376350e-01 1.90193832e-01 1.00813234e+00
5.42870939e-01 -7.95176774e-02 -8.94420922e-01 -6.65581822e-02
8.33412230e-01 1.51482657e-01 -3.96993786e-01 -2.42661402e-01
-1.28589261e+00 2.76637197e-01 3.07373136e-01 4.29608136e-01
-1.70635074e-01 6.55022085e-01 4.44257498e-01 1.59764215e-01
-1.13565244e-01 -6.29835606e-01 6.03113994e-02 -5.25034606e-01
-1.22073483e+00 2.61464030e-01 6.14920199e-01 5.81185520e-01
4.03107315e-01 4.80295241e-01 -1.19997755e-01 6.95963323e-01
6.99026048e-01 7.01334119e-01 6.64622247e-01 -6.75121725e-01
6.74546719e-01 4.59155709e-01 2.28420906e-02 -1.44233418e+00
-4.93464559e-01 -7.53218904e-02 -1.33205819e+00 -2.64582902e-01
3.78158391e-01 -8.11028630e-02 -8.50695372e-01 1.65813792e+00
4.72054705e-02 6.16355360e-01 -2.23536357e-01 8.42136085e-01
8.17262709e-01 2.77846187e-01 4.67690118e-02 1.55657664e-01
1.29177320e+00 -7.98236609e-01 -5.60623765e-01 -2.73779631e-01
6.52914524e-01 -6.02013111e-01 7.91529417e-01 2.01764077e-01
-7.63696134e-01 -8.83842289e-01 -1.40753293e+00 2.19254732e-01
-2.89595246e-01 3.85962635e-01 6.84298873e-01 1.10639763e+00
-1.17306674e+00 8.75527680e-01 -1.21574426e+00 -5.36183119e-01
3.47094446e-01 5.35447717e-01 -5.51966429e-01 -3.01930606e-01
-1.22472811e+00 3.87346238e-01 1.57911748e-01 6.15621328e-01
-4.81701314e-01 2.00620890e-01 -1.17922425e+00 -1.74619749e-01
-3.83409321e-01 -1.07860970e+00 7.26948202e-01 -3.03133756e-01
-1.26218474e+00 5.91500700e-01 -5.82907617e-01 -1.71711504e-01
5.55950940e-01 -1.64010718e-01 -7.36331940e-01 2.57390380e-01
8.10145736e-02 5.39129615e-01 1.15228713e+00 -7.67061770e-01
-2.75749564e-01 -4.75788742e-01 -4.51192200e-01 -2.53555384e-02
-7.59057581e-01 -1.76306605e-01 -8.23342323e-01 -9.55247879e-01
2.81490296e-01 -9.38369811e-01 -2.32674807e-01 -1.89521420e-03
-2.33293712e-01 5.60904667e-02 1.21109486e+00 -1.00755322e+00
1.66324699e+00 -2.09639382e+00 6.84425384e-02 3.06613803e-01
1.97361246e-01 4.14619327e-01 -2.22188205e-01 3.75819832e-01
1.50828063e-01 3.58188674e-02 -1.92607075e-01 -6.56344712e-01
5.82656153e-02 1.73378214e-01 1.68326721e-01 6.74395442e-01
2.14551345e-01 1.09846687e+00 -8.98930252e-01 -4.33391929e-01
2.20888779e-01 6.28079891e-01 -5.21104276e-01 -1.56551093e-01
6.32462382e-01 4.67665315e-01 -4.33867097e-01 1.09910762e+00
9.15866792e-01 -2.82856941e-01 2.84762025e-01 -4.43754494e-01
3.93357098e-01 -2.11400062e-01 -1.11257517e+00 2.11568165e+00
1.79625332e-01 4.56615269e-01 -2.14368612e-01 -9.78523433e-01
9.68878329e-01 3.02287191e-01 9.78290796e-01 -3.44289184e-01
-5.58999553e-03 1.20353408e-01 1.50168035e-02 -5.23037493e-01
6.44416511e-01 1.88939422e-01 -2.95654655e-01 3.61409873e-01
2.00872853e-01 7.01882064e-01 1.46998972e-01 -8.72463882e-02
1.05449045e+00 3.97221178e-01 1.99782208e-01 1.73927754e-01
7.47635484e-01 -4.82649237e-01 8.53410721e-01 6.21436596e-01
-9.44579482e-01 6.40321255e-01 -1.02731057e-01 -5.81369221e-01
-8.53286147e-01 -1.06556261e+00 4.85528827e-01 6.11784220e-01
3.35809678e-01 -7.11034417e-01 -6.99154496e-01 -6.06550276e-01
4.11329372e-03 -3.87935549e-01 -4.00907844e-01 -4.40828860e-01
-1.16950023e+00 -1.16737437e+00 1.06359828e+00 1.09506798e+00
1.38801539e+00 -7.06243217e-01 -5.55665314e-01 3.97349715e-01
-6.51970565e-01 -1.24277270e+00 -7.84252405e-01 -7.21727610e-01
-8.07412744e-01 -7.97706664e-01 -1.15304244e+00 -9.74911034e-01
2.17213660e-01 5.98613322e-01 7.78989077e-01 3.91220331e-01
-3.33112985e-01 6.04691565e-01 -1.49453416e-01 1.36344358e-01
2.25467905e-01 1.13881618e-01 4.96816546e-01 2.40195081e-01
7.99071252e-01 -7.61370480e-01 -9.86635387e-01 1.62785023e-01
-4.64603364e-01 -2.05694541e-01 3.90404522e-01 9.58226979e-01
-6.72226911e-03 -2.88368110e-02 4.69265163e-01 1.06924087e-01
5.12007236e-01 -2.82700241e-01 1.58353299e-01 1.62802622e-01
-9.25383866e-02 -1.48285329e-01 1.81763142e-01 -3.90514553e-01
-5.89442194e-01 2.36302316e-02 -2.46437535e-01 -3.97893965e-01
-1.18342549e-01 5.18255472e-01 -4.56871837e-01 -4.68180120e-01
2.93720998e-02 7.21645057e-01 6.38771942e-03 -4.04210269e-01
1.76659256e-01 4.60529417e-01 7.39955246e-01 -6.36206031e-01
1.03781652e+00 7.34649837e-01 2.76533455e-01 -1.08227658e+00
9.55034941e-02 -6.90412462e-01 -8.98425102e-01 -5.20429969e-01
7.69933343e-01 -8.65049303e-01 -1.04431152e+00 1.03342354e+00
-1.05422652e+00 1.84917673e-01 2.80131727e-01 4.73292053e-01
-5.91517210e-01 1.21374655e+00 -1.00609934e+00 -7.49058485e-01
-4.41502780e-01 -9.75434661e-01 1.09342539e+00 3.58521461e-01
-3.52639347e-01 -1.08796966e+00 -1.64652523e-02 1.82964742e-01
3.07778388e-01 3.56988519e-01 4.17356879e-01 1.78598743e-02
-4.54719573e-01 -3.17032635e-01 -2.31134929e-02 9.01275128e-03
4.56173629e-01 -3.36297005e-02 -9.73560452e-01 -6.95361257e-01
-1.55302286e-01 2.35315636e-02 1.10107350e+00 4.43210959e-01
8.47200036e-01 -7.45115355e-02 -5.73653162e-01 8.78538370e-01
1.16693377e+00 1.64252743e-01 7.68354237e-01 4.79450822e-01
1.11299598e+00 3.74541014e-01 3.72700036e-01 4.16675895e-01
5.01406014e-01 7.73864567e-01 -4.04702825e-03 7.78883547e-02
-1.14616014e-01 -3.43678385e-01 8.31900895e-01 1.28253698e+00
-5.20882130e-01 -2.95785945e-02 -9.23406839e-01 5.27423859e-01
-2.04879785e+00 -1.35827243e+00 3.33177656e-01 1.84914231e+00
2.53115982e-01 9.29730311e-02 5.54533362e-01 4.85062629e-01
7.94025838e-01 5.28411925e-01 -4.22307312e-01 4.41142768e-02
-2.54131883e-01 1.12759769e-02 4.02457684e-01 3.67483720e-02
-1.47917783e+00 8.34180653e-01 6.22957325e+00 6.52557075e-01
-1.08597541e+00 -1.94420382e-01 1.56259209e-01 9.65093374e-02
1.81751132e-01 -2.95758247e-01 -8.86414349e-01 6.22575283e-01
7.43666291e-01 -1.05918787e-01 -8.08947310e-02 5.87237000e-01
2.93453574e-01 4.25566316e-01 -8.47296119e-01 1.50452423e+00
1.02892555e-01 -1.17349529e+00 4.30246025e-01 2.41647497e-01
1.75234318e-01 -5.78348100e-01 2.96611011e-01 -1.60485078e-02
-2.56843809e-02 -1.03424358e+00 3.16420019e-01 4.13173229e-01
7.60517538e-01 -8.11028719e-01 7.09868491e-01 8.58630091e-02
-2.29508877e+00 -2.28390493e-03 -2.92897433e-01 2.30776872e-02
4.25184965e-01 1.23719603e-01 -2.87395656e-01 1.05476308e+00
7.16337204e-01 1.46492124e+00 -6.44007325e-01 1.08764625e+00
1.64767206e-01 6.08425796e-01 -3.78819145e-02 9.22490805e-02
3.69071394e-01 -5.86790070e-02 5.86996198e-01 1.51518846e+00
6.06702745e-01 -2.28168741e-01 2.99811155e-01 3.67958337e-01
1.27001867e-01 -2.96574175e-01 -8.95587087e-01 -6.41830862e-02
3.52182806e-01 1.00181305e+00 -7.45672643e-01 -6.55626953e-02
-6.93154812e-01 1.39997768e+00 -2.61712104e-01 4.08509612e-01
-7.00794637e-01 -3.20245892e-01 1.00451064e+00 -1.54191092e-01
3.05435181e-01 -8.41540873e-01 4.92807738e-02 -1.71403015e+00
1.54022038e-01 -6.63367629e-01 4.21176672e-01 -2.12642267e-01
-1.37389219e+00 1.88264295e-01 -2.99903899e-01 -1.86607754e+00
-4.36660528e-01 -8.59478235e-01 -7.19632268e-01 8.50430131e-01
-1.14217174e+00 -1.60985100e+00 -4.52612758e-01 8.87934387e-01
3.65127861e-01 -3.16394180e-01 7.37789989e-01 6.50015950e-01
-6.63342953e-01 9.60599661e-01 -3.19020003e-01 9.07124281e-01
6.22946799e-01 -8.67085516e-01 1.26898611e+00 1.36557508e+00
1.73423842e-01 9.69204307e-01 1.79254323e-01 -7.82980442e-01
-1.62982237e+00 -9.95826602e-01 6.08522952e-01 -2.10338727e-01
4.88061249e-01 -7.28779286e-02 -8.30417216e-01 4.98693585e-01
-3.29625234e-02 1.13662928e-01 8.94958317e-01 -4.76333618e-01
-3.97044957e-01 7.97235742e-02 -9.95397210e-01 8.38730931e-01
1.65687573e+00 -5.23116529e-01 -3.48571479e-01 -5.70350468e-01
2.16957018e-01 -2.44265818e-03 -7.90924549e-01 4.93235290e-01
1.05473423e+00 -9.06090498e-01 1.45523489e+00 -6.95093215e-01
-1.83452129e-01 -6.42322123e-01 -3.13262045e-01 -9.89022255e-01
-5.34599245e-01 -7.58843422e-01 -7.43163168e-01 1.19728911e+00
-4.04684007e-01 -4.09278512e-01 1.30339444e+00 2.01843992e-01
4.02984917e-01 -4.47639972e-01 -9.41318512e-01 -9.29887116e-01
-1.21489979e-01 -2.67766863e-01 5.96669078e-01 1.09141803e+00
8.64125714e-02 2.70414539e-02 -1.00469708e+00 1.32233158e-01
1.11686003e+00 7.93965608e-02 8.25180292e-01 -1.04708600e+00
-3.05310279e-01 -4.31476086e-01 -1.24575508e+00 -1.37719417e+00
-5.79143167e-02 -6.11977994e-01 -3.46307874e-01 -1.46422052e+00
1.35910958e-01 2.50936627e-01 -3.99137110e-01 3.43166411e-01
-4.18939292e-01 3.66884589e-01 4.64088082e-01 3.48415971e-01
-3.07910740e-01 5.67467093e-01 1.32953036e+00 -4.31793183e-01
2.66365781e-02 -1.37342751e-01 -4.05069172e-01 9.12860215e-01
8.94360542e-01 9.16777775e-02 -3.56382430e-01 -3.22177678e-01
-3.55078727e-01 -2.51986329e-02 5.21007359e-01 -1.42697310e+00
5.42140663e-01 -1.54245431e-02 9.35052752e-01 -7.72551954e-01
5.42521119e-01 -4.05643880e-01 3.60272340e-02 1.03496182e+00
4.19728965e-01 5.35183847e-01 1.21657655e-01 6.46930635e-01
-3.46086502e-01 2.93951362e-01 4.28266943e-01 -1.70553431e-01
-1.40985858e+00 6.40174508e-01 -4.53243971e-01 -3.16287547e-01
5.27184308e-01 -7.78269410e-01 -3.43891919e-01 -4.62043375e-01
-1.01126862e+00 9.12895203e-02 4.23975676e-01 6.09194160e-01
1.13310838e+00 -1.88409102e+00 -4.95411813e-01 3.51978332e-01
1.87069885e-02 -6.17209256e-01 3.14974874e-01 7.58477747e-01
-5.36998868e-01 3.47623497e-01 -5.97147346e-01 -7.55690694e-01
-1.39863813e+00 1.98269352e-01 3.76535803e-01 -2.72702187e-01
-7.30482936e-01 5.97754419e-01 -1.97154596e-01 -4.41156179e-02
7.05544874e-02 -1.27242565e-01 -3.31109136e-01 -9.03646797e-02
7.50675499e-01 4.43812460e-01 -2.60162473e-01 -1.02686417e+00
-7.37581670e-01 1.13976443e+00 4.37773690e-02 -2.62396187e-01
8.76997650e-01 -3.10412467e-01 1.73389837e-01 2.56875992e-01
1.11189032e+00 -2.58483410e-01 -1.17355061e+00 2.46946998e-02
3.64498347e-01 -5.54388940e-01 -6.61339819e-01 5.38185164e-02
-1.24117243e+00 1.13901007e+00 7.96084166e-01 -2.50320852e-01
1.21481001e+00 -8.31301570e-01 1.24937129e+00 4.26172853e-01
8.02218974e-01 -9.29710984e-01 4.49085474e-01 3.41430277e-01
2.48124704e-01 -8.27094853e-01 -2.15703204e-01 -3.25009674e-01
-3.47623736e-01 1.23024642e+00 7.25364566e-01 -6.81437701e-02
7.21278131e-01 3.25509697e-01 -1.27001328e-03 8.53977650e-02
-3.68969560e-01 -2.34380171e-01 3.16966265e-01 1.10221505e+00
5.30448198e-01 -1.15165770e-01 -1.81682810e-01 5.33280611e-01
-4.56843898e-03 2.74888456e-01 2.24468306e-01 1.36111891e+00
-3.18697602e-01 -1.30604959e+00 -3.52293551e-01 8.36774707e-02
-4.06831056e-01 7.75578916e-02 -2.95828670e-01 6.52008891e-01
7.75434598e-02 8.22213233e-01 -3.38850170e-01 -1.06076980e+00
6.40466139e-02 1.63373664e-01 7.42716551e-01 -3.93873490e-02
-2.21689403e-01 -4.90609230e-03 -2.92022619e-02 -8.30496550e-01
-8.65859807e-01 -6.61340714e-01 -1.06269312e+00 -8.18402112e-01
3.52653921e-01 -2.61628151e-01 9.74463671e-02 7.60974646e-01
2.95637399e-01 3.36502463e-01 3.92585486e-01 -1.12944293e+00
-3.88507873e-01 -5.98762035e-01 -6.65575504e-01 4.66630667e-01
4.60569531e-01 -7.80340314e-01 1.19394213e-01 2.96841532e-01]
|
[14.270843505859375, 1.4307408332824707]
|
a8449e73-6d90-4907-aa9f-dd09853c2532
|
self-taught-convolutional-neural-networks-for
|
1701.00185
| null |
http://arxiv.org/abs/1701.00185v1
|
http://arxiv.org/pdf/1701.00185v1.pdf
|
Self-Taught Convolutional Neural Networks for Short Text Clustering
|
Short text clustering is a challenging problem due to its sparseness of text
representation. Here we propose a flexible Self-Taught Convolutional neural
network framework for Short Text Clustering (dubbed STC^2), which can flexibly
and successfully incorporate more useful semantic features and learn non-biased
deep text representation in an unsupervised manner. In our framework, the
original raw text features are firstly embedded into compact binary codes by
using one existing unsupervised dimensionality reduction methods. Then, word
embeddings are explored and fed into convolutional neural networks to learn
deep feature representations, meanwhile the output units are used to fit the
pre-trained binary codes in the training process. Finally, we get the optimal
clusters by employing K-means to cluster the learned representations. Extensive
experimental results demonstrate that the proposed framework is effective,
flexible and outperform several popular clustering methods when tested on three
public short text datasets.
|
['Guanhua Tian', 'Bo Xu', 'Suncong Zheng', 'Peng Wang', 'Jun Zhao', 'Jiaming Xu']
|
2017-01-01
| null | null | null | null |
['text-clustering', 'short-text-clustering']
|
['natural-language-processing', 'natural-language-processing']
|
[-8.25753249e-03 -2.62086570e-01 -3.50478500e-01 -6.61831379e-01
-3.21216404e-01 -3.93559128e-01 5.96630573e-01 3.45446467e-02
-4.17846233e-01 1.16048656e-01 5.37745535e-01 -2.45499481e-02
-3.09728503e-01 -5.79038620e-01 -3.64765525e-01 -8.15984249e-01
3.03203940e-01 7.12714434e-01 -2.95404404e-01 5.84391020e-02
4.30489987e-01 5.83289191e-02 -1.55627251e+00 2.21759230e-01
1.05722058e+00 8.70233178e-01 2.01437354e-01 4.42784488e-01
-6.03135705e-01 7.51478374e-01 -1.80840477e-01 -1.74450740e-01
-6.23406656e-02 -8.43216106e-02 -9.07634139e-01 4.07162994e-01
-1.51494101e-01 -3.89202893e-01 -6.94138467e-01 9.59815085e-01
3.47349733e-01 4.26374674e-01 9.99847949e-01 -1.00349867e+00
-1.17293453e+00 1.03709912e+00 -6.28900528e-01 -1.59619674e-01
-3.06678675e-02 -1.71594843e-01 1.20286906e+00 -9.67082560e-01
2.97407448e-01 1.43617094e+00 5.48792601e-01 6.45234883e-01
-1.02174008e+00 -6.89788401e-01 2.54183501e-01 9.25792605e-02
-1.46110976e+00 -4.01113927e-01 9.19913709e-01 -3.49831372e-01
8.35104406e-01 -1.39950374e-02 3.24307591e-01 1.12163842e+00
-1.05912209e-01 1.00270402e+00 4.67550516e-01 -2.75329292e-01
1.56212091e-01 1.25785666e-02 5.84316850e-01 5.59223771e-01
2.15877101e-01 -5.49914062e-01 -1.80600747e-01 -9.28777903e-02
2.31049016e-01 8.87652516e-01 -4.02089916e-02 -3.38287979e-01
-1.08228397e+00 1.21832836e+00 7.89163888e-01 4.42169815e-01
-2.21526787e-01 4.73726988e-01 6.80549860e-01 4.66277227e-02
4.82718378e-01 -2.63787829e-03 -4.80421692e-01 2.04749852e-02
-9.36973989e-01 6.90722466e-03 5.52991927e-01 9.03408945e-01
8.58423531e-01 -7.84007236e-02 -1.78059444e-01 1.18687570e+00
6.94544315e-01 2.52839535e-01 1.41030145e+00 -5.93065739e-01
5.29531837e-01 1.05176604e+00 -2.53636271e-01 -1.12277687e+00
-5.07248819e-01 -1.74936816e-01 -1.17139566e+00 -5.66825271e-01
-2.33770147e-01 -3.01501185e-01 -1.00316048e+00 1.18782532e+00
1.57405406e-01 1.29588125e-02 3.03129613e-01 6.64308667e-01
9.80915725e-01 9.62966502e-01 7.55466381e-03 -2.66600261e-03
1.35624611e+00 -1.08258843e+00 -8.72186065e-01 -2.12179869e-01
9.40283597e-01 -5.54753125e-01 1.11855304e+00 1.68640047e-01
-4.26355749e-01 -5.21557748e-01 -9.15202200e-01 -2.94595420e-01
-7.15187430e-01 3.29939097e-01 5.91039479e-01 6.37227356e-01
-8.77244294e-01 3.80592316e-01 -7.90455699e-01 -4.11006749e-01
8.74748170e-01 5.66690803e-01 1.23189865e-02 -4.52404410e-01
-1.12358475e+00 1.16451681e-01 9.89084423e-01 -1.28295664e-02
-5.38247406e-01 -1.56918138e-01 -1.08407962e+00 5.09535611e-01
1.18404835e-01 -3.03005099e-01 7.92753041e-01 -1.02558458e+00
-1.68647730e+00 6.14943326e-01 -6.78768232e-02 -2.83300787e-01
4.55239303e-02 -1.56736150e-01 -4.08766121e-01 2.25676268e-01
2.12761492e-01 9.54402685e-01 1.05473745e+00 -1.12506497e+00
-4.17038321e-01 -3.69980395e-01 -4.53261465e-01 4.82016653e-01
-1.19322765e+00 -1.08844914e-01 -6.69104636e-01 -8.89359951e-01
2.45267719e-01 -8.00712526e-01 -3.69082749e-01 -3.25544238e-01
-5.37225366e-01 -9.02426779e-01 1.21303725e+00 -2.66808331e-01
1.33780992e+00 -2.32113171e+00 1.76632449e-01 3.39403868e-01
5.26407719e-01 6.22544885e-02 -7.59245008e-02 4.02729779e-01
-6.81379139e-02 9.16193873e-02 -3.49724114e-01 -4.48341161e-01
4.24324006e-01 3.68075311e-01 -2.67864197e-01 4.02940959e-01
-6.70984900e-03 1.05889690e+00 -6.44126475e-01 -7.17671216e-01
4.13806707e-01 4.26284760e-01 -5.70817947e-01 1.16517909e-01
-2.91633040e-01 -2.00383335e-01 -6.35953367e-01 2.28799447e-01
7.47832894e-01 -5.26997566e-01 1.94452181e-01 1.45322233e-01
1.94491640e-01 -1.88057367e-02 -1.12941539e+00 1.85085213e+00
-3.27237576e-01 5.64198136e-01 -2.84312010e-01 -1.49471331e+00
1.04261482e+00 -4.64525223e-02 4.27050591e-01 -3.29854727e-01
4.96246427e-01 -5.39212264e-02 -3.46388638e-01 -7.05390036e-01
6.64576888e-01 7.87588656e-02 -3.45574707e-01 8.53625178e-01
3.41174960e-01 1.69533625e-01 -6.49244413e-02 4.73802418e-01
6.75564468e-01 -3.98854256e-01 -1.79834887e-01 -3.64557266e-01
6.85000479e-01 -5.02502024e-02 2.07472414e-01 3.45876217e-01
-2.68032011e-02 6.25297725e-01 3.02811146e-01 -4.83658582e-01
-7.36441970e-01 -6.25068784e-01 -2.71640807e-01 1.43822753e+00
4.88276109e-02 -7.33253062e-01 -8.78399372e-01 -1.02079129e+00
1.32340521e-01 4.01186883e-01 -8.64230692e-01 -3.42541456e-01
-1.54942811e-01 -7.78892577e-01 5.06652296e-01 6.76648259e-01
4.47990447e-01 -9.73975003e-01 4.16109664e-03 1.25302404e-01
-2.10619211e-01 -7.99726188e-01 -6.30185604e-01 5.97444832e-01
-8.33816469e-01 -8.88218164e-01 -7.81713486e-01 -1.44176865e+00
8.89228046e-01 6.76427960e-01 5.58634222e-01 3.21827471e-01
-2.60735333e-01 3.58784765e-01 -6.76225722e-01 3.38265523e-02
7.15674274e-03 4.38053519e-01 1.16424844e-01 2.86413223e-01
1.04230845e+00 -2.62638360e-01 -8.33107591e-01 1.47054931e-02
-1.31195331e+00 -2.67907500e-01 5.29103220e-01 1.08910573e+00
3.57252359e-01 5.13594568e-01 6.39462113e-01 -1.00185049e+00
9.27086473e-01 -7.73894191e-01 -1.50680453e-01 -6.98335320e-02
-7.27396011e-01 2.74677753e-01 1.03615355e+00 -3.90849888e-01
-9.25281584e-01 2.51001805e-01 -8.48700628e-02 -5.13663948e-01
-3.02209228e-01 6.32731438e-01 -1.28713921e-01 3.34163636e-01
4.38239336e-01 6.20052993e-01 5.84977195e-02 -4.92241949e-01
8.60647678e-01 1.54096997e+00 2.77041018e-01 -4.96777326e-01
7.24541783e-01 5.26300609e-01 -7.04585731e-01 -7.47853935e-01
-7.97014594e-01 -7.68103123e-01 -9.76298869e-01 3.42245579e-01
1.14209223e+00 -1.09469950e+00 -5.04798353e-01 4.28674668e-01
-8.20655227e-01 -1.23689562e-01 1.12254068e-01 3.44443619e-01
-3.60283703e-01 6.67935312e-01 -7.18553662e-01 -3.45819533e-01
-5.40470362e-01 -1.10109830e+00 1.05226147e+00 2.73684382e-01
-6.02624156e-02 -1.26093042e+00 -1.25887141e-01 3.11070025e-01
1.67734891e-01 -2.60447711e-01 1.16973174e+00 -1.04748118e+00
6.25087619e-02 -3.33745033e-01 -4.66099083e-01 3.21098506e-01
3.15211654e-01 -5.10969497e-02 -8.19296658e-01 -4.80701536e-01
-1.81365013e-01 -7.16971099e-01 1.33303833e+00 3.67376536e-01
1.94486773e+00 -5.11895001e-01 -5.88547528e-01 7.57794082e-01
1.43543744e+00 -2.19122574e-01 3.69727135e-01 2.55637407e-01
1.31631041e+00 6.06460929e-01 1.99845091e-01 6.36431992e-01
6.18346572e-01 -2.54891776e-02 3.42683822e-01 1.17994092e-01
3.86298597e-01 -1.88148022e-01 1.83134392e-01 1.41859555e+00
5.71183026e-01 -1.08004726e-01 -9.07724023e-01 4.97547448e-01
-1.99694371e+00 -7.78676271e-01 -9.31728110e-02 1.42180967e+00
8.58169913e-01 2.07955930e-02 -1.19300172e-01 3.97960693e-01
1.02189147e+00 2.98980325e-01 -6.28294826e-01 -3.16023201e-01
1.94730490e-01 1.48419335e-01 2.67203540e-01 5.04574403e-02
-1.25050402e+00 1.26969469e+00 5.94923019e+00 1.08107388e+00
-9.74890709e-01 -8.12981371e-03 7.33140469e-01 -3.75229344e-02
-3.79555464e-01 -3.44266295e-01 -6.26435816e-01 7.93238819e-01
1.04148400e+00 -3.80022265e-02 3.55437964e-01 9.68033552e-01
-7.34925345e-02 5.26358247e-01 -8.18949044e-01 1.29823852e+00
3.52550119e-01 -1.43011796e+00 4.98649776e-01 -1.29592285e-01
9.69042003e-01 -2.48917267e-02 3.10952067e-01 6.66501105e-01
6.99841678e-01 -1.31730676e+00 1.79439679e-01 2.21038774e-01
9.06683207e-01 -1.21820915e+00 5.60266256e-01 2.34064102e-01
-1.15843534e+00 -4.10125554e-01 -9.17393506e-01 7.19997212e-02
-5.27443349e-01 4.69009310e-01 -4.59250391e-01 3.91993999e-01
8.80673110e-01 1.41960132e+00 -7.87042558e-01 4.69515890e-01
1.83847591e-01 6.96361899e-01 -4.22314592e-02 -5.16389608e-01
8.47631931e-01 -2.39750296e-01 -2.32353568e-01 1.37589216e+00
1.52271956e-01 1.19477458e-01 2.18852639e-01 6.90787852e-01
-5.56560218e-01 3.20953637e-01 -4.31889892e-01 -3.01621675e-01
5.32864392e-01 1.44286036e+00 -1.01554012e+00 -5.02874196e-01
-3.15932631e-01 1.26233220e+00 4.83515471e-01 4.47973400e-01
-5.39524615e-01 -1.12173629e+00 4.43117470e-01 -5.77192307e-01
6.80010855e-01 -4.48795259e-02 -2.28280127e-01 -1.35401118e+00
-1.92924932e-01 -6.79128706e-01 6.56737864e-01 -6.39156640e-01
-1.57290757e+00 3.74998301e-01 -5.36414504e-01 -1.19006741e+00
1.53219234e-02 -7.42768049e-01 -7.84436285e-01 4.65022296e-01
-1.48423815e+00 -1.06855953e+00 -3.89691859e-01 1.00514150e+00
8.47822011e-01 -5.69620967e-01 8.86587024e-01 1.21866383e-01
-9.60374475e-01 7.19660938e-01 1.37585604e+00 5.72666824e-01
6.72263563e-01 -1.39587033e+00 2.21768349e-01 4.23255920e-01
4.97438014e-02 7.53756642e-01 2.34107692e-02 -4.73053843e-01
-1.44269395e+00 -1.62264371e+00 4.84697938e-01 -3.59634697e-01
7.53038585e-01 -6.79215074e-01 -9.18063343e-01 6.93544626e-01
3.29515010e-01 1.46079659e-01 1.20368028e+00 7.82096013e-02
-4.95098084e-01 -1.33574098e-01 -9.41868663e-01 4.45536196e-01
7.11810827e-01 -5.78964174e-01 -9.03738976e-01 4.45275813e-01
1.24122119e+00 -1.59957539e-02 -8.01903367e-01 -3.16772759e-01
2.83007443e-01 -5.00676036e-01 8.19946647e-01 -6.34196997e-01
8.25594783e-01 -3.03905662e-02 -1.64468169e-01 -1.31999707e+00
-5.61300099e-01 -4.48146433e-01 -3.10618468e-02 1.48487592e+00
2.23082490e-02 -4.30307597e-01 8.30366492e-01 3.10169850e-02
-1.59107089e-01 -7.72258341e-01 -5.72511435e-01 -2.25550875e-01
5.92456400e-01 -2.95392126e-01 8.49729180e-01 1.50478351e+00
2.72087395e-01 4.83011633e-01 -9.77277011e-02 -5.79394214e-02
6.26103222e-01 1.84849218e-01 4.61021334e-01 -1.51363087e+00
2.31752038e-01 -5.39257824e-01 -2.95138597e-01 -1.42914701e+00
6.13188386e-01 -1.45348883e+00 1.38501793e-01 -1.61465168e+00
3.93175751e-01 -5.44012547e-01 -5.67577481e-01 3.28988820e-01
-4.55556750e-01 2.02988572e-02 -2.14097872e-01 5.09086967e-01
-1.08408403e+00 1.05643880e+00 7.32292175e-01 -4.81891274e-01
-6.86315745e-02 -2.21300021e-01 -1.18919337e+00 4.88504291e-01
7.88831413e-01 -5.75880408e-01 -3.72395009e-01 -6.23004317e-01
-4.29924205e-02 -4.65090185e-01 -1.20351262e-01 -7.75492787e-01
3.25201929e-01 8.88580680e-02 9.25433218e-01 -8.96940053e-01
-1.97199896e-01 -9.42014158e-01 -7.08137572e-01 3.76591742e-01
-7.62552202e-01 -2.66494840e-01 -1.81861669e-01 8.73173594e-01
-1.22805454e-01 -4.64449435e-01 5.87181032e-01 1.12056471e-01
-5.38402498e-01 4.58239853e-01 -5.64711452e-01 -3.21795382e-02
8.46991420e-01 -1.15709670e-01 -5.35337627e-02 2.07600780e-02
-4.62340862e-01 8.83918703e-01 2.63383836e-01 6.25500977e-01
8.18823993e-01 -1.79950261e+00 -4.47211504e-01 2.50568956e-01
3.91745329e-01 2.94771135e-01 3.77859533e-01 3.17200899e-01
-4.35936719e-01 5.62715888e-01 4.86703143e-02 -8.35078597e-01
-7.42375016e-01 8.83664370e-01 2.88189631e-02 3.35783474e-02
-7.52547503e-01 7.18019009e-01 4.50410359e-02 -9.44021642e-01
5.52564800e-01 -3.07367355e-01 -5.91918051e-01 1.88796729e-01
5.16274869e-01 1.53716311e-01 -1.67187855e-01 -5.87129235e-01
-2.79319435e-01 6.89070582e-01 -5.60412347e-01 4.33287084e-01
1.55652869e+00 -5.69648504e-01 -2.90832788e-01 3.17422152e-01
1.82562220e+00 -3.78461540e-01 -1.15143895e+00 -5.46711266e-01
1.16773598e-01 -1.40118495e-01 1.99032575e-01 -2.29571939e-01
-1.20937991e+00 1.12320292e+00 6.72371268e-01 2.29668602e-01
8.14056575e-01 -1.28233209e-02 1.03820658e+00 9.88433838e-01
-5.90431929e-01 -1.48771262e+00 5.90328932e-01 7.01163709e-01
2.52448261e-01 -1.29293120e+00 -1.53684899e-01 2.19077587e-01
-7.78080583e-01 1.44002867e+00 6.71339095e-01 -3.76254112e-01
8.50826025e-01 -1.25318915e-01 1.52563691e-01 -4.11823481e-01
-5.22328079e-01 -2.40232795e-01 -2.48693079e-02 6.24239326e-01
4.59793627e-01 -9.02667791e-02 -4.93061058e-02 1.10147810e+00
-2.30188355e-01 -3.82174581e-01 3.39790136e-01 6.56393170e-01
-7.50043869e-01 -7.48683691e-01 -2.52256930e-01 9.02692080e-01
-2.52712280e-01 -1.43131271e-01 -4.22929257e-01 3.21341544e-01
2.83717625e-02 9.27811921e-01 3.83846730e-01 -5.46343446e-01
-2.29903325e-01 2.46801317e-01 -1.59513801e-01 -8.75944018e-01
-3.92933279e-01 6.37776628e-02 -5.78861654e-01 -1.34262174e-01
-3.45166117e-01 -4.50287312e-01 -1.75292408e+00 -2.85495281e-01
-3.94946098e-01 4.97513473e-01 3.69765133e-01 1.05875087e+00
4.93887156e-01 5.87748885e-01 1.11179769e+00 -9.50803220e-01
-5.90692699e-01 -1.08319342e+00 -6.15784228e-01 7.47066617e-01
2.13197529e-01 -5.03227592e-01 -4.65737700e-01 2.70758063e-01]
|
[10.403873443603516, 6.751841068267822]
|
15ae1a6f-dc09-4a3e-ba3d-ed95fc991d1b
|
lambert-layout-aware-language-modeling-using
|
2002.08087
| null |
https://arxiv.org/abs/2002.08087v5
|
https://arxiv.org/pdf/2002.08087v5.pdf
|
LAMBERT: Layout-Aware (Language) Modeling for information extraction
|
We introduce a simple new approach to the problem of understanding documents where non-trivial layout influences the local semantics. To this end, we modify the Transformer encoder architecture in a way that allows it to use layout features obtained from an OCR system, without the need to re-learn language semantics from scratch. We only augment the input of the model with the coordinates of token bounding boxes, avoiding, in this way, the use of raw images. This leads to a layout-aware language model which can then be fine-tuned on downstream tasks. The model is evaluated on an end-to-end information extraction task using four publicly available datasets: Kleister NDA, Kleister Charity, SROIE and CORD. We show that our model achieves superior performance on datasets consisting of visually rich documents, while also outperforming the baseline RoBERTa on documents with flat layout (NDA \(F_{1}\) increase from 78.50 to 80.42). Our solution ranked first on the public leaderboard for the Key Information Extraction from the SROIE dataset, improving the SOTA \(F_{1}\)-score from 97.81 to 98.17.
|
['Michał Turski', 'Filip Graliński', 'Tomasz Stanisławek', 'Rafał Powalski', 'Łukasz Garncarek', 'Piotr Halama', 'Bartosz Topolski']
|
2020-02-19
| null | null | null | null |
['key-information-extraction']
|
['natural-language-processing']
|
[ 1.06722862e-01 1.56286910e-01 3.39849368e-02 -4.51136023e-01
-1.01301408e+00 -1.11548245e+00 8.17565918e-01 2.67669767e-01
-5.49634576e-01 4.56006289e-01 4.51903552e-01 -5.01157880e-01
-1.54662490e-01 -8.91874254e-01 -9.59362209e-01 -4.97247785e-01
-2.10319068e-02 4.98506457e-01 1.36294305e-01 -2.66047984e-01
5.83287477e-01 4.60813582e-01 -1.31986201e+00 8.68353069e-01
7.06547618e-01 7.56648481e-01 5.98701000e-01 8.91076028e-01
-4.51495856e-01 5.11397958e-01 -7.13993907e-01 -4.78948861e-01
8.70316997e-02 -6.88512698e-02 -1.08751678e+00 2.31339354e-02
8.22102845e-01 -2.83497989e-01 -2.43671224e-01 7.31480837e-01
2.39937231e-01 -2.20999166e-01 6.56674445e-01 -6.18476093e-01
-9.81352031e-01 1.00679588e+00 -7.03470349e-01 -2.04178737e-03
2.06189230e-01 -1.29727602e-01 1.46511877e+00 -9.85969424e-01
1.07610857e+00 1.32134545e+00 3.61813277e-01 2.09831893e-01
-1.12393308e+00 -3.45464110e-01 4.93034422e-01 2.09472463e-01
-1.42375803e+00 -2.51941532e-01 3.53327990e-01 -3.78170639e-01
1.18725359e+00 1.29058555e-01 3.07677776e-01 1.01970613e+00
-1.63728207e-01 1.05631518e+00 1.10972881e+00 -6.86839521e-01
-4.05413006e-03 -1.14021730e-02 3.46887320e-01 8.70023906e-01
2.28247046e-01 -5.61842978e-01 -6.25890076e-01 2.37571523e-01
3.64595950e-01 -4.54337388e-01 -2.22623050e-01 -2.92667687e-01
-1.27851367e+00 6.02351546e-01 5.99645257e-01 3.90937865e-01
1.25689104e-01 1.63687676e-01 4.51821655e-01 3.34238023e-01
2.74460644e-01 7.65110254e-01 -4.60645556e-01 -8.92959535e-02
-9.79046941e-01 3.15935165e-01 6.40465677e-01 1.16575062e+00
6.43945396e-01 -3.93707186e-01 -3.50379258e-01 7.82718718e-01
8.76268670e-02 5.56936443e-01 1.61276802e-01 -7.23858297e-01
8.00012529e-01 5.78050554e-01 9.60625149e-03 -7.46681094e-01
-4.90632594e-01 -5.85673869e-01 -6.68769896e-01 5.56052551e-02
6.58267856e-01 2.41144314e-01 -1.14478815e+00 1.52425325e+00
-1.31541476e-01 -6.56794429e-01 1.97278168e-02 6.40116930e-01
7.30036020e-01 9.19650733e-01 -2.75272399e-01 3.60976726e-01
1.64761686e+00 -1.10915077e+00 -5.69224358e-01 -3.14233840e-01
9.54512298e-01 -8.71763527e-01 1.47827899e+00 9.10151660e-01
-9.28883076e-01 -5.65254748e-01 -1.30349076e+00 -6.94747090e-01
-7.65499711e-01 5.41807592e-01 4.63757426e-01 3.89547467e-01
-1.20937240e+00 5.24175048e-01 -5.92383564e-01 -3.85308355e-01
4.58459079e-01 2.85994083e-01 -4.55837548e-01 -3.24440628e-01
-8.56563866e-01 7.66769171e-01 5.53729594e-01 -6.68713674e-02
-6.31503820e-01 -6.29438937e-01 -7.39791989e-01 1.79318726e-01
6.44027531e-01 -5.04168212e-01 1.15410411e+00 -6.29093230e-01
-1.28096557e+00 9.17431176e-01 -3.43000680e-01 -4.11536038e-01
6.77099764e-01 -6.04043663e-01 -5.60303293e-02 1.66891962e-01
1.59717992e-01 7.77127147e-01 5.37685573e-01 -1.40993822e+00
-6.25992298e-01 -3.68211895e-01 2.08956107e-01 -4.63956688e-03
-5.92089415e-01 -6.16011396e-02 -8.96875739e-01 -6.53469324e-01
-2.82468423e-02 -9.62380946e-01 2.29500979e-01 -4.68776189e-02
-8.68366182e-01 -2.67582804e-01 8.15436959e-01 -9.81084228e-01
1.19856155e+00 -2.08716536e+00 2.81703770e-01 4.76976961e-01
2.47724101e-01 5.80698922e-02 -3.79288018e-01 6.06102109e-01
5.18582128e-02 4.09114093e-01 -2.28765488e-01 -3.93057972e-01
9.94972438e-02 3.05284932e-03 -5.94882011e-01 1.99301332e-01
1.42706484e-01 9.11578953e-01 -8.10757220e-01 -3.82938921e-01
2.14755759e-02 3.29646170e-01 -5.30038118e-01 7.45361820e-02
-4.90762949e-01 -3.27728316e-02 -2.68256277e-01 4.47191179e-01
6.67097330e-01 -4.37594235e-01 3.17105263e-01 -1.28086358e-01
-3.19946110e-01 5.55701435e-01 -1.09540558e+00 1.98477972e+00
-7.06597507e-01 1.03243709e+00 -1.15706101e-01 -5.53277433e-01
9.75260794e-01 -1.65969387e-01 3.00594121e-02 -8.37886989e-01
-2.16712624e-01 1.85532019e-01 -5.15808500e-02 -2.37056315e-01
7.75540888e-01 5.08019745e-01 -3.67282778e-01 5.67810416e-01
8.11457634e-04 -1.27591491e-01 5.57028890e-01 6.62831187e-01
1.17481124e+00 1.81980491e-01 5.02354205e-02 -4.77497965e-01
5.77993929e-01 1.40272323e-02 -6.31442666e-02 8.94788504e-01
5.27433932e-01 7.04493105e-01 8.55284214e-01 -3.61822993e-01
-1.36687756e+00 -1.00970864e+00 1.06549300e-02 1.27161157e+00
-1.33670837e-01 -1.09301853e+00 -8.64125967e-01 -7.14667559e-01
-5.29459566e-02 8.49517047e-01 -8.35507870e-01 9.06571075e-02
-8.96947026e-01 -3.38916212e-01 6.47643924e-01 6.70250833e-01
4.62600291e-01 -7.03690946e-01 -2.97673762e-01 -8.88026357e-02
-1.78538889e-01 -1.20297492e+00 -4.04480100e-01 3.62367421e-01
-5.32750905e-01 -7.78272688e-01 -7.71923780e-01 -7.06583560e-01
6.80405259e-01 1.33570671e-01 1.21866965e+00 2.14829102e-01
-3.23625535e-01 -2.66901590e-03 -3.82616967e-01 -3.26222867e-01
-2.98630983e-01 7.01587975e-01 -5.03808141e-01 -3.72720301e-01
1.10488191e-01 -8.95436406e-02 -4.42479461e-01 -4.40677591e-02
-9.56216633e-01 4.02529120e-01 7.06133246e-01 7.05612600e-01
4.68496710e-01 -8.44614357e-02 3.03110391e-01 -1.07813370e+00
5.67624271e-01 -4.78849746e-02 -7.03670740e-01 4.94856358e-01
-5.49054742e-01 5.52086830e-01 9.81802344e-01 -4.10074219e-02
-1.08314347e+00 -4.46559004e-02 -5.01728915e-02 1.04504362e-01
-7.58258775e-02 4.14858788e-01 -4.20833588e-01 3.72463852e-01
4.59620565e-01 2.90640712e-01 -4.97689456e-01 -8.67222607e-01
7.78723955e-01 7.57827759e-01 8.81549656e-01 -6.88628733e-01
8.46756697e-01 5.67679524e-01 -1.39435276e-01 -8.34131896e-01
-9.67422068e-01 -3.82612765e-01 -9.46141481e-01 2.13754058e-01
6.54384375e-01 -8.13903928e-01 -5.11589646e-01 4.15337712e-01
-1.43342471e+00 -3.07107449e-01 1.70596857e-02 -4.77282666e-02
-2.99878418e-01 3.47751588e-01 -5.93354344e-01 -4.71576154e-01
-3.26958865e-01 -1.02339327e+00 1.36025727e+00 -3.48155275e-02
-3.51988256e-01 -6.60406172e-01 -1.61261350e-01 3.68693769e-01
1.35917231e-01 1.25316635e-01 1.45878267e+00 -4.95338827e-01
-9.11270380e-01 -1.19968401e-02 -6.00647211e-01 2.34401047e-01
-9.87259895e-02 2.25324541e-01 -1.05696356e+00 -2.61887968e-01
-6.94975197e-01 -2.34080076e-01 1.29077029e+00 -2.82221977e-02
1.51686919e+00 -2.32073799e-01 -3.94239128e-01 6.77468896e-01
1.41301513e+00 -8.00144076e-02 6.90657318e-01 7.94533014e-01
8.74585271e-01 6.94859624e-01 3.74135703e-01 4.18421417e-01
5.59225142e-01 6.75147533e-01 3.99461538e-01 -2.73198597e-02
-4.89821225e-01 -5.75261116e-01 3.67205262e-01 6.11485064e-01
9.69771221e-02 -6.33027613e-01 -1.02768290e+00 8.06575775e-01
-1.58771574e+00 -6.66251659e-01 -2.21728757e-01 1.87389529e+00
9.02587652e-01 3.39775145e-01 -1.32167533e-01 8.89007449e-02
3.81353915e-01 2.84850657e-01 -1.70019463e-01 -6.43433511e-01
-3.79597068e-01 2.49644980e-01 6.57364070e-01 5.19983232e-01
-1.00318444e+00 1.20748377e+00 5.75117302e+00 7.93535948e-01
-1.01965177e+00 -1.63754478e-01 4.86860067e-01 -2.27283388e-01
-6.22535944e-01 1.17820971e-01 -9.84808326e-01 4.39896107e-01
8.78573060e-01 7.90424719e-02 4.13491666e-01 6.26334906e-01
1.03574291e-01 -2.85554379e-01 -1.28309977e+00 6.55990362e-01
1.35273352e-01 -1.48932934e+00 1.69750839e-01 1.13567941e-01
5.76160014e-01 -5.37421517e-02 1.44027919e-01 1.65084094e-01
4.43161875e-01 -1.27901149e+00 1.10641134e+00 4.15506095e-01
9.91192818e-01 -8.08989346e-01 4.08907503e-01 2.29887426e-01
-1.00741696e+00 3.95918675e-02 -2.92479455e-01 1.79397210e-01
2.04850193e-02 7.05264211e-01 -1.09398794e+00 5.34099519e-01
8.59954953e-01 5.47070503e-01 -9.99950707e-01 6.73427641e-01
-5.11928260e-01 3.97305816e-01 -3.61319244e-01 -2.02205837e-01
5.02592564e-01 1.66634798e-01 2.69903362e-01 1.76742899e+00
2.97786742e-01 -3.15220088e-01 -9.90309790e-02 8.20899725e-01
-4.89943087e-01 1.32945001e-01 -6.10015512e-01 -4.71728779e-02
3.68994981e-01 1.20345819e+00 -8.22054982e-01 -3.45574081e-01
-2.98183441e-01 1.14646769e+00 6.05687976e-01 2.36051098e-01
-6.21923149e-01 -8.30291033e-01 2.77449936e-01 1.76237196e-01
7.21395791e-01 -4.70873982e-01 -4.18230146e-01 -1.00747585e+00
2.99005479e-01 -9.58424151e-01 3.83358389e-01 -9.14449811e-01
-8.23176146e-01 6.41412854e-01 1.27007946e-01 -7.30376184e-01
-2.92389840e-01 -9.96823132e-01 -2.57573485e-01 7.99989522e-01
-1.59303606e+00 -1.38568318e+00 -2.79265136e-01 2.53519982e-01
5.31929493e-01 -1.33293584e-01 7.54377902e-01 1.43955976e-01
-3.50003421e-01 7.49556243e-01 4.48010176e-01 4.18731809e-01
7.97626555e-01 -1.58922338e+00 7.90226698e-01 9.18823719e-01
4.84567106e-01 8.10698450e-01 6.17984056e-01 -4.01380420e-01
-1.36718178e+00 -9.11601245e-01 1.15017653e+00 -5.67138433e-01
6.94714725e-01 -1.08247030e+00 -8.18409920e-01 6.43473148e-01
5.29316366e-01 -3.80765557e-01 4.28072572e-01 3.35924208e-01
-8.56682301e-01 2.96041612e-02 -6.65092349e-01 7.42618382e-01
1.15525723e+00 -7.12414324e-01 -4.93005902e-01 3.46451432e-01
6.63226247e-01 -3.37101668e-01 -8.08608592e-01 2.46058945e-02
5.21948576e-01 -7.64447153e-01 9.20554876e-01 -4.80882674e-01
7.95604110e-01 -3.38969737e-01 -1.44399971e-01 -1.20643413e+00
-3.52546126e-01 -6.15075052e-01 -2.36645844e-02 1.22572672e+00
6.52127445e-01 -1.12107962e-01 6.18796349e-01 9.18344222e-03
-1.33827433e-01 -6.33199334e-01 -6.80405140e-01 -7.15117097e-01
3.11548382e-01 -4.38010663e-01 5.16301036e-01 5.18345654e-01
-1.42274529e-01 4.28353846e-01 -1.39147222e-01 -5.46324626e-03
3.14037710e-01 4.15787518e-01 8.43966663e-01 -9.78337348e-01
-3.55907947e-01 -5.63630700e-01 1.87656879e-02 -1.39005744e+00
1.77294612e-01 -1.13535571e+00 -9.68648419e-02 -1.73851538e+00
3.27897400e-01 -3.18518132e-01 -8.01900029e-02 7.35478640e-01
1.79726273e-01 1.92402348e-01 4.20740008e-01 8.17701742e-02
-6.09642029e-01 1.96989030e-01 1.16382480e+00 -4.68476653e-01
1.09661683e-01 -6.06476426e-01 -9.92735624e-01 6.80204451e-01
6.19499445e-01 -2.59707391e-01 -2.64898300e-01 -9.41371083e-01
4.95297343e-01 -5.67579210e-01 3.15722436e-01 -6.43044412e-01
1.10263005e-01 1.43414676e-01 5.91727972e-01 -8.45440090e-01
1.41756773e-01 -4.04911965e-01 -5.65491974e-01 1.67419225e-01
-6.87430859e-01 2.81777591e-01 2.37307116e-01 2.34362423e-01
-6.65124357e-02 -4.44682151e-01 4.91416276e-01 -9.52957049e-02
-8.78339291e-01 -2.37699315e-01 -2.57175595e-01 9.71835405e-02
6.45832360e-01 -2.69135647e-02 -8.29349458e-01 -3.42897266e-01
-3.49767148e-01 1.09429456e-01 6.48800433e-01 6.50157809e-01
4.26639616e-01 -1.00177193e+00 -6.78281426e-01 2.01802522e-01
4.00125355e-01 2.33549997e-01 -7.21231699e-02 3.28385532e-01
-7.79598057e-01 8.57363701e-01 -2.06600931e-02 -4.86269444e-01
-1.24657702e+00 4.21297342e-01 1.43721942e-02 -6.35493159e-01
-6.91021919e-01 8.96259904e-01 2.41806597e-01 -3.08423400e-01
2.25596681e-01 -6.36630595e-01 -2.38496233e-02 1.23681270e-01
6.32533967e-01 2.17062652e-01 4.81611073e-01 -3.97957891e-01
-3.36919636e-01 6.38016522e-01 -6.11687064e-01 -3.85114640e-01
1.42241752e+00 -8.41792598e-02 -2.34112561e-01 2.26704180e-01
1.49298000e+00 3.64222080e-01 -1.01116216e+00 -1.44320711e-01
5.23886323e-01 -3.85091931e-01 6.80566803e-02 -1.09952605e+00
-8.25518012e-01 1.02743995e+00 2.62355208e-01 3.74045526e-03
9.51783001e-01 2.06453383e-01 6.34309709e-01 9.42536712e-01
1.54266074e-01 -1.28002310e+00 2.61279076e-01 8.56932104e-01
1.06412327e+00 -9.44069147e-01 1.88225508e-01 -2.45854646e-01
-4.70843881e-01 1.35799122e+00 2.78819531e-01 -3.11653502e-02
2.71672100e-01 5.76264322e-01 -7.15392381e-02 -1.69556960e-01
-8.36012542e-01 -1.40858039e-01 5.84251285e-01 3.69181037e-01
6.55353069e-01 -1.74070708e-02 -7.37215802e-02 5.85251570e-01
-6.18363917e-01 -2.43094355e-01 5.58488548e-01 9.21719015e-01
-3.80755693e-01 -1.32275283e+00 -2.40414545e-01 3.20932597e-01
-3.90777797e-01 -3.08284640e-01 -7.04133034e-01 9.89216089e-01
1.92742236e-02 6.35817885e-01 1.56528160e-01 -1.75304204e-01
2.85240918e-01 5.62650561e-02 4.37443227e-01 -6.12443507e-01
-4.05883670e-01 1.02709174e-01 2.86625773e-01 -4.05181944e-01
8.64740014e-02 -6.02282226e-01 -1.42431557e+00 -2.87024170e-01
-1.12060569e-02 -1.91459090e-01 7.57111609e-01 7.88230836e-01
2.98046499e-01 6.85738921e-01 2.97468007e-01 -3.95816267e-01
-3.90799969e-01 -8.05333793e-01 -3.98352385e-01 3.92113179e-01
3.22969317e-01 -3.07150036e-01 1.18177794e-02 1.46803334e-01]
|
[11.623963356018066, 2.597076177597046]
|
53e8c822-764e-48cc-b88d-68bcd53bd0f4
|
learning-rich-features-for-gait-recognition
|
2110.13408
| null |
https://arxiv.org/abs/2110.13408v2
|
https://arxiv.org/pdf/2110.13408v2.pdf
|
Learning Rich Features for Gait Recognition by Integrating Skeletons and Silhouettes
|
Gait recognition captures gait patterns from the walking sequence of an individual for identification. Most existing gait recognition methods learn features from silhouettes or skeletons for the robustness to clothing, carrying, and other exterior factors. The combination of the two data modalities, however, is not fully exploited. Previous multimodal gait recognition methods mainly employ the skeleton to assist the local feature extraction where the intrinsic discrimination of the skeleton data is ignored. This paper proposes a simple yet effective Bimodal Fusion (BiFusion) network which mines discriminative gait patterns in skeletons and integrates with silhouette representations to learn rich features for identification. Particularly, the inherent hierarchical semantics of body joints in a skeleton is leveraged to design a novel Multi-Scale Gait Graph (MSGG) network for the feature extraction of skeletons. Extensive experiments on CASIA-B and OUMVLP demonstrate both the superiority of the proposed MSGG network in modeling skeletons and the effectiveness of the bimodal fusion for gait recognition. Under the most challenging condition of walking in different clothes on CASIA-B, our method achieves the rank-1 accuracy of 92.1%.
|
['Zhiqiang He', 'Yang Zhang', 'Kang Ma', 'Yunjie Peng']
|
2021-10-26
| null | null | null | null |
['gait-identification']
|
['computer-vision']
|
[ 0.01880165 -0.45592475 -0.38836527 -0.26863524 -0.39564875 -0.03645656
0.18193781 -0.2381882 -0.14876387 0.51234525 0.35918054 0.5657741
-0.01443955 -0.7811022 -0.19182831 -0.9448516 -0.21427526 0.19661665
0.27175543 -0.3000722 -0.12223334 0.3175098 -1.5962294 -0.05692229
0.6216201 1.0823066 -0.150865 0.55473703 0.14473405 0.36787474
-0.3113487 -0.4998693 0.12004369 -0.2327404 -0.22360305 0.33802634
0.40940702 -0.32724246 -0.67008716 0.78945106 0.73245335 -0.11927848
0.7357017 -1.3066223 -0.17067651 0.20652652 -1.1925898 0.00694574
0.55322856 0.29359785 1.1195259 -0.74631506 0.28837094 1.3735402
0.8393112 0.33539587 -0.9694106 -0.58788383 -0.14234911 0.37066922
-1.4907039 -0.38792986 1.128468 -0.19482878 0.49520138 0.02606843
0.83917886 1.2666537 0.18880549 1.0255015 0.80336654 -0.2167063
-0.13948743 -0.65871763 0.33746386 1.2996916 0.6411646 -0.18606746
-0.9728411 -0.17324294 0.8534976 0.1532965 -0.02164435 -0.45327434
-1.1377044 0.36532304 0.20633404 0.1400711 -0.34976593 0.21216187
0.6450253 0.02448319 -0.04851498 -0.4531996 0.15720914 -0.2281022
-1.0429754 0.11083039 0.54658824 0.67913705 0.649448 0.41213983
-0.09441626 1.0425799 0.6344584 0.8915244 0.66200286 -0.5918595
0.6973921 0.92002875 -0.16241212 -1.5862837 -0.47290245 -0.22677906
-1.0437466 -0.35998932 0.42198947 0.10055406 -0.8236628 1.6290977
0.23202643 0.20875718 -0.17567639 1.0236696 0.8384523 0.02270339
0.10921789 0.23760216 1.5478597 -0.67720205 -0.44131416 -0.37346515
0.33094174 -0.388769 0.80193704 0.10789566 -0.7822298 -0.7347392
-1.4185721 0.1770508 -0.13671826 0.7300329 0.43982723 0.9555301
-0.6371541 0.45952985 -1.1236379 -0.6009467 0.2739985 0.5815776
-0.6562604 -0.25166574 -1.1612787 0.42682764 0.273587 0.32925564
-0.5213424 0.13219097 -1.1424122 -0.16578911 0.22834663 -0.7452117
0.3052774 -0.27469477 -1.1812358 0.7232182 -0.18741466 -0.26441464
0.5238842 -0.18835215 -0.5362913 0.53048366 0.3716646 0.35756037
1.2946459 -0.8968045 -0.46376956 -0.85201067 -0.50122 0.16842692
-0.605544 -0.4947749 -0.7574893 -0.7832824 0.35067523 -0.8425464
0.16654712 -0.11831994 -0.46337917 -0.17325456 1.0160888 -0.9666918
1.3643169 -2.178747 0.20332246 0.51942027 0.15902321 0.01311094
-0.12708859 0.40028423 0.148818 -0.3316741 -0.22487843 -0.5041292
0.13977726 0.573379 0.224829 0.71148926 0.20765817 0.83044285
-0.74806523 -1.1091785 0.307711 0.37136185 -0.26757097 0.09063585
0.59348285 0.22567338 -0.6807239 1.3065785 0.5920377 -0.09194131
0.46970725 -0.6967406 0.43147677 -0.45429704 -1.3750039 1.7099539
0.12054137 -0.13943358 0.17771487 -1.1525544 1.1882753 0.17537431
0.97365916 -0.48163268 0.02618737 0.26746565 -0.24267614 -0.55679107
0.2820377 0.09305758 -0.4248398 0.19390853 0.2680575 0.49599537
0.5174104 0.01360028 1.0540702 0.4184375 0.4241447 -0.0899688
0.8281434 -0.45349976 0.8724585 0.5372187 -0.6284674 0.5686208
0.02909754 -0.42423066 -0.7147025 -1.391557 0.27423012 0.9345326
0.31931347 -0.71523803 -0.4655287 -0.6701392 0.23523974 -0.37124896
-0.47755566 -0.45667294 -0.77946746 -0.9344161 1.1193769 0.92926705
1.2010098 -0.4218505 -0.64622927 0.1370883 -0.6298408 -1.1829475
-0.64176214 -0.3919633 -0.89166623 -1.1870909 -0.9273437 -0.67157984
0.40367863 0.2145637 0.5979382 0.17859954 -0.5559483 0.7000651
-0.13250467 0.4382884 0.18407755 0.04479954 0.41372734 0.50154305
0.47913036 -0.947754 -0.6726012 0.5178542 -0.5031823 -0.11485436
0.7380683 1.1263903 0.2882195 0.11347952 0.45347613 -0.14385979
0.41709593 -0.4003512 0.18767756 0.3519547 -0.22496912 0.01959045
0.301154 -0.23864911 -0.7941294 0.26075816 0.04075679 -0.4479766
-0.17903452 0.48790514 -0.4116351 -0.17785327 0.12945044 0.6284631
0.23435867 -0.5866553 0.07880314 0.43489602 0.9154901 -1.119875
0.83758515 0.5722967 0.58590746 -1.1722461 -0.12892689 -0.7092388
-0.9188176 -0.6112062 0.8043706 -0.9702479 -0.91901994 0.7102313
-0.56648916 0.543502 0.09320044 0.3631475 -0.6529018 1.248103
-0.8258301 -0.8840775 -0.38398233 -0.9002752 1.3297749 0.22116558
-0.3273919 -0.83913887 -0.09945557 0.7978151 -0.17697288 0.5756394
0.7512984 -0.25388125 -0.10089685 -0.44244745 -0.26981387 0.12171719
0.50393534 0.0691671 -0.70326495 -0.40965313 -0.44577575 -0.3825841
1.0513598 0.16426444 0.49027133 -0.07709847 -0.31694996 0.36858475
1.2520603 -0.12166115 0.45121914 0.12413769 1.0716544 0.566673
0.561267 0.7455419 0.6297921 0.6549372 0.28520933 0.1196989
-0.17811905 -0.36266178 0.8146523 1.228286 -0.69486475 0.21629816
-0.93200576 0.6868636 -2.08204 -0.79658186 0.02576219 2.046335
0.32739514 0.06666134 0.6440213 0.4447085 0.95664346 0.3072624
-0.4451892 0.29485822 -0.32061368 0.11802676 0.48616937 -0.14120702
-1.3772129 0.5883127 5.715454 0.8795729 -0.7655494 -0.22482818
0.00812172 0.31029305 0.44214037 -0.38438347 -0.7210468 0.59474915
0.57750654 0.16101508 -0.04909715 0.7088733 0.13476114 0.04887043
-0.6654841 1.2323097 0.10218882 -0.57788813 0.2907935 0.14692283
0.19197477 -0.41794237 0.08470789 0.03159718 0.08230992 -0.65263605
0.3974381 0.579296 0.66109 -0.79388034 0.81628335 0.05677367
-2.087472 -0.20948352 -0.14650212 0.12273883 0.1632826 0.30234602
-0.5592634 1.1701493 0.6644756 1.4594411 -0.81559294 0.88265616
0.01820876 0.74921924 -0.5828902 0.36101937 0.1748153 -0.28669748
0.621637 1.3766938 0.40640238 -0.30804282 0.5885781 0.43469143
0.22617966 0.21961096 -0.5666967 -0.3789938 0.07593443 1.1067787
-0.685678 -0.10592849 -0.5521386 0.93126184 -0.02003165 0.22827426
-0.6201129 -0.18989347 0.76494306 0.01384948 0.3728354 -0.61821306
-0.25198373 -1.5481813 0.16584729 -0.7067715 0.8285294 -0.26120505
-1.6271379 -0.08701515 -0.00948569 -1.3680564 -0.20721461 -0.65135646
-0.63600326 0.54302776 -0.97009265 -1.8277069 -0.44554758 0.9650471
0.36552206 -0.32817617 0.551205 0.6183527 -0.8842606 0.6859406
-0.29684922 0.70203406 0.6944678 -1.014331 0.21598585 0.9289259
-0.20197019 0.7521203 0.41503987 -1.0761868 -1.742177 -0.77084655
0.5320758 0.07414413 0.6004171 -0.0430651 -0.6861369 0.3132429
-0.34516245 -0.12279648 0.79326075 -0.07607205 -0.44400048 -0.38037625
-0.96992105 0.57551455 1.165546 -0.44338524 -0.61406606 -0.32430503
-0.12947868 0.08856533 -1.2417046 0.6447811 1.1665478 -0.7053773
1.435278 -0.5140895 0.1938396 -0.48039 -0.5621796 -0.73971057
-0.21450183 -0.07601224 -0.36454463 1.2244579 -0.47717962 -0.47369
1.0971397 0.25237536 0.30527708 -0.39927492 -1.0585558 -0.8242245
-0.54768187 -0.23170282 0.46745646 0.66332537 0.02732917 0.25728166
-0.9657874 0.01892724 1.3062673 0.05916327 0.85906637 -1.2772686
-0.17719595 -0.23721083 -1.2888117 -0.9160258 0.07903817 -0.7352449
-0.24732792 -1.3312942 0.18916176 0.09215576 -0.43197346 0.5787806
-0.07678629 0.45901304 0.22566898 0.3715671 -0.57081765 0.7954269
1.1846251 -0.51662356 0.20484471 0.01611581 -0.22292905 0.8371581
0.56593233 0.05650461 -0.34361374 0.00895659 -0.3285544 0.23200604
0.5408229 -1.3318347 0.2705053 0.02618687 0.724236 -0.77577156
0.68542635 -0.68834007 0.04329021 0.6273457 0.35205057 0.1486937
-0.1363351 0.92645705 -0.36879694 0.22464485 0.5644733 -0.24620782
-1.0661544 0.28362823 -0.47436884 -0.19642156 0.63912845 -0.81127304
-0.10452068 -0.2995344 -1.0992653 0.2737587 0.3954178 0.37544337
0.93914604 -1.6557919 -0.6050304 0.38804382 0.2645755 -0.603918
0.3400391 1.0200448 -0.31955367 0.18718317 -0.73679876 -0.80212
-1.5002565 -0.04153939 0.19476575 -0.43962225 -0.54583037 0.27571362
-0.07500415 -0.16129662 -0.06363105 0.03020512 -0.4144146 0.24463017
0.19044496 0.8591747 -0.20616356 -1.1934469 -0.66569656 1.1256409
0.11829229 0.05095429 1.0574507 -0.33883333 -0.0400764 0.37329823
0.9598987 -0.3015868 -1.0374198 -0.2600579 0.09468479 -0.4586828
-0.5161787 -0.2985842 -1.2008648 0.9755535 0.6393578 -0.36169994
1.1164787 -0.4931729 1.1232895 0.34135774 0.7332244 -1.1908497
0.3929954 0.16397332 0.499396 -1.0838381 0.0581957 -0.45157945
-0.61018926 1.3100381 0.59040433 -0.18068719 0.54288757 0.02929635
-0.25946918 -0.06325258 -0.23685254 -0.30521792 0.42849043 0.7821748
0.17927237 0.24923892 -0.3702928 0.7342371 0.127599 -0.02228763
0.03213182 1.2452538 -0.4633723 -1.08592 -0.52004415 0.41613108
-0.27826852 0.45362684 -0.41511902 0.7109514 0.1814025 0.9140136
-0.45225728 -0.77424216 0.22886594 0.28375947 0.51236403 -0.16053374
-0.23285766 0.19924097 0.15974213 -0.7081493 -0.68980515 -0.8539321
-1.0367693 -0.33887398 0.20523933 -0.03371152 0.20693542 0.8928685
0.15824807 0.21252544 0.35999164 -0.8128647 -0.54830235 -0.76584554
-1.1081494 0.62055635 0.18941706 -1.3018007 -0.01448339 0.18175137]
|
[14.291923522949219, 1.4104026556015015]
|
2e4fd196-8b8f-4b54-a227-4ac743e8ebc7
|
functional-intrusive-load-monitor-film-a
|
1809.08910
| null |
http://arxiv.org/abs/1809.08910v1
|
http://arxiv.org/pdf/1809.08910v1.pdf
|
Functional Intrusive Load Monitor (FILM): A Model-based Platform for Non-Intrusive Load Monitoring System Development
|
Non-Intrusive Load Monitoring (NILM) is an important application to monitor
household appliance activities and provide related information to house owner
or/and utility company via a single sensor installed at the electrical entry of
the house. It can be used for different purposes in residential and industrial
sectors. Thus, an increasing number of new algorithms have been developed in
recent years. In these algorithms, researchers either use existing public
datasets or collect their own data which causes such problems as insufficiency
of electrical parameters, missing of ground-truth data, absence of many
appliances, and lack of appliance information. To solve these problems, this
paper presents a model-based platform for NILM system development, namely
Functional Intrusive Load Monitor (FILM). By using this platform, the state
transitions and activities of all the involved appliances can be preset by
researchers, and multiple electrical parameters such as harmonics and power
factor can be monitored or calculated. This platform will help researchers save
the time of collecting experimental data, utilize precise control of individual
appliance activities, and develop load signatures of devices. This paper
describes the steps, structure, and requirements of building this platform.
Case study is presented to help understand this platform.
|
[]
|
2018-09-19
| null | null | null | null |
['non-intrusive-load-monitoring', 'non-intrusive-load-monitoring', 'non-intrusive-load-monitoring']
|
['knowledge-base', 'miscellaneous', 'time-series']
|
[ 1.70421064e-01 -2.56543458e-01 -2.74102122e-01 -3.42042893e-01
-1.96678832e-01 -5.26811242e-01 2.94741690e-01 -7.19037503e-02
4.67138171e-01 7.34130621e-01 -1.82472333e-01 4.39105034e-02
-2.53866225e-01 -9.99711692e-01 1.33774251e-01 -9.50620592e-01
1.61066815e-01 4.80056435e-01 1.13790870e-01 -5.58518060e-02
-6.57283589e-02 4.83048081e-01 -1.63028169e+00 -4.86597382e-02
7.28050113e-01 9.22566473e-01 4.15459007e-01 6.36170879e-02
2.62137979e-01 6.76740110e-01 -1.01747477e+00 5.27011096e-01
-1.51309624e-01 -5.31359792e-01 -5.22201359e-01 3.57789427e-01
-7.17108071e-01 -6.15447402e-01 -7.09819188e-03 1.14003563e+00
7.07912564e-01 -1.00876704e-01 4.37642992e-01 -1.88702369e+00
-2.03969702e-01 7.59548545e-01 -6.62998915e-01 1.50628984e-01
7.95498013e-01 2.81598717e-01 4.10115749e-01 8.76353458e-02
-2.36539040e-02 8.79968703e-01 4.17582363e-01 4.74549793e-02
-1.29200923e+00 -9.52215731e-01 -6.97198361e-02 8.28478396e-01
-1.54526341e+00 -2.77886122e-01 1.36340916e+00 -2.86479056e-01
1.04980171e+00 8.85527968e-01 6.62125766e-01 1.08716631e+00
6.21995423e-03 6.90967023e-01 1.04365385e+00 -4.56190497e-01
3.75994563e-01 4.25329298e-01 3.03559572e-01 8.79120007e-02
4.90865856e-01 -1.83157697e-01 2.39513163e-02 -3.78533095e-01
2.36091390e-01 2.62004912e-01 -4.81404305e-01 -3.19592748e-03
-6.98002756e-01 3.61223727e-01 -2.41668954e-01 9.00811017e-01
-3.49406809e-01 -4.97626126e-01 4.07894075e-01 4.49835397e-02
1.34500086e-01 3.28854918e-02 -5.12983799e-01 -2.07869962e-01
-7.43880332e-01 -1.46039769e-01 8.20827127e-01 1.03139508e+00
6.98748529e-01 5.45236915e-02 5.17955422e-02 5.38388014e-01
4.92589414e-01 5.77700794e-01 8.09419513e-01 -7.68814981e-01
2.62314469e-01 1.05295122e+00 1.43669546e-01 -9.01206434e-01
-8.17766011e-01 2.45057523e-01 -1.11860168e+00 -5.13708219e-02
-2.84755975e-01 -9.00983512e-02 -3.23320180e-01 1.39979804e+00
4.62541133e-01 2.41533428e-01 -3.20847392e-01 5.56470692e-01
5.46591759e-01 8.93673062e-01 1.09963948e-02 -1.00736427e+00
1.36731064e+00 -2.58610338e-01 -1.48793852e+00 2.17587531e-01
4.56095010e-01 -7.14560091e-01 9.69440162e-01 4.30933058e-01
-7.13808894e-01 -4.71531421e-01 -1.16660190e+00 3.55885297e-01
-5.73014319e-01 1.30157724e-01 3.49558234e-01 7.59349048e-01
-5.18912435e-01 6.04555547e-01 -9.32971656e-01 -5.57957828e-01
2.24194571e-01 4.10725743e-01 8.09336975e-02 1.76766202e-01
-1.06431997e+00 8.87028873e-01 5.06594718e-01 2.47534081e-01
-4.66310054e-01 -5.62098682e-01 -7.04898119e-01 -4.73189130e-02
8.92745927e-02 -1.69916555e-01 1.11812699e+00 -2.67772287e-01
-1.46611226e+00 3.59093994e-01 7.23608956e-02 -3.37781720e-02
3.00317824e-01 1.14244133e-01 -1.10567498e+00 -1.00630626e-01
2.50562966e-01 -4.37250465e-01 4.26322609e-01 -9.82904077e-01
-7.41521955e-01 -5.77077746e-01 -4.14204091e-01 -3.05191100e-01
-2.45425746e-01 -8.61777365e-03 1.47823125e-01 -1.30239069e-01
1.42265990e-01 -6.63887799e-01 2.48729154e-01 -7.42680728e-01
-7.63829052e-01 -5.01131594e-01 1.77473187e+00 -9.49128211e-01
1.47073257e+00 -2.07527947e+00 -4.99712437e-01 5.23082733e-01
-2.87212253e-01 2.73536056e-01 4.03424412e-01 5.18637598e-01
-2.65332490e-01 -8.73384327e-02 -2.56553233e-01 -5.44041321e-02
2.31740087e-01 4.64963853e-01 6.64245188e-02 6.85906649e-01
-3.19823861e-01 7.16754496e-01 -6.49130106e-01 -5.01106858e-01
1.18636215e+00 4.16374296e-01 2.27501407e-01 2.95354366e-01
-4.15617712e-02 5.83763242e-01 -5.10738134e-01 8.94375622e-01
7.34151602e-01 3.02297585e-02 5.04300296e-01 -6.07024252e-01
-4.35909368e-02 3.01297516e-01 -1.65468383e+00 1.04013479e+00
-6.02125525e-01 3.88422489e-01 3.35739926e-02 -1.15194285e+00
6.68259203e-01 9.35142517e-01 1.03016019e+00 -6.69193506e-01
4.75124031e-01 -1.41866431e-01 -3.32677662e-01 -1.01965249e+00
-2.10472956e-01 4.46695209e-01 9.92328767e-03 7.51048028e-01
-2.38959566e-01 -4.80117947e-02 4.29583549e-01 -5.10257661e-01
1.07221401e+00 -9.08991769e-02 6.77288711e-01 -3.32737535e-01
8.37958753e-01 -3.85097116e-01 6.84866071e-01 -2.63910219e-02
-1.36321232e-01 -1.57320097e-01 -6.19670637e-02 -2.43133381e-01
-7.81137586e-01 -7.48856127e-01 -2.17128724e-01 5.67234814e-01
2.32529938e-01 -2.75358409e-01 -6.46483600e-01 -1.13378875e-01
-1.19592182e-01 1.06884122e+00 -1.40981138e-01 -2.69731641e-01
-6.81593001e-01 -1.01914644e+00 -4.09303904e-02 4.17716950e-01
8.00692379e-01 -1.27157342e+00 -6.87967837e-01 5.83312392e-01
-4.56599802e-01 -1.03831983e+00 -4.07869816e-02 1.52181849e-01
-6.35537148e-01 -1.50010109e+00 1.75926477e-01 -6.89111948e-01
7.49386966e-01 1.99092597e-01 9.67107117e-01 1.18953481e-01
-5.75358570e-01 7.29326382e-02 -1.13587618e-01 -5.33862412e-01
-4.19563979e-01 1.45463407e-01 1.99546069e-01 -2.24086437e-02
8.51155996e-01 -1.13062966e+00 -3.53834808e-01 6.90534055e-01
-6.65275693e-01 -3.32777947e-01 1.41852377e-02 1.17393374e-01
3.76854897e-01 1.10647464e+00 6.97553515e-01 -5.85592926e-01
6.86900258e-01 -5.81204355e-01 -1.14741766e+00 2.20629498e-01
-1.01971078e+00 -4.83997136e-01 6.59069300e-01 -4.87157285e-01
-1.01116705e+00 9.10435989e-02 8.66406504e-03 7.54106417e-02
-6.49483502e-01 -1.99932486e-01 -1.14891791e+00 4.22639728e-01
2.22966485e-02 9.85722095e-02 -2.67434657e-01 -9.11099792e-01
-1.60686716e-01 1.10525358e+00 5.88238537e-01 -2.91936487e-01
8.29440832e-01 2.93130815e-01 3.06400657e-03 -6.98059857e-01
-3.88294697e-01 -5.13027787e-01 -5.53244591e-01 -1.34134248e-01
6.88626111e-01 -6.66037858e-01 -1.34367764e+00 6.62721395e-01
-1.08946586e+00 1.91897259e-03 -5.13539195e-01 3.00091624e-01
-3.18594664e-01 3.23824316e-01 -5.87448359e-01 -1.02172852e+00
-5.40714324e-01 -1.14417696e+00 9.46499109e-01 3.34170759e-01
-7.62811124e-01 -8.50135386e-01 -1.62171662e-01 3.29707593e-01
2.72967786e-01 4.51464474e-01 9.74290550e-01 -2.66746163e-01
-5.04466891e-01 -4.90795135e-01 2.86248982e-01 4.31903809e-01
9.80640292e-01 -5.85903705e-04 -1.17907858e+00 -3.90563846e-01
6.06529295e-01 4.33505684e-01 -2.37942591e-01 2.71402538e-01
1.27493858e+00 -6.20209098e-01 -7.79465735e-01 3.32115233e-01
1.53315854e+00 8.10930014e-01 7.22353339e-01 1.60975322e-01
3.84102434e-01 3.37323874e-01 3.76104295e-01 6.13196194e-01
3.39807957e-01 5.56216419e-01 4.37667847e-01 -2.12373096e-03
3.15495938e-01 -1.60774678e-01 2.43592054e-01 9.49731410e-01
-9.10901576e-02 -2.76618332e-01 -1.94136754e-01 3.95209372e-01
-1.82882738e+00 -1.13187659e+00 -2.94900209e-01 1.77320731e+00
6.98594689e-01 -5.44855744e-02 1.88888133e-01 1.15521407e+00
9.17265832e-01 -9.27586108e-02 -7.49531806e-01 2.38488801e-02
1.65962785e-01 -4.21874300e-02 3.35120499e-01 2.87088275e-01
-7.37437546e-01 -6.30929247e-02 6.48093653e+00 5.17105281e-01
-6.62031472e-01 3.06933284e-01 2.19500571e-01 1.03104733e-01
1.34147942e-01 -2.87788123e-01 -7.45444417e-01 1.11161494e+00
8.46034050e-01 -3.47045124e-01 7.05421746e-01 1.01470411e+00
8.12936902e-01 -5.07354438e-01 -1.32780612e+00 1.26873946e+00
-1.09254666e-01 -5.62685072e-01 -4.54018176e-01 1.51575789e-01
6.31220877e-01 -3.27491701e-01 -6.34863675e-01 -1.09952256e-01
1.56865627e-01 -3.89461875e-01 1.35176554e-01 4.11431789e-01
2.62270927e-01 -8.00091028e-01 1.02012801e+00 6.42662406e-01
-1.47272110e+00 -3.19544107e-01 1.69757642e-02 -1.63925439e-01
5.04799604e-01 9.08016562e-01 -5.18795431e-01 6.86203241e-01
9.02732730e-01 1.91402391e-01 -3.46104681e-01 7.36207306e-01
-3.58848959e-01 6.60432577e-01 -8.69703472e-01 1.33855030e-01
-7.20414937e-01 -5.61125696e-01 2.53784031e-01 5.60702026e-01
4.25189853e-01 1.47088036e-01 2.40469724e-01 9.57355261e-01
2.74344295e-01 -9.00737271e-02 -5.41163266e-01 4.42311466e-01
7.03454673e-01 1.44558585e+00 -4.78699833e-01 -4.05519605e-01
-2.65184462e-01 8.28005493e-01 -5.98584592e-01 1.47842318e-01
-8.30633342e-01 -3.71109486e-01 6.72917068e-01 3.42467934e-01
-2.30021209e-01 5.26015870e-02 -1.54939979e-01 -6.74294651e-01
2.52204597e-01 -7.37338364e-01 3.34562808e-01 -8.50206375e-01
-1.21224952e+00 -4.50899117e-02 2.92803138e-01 -9.77653563e-01
-7.22773850e-01 -4.80744392e-02 -1.06924927e+00 7.07091391e-01
-1.20762074e+00 -9.70432460e-01 -6.38391495e-01 7.80446351e-01
4.86620039e-01 -1.31836995e-01 9.84776556e-01 7.41250992e-01
-9.86193776e-01 8.94583538e-02 2.51227111e-01 -9.02086347e-02
8.27746019e-02 -8.95479739e-01 -1.36866728e-02 6.88956678e-01
-2.11439639e-01 -1.08550023e-02 8.32924485e-01 -5.99075556e-01
-1.26469505e+00 -9.89118934e-01 7.45262682e-01 -4.66467083e-01
3.31222355e-01 -2.88178891e-01 -6.73111975e-01 8.61000776e-01
3.38835120e-01 -2.49688789e-01 7.19987333e-01 -2.44527087e-01
5.73509455e-01 -6.45646334e-01 -1.76803064e+00 1.10811509e-01
6.65778458e-01 -4.04010355e-01 -4.98063475e-01 6.45641327e-01
3.21421951e-01 1.07410178e-01 -1.06366754e+00 3.23913574e-01
6.71554655e-02 -9.64029312e-01 6.20693386e-01 1.98373333e-01
-6.85068488e-01 -7.46915638e-01 -1.64752290e-01 -1.06330538e+00
-3.82804334e-01 -8.41232896e-01 -5.93718350e-01 2.09314966e+00
-2.40948066e-01 -9.96759176e-01 3.74640137e-01 6.73041344e-01
2.10931167e-01 -1.68230772e-01 -7.84186244e-01 -7.02872336e-01
-9.52092528e-01 -5.34200549e-01 1.67190194e+00 9.68930066e-01
3.47093493e-01 4.21569735e-01 -1.29734606e-01 6.02575421e-01
6.85846090e-01 1.08104609e-01 5.06255388e-01 -1.37820625e+00
1.28847554e-01 -8.83017629e-02 -5.08054316e-01 -3.10529888e-01
-2.79466435e-02 -3.88347715e-01 -2.59343773e-01 -1.77669847e+00
2.71437271e-03 -1.72753707e-01 -2.94528455e-01 7.41419137e-01
3.37025255e-01 1.88913152e-01 -1.53416723e-01 1.86005175e-01
-1.29601255e-01 2.01366007e-01 5.92880905e-01 -3.30593526e-01
-2.79982507e-01 3.18151087e-01 -2.59100497e-01 9.83042300e-01
1.26793432e+00 -1.57923535e-01 -6.31632686e-01 8.19139555e-03
-2.46457413e-01 -1.95998356e-01 3.70344311e-01 -1.28372145e+00
4.76273634e-02 -1.42567381e-01 6.23987317e-01 -1.18541300e+00
1.65891722e-01 -1.79906237e+00 9.64406729e-01 4.98750687e-01
4.86990422e-01 1.94076881e-01 -9.19044837e-02 -7.31528774e-02
1.16555579e-01 -2.82558501e-01 5.99272013e-01 -7.90486559e-02
-3.61091673e-01 -3.00979968e-02 -4.14250553e-01 -7.32724309e-01
1.52612805e+00 -7.47451335e-02 -2.61565953e-01 -1.38209537e-01
-4.79731113e-01 4.12603319e-01 2.43387401e-01 4.11868215e-01
-6.83032200e-02 -1.60108054e+00 -7.50209093e-02 6.49267495e-01
-1.87835798e-01 1.07227318e-01 1.05084971e-01 5.32064319e-01
-1.57483965e-01 4.80243713e-01 1.84187181e-02 -6.65090322e-01
-1.41611755e+00 8.49793017e-01 2.53503114e-01 -1.46091744e-01
-5.54920614e-01 -3.63241881e-01 -3.12351108e-01 -8.89610797e-02
5.61228655e-02 -5.71637213e-01 -3.09100270e-01 2.86038548e-01
7.48468161e-01 1.04680800e+00 3.14896554e-01 -6.59791708e-01
-3.96548808e-01 6.79565430e-01 5.20168602e-01 5.30487895e-01
1.34386146e+00 -6.91612363e-01 -4.10826772e-01 6.47922277e-01
1.18177104e+00 -3.03693116e-01 -5.89861095e-01 2.51534551e-01
-6.80158958e-02 -1.92939922e-01 -8.15408081e-02 -7.28112936e-01
-1.20494485e+00 3.25432092e-01 1.15633237e+00 9.86218870e-01
1.55148089e+00 1.29677737e-02 9.89545286e-01 7.84756616e-02
7.34183609e-01 -1.54081905e+00 -7.28549719e-01 -4.97471422e-01
5.19101024e-01 -1.05669391e+00 1.91106990e-01 -5.00791430e-01
2.01679200e-01 7.87895858e-01 4.64920819e-01 4.62032139e-01
9.85881567e-01 7.18763530e-01 -1.76932558e-01 -4.44688387e-02
-3.75413299e-01 6.81977794e-02 -3.54000986e-01 1.08191335e+00
4.06285375e-02 3.06151420e-01 -2.89221853e-01 6.84033573e-01
-2.61019021e-01 3.06106985e-01 2.25736037e-01 1.25838971e+00
-3.66005212e-01 -1.25505090e+00 -8.00901771e-01 6.03224397e-01
-4.00877923e-01 6.42170608e-01 3.15430552e-01 7.08502769e-01
5.84557891e-01 1.47076857e+00 2.70884857e-02 -1.62168875e-01
7.51813114e-01 1.08739875e-01 3.21642846e-01 -1.65553227e-01
-7.75377750e-02 -3.87058482e-02 -2.92447247e-02 -5.26126206e-01
-5.21107435e-01 -6.49990618e-01 -1.18356240e+00 -5.87099612e-01
-6.18955374e-01 1.81726426e-01 7.56603241e-01 9.81034994e-01
1.00168914e-01 6.83173776e-01 1.16953325e+00 -8.01653147e-01
-2.79693425e-01 -1.20202851e+00 -1.08402276e+00 5.75395465e-01
1.32689655e-01 -6.07673109e-01 -5.39926827e-01 2.23448560e-01]
|
[5.984971523284912, 2.565417766571045]
|
10f0a349-9087-4718-82fa-42be19915b3b
|
non-autoregressive-conditional-diffusion
|
2306.05043
| null |
https://arxiv.org/abs/2306.05043v1
|
https://arxiv.org/pdf/2306.05043v1.pdf
|
Non-autoregressive Conditional Diffusion Models for Time Series Prediction
|
Recently, denoising diffusion models have led to significant breakthroughs in the generation of images, audio and text. However, it is still an open question on how to adapt their strong modeling ability to model time series. In this paper, we propose TimeDiff, a non-autoregressive diffusion model that achieves high-quality time series prediction with the introduction of two novel conditioning mechanisms: future mixup and autoregressive initialization. Similar to teacher forcing, future mixup allows parts of the ground-truth future predictions for conditioning, while autoregressive initialization helps better initialize the model with basic time series patterns such as short-term trends. Extensive experiments are performed on nine real-world datasets. Results show that TimeDiff consistently outperforms existing time series diffusion models, and also achieves the best overall performance across a variety of the existing strong baselines (including transformers and FiLM).
|
['James Kwok', 'Lifeng Shen']
|
2023-06-08
| null | null | null | null |
['open-question', 'time-series-prediction']
|
['natural-language-processing', 'time-series']
|
[-1.11888401e-01 -3.28700870e-01 -1.48398429e-01 -5.02790749e-01
-9.50729132e-01 -6.92590952e-01 1.13491261e+00 -1.02731757e-01
-2.17485696e-01 2.46360555e-01 7.70104706e-01 -3.92184317e-01
-1.09692395e-01 -7.34396398e-01 -6.49293303e-01 -7.07925797e-01
-3.40103269e-01 3.82611573e-01 3.06635410e-01 -3.50658894e-01
-5.39943054e-02 9.47551653e-02 -1.17634380e+00 3.90786350e-01
7.99082160e-01 1.05555737e+00 -2.11057737e-02 7.78592944e-01
4.92842216e-03 1.14792907e+00 -7.24294066e-01 -5.05732000e-01
9.92147177e-02 -2.84469217e-01 -3.93223614e-01 -4.88571934e-02
1.54263884e-01 -4.91549402e-01 -7.22270131e-01 6.82787299e-01
4.06900406e-01 2.64986366e-01 6.82012677e-01 -1.00953770e+00
-1.02162516e+00 1.11082304e+00 -7.89230108e-01 6.24266148e-01
2.81990059e-02 4.22473967e-01 9.50032294e-01 -5.85956693e-01
5.45106053e-01 1.38634729e+00 1.02846730e+00 3.52505207e-01
-1.42108119e+00 -8.41336668e-01 5.97225308e-01 2.97580093e-01
-1.04534507e+00 -2.45936304e-01 1.01253128e+00 -4.48069930e-01
1.09451652e+00 2.22857874e-02 6.15517676e-01 1.65686393e+00
3.74798983e-01 8.82526755e-01 1.27508557e+00 -1.58763871e-01
2.30563030e-01 -3.74945879e-01 3.47788423e-01 6.01256043e-02
-6.74506009e-01 1.70096204e-01 -6.17237151e-01 -2.00323075e-01
6.91676855e-01 -6.31898642e-02 -2.33866557e-01 4.08978850e-01
-1.13501775e+00 6.84337199e-01 1.94169641e-01 3.99140805e-01
-7.63090432e-01 4.41431463e-01 2.56224215e-01 5.31493187e-01
1.19066381e+00 1.84406251e-01 -5.66388249e-01 -5.74394822e-01
-1.26911163e+00 3.40333045e-01 5.40661514e-01 7.10023224e-01
2.78408468e-01 5.69384456e-01 -3.84929270e-01 7.86863804e-01
1.12492405e-01 6.55649900e-01 6.01064861e-01 -1.20706689e+00
3.93522143e-01 -6.57131299e-02 1.69984773e-02 -8.16739500e-01
-1.93711117e-01 -5.79787016e-01 -8.62890899e-01 -2.96966434e-01
3.57360572e-01 -5.31721175e-01 -1.33521342e+00 1.86926115e+00
1.39729664e-01 1.03858519e+00 -1.97015867e-01 5.03838956e-01
4.39699590e-01 1.37713802e+00 1.74368128e-01 -3.06422353e-01
7.81146705e-01 -8.28704417e-01 -8.68945062e-01 -1.09064125e-01
1.40726909e-01 -9.63450611e-01 9.86944497e-01 7.39926755e-01
-1.13607192e+00 -5.20063937e-01 -6.40872061e-01 5.17073274e-02
-1.35354519e-01 -2.66223401e-01 8.77995491e-01 3.68379623e-01
-1.18534207e+00 9.16592300e-01 -1.34988594e+00 3.20247225e-02
1.61300436e-01 4.41020690e-02 8.00323784e-02 -3.44492532e-02
-1.29158151e+00 4.79025066e-01 -1.11929923e-01 -8.32974762e-02
-1.25858986e+00 -1.26554680e+00 -5.07584810e-01 -9.63249616e-03
1.06494859e-01 -5.65140188e-01 1.70769501e+00 -8.88768435e-01
-1.65397084e+00 3.33827555e-01 -2.79742897e-01 -9.77828443e-01
5.57986856e-01 -3.70271236e-01 -7.44816780e-01 3.54864188e-02
-1.29897654e-01 6.27235472e-01 1.20176232e+00 -8.73660743e-01
-4.48387712e-01 -1.31976902e-01 -1.98322445e-01 -3.11333477e-01
-7.32945442e-01 1.94190577e-01 -6.60653949e-01 -1.27582490e+00
-1.40002057e-01 -9.71747756e-01 -4.92464900e-01 -2.53635794e-01
-1.30227119e-01 -4.29703385e-01 8.18973780e-01 -7.72689223e-01
1.54987788e+00 -2.13727188e+00 -2.44616196e-02 1.78876773e-01
2.76216539e-03 -3.18126589e-01 -3.98555547e-01 6.35420144e-01
-3.27607304e-01 1.41488791e-01 8.04950763e-03 -7.45767236e-01
6.35055313e-03 4.80615944e-01 -1.25675905e+00 2.79855430e-01
1.21125355e-01 7.55305409e-01 -9.51621890e-01 -1.20720483e-01
6.09849095e-02 7.54513681e-01 -5.92989504e-01 4.77696061e-02
-6.11243665e-01 5.91746628e-01 -2.41494194e-01 2.61557847e-01
4.29261327e-01 -3.28036964e-01 -8.71269330e-02 -1.37446644e-02
-1.63383499e-01 4.81566936e-01 -9.82369423e-01 1.58378184e+00
-2.93559134e-01 7.41006672e-01 -2.94424206e-01 -6.01866364e-01
6.81858420e-01 4.72391993e-01 8.01820517e-01 -6.43823266e-01
2.60365494e-02 -4.38893884e-02 -1.76269814e-01 -1.74756601e-01
6.22811317e-01 -2.10406005e-01 1.06540844e-01 5.91889739e-01
2.51361787e-01 -4.93595630e-01 3.63443941e-01 3.16988111e-01
1.11547196e+00 1.37801141e-01 -6.21546745e-01 -2.34246552e-02
-4.93643284e-02 3.22589977e-03 6.81008339e-01 7.10166693e-01
1.56428561e-01 9.30717409e-01 3.75627995e-01 -2.76635915e-01
-1.02678883e+00 -1.10002899e+00 1.25053972e-01 1.10570037e+00
-4.79744434e-01 -8.31410110e-01 -5.29976189e-01 -4.12185311e-01
-1.66746765e-01 9.37470794e-01 -8.06190789e-01 -4.81381156e-02
-6.38692915e-01 -1.05459607e+00 5.32268167e-01 8.04128110e-01
1.59983009e-01 -7.25297272e-01 1.99241549e-01 5.34246385e-01
-2.37536281e-01 -8.85091364e-01 -7.34018564e-01 1.32936835e-01
-1.11856067e+00 -4.31067675e-01 -9.88519490e-01 -3.29231918e-01
1.06040813e-01 2.49175578e-01 1.23515439e+00 -3.96540791e-01
2.44560421e-01 5.53548753e-01 -1.86098576e-01 -6.92258000e-01
-3.20242643e-01 -1.20562881e-01 -1.12199197e-02 8.51088464e-02
1.56477869e-01 -9.99161541e-01 -7.58758903e-01 1.78159773e-01
-9.99982476e-01 -2.72216827e-01 -3.27506736e-02 7.03079998e-01
6.61322474e-01 3.71160507e-01 4.64614660e-01 -7.56861985e-01
8.79014492e-01 -7.06672072e-01 -6.25230908e-01 2.40064245e-02
-8.32849562e-01 1.77279040e-02 5.45492828e-01 -1.20163310e+00
-1.44793773e+00 -5.36486328e-01 -4.06830043e-01 -6.46406412e-01
3.55368853e-01 9.15556610e-01 5.30310512e-01 4.91190553e-01
7.31002510e-01 1.96414903e-01 -4.26574528e-01 -6.73770368e-01
7.05241621e-01 2.92714298e-01 7.32796967e-01 -8.28545034e-01
7.34582543e-01 7.05923557e-01 -5.42820692e-01 -7.91247427e-01
-8.37564111e-01 -3.37205559e-01 -2.33513892e-01 -2.12948531e-01
6.04956686e-01 -1.19206262e+00 -1.18287476e-02 8.58908236e-01
-1.10320592e+00 -9.14815128e-01 -5.83882391e-01 5.35482585e-01
-4.55023408e-01 3.21672559e-01 -1.21196473e+00 -8.05536509e-01
-2.78735757e-01 -7.22825170e-01 1.01991665e+00 -6.24354146e-02
-5.43050349e-01 -1.29100215e+00 3.04453105e-01 -6.48547262e-02
6.14354491e-01 6.37027025e-02 8.38993251e-01 -2.54190743e-01
-5.55857837e-01 -1.10739097e-01 3.65196615e-01 5.49441278e-01
-2.16040239e-01 6.30041122e-01 -1.06237209e+00 -7.48199970e-02
3.30397248e-01 -8.29065815e-02 1.07656062e+00 8.08877289e-01
1.22365129e+00 -1.01821169e-01 -1.65890306e-01 7.70465672e-01
9.17255938e-01 3.60874802e-01 7.19674110e-01 1.92438439e-01
2.43392080e-01 3.88482958e-01 3.93223047e-01 6.99198782e-01
6.34969473e-01 1.24891251e-01 6.72857463e-02 -8.99946839e-02
-1.98895738e-01 -5.39890826e-01 5.72705507e-01 1.38569224e+00
-1.01430967e-01 -5.09481251e-01 -1.03542435e+00 8.69468987e-01
-1.69901323e+00 -1.30728161e+00 -4.50389802e-01 1.88363767e+00
1.03631377e+00 3.57409865e-01 1.03046581e-01 1.24089047e-01
2.79510021e-01 4.82104868e-01 -5.35059690e-01 6.11406006e-02
-4.63106900e-01 3.19629997e-01 2.59295732e-01 3.69248152e-01
-1.14369226e+00 8.49875808e-01 7.69753742e+00 8.41205597e-01
-1.38916445e+00 2.24987537e-01 8.79584610e-01 -2.11299807e-01
-5.09349406e-01 -4.12215702e-02 -7.33533919e-01 5.52633941e-01
1.35179341e+00 -5.20058215e-01 3.70090783e-01 4.85622168e-01
6.15924299e-01 3.41724157e-01 -8.86912763e-01 8.02834511e-01
-2.21905261e-01 -1.28347766e+00 6.92626368e-03 -2.56028593e-01
1.18327928e+00 5.20239532e-01 6.60308838e-01 5.51095009e-01
1.01800799e+00 -5.92557371e-01 9.90266144e-01 7.89482355e-01
3.97041321e-01 -4.43422258e-01 1.57409251e-01 3.96869123e-01
-9.73774910e-01 -2.59785563e-01 -6.15218095e-02 -1.86647289e-03
5.38990200e-01 7.46796548e-01 -4.64292765e-01 2.29444861e-01
9.09153044e-01 1.14189911e+00 -4.90048975e-01 1.02491891e+00
-3.99224460e-01 1.59331059e+00 -6.17673039e-01 4.62632418e-01
4.50475007e-01 -1.56024873e-01 4.97930348e-01 1.12451494e+00
5.76116383e-01 8.70905891e-02 8.19911659e-02 6.45466208e-01
-6.02950715e-02 -1.81979567e-01 -3.33984464e-01 -2.38820136e-01
3.54620308e-01 8.31472039e-01 -4.20406938e-01 -6.28190875e-01
-5.62348723e-01 7.78157473e-01 -5.89090921e-02 8.84293139e-01
-1.00138509e+00 1.35688931e-01 5.89736462e-01 1.71244651e-01
3.51848572e-01 -5.30930817e-01 -2.27188822e-02 -1.33064508e+00
-1.89209640e-01 -1.10884559e+00 7.48665631e-01 -1.18633771e+00
-1.74050868e+00 6.77599430e-01 2.22795963e-01 -1.07391942e+00
-5.98976910e-01 -2.57151157e-01 -9.26968873e-01 8.46454799e-01
-1.45019722e+00 -1.23510218e+00 8.11029822e-02 7.17461109e-01
8.43623698e-01 2.25783020e-01 4.26226497e-01 3.03676754e-01
-3.98965955e-01 1.50676966e-01 4.11125392e-01 -1.16062164e-01
9.24036443e-01 -1.42274821e+00 8.09747934e-01 1.04017746e+00
3.42352718e-01 7.94659376e-01 1.00186074e+00 -7.67545998e-01
-1.22707939e+00 -1.18227160e+00 6.82398796e-01 -4.85755831e-01
1.42234492e+00 -1.08348407e-01 -1.25082278e+00 1.03418136e+00
6.39123023e-01 -3.24113727e-01 4.86132503e-01 3.07923347e-01
-4.80443954e-01 -3.90397221e-01 -4.74032462e-01 7.41581976e-01
6.65269375e-01 -7.29951978e-01 -5.20059764e-01 5.02904475e-01
1.04310822e+00 -3.05809498e-01 -1.11238313e+00 1.87051386e-01
3.77821892e-01 -6.16279066e-01 1.02831209e+00 -4.78760511e-01
5.38087070e-01 -8.63351021e-03 -8.91180988e-03 -1.57369804e+00
-5.02884150e-01 -1.39835811e+00 -5.47075391e-01 1.68719769e+00
4.65440035e-01 -4.03893739e-01 7.19324946e-01 7.05646157e-01
-1.11224070e-01 -3.60800087e-01 -5.97202539e-01 -9.07097280e-01
4.98178422e-01 -1.08868849e+00 7.94656754e-01 1.01685894e+00
-3.29883665e-01 3.49934846e-01 -5.66938877e-01 8.46492127e-02
5.07956147e-01 3.90654542e-02 7.24432945e-01 -1.11176777e+00
-5.32025456e-01 -5.92431605e-01 1.74657583e-01 -1.72928476e+00
1.13784768e-01 -4.28154022e-01 -1.46880910e-01 -1.38378990e+00
-2.20351636e-01 -3.29213411e-01 -3.96505088e-01 2.20282868e-01
-2.12651581e-01 -1.73671413e-02 -4.49028984e-02 2.23300442e-01
-2.91818112e-01 8.23433995e-01 1.09897327e+00 -4.30294275e-01
-2.55979627e-01 2.47390941e-01 -5.21980941e-01 7.50598192e-01
6.50966644e-01 -5.06166101e-01 -7.95620859e-01 -9.22973275e-01
1.16150454e-01 2.91468263e-01 1.12512842e-01 -7.90216267e-01
4.50377136e-01 -2.53646016e-01 2.04584867e-01 -8.89775276e-01
5.84264040e-01 -5.39846182e-01 3.63315046e-01 6.02094829e-03
-4.51619416e-01 6.20437860e-01 2.93165475e-01 9.02471125e-01
-4.24473494e-01 1.52465060e-01 5.64837396e-01 1.12685911e-01
-6.25092268e-01 5.42149007e-01 -5.40540576e-01 -6.31327694e-03
6.12278342e-01 1.91126451e-01 -3.80984038e-01 -1.05624735e+00
-8.49298716e-01 3.40324730e-01 6.79313093e-02 6.29720151e-01
3.84632915e-01 -1.13536310e+00 -7.00449228e-01 -6.04890008e-03
-3.16636056e-01 -2.75354862e-01 4.15694922e-01 8.43976319e-01
1.10066786e-01 1.11798786e-01 5.40246010e-01 -6.86009109e-01
-1.04867828e+00 7.48786151e-01 2.21718296e-01 -4.31739837e-01
-6.06900871e-01 8.70087743e-01 1.82112634e-01 -6.50812015e-02
3.74066293e-01 -8.80591333e-01 -2.26222929e-02 1.59089208e-01
7.15727210e-01 2.41004080e-01 -1.01826571e-01 -4.47259277e-01
2.77053416e-01 3.77685070e-01 -1.29320443e-01 -5.72090685e-01
1.89811647e+00 -1.52837798e-01 5.33066243e-02 8.14343393e-01
8.42018366e-01 1.91160459e-02 -1.61408758e+00 -4.38108891e-01
1.06807224e-01 -1.81578845e-01 3.36960196e-01 -7.67015755e-01
-1.00496936e+00 1.12817395e+00 4.31987345e-01 4.74475145e-01
1.29029214e+00 -2.59621382e-01 9.87377703e-01 7.00553581e-02
2.65543312e-01 -1.05737686e+00 3.11770618e-01 7.60356367e-01
9.27729607e-01 -8.26471925e-01 -1.33317634e-01 -1.94311500e-01
-6.82985008e-01 9.59294498e-01 3.77984464e-01 -1.19121373e-01
1.06450891e+00 5.11458099e-01 2.97679394e-01 -1.44014299e-01
-1.53228033e+00 -9.22538117e-02 3.55976820e-01 6.58822715e-01
5.30787826e-01 -2.32620418e-01 9.69295204e-02 7.23877370e-01
-2.84222871e-01 -2.11090641e-03 3.05941284e-01 6.50921404e-01
-5.09745851e-02 -1.09178591e+00 -3.68998766e-01 3.94069076e-01
-7.85476089e-01 -1.34737402e-01 -1.24610933e-02 5.52070022e-01
-3.07377189e-01 1.15912020e+00 1.39389202e-01 -1.75543085e-01
4.02072012e-01 4.14323341e-03 2.55800188e-01 -3.10842842e-01
-8.59535336e-01 7.89150715e-01 1.47633115e-02 -3.31892997e-01
-2.76988089e-01 -8.74718964e-01 -9.37864423e-01 -4.32410538e-01
-3.32319170e-01 1.16130345e-01 4.70532089e-01 5.48201740e-01
3.72596622e-01 6.84726179e-01 6.21183217e-01 -7.12732553e-01
-7.72439420e-01 -1.23680770e+00 -3.95552516e-01 5.09941518e-01
3.73674244e-01 -3.19899023e-01 -2.96337366e-01 6.25322282e-01]
|
[7.1767578125, 3.2450811862945557]
|
1bb198e1-111a-42dc-a5f4-2694a6ec5cd8
|
simple-and-deep-graph-convolutional-networks-1
|
2007.02133
| null |
https://arxiv.org/abs/2007.02133v1
|
https://arxiv.org/pdf/2007.02133v1.pdf
|
Simple and Deep Graph Convolutional Networks
|
Graph convolutional networks (GCNs) are a powerful deep learning approach for graph-structured data. Recently, GCNs and subsequent variants have shown superior performance in various application areas on real-world datasets. Despite their success, most of the current GCN models are shallow, due to the {\em over-smoothing} problem. In this paper, we study the problem of designing and analyzing deep graph convolutional networks. We propose the GCNII, an extension of the vanilla GCN model with two simple yet effective techniques: {\em Initial residual} and {\em Identity mapping}. We provide theoretical and empirical evidence that the two techniques effectively relieves the problem of over-smoothing. Our experiments show that the deep GCNII model outperforms the state-of-the-art methods on various semi- and full-supervised tasks. Code is available at https://github.com/chennnM/GCNII .
|
['Yaliang Li', 'Ming Chen', 'Zhewei Wei', 'Zengfeng Huang', 'Bolin Ding']
|
2020-07-04
|
simple-and-deep-graph-convolutional-networks
|
https://proceedings.icml.cc/static/paper_files/icml/2020/2172-Paper.pdf
|
https://proceedings.icml.cc/static/paper_files/icml/2020/2172-Paper.pdf
|
icml-2020-1
|
['node-classification-on-non-homophilic']
|
['graphs']
|
[ 9.04712901e-02 4.58349675e-01 -2.77143091e-01 -4.06458080e-01
-3.15012991e-01 -3.02064955e-01 7.03639686e-01 -1.87636372e-02
-3.44372302e-01 5.10318935e-01 -4.51778388e-03 -6.77219033e-01
-8.58160183e-02 -9.03826535e-01 -8.34557474e-01 -4.47562456e-01
-3.53631049e-01 5.72504818e-01 1.14585049e-01 -8.13977420e-02
4.90886904e-02 3.46959978e-01 -1.09828484e+00 1.09850645e-01
9.48496580e-01 8.02400351e-01 4.31582611e-03 6.29274428e-01
-3.57848734e-01 1.02572346e+00 -3.41358393e-01 -7.46539533e-01
3.43946934e-01 -2.86598176e-01 -9.75462615e-01 -8.58144835e-02
6.31329060e-01 -3.48359615e-01 -8.40676367e-01 1.42136610e+00
4.30732340e-01 1.72572494e-01 3.29983950e-01 -1.60100734e+00
-1.01752794e+00 1.00239062e+00 -6.67630613e-01 1.81381151e-01
-1.11595005e-01 -1.74360171e-01 1.11940110e+00 -8.55577111e-01
6.41592026e-01 1.25659478e+00 9.89246488e-01 4.58566695e-01
-1.17178130e+00 -7.10220516e-01 2.24938959e-01 2.30643630e-01
-1.48517644e+00 -1.66114926e-01 7.33713448e-01 -2.16168299e-01
1.06624520e+00 -2.19162419e-01 4.02811259e-01 9.43372905e-01
2.36926079e-02 1.05252600e+00 7.35621810e-01 -4.09932256e-01
-1.08502090e-01 -3.41703832e-01 2.94135660e-01 1.01823461e+00
5.47258258e-01 -3.44481021e-02 -1.85953900e-01 -9.07484218e-02
8.68448794e-01 2.05396980e-01 -1.61817193e-01 -4.50180918e-01
-7.32321501e-01 9.66212273e-01 7.37694561e-01 3.86232406e-01
-1.29476324e-01 4.47381198e-01 4.62055951e-01 2.62979209e-01
8.11234713e-01 1.61793575e-01 -3.66675586e-01 1.34309694e-01
-8.18540335e-01 1.38976395e-01 8.35918069e-01 1.27153480e+00
9.95338976e-01 1.83540568e-01 7.33070076e-02 7.48854041e-01
2.36951768e-01 1.94237620e-01 1.25075996e-01 -7.34858632e-01
5.80618024e-01 9.21043336e-01 -5.13333917e-01 -1.01287997e+00
-6.85228467e-01 -7.03870475e-01 -1.36912501e+00 -2.02467032e-02
4.38326657e-01 -9.16250870e-02 -1.28040707e+00 1.77296853e+00
3.02815437e-02 5.04553556e-01 -2.74525464e-01 5.21809220e-01
1.25646877e+00 3.49436641e-01 9.01965126e-02 3.89673382e-01
8.42361152e-01 -1.28924513e+00 -7.37819135e-01 -4.78421718e-01
9.72758114e-01 -4.25073594e-01 1.05510151e+00 2.93282896e-01
-9.85337555e-01 -4.54889119e-01 -8.71535778e-01 -3.66583675e-01
-6.22365057e-01 1.25513911e-01 1.02154100e+00 2.87283540e-01
-1.65426433e+00 1.02888083e+00 -8.58356416e-01 -4.13607717e-01
7.49597073e-01 4.49298829e-01 -6.10861659e-01 -3.47470194e-01
-1.12951016e+00 5.47221303e-01 4.33728397e-01 4.00135785e-01
-8.75859082e-01 -6.03402972e-01 -1.08335304e+00 2.76091009e-01
5.64464331e-01 -4.66525137e-01 1.26353884e+00 -8.51752818e-01
-9.08812821e-01 1.05629468e+00 -1.24447234e-02 -7.42848575e-01
3.04472357e-01 -2.93139964e-01 -3.44084740e-01 -5.01729436e-02
6.06856905e-02 3.82540286e-01 5.37258744e-01 -9.59295511e-01
-2.53855020e-01 -3.91728282e-01 1.37370378e-01 -1.44322261e-01
-2.63382584e-01 -7.48501793e-02 -7.38965988e-01 -6.84089601e-01
-7.62091279e-02 -9.96286154e-01 -3.97510797e-01 -8.41086954e-02
-6.97391570e-01 -5.68719745e-01 7.80494153e-01 -6.39229357e-01
1.40088487e+00 -2.00350976e+00 -2.36313716e-02 1.89662263e-01
9.39582169e-01 8.23499382e-01 -4.48703259e-01 6.40666842e-01
-2.65022427e-01 4.03665215e-01 -4.22926426e-01 -6.62600338e-01
5.31884916e-02 3.61461520e-01 1.13258608e-01 4.73073900e-01
6.21271916e-02 1.39941907e+00 -9.77419496e-01 -4.15093303e-01
1.85390085e-01 6.75104916e-01 -3.14468294e-01 1.53486669e-01
-2.98277438e-01 3.82693931e-02 -1.03080831e-01 3.63085926e-01
9.44895506e-01 -8.59595180e-01 4.88682032e-01 -3.08846273e-02
2.74848312e-01 4.03035730e-01 -9.68974292e-01 1.84193850e+00
-1.49292544e-01 8.30073178e-01 2.63111562e-01 -1.29901171e+00
8.22519183e-01 4.31944663e-03 2.25138023e-01 -7.40949154e-01
1.69053808e-01 1.01120770e-01 1.00233257e-01 -1.26516610e-01
4.03806537e-01 2.59259969e-01 2.22714067e-01 6.32904708e-01
3.79160106e-01 2.54168272e-01 4.22693819e-01 8.11903656e-01
1.26139832e+00 1.04885928e-01 2.40680888e-01 -5.54928482e-01
3.23375225e-01 -3.79119903e-01 4.90900248e-01 9.48634923e-01
-3.30285817e-01 6.26542747e-01 8.17020357e-01 -7.30294466e-01
-9.68702555e-01 -8.09335709e-01 3.61585826e-01 1.15810323e+00
-4.64929007e-02 -8.07467520e-01 -9.79641557e-01 -8.21121156e-01
-3.88597995e-02 4.73023117e-01 -8.15312803e-01 -1.81853265e-01
-6.74543738e-01 -6.31947458e-01 6.84147060e-01 8.67070198e-01
7.10242808e-01 -1.16372144e+00 1.50190175e-01 1.90900058e-01
-4.10576425e-02 -1.23488486e+00 -4.32191551e-01 5.46833910e-02
-7.95167208e-01 -1.15282786e+00 -4.70074952e-01 -8.74823034e-01
6.65976524e-01 5.21139741e-01 1.64533424e+00 8.38241279e-01
-2.24323511e-01 3.16698290e-03 -3.36031646e-01 -2.81224430e-01
-4.06608135e-01 5.33851326e-01 -3.19392532e-01 -2.65153348e-01
5.64661264e-01 -7.60073960e-01 -5.69958270e-01 2.24840101e-02
-1.04552722e+00 2.14924470e-01 5.06666124e-01 7.17652917e-01
5.39786994e-01 1.43064737e-01 3.81716192e-01 -1.56217074e+00
7.84355342e-01 -3.69555771e-01 -6.07662082e-01 1.66677773e-01
-8.69178474e-01 2.10184455e-01 7.40408778e-01 -1.79034751e-02
-6.47089899e-01 -2.81794667e-01 -2.77008265e-01 -4.48169380e-01
-9.82661322e-02 8.57732058e-01 -8.53213593e-02 -1.62911519e-01
4.82352346e-01 -3.61283161e-02 8.71426463e-02 -5.32081842e-01
4.65320617e-01 2.60914564e-01 5.11958063e-01 -3.63481730e-01
8.12030613e-01 4.82907116e-01 1.74103394e-01 -6.79412365e-01
-1.17929697e+00 -4.62014824e-01 -7.76905000e-01 -7.70081356e-02
7.03357041e-01 -8.69481981e-01 -4.83184338e-01 7.32737720e-01
-1.14736366e+00 -9.58502531e-01 7.07426071e-02 8.70301351e-02
-2.54111975e-01 6.80805624e-01 -8.55924845e-01 -4.64734435e-01
-6.00442290e-01 -9.09486771e-01 1.03079534e+00 2.47924685e-01
1.13747269e-01 -1.47188294e+00 -7.46483579e-02 7.74363130e-02
5.39736927e-01 2.38657549e-01 7.55726278e-01 -9.21198130e-01
-4.99187231e-01 -1.30874410e-01 -7.30261564e-01 4.35994714e-01
5.02522849e-03 -1.00194355e-02 -8.15291464e-01 -5.55860400e-01
-6.39659524e-01 -2.23341420e-01 1.10815668e+00 5.14579356e-01
1.46233094e+00 -2.21638337e-01 -3.17370653e-01 9.76738870e-01
1.52622187e+00 -1.53178811e-01 8.64075601e-01 1.55692428e-01
1.24709678e+00 1.86070710e-01 -1.71571821e-02 1.00947313e-01
5.50518334e-01 2.66793728e-01 7.18576908e-01 -6.15800083e-01
-2.98271686e-01 -4.01153892e-01 -2.20586248e-02 8.46146107e-01
-2.24619567e-01 -6.45602047e-01 -1.20888376e+00 7.04280436e-01
-2.21145535e+00 -8.22399795e-01 -3.43901813e-01 1.78756058e+00
3.74439597e-01 2.24326208e-01 8.48562270e-02 -4.32198532e-02
8.44430923e-01 4.27475572e-01 -4.81402338e-01 -4.21063930e-01
-2.03142956e-01 5.73896527e-01 6.75514221e-01 5.57472229e-01
-1.28499079e+00 1.23779953e+00 5.76094627e+00 7.05993652e-01
-6.71709895e-01 1.40492404e-02 5.48364341e-01 2.73452967e-01
-2.75655597e-01 6.09987043e-02 -6.97367787e-01 7.67373517e-02
1.01245618e+00 -1.97333425e-01 7.74035990e-01 8.45007181e-01
-2.29254812e-01 3.07591796e-01 -1.02639925e+00 7.78180301e-01
8.53060707e-02 -1.64842308e+00 3.20795253e-02 1.96777910e-01
7.69524515e-01 6.65025830e-01 -6.88860491e-02 4.34310764e-01
9.23370600e-01 -1.24511242e+00 3.02499086e-01 2.87191421e-01
8.50081146e-01 -6.97148919e-01 8.60605299e-01 1.30381092e-01
-1.57404208e+00 1.15122482e-01 -5.20039797e-01 -2.09024042e-01
-4.35245410e-03 7.76884556e-01 -6.05618358e-01 7.97199368e-01
7.90168226e-01 1.06413972e+00 -8.61851633e-01 7.61192083e-01
-4.76885617e-01 8.08091402e-01 -2.33971834e-01 1.33268967e-01
5.11369348e-01 -4.39258039e-01 2.74618030e-01 1.24296260e+00
1.05841728e-02 -8.27767849e-02 2.42807686e-01 1.08840358e+00
-5.60550809e-01 -8.88275132e-02 -9.21360433e-01 -3.86604786e-01
3.20451856e-01 1.35718083e+00 -8.83459687e-01 -4.11544800e-01
-7.24547088e-01 7.77593851e-01 1.03848028e+00 5.70037663e-01
-7.74072468e-01 -6.83247685e-01 5.70271611e-01 1.41534418e-01
4.06357586e-01 -3.64734858e-01 -1.47730976e-01 -1.19620633e+00
-1.58107150e-02 -8.28021646e-01 7.43865669e-01 -6.29259586e-01
-1.41095448e+00 8.09263885e-01 -1.88883990e-01 -6.21612310e-01
-8.72901753e-02 -8.59669149e-01 -8.09896231e-01 7.26072311e-01
-1.59339607e+00 -1.29734433e+00 -4.38580990e-01 7.12823927e-01
1.31511986e-01 -7.54051283e-02 6.25277162e-01 5.47068059e-01
-6.72436416e-01 8.50097179e-01 9.08945724e-02 6.52451098e-01
4.21318293e-01 -1.32858288e+00 1.18952358e+00 1.25519848e+00
2.60783702e-01 7.82723963e-01 3.45328182e-01 -7.24858105e-01
-1.31857514e+00 -1.46856499e+00 9.20048654e-01 -1.74419463e-01
9.37674284e-01 -8.03714216e-01 -1.18902016e+00 1.25267255e+00
4.18046176e-01 3.16623181e-01 3.84944767e-01 3.08703542e-01
-4.97887850e-01 8.93732905e-02 -8.26853633e-01 5.92660308e-01
1.59152663e+00 -5.59866846e-01 -7.12934658e-02 4.83091176e-01
8.86690915e-01 -3.36327046e-01 -7.31474578e-01 4.18417871e-01
1.51967362e-01 -1.12477028e+00 7.52063692e-01 -7.53340602e-01
3.53458047e-01 -1.72868371e-02 1.05143875e-01 -1.24929380e+00
-5.45725346e-01 -8.88178647e-01 -2.57786632e-01 1.05373096e+00
4.57056493e-01 -8.13193202e-01 1.04378128e+00 4.53762650e-01
-3.54325116e-01 -6.39596879e-01 -6.98759556e-01 -8.14570189e-01
2.25925088e-01 -3.40954274e-01 7.07696021e-01 9.63101506e-01
-2.59675562e-01 3.70822161e-01 -3.58061880e-01 6.86419308e-02
8.51100147e-01 4.56635095e-02 9.98847246e-01 -1.45328891e+00
-2.26727612e-02 -5.88968754e-01 -3.96397650e-01 -9.46454406e-01
6.00803792e-01 -1.28061020e+00 -1.79050013e-01 -2.07735205e+00
3.23969096e-01 -2.23797783e-01 -2.90846765e-01 8.17518890e-01
-2.31596246e-01 1.93312049e-01 1.97437946e-02 -1.02729149e-01
-8.31109762e-01 3.45596462e-01 1.02728665e+00 -1.17311008e-01
9.74522009e-02 -1.03013627e-01 -6.99829280e-01 6.23788595e-01
1.19204545e+00 -4.03810650e-01 -5.98504364e-01 -7.30664194e-01
4.69121158e-01 -4.11644638e-01 5.64170122e-01 -8.17369580e-01
2.20012426e-01 2.33336553e-01 -3.45440805e-02 -6.67579830e-01
-2.66055077e-01 -5.10468543e-01 1.11731395e-01 3.13704610e-01
-2.16918558e-01 3.85540277e-01 3.23606342e-01 5.90726316e-01
-1.12702399e-01 9.23042297e-02 7.97826290e-01 -2.88315207e-01
-7.38463342e-01 9.49236214e-01 -5.81751764e-02 3.72363538e-01
4.34895426e-01 7.23707676e-02 -7.59281218e-01 -6.89788878e-01
-5.46748340e-01 3.14310282e-01 3.24307591e-01 3.35984349e-01
4.34774131e-01 -1.31044340e+00 -6.09539986e-01 2.62214720e-01
-6.03549518e-02 4.09130335e-01 1.58986777e-01 8.23461413e-01
-6.45793259e-01 3.18357080e-01 3.94132957e-02 -3.13788652e-01
-1.01699281e+00 7.08607614e-01 4.81313616e-01 -6.92179739e-01
-1.00919402e+00 9.98312593e-01 4.14574772e-01 -7.48484135e-01
3.63887042e-01 -3.36889088e-01 1.39752194e-01 -5.17526627e-01
3.31009775e-01 3.78972322e-01 1.63701162e-01 -4.12422210e-01
-2.76261866e-01 1.99685290e-01 -3.63780230e-01 4.50873345e-01
1.55268574e+00 -7.58326054e-02 -4.58295763e-01 -9.11655426e-02
1.37831545e+00 -5.31920731e-01 -1.10910332e+00 -6.42954588e-01
1.46511257e-01 -8.28993917e-02 9.08562094e-02 -4.87356991e-01
-1.45896721e+00 1.14579821e+00 9.24361125e-02 4.78378385e-01
1.07426786e+00 9.10766944e-02 9.39278841e-01 3.36466253e-01
8.27160180e-02 -9.92224395e-01 -1.44947559e-01 6.61119044e-01
6.79454684e-01 -1.26548839e+00 1.62453145e-01 -4.55896556e-01
-3.59050304e-01 9.29020345e-01 7.74751902e-01 -3.44783902e-01
9.73330796e-01 3.21124941e-01 -2.96557434e-02 -6.59105420e-01
-6.16400361e-01 -3.14178795e-01 2.42422789e-01 7.05148280e-01
5.41080058e-01 1.92083597e-01 -1.04124658e-01 4.83238846e-01
-1.51570335e-01 -4.63212095e-02 4.86439824e-01 7.46585667e-01
-8.99449661e-02 -1.06860578e+00 4.09565002e-01 6.14577353e-01
-4.75042284e-01 -5.10390699e-01 -7.03643501e-01 1.15064180e+00
-3.14706326e-01 7.97957599e-01 1.66179482e-02 -4.03776348e-01
9.18566510e-02 -1.59266517e-01 3.36103857e-01 -5.15476406e-01
-4.58301306e-01 -1.33612901e-01 2.90535688e-01 -8.52132440e-01
-3.88171554e-01 -2.39190951e-01 -1.35986114e+00 -7.52246022e-01
-3.24208319e-01 -1.23629961e-02 3.28805029e-01 7.55256414e-01
6.37724459e-01 5.41903615e-01 1.35948911e-01 -8.67160261e-01
-3.33685219e-01 -1.05894971e+00 -6.94791079e-01 5.52125990e-01
2.17004329e-01 -5.79293966e-01 -4.75791454e-01 -3.37488472e-01]
|
[7.002048969268799, 6.219116687774658]
|
fec4f6b8-eb1a-4a7b-a2c7-a11fbd0aaf39
|
differentiable-programming-for-hyperspectral
|
2007.05996
| null |
https://arxiv.org/abs/2007.05996v1
|
https://arxiv.org/pdf/2007.05996v1.pdf
|
Differentiable Programming for Hyperspectral Unmixing using a Physics-based Dispersion Model
|
Hyperspectral unmixing is an important remote sensing task with applications including material identification and analysis. Characteristic spectral features make many pure materials identifiable from their visible-to-infrared spectra, but quantifying their presence within a mixture is a challenging task due to nonlinearities and factors of variation. In this paper, spectral variation is considered from a physics-based approach and incorporated into an end-to-end spectral unmixing algorithm via differentiable programming. The dispersion model is introduced to simulate realistic spectral variation, and an efficient method to fit the parameters is presented. Then, this dispersion model is utilized as a generative model within an analysis-by-synthesis spectral unmixing algorithm. Further, a technique for inverse rendering using a convolutional neural network to predict parameters of the generative model is introduced to enhance performance and speed when training data is available. Results achieve state-of-the-art on both infrared and visible-to-near-infrared (VNIR) datasets, and show promise for the synergy between physics-based models and deep learning in hyperspectral unmixing in the future.
|
['Suren Jayasuriya', 'Philip Christensen', 'John Janiczek', 'Gautam Dasarathy', 'Christopher S. Edwards', 'Parth Thaker']
|
2020-07-12
| null |
https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/5893_ECCV_2020_paper.php
|
https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123720647.pdf
|
eccv-2020-8
|
['hyperspectral-unmixing']
|
['computer-vision']
|
[ 7.35784292e-01 -7.45226085e-01 7.62519985e-02 -1.34713128e-01
-9.30036426e-01 -6.63688600e-01 5.88255465e-01 -3.25505793e-01
1.17265701e-01 4.81563330e-01 -1.06523614e-02 -2.51566768e-01
-3.78941029e-01 -9.02393639e-01 -5.31014264e-01 -1.14375889e+00
1.36885911e-01 3.74013633e-01 -5.75493693e-01 -2.94342816e-01
-4.15974915e-01 8.04029763e-01 -1.64985561e+00 1.31903023e-01
1.28282166e+00 9.15724337e-01 3.81252408e-01 7.82722533e-01
-2.23745883e-01 5.19655704e-01 -1.50289938e-01 2.44251564e-01
6.80524766e-01 -3.95317018e-01 -3.48915458e-01 4.66096938e-01
7.29041278e-01 -3.51396412e-01 -4.21827793e-01 1.25274181e+00
4.77022916e-01 4.38993752e-01 8.48154187e-01 -1.08928871e+00
-7.76905954e-01 4.08131391e-01 -8.43336165e-01 -3.17155719e-01
-7.87952840e-02 3.63376588e-01 8.78075600e-01 -5.97026169e-01
8.02130252e-02 1.07827294e+00 6.85563624e-01 4.49562930e-02
-1.48959458e+00 -6.56898022e-01 -3.03282477e-02 -1.38947926e-02
-1.43252611e+00 -2.14954183e-01 1.18130243e+00 -8.19764137e-01
6.33102894e-01 5.18822908e-01 8.93032908e-01 6.93584502e-01
-1.09493762e-01 4.36328292e-01 1.34138501e+00 -4.24246609e-01
-6.87723085e-02 -2.06347525e-01 -8.36961493e-02 4.12253797e-01
-1.17626013e-02 5.09919524e-01 -2.09250227e-01 -1.78794399e-01
6.81344569e-01 2.16784135e-01 -5.04916131e-01 -1.53056964e-01
-9.84641910e-01 7.10932314e-01 8.50591660e-01 -1.62804425e-01
-6.84856057e-01 -3.83594893e-02 -3.01189601e-01 1.38439313e-01
9.87648249e-01 4.69911009e-01 -2.32369334e-01 5.32707572e-01
-1.26887250e+00 3.80604327e-01 4.21252012e-01 5.53759038e-01
1.21160233e+00 2.83447117e-01 5.28472150e-03 1.35111463e+00
7.10737884e-01 1.10580754e+00 1.13553077e-01 -9.33606744e-01
1.40872940e-01 4.30372506e-01 7.80477077e-02 -6.26477599e-01
-5.35479605e-01 -7.99221635e-01 -9.55783963e-01 4.93380696e-01
3.87682836e-03 -7.58554861e-02 -1.11006331e+00 1.51152945e+00
3.38415235e-01 3.08773786e-01 1.31343529e-01 1.09526575e+00
9.32044089e-01 1.09668410e+00 1.91955920e-02 -3.25024784e-01
1.17324293e+00 -1.14746368e+00 -5.28758347e-01 -5.37109673e-01
1.59776181e-01 -8.33157063e-01 8.89865398e-01 2.12992117e-01
-7.67253935e-01 -5.12352288e-01 -9.47883070e-01 5.61272874e-02
-2.34589368e-01 2.51703739e-01 6.27385080e-01 6.20572150e-01
-8.62889469e-01 7.92828202e-01 -7.44392037e-01 -6.30885139e-02
3.75716776e-01 3.21116298e-02 1.24460571e-01 -7.62649477e-02
-9.53485906e-01 5.72855055e-01 4.16566789e-01 3.59919548e-01
-6.03557467e-01 -1.27324688e+00 -6.41891479e-01 5.27915955e-02
1.02228820e-01 -8.74639750e-01 1.00381827e+00 -1.22764337e+00
-1.70230913e+00 7.40433633e-01 -1.40371993e-01 1.21473424e-01
5.07438958e-01 -3.57656777e-02 -8.45177412e-01 1.33047447e-01
-1.66746721e-01 3.09281945e-01 1.10531509e+00 -1.71754146e+00
-2.74519295e-01 -2.13039055e-01 -1.87347874e-01 4.21316028e-01
-1.16899461e-01 -6.88233823e-02 -1.09721497e-01 -6.47240043e-01
5.69708228e-01 -1.07372797e+00 -2.28550449e-01 2.75582224e-01
-5.06248951e-01 5.50769150e-01 7.81900406e-01 -1.15273392e+00
7.07240939e-01 -2.02966785e+00 4.93965000e-02 3.44419211e-01
3.35097790e-01 3.36558610e-01 -3.65420282e-01 3.80245358e-01
-5.34048140e-01 -9.81261060e-02 -7.91294634e-01 -2.98990667e-01
1.83924027e-02 -1.59389868e-01 -3.10232282e-01 6.07887924e-01
2.13971689e-01 9.35939133e-01 -7.93499112e-01 2.21414909e-01
6.12467647e-01 7.67205954e-01 -1.44861147e-01 2.98485070e-01
-5.44414282e-01 7.03173697e-01 -5.02850227e-02 8.75337005e-01
1.28654635e+00 -2.89883643e-01 6.53906390e-02 -7.50969589e-01
-2.24242553e-01 -1.35546967e-01 -1.09968591e+00 1.46823823e+00
-6.06334627e-01 5.13374448e-01 5.35730720e-01 -8.11499655e-01
8.79055083e-01 1.49535358e-01 7.15487421e-01 -5.33736229e-01
-8.02496523e-02 4.11344469e-01 6.43608570e-02 -3.81471336e-01
6.17016137e-01 -5.34123123e-01 7.38795936e-01 4.15852398e-01
-3.37759644e-01 -6.87650084e-01 -1.45070285e-01 -2.82454014e-01
2.59772211e-01 4.14632618e-01 1.08835129e-02 -2.43272156e-01
5.86075187e-01 1.81613743e-01 1.78520903e-01 4.70080465e-01
2.91314989e-01 7.34348774e-01 -3.28871280e-01 -2.57036328e-01
-1.07279706e+00 -1.07783639e+00 -3.25066537e-01 9.22546804e-01
-6.73160329e-02 1.78293735e-01 -4.68089044e-01 2.46452168e-01
1.58527970e-01 6.47841215e-01 -2.24001721e-01 -2.92130988e-02
-3.76820788e-02 -1.40354908e+00 2.44966671e-01 2.55439252e-01
6.78240836e-01 -7.70704567e-01 7.20265955e-02 3.16952825e-01
-2.85267413e-01 -1.15406418e+00 -7.95519277e-02 -4.00291085e-02
-7.74954200e-01 -8.86112034e-01 -7.28782773e-01 -2.96355605e-01
3.31761420e-01 6.40105963e-01 9.25800264e-01 -3.27097088e-01
-4.21301335e-01 2.41592064e-01 -7.82991797e-02 -5.07988513e-01
-9.70758557e-01 -1.59594148e-01 -1.85912564e-01 2.25853473e-01
2.81785172e-03 -7.52845228e-01 -5.90972185e-01 1.48939028e-01
-1.17747235e+00 3.80481511e-01 5.56651533e-01 5.29649377e-01
7.10639656e-01 3.29119891e-01 2.39849612e-02 -6.50182903e-01
3.15790147e-01 -3.87022972e-01 -8.00086081e-01 3.23457360e-01
-5.65892518e-01 -1.34545460e-01 5.07841885e-01 -4.61197466e-01
-1.30753720e+00 1.15643196e-01 -5.23876073e-03 -6.36597157e-01
-2.63600588e-01 9.45788026e-01 -4.45137084e-01 -3.29457283e-01
8.04809570e-01 3.23813617e-01 1.99850470e-01 -5.89308918e-01
4.77892816e-01 8.61446261e-01 7.31299281e-01 -5.38983703e-01
1.21777534e+00 7.68541813e-01 2.90959060e-01 -1.44688606e+00
-9.76731956e-01 -7.48739958e-01 -4.91857052e-01 -3.41788411e-01
5.22611320e-01 -1.37387311e+00 -2.35235706e-01 1.09075940e+00
-8.56844127e-01 -5.69652677e-01 -1.30706921e-01 4.63811636e-01
-4.73046273e-01 5.66138387e-01 -3.84382665e-01 -1.03565490e+00
-5.09706914e-01 -1.14472091e+00 1.10364461e+00 1.25103280e-01
2.69367009e-01 -1.13603485e+00 7.66239017e-02 6.78993285e-01
5.12266755e-01 5.14167249e-01 8.48059416e-01 1.40896186e-01
-6.01515114e-01 -8.09416994e-02 -5.95172524e-01 6.07337713e-01
5.74529052e-01 4.79330838e-01 -1.51759684e+00 -3.24173301e-01
1.12789221e-01 2.41983049e-02 1.02813554e+00 8.18153620e-01
1.31809425e+00 -1.68608874e-01 -1.03236757e-01 1.33489859e+00
1.73706293e+00 -9.74132642e-02 6.16954267e-01 3.13634038e-01
1.31153047e+00 7.23900080e-01 2.85371356e-02 3.61985028e-01
5.99002838e-02 4.07256901e-01 7.30121017e-01 -6.32684052e-01
-2.47697040e-01 2.11716399e-01 1.54806674e-01 5.31943798e-01
-2.26090997e-01 -2.91936189e-01 -8.63452077e-01 5.15854880e-02
-1.58347344e+00 -1.12664795e+00 -6.71245515e-01 2.23046279e+00
7.40324020e-01 -6.03578329e-01 -8.72183889e-02 4.26256061e-02
9.05546010e-01 5.62766075e-01 -7.54619777e-01 3.13033700e-01
-6.39819801e-01 1.96667284e-01 7.32571781e-01 6.98754728e-01
-1.27510464e+00 7.24916279e-01 6.29929495e+00 6.30227685e-01
-1.49895918e+00 -2.27727424e-02 4.65732694e-01 1.71606854e-01
-6.28292382e-01 -1.40568823e-01 -2.58230537e-01 2.19797820e-01
8.07319164e-01 1.73935965e-01 1.07729387e+00 1.75932750e-01
4.72218573e-01 6.23599142e-02 -6.59176886e-01 1.11236608e+00
-1.41263947e-01 -1.25760257e+00 7.40602538e-02 1.91566542e-01
1.02905619e+00 4.05442029e-01 2.83877641e-01 -2.63799310e-01
2.09433809e-01 -1.09513712e+00 7.38280535e-01 8.17756891e-01
9.25870538e-01 -5.63452303e-01 1.95511609e-01 3.77615213e-01
-1.33830118e+00 -6.31293878e-02 -3.62598449e-01 1.29849002e-01
1.13874376e-01 9.35469449e-01 -4.71183002e-01 9.52862084e-01
3.10421795e-01 8.62778485e-01 -2.37645671e-01 1.01665473e+00
-2.61443704e-01 6.09581470e-01 -4.39973563e-01 5.17004967e-01
3.80394086e-02 -1.00134456e+00 6.94817483e-01 1.14925575e+00
6.49722338e-01 7.25655183e-02 3.26376259e-01 1.49894321e+00
-3.16412747e-02 -6.20702691e-02 -9.11249667e-02 -5.54467797e-01
1.61954269e-01 1.59916389e+00 -4.50997591e-01 -1.31863728e-01
-1.79649860e-01 7.18831599e-01 -2.97136664e-01 7.50597537e-01
-7.40858078e-01 2.66817212e-01 9.40053284e-01 1.59759045e-01
-7.54002109e-02 -2.60838091e-01 -2.35641509e-01 -1.16459358e+00
-6.62665740e-02 -9.19618428e-01 1.64140034e-02 -1.29894531e+00
-1.64663792e+00 2.92622596e-01 -6.52531460e-02 -1.32873666e+00
1.33324459e-01 -9.66657400e-01 -6.41013205e-01 1.59529042e+00
-2.09108424e+00 -1.57741117e+00 -1.07788634e+00 2.48124868e-01
8.30239877e-02 7.90271088e-02 8.84264588e-01 6.14644587e-02
-5.38777232e-01 -1.21100657e-01 7.45548010e-01 -2.87222296e-01
4.94395226e-01 -1.10296750e+00 2.84660190e-01 9.57049012e-01
-1.03045411e-01 1.94015086e-01 7.84294188e-01 -6.17256224e-01
-1.46815765e+00 -1.50313342e+00 -4.46977206e-02 5.65696247e-02
8.95269811e-01 1.28504038e-01 -1.10834634e+00 3.43116313e-01
-2.38240734e-02 2.16079116e-01 1.01198065e+00 -2.81240910e-01
-4.12868947e-01 -1.59562767e-01 -1.02315485e+00 4.18716013e-01
8.70154262e-01 -8.55635464e-01 4.03748639e-02 6.70069635e-01
5.56782067e-01 -5.72906435e-01 -8.33881199e-01 4.40600395e-01
6.43218994e-01 -7.87102878e-01 1.34435415e+00 -2.77166128e-01
6.45781100e-01 -5.89829445e-01 -4.65969503e-01 -1.76424420e+00
-7.19959199e-01 -7.91032434e-01 -1.63133055e-01 1.03035915e+00
4.30043846e-01 -6.92942202e-01 5.23101270e-01 4.45959240e-01
-2.44539067e-01 -2.45246794e-02 -4.08987910e-01 -8.61574769e-01
1.86290115e-01 -4.07741517e-01 9.57022250e-01 1.22005999e+00
-8.45840156e-01 2.30488777e-02 -2.72395104e-01 9.27086115e-01
9.10686255e-01 4.02210832e-01 6.25864804e-01 -1.60877216e+00
-4.89581615e-01 -7.93208539e-01 1.81242958e-01 -7.82159865e-01
3.33462387e-01 -1.02897298e+00 5.72889484e-02 -1.75592196e+00
1.07556373e-01 -4.37772483e-01 -2.07846329e-01 1.63450807e-01
-4.43407923e-01 2.29466423e-01 9.16007832e-02 5.42147636e-01
4.87386853e-01 7.11346149e-01 1.37058115e+00 -7.25655675e-01
-4.33787227e-01 3.99888083e-02 -6.28678918e-01 5.31156540e-01
8.87449503e-01 -1.10374369e-01 -3.39456707e-01 -4.21862006e-01
1.58129454e-01 3.90739180e-02 6.85745537e-01 -1.06629193e+00
-1.79733038e-01 -4.46373433e-01 3.27780277e-01 -5.61714411e-01
6.68540597e-01 -9.63811338e-01 8.32274020e-01 1.15513690e-01
-4.22641672e-02 -5.97111285e-01 4.56951171e-01 3.69296461e-01
8.04862455e-02 -1.66879311e-01 9.33123887e-01 5.11533581e-02
-6.25315368e-01 7.71164238e-01 -1.41321391e-01 -3.91772300e-01
5.74235499e-01 -2.73499995e-01 -4.93209898e-01 -3.54755819e-01
-5.39133906e-01 -1.55306295e-01 6.42433643e-01 -3.10895517e-02
2.34857962e-01 -1.16278815e+00 -1.10905182e+00 3.25797766e-01
3.07682902e-01 1.03631541e-01 6.72369242e-01 6.27931058e-01
-9.90834355e-01 -2.01908380e-01 2.43232609e-03 -7.43340731e-01
-1.10238385e+00 3.61744165e-01 9.46398377e-01 1.36411777e-02
-3.79233390e-01 7.52410948e-01 2.38854155e-01 -9.11119103e-01
-3.83548766e-01 -3.06539118e-01 1.06078938e-01 1.64250769e-02
4.97146785e-01 5.53450525e-01 1.51849344e-01 -9.02637005e-01
7.31350482e-03 5.91269493e-01 6.52078867e-01 -9.38346703e-03
1.63100851e+00 -9.89563018e-02 -5.84828436e-01 4.60417479e-01
1.16247308e+00 -9.48545709e-02 -1.63977325e+00 -6.30813718e-01
-6.87752187e-01 -3.98204565e-01 6.83236539e-01 -9.07365739e-01
-1.23583913e+00 9.61975873e-01 6.93154097e-01 3.24975610e-01
1.23606455e+00 -3.22327346e-01 4.91081566e-01 2.78862089e-01
-2.63637304e-01 -8.17187786e-01 -4.17251319e-01 3.95874381e-01
9.91898298e-01 -1.38782370e+00 3.48515779e-01 -7.94757783e-01
-3.17142606e-01 1.36524618e+00 1.86931580e-01 2.26099402e-01
7.71401167e-01 1.62541375e-01 4.19897646e-01 -1.45115227e-01
7.78069794e-02 -2.56047785e-01 7.53933251e-01 8.01123917e-01
4.66288596e-01 4.66156006e-01 5.34061491e-01 -4.83562648e-02
-2.75392413e-01 -2.54854620e-01 2.55050004e-01 5.16785383e-01
-3.85265976e-01 -8.69891107e-01 -8.13813806e-01 5.68185627e-01
8.18558782e-02 -4.09316659e-01 -1.05964832e-01 2.15448752e-01
6.22984692e-02 1.08582473e+00 1.07316360e-01 -8.39443356e-02
-4.57983837e-02 1.63809255e-01 4.31203485e-01 -5.12430131e-01
-2.50174463e-01 5.13509154e-01 -2.17393398e-01 -1.44050166e-01
-9.96943474e-01 -5.76302111e-01 -7.93896735e-01 -5.00725031e-01
-4.35582787e-01 -3.29018414e-01 8.62467647e-01 8.43808353e-01
-1.01475865e-02 7.41726220e-01 8.21632445e-01 -1.34541166e+00
-5.34435749e-01 -9.95316505e-01 -1.09937680e+00 1.94823936e-01
5.60113251e-01 -4.62260067e-01 -4.60064799e-01 9.32477340e-02]
|
[10.083846092224121, -2.0253851413726807]
|
7f9d0a63-1e23-4210-90c1-1475b2df6ac6
|
efficient-plane-based-optimization-of
|
1905.08853
| null |
https://arxiv.org/abs/1905.08853v1
|
https://arxiv.org/pdf/1905.08853v1.pdf
|
Efficient Plane-Based Optimization of Geometry and Texture for Indoor RGB-D Reconstruction
|
We propose a novel approach to reconstruct RGB-D indoor scene based on plane primitives. Our approach takes as input a RGB-D sequence and a dense coarse mesh reconstructed from it, and generates a lightweight, low-polygonal mesh with clear face textures and sharp features without losing geometry details from the original scene. Compared to existing methods which only cover large planar regions in the scene, our method builds the entire scene by adaptive planes without losing geometry details and also preserves sharp features in the mesh. Experiments show that our method is more efficient to generate textured mesh from RGB-D data than state-of-the-arts.
|
['Chao Wang', 'Xiaohu Guo']
|
2019-05-21
| null | null | null | null |
['rgb-d-reconstruction']
|
['computer-vision']
|
[ 4.62387204e-01 3.55388552e-01 6.12309396e-01 -2.82359391e-01
-5.59129298e-01 -5.00543118e-01 2.74915248e-01 -1.03075795e-01
2.22132921e-01 6.62501991e-01 -1.08988456e-01 1.05462754e-02
3.07838209e-02 -1.48840380e+00 -9.12789524e-01 -3.43931109e-01
3.48014802e-01 6.74930930e-01 5.86308002e-01 -2.11976692e-01
1.33397639e-01 9.55289125e-01 -1.73743522e+00 3.48583430e-01
6.49308443e-01 1.25336742e+00 1.08023815e-01 4.87216860e-01
-3.74997228e-01 5.69931507e-01 -2.02631578e-01 -3.99371415e-01
6.13452494e-01 -2.35441998e-01 -8.55045617e-01 6.16935372e-01
6.61812365e-01 -3.63245636e-01 -5.29683590e-01 7.17446804e-01
1.56061456e-01 -1.02176398e-01 3.13311696e-01 -6.27629697e-01
-1.62273437e-01 -4.66090888e-01 -6.68410301e-01 -8.42862129e-01
1.15356755e+00 -1.11942455e-01 1.83783457e-01 -1.07122350e+00
1.09811091e+00 1.31403553e+00 1.05118573e+00 4.17325586e-01
-1.40241539e+00 -4.67060506e-01 -2.43108108e-04 -4.78704989e-01
-1.59765458e+00 -2.73737192e-01 1.25860536e+00 -3.45860869e-02
7.08435774e-01 5.94695926e-01 1.11771297e+00 7.45334089e-01
2.81384140e-01 3.05427700e-01 1.46356893e+00 -4.12458122e-01
4.00504738e-01 -4.22300696e-01 -4.90561545e-01 1.01044357e+00
-6.76684603e-02 4.10066657e-02 -6.30423665e-01 -3.57598543e-01
1.59705102e+00 2.49679074e-01 -3.64531428e-01 -7.64588594e-01
-1.28172970e+00 1.48212016e-01 4.96030360e-01 -2.35985219e-01
-4.63773280e-01 3.50250423e-01 -2.35882699e-01 1.20138347e-01
7.91768014e-01 -1.14992388e-01 -3.22560459e-01 -2.33147472e-01
-7.68637419e-01 2.40231976e-01 6.44401610e-01 1.29248023e+00
1.30672812e+00 -2.29113445e-01 6.29311740e-01 3.36309224e-01
1.75832048e-01 7.42367923e-01 -4.32695180e-01 -1.47882128e+00
4.01385903e-01 8.02256405e-01 1.80707082e-01 -1.07727110e+00
-1.36383027e-01 5.23617566e-01 -8.35970044e-01 6.77069902e-01
3.07539236e-02 2.39804462e-01 -1.01716292e+00 6.81506991e-01
9.07413900e-01 4.52999882e-02 -2.62529016e-01 4.89240915e-01
1.03449690e+00 5.85609138e-01 -7.54280508e-01 1.60212830e-01
9.60471094e-01 -5.59996367e-01 -5.61782002e-01 6.53277561e-02
-1.10457756e-01 -8.46258104e-01 1.26096034e+00 7.29341805e-01
-1.25812209e+00 -3.67017955e-01 -7.67008483e-01 -5.26497543e-01
7.92922378e-02 -3.32328171e-01 7.51851380e-01 6.74354494e-01
-1.07137656e+00 8.20917487e-01 -8.28180134e-01 -1.73362523e-01
5.74469984e-01 2.91639715e-01 -5.83612263e-01 -5.42314053e-01
-2.51604974e-01 3.54771227e-01 3.98264155e-02 -1.18814960e-01
-6.59911394e-01 -9.02900577e-01 -8.64083946e-01 -5.10475457e-01
3.53417337e-01 -1.00504315e+00 1.10173452e+00 -6.75467491e-01
-2.21060491e+00 9.75904167e-01 -3.77565950e-01 4.31109399e-01
7.31205285e-01 -1.68451771e-01 9.20124575e-02 3.14450771e-01
-2.00917527e-01 2.00433001e-01 7.46381879e-01 -2.01815963e+00
-3.62405598e-01 -4.83163267e-01 2.68523633e-01 2.52464116e-01
6.30041659e-01 -5.50386727e-01 -6.03209376e-01 -5.74744761e-01
8.34739387e-01 -6.22085929e-01 -4.85908717e-01 8.20633888e-01
-5.87166071e-01 4.40176308e-01 9.84285593e-01 -5.36410987e-01
4.61542517e-01 -2.20410585e+00 1.78450763e-01 6.49587035e-01
2.55995482e-01 -5.81156254e-01 3.83907735e-01 4.23756033e-01
4.18660372e-01 2.47855880e-03 -5.28303087e-01 -8.78915608e-01
-1.27125174e-01 6.30101323e-01 -2.57945210e-01 6.64960146e-01
-3.35231930e-01 3.60937953e-01 -9.51509953e-01 -4.58362013e-01
6.87905908e-01 1.06886518e+00 -6.85620189e-01 1.33545339e-01
-3.73331428e-01 9.12591159e-01 -6.59008503e-01 1.10061061e+00
1.01056242e+00 6.90179542e-02 1.33788854e-01 -2.24108025e-01
-3.16397935e-01 2.40105823e-01 -1.49917257e+00 2.31561136e+00
-5.56279063e-01 1.47266924e-01 4.26972806e-01 -2.43858725e-01
9.74616587e-01 2.39966661e-01 5.67091167e-01 -7.30481923e-01
5.36731398e-03 3.09662461e-01 -1.14021385e+00 -1.26778379e-01
4.01720434e-01 -3.54996681e-01 -2.64966916e-02 1.95029348e-01
-5.17408133e-01 -1.29217076e+00 -7.26429760e-01 -1.69230580e-01
1.06478870e+00 8.58785629e-01 -7.78563274e-03 -1.67584196e-01
2.12096259e-01 2.68889815e-01 4.09390539e-01 1.51965737e-01
6.75329268e-01 1.09858918e+00 1.94080770e-02 -8.11604261e-01
-1.15370643e+00 -1.51345921e+00 -3.64598453e-01 8.32432434e-02
6.82474554e-01 -5.71769953e-01 -8.73913169e-01 -3.08146745e-01
1.15249306e-01 3.83156359e-01 -6.91596568e-01 5.69473803e-01
-6.79423630e-01 3.23032625e-02 3.55266705e-02 2.05700636e-01
1.03533649e+00 -6.55439496e-01 -8.43036830e-01 5.43345772e-02
-1.33335710e-01 -1.14264059e+00 -1.78368181e-01 -2.55514592e-01
-1.29060137e+00 -1.19131565e+00 -4.30200160e-01 -6.70345247e-01
1.19781089e+00 2.96814531e-01 1.43550575e+00 1.17132470e-01
-4.00925815e-01 7.97779143e-01 -2.11372226e-01 -1.85595781e-01
-4.72157262e-02 -6.12885237e-01 -2.22381979e-01 -2.62921005e-02
-5.11783361e-01 -8.56114924e-01 -6.29940152e-01 3.76799822e-01
-8.58013093e-01 7.57357001e-01 -3.00125405e-02 2.20916182e-01
1.72549605e+00 2.92335570e-01 -7.18867958e-01 -9.39339697e-01
-2.10025296e-01 2.53402088e-02 -7.39386380e-01 -6.05965182e-02
1.49305835e-01 -1.41223207e-01 4.57525849e-01 -3.93595882e-02
-1.24471533e+00 5.12522817e-01 -1.92579448e-01 -5.61042190e-01
-1.69217557e-01 -2.48173118e-01 -3.65393579e-01 -6.15482986e-01
2.69360304e-01 2.05612391e-01 -4.00717705e-01 -9.97099996e-01
2.73434997e-01 1.05177544e-01 5.06570399e-01 -8.67471635e-01
1.05202675e+00 1.29567707e+00 4.48466986e-01 -1.17655146e+00
-5.10847926e-01 1.21564148e-02 -1.09602439e+00 -2.98624247e-01
7.34216094e-01 -7.11316228e-01 -8.06523502e-01 5.02560496e-01
-1.37929463e+00 -7.59498835e-01 -8.12821984e-01 1.04532994e-01
-9.79048967e-01 1.67115599e-01 -5.38428903e-01 -6.26301646e-01
1.49109945e-01 -9.37057912e-01 1.62502098e+00 -1.61445022e-01
-1.38696097e-02 -7.19889820e-01 2.58271664e-01 2.55539536e-01
1.28399625e-01 1.20581210e+00 6.33912027e-01 1.15231371e+00
-1.21523011e+00 7.05506951e-02 1.85647234e-02 -1.49418652e-01
4.70861197e-01 2.44611278e-01 -1.17674708e+00 2.81329483e-01
1.38153374e-01 -8.33135620e-02 1.97003633e-01 1.57970518e-01
1.46002591e+00 -2.94135332e-01 -3.01092416e-01 1.25362957e+00
1.80663085e+00 -1.08836472e-01 1.10917318e+00 2.87828594e-01
1.08107722e+00 3.48509669e-01 4.26487297e-01 4.62446749e-01
6.90240920e-01 6.62494838e-01 6.72907174e-01 -4.65061754e-01
-3.33573222e-01 -5.98625183e-01 -3.26753035e-02 7.23038554e-01
-7.64580429e-01 1.81700155e-01 -8.86033893e-01 1.01168849e-01
-1.29713106e+00 -6.19942129e-01 -6.01670980e-01 2.40563869e+00
9.78591919e-01 -1.14824347e-01 -1.73618346e-01 5.15924096e-01
1.36367336e-01 2.79990141e-03 -2.24363536e-01 -4.48422372e-01
-9.29713845e-02 7.55850554e-01 5.52274585e-01 8.48986328e-01
-6.84464455e-01 8.64032090e-01 6.61851406e+00 5.51795483e-01
-9.13031995e-01 1.06125936e-01 4.14865285e-01 -4.91386615e-02
-7.18202949e-01 -1.97151434e-02 -2.75761276e-01 6.81168446e-03
3.71571690e-01 4.49662469e-02 4.92066264e-01 1.02336872e+00
-1.95566807e-02 -4.59101021e-01 -9.90012527e-01 1.32646644e+00
-1.40073970e-01 -1.51234663e+00 5.36465012e-02 1.93859816e-01
1.21692884e+00 -2.28967562e-01 -2.44130060e-01 -5.49688637e-01
5.17806768e-01 -1.06332266e+00 1.22018719e+00 8.69680464e-01
1.24076569e+00 -8.51290107e-01 1.30540550e-01 3.10450852e-01
-1.50637293e+00 6.89243495e-01 -5.35422862e-01 -2.58850694e-01
1.49005920e-01 1.02116787e+00 -2.78162211e-01 8.44143391e-01
9.67870533e-01 5.17153740e-01 -2.07094550e-01 9.50995743e-01
-2.44996086e-01 2.58262977e-02 -6.39462948e-01 4.72672552e-01
-1.01958416e-01 -4.87278253e-01 1.44552767e-01 7.35785663e-01
4.77148622e-01 6.53168321e-01 4.27105397e-01 7.83862889e-01
-2.68214971e-01 3.44167463e-02 -8.93064499e-01 6.55413091e-01
2.19491571e-01 7.61989653e-01 -1.01446521e+00 -3.79367441e-01
-2.85454452e-01 1.53149354e+00 1.48408934e-02 2.58660913e-01
-5.99929571e-01 -2.46192425e-01 6.16986036e-01 6.67417228e-01
2.46904448e-01 -7.19488204e-01 -7.24705815e-01 -1.03513622e+00
1.38124377e-01 -4.81880009e-01 -2.87588954e-01 -9.83747482e-01
-9.31543767e-01 7.22824335e-01 -1.59150854e-01 -1.35483193e+00
5.08359671e-01 -5.04314482e-01 -1.34733841e-01 8.38244915e-01
-1.49377561e+00 -1.32249749e+00 -7.63964057e-01 1.12849772e+00
4.77340639e-01 8.77738535e-01 1.16975796e+00 -4.21058834e-02
1.29125550e-01 -2.34959722e-01 1.82026282e-01 -1.74452156e-01
5.91968559e-02 -1.10543752e+00 5.15790045e-01 4.04280424e-01
-2.52249837e-01 3.25776458e-01 3.86968464e-01 -9.65440691e-01
-1.97500074e+00 -9.57305074e-01 5.03324509e-01 -6.27078652e-01
-1.08846918e-01 -6.37393713e-01 -5.47190249e-01 7.14101374e-01
-1.25801355e-01 3.31290483e-01 4.00242746e-01 -3.95472080e-01
-3.11535686e-01 -2.78510689e-03 -1.83669698e+00 4.55133557e-01
1.57885456e+00 -6.33153856e-01 -4.45877403e-01 3.42150152e-01
6.24594510e-01 -1.31818473e+00 -1.10947752e+00 3.85787666e-01
7.30086029e-01 -1.33750594e+00 1.33573043e+00 3.28628838e-01
4.56157327e-01 -5.40075004e-01 -4.42968845e-01 -1.08182919e+00
-1.59835324e-01 -8.05495977e-01 -1.02163047e-01 7.54229724e-01
-1.37453511e-01 -4.49136257e-01 9.03680265e-01 8.15159380e-01
-3.49471509e-01 -7.11018503e-01 -1.13945484e+00 -5.62281728e-01
-4.12418813e-01 -7.31973708e-01 1.04589665e+00 8.98880839e-01
-5.49523830e-01 -5.18109441e-01 -1.29567474e-01 3.07124883e-01
1.10790873e+00 5.03989637e-01 1.17299926e+00 -1.33489442e+00
9.71912965e-02 1.19085662e-01 -3.38139743e-01 -1.00955355e+00
-2.18576610e-01 -4.06207830e-01 1.21638544e-01 -1.85324001e+00
-3.77628833e-01 -9.81572926e-01 6.46529913e-01 4.31204855e-01
4.04367656e-01 8.54582012e-01 -2.57697135e-01 1.55085465e-03
-2.94280440e-01 7.19602048e-01 1.69962537e+00 1.42507106e-01
-2.77273953e-01 -1.91363826e-01 -1.23414241e-01 1.22136760e+00
3.32769811e-01 -3.11125129e-01 -3.13638061e-01 -7.49394417e-01
2.33797103e-01 1.16128378e-01 4.55403656e-01 -1.12802303e+00
-2.59989530e-01 -5.79838693e-01 9.28105772e-01 -8.05877566e-01
9.55707967e-01 -1.30015099e+00 9.54147696e-01 2.67371774e-01
4.66578871e-01 1.09736331e-01 3.77005100e-01 3.69184285e-01
3.10747355e-01 4.91112173e-01 7.07517684e-01 -6.38778150e-01
-3.34497422e-01 5.72470427e-01 1.07026391e-01 -2.49909759e-01
9.84878719e-01 -7.93805897e-01 1.50571719e-01 -8.70632827e-02
-7.72145450e-01 -5.19064128e-01 1.66177201e+00 -1.65685013e-01
1.13783479e+00 -1.56191850e+00 -3.96489292e-01 6.73822820e-01
-1.74222723e-01 9.70735133e-01 2.27364004e-01 3.67938951e-02
-1.53570235e+00 1.87831763e-02 -1.04325250e-01 -6.37820840e-01
-1.09121132e+00 2.39288270e-01 3.78960580e-01 1.96821287e-01
-1.45925331e+00 8.74009728e-01 3.81271601e-01 -6.90708399e-01
5.44276647e-03 -8.73438478e-01 8.00332427e-01 -6.14498258e-01
3.79287422e-01 4.12275970e-01 2.52762467e-01 -7.38616824e-01
-4.15528983e-01 1.38392329e+00 9.43041921e-01 -1.56477645e-01
1.59268939e+00 -1.49238259e-01 -3.99077386e-01 6.39380515e-01
7.77401984e-01 6.91144764e-01 -1.54301190e+00 -1.66446909e-01
-6.46202862e-01 -1.35678542e+00 1.73477784e-01 -3.57041061e-01
-1.05777097e+00 6.78678751e-01 1.42920524e-01 -1.44824624e-01
1.17107356e+00 7.31656104e-02 8.20649981e-01 2.49197468e-01
1.57400250e+00 -8.23202312e-01 4.78581935e-02 4.36242402e-01
1.19945443e+00 -5.60745120e-01 5.00717342e-01 -1.21229589e+00
-1.21061094e-01 1.34060788e+00 1.84300497e-01 -4.34866220e-01
8.56571138e-01 4.59604889e-01 -2.04653010e-01 -5.95255077e-01
-1.33676920e-02 8.95620212e-02 1.33275136e-01 9.15366232e-01
2.69401502e-02 8.98526087e-02 4.59911644e-01 -7.28214011e-02
-6.11215472e-01 1.27198026e-01 3.92227173e-01 1.39680004e+00
-2.85211384e-01 -1.23750854e+00 -9.31637287e-01 1.47200763e-01
-3.43434652e-03 1.44411102e-01 -3.26424897e-01 7.70770073e-01
3.16531211e-01 5.65334439e-01 2.11484671e-01 -4.09904242e-01
8.18046153e-01 -2.46950492e-01 1.15267849e+00 -7.22571433e-01
-2.78913289e-01 -1.81336666e-03 4.84576412e-02 -1.24399579e+00
-4.58683223e-01 -5.98055840e-01 -1.32050765e+00 -8.06900024e-01
2.11565718e-01 -3.70109648e-01 7.92289793e-01 5.47426701e-01
3.13591778e-01 3.71335953e-01 8.38362098e-01 -1.56806159e+00
6.58282638e-01 -3.07105958e-01 -7.62772560e-01 3.35833460e-01
3.93978029e-01 -7.46524572e-01 -5.38094118e-02 2.29668602e-01]
|
[9.053975105285645, -3.106006145477295]
|
1c73a750-69d5-4eb5-8275-5e5cba59bb2d
|
cdjur-br-a-golden-collection-of-legal
|
2305.18315
| null |
https://arxiv.org/abs/2305.18315v1
|
https://arxiv.org/pdf/2305.18315v1.pdf
|
CDJUR-BR -- A Golden Collection of Legal Document from Brazilian Justice with Fine-Grained Named Entities
|
A basic task for most Legal Artificial Intelligence (Legal AI) applications is Named Entity Recognition (NER). However, texts produced in the context of legal practice make references to entities that are not trivially recognized by the currently available NERs. There is a lack of categorization of legislation, jurisprudence, evidence, penalties, the roles of people in a legal process (judge, lawyer, victim, defendant, witness), types of locations (crime location, defendant's address), etc. In this sense, there is still a need for a robust golden collection, annotated with fine-grained entities of the legal domain, and which covers various documents of a legal process, such as petitions, inquiries, complaints, decisions and sentences. In this article, we describe the development of the Golden Collection of the Brazilian Judiciary (CDJUR-BR) contemplating a set of fine-grained named entities that have been annotated by experts in legal documents. The creation of CDJUR-BR followed its own methodology that aimed to attribute a character of comprehensiveness and robustness. Together with the CDJUR-BR repository we provided a NER based on the BERT model and trained with the CDJUR-BR, whose results indicated the prevalence of the CDJUR-BR.
|
['Nilsiton Aragão', 'Raquel Silveira', 'André Câmara Ferreira da Costa', 'Francisco das Chagas Jucá Bomfim', 'João Araújo Monteiro Neto', 'Vasco Furtado', 'Vladia Pinheiro', 'Antonio Mauricio']
|
2023-05-20
| null | null | null | null |
['jurisprudence', 'named-entity-recognition-ner']
|
['miscellaneous', 'natural-language-processing']
|
[-3.05202454e-01 2.17454702e-01 -1.20315351e-01 -2.98146993e-01
-7.55374074e-01 -7.94394791e-01 6.60768926e-01 7.64142811e-01
-8.84615541e-01 1.02470720e+00 5.32556474e-01 -8.06295455e-01
-6.21751130e-01 -8.98789048e-01 -3.16547990e-01 -2.25025713e-01
4.81250137e-01 5.81167758e-01 2.19436228e-01 -4.71155018e-01
5.86646259e-01 1.04977131e+00 -9.64563847e-01 2.26152852e-01
1.05168927e+00 3.63912553e-01 -1.97205633e-01 2.57898420e-01
-2.28955522e-01 1.10635221e+00 -9.89263415e-01 -1.00848770e+00
1.04718179e-01 -2.33068153e-01 -1.01046777e+00 -4.98251438e-01
2.52347887e-01 -2.67573118e-01 -4.20814395e-01 9.82241154e-01
3.95961642e-01 7.13164583e-02 8.35658729e-01 -7.00416148e-01
-7.34450221e-01 6.62797868e-01 -1.24527678e-01 4.93354797e-01
5.09709537e-01 -1.54493526e-01 7.68386245e-01 -4.12549436e-01
1.36333549e+00 1.01964903e+00 4.89744455e-01 4.40118551e-01
-6.44469082e-01 -2.95022637e-01 -2.68242300e-01 4.62648064e-01
-1.26644862e+00 -3.17959279e-01 2.88229257e-01 -7.87481666e-01
7.52438307e-01 -4.48500402e-02 8.06812048e-02 9.02607679e-01
2.50632852e-01 1.20196573e-01 8.76296043e-01 -6.46114290e-01
5.37982941e-01 2.73989081e-01 6.38876259e-01 1.72013909e-01
8.07328582e-01 -8.28990489e-02 9.44838524e-02 -6.06951773e-01
4.66570765e-01 -2.95339555e-01 -1.01069182e-01 1.93906784e-01
-6.05807424e-01 7.09730089e-01 -1.30797878e-01 1.14645886e+00
-6.73487902e-01 -6.05499446e-01 7.75448620e-01 8.43735710e-02
-7.27560883e-03 5.01030743e-01 -4.30280954e-01 -4.97042656e-01
-6.86412334e-01 2.50502974e-01 1.11007833e+00 4.41926837e-01
3.29331487e-01 -4.11098361e-01 -3.62536430e-01 7.31155515e-01
3.03915769e-01 1.89910114e-01 3.44197005e-01 -7.33567834e-01
7.49433994e-01 1.06685102e+00 2.77551562e-01 -1.08815408e+00
-2.65180230e-01 2.80480646e-02 -5.37574291e-01 2.11024120e-01
4.94131863e-01 -1.71448410e-01 -7.70111561e-01 1.32856774e+00
4.00784701e-01 -4.19603437e-01 4.28514063e-01 3.74503732e-01
9.73963141e-01 3.01119417e-01 5.09216309e-01 -9.58384573e-02
1.32085800e+00 2.61851717e-02 -9.37247217e-01 3.03771675e-01
6.10951781e-01 -6.38799250e-01 3.12411219e-01 1.24579899e-01
-7.91044533e-01 -1.70910344e-01 -5.51047623e-01 -2.03767821e-01
-8.59362245e-01 1.45432159e-01 4.92040426e-01 8.33957732e-01
-4.47123230e-01 5.22908688e-01 -2.20282987e-01 -4.40692276e-01
4.30204034e-01 -2.69989640e-01 -6.36492729e-01 -1.63217541e-02
-1.42256641e+00 1.37909615e+00 6.10484064e-01 1.56766176e-01
-3.99584085e-01 -3.90604377e-01 -9.39635813e-01 2.77204122e-02
5.66478431e-01 -9.98384580e-02 8.08843017e-01 -5.19285761e-02
-6.69334173e-01 1.33573020e+00 -8.06328505e-02 -6.69345737e-01
6.50836468e-01 8.62628445e-02 -1.22329676e+00 2.51414686e-01
7.66443253e-01 -1.83549792e-01 1.47808308e-03 -8.91470015e-01
-7.10891068e-01 -5.05618095e-01 4.82436806e-01 -5.26180148e-01
9.76534784e-02 7.73852110e-01 -1.37463526e-03 -5.81849873e-01
-3.47511858e-01 -4.64751631e-01 -2.03509808e-01 -6.16297603e-01
-3.38186026e-01 -5.72384238e-01 5.17373502e-01 -8.24101150e-01
1.65583205e+00 -2.18581033e+00 -4.30922747e-01 3.28519076e-01
1.77686408e-01 8.69870782e-01 3.31913024e-01 8.37639391e-01
-7.35458136e-02 6.16596520e-01 -2.82509774e-01 3.93157721e-01
2.13007718e-01 4.44821358e-01 -3.34972322e-01 3.68405759e-01
-1.26833498e-01 6.51767015e-01 -7.61975765e-01 -7.44510174e-01
1.30208626e-01 2.19810396e-01 -1.11528672e-01 -4.22396421e-01
2.10248202e-01 2.03783616e-01 -9.21707034e-01 7.35927165e-01
3.14442903e-01 1.30318403e-01 2.05111653e-01 -8.08052793e-02
-5.76035798e-01 4.02710646e-01 -1.04185820e+00 1.19273162e+00
-2.28971228e-01 5.14373541e-01 1.01522245e-01 -7.95622051e-01
8.92652512e-01 7.45901287e-01 4.31139439e-01 -5.22026777e-01
3.92276257e-01 6.29522264e-01 2.73715029e-03 -8.30423594e-01
6.36361241e-01 -1.65307429e-02 -1.60778508e-01 3.47477913e-01
-4.17833822e-03 2.80425072e-01 1.02255869e+00 2.61364162e-01
1.29676402e+00 4.73843478e-02 8.83001626e-01 5.71144894e-02
8.72067928e-01 2.60187924e-01 7.47527301e-01 5.69680989e-01
-4.21786517e-01 1.68444723e-01 5.81816018e-01 -5.45249879e-01
-1.07497823e+00 -5.30032396e-01 -7.55827248e-01 3.48900527e-01
-3.01765293e-01 -2.27308571e-01 -7.80780435e-01 -8.76364410e-01
-1.05044179e-01 8.81532371e-01 -5.75489223e-01 1.95242569e-01
-5.33596694e-01 -3.92759562e-01 9.00818288e-01 2.63921265e-02
5.33977926e-01 -1.48372793e+00 -9.79232550e-01 4.73509371e-01
4.92443889e-02 -1.30434859e+00 -5.87050766e-02 -1.82751492e-01
-2.86335945e-01 -1.69074953e+00 -3.71495038e-01 -3.31418782e-01
4.75024968e-01 -7.13202596e-01 7.26836443e-01 8.95557106e-02
-4.26073134e-01 1.91098273e-01 -6.80919051e-01 -4.82946038e-01
-9.91587639e-01 -3.62746194e-02 -3.77094418e-01 -9.07489508e-02
6.57822430e-01 -3.38129312e-01 1.76268160e-01 3.26363221e-02
-1.31574309e+00 -9.20244038e-01 4.64352846e-01 2.53322870e-01
1.31754398e-01 -2.06383783e-02 7.94971347e-01 -1.33605731e+00
1.09265089e+00 -5.52031994e-01 -6.15207076e-01 6.69593692e-01
-4.29648727e-01 -1.94923222e-01 5.11973619e-01 1.23263150e-02
-1.37016165e+00 -5.76963007e-01 -6.03553712e-01 1.47901654e-01
-1.15057290e+00 7.74747193e-01 -2.91000187e-01 1.02227047e-01
9.43228185e-01 -2.22206727e-01 -5.83555877e-01 -6.47961318e-01
1.81081027e-01 1.04401529e+00 8.24511290e-01 -7.43859053e-01
5.54053009e-01 1.49232358e-01 -5.79756834e-02 -6.70151651e-01
-9.42557096e-01 -6.52518332e-01 -8.70318294e-01 -1.83724239e-01
1.02524102e+00 -3.73025537e-01 -7.03092515e-01 3.35501432e-02
-1.55180562e+00 2.52545625e-01 -5.11415839e-01 6.93178952e-01
-1.21222641e-02 5.69692552e-01 -5.64376175e-01 -1.09153831e+00
-3.60408455e-01 -8.18107486e-01 4.71451998e-01 2.23944619e-01
-4.04868454e-01 -1.00951660e+00 3.16941202e-01 7.37721205e-01
2.96737939e-01 6.19582951e-01 1.09988654e+00 -1.03911126e+00
1.50420576e-01 -5.93604803e-01 -6.44204617e-02 3.46685827e-01
1.15295753e-01 1.51000157e-01 -4.53015685e-01 4.99827415e-01
-7.48551711e-02 1.81763366e-01 4.57880110e-01 -4.70316447e-02
2.28852108e-01 -4.47679162e-01 -7.75139928e-02 -1.84920788e-01
1.44027853e+00 9.32630479e-01 1.03963411e+00 7.99065650e-01
1.92768529e-01 6.43121779e-01 5.55188417e-01 4.26852167e-01
3.53747755e-01 5.18291414e-01 3.28552797e-02 2.55147308e-01
4.39942777e-02 4.38544340e-02 1.76988930e-01 -2.20375918e-02
-6.86042309e-01 -5.63169569e-02 -1.19059598e+00 7.31627226e-01
-1.61020565e+00 -1.40834773e+00 -4.18342888e-01 1.97993219e+00
6.73073828e-01 1.27510324e-01 -1.04059532e-01 2.58628666e-01
9.18398619e-01 -1.29553542e-01 -2.09232137e-01 -7.27877140e-01
-2.24829406e-01 3.81611437e-01 3.73902202e-01 2.90523142e-01
-9.78427052e-01 8.45395803e-01 5.88823986e+00 9.31934297e-01
-6.59767747e-01 1.30234823e-01 3.37916791e-01 5.69797695e-01
-1.48972645e-01 8.67562369e-02 -9.85503554e-01 5.57451367e-01
9.65275347e-01 -4.73117381e-01 -4.23346758e-02 9.23007667e-01
5.04722595e-01 -1.00919105e-01 -6.47357643e-01 4.85463262e-01
8.84158313e-02 -1.45304561e+00 -4.27713767e-02 3.91442716e-01
5.93860328e-01 -1.99882239e-01 -6.04033172e-01 3.39197874e-01
3.55710864e-01 -6.24723256e-01 8.33336055e-01 5.40097713e-01
5.68558633e-01 -4.76672202e-01 1.25598300e+00 2.94511259e-01
-6.28671408e-01 -1.71591982e-01 -4.11152780e-01 2.02977791e-01
6.90059483e-01 6.89598203e-01 -5.93686342e-01 1.02841568e+00
5.35050035e-01 3.40345711e-01 -3.75058860e-01 1.14453197e+00
-5.08876503e-01 5.40304244e-01 1.24509610e-01 2.90420558e-02
3.38162512e-01 -3.70258778e-01 5.72173774e-01 1.07036769e+00
1.46510258e-01 4.09424961e-01 -1.36351541e-01 7.69802034e-01
-1.69294044e-01 5.77999175e-01 -5.68270326e-01 1.04800127e-02
6.70718133e-01 1.14295781e+00 -4.79977161e-01 -5.20540178e-01
-1.81410685e-01 3.33910197e-01 1.10552078e-02 2.98285753e-01
-5.35904884e-01 -4.67014343e-01 2.56227970e-01 3.15738559e-01
1.02311239e-01 1.91340402e-01 1.57977212e-02 -9.22471106e-01
1.10096611e-01 -8.56229901e-01 6.84429348e-01 -7.11367905e-01
-1.23742223e+00 7.98190117e-01 3.86421770e-01 -8.06323946e-01
-2.82561749e-01 -5.74620485e-01 -4.96223956e-01 7.37999082e-01
-1.16715407e+00 -9.89482164e-01 2.76984721e-01 4.88451093e-01
-1.87598299e-02 -4.09518719e-01 9.41749930e-01 7.88938403e-01
-4.62844729e-01 -2.18256786e-02 -1.56382740e-01 6.47189438e-01
5.37537098e-01 -1.01864946e+00 2.30306275e-02 8.17106426e-01
9.26921144e-04 9.43562806e-01 6.36752903e-01 -9.17402446e-01
-2.67886132e-01 -6.57917202e-01 1.64790893e+00 -3.72814417e-01
1.00368249e+00 3.32678795e-01 -8.27381790e-01 6.74557030e-01
1.12714916e-01 -3.98516595e-01 7.59666800e-01 2.23942790e-02
-4.00790781e-01 1.15959182e-01 -1.64387810e+00 4.94299024e-01
7.95498550e-01 -6.21699095e-01 -1.27062309e+00 3.46447319e-01
4.72539365e-02 -1.89520940e-01 -1.19418764e+00 -5.79145700e-02
2.87712753e-01 -7.37261772e-01 5.83012044e-01 -8.57490420e-01
1.29771456e-01 -2.38178000e-01 1.52505443e-01 -6.74010575e-01
1.26101732e-01 -3.13295305e-01 5.18997192e-01 1.62812722e+00
7.47921586e-01 -9.91253555e-01 2.24605471e-01 1.24404013e+00
-2.93694854e-01 -2.15950206e-01 -1.42031109e+00 -6.44342661e-01
4.77949530e-02 -5.42222321e-01 6.26245975e-01 1.25691152e+00
2.44620278e-01 2.04798877e-01 6.10139861e-04 -4.65044705e-03
3.05628926e-01 -7.39466548e-02 4.19364572e-01 -1.57565951e+00
1.81072235e-01 -3.63977790e-01 -7.59621799e-01 -5.01172692e-02
1.85444295e-01 -5.19085944e-01 -4.79398251e-01 -2.00786591e+00
4.48745023e-03 -3.29858005e-01 1.11091144e-01 8.23875546e-01
9.86307040e-02 -4.45821822e-01 2.64301836e-01 3.98931324e-01
-4.05851036e-01 -1.25783801e-01 9.30223286e-01 -2.32152548e-02
-2.90561896e-02 -7.61345252e-02 -7.12946415e-01 9.04720962e-01
7.19309449e-01 -7.48737752e-01 3.33257973e-01 -1.44732101e-02
4.83560234e-01 -6.88732713e-02 1.19271867e-01 -8.30377281e-01
5.33350110e-01 -2.13638753e-01 -1.51544034e-01 -5.50033629e-01
-8.48645046e-02 -9.70280647e-01 3.66336435e-01 2.71122515e-01
-2.88605809e-01 -1.80602431e-01 -9.46126059e-02 1.86476126e-01
-6.09039843e-01 -1.07312250e+00 4.75888908e-01 -2.48868942e-01
-5.64696014e-01 -6.47286251e-02 -6.58238292e-01 1.23300411e-01
1.33142447e+00 -3.08099478e-01 -7.19192326e-01 1.80193353e-02
-8.17368448e-01 -1.25251159e-01 3.45599562e-01 9.48853865e-02
2.29714572e-01 -8.20182800e-01 -9.25009549e-01 -4.81919199e-01
4.33364697e-02 -2.94599712e-01 4.42152679e-01 6.22648239e-01
-6.40452206e-01 7.16893375e-01 -1.91987306e-01 5.06178439e-01
-1.12272990e+00 5.86607873e-01 1.64355665e-01 -7.04219759e-01
-4.60423887e-01 -1.32711872e-01 -5.13620555e-01 -2.72828639e-01
-2.15253294e-01 -1.31691158e-01 -1.16057026e+00 3.30946594e-01
4.07973379e-01 5.77464223e-01 2.12036729e-01 -1.24993324e+00
-3.76967996e-01 5.34224689e-01 -1.21284025e-02 -6.41389564e-02
1.50028384e+00 2.73294926e-01 -5.45794547e-01 1.64575920e-01
7.55437016e-01 5.45662820e-01 -7.41223469e-02 2.09268764e-01
5.60892284e-01 -3.25134158e-01 -3.51407737e-01 -1.04873836e+00
-8.51941705e-01 3.35059524e-01 1.34113997e-01 5.71474314e-01
7.67251551e-01 1.88119426e-01 3.22206885e-01 3.94350380e-01
6.60008371e-01 -1.19245434e+00 -9.81995285e-01 5.77395737e-01
1.00737727e+00 -8.14860344e-01 -7.32938573e-02 -3.96458536e-01
-7.03495800e-01 1.17414856e+00 -1.37702823e-01 1.60866946e-01
6.04420781e-01 8.77804682e-02 3.06229204e-01 -5.99861979e-01
-1.29479483e-01 -3.27403665e-01 2.26640493e-01 4.42155570e-01
3.18786502e-01 -1.70952007e-01 -1.52500439e+00 6.15799308e-01
-1.57673359e-02 3.36043268e-01 6.82984233e-01 9.95312929e-01
-3.91233534e-01 -1.37382424e+00 -7.13061035e-01 2.45631829e-01
-1.13958573e+00 -2.11282849e-01 -6.58032894e-01 1.10698044e+00
7.29943991e-01 1.23069167e+00 -3.82787883e-01 3.43188107e-01
7.96333551e-01 -9.09628868e-02 -1.16880745e-01 -7.52903402e-01
-1.06156909e+00 -3.42035651e-01 7.76036024e-01 -1.82939991e-01
-6.42477572e-01 -5.58600783e-01 -1.36679482e+00 -1.24282777e-01
-1.21507391e-01 8.14182580e-01 8.45039189e-01 1.27919316e+00
1.95833147e-01 2.82671630e-01 -3.12925316e-02 -7.74361473e-03
-2.01917753e-01 -9.92550552e-01 -8.96935165e-01 6.31720722e-01
-2.08832875e-01 -3.68930638e-01 -3.35476995e-01 4.05633114e-02]
|
[9.704315185546875, 9.243600845336914]
|
b76f733b-5d3d-48fa-81cf-0282e924d65a
|
pro-uigan-progressive-face-hallucination-from
|
2108.00602
| null |
https://arxiv.org/abs/2108.00602v6
|
https://arxiv.org/pdf/2108.00602v6.pdf
|
Pro-UIGAN: Progressive Face Hallucination from Occluded Thumbnails
|
In this paper, we study the task of hallucinating an authentic high-resolution (HR) face from an occluded thumbnail. We propose a multi-stage Progressive Upsampling and Inpainting Generative Adversarial Network, dubbed Pro-UIGAN, which exploits facial geometry priors to replenish and upsample (8*) the occluded and tiny faces (16*16 pixels). Pro-UIGAN iteratively (1) estimates facial geometry priors for low-resolution (LR) faces and (2) acquires non-occluded HR face images under the guidance of the estimated priors. Our multi-stage hallucination network super-resolves and inpaints occluded LR faces in a coarse-to-fine manner, thus reducing unwanted blurriness and artifacts significantly. Specifically, we design a novel cross-modal transformer module for facial priors estimation, in which an input face and its landmark features are formulated as queries and keys, respectively. Such a design encourages joint feature learning across the input facial and landmark features, and deep feature correspondences will be discovered by attention. Thus, facial appearance features and facial geometry priors are learned in a mutual promotion manner. Extensive experiments demonstrate that our Pro-UIGAN achieves visually pleasing HR faces, reaching superior performance in downstream tasks, i.e., face alignment, face parsing, face recognition and expression classification, compared with other state-of-the-art (SotA) methods.
|
['Ping Liu', 'Xiaobo Lu', 'Xin Yu', 'Yang Zhang']
|
2021-08-02
| null | null | null | null |
['face-alignment', 'face-parsing', 'face-hallucination']
|
['computer-vision', 'computer-vision', 'computer-vision']
|
[ 4.03445750e-01 3.90941232e-01 1.05285816e-01 -5.32063007e-01
-1.04412782e+00 -4.44008708e-01 2.45822594e-01 -1.06195199e+00
1.62819341e-01 7.10577369e-01 2.81741977e-01 3.15656960e-01
3.16687614e-01 -7.59636581e-01 -1.03895152e+00 -6.83423698e-01
3.47276151e-01 2.70774066e-01 -5.88089347e-01 2.69036852e-02
-1.80856273e-01 9.37943459e-01 -1.67657876e+00 3.93838555e-01
6.87000811e-01 1.05078399e+00 2.11171787e-02 4.29482728e-01
1.60006478e-01 8.06511998e-01 -3.99247795e-01 -6.22738957e-01
4.84006792e-01 -4.87130940e-01 -5.90251744e-01 5.18315196e-01
8.94355297e-01 -8.82100701e-01 -6.20957732e-01 1.01536739e+00
4.53670323e-01 -7.19937757e-02 4.37584668e-01 -1.13954270e+00
-1.01703656e+00 1.43552199e-01 -1.32446015e+00 -2.56840557e-01
5.73729634e-01 3.31112981e-01 5.64068198e-01 -1.42699277e+00
7.32841015e-01 1.82473350e+00 5.63988388e-01 9.40467596e-01
-1.41453719e+00 -1.10514140e+00 -7.53028132e-03 -2.24779516e-01
-1.67086756e+00 -9.96638238e-01 1.08402157e+00 -1.17350951e-01
2.96757936e-01 4.12873030e-01 4.16778326e-01 1.07983863e+00
-1.23806365e-01 6.07755184e-01 1.13360441e+00 -1.53456017e-01
-1.14346981e-01 -2.17769787e-01 -6.82326972e-01 1.03630555e+00
-2.57978201e-01 1.76244855e-01 -5.26696801e-01 -2.20661283e-01
1.33240676e+00 2.28788093e-01 -4.88765895e-01 2.10006312e-02
-5.85960209e-01 6.29431009e-01 2.64234602e-01 -2.80748401e-02
-4.84921575e-01 5.17169870e-02 -1.03872687e-01 2.41413489e-01
7.16405332e-01 1.98078111e-01 -3.96368504e-01 4.22489554e-01
-9.72227335e-01 1.22225456e-01 3.80306154e-01 1.02697587e+00
1.06295693e+00 3.03635269e-01 -2.82927364e-01 1.11859465e+00
3.27970058e-01 6.85989618e-01 -9.68933925e-02 -1.44917476e+00
2.83192396e-01 2.04116344e-01 6.63609356e-02 -9.99117732e-01
9.66711640e-02 4.60503660e-02 -1.06472993e+00 2.71633327e-01
1.93782672e-01 -2.57017583e-01 -1.18622077e+00 2.11186647e+00
6.35020912e-01 3.00048411e-01 5.29151522e-02 1.02314806e+00
9.36432302e-01 7.45730877e-01 7.02768415e-02 -5.90649962e-01
1.52471340e+00 -7.08829045e-01 -6.66347623e-01 -2.59562284e-01
-4.56332862e-01 -1.00712323e+00 9.05099988e-01 2.96398252e-01
-1.43581116e+00 -8.00827980e-01 -5.31798840e-01 -5.06003976e-01
2.78656274e-01 3.36345524e-01 5.29502153e-01 3.96620184e-01
-1.12962353e+00 5.15301585e-01 -4.65737551e-01 1.34853199e-01
1.07111204e+00 4.81761843e-01 -6.97425663e-01 -4.15600806e-01
-8.82879615e-01 4.22115713e-01 -3.44933450e-01 3.36859494e-01
-1.19647682e+00 -1.03763175e+00 -1.18086374e+00 2.60144398e-02
2.32726768e-01 -7.51716256e-01 9.13352787e-01 -1.21323574e+00
-1.82568848e+00 1.11235452e+00 -4.44752306e-01 2.13563070e-01
4.97522295e-01 -2.41006732e-01 -3.38445306e-01 2.61671722e-01
1.28479123e-01 1.11050069e+00 1.70350814e+00 -1.46088231e+00
-2.00916901e-01 -6.38857424e-01 -2.12093368e-01 1.99448943e-01
-5.99027090e-02 2.22375333e-01 -6.50567234e-01 -8.98380280e-01
1.07591078e-01 -5.48362494e-01 7.12104142e-02 4.44114149e-01
-3.61790627e-01 1.15862183e-01 1.09329605e+00 -1.10784996e+00
4.88795936e-01 -2.34084678e+00 2.89189726e-01 2.92225741e-02
3.35872471e-01 5.53291924e-02 -4.39216048e-01 -1.96691573e-01
-3.91760260e-01 -9.17894021e-02 -3.22883308e-01 -7.61195660e-01
-2.74279147e-01 1.06215745e-01 -7.08301365e-01 6.51344538e-01
5.26114047e-01 1.02612519e+00 -6.27337039e-01 -4.99732345e-01
2.18616322e-01 1.07564008e+00 -5.15522838e-01 6.40815258e-01
-2.30232343e-01 8.34155560e-01 -3.29977095e-01 1.02316427e+00
1.33037031e+00 -8.98219179e-03 1.14424199e-01 -6.01749599e-01
6.20578639e-02 -2.53746301e-01 -8.04463446e-01 1.68400085e+00
-5.61537325e-01 4.29896295e-01 4.63546187e-01 -3.32181036e-01
1.12977731e+00 2.76744336e-01 4.42340821e-01 -5.92160583e-01
5.70161752e-02 -1.27668679e-01 -5.52643061e-01 -4.58397865e-01
1.43471152e-01 -2.08478093e-01 2.25722790e-01 2.89300323e-01
1.09338194e-01 -2.12416202e-01 -4.33698684e-01 -9.32299271e-02
7.04190612e-01 3.80033493e-01 1.36968553e-01 1.21408395e-01
6.60664618e-01 -8.46411109e-01 7.13684797e-01 -3.50354798e-02
9.85746607e-02 1.05693769e+00 5.08660376e-01 -3.50422919e-01
-1.07431638e+00 -1.16447675e+00 -4.37703282e-02 1.03384578e+00
-7.87557438e-02 -5.49247935e-02 -1.00197220e+00 -5.50157785e-01
-3.42789590e-02 3.34774137e-01 -9.25522268e-01 9.02672336e-02
-7.12880373e-01 -3.62103075e-01 5.10242939e-01 4.47258890e-01
6.33959949e-01 -1.27830791e+00 -1.16752639e-01 -6.21862337e-02
-1.06277533e-01 -1.05681694e+00 -1.02114737e+00 -4.74121422e-01
-5.61492860e-01 -9.07486081e-01 -8.36014092e-01 -8.82413208e-01
1.10582817e+00 2.02674419e-01 9.07283187e-01 -3.60153653e-02
-4.14295107e-01 1.20963760e-01 1.32987589e-01 1.38264462e-01
-2.05089554e-01 -5.27565479e-01 3.84309515e-02 5.76640666e-01
-9.55960825e-02 -8.99601579e-01 -7.50897765e-01 3.13845932e-01
-9.07842159e-01 2.32135430e-01 8.51945996e-01 9.31566834e-01
9.18647945e-01 -1.87686712e-01 4.04624254e-01 -9.33448553e-01
4.85906154e-02 -2.94359982e-01 -7.04373419e-01 2.24802971e-01
-1.27668336e-01 -1.05113521e-01 7.94981241e-01 -5.85464895e-01
-1.42323351e+00 3.03683281e-01 -2.51593411e-01 -1.17088318e+00
-1.22942947e-01 -3.54307145e-01 -9.80647147e-01 -3.60676259e-01
2.33316123e-01 2.82503009e-01 2.24743471e-01 -4.33874726e-01
6.72903657e-01 4.67486143e-01 8.30510318e-01 -6.95493519e-01
1.06636000e+00 6.78947210e-01 -4.46943417e-02 -7.54361570e-01
-8.63699019e-01 1.59895241e-01 -4.01815832e-01 -1.37279242e-01
7.67390311e-01 -1.19104862e+00 -8.61994803e-01 6.77722216e-01
-1.29489219e+00 -2.26882666e-01 -3.27474385e-01 -1.41418397e-01
-7.12022901e-01 2.15323970e-01 -7.99791753e-01 -7.43311346e-01
-5.22859931e-01 -1.06646097e+00 1.56790209e+00 5.55644810e-01
-7.73927663e-03 -3.29347998e-01 -3.72357398e-01 5.21915853e-01
1.96856722e-01 4.48882222e-01 5.42565107e-01 2.92342991e-01
-7.92328119e-01 3.54670227e-01 -5.54475665e-01 3.17297012e-01
3.61846089e-01 -3.07120625e-02 -1.33800805e+00 -3.33529919e-01
6.84869960e-02 -5.06744683e-01 5.44536710e-01 2.82134950e-01
1.62982011e+00 -6.61901414e-01 -1.35333702e-01 1.19401491e+00
1.14172697e+00 -2.18634419e-02 1.05601180e+00 -4.72644240e-01
9.57490087e-01 7.98395932e-01 6.31751955e-01 5.12216389e-01
1.45014420e-01 5.88980377e-01 4.38230157e-01 -3.96814615e-01
-3.23295921e-01 -5.76891661e-01 3.62913400e-01 1.41930521e-01
-2.20254203e-03 2.21928746e-01 -7.91895241e-02 1.85136765e-01
-1.28744555e+00 -9.67122555e-01 4.59968835e-01 1.77222013e+00
1.23699713e+00 -6.88939929e-01 -1.41108915e-01 -3.15554112e-01
8.78583729e-01 3.81103933e-01 -9.14761484e-01 -2.66918652e-02
-1.42999932e-01 7.22061157e-01 1.22204013e-01 5.37467062e-01
-8.39422464e-01 1.24575949e+00 5.29600382e+00 9.24553394e-01
-1.03871131e+00 2.56358415e-01 1.38042223e+00 -4.10969816e-02
-4.67605740e-01 -2.99597174e-01 -6.16912723e-01 3.68435323e-01
3.41654927e-01 2.81863630e-01 9.40214515e-01 9.39234436e-01
1.44018698e-02 3.26458544e-01 -9.39634442e-01 1.27927887e+00
3.59648198e-01 -1.32490897e+00 1.77630961e-01 6.86088949e-02
7.44826913e-01 -5.79138100e-01 5.05113244e-01 4.46036719e-02
2.29553387e-01 -1.39242172e+00 6.73393071e-01 7.60220706e-01
1.81279385e+00 -1.12392282e+00 9.75945368e-02 -2.33365759e-01
-1.17733824e+00 -5.23691997e-03 -3.48185927e-01 4.67442691e-01
1.06310971e-01 3.89828324e-01 -3.08228165e-01 4.08081681e-01
8.34858298e-01 5.80042899e-01 -1.13967933e-01 9.24254060e-02
-4.15006816e-01 1.09541647e-01 -1.48364857e-01 8.04517150e-01
-3.87600183e-01 -2.85067499e-01 3.10627103e-01 7.45023847e-01
3.45551968e-01 7.32745528e-01 -1.88096642e-01 1.42012179e+00
-6.88574493e-01 -2.24565491e-01 -4.86507386e-01 1.84993774e-01
5.62536418e-01 1.64464700e+00 -2.52719909e-01 -8.90159383e-02
-2.90939212e-01 1.39081442e+00 4.19287175e-01 4.19066012e-01
-9.53350306e-01 -1.29395589e-01 1.15514576e+00 1.66100159e-01
4.08065557e-01 2.65184700e-01 -5.93944406e-03 -1.00545549e+00
1.06614046e-01 -1.03892231e+00 1.06669776e-02 -1.05015934e+00
-1.33430207e+00 7.37228692e-01 -3.96751910e-01 -9.59934711e-01
-1.77741662e-01 -2.47252181e-01 -5.11995971e-01 1.04522657e+00
-1.54477072e+00 -1.61183667e+00 -5.21107078e-01 9.33149040e-01
5.64607263e-01 -1.18876107e-01 6.55451715e-01 2.97930449e-01
-6.04845285e-01 9.77716386e-01 -4.71434504e-01 3.21987510e-01
8.37997615e-01 -7.54142523e-01 3.64520460e-01 7.32226789e-01
-4.46875915e-02 5.73411524e-01 4.96322900e-01 -5.82618117e-01
-1.68219650e+00 -1.63316000e+00 3.71985435e-01 -1.74488172e-01
-1.07651586e-02 -3.66518676e-01 -9.70073521e-01 7.12399662e-01
-5.40929586e-02 5.68914354e-01 4.28306818e-01 -4.86914694e-01
-6.31202221e-01 -4.03119117e-01 -1.69397020e+00 6.51747167e-01
1.24785960e+00 -6.86576009e-01 -2.15263560e-01 2.19211444e-01
8.08340371e-01 -6.13153994e-01 -9.29857552e-01 4.49054807e-01
7.31725037e-01 -1.00805581e+00 1.19978511e+00 -3.17179143e-01
5.83248377e-01 -3.59372199e-01 -1.54725984e-01 -9.72973228e-01
-2.31108829e-01 -1.28476620e+00 -2.40104511e-01 1.48508453e+00
-1.28416330e-01 -4.77694243e-01 8.14742744e-01 5.21811903e-01
1.13267973e-02 -8.48378599e-01 -7.59351909e-01 -1.48997307e-01
-2.35333040e-01 2.61089448e-02 8.91835809e-01 8.79660904e-01
-6.75805986e-01 1.42516017e-01 -7.53125608e-01 4.49102044e-01
8.76010001e-01 2.85851538e-01 8.80166113e-01 -8.21747482e-01
-2.20461324e-01 -3.28746773e-02 -3.94952446e-02 -1.08942842e+00
6.04405344e-01 -4.47300255e-01 7.25683523e-03 -7.22569585e-01
1.99933603e-01 -2.71846622e-01 3.31381887e-01 6.49410367e-01
-1.52379319e-01 7.42539525e-01 5.27253561e-03 1.25083372e-01
-1.52523041e-01 7.25377858e-01 1.62369370e+00 1.49679072e-02
-8.40115100e-02 -2.53092319e-01 -1.08636069e+00 6.97714627e-01
3.73300642e-01 -1.91171765e-01 -2.10217714e-01 -3.16723168e-01
-2.72918046e-01 5.04807889e-01 5.44869065e-01 -6.16666079e-01
-1.05939209e-01 -2.97254682e-01 1.21704102e+00 -3.87258351e-01
7.72468805e-01 -5.63560307e-01 5.55635750e-01 -9.10204872e-02
-1.34761065e-01 -2.19550490e-01 1.67864859e-01 4.27282482e-01
-1.44098446e-01 4.24819618e-01 1.22882044e+00 -8.17927718e-02
-3.69032621e-01 1.01859796e+00 2.57013410e-01 -1.04249761e-01
9.90286231e-01 -1.55311212e-01 -1.53388798e-01 -4.57509756e-01
-6.64563298e-01 -6.38210550e-02 7.00184703e-01 3.52962255e-01
1.08905292e+00 -1.30780530e+00 -8.34412456e-01 8.21817219e-01
-2.24960983e-01 3.77709031e-01 5.82598269e-01 5.06235719e-01
-3.23468804e-01 -1.32904232e-01 -4.98842955e-01 -4.00624543e-01
-1.33474982e+00 5.77250719e-01 3.53784055e-01 1.01708293e-01
-6.65283918e-01 1.12021005e+00 7.09940434e-01 -2.74250776e-01
3.30646224e-02 2.87118316e-01 -3.48274335e-02 -7.95800015e-02
8.74655485e-01 4.58238199e-02 -4.34616953e-01 -8.42308283e-01
-1.54031023e-01 8.76421094e-01 -3.10322613e-01 -1.30752429e-01
1.45204866e+00 -2.23424524e-01 -4.96345162e-01 -2.76100904e-01
1.36920798e+00 3.64384741e-01 -1.94474125e+00 -1.79771394e-01
-8.19093406e-01 -8.97084177e-01 -1.96860000e-01 -4.44903255e-01
-1.75673580e+00 5.93882501e-01 4.03612822e-01 -5.94892800e-01
1.56329668e+00 2.15676963e-01 8.94329548e-01 -3.47739995e-01
1.87699914e-01 -4.89431977e-01 3.06079209e-01 7.40221292e-02
1.24426568e+00 -9.42896962e-01 -5.65365814e-02 -7.88947582e-01
-5.89910924e-01 1.12965989e+00 9.52450037e-01 -2.30140522e-01
4.76101041e-01 4.44495201e-01 -9.42007154e-02 -5.36395796e-02
-5.41983008e-01 7.54118478e-03 2.45126978e-01 6.82816505e-01
1.47377461e-01 -8.29574093e-02 3.18573982e-01 6.05748117e-01
-2.54302025e-01 -3.92499082e-02 8.36091340e-02 3.44045311e-01
-4.41571325e-02 -8.57279778e-01 -6.74244225e-01 1.23521000e-01
-4.22675550e-01 -1.48147076e-01 -1.62393793e-01 6.60654962e-01
4.14407670e-01 5.93226194e-01 1.47089064e-01 -3.06130469e-01
1.07134856e-01 -2.73069352e-01 7.94595420e-01 -5.34836650e-01
-1.11124367e-01 2.41715655e-01 -3.14469635e-01 -9.24619913e-01
-9.75441486e-02 -5.75840056e-01 -1.09049225e+00 -4.14615929e-01
2.13168152e-02 -1.57647252e-01 2.23093227e-01 7.22474277e-01
6.20219052e-01 4.36570883e-01 1.06341994e+00 -1.22951615e+00
-3.58231127e-01 -9.56303895e-01 -5.51729083e-01 3.40126306e-01
5.19434214e-01 -6.58576369e-01 -3.48089755e-01 2.92368591e-01]
|
[12.783344268798828, -0.11673842370510101]
|
dd3e431c-e623-492e-a364-42785ec53b6a
|
guided-adaptive-credit-assignment-for-sample
| null | null |
https://openreview.net/forum?id=SyxBgkBFPS
|
https://openreview.net/pdf?id=SyxBgkBFPS
|
Guided Adaptive Credit Assignment for Sample Efficient Policy Optimization
|
Policy gradient methods have achieved remarkable successes in solving challenging reinforcement learning problems. However, it still often suffers from sparse reward tasks, which leads to poor sample efficiency during training. In this work, we propose a guided adaptive credit assignment method to do effectively credit assignment for policy gradient methods. Motivated by entropy regularized policy optimization, our method extends the previous credit assignment methods by introducing more general guided adaptive credit assignment(GACA). The benefit of GACA is a principled way of utilizing off-policy samples. The effectiveness of proposed algorithm is demonstrated on the challenging \textsc{WikiTableQuestions} and \textsc{WikiSQL} benchmarks and an instruction following environment. The task is generating action sequences or program sequences from natural language questions or instructions, where only final binary success-failure execution feedback is available. Empirical studies show that our method significantly improves the sample efficiency of the state-of-the-art policy optimization approaches.
|
['Caiming Xiong', 'Richard Socher', 'Hao liu']
|
2019-09-25
| null | null | null | null |
['policy-gradient-methods']
|
['methodology']
|
[ 3.13653558e-01 -1.24498367e-01 -9.42198575e-01 -3.41300040e-01
-7.46219456e-01 -3.22508961e-01 5.09310186e-01 1.81243196e-03
-7.74742365e-01 1.35885191e+00 1.41878769e-01 -7.84591913e-01
-1.06616460e-01 -5.00658989e-01 -7.34745681e-01 -5.50908029e-01
1.75476387e-01 3.58623177e-01 1.13291137e-01 -1.78508043e-01
8.86423111e-01 1.11357421e-01 -1.61669838e+00 -4.29592133e-02
1.41450202e+00 1.05599892e+00 5.19905508e-01 5.85625112e-01
-3.86523098e-01 1.12800169e+00 -4.84735698e-01 3.85691710e-05
8.19761530e-02 -5.57628572e-01 -7.83267975e-01 -3.42882544e-01
1.52359471e-01 -5.76320469e-01 8.39323401e-02 1.18793595e+00
4.80159193e-01 6.38881445e-01 6.80252373e-01 -1.30978417e+00
-5.38692713e-01 7.14867353e-01 -5.38782895e-01 3.43091875e-01
4.01321024e-01 3.66195291e-01 1.04329407e+00 -3.86408687e-01
4.22927290e-01 1.20368159e+00 4.44817990e-02 9.29214180e-01
-8.35018873e-01 -6.70520484e-01 5.29450834e-01 4.39433485e-01
-5.26578486e-01 -2.27838293e-01 6.08520269e-01 -2.97844529e-01
1.33002794e+00 3.71948183e-02 5.43768108e-01 1.11494267e+00
5.57764173e-02 1.41009331e+00 1.49281788e+00 -5.58166981e-01
7.00899780e-01 1.19281150e-01 -1.01203015e-02 8.90196145e-01
-9.89257470e-02 6.44262433e-01 -6.04327679e-01 -4.23880696e-01
3.52920800e-01 1.43464783e-03 -2.85736710e-01 -3.55651706e-01
-1.01778972e+00 1.01557398e+00 -9.83544365e-02 -1.30805805e-01
-4.14067864e-01 4.33797985e-01 5.21346509e-01 3.93974215e-01
2.18142629e-01 6.06141627e-01 -6.56816900e-01 -9.37796831e-01
-8.24490547e-01 4.72968191e-01 8.55633914e-01 1.14547694e+00
5.82609951e-01 4.18886870e-01 -6.44250572e-01 7.09245384e-01
2.16578126e-01 6.28246844e-01 1.04504061e+00 -1.15951622e+00
7.02558696e-01 3.15341026e-01 5.48949420e-01 -3.94973040e-01
-8.28902051e-02 -9.01865482e-04 -3.27829987e-01 5.41513860e-01
2.32450172e-01 -3.69590521e-01 -7.79190242e-01 1.76771271e+00
2.95531601e-01 5.21166734e-02 1.01755500e-01 8.03408384e-01
1.83210969e-01 6.71682894e-01 3.14203531e-01 -5.32093108e-01
9.39763069e-01 -1.21183646e+00 -7.99246490e-01 -2.83773214e-01
4.49179381e-01 -5.67098916e-01 1.74790847e+00 5.20743608e-01
-9.71256435e-01 -2.70396441e-01 -1.04781127e+00 4.02042389e-01
-7.63536245e-02 1.06741108e-01 8.66762459e-01 7.08207309e-01
-5.99077642e-01 9.84709501e-01 -8.89413357e-01 1.19165324e-01
5.03957927e-01 3.12139153e-01 2.79571980e-01 2.35412121e-01
-1.06718886e+00 7.51799405e-01 4.66849059e-01 -4.83024031e-01
-1.20344746e+00 -6.00693643e-01 -6.64303839e-01 2.86557436e-01
9.30897295e-01 -1.37585357e-01 1.78931010e+00 -7.79812336e-01
-2.51939201e+00 2.37484023e-01 6.96046697e-03 -5.43123901e-01
5.88691056e-01 -6.48866832e-01 8.32678005e-02 -1.19443029e-01
-1.76404074e-01 4.84408379e-01 1.16668510e+00 -8.11324418e-01
-1.07266009e+00 -7.52704069e-02 4.52406742e-02 4.74050432e-01
-4.83177513e-01 1.77806746e-02 1.82208996e-02 -5.31974852e-01
-7.74434328e-01 -8.94756496e-01 -4.31416512e-01 -5.79732001e-01
-1.87263429e-01 -6.11403584e-01 6.53527677e-01 -3.27981770e-01
1.42415571e+00 -1.75128400e+00 4.08897810e-02 1.16033060e-02
-2.19035700e-01 5.25988400e-01 -7.42926076e-02 1.85187921e-01
2.99122930e-01 1.08354174e-01 -2.70794362e-01 -1.65454164e-01
4.93669271e-01 2.68643230e-01 -5.72989762e-01 1.26449212e-01
-2.90114488e-02 7.13878691e-01 -1.27718341e+00 -5.82120478e-01
1.34599745e-01 -2.74888068e-01 -8.57700169e-01 4.96021032e-01
-8.35389316e-01 3.08482945e-01 -8.24535787e-01 6.95772350e-01
1.95770577e-01 -1.29294738e-01 7.73143247e-02 7.18971729e-01
-1.03104196e-01 3.92450064e-01 -1.01819706e+00 1.72737586e+00
-5.53370416e-01 2.73279995e-01 -1.73923656e-01 -1.14429533e+00
9.05489206e-01 1.78151131e-01 3.91587585e-01 -7.48466969e-01
1.38989761e-02 2.71584243e-01 -1.01421855e-01 -5.30837774e-01
7.19744980e-01 1.11108027e-01 -3.90971266e-02 7.61627018e-01
9.23633948e-03 -3.38938236e-01 4.14106131e-01 7.23339245e-02
9.70300376e-01 7.12687314e-01 5.48455656e-01 -2.65434325e-01
6.48562014e-01 3.46756019e-02 7.80574143e-01 1.08956707e+00
-6.55053318e-01 -1.49237677e-01 6.63155496e-01 -1.67777017e-01
-8.04168224e-01 -6.44460380e-01 3.28157753e-01 1.59378016e+00
-1.68309882e-01 -2.54025191e-01 -8.44786644e-01 -1.26657999e+00
2.43812665e-01 1.15811932e+00 -5.25984824e-01 -2.98366815e-01
-6.65146828e-01 -7.29481936e-01 2.81634837e-01 4.98969406e-01
5.00434518e-01 -1.71433485e+00 -9.27885234e-01 4.44939703e-01
-6.76042214e-02 -7.36799479e-01 -9.68572319e-01 2.26778641e-01
-1.04673326e+00 -9.52442467e-01 -5.86198509e-01 -5.58715403e-01
4.29135948e-01 -1.41897470e-01 1.27248442e+00 -5.96265979e-02
-1.19907565e-01 3.41691345e-01 -4.01445836e-01 -6.86789691e-01
-5.30589163e-01 2.64963031e-01 2.91792244e-01 -4.59074587e-01
4.82747555e-01 -3.45696360e-01 -7.28173971e-01 2.81193197e-01
-6.55614078e-01 -1.71805188e-01 5.50228894e-01 1.23822773e+00
7.53422499e-01 -4.96265888e-01 1.05724895e+00 -1.10705531e+00
1.24986899e+00 -3.53103518e-01 -1.14879632e+00 3.33371252e-01
-1.51224959e+00 6.36645854e-01 1.14505184e+00 -6.36148751e-01
-1.24838817e+00 1.08725578e-03 1.23728432e-01 -2.76836514e-01
-7.07668588e-02 3.06774467e-01 2.31043383e-01 1.08922340e-01
6.85020030e-01 4.52273875e-01 1.26786068e-01 -3.59225720e-01
2.20992923e-01 6.75905406e-01 2.64222175e-01 -1.21279812e+00
2.84311920e-01 -5.93380891e-02 -1.67489663e-01 -2.05406532e-01
-7.49931335e-01 -3.36902380e-01 1.29250824e-01 -3.50629948e-02
5.60916901e-01 -3.57432514e-01 -1.21189225e+00 2.16206789e-01
-5.42883277e-01 -8.92193794e-01 -5.02729177e-01 6.28870487e-01
-1.26700580e+00 2.94946045e-01 -5.48288047e-01 -1.04866123e+00
-6.53297007e-01 -1.42366028e+00 7.29836881e-01 6.24774158e-01
1.07052594e-01 -8.75860214e-01 1.69113711e-01 7.86597431e-02
5.61868131e-01 -7.84340724e-02 1.11855412e+00 -5.87408781e-01
-4.28738981e-01 8.30643848e-02 2.88866311e-01 5.46843708e-01
9.25245881e-02 -2.38080785e-01 -4.81100708e-01 -4.23568577e-01
-1.46741318e-02 -9.45203304e-01 7.04651296e-01 2.90963888e-01
1.61307967e+00 -3.97713572e-01 -8.11397806e-02 5.00228405e-01
1.39506197e+00 4.46751237e-01 3.80955219e-01 5.03128469e-01
2.83046424e-01 1.16517834e-01 1.30546141e+00 9.21379089e-01
-4.03310321e-02 4.66669917e-01 5.21383345e-01 4.70827609e-01
3.15160751e-01 -4.31211919e-01 7.72677422e-01 6.65100515e-01
-9.36756358e-02 -2.28435453e-02 -5.89885473e-01 4.72420305e-01
-2.07468152e+00 -9.41981494e-01 5.46489060e-01 2.33528161e+00
1.37055528e+00 6.00941963e-02 4.02479433e-02 -2.26642564e-01
3.96752596e-01 7.49665424e-02 -1.22177923e+00 -8.59953880e-01
4.31071609e-01 6.11809254e-01 5.64093173e-01 5.63546121e-01
-9.65402782e-01 1.15096080e+00 5.93423080e+00 1.32310784e+00
-8.01506698e-01 -1.20702326e-01 5.53742826e-01 -1.95436835e-01
-1.18031293e-01 4.56773490e-03 -1.29892945e+00 5.86259842e-01
9.62136924e-01 -4.25541252e-01 8.86709392e-01 1.33372629e+00
2.06868961e-01 -4.53648090e-01 -9.59978819e-01 7.37329423e-01
-3.14010352e-01 -1.15704572e+00 -4.03675944e-01 -1.39265820e-01
9.92509902e-01 -4.17380184e-02 2.63718009e-01 1.03452194e+00
9.50308084e-01 -9.71338689e-01 4.58754182e-01 4.05171305e-01
8.03295374e-01 -9.65926170e-01 3.70337814e-01 7.86215544e-01
-6.58089578e-01 -3.58887553e-01 -5.99474847e-01 -6.54269159e-02
-4.46791649e-02 1.60386130e-01 -9.75349128e-01 1.81370780e-01
5.01779258e-01 4.79535222e-01 -1.92934871e-01 8.19892883e-01
-6.13119781e-01 9.53003883e-01 -1.66290924e-01 -8.89998019e-01
5.59738457e-01 -3.76209348e-01 3.63868952e-01 9.73736525e-01
2.47569039e-01 1.20533623e-01 3.86242777e-01 7.89836645e-01
-1.31998226e-01 4.04205441e-01 -3.91325235e-01 -2.05108956e-01
4.18439955e-01 1.14267480e+00 -3.34383398e-01 -4.66874272e-01
-2.33947039e-01 7.88528442e-01 5.90597093e-01 3.57129902e-01
-8.96662652e-01 -5.34186244e-01 5.11269510e-01 -4.62237448e-01
4.14153576e-01 -1.20255537e-02 7.56230131e-02 -9.71277416e-01
-1.14964480e-02 -1.20691252e+00 2.91943640e-01 -3.49980921e-01
-9.86749470e-01 1.64678827e-01 1.53521478e-01 -1.09249139e+00
-7.84655511e-01 -6.16392851e-01 -5.56273699e-01 8.13391685e-01
-1.89040661e+00 -1.85801342e-01 -8.30488652e-03 5.82416594e-01
6.56113446e-01 -5.69132984e-01 8.76593411e-01 -3.04312091e-02
-5.89044094e-01 8.01807821e-01 6.21855557e-01 -3.82713079e-01
7.75302231e-01 -1.71247828e+00 7.49009177e-02 5.05183399e-01
-2.50354022e-01 4.25767273e-01 5.64703703e-01 -6.72835648e-01
-1.41903818e+00 -8.69161367e-01 2.86919117e-01 1.41592398e-02
5.51826954e-01 5.67227369e-03 -5.29294670e-01 3.49250108e-01
3.09557587e-01 -1.48436561e-01 5.47161043e-01 -6.68496937e-02
9.36738923e-02 8.71773362e-02 -1.17508996e+00 6.62683487e-01
8.20500612e-01 -2.17369199e-01 -5.10392845e-01 5.29805899e-01
6.83110535e-01 -4.86520618e-01 -6.70000553e-01 2.52072245e-01
2.53958851e-01 -7.64406979e-01 5.94159603e-01 -1.05508053e+00
5.31105757e-01 4.67821546e-02 8.65639001e-03 -1.65878332e+00
5.07197492e-02 -1.08280885e+00 -6.85103714e-01 8.58958244e-01
3.59131336e-01 -5.88003397e-01 1.22801888e+00 5.83984375e-01
-2.24520311e-01 -1.03950512e+00 -6.87292635e-01 -1.01817536e+00
3.28584552e-01 -6.94373846e-02 6.70035183e-01 6.03196800e-01
3.33240777e-01 8.45815763e-02 -6.23769522e-01 -6.20599449e-01
6.34839237e-01 3.16464812e-01 6.07648909e-01 -7.50474513e-01
-7.77230740e-01 -5.76665759e-01 4.66039628e-01 -1.30810225e+00
5.39062083e-01 -6.84817135e-01 4.57755923e-01 -9.86641884e-01
1.04907118e-01 -5.93713701e-01 -4.83234674e-01 4.75585550e-01
-6.82808340e-01 -7.10864484e-01 2.28144363e-01 -6.38805032e-02
-8.92856359e-01 1.03073895e+00 1.43881416e+00 2.08862200e-02
-4.61908340e-01 2.85731465e-01 -4.96044159e-01 4.78657007e-01
1.23642778e+00 -6.22182071e-01 -7.67025173e-01 7.10383281e-02
1.08188540e-01 3.18048090e-01 -3.20454597e-01 -8.24129581e-01
7.16170967e-02 -7.90341914e-01 -8.88765082e-02 -2.62754828e-01
-2.09724754e-02 -3.57460946e-01 -6.81303918e-01 6.67094052e-01
-7.91124344e-01 1.52142048e-01 1.96950421e-01 7.84628570e-01
-1.65302530e-01 -8.14382553e-01 7.44262934e-01 -3.62920851e-01
-7.29245365e-01 4.41869885e-01 -5.16043901e-01 4.80531007e-01
9.03786480e-01 2.02321455e-01 -3.30949813e-01 -3.09020877e-01
-3.76755446e-01 5.84504247e-01 2.25846291e-01 2.95058966e-01
6.77571476e-01 -1.26513505e+00 -4.62776750e-01 -9.04718339e-02
-5.85468300e-02 -2.32196048e-01 -2.04973295e-01 6.61567986e-01
-2.19320953e-01 5.95943391e-01 -2.45495304e-01 -2.17332125e-01
-1.06575465e+00 5.70128977e-01 1.38380766e-01 -9.99643326e-01
-2.62411952e-01 8.35747719e-01 -2.33746171e-01 -5.26480854e-01
5.87784767e-01 -4.82977301e-01 -2.80564904e-01 -3.90245825e-01
5.27097940e-01 4.74506408e-01 -1.64235339e-01 3.60290915e-01
-7.22236698e-03 8.90872255e-02 -4.87029403e-01 -4.23955828e-01
1.15813720e+00 3.33409280e-01 3.85889441e-01 2.99582273e-01
7.99874783e-01 -1.95100725e-01 -1.66106570e+00 -3.05947989e-01
2.11309835e-01 -5.01752138e-01 3.48015614e-02 -9.99280751e-01
-8.33982468e-01 8.39684188e-01 6.99041843e-01 -2.06899568e-01
9.50933874e-01 -7.06350565e-01 8.10132682e-01 7.68463612e-01
5.89897871e-01 -1.72648418e+00 2.64193803e-01 7.95054913e-01
5.94050825e-01 -1.57475138e+00 7.70910904e-02 2.86728382e-01
-8.44245136e-01 9.59836185e-01 9.80753720e-01 -1.31173968e-01
3.73799533e-01 -9.59842000e-03 -2.96042949e-01 2.68344134e-01
-1.04109728e+00 -4.11504246e-02 -1.20514236e-01 5.75053871e-01
3.45473289e-01 2.86071301e-01 -7.03479528e-01 5.44215024e-01
-1.45324945e-01 3.54399711e-01 5.17121553e-01 1.37243140e+00
-6.51489377e-01 -1.50530803e+00 -2.24206045e-01 7.56671727e-01
-6.40929103e-01 -3.12294334e-01 -5.89786880e-02 5.70306301e-01
-5.32505751e-01 7.86272049e-01 -3.22708726e-01 -1.11213714e-01
2.24273965e-01 2.49583498e-01 5.37976444e-01 -4.64886039e-01
-8.38079154e-01 -1.97614148e-01 -6.16921596e-02 -7.72337973e-01
-1.40897259e-01 -8.02290142e-01 -1.50434816e+00 -9.84925106e-02
-3.15648764e-01 6.03658140e-01 7.23680794e-01 7.55327642e-01
3.59051555e-01 5.68707943e-01 7.79850781e-01 -4.94459301e-01
-1.59053278e+00 -1.10974753e+00 -4.87654358e-01 3.43632132e-01
3.29669148e-01 -8.69728744e-01 -2.70005375e-01 -2.37609699e-01]
|
[4.12155294418335, 2.220228433609009]
|
9d546522-9f3e-4782-a73b-3390a703d788
|
joint-lemmatization-and-morphological-tagging
| null | null |
https://aclanthology.org/D15-1272
|
https://aclanthology.org/D15-1272.pdf
|
Joint Lemmatization and Morphological Tagging with Lemming
| null |
['Hinrich Sch{\\"u}tze', 'er', 'Thomas M{\\"u}ller', 'Alex Fraser', 'Ryan Cotterell']
|
2015-09-01
| null | null | null |
emnlp-2015-9
|
['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]
|
[-7.376143932342529, 3.7354094982147217]
|
ae86d3c2-db0d-4887-8e1f-74cff0140b63
|
deep-curvilinear-editing-commutative-and
|
2211.14573
| null |
https://arxiv.org/abs/2211.14573v2
|
https://arxiv.org/pdf/2211.14573v2.pdf
|
Deep Curvilinear Editing: Commutative and Nonlinear Image Manipulation for Pretrained Deep Generative Model
|
Semantic editing of images is the fundamental goal of computer vision. Although deep learning methods, such as generative adversarial networks (GANs), are capable of producing high-quality images, they often do not have an inherent way of editing generated images semantically. Recent studies have investigated a way of manipulating the latent variable to determine the images to be generated. However, methods that assume linear semantic arithmetic have certain limitations in terms of the quality of image editing, whereas methods that discover nonlinear semantic pathways provide non-commutative editing, which is inconsistent when applied in different orders. This study proposes a novel method called deep curvilinear editing (DeCurvEd) to determine semantic commuting vector fields on the latent space. We theoretically demonstrate that owing to commutativity, the editing of multiple attributes depends only on the quantities and not on the order. Furthermore, we experimentally demonstrate that compared to previous methods, the nonlinear and commutative nature of DeCurvEd facilitates the disentanglement of image attributes and provides higher-quality editing.
|
['Takashi Matsubara', 'Takehiro Aoshima']
|
2022-11-26
| null |
http://openaccess.thecvf.com//content/CVPR2023/html/Aoshima_Deep_Curvilinear_Editing_Commutative_and_Nonlinear_Image_Manipulation_for_Pretrained_CVPR_2023_paper.html
|
http://openaccess.thecvf.com//content/CVPR2023/papers/Aoshima_Deep_Curvilinear_Editing_Commutative_and_Nonlinear_Image_Manipulation_for_Pretrained_CVPR_2023_paper.pdf
|
cvpr-2023-1
|
['image-manipulation']
|
['computer-vision']
|
[ 4.76682842e-01 6.68016747e-02 1.07695013e-01 -2.52066970e-01
-1.74409613e-01 -8.57948422e-01 9.99462545e-01 -4.20554787e-01
-2.62860388e-01 7.89187849e-01 1.28707783e-02 2.35602446e-02
-2.34158933e-01 -1.05639124e+00 -1.00680506e+00 -8.72133434e-01
3.80421638e-01 2.24763423e-01 -2.83999443e-01 -2.30498746e-01
1.56648934e-01 6.02860272e-01 -1.52476025e+00 1.06078744e-01
1.01614273e+00 5.87662101e-01 -7.95018673e-02 5.48147798e-01
-9.03996080e-02 6.61551654e-01 -7.17138410e-01 -4.90735412e-01
4.51530635e-01 -9.04177964e-01 -5.68547070e-01 -1.67891290e-02
3.52495760e-01 -3.07066381e-01 -2.85298973e-01 1.39985693e+00
1.83654785e-01 7.15545863e-02 1.03269160e+00 -1.62137341e+00
-1.21699572e+00 4.49019253e-01 -2.13549152e-01 -2.84057021e-01
1.48809657e-01 3.10309175e-02 8.27267826e-01 -5.69382012e-01
8.09330523e-01 1.29970372e+00 4.17096585e-01 6.95625663e-01
-1.58228266e+00 -8.92682731e-01 -2.28857011e-01 1.50499567e-01
-1.32342684e+00 -3.80944461e-01 1.07979548e+00 -4.72088128e-01
3.18800569e-01 3.59348327e-01 6.39211595e-01 1.17798662e+00
2.51012802e-01 2.41926730e-01 1.44144714e+00 -6.24386430e-01
1.83666036e-01 1.53864026e-01 -5.73202372e-01 6.87849402e-01
1.47153363e-01 3.69976550e-01 -4.56218988e-01 1.12738080e-01
1.18196261e+00 -1.77967295e-01 -2.15126321e-01 -6.24168038e-01
-1.50787663e+00 1.00610852e+00 4.82753992e-01 3.67850691e-01
-3.25019717e-01 3.34398806e-01 1.11765638e-01 4.12069380e-01
2.00419858e-01 9.60746646e-01 6.57896250e-02 9.78542641e-02
-7.25438297e-01 1.83644339e-01 5.45744658e-01 8.51823688e-01
7.57057071e-01 3.27606589e-01 -2.45211914e-01 5.46499610e-01
-1.19046740e-01 4.65072155e-01 3.58462393e-01 -1.21304786e+00
1.66154608e-01 4.65247095e-01 -1.49357617e-02 -1.08177376e+00
2.31574476e-03 -2.27111548e-01 -1.07353771e+00 7.10529804e-01
4.80067730e-01 8.12087059e-02 -9.22314584e-01 2.15698910e+00
2.93009765e-02 -1.90848522e-02 7.88251907e-02 8.38501513e-01
4.84604746e-01 5.41499555e-01 1.69351012e-01 -1.05088800e-01
1.12647021e+00 -5.90359151e-01 -9.54688966e-01 2.61866748e-01
2.70564109e-01 -7.54837990e-01 1.26796091e+00 2.41840363e-01
-1.07135665e+00 -5.19764364e-01 -1.16587102e+00 -2.63917714e-01
-4.84319568e-01 -7.49151558e-02 7.60594547e-01 7.18457878e-01
-1.10376608e+00 5.68876386e-01 -6.54520869e-01 -2.80243251e-02
3.95540804e-01 2.87699342e-01 -5.43972969e-01 2.12970883e-01
-1.35851228e+00 9.68259633e-01 2.71979243e-01 1.25707379e-02
-6.94015741e-01 -8.14826071e-01 -6.69000864e-01 2.16641247e-01
1.69678122e-01 -1.01102817e+00 7.30078042e-01 -1.48532629e+00
-1.73437774e+00 8.33154976e-01 3.72259878e-02 -2.78632581e-01
8.94674063e-01 8.54613930e-02 -2.43528262e-01 1.70827881e-01
1.31440043e-01 1.09501231e+00 1.14830065e+00 -1.38376653e+00
-4.09713387e-02 -1.52823269e-01 3.60658556e-01 1.83505490e-01
-3.44027728e-01 -2.95227349e-01 -6.22302070e-02 -9.63381171e-01
1.95443779e-02 -9.72107530e-01 -1.75000906e-01 2.85333872e-01
-5.52342176e-01 2.67939568e-01 6.74793482e-01 -4.11500901e-01
6.76568866e-01 -2.22887707e+00 5.32823086e-01 1.67228267e-01
2.88909674e-01 1.98742095e-02 -1.40636966e-01 3.60938847e-01
-2.01246679e-01 3.52875054e-01 -4.03697342e-01 -3.20750214e-02
6.28640801e-02 3.10015470e-01 -3.85238349e-01 3.78089339e-01
2.35408276e-01 1.18954504e+00 -8.53207350e-01 -3.37061286e-01
2.93783605e-01 7.06135988e-01 -7.10189641e-01 2.29839515e-03
-2.10384876e-01 7.32039273e-01 -3.59316677e-01 9.50307325e-02
5.28024316e-01 -1.43046916e-01 1.25604302e-01 -5.09314299e-01
-4.40407842e-02 -7.49351755e-02 -8.79737139e-01 1.60690844e+00
-4.73279983e-01 6.67238951e-01 -2.20406637e-01 -1.06769538e+00
7.86699057e-01 1.16097301e-01 4.06545520e-01 -8.42363179e-01
2.58545335e-02 8.19708873e-03 8.92197862e-02 -1.51864678e-01
5.23787260e-01 -4.70763206e-01 -2.86454521e-03 4.72816288e-01
-5.18296286e-03 -3.87614310e-01 1.14649490e-01 2.68849611e-01
7.74882793e-01 1.80094495e-01 1.64196745e-01 -3.75541866e-01
3.03297400e-01 -1.25123724e-01 3.06082636e-01 6.75737262e-01
1.22793280e-01 5.76774836e-01 8.09016883e-01 -2.58674592e-01
-1.51805139e+00 -1.17919648e+00 -9.44876149e-02 5.15112400e-01
3.59274209e-01 -6.56725615e-02 -1.00360644e+00 -4.04591829e-01
-1.43781498e-01 9.95218992e-01 -7.69117832e-01 -5.54291606e-01
-4.09234345e-01 -5.44997096e-01 7.53005922e-01 3.47183406e-01
6.04766846e-01 -1.12115180e+00 -5.36750495e-01 -2.81219240e-02
-2.03144848e-01 -1.18331063e+00 -6.10862494e-01 -2.45121852e-01
-6.48598254e-01 -8.88666630e-01 -6.44951284e-01 -4.67727572e-01
1.11082137e+00 -5.58562391e-02 8.71526718e-01 -6.62287548e-02
-2.03418598e-01 3.00307751e-01 -6.50853589e-02 -3.15027952e-01
-8.22844207e-01 -1.74694940e-01 -4.47399989e-02 2.32391730e-01
-4.21375930e-02 -8.13372910e-01 -5.38469791e-01 2.27070943e-01
-1.35304904e+00 4.03746724e-01 6.26940906e-01 9.06135142e-01
5.36992252e-01 2.59381533e-01 5.85611880e-01 -1.10966420e+00
6.07822835e-01 -4.30317633e-02 -6.59382164e-01 3.32551032e-01
-8.02866042e-01 4.82263058e-01 8.22359145e-01 -6.10320210e-01
-1.21383619e+00 -2.31192801e-02 9.84142870e-02 -4.32538450e-01
-2.85514407e-02 1.22316547e-01 -3.50859135e-01 -2.19376430e-01
6.80883765e-01 2.97021955e-01 3.01220685e-01 -1.20266363e-01
7.57197857e-01 -1.07341409e-01 5.86392522e-01 -4.87649113e-01
1.09450769e+00 7.51316547e-01 3.39287907e-01 -5.62864900e-01
-3.63505363e-01 3.86256099e-01 -7.46700883e-01 -2.30727360e-01
1.17554092e+00 -6.41202569e-01 -5.91924667e-01 7.12949574e-01
-9.73320663e-01 -2.46135876e-01 -5.66371739e-01 3.76726806e-01
-7.50369430e-01 2.50095218e-01 -4.58176076e-01 -3.09560180e-01
5.40959388e-02 -1.29048288e+00 7.31373370e-01 -9.09144506e-02
-1.99882507e-01 -1.04357207e+00 -3.94434810e-01 1.87854677e-01
4.72474575e-01 5.12738526e-01 1.37340379e+00 -1.07918218e-01
-8.71729910e-01 8.67730156e-02 -1.07725546e-01 4.20912445e-01
3.58222634e-01 3.74592990e-02 -6.97762311e-01 -3.06946672e-02
3.46893519e-02 1.03626484e-02 6.40976727e-01 4.46208030e-01
1.11943841e+00 -5.66084146e-01 -4.50097732e-02 8.48238468e-01
1.39206672e+00 3.05768251e-01 1.05867565e+00 3.38524580e-01
9.34128761e-01 5.34290791e-01 1.05023660e-01 -6.47924170e-02
1.12269469e-01 8.28296840e-01 3.70168984e-01 -3.47422093e-01
-2.91449338e-01 -4.19759840e-01 2.56684661e-01 6.52507186e-01
-2.31528640e-01 -1.11571610e-01 -4.61625397e-01 2.04336420e-01
-1.39778411e+00 -9.89787877e-01 -1.17845666e-02 2.09072614e+00
9.93120253e-01 5.36878686e-03 -4.12750155e-01 1.91093177e-01
7.56414652e-01 8.32039714e-02 -5.23571074e-01 -4.88525361e-01
-3.52675706e-01 3.20553720e-01 6.63784683e-01 4.34858173e-01
-6.89611256e-01 9.28674877e-01 6.88371229e+00 8.29343736e-01
-1.22241449e+00 2.07521021e-01 3.81923944e-01 1.01096958e-01
-9.01000202e-01 8.14069957e-02 -1.88757494e-01 4.73878413e-01
4.00294274e-01 -1.33089796e-01 6.71754658e-01 5.43818355e-01
7.59459734e-02 4.67236862e-02 -1.09201169e+00 7.51323819e-01
6.44849092e-02 -1.47699130e+00 6.31463110e-01 1.22282542e-01
9.88374174e-01 -7.95504034e-01 3.10461760e-01 -8.25341195e-02
2.04683036e-01 -1.26278412e+00 9.67357635e-01 7.38500297e-01
1.15627313e+00 -8.73885870e-01 3.06088895e-01 1.55955479e-01
-5.78596473e-01 2.28294820e-01 -1.52072489e-01 1.33392692e-01
1.78444043e-01 6.19605243e-01 -4.15701330e-01 3.97711724e-01
4.39163893e-01 4.31580722e-01 -4.11167622e-01 2.56178737e-01
-6.08639300e-01 3.04051042e-01 2.19162963e-02 2.28297994e-01
-5.67733645e-02 -6.95008576e-01 6.88907206e-01 5.63766181e-01
4.29460347e-01 1.98909968e-01 -4.68000054e-01 1.52187324e+00
-2.54194260e-01 -1.62022874e-01 -7.49915063e-01 -3.81523162e-01
3.91618788e-01 9.14063394e-01 -7.68490970e-01 -2.51327604e-01
-2.15327069e-01 1.20585275e+00 -4.10262011e-02 5.95561266e-01
-9.58843231e-01 -2.97542155e-01 7.29178131e-01 2.93778945e-02
2.44261026e-01 -3.54664236e-01 -4.82922852e-01 -1.11167169e+00
5.36545441e-02 -9.70743418e-01 -2.69487798e-01 -1.03418303e+00
-1.10126185e+00 4.43722427e-01 6.17434606e-02 -1.12725425e+00
-1.90793321e-01 -3.93181831e-01 -2.81052113e-01 9.69834030e-01
-1.19020581e+00 -1.46236324e+00 -2.51518905e-01 7.35770106e-01
2.53565550e-01 -1.03416666e-01 8.52231026e-01 1.64019302e-01
-1.05004802e-01 6.73428833e-01 1.53531224e-01 3.94630432e-02
6.38203740e-01 -1.11045194e+00 1.23719975e-01 8.80214989e-01
3.21368724e-01 7.03670859e-01 8.30249667e-01 -4.31827724e-01
-1.31685233e+00 -8.97788942e-01 6.39019310e-01 -3.45760554e-01
2.77815759e-01 -4.15667176e-01 -8.76386583e-01 6.51445866e-01
1.81821555e-01 -2.23190248e-01 3.83949876e-01 -2.52164543e-01
-5.50573587e-01 -1.28552943e-01 -1.13617122e+00 8.64678085e-01
1.12121761e+00 -8.22086573e-01 -2.81390339e-01 4.11622003e-02
8.52248609e-01 -2.31689736e-01 -8.69795859e-01 5.02919555e-01
5.47065020e-01 -1.03060937e+00 1.12758350e+00 -5.16681373e-01
7.62244642e-01 -2.57536590e-01 5.80325387e-02 -1.49458587e+00
-3.44444394e-01 -3.80390972e-01 3.20682317e-01 1.25848150e+00
1.96426883e-01 -8.50585580e-01 4.86170113e-01 5.55648625e-01
1.08688794e-01 -3.47886473e-01 -6.36658788e-01 -8.87067020e-01
2.26285368e-01 -1.07900374e-01 7.91991651e-01 1.27926564e+00
-4.97923762e-01 1.82192981e-01 -4.73567188e-01 1.67964380e-02
5.77499926e-01 3.18312682e-02 6.20824516e-01 -9.12232816e-01
-1.32341668e-01 -6.11768007e-01 -6.10526741e-01 -6.03065670e-01
3.98592442e-01 -9.04644608e-01 -1.68269753e-01 -1.19175386e+00
9.39054415e-02 -6.17186785e-01 -1.78674892e-01 3.39019328e-01
1.06463678e-01 3.97665918e-01 1.50295600e-01 3.77479225e-01
1.02992259e-01 7.30072320e-01 1.69953120e+00 -1.33931711e-01
-7.52085596e-02 -3.97600263e-01 -7.06401110e-01 4.91743356e-01
7.53250182e-01 -4.51101750e-01 -6.51190221e-01 -5.19631803e-01
4.30476069e-01 -1.14057384e-01 7.78659463e-01 -7.51531422e-01
1.10957734e-01 -2.40537152e-01 3.76380473e-01 6.05246909e-02
3.38626951e-01 -5.87715328e-01 7.85085499e-01 3.22930574e-01
-6.36039853e-01 3.23100761e-02 -1.24719907e-02 4.00095224e-01
-3.59536231e-01 -1.62487015e-01 8.66648734e-01 -2.80860871e-01
-5.62402189e-01 1.47362202e-01 -3.80544215e-01 -1.28193319e-01
1.08975077e+00 -2.69040525e-01 -2.28539541e-01 -4.97018337e-01
-4.88795549e-01 -4.69933122e-01 7.75487006e-01 4.22305733e-01
3.52301240e-01 -1.55384922e+00 -4.58076507e-01 4.14069563e-01
-2.45456681e-01 -1.91028073e-01 2.38105282e-01 5.62552214e-01
-5.73461831e-01 3.22614700e-01 -7.43463874e-01 -3.88523549e-01
-8.83837163e-01 6.51444912e-01 2.87674665e-01 -2.83483900e-02
-6.11924767e-01 6.46844625e-01 7.27720141e-01 -3.33810598e-01
-2.71365285e-01 -3.03864088e-02 -4.69425172e-02 -1.23799080e-02
7.70978928e-02 1.08378276e-01 -6.96617588e-02 -7.17888832e-01
-3.99728492e-02 6.23570025e-01 3.13814878e-01 -3.22985500e-01
1.01443243e+00 -1.72315910e-03 -3.97521406e-01 2.10113898e-01
1.07589614e+00 1.91606998e-01 -1.32851028e+00 7.54361376e-02
-5.31804800e-01 -6.00199044e-01 -4.96413335e-02 -5.71117520e-01
-1.23901010e+00 7.32623756e-01 4.02171403e-01 2.75283635e-01
1.30508482e+00 -1.82430655e-01 3.83695245e-01 2.51675993e-02
2.61456370e-01 -9.17394876e-01 2.04984128e-01 5.08006103e-02
1.06281006e+00 -9.02107894e-01 -2.07021222e-01 -5.61636686e-01
-6.25053346e-01 8.85833383e-01 4.68108177e-01 -4.38346192e-02
5.00020862e-01 1.41554102e-01 4.66026962e-02 -8.19438696e-02
-3.94655824e-01 8.31561312e-02 2.76055545e-01 6.45968258e-01
1.20918334e-01 2.11637899e-01 -4.06493604e-01 7.09293410e-02
-5.25784373e-01 -1.50933549e-01 6.48292482e-01 5.52398860e-01
3.22853357e-01 -1.32681262e+00 -3.50546837e-01 1.83528811e-01
-3.77324194e-01 -1.13942750e-01 -2.68321872e-01 8.94338608e-01
3.71369541e-01 4.92093831e-01 6.13280982e-02 -1.70526102e-01
1.68195769e-01 3.97361778e-02 8.24736953e-01 -2.19062001e-01
-2.27539331e-01 -3.43079060e-01 -2.31276006e-01 -4.74039942e-01
-6.77624464e-01 -4.58522469e-01 -1.17943072e+00 -3.70563447e-01
-2.48788312e-01 1.00263171e-01 8.34886014e-01 9.11899984e-01
3.62024009e-01 6.67095721e-01 4.45235163e-01 -6.02942646e-01
-3.98314267e-01 -5.99562824e-01 -5.71698785e-01 9.13793206e-01
1.48894891e-01 -9.45088565e-01 -3.58946919e-01 6.19009018e-01]
|
[11.708251953125, -0.3684540390968323]
|
cf43e438-1a61-4db6-a488-4068bc301495
|
eliminating-gradient-conflict-in-reference
|
2207.06095
| null |
https://arxiv.org/abs/2207.06095v3
|
https://arxiv.org/pdf/2207.06095v3.pdf
|
Eliminating Gradient Conflict in Reference-based Line-Art Colorization
|
Reference-based line-art colorization is a challenging task in computer vision. The color, texture, and shading are rendered based on an abstract sketch, which heavily relies on the precise long-range dependency modeling between the sketch and reference. Popular techniques to bridge the cross-modal information and model the long-range dependency employ the attention mechanism. However, in the context of reference-based line-art colorization, several techniques would intensify the existing training difficulty of attention, for instance, self-supervised training protocol and GAN-based losses. To understand the instability in training, we detect the gradient flow of attention and observe gradient conflict among attention branches. This phenomenon motivates us to alleviate the gradient issue by preserving the dominant gradient branch while removing the conflict ones. We propose a novel attention mechanism using this training strategy, Stop-Gradient Attention (SGA), outperforming the attention baseline by a large margin with better training stability. Compared with state-of-the-art modules in line-art colorization, our approach demonstrates significant improvements in Fr\'echet Inception Distance (FID, up to 27.21%) and structural similarity index measure (SSIM, up to 25.67%) on several benchmarks. The code of SGA is available at https://github.com/kunkun0w0/SGA .
|
['Yibo Yang', 'Wenyu Chen', 'Zhao Kang', 'Zhengyang Geng', 'Zekun Li']
|
2022-07-13
| null | null | null | null |
['colorization', 'line-art-colorization']
|
['computer-vision', 'computer-vision']
|
[ 8.95819515e-02 -3.16824406e-01 1.42682577e-02 -2.44490594e-01
-6.26243472e-01 -5.23222506e-01 5.11167705e-01 -2.64467865e-01
-2.07634449e-01 6.03370428e-01 -5.32061458e-02 -2.28679687e-01
6.99722255e-03 -6.98663116e-01 -8.50282669e-01 -7.73108125e-01
3.51946026e-01 -6.08610623e-02 2.26492584e-01 -3.98763001e-01
4.94072944e-01 5.36405981e-01 -1.22180498e+00 2.12016612e-01
1.22418213e+00 1.05194032e+00 3.86572704e-02 4.61305022e-01
-5.61695457e-01 4.38428760e-01 -5.19201100e-01 -6.80235207e-01
3.37525934e-01 -6.58463597e-01 -4.02855396e-01 -2.58382596e-02
7.24157512e-01 -2.40750074e-01 -2.04179212e-01 1.16818738e+00
5.46509087e-01 2.36863084e-02 5.12759745e-01 -1.21324098e+00
-9.25313890e-01 2.80972779e-01 -1.22128427e+00 7.18705952e-02
9.27091613e-02 1.13376625e-01 1.04899228e+00 -1.06231713e+00
4.72297877e-01 1.24768662e+00 6.70225263e-01 5.25473595e-01
-1.28097761e+00 -7.81900406e-01 4.18984771e-01 2.01731771e-01
-1.32085872e+00 -2.17480302e-01 1.23709345e+00 -3.82954419e-01
3.54141623e-01 3.68215412e-01 5.17726362e-01 1.06543922e+00
6.45298627e-04 8.28869641e-01 1.12976015e+00 -4.70227480e-01
-4.63221632e-02 -8.83788094e-02 1.02973901e-01 8.52813840e-01
1.49787843e-01 4.99621220e-02 -5.16720891e-01 1.96618319e-01
1.04912698e+00 3.17539014e-02 -4.62462604e-01 -3.70662898e-01
-9.46783781e-01 5.71193755e-01 7.72380412e-01 2.69671082e-01
1.10255985e-03 3.38596493e-01 2.60281801e-01 1.51035666e-01
4.64525104e-01 3.97807926e-01 -2.53983289e-01 -6.69455752e-02
-8.35275531e-01 -1.04228772e-01 2.72932023e-01 8.13188493e-01
8.70491743e-01 1.55153647e-01 -3.90243292e-01 1.01208270e+00
1.48494333e-01 4.66744125e-01 1.79982886e-01 -8.42868507e-01
5.36606610e-01 7.88715661e-01 -4.00912091e-02 -1.01477492e+00
-3.51717994e-02 -6.44038200e-01 -1.10665214e+00 6.01252437e-01
5.93118489e-01 -4.69700247e-02 -9.31554079e-01 1.62924886e+00
1.99071392e-01 1.22512192e-01 -4.44770068e-01 1.06259894e+00
6.12844408e-01 4.68520463e-01 4.00664890e-03 4.66460995e-02
1.19408703e+00 -1.22054696e+00 -5.54554045e-01 -3.49521376e-02
1.82936445e-01 -9.54766631e-01 1.58131301e+00 2.31674746e-01
-1.09583342e+00 -7.81627476e-01 -1.15556228e+00 -3.03327531e-01
-3.41597199e-01 3.43182951e-01 6.50524020e-01 5.38672447e-01
-8.26839328e-01 7.96046078e-01 -5.91676831e-01 -1.84030935e-01
5.96291363e-01 3.00100353e-02 -1.12386078e-01 -1.90770090e-01
-9.23722863e-01 3.62171113e-01 -5.56753092e-02 2.88702995e-01
-3.87106419e-01 -9.74413395e-01 -5.63029885e-01 1.20499887e-01
2.79795021e-01 -5.95124662e-01 7.01163054e-01 -1.37770617e+00
-1.71692598e+00 7.49509096e-01 -2.03156918e-02 -2.21755765e-02
8.72544229e-01 -5.20156980e-01 -1.20253742e-01 -8.25688317e-02
-7.27628618e-02 6.13749683e-01 9.91539896e-01 -1.49793839e+00
-3.91994566e-01 -3.67425203e-01 1.15462996e-01 2.05071896e-01
-2.72581935e-01 -1.93406507e-01 -8.32557380e-01 -9.75506604e-01
5.81512861e-02 -7.66997993e-01 1.11825466e-01 3.06401521e-01
-5.02507865e-01 -2.97486335e-02 8.68151307e-01 -4.81979042e-01
1.35472298e+00 -2.19177246e+00 8.76113996e-02 1.38222933e-01
1.68929264e-01 3.36887360e-01 -1.77238196e-01 3.07208896e-01
-1.40055686e-01 2.01409444e-01 -4.24515873e-01 -3.90822798e-01
-5.58516793e-02 2.66788155e-02 -4.65467006e-01 3.25656503e-01
2.79649645e-01 8.63535285e-01 -8.32197845e-01 -4.30921763e-01
2.50972271e-01 6.26734138e-01 -5.22119582e-01 2.32678115e-01
-9.69235897e-02 5.28560460e-01 -3.33178997e-01 7.08373725e-01
9.08204854e-01 -1.13487758e-01 -3.41994278e-02 -5.49688697e-01
-2.59392798e-01 -1.51938379e-01 -9.97331560e-01 2.01659012e+00
-4.18692023e-01 6.68358922e-01 -2.83089299e-02 -6.53842628e-01
1.06123769e+00 -1.66964531e-01 3.07703316e-01 -1.03775346e+00
4.95830104e-02 1.20863914e-01 -8.96545053e-02 -1.76489800e-01
3.07270736e-01 1.22599110e-01 1.88428059e-01 1.96007475e-01
-3.47305596e-01 6.81250542e-02 1.26366645e-01 2.08993942e-01
7.04141319e-01 5.23001015e-01 -2.94408560e-01 -3.57295990e-01
6.02161169e-01 -3.77027869e-01 5.64093471e-01 5.70821047e-01
-1.26583260e-02 8.97516429e-01 6.75514877e-01 -3.81893992e-01
-9.87277865e-01 -9.34363425e-01 -2.88343634e-02 1.21565270e+00
4.28316206e-01 -3.31404835e-01 -6.59718335e-01 -7.09519506e-01
8.45958143e-02 6.15273297e-01 -8.40669692e-01 -2.09507152e-01
-6.45251572e-01 -6.56284750e-01 3.89738619e-01 5.91723144e-01
7.93683708e-01 -9.74898517e-01 -4.08235103e-01 -7.07300305e-02
7.28513524e-02 -7.21188784e-01 -8.61241996e-01 -1.54431596e-01
-7.39961922e-01 -1.01669705e+00 -1.08999372e+00 -4.23492432e-01
8.53208005e-01 2.84919888e-01 1.24793983e+00 3.31214100e-01
-3.43540519e-01 2.52377540e-01 -2.54801631e-01 -1.61232740e-01
8.80258679e-02 1.13591529e-01 -5.83725452e-01 3.05059820e-01
-7.05360547e-02 -5.66524446e-01 -1.00222170e+00 2.97087431e-01
-7.32996821e-01 4.00819272e-01 7.32453525e-01 1.05619991e+00
5.30904949e-01 -4.26908642e-01 3.42314243e-01 -7.78869927e-01
5.59352875e-01 -1.87365934e-02 -7.72089481e-01 4.49928194e-01
-6.72862947e-01 2.43239403e-01 6.74615562e-01 -3.25434119e-01
-1.22746980e+00 -8.57006088e-02 -1.38589635e-01 -7.41085589e-01
2.74856798e-02 5.57568558e-02 -2.81222761e-01 -1.59537420e-01
3.59541208e-01 1.26193404e-01 -2.06032708e-01 -4.87860829e-01
4.13118541e-01 1.99244872e-01 4.91483837e-01 -7.74245203e-01
8.70979488e-01 5.68783045e-01 1.00705745e-02 -5.98131359e-01
-8.57348084e-01 -2.07597107e-01 -5.00795841e-01 -1.85967118e-01
7.62660027e-01 -5.50287127e-01 -6.81430280e-01 5.03847420e-01
-1.15361476e+00 -5.83634079e-01 -1.33627042e-01 1.10205449e-01
-1.71914458e-01 4.64309573e-01 -6.09501004e-01 -8.04557502e-01
-5.16055942e-01 -9.19390023e-01 1.06470680e+00 4.18665648e-01
2.24495575e-01 -8.20055842e-01 -2.11466197e-02 8.72154385e-02
5.40176034e-01 4.03119177e-01 9.47154701e-01 1.22502655e-01
-7.29789436e-01 1.56887457e-01 -7.56451607e-01 2.79397696e-01
1.48745984e-01 3.78543705e-01 -9.46571887e-01 -3.20503622e-01
-4.15571690e-01 -9.54808369e-02 1.09797990e+00 3.40422869e-01
1.36029637e+00 4.96078581e-02 -2.20723618e-02 9.63450730e-01
1.64704442e+00 2.45598391e-01 9.08542514e-01 2.91280419e-01
1.08664179e+00 3.76405239e-01 3.66153926e-01 3.12393576e-01
1.57605514e-01 7.17167199e-01 3.22449028e-01 -5.65075874e-01
-4.99034435e-01 -3.72317791e-01 8.60698968e-02 6.06059194e-01
-2.21290261e-01 -1.03137009e-02 -7.14028060e-01 3.19691539e-01
-1.70676732e+00 -7.24467397e-01 -1.56430706e-01 2.33915949e+00
6.75429881e-01 2.26630330e-01 -6.03499152e-02 9.67719778e-02
7.27206767e-01 2.04540342e-01 -6.72957957e-01 -3.56554478e-01
-2.31606513e-01 1.03780597e-01 4.57200199e-01 5.31252205e-01
-8.60336959e-01 9.75022674e-01 4.95505619e+00 9.27909076e-01
-1.47525132e+00 -1.16052762e-01 8.83716762e-01 1.60666138e-01
-5.45846641e-01 -9.47627574e-02 -5.08782089e-01 6.38070345e-01
1.14738740e-01 1.01855271e-01 5.73767662e-01 5.57226896e-01
1.42950928e-02 -4.30331156e-02 -9.11187649e-01 1.14129436e+00
1.36726931e-01 -1.14352512e+00 6.39479384e-02 -1.69928163e-01
8.11876476e-01 -2.02769980e-01 2.77117640e-01 2.59493977e-01
6.59125019e-03 -7.87051857e-01 6.57489538e-01 5.56218922e-01
1.08162415e+00 -7.60340452e-01 3.89481008e-01 -2.32096449e-01
-1.34751272e+00 1.34246752e-01 -2.88866401e-01 2.78591603e-01
-3.01099513e-02 7.24892676e-01 -8.40208754e-02 7.87378609e-01
7.29405105e-01 7.60749817e-01 -8.03788006e-01 9.53703523e-01
-3.49840432e-01 5.12046218e-01 -7.37951919e-02 2.41176382e-01
2.46625543e-01 -6.94050491e-01 4.40330058e-01 1.19702339e+00
2.27708966e-01 -8.88055563e-02 -1.87191352e-01 1.17727244e+00
-2.04870492e-01 9.38484743e-02 -3.25337946e-01 1.18803203e-01
1.68693140e-01 1.27156830e+00 -6.89371943e-01 -2.56957382e-01
-4.66468364e-01 1.39129722e+00 4.41999972e-01 6.58289671e-01
-1.18012762e+00 -6.46787286e-01 5.16301811e-01 1.38805822e-01
4.39484775e-01 -2.60552466e-01 -6.28954113e-01 -1.12621331e+00
2.28072405e-01 -6.56884015e-01 2.46547401e-01 -8.79394233e-01
-1.37058473e+00 7.28617787e-01 -4.05437142e-01 -1.20836484e+00
3.03508848e-01 -5.08926630e-01 -1.00757217e+00 8.70219707e-01
-1.74869561e+00 -1.19261622e+00 -7.06808031e-01 6.60456061e-01
5.62519431e-01 6.09242357e-02 4.61060017e-01 5.53043485e-01
-8.77327383e-01 9.45024312e-01 2.37048700e-01 1.61033645e-01
9.61221516e-01 -1.26671290e+00 3.57912928e-01 8.93964469e-01
1.46757528e-01 5.35081446e-01 3.40766340e-01 -4.81295019e-01
-1.38519025e+00 -9.83745813e-01 3.25022727e-01 -1.80027753e-01
5.55223525e-01 -3.55044395e-01 -1.06381786e+00 2.38623217e-01
3.20915759e-01 1.42944261e-01 2.22659081e-01 6.76293001e-02
-7.53950834e-01 -5.56466222e-01 -8.11609626e-01 8.35958719e-01
1.19471705e+00 -3.89304727e-01 -1.76914439e-01 -1.06451333e-01
4.35106426e-01 -3.84487540e-01 -5.82255661e-01 3.28050762e-01
8.17380905e-01 -1.33479297e+00 8.98708045e-01 -2.62005627e-01
6.86090112e-01 -4.52844888e-01 1.55576482e-01 -1.10251975e+00
-3.85854304e-01 -6.08187675e-01 7.06855506e-02 1.42303383e+00
2.46864915e-01 -5.40183365e-01 7.32298911e-01 5.75030267e-01
-2.53284007e-01 -9.62520063e-01 -7.08428264e-01 -5.69291413e-01
1.41958028e-01 -3.01394872e-02 4.74567533e-01 1.01231349e+00
-4.65941340e-01 3.19436908e-01 -4.00300533e-01 -1.43829789e-02
7.90522754e-01 5.43893397e-01 7.48646557e-01 -1.00681162e+00
-1.62221029e-01 -8.19365978e-01 -2.00754069e-02 -1.15120971e+00
-6.91024400e-03 -6.38884425e-01 -1.86410218e-01 -1.55812776e+00
-1.66237839e-02 -5.61066628e-01 -4.99085903e-01 4.04231578e-01
-4.51579809e-01 4.03101504e-01 4.30476755e-01 1.50996879e-01
-5.39933443e-01 8.10191214e-01 1.46238029e+00 -2.08591789e-01
-1.90106034e-01 -3.54636669e-01 -6.05128765e-01 6.01163507e-01
9.11672771e-01 -1.30101651e-01 -3.67445141e-01 -8.56420100e-01
1.44888507e-02 -3.06350976e-01 3.67693156e-01 -9.85482574e-01
2.33484194e-01 -1.31326109e-01 6.00468755e-01 -5.50950706e-01
1.77891463e-01 -7.63396323e-01 1.19733132e-01 3.93886775e-01
-2.79182851e-01 1.38413191e-01 2.20263675e-01 4.34316367e-01
-1.78793207e-01 8.24511126e-02 9.16635334e-01 2.62011345e-02
-6.97405875e-01 3.75939339e-01 2.62219578e-01 1.56490207e-01
7.37552404e-01 -2.16104582e-01 -4.13979024e-01 -1.56050026e-01
-2.40512788e-01 1.14140458e-01 6.06556535e-01 4.23047364e-01
5.73537529e-01 -1.46138322e+00 -6.39202595e-01 4.49792266e-01
5.60361743e-02 -1.73923485e-02 4.66975689e-01 7.43275225e-01
-4.64406312e-01 6.73574284e-02 -4.43241984e-01 -4.62421864e-01
-9.47407186e-01 5.16413689e-01 3.88974011e-01 -8.45007151e-02
-6.25764906e-01 9.62377191e-01 5.59285164e-01 9.95774418e-02
2.74092972e-01 -2.89561689e-01 2.06604883e-01 -1.27650648e-01
2.82206744e-01 4.11120266e-01 -6.52067736e-02 -2.88576454e-01
-2.66313970e-01 1.04228342e+00 -8.31082240e-02 6.79576024e-02
1.03159177e+00 -1.33423030e-01 -1.98669191e-02 3.50422025e-01
1.13027585e+00 3.35772365e-01 -1.60007763e+00 1.50467840e-03
-2.39842057e-01 -7.83466041e-01 4.43047173e-02 -8.96832705e-01
-1.49892032e+00 1.06443512e+00 7.47498930e-01 6.83606043e-02
1.29742849e+00 -2.83558905e-01 7.79394209e-01 -3.20745930e-02
1.28762275e-01 -9.64006901e-01 2.50315696e-01 2.52896130e-01
1.30821443e+00 -1.28598344e+00 5.89454100e-02 -5.08872330e-01
-6.74769998e-01 1.12759483e+00 8.98821294e-01 -2.81587690e-01
4.38544124e-01 8.73349309e-02 1.97769359e-01 -7.97296539e-02
-4.04582590e-01 -2.02511311e-01 5.01064301e-01 2.65272170e-01
6.33675516e-01 -1.61884874e-01 -3.41819435e-01 3.65806729e-01
1.81123331e-01 -2.31425360e-01 5.21334037e-02 5.03626585e-01
6.24330454e-02 -1.14648187e+00 -2.86735803e-01 2.01188400e-01
-3.29240263e-01 -2.69051790e-01 -4.01777476e-01 7.28018641e-01
1.58554360e-01 4.93059099e-01 1.26344427e-01 -2.74616092e-01
5.19555569e-01 -1.22131892e-01 4.31307495e-01 9.76912584e-03
-4.73789871e-01 3.06290418e-01 -2.58949727e-01 -6.39544904e-01
-3.38649213e-01 -1.75670847e-01 -9.57154393e-01 -3.28307152e-01
-2.75868207e-01 -1.10553071e-01 4.93418336e-01 3.50191534e-01
5.61566055e-01 6.74462974e-01 7.29367375e-01 -8.75628710e-01
-1.80915043e-01 -7.30284214e-01 -4.40623760e-01 7.64167249e-01
1.99444398e-01 -6.94365263e-01 -4.96540844e-01 -6.60016015e-02]
|
[11.422110557556152, -0.8411617279052734]
|
a70e7b74-b7e6-4c01-bd44-255bbaf97f0e
|
attention-based-aspect-reasoning-for
|
2108.00513
| null |
https://arxiv.org/abs/2108.00513v2
|
https://arxiv.org/pdf/2108.00513v2.pdf
|
Attention-based Aspect Reasoning for Knowledge Base Question Answering on Clinical Notes
|
Question Answering (QA) in clinical notes has gained a lot of attention in the past few years. Existing machine reading comprehension approaches in clinical domain can only handle questions about a single block of clinical texts and fail to retrieve information about multiple patients and their clinical notes. To handle more complex questions, we aim at creating knowledge base from clinical notes to link different patients and clinical notes, and performing knowledge base question answering (KBQA). Based on the expert annotations available in the n2c2 dataset, we first created the ClinicalKBQA dataset that includes around 9K QA pairs and covers questions about seven medical topics using more than 300 question templates. Then, we investigated an attention-based aspect reasoning (AAR) method for KBQA and analyzed the impact of different aspects of answers (e.g., entity, type, path, and context) for prediction. The AAR method achieves better performance due to the well-designed encoder and attention mechanism. From our experiments, we find that both aspects, type and path, enable the model to identify answers satisfying the general conditions and produce lower precision and higher recall. On the other hand, the aspects, entity and context, limit the answers by node-specific information and lead to higher precision and lower recall.
|
['Chandan K. Reddy', 'Sutanay Choudhury', 'Khushbu Agarwal', 'Tian Shi', 'Ping Wang']
|
2021-08-01
| null | null | null | null |
['knowledge-base-question-answering']
|
['natural-language-processing']
|
[ 5.92540298e-03 6.50967240e-01 -1.68452218e-01 -3.31751466e-01
-1.04332423e+00 -3.30131024e-01 1.93071187e-01 7.82320917e-01
-2.14399397e-01 7.14348674e-01 7.97001660e-01 -5.75106382e-01
-4.95916486e-01 -8.37705731e-01 -3.27893496e-01 -1.53201684e-01
2.29341209e-01 8.09846282e-01 5.91991961e-01 -3.35075766e-01
-1.35512263e-01 -2.03300908e-01 -8.24431419e-01 8.64081740e-01
1.23241210e+00 1.03048337e+00 1.02520853e-01 6.80200636e-01
-4.56453800e-01 1.24806654e+00 -6.82702661e-01 -9.67482686e-01
-2.94626802e-01 -7.76937842e-01 -1.37367761e+00 -4.78557765e-01
1.77508332e-02 -2.81355143e-01 -3.47997725e-01 7.44027317e-01
6.96278870e-01 -2.67432541e-01 5.07938981e-01 -8.52692842e-01
-8.85998726e-01 6.67418420e-01 -5.72704077e-02 4.78875518e-01
6.18225515e-01 1.72132850e-01 1.25973666e+00 -3.75929236e-01
6.10206246e-01 1.02438140e+00 6.71326697e-01 7.40408838e-01
-5.74608803e-01 -2.54189998e-01 7.12716728e-02 6.34037077e-01
-1.12012458e+00 -1.69089008e-02 3.08945984e-01 -1.43851295e-01
1.19409513e+00 5.11219263e-01 5.93224585e-01 8.18427444e-01
3.60712081e-01 8.25504661e-01 6.21969879e-01 -2.35704035e-01
1.94150254e-01 1.59220621e-01 4.50216860e-01 6.45607471e-01
1.65941536e-01 -4.54219460e-01 -1.48699492e-01 -5.54450095e-01
4.11493570e-01 3.96050140e-02 -7.34841287e-01 1.68416306e-01
-1.23740625e+00 7.29649663e-01 5.97313106e-01 3.04777384e-01
-4.10774142e-01 -2.67883182e-01 5.76740980e-01 1.24835603e-01
-6.21713549e-02 8.22220683e-01 -9.56191361e-01 -1.55379232e-02
-3.05529296e-01 1.94300517e-01 1.21138060e+00 1.07730877e+00
2.78558433e-01 -8.73823225e-01 -9.09204245e-01 6.40509903e-01
1.45025119e-01 4.42509949e-01 6.77973032e-01 -6.75787508e-01
7.73319721e-01 1.10431004e+00 -1.02380008e-01 -9.19759572e-01
-5.90970576e-01 -4.59649384e-01 -8.27037990e-01 -1.04561853e+00
2.87987143e-01 -1.93578959e-01 -8.91737700e-01 1.60446966e+00
3.16743612e-01 -4.14928533e-02 4.94729012e-01 7.40948498e-01
1.52983761e+00 4.38647091e-01 5.05329549e-01 -1.44000709e-01
2.05799103e+00 -1.09186423e+00 -1.14680779e+00 -9.66918170e-02
9.56102729e-01 -5.61359227e-01 9.23946202e-01 -1.08067565e-01
-8.22411418e-01 -1.19675353e-01 -5.48632503e-01 -3.08835864e-01
-3.32563251e-01 -1.94627903e-02 4.20644194e-01 1.75225630e-01
-6.88632488e-01 1.40655965e-01 -8.17510247e-01 -3.14600319e-01
4.28767592e-01 8.23770091e-02 -7.30139092e-02 -3.55549872e-01
-1.66204345e+00 1.01746464e+00 3.00859660e-01 -3.52118686e-02
-6.15219772e-01 -1.11722291e+00 -7.67945766e-01 4.19146091e-01
5.83274186e-01 -1.35489786e+00 1.36308396e+00 -3.89120191e-01
-1.10126305e+00 5.22448719e-01 -2.58648515e-01 -3.24310899e-01
1.92903072e-01 -2.40993261e-01 -6.47349775e-01 5.41254461e-01
-2.44640578e-02 5.21594226e-01 8.40160176e-02 -7.33437538e-01
-4.89909559e-01 -4.65035558e-01 5.62339962e-01 2.71048069e-01
-1.54789478e-01 -2.38613501e-01 -5.72601616e-01 -3.86266440e-01
-1.08690366e-01 -5.91057479e-01 -3.30581039e-01 -2.22004473e-01
-5.75460851e-01 -3.13951790e-01 3.18713576e-01 -1.20415187e+00
1.61172974e+00 -1.79008293e+00 3.74956876e-02 -9.30043235e-02
4.98755664e-01 3.76033336e-01 -3.54960449e-02 5.91193080e-01
8.29207748e-02 2.38611504e-01 -1.77123770e-01 3.51982027e-01
-2.80412018e-01 3.75402480e-01 -2.24952340e-01 -3.82548124e-01
4.24892604e-01 1.34175813e+00 -9.16534066e-01 -8.92517924e-01
-4.63512659e-01 2.80663282e-01 -8.94771338e-01 4.07334238e-01
-6.34083092e-01 2.46055603e-01 -8.52198839e-01 8.79016936e-01
3.20317984e-01 -9.46452320e-01 1.35676011e-01 -3.70696336e-01
7.54145801e-01 7.14263678e-01 -4.49261397e-01 1.40750372e+00
-4.54555988e-01 1.02069415e-01 -1.58577949e-01 -4.97247130e-01
6.76261127e-01 7.27875471e-01 3.72776091e-01 -7.38626540e-01
-1.87301775e-03 1.18142426e-01 3.45943689e-01 -1.11596155e+00
1.70844764e-01 -1.58125043e-01 1.16402969e-01 2.66477793e-01
-7.14436173e-02 7.49271885e-02 6.40096888e-02 4.87840533e-01
1.61890173e+00 -3.66396368e-01 6.88649595e-01 -3.52988727e-02
8.03396463e-01 3.30725431e-01 5.83430290e-01 6.25401378e-01
-2.43046775e-01 3.97644728e-01 7.41041720e-01 -3.17777783e-01
-4.60280359e-01 -7.06968009e-01 -1.61855116e-01 7.39108503e-01
1.23016171e-01 -7.15419412e-01 -5.93094766e-01 -1.03870118e+00
-1.04939133e-01 6.89688683e-01 -5.97090721e-01 -4.44337606e-01
-5.22654235e-01 -6.95763886e-01 5.56083798e-01 7.50211239e-01
5.59279501e-01 -1.16104329e+00 -5.46594143e-01 2.24247083e-01
-7.99949348e-01 -1.18578362e+00 -4.64203179e-01 -3.96983661e-02
-9.11331713e-01 -1.46376991e+00 -7.57677257e-01 -6.91540003e-01
5.36426425e-01 -2.26655975e-01 1.62702131e+00 3.44012439e-01
-1.25126898e-01 5.46378016e-01 -7.54088819e-01 -4.71195549e-01
-3.92858297e-01 3.85966331e-01 -7.37708569e-01 -4.61069316e-01
7.19890177e-01 -1.11915454e-01 -9.68741059e-01 2.93739051e-01
-9.40228641e-01 -6.21309578e-02 8.68487895e-01 9.05803859e-01
5.55663824e-01 -3.73200893e-01 7.82889545e-01 -1.06386578e+00
9.68500614e-01 -7.14961171e-01 -1.03822909e-02 8.55552316e-01
-6.75646961e-01 1.58959329e-01 5.66542864e-01 -1.50346011e-01
-1.04410648e+00 -4.16775703e-01 -6.41866803e-01 1.07624196e-01
-1.18026704e-01 8.55973542e-01 -2.68579155e-01 4.22832698e-01
6.51459575e-01 1.82814032e-01 -1.60408318e-01 -4.13112611e-01
1.34604752e-01 6.59001589e-01 1.77040353e-01 -3.51506501e-01
1.52027264e-01 1.17509775e-01 -2.07407519e-01 -1.62556678e-01
-1.29456949e+00 -5.64446688e-01 -1.93507493e-01 1.94274679e-01
1.12717259e+00 -7.24670172e-01 -7.63213158e-01 -1.30891144e-01
-1.21440458e+00 1.40711870e-02 -3.21863055e-01 4.42665845e-01
-3.53209823e-01 3.42217237e-01 -8.52086604e-01 -3.45927984e-01
-8.10442448e-01 -1.12911117e+00 1.09532833e+00 1.50852516e-01
-2.90320903e-01 -9.69899476e-01 1.31932601e-01 6.67802334e-01
4.18735623e-01 -1.60533741e-01 1.73832166e+00 -1.09311306e+00
-5.49740493e-01 -1.43513650e-01 -3.51681232e-01 -1.13594085e-01
2.85984963e-01 -5.62057555e-01 -5.51687896e-01 9.72400308e-02
4.54602651e-02 -1.76530808e-01 7.14152992e-01 7.47887418e-02
1.21584070e+00 -5.85329771e-01 -4.92856443e-01 3.16183358e-01
1.23303986e+00 3.82084697e-01 5.52593231e-01 1.88791782e-01
5.13146162e-01 5.97732186e-01 6.37476563e-01 1.74796999e-01
8.66744936e-01 4.74812418e-01 4.19679761e-01 2.69341469e-03
-5.97599968e-02 -3.58852684e-01 -2.58484334e-01 1.08483768e+00
1.05830140e-01 -3.62239748e-01 -1.21044457e+00 6.68524027e-01
-1.65916431e+00 -4.31736022e-01 3.85351516e-02 1.61427188e+00
1.32988572e+00 -1.49301767e-01 -2.20079586e-01 -1.11061260e-01
2.72986203e-01 -2.79539227e-01 -6.95122063e-01 -1.20160148e-01
1.69715583e-01 1.52226180e-01 2.18715236e-01 3.93058598e-01
-5.24121940e-01 5.69805205e-01 6.12338257e+00 6.70684397e-01
-5.73311806e-01 -1.98463611e-02 7.50150979e-01 2.14042425e-01
-7.19715655e-01 -1.28852174e-01 -9.92119133e-01 4.47946995e-01
8.12374115e-01 -1.27802774e-01 -7.74358064e-02 6.30100727e-01
-1.17414944e-01 3.76111940e-02 -1.17188692e+00 5.29411674e-01
9.48745310e-02 -1.38976479e+00 5.29683411e-01 -1.47121325e-01
4.44360226e-01 -2.82702327e-01 -3.60938698e-01 6.17550552e-01
3.21498990e-01 -1.01627100e+00 -1.12414509e-01 8.34473431e-01
4.37763840e-01 -3.16937715e-01 1.46917534e+00 4.48016614e-01
-8.98117542e-01 -1.96194991e-01 -3.61725092e-01 3.56695682e-01
1.86464250e-01 5.75114727e-01 -1.17119551e+00 9.30979252e-01
7.71033704e-01 3.41648877e-01 -5.58570266e-01 1.05648839e+00
-2.69393921e-01 6.94968283e-01 -5.22853322e-02 -2.07951531e-01
1.86419860e-01 1.45587116e-01 2.17750579e-01 1.04388714e+00
1.70385987e-01 8.46086323e-01 -8.67642909e-02 5.58980405e-01
-7.06418976e-02 4.17926878e-01 -1.37193024e-01 -6.14351965e-02
4.68993872e-01 9.53614116e-01 -3.37161809e-01 -5.59049606e-01
-5.25219321e-01 5.89586556e-01 3.83836091e-01 3.45267326e-01
-7.78595030e-01 -3.16127062e-01 3.27794284e-01 7.92128444e-02
3.22323740e-01 4.86492574e-01 -1.58348352e-01 -1.12070894e+00
8.87771025e-02 -1.27589500e+00 9.95625615e-01 -7.93730676e-01
-1.28024781e+00 9.32682455e-01 -2.04839215e-01 -8.43736112e-01
-3.17117065e-01 -4.42347854e-01 -3.75398517e-01 7.10086823e-01
-1.70628250e+00 -9.90441799e-01 -4.80791330e-01 6.57984972e-01
3.27116162e-01 1.11060236e-02 1.11357212e+00 4.17671233e-01
-4.28738564e-01 7.13625193e-01 -3.17367107e-01 3.41796875e-01
7.14322150e-01 -1.24825573e+00 -1.38593279e-02 8.10679719e-02
-2.57431328e-01 9.09625590e-01 4.70373929e-01 -5.60331821e-01
-1.08130240e+00 -1.02431464e+00 1.34423351e+00 -6.50192976e-01
3.09335798e-01 2.68810987e-01 -1.32088077e+00 6.71733320e-01
2.79472679e-01 -2.99632818e-01 1.17279589e+00 1.03951082e-01
-3.30586076e-01 -7.04416260e-02 -1.12254179e+00 5.95413864e-01
7.99791813e-01 -5.76791584e-01 -1.02091563e+00 3.11760575e-01
1.35775912e+00 -4.59759414e-01 -1.33321226e+00 8.03989589e-01
1.43378705e-01 -5.63589752e-01 8.94573331e-01 -1.19901001e+00
7.38126159e-01 -2.58861035e-01 -9.86424163e-02 -1.16296649e+00
-3.15182060e-01 -2.76033897e-02 -3.24381351e-01 1.02661467e+00
9.42589045e-01 -5.06823659e-01 5.83752692e-01 7.41399884e-01
-1.85564324e-01 -1.53590453e+00 -8.04550469e-01 -1.36011198e-01
1.21693268e-01 1.27607062e-01 9.78259861e-01 8.60882342e-01
4.78810698e-01 6.49636447e-01 9.14920866e-02 2.31147707e-01
-2.05126449e-01 4.57796633e-01 2.17762187e-01 -9.85189259e-01
-5.74604452e-01 -1.52721837e-01 -9.39456969e-02 -1.23910820e+00
-2.87018299e-01 -9.00451183e-01 -1.47064567e-01 -2.15057635e+00
5.66914976e-01 -4.23802704e-01 -2.69857496e-01 6.70765281e-01
-8.91064227e-01 -2.87950933e-01 -8.22485387e-02 1.64717287e-01
-8.97743583e-01 4.65880841e-01 1.65040743e+00 -2.65680552e-01
-4.57393602e-02 6.85878024e-02 -9.83475029e-01 5.40149152e-01
5.65500677e-01 -3.70610356e-01 -4.12767470e-01 -6.22968137e-01
6.10464513e-01 6.78213596e-01 4.72800992e-02 -6.23869300e-01
4.45703298e-01 2.72619613e-02 6.49593696e-02 -5.65111458e-01
1.16683267e-01 -9.71014202e-01 -2.30837300e-01 7.17984855e-01
-7.09665895e-01 1.60598725e-01 1.79736465e-01 6.39995992e-01
-5.42002141e-01 -3.44013542e-01 3.09261084e-01 -4.29498047e-01
-3.17755580e-01 2.26300374e-01 -1.64333120e-01 5.88196158e-01
7.26629734e-01 2.75581479e-01 -6.23947918e-01 -5.45178354e-01
-8.21417689e-01 7.31081486e-01 -1.29966587e-01 2.56483912e-01
5.25050402e-01 -1.02805746e+00 -7.93984830e-01 -1.94483459e-01
4.74958062e-01 3.03897619e-01 5.50433695e-01 9.57625329e-01
-5.64538777e-01 8.56882334e-01 2.92611588e-02 -5.31284332e-01
-1.19622326e+00 5.98251641e-01 4.86060083e-01 -9.59895372e-01
-5.18661618e-01 7.86952853e-01 2.34033197e-01 -5.90987980e-01
1.95747495e-01 -6.24514759e-01 -5.85512638e-01 -9.43368822e-02
6.96504772e-01 -3.09880618e-02 3.42683166e-01 -1.60502091e-01
-4.14965391e-01 3.82097125e-01 -4.57813829e-01 3.86320829e-01
1.06970167e+00 4.94238269e-03 -1.03814028e-01 -1.03554100e-01
8.23038816e-01 -8.58406052e-02 -4.87833738e-01 -3.94421041e-01
1.45920247e-01 -7.75330663e-02 -2.41097495e-01 -1.44056296e+00
-9.52234149e-01 8.12074184e-01 2.12666169e-01 2.37348080e-01
1.25971746e+00 3.88365805e-01 1.07407892e+00 5.03931403e-01
1.57102440e-02 -6.87833786e-01 1.44601345e-01 6.08425498e-01
7.56728947e-01 -1.15827262e+00 -1.94376484e-01 -7.57234573e-01
-8.94040346e-01 8.34476173e-01 8.50898147e-01 6.83319807e-01
5.60543299e-01 9.94737372e-02 2.90085703e-01 -5.46110451e-01
-1.15304124e+00 -2.05025434e-01 4.09615010e-01 4.67646480e-01
6.20018721e-01 -8.26897472e-02 -4.21768725e-01 1.12889254e+00
-1.77543670e-01 1.23395570e-01 2.47893631e-01 7.90651262e-01
-3.33289504e-01 -8.90780687e-01 -1.90463550e-02 8.92748535e-01
-8.34455192e-01 -4.87233549e-01 -4.10017818e-01 6.22542202e-01
8.31748694e-02 9.45443451e-01 -2.03733563e-01 -1.73955962e-01
5.53236127e-01 2.15979934e-01 9.38191712e-02 -6.65653110e-01
-8.47763360e-01 -3.58655512e-01 2.84083873e-01 -4.30341274e-01
-2.01687708e-01 -2.20647469e-01 -1.35007191e+00 8.01321343e-02
-5.39773166e-01 5.52102923e-01 5.83932064e-02 9.94973838e-01
8.43471408e-01 9.71924603e-01 2.17290688e-02 6.54996157e-01
-6.66493833e-01 -1.10047865e+00 -3.56078446e-02 3.48423392e-01
2.51585841e-01 -1.68852836e-01 -8.36905390e-02 -9.18741077e-02]
|
[8.769164085388184, 8.5711030960083]
|
13853259-bed4-4dd8-92c9-bfb922466ab9
|
driveability-constrained-models-for-optimal
|
2303.12603
| null |
https://arxiv.org/abs/2303.12603v1
|
https://arxiv.org/pdf/2303.12603v1.pdf
|
Driveability Constrained Models for Optimal Control of Hybrid Electric Vehicles
|
This work investigates the effect of three different driveability constraints on the optimal energy management strategy for a p2 parallel hybrid. Two of these constraints are used to prevent frequent gear shifting and engine start/stops, while the third is used to increase the sportiness of the vehicle by maximizing the available torque reserve at all times. The constraints are imposed by reformulating them as penalty terms to be added to the base running cost of the control strategy, which is fuel consumption. Dynamic programming, a popular optimal control technique, is then used to design the energy management strategy that minimizes the total cost. A case study is developed for a p2 parallel hybrid and simulated on a combination of the Artemis driving cycles. The impact of each driveability constraint is analyzed with respect to a set of relevant features of the control strategy, such as the choice of engine operating points and the gear shift pattern. The resulting discussion provides some useful insight for the design of real-time, rule-based control strategies.
|
['Daniela Misul', 'Federico Miretti']
|
2023-03-22
| null | null | null | null |
['energy-management']
|
['time-series']
|
[ 1.19739033e-01 2.38269180e-01 -5.50987959e-01 1.60985425e-01
3.90681744e-01 -4.67047006e-01 4.89475727e-01 2.33503833e-01
-2.86092758e-01 5.32316387e-01 -3.43757212e-01 -5.46666980e-01
-1.00777352e+00 -8.10130835e-01 -2.11644545e-01 -8.05079043e-01
1.93019062e-01 2.09877059e-01 3.07827204e-01 -3.03601354e-01
4.42219257e-01 9.81327832e-01 -1.85487282e+00 -7.11460769e-01
4.44369286e-01 8.93314242e-01 6.24021292e-01 4.75337744e-01
2.06306994e-01 2.85725236e-01 -4.38645661e-01 1.32318586e-01
3.62330019e-01 -4.09558564e-01 -3.80640090e-01 5.31148493e-01
-6.32845223e-01 -1.11487612e-01 1.12333342e-01 6.19636953e-01
3.98953915e-01 6.95347011e-01 6.24047637e-01 -1.60564506e+00
5.92952907e-01 -1.12605281e-01 -1.03595205e-01 4.17614877e-01
-2.62245446e-01 2.40583256e-01 6.19525075e-01 -1.13345101e-01
3.08526576e-01 8.36432159e-01 -1.50343448e-01 5.47140837e-02
-1.00992143e+00 -2.24099994e-01 -9.28731114e-02 3.56411129e-01
-1.34755492e+00 -4.71374512e-01 8.38295817e-01 -3.37689698e-01
1.08083689e+00 5.28029740e-01 9.40958023e-01 -1.01490624e-01
7.07949698e-01 -8.32752138e-02 9.87630844e-01 -6.89979076e-01
4.19222206e-01 8.83733109e-02 -1.95850894e-01 3.68234694e-01
4.67105031e-01 3.90440822e-01 8.23545605e-02 1.02992922e-01
3.17968905e-01 -2.56567806e-01 8.06053693e-04 -5.61453879e-01
-3.73433650e-01 7.40583003e-01 -1.27217457e-01 2.78938383e-01
-5.19351721e-01 -9.38828364e-02 2.85496265e-01 3.73757333e-02
4.48862836e-02 6.12046063e-01 -3.55150431e-01 -1.71489432e-01
-6.21518731e-01 5.20497680e-01 7.38908768e-01 6.42563462e-01
5.94906509e-01 1.52004659e-01 -1.60365507e-01 7.57766962e-01
3.87349278e-01 4.39623028e-01 -9.93541703e-02 -1.19408464e+00
2.32802853e-01 6.48778915e-01 4.21528965e-01 -7.99907267e-01
-3.59322935e-01 -1.86210826e-01 1.51979372e-01 3.60989749e-01
1.15215346e-01 -4.95208055e-01 -6.81841969e-01 1.19418609e+00
4.31508064e-01 -5.46025813e-01 -2.10479394e-01 7.93484509e-01
-2.28406653e-01 7.84996331e-01 3.65963429e-02 -6.25956357e-01
1.37676561e+00 -4.23074454e-01 -9.31961954e-01 -2.63713211e-01
4.52473432e-01 -7.82008469e-01 3.94423813e-01 8.66859108e-02
-1.24286628e+00 -2.45907381e-01 -1.32658792e+00 4.60071504e-01
-4.39367086e-01 3.55392992e-01 -1.03987999e-01 6.42406464e-01
-4.22798514e-01 6.90930068e-01 -7.62135029e-01 -1.23366848e-01
-5.77630699e-01 4.75366384e-01 2.62679100e-01 2.20006824e-01
-1.03573990e+00 1.54371440e+00 4.55684245e-01 4.08224672e-01
-4.67353970e-01 -4.18080896e-01 -6.08439744e-01 1.35647699e-01
9.17198241e-01 -3.15391898e-01 1.20272708e+00 -5.40806293e-01
-1.85989320e+00 3.09700638e-01 2.10631207e-01 -1.73039615e-01
5.36361754e-01 2.48641014e-01 -4.36339945e-01 1.61348000e-01
-6.55027246e-03 1.42862976e-01 4.44399089e-01 -8.76830518e-01
-7.65206039e-01 -7.82536715e-03 1.45810798e-01 6.11471176e-01
4.17323075e-02 2.81805508e-02 -3.87961566e-01 -8.31732601e-02
-6.47078753e-01 -1.51358902e+00 -2.80843943e-01 -5.37406087e-01
-1.84858352e-01 -5.19179940e-01 1.02999735e+00 -4.75145668e-01
1.28710008e+00 -2.04985547e+00 5.24790227e-01 6.08715296e-01
-6.27220511e-01 1.80538103e-01 3.82815272e-01 7.04627395e-01
5.59329204e-02 -1.15595795e-01 -1.21321850e-01 2.38654971e-01
-5.31145595e-02 6.79511428e-01 2.94677138e-01 2.62531787e-01
3.44334811e-01 -1.74761936e-02 -5.21827340e-01 8.06083530e-03
7.31790066e-01 1.68671057e-01 -1.35199606e-01 2.92358428e-01
-2.39128053e-01 -1.85422152e-01 -6.16789520e-01 -1.21827111e-01
2.47385100e-01 6.21473134e-01 3.69902194e-01 -4.49804291e-02
-8.66901934e-01 1.16399631e-01 -1.30662465e+00 6.16255879e-01
-7.00417578e-01 3.93478781e-01 3.74809295e-01 -9.97829556e-01
1.08765411e+00 1.59692436e-01 6.46648765e-01 -7.80481994e-01
4.43406999e-01 2.15828821e-01 2.84118533e-01 -6.66530728e-01
7.99039781e-01 -1.45613283e-01 6.97177798e-02 8.39745346e-03
-3.50654721e-01 -5.39673924e-01 9.00209844e-01 -3.16724360e-01
5.55755436e-01 6.00972585e-03 5.12417555e-02 -8.37198436e-01
7.41607845e-01 3.01708013e-01 7.31378853e-01 -2.76797861e-01
1.37829352e-02 -4.58453447e-01 9.10399020e-01 1.59154534e-01
-1.19516730e+00 -2.75727838e-01 -3.10589299e-02 6.63995326e-01
5.60714424e-01 -6.59657791e-02 -7.04169810e-01 -6.32595643e-02
1.34085998e-01 1.20402646e+00 -1.50880530e-01 -6.23973429e-01
-7.34462202e-01 -5.62369227e-01 -4.05331403e-01 3.21109116e-01
3.57211471e-01 -4.04350817e-01 -1.50237024e+00 2.11847693e-01
2.19656676e-01 -7.80800760e-01 -4.42592651e-01 3.04833025e-01
-6.40737236e-01 -1.39200830e+00 -2.25528121e-01 -3.81478786e-01
6.76066875e-01 1.71640053e-01 7.40952790e-01 4.04060006e-01
-1.41739741e-01 2.94947773e-01 -3.70815508e-02 -7.42938697e-01
-5.23839355e-01 1.00308210e-01 1.79237220e-03 -3.93291831e-01
1.13807470e-01 2.13028744e-01 -3.82775307e-01 8.46782386e-01
-7.81751692e-01 4.39551957e-02 2.74617463e-01 5.50265729e-01
7.34765649e-01 1.04931700e+00 5.34281671e-01 -2.39700839e-01
6.88259423e-01 -2.25120202e-01 -1.02876377e+00 1.80527419e-01
-8.42033207e-01 1.48412481e-01 5.11447132e-01 -3.03327024e-01
-1.05607462e+00 -4.72434424e-03 1.66563526e-01 -2.91935802e-01
1.21506810e-01 3.05933148e-01 -5.62207639e-01 -4.89644855e-02
-2.97037542e-01 9.20068473e-03 4.90182400e-01 -3.72136325e-01
-1.16053067e-01 3.13042670e-01 2.60717779e-01 -4.56396133e-01
7.32956290e-01 -1.27292439e-01 6.67408407e-01 -1.01994443e+00
-1.13559410e-01 -2.24112079e-01 -4.14551497e-01 -6.31888449e-01
7.64639139e-01 -3.52362275e-01 -6.77633941e-01 3.04559916e-01
-4.46577907e-01 -4.77564454e-01 -1.55958384e-01 4.38486487e-01
-7.87654400e-01 -1.13723084e-01 1.65656149e-01 -9.37393010e-01
2.84570549e-02 -1.27865887e+00 3.27856332e-01 3.99586827e-01
-2.02937126e-01 -8.68939221e-01 -1.01348512e-01 1.97786987e-01
3.55804414e-01 4.52503532e-01 1.08784103e+00 -1.07952081e-01
-3.30730110e-01 -3.49009521e-02 3.21177751e-01 5.37755013e-01
9.43301320e-02 3.01953137e-01 -4.15187068e-02 -4.40733403e-01
8.27129371e-03 4.08011407e-01 1.98052973e-01 3.29761744e-01
6.97490215e-01 -3.26138407e-01 -5.07842064e-01 -7.91861117e-02
1.76360834e+00 8.59565556e-01 3.56123388e-01 5.16928256e-01
1.43250093e-01 1.04977000e+00 1.18292296e+00 4.08116937e-01
1.70060083e-01 1.26550996e+00 6.26276076e-01 -6.50778189e-02
9.76206511e-02 6.17420748e-02 1.64617583e-01 1.79380670e-01
-2.33777508e-01 -1.46176338e-01 -6.30700171e-01 5.86763680e-01
-1.44643557e+00 -5.92571676e-01 2.63225175e-02 2.44921350e+00
2.90917963e-01 2.39718199e-01 3.60445440e-01 4.62215245e-01
8.34854305e-01 6.70511276e-02 -2.58989602e-01 -1.03416836e+00
4.99701500e-01 -3.31412286e-01 1.22171307e+00 4.05654401e-01
-5.25746644e-01 -3.29525657e-02 6.30428028e+00 7.51452327e-01
-1.12315786e+00 -2.86813080e-01 3.56079876e-01 -5.05438745e-01
-1.18422396e-01 2.93240279e-01 -8.06620836e-01 8.82224560e-01
1.35723567e+00 -7.32342005e-01 5.38562357e-01 5.63003778e-01
9.17080462e-01 -6.92246258e-01 -7.47080505e-01 8.19664374e-02
-3.22733730e-01 -9.87937868e-01 -5.25706768e-01 3.03533643e-01
3.34344596e-01 -3.77546787e-01 -4.31027055e-01 1.66066542e-01
-2.69438356e-01 -3.19809049e-01 1.03946733e+00 3.21453393e-01
3.65516245e-01 -1.58478749e+00 7.31347859e-01 3.35908502e-01
-1.17916930e+00 -5.47875166e-01 1.17194891e-01 -9.20381621e-02
6.25787735e-01 3.34892452e-01 -5.95389664e-01 5.16734004e-01
2.79429525e-01 -1.41200528e-01 -1.27123386e-01 8.54448915e-01
-5.95288798e-02 8.23369622e-02 -6.28479600e-01 -5.41052163e-01
2.66771227e-01 -4.87897903e-01 8.54503989e-01 6.40151262e-01
2.76308477e-01 2.62091644e-02 1.79537058e-01 8.26028049e-01
6.27656579e-01 -5.12572192e-02 -6.13975823e-01 -1.69877693e-01
7.09451795e-01 1.20350873e+00 -9.05096769e-01 2.69426033e-02
-1.14731364e-01 1.19199708e-01 -5.30740917e-01 1.80268139e-01
-9.29184318e-01 -6.65892243e-01 8.53504717e-01 7.42305338e-01
3.54122907e-01 -3.65836143e-01 5.71858399e-02 -2.71620974e-02
8.05239081e-02 -2.16115519e-01 3.31753455e-02 -5.60484529e-01
-4.95433331e-01 -1.73870221e-01 1.02898788e+00 -9.86416221e-01
-5.82184315e-01 -7.68606186e-01 -9.00850534e-01 1.01289999e+00
-1.55068159e+00 -4.25837189e-01 3.12092602e-01 -1.52784800e-02
5.56026578e-01 -5.16146570e-02 1.41585156e-01 3.03152412e-01
-1.14681900e+00 -1.85951710e-01 2.96512425e-01 -5.88143170e-01
-1.16800621e-01 -7.45040536e-01 -3.75389189e-01 1.02494347e+00
-1.07476914e+00 1.94714457e-01 1.16780281e+00 -6.28174007e-01
-1.80586445e+00 -7.79534876e-01 8.08075190e-01 3.49542499e-01
5.51523805e-01 -2.00569443e-02 -5.75691283e-01 -2.97780205e-02
2.77041644e-01 -5.79478145e-01 7.65240490e-02 -3.59438419e-01
9.06391263e-01 -1.97151721e-01 -1.09701359e+00 5.99366069e-01
1.97252095e-01 -7.68285543e-02 -1.72665164e-01 4.11312245e-02
3.46422732e-01 -3.96826088e-01 -1.05450344e+00 4.84883040e-01
4.08205718e-01 -3.34719181e-01 6.40800118e-01 -2.99707316e-02
-1.69764265e-01 -4.07260835e-01 3.66804957e-01 -1.48881698e+00
-2.73564070e-01 -6.32701397e-01 -3.29354629e-02 1.22000933e+00
3.08294833e-01 -5.27674317e-01 3.33492577e-01 9.85724032e-01
-2.89672971e-01 -1.07905293e+00 -1.26991701e+00 -9.73779559e-01
-1.50950864e-01 7.91825354e-02 5.12967765e-01 1.96246713e-01
-1.15601361e-01 1.75917730e-01 -6.12907223e-02 1.21814564e-01
2.51458198e-01 -2.08550721e-01 5.38231254e-01 -8.46720338e-01
-1.38623044e-01 -5.51928520e-01 -1.91341266e-01 -6.33861423e-02
1.49418274e-02 -1.74504742e-01 2.16816127e-01 -1.30889463e+00
-3.42046678e-01 -2.29880065e-01 1.87794380e-02 2.06829354e-01
1.60316780e-01 -4.80794013e-01 4.59046394e-01 -1.07764415e-01
2.62831271e-01 4.73048419e-01 9.13124442e-01 7.15751052e-02
-5.03220975e-01 3.75186205e-01 -2.37790689e-01 4.39955980e-01
7.93874741e-01 -3.84530991e-01 -8.69305611e-01 3.51564139e-02
2.23693386e-01 2.65636146e-01 7.65528381e-02 -7.88754165e-01
-1.08726338e-01 -6.80378914e-01 -4.36459184e-01 -6.36527359e-01
2.01543525e-01 -1.33707428e+00 7.76611567e-01 7.95876026e-01
-6.54326752e-02 4.09974605e-01 4.09892082e-01 5.26479006e-01
6.73441216e-02 -6.53488219e-01 9.54009473e-01 3.58333230e-01
-7.93245733e-01 -3.74311596e-01 -8.21138024e-01 -6.52741253e-01
1.80405796e+00 -4.18027073e-01 3.12735513e-02 1.78333506e-01
-3.77881914e-01 7.90377796e-01 4.71240401e-01 7.39139974e-01
-1.28308505e-01 -1.04052937e+00 -2.96133578e-01 1.17994174e-01
-1.62283823e-01 -4.56686884e-01 2.47999966e-01 8.83876860e-01
-5.22799075e-01 6.96637809e-01 -5.87893546e-01 -2.32283473e-01
-1.42992890e+00 5.40254474e-01 6.57341361e-01 -3.49118054e-01
-3.42219979e-01 -1.03145666e-01 -5.86775839e-01 3.04260939e-01
-2.96617359e-01 -1.85646504e-01 -3.66153032e-01 9.26654935e-02
1.83982059e-01 1.11053753e+00 4.20621902e-01 -7.16317236e-01
-3.98472458e-01 5.58020175e-01 5.70439339e-01 -1.82844505e-01
9.30236340e-01 -3.54724675e-01 3.74142081e-02 3.24998468e-01
6.48591161e-01 -1.85616165e-01 -1.32254493e+00 4.18298990e-01
1.04013532e-01 -4.21169579e-01 5.85694492e-01 -5.99112749e-01
-1.11502969e+00 2.03434646e-01 6.15587533e-01 5.70658326e-01
1.36796129e+00 -5.21733761e-01 2.74892598e-01 3.25096138e-02
2.93027490e-01 -1.76043582e+00 -5.81782997e-01 3.90471965e-01
7.21785128e-01 -3.01020950e-01 2.30588138e-01 -4.49736655e-01
-5.27391434e-01 9.80947196e-01 4.57037568e-01 -1.55506819e-01
5.50308585e-01 2.83688366e-01 -5.59367537e-01 -3.58998440e-02
-7.53785849e-01 -4.34646979e-02 1.83108419e-01 -3.87000255e-02
6.87747151e-02 2.38030776e-01 -1.19152725e+00 3.87962945e-02
2.68486619e-01 1.04877338e-01 4.85344350e-01 1.06553519e+00
-5.80558062e-01 -1.16849041e+00 -4.47461545e-01 3.55146855e-01
-2.34230697e-01 7.68729985e-01 -2.22863201e-02 1.24206066e+00
5.49473226e-01 1.03764904e+00 3.90808403e-01 -7.07673803e-02
9.18189287e-01 7.77739733e-02 2.21239015e-01 -4.00799483e-01
-4.27878648e-01 8.30167253e-03 7.77786076e-01 -1.85495690e-01
-3.26065361e-01 -7.17778087e-01 -1.29581773e+00 -3.82159591e-01
-7.69052505e-01 5.36463857e-01 1.10498548e+00 1.17499113e+00
3.96041065e-01 7.76835084e-01 7.80386686e-01 -8.53979349e-01
-4.83282626e-01 -6.45744741e-01 -7.95780838e-01 -1.36753306e-01
-1.42666146e-01 -1.36173201e+00 -3.70825917e-01 -3.69125068e-01]
|
[5.497677803039551, 2.0799129009246826]
|
50e0cd8e-0c02-47b7-8652-f46b722a23fa
|
diffdock-diffusion-steps-twists-and-turns-for
|
2210.01776
| null |
https://arxiv.org/abs/2210.01776v2
|
https://arxiv.org/pdf/2210.01776v2.pdf
|
DiffDock: Diffusion Steps, Twists, and Turns for Molecular Docking
|
Predicting the binding structure of a small molecule ligand to a protein -- a task known as molecular docking -- is critical to drug design. Recent deep learning methods that treat docking as a regression problem have decreased runtime compared to traditional search-based methods but have yet to offer substantial improvements in accuracy. We instead frame molecular docking as a generative modeling problem and develop DiffDock, a diffusion generative model over the non-Euclidean manifold of ligand poses. To do so, we map this manifold to the product space of the degrees of freedom (translational, rotational, and torsional) involved in docking and develop an efficient diffusion process on this space. Empirically, DiffDock obtains a 38% top-1 success rate (RMSD<2A) on PDBBind, significantly outperforming the previous state-of-the-art of traditional docking (23%) and deep learning (20%) methods. Moreover, while previous methods are not able to dock on computationally folded structures (maximum accuracy 10.4%), DiffDock maintains significantly higher precision (21.7%). Finally, DiffDock has fast inference times and provides confidence estimates with high selective accuracy.
|
['Tommi Jaakkola', 'Regina Barzilay', 'Bowen Jing', 'Hannes Stärk', 'Gabriele Corso']
|
2022-10-04
| null | null | null | null |
['blind-docking', 'molecular-docking']
|
['medical', 'medical']
|
[-1.90679342e-01 -7.35774636e-02 -1.11263938e-01 -2.03972921e-01
-1.15829957e+00 -8.15872967e-01 3.88097882e-01 7.30019882e-02
-4.91477787e-01 1.35255623e+00 -1.13483509e-02 -6.97274983e-01
5.13524823e-02 -6.26006246e-01 -1.13909805e+00 -1.15709662e+00
-3.61101925e-01 9.83642876e-01 -7.74660707e-02 -9.59611759e-02
2.79574811e-01 7.27988124e-01 -8.03476036e-01 1.11636490e-01
9.50630486e-01 2.70220131e-01 7.69253448e-03 4.61822808e-01
1.39664322e-01 1.81206837e-01 -6.00727499e-01 -4.34532046e-01
-1.07035078e-01 -5.93613029e-01 -8.64128947e-01 -6.42299473e-01
3.95140052e-01 -7.00375438e-02 -1.40724942e-01 8.57469082e-01
1.05563223e+00 2.92091090e-02 1.02871239e+00 -4.92510051e-01
-6.96938634e-01 -4.90272120e-02 -4.44479376e-01 1.04817264e-01
6.11235678e-01 2.10703418e-01 9.06493902e-01 -1.17359424e+00
1.05141127e+00 1.11676490e+00 7.81645298e-01 5.34524500e-01
-1.78328550e+00 -8.50041449e-01 -1.37292370e-01 -4.08495963e-02
-1.68249619e+00 -9.93430689e-02 -6.12379843e-03 -8.08606565e-01
1.58468843e+00 1.49166584e-01 5.94317436e-01 1.31758606e+00
1.02055061e+00 3.71734127e-02 7.74865329e-01 5.98792993e-02
6.49763346e-01 -4.95965660e-01 -2.25763664e-01 6.03653431e-01
9.34018493e-02 3.02042775e-02 -5.52570224e-01 -7.42100239e-01
7.61795104e-01 1.40813470e-01 -3.07335705e-01 -5.06825924e-01
-1.00899947e+00 1.21622908e+00 6.59322381e-01 -1.10894836e-01
-6.41764641e-01 2.10821390e-01 -2.97086220e-02 6.64066374e-02
3.83148968e-01 6.20444715e-01 -6.23521388e-01 -1.88830554e-01
-6.92200363e-01 7.42042482e-01 1.06774282e+00 5.88098109e-01
5.70930779e-01 -2.93733150e-01 2.25549370e-01 2.06611991e-01
5.29371321e-01 1.72798187e-01 5.96222207e-02 -4.49613214e-01
1.28169924e-01 2.62161344e-01 3.75381202e-01 -6.04720414e-01
-7.76832044e-01 -4.03671205e-01 -6.91744268e-01 4.83118087e-01
4.39821571e-01 -3.97058219e-01 -9.84968603e-01 1.65732586e+00
6.07926250e-01 -1.60893612e-02 1.10662945e-01 8.42085779e-01
7.54292250e-01 8.40741217e-01 3.04179102e-01 -4.95746464e-01
1.01899230e+00 -5.59279799e-01 -4.09237057e-01 2.99601585e-01
7.88258314e-01 -9.10410285e-01 6.12371385e-01 6.40750349e-01
-9.55098629e-01 -1.79056168e-01 -1.11547303e+00 -1.09775318e-03
-2.63867170e-01 -3.36826026e-01 8.26199353e-01 5.58357537e-01
-1.00342286e+00 9.88481641e-01 -1.08023381e+00 -9.36538726e-02
4.47927117e-01 1.16454613e+00 -6.46567523e-01 5.04944921e-02
-9.41391528e-01 1.17704368e+00 2.92645931e-01 -3.10954392e-01
-1.02183962e+00 -9.30908918e-01 -4.30154890e-01 -3.29803884e-01
8.91826823e-02 -8.56691360e-01 9.89611328e-01 -3.79880816e-01
-1.49058986e+00 4.66088742e-01 -4.74099219e-01 -3.09204996e-01
4.85785484e-01 -2.61204958e-01 2.15609018e-02 -3.25567573e-01
-8.82619619e-02 7.72422075e-01 2.35519633e-01 -9.56360996e-01
-7.03605637e-02 -4.46873039e-01 -1.39733225e-01 3.24065119e-01
4.09144163e-01 -9.14583057e-02 -1.67223319e-01 -3.90250772e-01
1.69937804e-01 -1.18017542e+00 -5.76729417e-01 -2.80573577e-01
-5.25484860e-01 -3.24791670e-01 1.55395970e-01 -5.11369944e-01
1.04987586e+00 -1.55248535e+00 8.32081318e-01 4.25550908e-01
5.77733278e-01 1.85450539e-01 5.34499362e-02 8.52754712e-01
-4.78878409e-01 -5.99893034e-02 -5.41257784e-02 -1.22972894e-02
-1.23001114e-01 9.79514420e-02 -1.13759272e-01 9.51050639e-01
3.47589031e-02 8.60106647e-01 -8.87973726e-01 6.07935302e-02
-3.37686166e-02 1.02594888e+00 -1.01105785e+00 -3.71377654e-02
-5.49914777e-01 7.14422584e-01 -5.90028167e-01 6.59415483e-01
8.25932741e-01 -5.10836959e-01 7.36422479e-01 -1.23329341e-01
-1.39214337e-01 4.30178672e-01 -8.73866677e-01 1.81931436e+00
1.23769566e-01 9.85447839e-02 -4.38113928e-01 -4.77333516e-01
8.80636334e-01 1.76467419e-01 7.50045478e-01 -4.50413436e-01
-1.78741068e-02 3.63822222e-01 3.94981146e-01 4.34810156e-03
6.54902905e-02 -3.33374798e-01 3.60605955e-01 1.29510194e-01
3.08789760e-01 3.12061340e-01 -2.03682914e-01 -2.09375229e-02
1.10034907e+00 5.45230806e-01 2.63226211e-01 -3.09144408e-01
5.11428900e-02 1.70159444e-01 4.31179315e-01 3.07912916e-01
3.38056684e-01 3.40386689e-01 5.90958118e-01 -8.38064551e-01
-1.03831899e+00 -8.70368838e-01 -4.66874599e-01 1.04991508e+00
-3.18191238e-02 -8.58751833e-01 -1.03891468e+00 -5.59704006e-01
3.76150161e-01 3.78584206e-01 -7.44823575e-01 -1.67591870e-01
-3.58535886e-01 -1.45346081e+00 3.83831114e-01 3.11403126e-01
-4.19673771e-01 -7.91416824e-01 -1.71631165e-02 7.46788383e-01
2.98597962e-01 -2.68511117e-01 -3.46258283e-01 9.24106181e-01
-6.87546909e-01 -1.28246725e+00 -7.34598279e-01 -6.21262491e-01
4.56889659e-01 -2.43721291e-01 1.14405727e+00 -2.89222896e-01
-4.91219401e-01 -3.67307037e-01 1.34519815e-01 -2.33836204e-01
-3.00275207e-01 4.08712253e-02 3.49697798e-01 -4.82374489e-01
8.22594047e-01 -6.93989396e-01 -9.62369382e-01 3.04443687e-01
-5.26337564e-01 -1.95351720e-01 3.97247493e-01 1.00158632e+00
1.21336627e+00 -3.65387946e-01 3.21179628e-01 -6.57845378e-01
6.27111852e-01 -5.44942915e-01 -6.85419083e-01 -2.84250490e-02
-6.61308289e-01 3.03265929e-01 3.49351823e-01 -5.06821334e-01
-4.25618172e-01 5.39578617e-01 -5.80931604e-01 -1.07054815e-01
3.55273783e-02 6.35985017e-01 -6.31185919e-02 -5.71341753e-01
9.51681674e-01 2.14105874e-01 6.90041855e-02 -7.99571693e-01
2.28801087e-01 2.10420281e-01 1.48194924e-01 -5.89196086e-01
4.90411311e-01 1.68123692e-01 4.35167134e-01 -5.11114717e-01
-5.52580297e-01 -2.41136372e-01 -8.46066415e-01 3.45776528e-01
1.11208844e+00 -9.36976373e-01 -1.43070257e+00 4.25205193e-02
-1.21936584e+00 -3.40331525e-01 4.55484271e-01 6.30200744e-01
-6.17238104e-01 4.83767241e-01 -5.88374794e-01 -3.00389796e-01
-3.06385189e-01 -1.65121841e+00 1.13215828e+00 -1.23441070e-01
-5.22079110e-01 -1.03174579e+00 6.59696162e-01 1.72425210e-01
1.72211409e-01 8.22754920e-01 1.07235026e+00 -8.43689978e-01
-5.99356830e-01 -1.23067662e-01 2.43366182e-01 -2.03184173e-01
5.94852194e-02 -2.91974796e-03 -6.42159104e-01 -4.82858986e-01
-3.57986391e-01 -3.46735328e-01 7.19455481e-01 6.76603258e-01
7.58327603e-01 -2.57470399e-01 -7.55230486e-01 8.86510372e-01
1.30761278e+00 5.11436701e-01 6.97655559e-01 4.49131876e-01
6.73030674e-01 -8.73050690e-02 4.88201380e-01 5.26741803e-01
1.05012849e-01 9.98076022e-01 7.39458501e-01 -3.12341392e-01
3.55969161e-01 -1.43280610e-01 3.17521423e-01 2.09601238e-01
-5.08098722e-01 -2.30793238e-01 -1.03061223e+00 -3.65641296e-01
-1.85413814e+00 -9.85134602e-01 -3.05295974e-01 2.33966470e+00
1.36849511e+00 9.60341375e-03 3.03148061e-01 -5.39240003e-01
2.10896984e-01 -3.28485817e-01 -1.04780805e+00 -2.11908385e-01
-1.22258082e-01 5.82695305e-01 5.70602417e-01 7.82841384e-01
-1.09241056e+00 1.05827427e+00 7.56157875e+00 7.70806730e-01
-1.12696302e+00 -1.46290213e-01 5.71922004e-01 -1.41392782e-01
-2.13346750e-01 -2.55128276e-02 -1.16632402e+00 4.30449605e-01
1.14353824e+00 3.96913104e-02 4.07995552e-01 8.83533120e-01
1.06270127e-01 6.38708994e-02 -1.40097690e+00 9.50667441e-01
-3.02211285e-01 -1.82203555e+00 2.51163870e-01 8.00863862e-01
5.37776291e-01 3.47844332e-01 1.93633810e-01 7.29475841e-02
4.15040761e-01 -1.83634543e+00 2.16727227e-01 6.03540480e-01
8.52979243e-01 -1.17994702e+00 5.68064988e-01 2.43894011e-01
-7.16410220e-01 5.58194637e-01 -7.35113919e-01 1.77305654e-01
-3.63730416e-02 2.63045639e-01 -1.03766060e+00 4.00063723e-01
3.67954969e-01 6.56456828e-01 -2.04595700e-01 9.69760001e-01
1.21074215e-01 2.84889221e-01 -3.67881864e-01 -1.72974497e-01
4.17349011e-01 -5.47685921e-01 2.00571746e-01 1.13344383e+00
1.24763556e-01 1.43252298e-01 3.56885970e-01 8.40296566e-01
-2.55683422e-01 9.80402827e-02 -3.52282435e-01 -3.05474200e-03
3.90161723e-01 7.18827069e-01 -3.73557389e-01 2.28745844e-02
-1.61721841e-01 1.07471752e+00 4.46925372e-01 3.82533640e-01
-1.06834638e+00 -3.63089859e-01 1.38354957e+00 1.06554799e-01
2.53455669e-01 -3.92345756e-01 3.78934771e-01 -8.96923959e-01
-3.45260173e-01 -1.01525199e+00 3.05430554e-02 -5.49306810e-01
-1.12455893e+00 5.94850659e-01 -4.24189657e-01 -7.94347405e-01
-1.55664146e-01 -1.02135122e+00 -2.39090160e-01 1.38658428e+00
-1.10215485e+00 -7.08942533e-01 2.31251180e-01 4.91505146e-01
1.93329468e-01 -1.60264269e-01 1.35182917e+00 2.42819011e-01
-4.57694650e-01 5.86906195e-01 8.52154493e-01 -4.21759665e-01
9.43248630e-01 -1.38116956e+00 6.42269731e-01 -2.07539443e-02
-1.41200513e-01 1.23883414e+00 8.31815839e-01 -9.25192833e-01
-1.99582148e+00 -9.84073162e-01 6.94902301e-01 -8.74639571e-01
6.00597203e-01 -3.00790519e-01 -1.02231407e+00 4.66545552e-01
-1.00077234e-01 -1.91495001e-01 1.37025106e+00 2.94894934e-01
-3.74167830e-01 5.36117911e-01 -9.22913730e-01 5.54423571e-01
9.54379261e-01 -3.35225463e-01 -2.49736354e-01 9.85863805e-01
6.36456132e-01 -8.78442883e-01 -1.26676464e+00 1.97561055e-01
5.69308877e-01 -8.24531853e-01 1.31908596e+00 -1.19936192e+00
-1.33829311e-01 -5.02181709e-01 -1.25612527e-01 -1.12654078e+00
-7.82546639e-01 -9.29559231e-01 -1.60769850e-01 2.75289029e-01
6.21892929e-01 -3.66886646e-01 9.60230172e-01 4.04214382e-01
-2.00586513e-01 -1.17478800e+00 -1.12569749e+00 -6.50527060e-01
5.36051631e-01 1.08738661e-01 4.62129861e-01 9.36645508e-01
9.12846252e-02 5.31784117e-01 -3.23492259e-01 1.63083792e-01
4.51526523e-01 -7.76444376e-02 7.82057941e-01 -1.43642056e+00
-3.80756021e-01 -2.04123989e-01 -4.96942669e-01 -1.01822102e+00
7.10225180e-02 -1.09029603e+00 -4.23363447e-01 -1.55391777e+00
4.50033993e-01 -3.90245430e-02 -9.35539529e-02 6.93736434e-01
1.88707467e-02 1.45630747e-01 -5.47111869e-01 3.62841129e-01
-5.84449530e-01 6.05775535e-01 1.13011026e+00 4.69199046e-02
-3.75870317e-01 -1.67155027e-01 -4.95396256e-01 4.20770705e-01
4.82251495e-01 -6.17779851e-01 5.40861785e-02 1.12362847e-01
4.86060977e-01 -8.14193413e-02 1.16698548e-01 -6.32840097e-01
-1.86670441e-02 -3.67273927e-01 7.92569518e-01 -6.39957249e-01
3.88690263e-01 -3.29427540e-01 7.19404519e-01 7.92157590e-01
1.88526437e-01 2.03204736e-01 2.66669512e-01 7.82613814e-01
1.24944367e-01 2.66912937e-01 7.87629604e-01 5.69303073e-02
-2.83586700e-02 5.04865766e-01 -5.64914167e-01 -2.94469744e-01
9.78764176e-01 -1.53529763e-01 -1.03843786e-01 -2.26357728e-02
-1.18548763e+00 -1.30574390e-01 7.24978149e-01 -4.71870229e-02
5.18876135e-01 -1.09820485e+00 -4.46879506e-01 -4.50707749e-02
-9.45477709e-02 -3.36706527e-02 -4.45318557e-02 7.12858260e-01
-1.02694261e+00 6.64533854e-01 -1.37283523e-02 -5.63436508e-01
-1.35261142e+00 5.72852850e-01 6.94346786e-01 -2.09783651e-02
-4.39997256e-01 9.92903590e-01 2.19096333e-01 -2.81358540e-01
3.32739830e-01 -1.42704651e-01 1.73641577e-01 3.56761850e-02
3.93040001e-01 -9.96109843e-03 3.98386717e-01 -5.46861768e-01
-6.97122395e-01 5.99478066e-01 -4.68322486e-01 4.51174915e-01
1.60442531e+00 7.39678562e-01 -1.18123591e-01 -1.86176136e-01
1.22394359e+00 -8.30187351e-02 -1.42875981e+00 1.04125671e-01
-4.23242934e-02 -3.79304811e-02 -5.91515675e-02 -1.00378788e+00
-3.24033707e-01 6.84212029e-01 7.40107954e-01 -4.62678730e-01
2.20035642e-01 -1.76076163e-02 4.23017293e-01 8.38997006e-01
4.21952754e-01 -5.40075421e-01 3.18891369e-02 6.01013660e-01
7.63066590e-01 -1.11301231e+00 2.04585090e-01 -8.31345692e-02
-4.16846424e-01 1.22184956e+00 2.84881741e-01 -9.78216156e-02
5.60142696e-01 2.46784344e-01 -1.19238809e-01 -5.19598305e-01
-7.28816330e-01 1.84927270e-01 4.87788022e-01 6.09211206e-01
1.02514088e+00 -6.42905384e-02 -2.12325215e-01 4.87923414e-01
-1.24185383e-01 -1.51917234e-01 6.50883913e-02 9.90394115e-01
-6.02861524e-01 -1.48862720e+00 -2.68069625e-01 -8.16933960e-02
-5.93714476e-01 -3.31889242e-01 -8.39067161e-01 6.87352359e-01
-5.02881035e-02 5.59391260e-01 -3.36654216e-01 -3.59271199e-01
2.34736592e-01 1.49986103e-01 7.12430537e-01 -6.21825695e-01
-5.82211554e-01 4.78272259e-01 -1.63854644e-01 -6.46081448e-01
-9.16283876e-02 -3.64740878e-01 -1.51510954e+00 -6.40884519e-01
-4.11534280e-01 6.13929749e-01 7.46551394e-01 7.13352680e-01
8.43410552e-01 3.29545617e-01 2.74739981e-01 -1.20071030e+00
-7.41014063e-01 -6.87441587e-01 -5.99637568e-01 -1.47538289e-01
3.43910784e-01 -8.63355875e-01 -1.12980357e-04 -1.26208350e-01]
|
[4.87790584564209, 5.597963333129883]
|
9e806fcf-47e1-48fa-846b-ca5745d0643c
|
part-of-speech-tagging-of-swedish-texts-in
| null | null |
https://aclanthology.org/2021.nodalida-main.20
|
https://aclanthology.org/2021.nodalida-main.20.pdf
|
Part-of-speech tagging of Swedish texts in the neural era
|
We train and test five open-source taggers, which use different methods, on three Swedish corpora, which are of comparable size but use different tagsets. The KB-Bert tagger achieves the highest accuracy for part-of-speech and morphological tagging, while being fast enough for practical use. We also compare the performance across tagsets and across different genres in one of the corpora. We perform manual error analysis and perform a statistical analysis of factors which affect how difficult specific tags are. Finally, we test ensemble methods, showing that a small (but not significant) improvement over the best-performing tagger can be achieved.
|
['Aleksandrs Berdicevskis', 'Yvonne Adesam']
| null | null | null | null |
nodalida-2021-5
|
['morphological-tagging']
|
['natural-language-processing']
|
[-2.30211057e-02 7.37726912e-02 1.80411572e-03 -2.87002325e-01
-1.49110985e+00 -1.05925810e+00 5.45464158e-01 3.37288171e-01
-8.72045815e-01 8.45225751e-01 4.89555866e-01 -4.43979710e-01
-9.72399712e-02 -3.66305977e-01 -3.02347809e-01 -6.41866446e-01
-1.75490618e-01 8.98700893e-01 7.49804795e-01 -1.09036170e-01
1.37257397e-01 -1.27953991e-01 -1.19501877e+00 3.78737420e-01
8.34812999e-01 4.91193473e-01 4.49363589e-01 5.99523604e-01
-3.25537026e-01 4.22087967e-01 -8.72542739e-01 -6.68876112e-01
-1.72016043e-02 -1.05901495e-01 -1.21507728e+00 -2.28840768e-01
4.03790623e-01 4.51429963e-01 1.49172485e-01 9.03259635e-01
6.66505098e-01 3.21586281e-02 5.23942351e-01 -3.81900400e-01
-2.95926064e-01 1.18291008e+00 1.32457567e-02 3.84748280e-01
4.34230387e-01 -4.60590005e-01 1.29473281e+00 -4.52088714e-01
6.91887856e-01 9.64892745e-01 1.13569295e+00 5.60208261e-01
-1.29391706e+00 -5.63500464e-01 -9.89124104e-02 -2.15854019e-01
-1.47601485e+00 -6.60525322e-01 -9.91283823e-03 -3.86609018e-01
1.18051159e+00 3.01868290e-01 4.25123215e-01 1.07992947e+00
3.52539659e-01 4.96023476e-01 1.38489711e+00 -6.64145410e-01
4.91739400e-02 2.53480412e-02 2.00141847e-01 5.05101681e-01
3.57613742e-01 -1.48639381e-01 -3.46817553e-01 -4.69528735e-01
3.42635274e-01 -5.38309157e-01 -1.12776361e-01 7.67260045e-02
-1.44112468e+00 7.54005134e-01 -2.63445407e-01 9.07817841e-01
-2.12926313e-01 -8.67785513e-02 6.42056108e-01 2.52667099e-01
8.76695871e-01 7.28805065e-01 -1.20344198e+00 -7.26545155e-01
-1.13220119e+00 -1.19701244e-01 1.18007731e+00 9.22428787e-01
3.87532234e-01 -1.16700061e-01 -4.44481149e-02 1.25499570e+00
2.09178537e-01 5.35506845e-01 8.35227668e-01 -6.55703902e-01
4.94954616e-01 2.27037862e-01 2.16830418e-01 -4.81295049e-01
-6.19552135e-01 -2.19436765e-01 -1.02431148e-01 -3.52085382e-01
6.29259348e-01 -3.04948688e-01 -1.11664712e+00 1.45242763e+00
-1.54362053e-01 1.11169875e-01 9.22727510e-02 3.51566344e-01
8.42200100e-01 5.61811328e-01 7.99962580e-01 -3.70322108e-01
1.61332715e+00 -6.08087063e-01 -6.74444854e-01 -6.37236059e-01
1.18740857e+00 -1.15619767e+00 8.15937579e-01 3.09224308e-01
-9.23537195e-01 -2.88036972e-01 -8.08710694e-01 1.76846161e-01
-4.94548023e-01 1.39088914e-01 7.50712693e-01 8.04541886e-01
-1.07922792e+00 8.33804607e-01 -8.91020536e-01 -8.65605712e-01
-2.20472127e-01 3.84981185e-01 -5.84325075e-01 2.24844486e-01
-1.15070713e+00 1.11141872e+00 7.88653314e-01 -4.65311885e-01
-3.35789800e-01 -2.64740795e-01 -7.78198183e-01 -1.69095963e-01
2.62479603e-01 -8.15623775e-02 1.55490673e+00 -7.23594069e-01
-1.21538436e+00 1.25928676e+00 9.59418565e-02 -3.86395633e-01
-2.15439126e-02 -8.43953714e-02 -9.73791480e-01 -1.99251488e-01
3.56574833e-01 3.80973577e-01 2.07259133e-01 -1.01719940e+00
-1.19366169e+00 3.19666527e-02 -2.86137015e-01 3.75387110e-02
-3.83023471e-01 8.10746670e-01 -3.15649390e-01 -7.85690427e-01
1.48963436e-01 -1.07075989e+00 -1.86209887e-01 -1.24770415e+00
-1.42091006e-01 -2.61404425e-01 1.25338331e-01 -1.07931638e+00
1.60264313e+00 -2.10620713e+00 -1.53145730e-01 -6.67288750e-02
-1.55146033e-01 3.73675853e-01 -1.07078418e-01 4.87200737e-01
1.49833068e-01 6.51460588e-01 -8.52409936e-03 -1.35941818e-01
8.87484327e-02 6.51859343e-01 2.61914283e-02 1.14947796e-01
-1.00892887e-01 8.14493239e-01 -1.06056464e+00 -7.82851100e-01
-1.57906041e-01 6.70926273e-02 -1.93782479e-01 -1.25620037e-01
1.94775283e-01 8.05090070e-02 -3.54498565e-01 5.27821302e-01
2.34214067e-01 1.82011709e-01 6.80397451e-01 8.58449638e-02
-3.16541910e-01 1.07605267e+00 -9.37320709e-01 1.59625864e+00
-6.13858104e-01 5.32335520e-01 -3.16106994e-03 -6.34765208e-01
8.41905236e-01 5.60883403e-01 4.48234499e-01 -1.71832025e-01
1.67554453e-01 6.24878526e-01 2.48498186e-01 -2.61315584e-01
8.24265003e-01 -3.95856678e-01 -7.58446455e-01 3.49542201e-01
4.82398838e-01 3.15992236e-02 4.84364867e-01 5.07466197e-02
1.53665924e+00 8.03718865e-02 5.71951807e-01 -7.84967601e-01
-2.00181529e-02 2.29111895e-01 8.28646183e-01 7.24756241e-01
-1.65160134e-01 4.15666223e-01 1.11932620e-01 -3.33160639e-01
-8.88721466e-01 -9.03531551e-01 -3.43891591e-01 1.60408986e+00
-2.80570209e-01 -7.20396876e-01 -7.39976287e-01 -1.05431354e+00
-2.14822695e-01 6.35693848e-01 -4.55959201e-01 1.80911660e-01
-5.78063965e-01 -9.04656708e-01 1.09845757e+00 6.48988605e-01
4.23187651e-02 -1.24053061e+00 -2.15708375e-01 5.52523375e-01
-5.98209500e-01 -1.38537252e+00 -4.46832180e-01 9.45019424e-01
-9.09348547e-01 -8.73371422e-01 -3.64200860e-01 -1.18991470e+00
2.20024467e-01 2.12947628e-03 1.65210390e+00 1.15845248e-01
4.04695332e-01 1.65566728e-01 -9.86687183e-01 -4.58899438e-01
-8.57683361e-01 6.68830752e-01 5.55813573e-02 -6.33965850e-01
8.49789381e-01 -2.53584027e-01 2.90909350e-01 5.69636047e-01
-4.78960544e-01 -3.87534559e-01 6.52785897e-01 8.44000697e-01
3.46973956e-01 9.31788906e-02 1.58484727e-01 -1.01830459e+00
6.10292792e-01 -3.97445619e-01 -1.48351118e-01 2.03411415e-01
-4.98723716e-01 1.07935429e-01 2.76292086e-01 -3.04723740e-01
-9.02247250e-01 8.74767900e-02 -4.43407804e-01 3.61099809e-01
-3.72663707e-01 4.96423811e-01 -2.31550470e-01 -2.18685213e-02
5.27532995e-01 -1.64535582e-01 -4.29389656e-01 -8.52143288e-01
1.52326515e-02 9.59320605e-01 4.65853810e-01 -6.30506575e-01
5.60305953e-01 -1.48020551e-01 -5.61208189e-01 -4.98839259e-01
-1.02987456e+00 -8.94804835e-01 -9.21867907e-01 1.96986824e-01
8.66836190e-01 -8.50912869e-01 -8.22414756e-02 3.33031535e-01
-9.03989553e-01 -6.31503761e-01 -1.98164567e-01 5.24131238e-01
-4.37136501e-01 2.88601607e-01 -8.45166683e-01 -4.19915915e-01
-2.03509331e-01 -7.02613413e-01 1.09927738e+00 -8.73380825e-02
-5.33579528e-01 -1.36177361e+00 2.35722557e-01 1.40080407e-01
-1.14121959e-01 -1.91520855e-01 5.37345767e-01 -1.37298965e+00
3.17454606e-01 -1.57002747e-01 2.05844492e-01 2.63870597e-01
2.35581562e-01 -8.33831087e-04 -7.47721732e-01 -2.76463747e-01
-5.28844774e-01 -9.58436504e-02 8.80199552e-01 3.35894853e-01
4.45020467e-01 -1.27074957e-01 -7.37509429e-01 2.12035164e-01
1.12772179e+00 3.30131829e-01 6.64724171e-01 7.42595911e-01
4.72396195e-01 3.89783084e-01 8.94388080e-01 1.19213447e-01
4.41829234e-01 9.38383579e-01 -3.05086911e-01 2.16210082e-01
8.03349242e-02 -3.01787853e-01 5.78213930e-01 1.45762920e+00
-3.32885981e-01 -4.40122455e-01 -1.12558651e+00 8.83359134e-01
-1.77810454e+00 -9.90357399e-01 -3.91404033e-01 2.19816113e+00
1.05333579e+00 3.57885152e-01 2.46551096e-01 2.48311684e-01
8.64691734e-01 -1.73608866e-02 4.91346478e-01 -7.51977265e-01
-1.93740770e-01 5.64441562e-01 9.02775049e-01 4.26332206e-01
-1.43184364e+00 1.46086013e+00 8.23303127e+00 1.01987588e+00
-6.60965860e-01 4.87956464e-01 1.47264525e-01 1.57452792e-01
-5.42843752e-02 4.21860255e-02 -1.00382602e+00 7.14102447e-01
1.61951280e+00 6.70062974e-02 8.68330747e-02 4.43975747e-01
1.32496597e-03 -2.32190549e-01 -6.29854083e-01 3.69139731e-01
-1.35819361e-01 -7.83686876e-01 -4.28261757e-01 4.01638210e-01
6.37458265e-01 3.84193540e-01 -5.80405653e-01 5.06533444e-01
9.99949634e-01 -4.59283173e-01 9.54739809e-01 1.70589060e-01
7.96571910e-01 -6.18233263e-01 1.11961401e+00 1.79866031e-01
-1.22296333e+00 4.00169939e-02 -5.37833869e-01 -4.47761361e-03
2.99689054e-01 5.78677416e-01 -7.83487201e-01 4.47495878e-01
9.55030560e-01 3.37292969e-01 -5.98999500e-01 1.05073941e+00
-5.02462149e-01 1.30288541e+00 -2.97889531e-01 -1.17620647e-01
2.97261417e-01 1.92342073e-01 4.96666789e-01 1.83752203e+00
6.07105255e-01 5.81192188e-02 3.36262614e-01 -1.45132035e-01
1.57420084e-01 3.68442148e-01 -3.17757696e-01 -1.78427204e-01
9.39481735e-01 1.45766759e+00 -9.59449232e-01 -3.72969449e-01
-4.06001031e-01 9.34218466e-01 2.86037534e-01 -1.64575428e-01
-6.53040111e-01 -4.22507495e-01 7.37944126e-01 1.46104053e-01
5.39845407e-01 -3.56267601e-01 -7.32417479e-02 -9.60438430e-01
-3.09348077e-01 -6.45208478e-01 8.12467575e-01 -5.45504272e-01
-1.14954162e+00 6.37659371e-01 -6.47458360e-02 -1.14451814e+00
-5.76500952e-01 -8.43694746e-01 -3.22084308e-01 6.98038280e-01
-1.04338229e+00 -1.14442170e+00 1.77472591e-01 1.43061787e-01
3.03187460e-01 -2.46163368e-01 1.28133392e+00 4.78195667e-01
-3.23185414e-01 6.72726393e-01 3.74586642e-01 4.26682770e-01
1.02573979e+00 -1.59357786e+00 7.95711637e-01 7.19170034e-01
5.68393052e-01 5.61647058e-01 5.87832153e-01 -9.39056754e-01
-6.25227332e-01 -1.12100768e+00 1.66457391e+00 -7.36201346e-01
9.68388438e-01 -3.83984029e-01 -7.49357462e-01 8.98652196e-01
1.72624081e-01 -1.75570443e-01 9.76170480e-01 8.81459355e-01
-2.95038342e-01 2.53636301e-01 -1.07340419e+00 2.72235349e-02
1.32049763e+00 -1.86790764e-01 -9.35612619e-01 1.62392914e-01
4.76493329e-01 -3.50658208e-01 -1.28376079e+00 2.06307024e-01
5.86493611e-01 -7.27382123e-01 5.39568305e-01 -6.00000799e-01
-2.75960602e-02 -4.37935768e-03 -1.73018873e-01 -1.90201068e+00
-9.54260588e-01 -6.40178800e-01 5.30988514e-01 1.85077500e+00
8.06579053e-01 -7.10541248e-01 3.78729165e-01 1.65445194e-01
-6.13036156e-01 -1.68997616e-01 -1.06479812e+00 -1.23997140e+00
2.85991699e-01 -5.13387501e-01 5.24523377e-01 1.06895399e+00
5.38877785e-01 3.56776685e-01 -2.31758267e-01 -4.06839885e-02
7.79490322e-02 -4.13912743e-01 2.28058785e-01 -1.65084362e+00
-2.41628885e-01 -3.25897783e-01 -7.34457850e-01 -6.68795824e-01
3.00830394e-01 -8.44558120e-01 5.21592975e-01 -1.47205544e+00
3.55938047e-01 -6.60772860e-01 -4.99414712e-01 1.06001770e+00
-2.67372549e-01 7.42616951e-01 1.34666905e-01 3.64036560e-01
-7.54477441e-01 -3.30094397e-01 5.95205247e-01 5.92351612e-03
2.33251099e-02 -2.65615314e-01 -8.24042022e-01 7.46326864e-01
8.26764226e-01 -7.93980896e-01 3.27165544e-01 -5.22151768e-01
2.33910367e-01 -4.33894157e-01 -1.42244935e-01 -1.16381907e+00
-3.03311199e-01 -4.50998582e-02 7.48904347e-02 -6.40070587e-02
-6.44022450e-02 -5.18282533e-01 2.20332339e-01 6.65883794e-02
-2.20136307e-02 2.29912773e-01 4.84623760e-01 1.06784351e-01
-3.47494483e-02 -6.92915916e-01 6.02504075e-01 -2.76628613e-01
-8.97338212e-01 -1.29774928e-01 -8.11628461e-01 3.18909168e-01
6.35277331e-01 -2.46977806e-01 -1.01240732e-01 5.32237589e-02
-8.48768115e-01 -2.82553762e-01 5.25594413e-01 4.20653343e-01
-3.29706281e-01 -1.32616484e+00 -7.46732354e-01 -2.23706678e-01
3.82710487e-01 -7.67561257e-01 -3.91416550e-01 6.52504861e-01
-2.94815481e-01 5.28596759e-01 -3.68774012e-02 -2.20190719e-01
-1.45292425e+00 2.19772860e-01 1.84690461e-01 -7.09632635e-01
-2.15430796e-01 7.29436696e-01 -4.11634237e-01 -5.85634589e-01
-2.66126543e-01 -2.47244462e-01 -3.41694474e-01 2.78479397e-01
2.94923604e-01 4.08055156e-01 4.96769994e-01 -1.01497257e+00
-6.07256353e-01 5.77023029e-01 8.57793242e-02 -3.28338295e-01
1.13720238e+00 -8.02828446e-02 9.65155754e-03 6.27474785e-01
6.87509894e-01 5.28266668e-01 -5.56827426e-01 -7.16230273e-02
4.48324919e-01 -1.91013291e-01 1.69268385e-01 -1.03355229e+00
-6.53876662e-01 3.74433756e-01 3.61023337e-01 7.56743431e-01
7.63316870e-01 2.43432224e-01 8.32375646e-01 9.01016593e-02
7.92361915e-01 -1.42466640e+00 -7.51075447e-01 9.05782044e-01
2.02604964e-01 -9.51893389e-01 -1.05488993e-01 -4.59960103e-01
-6.15577996e-01 6.39155805e-01 1.89016253e-01 2.11618766e-01
4.97395009e-01 5.96235156e-01 4.86680984e-01 -9.27596539e-02
-6.68829918e-01 -6.70165896e-01 3.33021492e-01 7.25023568e-01
1.00798118e+00 5.31586826e-01 -1.02306497e+00 8.00370157e-01
-5.94329774e-01 -3.93682212e-01 3.05561244e-01 8.89115930e-01
-6.71616495e-01 -1.58520246e+00 -3.68176579e-01 6.51388109e-01
-1.10909247e+00 -3.75802845e-01 -6.32325232e-01 8.88501644e-01
3.65294337e-01 1.38528192e+00 8.07572603e-02 -7.46860027e-01
1.36992440e-01 5.83710849e-01 5.73956072e-01 -1.02294314e+00
-8.45870972e-01 2.45235592e-01 9.65086699e-01 -2.43644625e-01
-6.15615904e-01 -1.21959674e+00 -1.13094771e+00 -2.69319236e-01
-8.23507249e-01 8.01562250e-01 4.57153767e-01 1.11480594e+00
5.31107038e-02 2.20706388e-01 2.80619621e-01 -5.96292078e-01
-3.28364521e-01 -1.68900943e+00 -9.37514007e-01 3.99484217e-01
-5.64281464e-01 -7.67630398e-01 -3.11200827e-01 2.15785235e-01]
|
[10.329988479614258, 10.046083450317383]
|
922e494f-b92c-4db5-979a-57d39c3e0af1
|
deeplung-deep-3d-dual-path-nets-for-automated
|
1801.09555
| null |
http://arxiv.org/abs/1801.09555v1
|
http://arxiv.org/pdf/1801.09555v1.pdf
|
DeepLung: Deep 3D Dual Path Nets for Automated Pulmonary Nodule Detection and Classification
|
In this work, we present a fully automated lung computed tomography (CT)
cancer diagnosis system, DeepLung. DeepLung consists of two components, nodule
detection (identifying the locations of candidate nodules) and classification
(classifying candidate nodules into benign or malignant). Considering the 3D
nature of lung CT data and the compactness of dual path networks (DPN), two
deep 3D DPN are designed for nodule detection and classification respectively.
Specifically, a 3D Faster Regions with Convolutional Neural Net (R-CNN) is
designed for nodule detection with 3D dual path blocks and a U-net-like
encoder-decoder structure to effectively learn nodule features. For nodule
classification, gradient boosting machine (GBM) with 3D dual path network
features is proposed. The nodule classification subnetwork was validated on a
public dataset from LIDC-IDRI, on which it achieved better performance than
state-of-the-art approaches and surpassed the performance of experienced
doctors based on image modality. Within the DeepLung system, candidate nodules
are detected first by the nodule detection subnetwork, and nodule diagnosis is
conducted by the classification subnetwork. Extensive experimental results
demonstrate that DeepLung has performance comparable to experienced doctors
both for the nodule-level and patient-level diagnosis on the LIDC-IDRI
dataset.\footnote{https://github.com/uci-cbcl/DeepLung.git}
|
['Wei Fan', 'Wentao Zhu', 'Chaochun Liu', 'Xiaohui Xie']
|
2018-01-25
| null | null | null | null |
['lung-nodule-classification']
|
['medical']
|
[-1.16232544e-01 4.84853864e-01 -5.97579539e-01 -4.34349589e-02
-9.77109969e-01 -2.15558961e-01 3.12444299e-01 -2.87832081e-01
-1.28645167e-01 2.26500019e-01 2.38173395e-01 -6.96495831e-01
-3.44805866e-02 -8.66668284e-01 -4.96875644e-01 -6.54143989e-01
-5.92396781e-02 8.38226736e-01 7.66554236e-01 3.18837315e-01
-4.51191247e-01 5.71373641e-01 -9.59898591e-01 6.05149448e-01
4.21898127e-01 1.23626876e+00 5.67308187e-01 7.94590235e-01
1.74206734e-01 1.21028113e+00 2.14730963e-01 -2.35576034e-02
4.83605206e-01 -5.08197725e-01 -9.16008770e-01 1.84914544e-01
4.44695726e-02 -6.31739914e-01 -6.64114356e-01 6.62114382e-01
4.74948555e-01 -6.67455971e-01 8.84171963e-01 -8.51543128e-01
-3.62536609e-01 4.74339962e-01 -3.85285646e-01 2.46395975e-01
-3.85088354e-01 2.55886793e-01 1.02190924e+00 -1.12698734e+00
4.31061119e-01 5.40092468e-01 9.05071616e-01 6.61603332e-01
-5.17344773e-01 -4.81673270e-01 -4.38498259e-01 -1.80196925e-03
-1.33092773e+00 2.96154618e-01 8.29650685e-02 -5.72138071e-01
5.79115152e-01 4.54202861e-01 1.08138943e+00 8.02968860e-01
4.66512352e-01 8.01456094e-01 6.76231742e-01 -3.44291143e-02
-1.16410121e-01 1.40683070e-01 -2.87099302e-01 1.43317378e+00
4.06182438e-01 2.34002531e-01 2.55298644e-01 -2.68912345e-01
1.41449022e+00 5.05096853e-01 -2.75532722e-01 -4.59333032e-01
-1.52314234e+00 8.69736075e-01 1.20037532e+00 4.56694603e-01
-5.54724991e-01 2.52171874e-01 3.59233558e-01 -2.11823329e-01
1.65648535e-01 -6.15100702e-03 -2.41165161e-01 4.51654315e-01
-7.97623992e-01 -2.61729330e-01 8.14383209e-01 7.68945634e-01
2.61673778e-01 -3.24205905e-01 -8.83111119e-01 6.77276611e-01
4.42808241e-01 4.76677060e-01 8.44448090e-01 -4.22351211e-01
5.58138900e-02 8.10568571e-01 -3.42989862e-01 -3.33040416e-01
-7.08853960e-01 -1.00102699e+00 -1.31105328e+00 -1.09958157e-01
8.16004425e-02 -4.99206893e-02 -1.22100711e+00 9.39966142e-01
2.54488587e-01 1.74264669e-01 -1.96104661e-01 1.00282407e+00
1.34211254e+00 1.26905560e-01 -1.36828333e-01 2.84497440e-01
1.89003694e+00 -1.38851535e+00 6.44388422e-03 -4.25329618e-02
9.68646765e-01 -4.96354848e-01 7.58572996e-01 -2.06914693e-01
-9.72244442e-01 -4.64919090e-01 -6.02510691e-01 -2.50468664e-02
1.42274022e-01 9.59351242e-01 3.94527346e-01 4.29087281e-01
-1.21913409e+00 1.75808907e-01 -1.13310993e+00 -4.90376264e-01
7.23712862e-01 5.00012219e-01 -1.64921850e-01 -1.75027192e-01
-9.42011356e-01 8.20216477e-01 2.85653561e-01 -7.03606978e-02
-1.36365187e+00 -9.11200285e-01 -5.18706858e-01 2.49315232e-01
4.39987510e-01 -1.12542284e+00 1.71826339e+00 -5.09128094e-01
-1.02883661e+00 1.08847737e+00 8.90048742e-02 -5.99309027e-01
6.64625883e-01 4.33861434e-01 -5.49157448e-02 4.25525606e-01
3.72813642e-01 9.96836543e-01 6.97723925e-01 -6.66143954e-01
-9.17749643e-01 -1.01225965e-01 -3.13493133e-01 1.96550474e-01
-1.28282890e-01 -4.44497913e-01 -8.31578076e-01 -5.39494216e-01
2.42089555e-01 -1.15804148e+00 -5.08328617e-01 4.05285925e-01
-7.20657706e-01 -1.91683337e-01 8.94979060e-01 -6.56608701e-01
1.08893538e+00 -1.96327424e+00 -2.67531425e-01 4.06626880e-01
7.84262836e-01 2.62731314e-01 1.56812012e-01 -1.43549830e-01
-1.80927008e-01 1.93156749e-01 5.58152869e-02 1.02102436e-01
-2.79488593e-01 6.15480505e-02 5.96292138e-01 3.91850680e-01
2.71040678e-01 1.48961294e+00 -6.74477518e-01 -8.40506375e-01
1.70661494e-01 3.29154193e-01 -6.13697588e-01 2.17875972e-01
-1.39712282e-02 3.39915752e-01 -7.02828526e-01 1.07306886e+00
2.51609802e-01 -1.15398979e+00 5.62679097e-02 -3.20683479e-01
2.67350245e-02 2.25861847e-01 -4.29165453e-01 1.14179087e+00
-3.58299404e-01 2.66044945e-01 1.17319606e-01 -4.77889210e-01
5.54450929e-01 8.65344107e-01 7.91008413e-01 -3.37368190e-01
2.15442955e-01 5.87072849e-01 5.20415902e-01 -5.52708924e-01
-3.15388680e-01 -7.37946257e-02 2.50192255e-01 4.06730294e-01
-1.29362106e-01 -1.41566902e-01 2.97035519e-02 -4.86093983e-02
1.76334631e+00 -4.68235582e-01 6.60745084e-01 -2.46287614e-01
5.23025155e-01 4.61247504e-01 2.75862992e-01 7.21044719e-01
-3.88478339e-01 8.55117202e-01 5.99201083e-01 -2.68915504e-01
-8.76416385e-01 -1.15928805e+00 -3.38482022e-01 6.15633547e-01
-2.54656136e-01 -6.69825897e-02 -4.20901358e-01 -1.20220196e+00
1.01271920e-01 -1.81949828e-02 -9.31249976e-01 -1.07383942e-02
-5.21093428e-01 -6.41050756e-01 5.98852873e-01 8.53150606e-01
8.33746314e-01 -1.00557387e+00 -6.35555446e-01 1.78746790e-01
-1.80807561e-01 -8.82307649e-01 -6.15336239e-01 5.66649079e-01
-1.03051353e+00 -1.53518689e+00 -1.13757825e+00 -1.12364221e+00
9.81929302e-01 4.97478604e-01 1.13838506e+00 3.64377022e-01
-9.19990003e-01 3.50383818e-01 -1.26107007e-01 -3.59012872e-01
-5.63331604e-01 3.50324512e-01 -4.69666809e-01 -4.48495001e-01
1.24789052e-01 -1.48912534e-01 -1.03292418e+00 5.07285535e-01
-6.71786129e-01 3.50864589e-01 1.64806247e+00 1.02735519e+00
8.04991722e-01 -5.14678098e-02 2.20293686e-01 -9.67800200e-01
6.19418267e-03 -7.83520699e-01 -2.68475294e-01 1.56016290e-01
-1.67775691e-01 -3.65226746e-01 2.97576129e-01 -1.02193862e-01
-6.88449681e-01 4.69078720e-01 -3.66900027e-01 -5.05232871e-01
-5.94910653e-03 3.93051594e-01 4.60069835e-01 -9.91038606e-02
7.60635018e-01 2.01378167e-01 1.85017660e-01 -6.23547472e-02
-2.62175977e-01 6.65256202e-01 2.72074401e-01 1.89653426e-01
8.86034548e-01 5.50359547e-01 1.98220432e-01 -4.68708038e-01
-1.15048373e+00 -8.46827924e-01 -8.32719564e-01 -2.72675484e-01
1.13647306e+00 -1.22299802e+00 -2.75537640e-01 5.82791716e-02
-6.73978806e-01 -2.95368463e-01 -5.68488061e-01 7.61285424e-01
-4.30094659e-01 -1.71626762e-01 -9.36712325e-01 -7.48914853e-02
-6.50563657e-01 -1.30651975e+00 1.16452038e+00 -4.27874830e-03
-5.59334867e-02 -9.25754666e-01 -1.09134771e-01 3.39829832e-01
5.81886232e-01 -1.23821393e-01 1.03573120e+00 -8.72501135e-01
-9.05760884e-01 -5.28979838e-01 -6.32810891e-01 2.54143506e-01
3.77183199e-01 -1.30613163e-01 -7.94679046e-01 -1.93734661e-01
-1.68733269e-01 -2.30410233e-01 9.91548121e-01 7.67144442e-01
1.35297894e+00 -9.56497807e-03 -8.78662646e-01 5.87850153e-01
1.34906876e+00 -4.20048088e-02 2.79197186e-01 1.09383427e-01
7.61066079e-01 -1.48360357e-01 2.85109460e-01 3.90366077e-01
1.71744153e-01 1.57681718e-01 9.45818126e-01 -6.35235906e-01
-5.55284977e-01 -1.44615233e-01 -9.87408385e-02 6.18286014e-01
-1.23916171e-01 -1.31576583e-01 -1.16727102e+00 6.11862421e-01
-1.59286773e+00 -5.61546981e-01 -2.81730205e-01 1.80614591e+00
5.08312523e-01 1.06361575e-01 2.75000315e-02 -1.84953719e-01
6.54582143e-01 -1.43965870e-01 -5.86575985e-01 3.53484541e-01
4.04451907e-01 1.32245913e-01 7.11375475e-01 -2.11800665e-01
-1.39504826e+00 3.41498911e-01 5.68790770e+00 9.76842642e-01
-1.20255625e+00 4.46224749e-01 9.72725630e-01 -7.68817142e-02
1.24285243e-01 -5.70958853e-01 -7.25988209e-01 5.47690950e-02
4.66005981e-01 1.50991559e-01 -3.79000545e-01 1.16497123e+00
1.40295029e-01 5.59970178e-02 -1.06862473e+00 5.14530838e-01
-2.13632122e-01 -1.71475911e+00 -7.56626055e-02 3.76117915e-01
6.39614463e-01 7.85484135e-01 1.12037905e-01 5.82650602e-01
2.76050717e-01 -1.07942343e+00 8.87086019e-02 2.42367283e-01
1.11869383e+00 -3.00526410e-01 1.17715454e+00 5.01816630e-01
-1.48533618e+00 -1.53908402e-01 -4.11853999e-01 6.23981357e-01
-3.90220612e-01 6.78316593e-01 -1.72235620e+00 5.57775676e-01
7.92374313e-01 8.28398287e-01 -7.49795556e-01 1.20669174e+00
-7.75003210e-02 9.60150242e-01 -3.59420657e-01 -7.28233084e-02
5.09394884e-01 2.80868739e-01 2.64622509e-01 1.05854666e+00
6.74690008e-01 1.33480996e-01 2.48574138e-01 8.42976987e-01
-3.87279779e-01 2.91072167e-02 -4.39705819e-01 1.59254700e-01
2.49422684e-01 1.70776594e+00 -9.78639424e-01 -3.15762043e-01
-6.44926906e-01 7.19760716e-01 3.64871249e-02 -2.08402932e-01
-1.07890141e+00 2.35078722e-01 1.25257913e-02 6.77306712e-01
6.89158261e-01 5.26596785e-01 -3.89601700e-02 -8.37364733e-01
-1.40003920e-01 -4.10952240e-01 5.63793719e-01 -6.38118744e-01
-1.30293059e+00 7.46018708e-01 -4.44266707e-01 -1.73553133e+00
-7.45556876e-02 -8.31961334e-01 -9.60936844e-01 4.50659871e-01
-1.42864203e+00 -1.29844058e+00 -5.79089105e-01 5.67907035e-01
3.54836404e-01 -1.97844282e-01 6.51156247e-01 1.60433695e-01
-4.08511519e-01 3.76444072e-01 -1.20505085e-02 6.31144166e-01
4.79214966e-01 -1.30593920e+00 1.47060081e-01 2.48233542e-01
-3.46040577e-01 -2.25963131e-01 -4.12484854e-01 -8.08040082e-01
-1.21333683e+00 -1.98003471e+00 5.95737994e-01 -3.07195067e-01
6.54094398e-01 1.23910680e-01 -6.11173630e-01 7.77274251e-01
-1.84781373e-01 6.01351142e-01 5.78616679e-01 -6.97194874e-01
7.62274116e-02 1.67862073e-01 -1.20436060e+00 4.70371395e-01
8.58631551e-01 -2.23343194e-01 -6.72731474e-02 8.53485107e-01
6.08294785e-01 -6.49525225e-01 -9.89486873e-01 8.40941906e-01
3.96555275e-01 -1.00379634e+00 1.00542402e+00 -5.81911057e-02
6.08084738e-01 -1.49089858e-01 1.46381035e-01 -8.51355493e-01
-8.62829447e-01 2.64084250e-01 3.55791785e-02 2.68987536e-01
7.69195020e-01 -3.60949546e-01 1.27083564e+00 -8.83055106e-03
-3.78685653e-01 -1.50254285e+00 -8.32430124e-01 -4.67318684e-01
-4.10734862e-02 -1.21941760e-01 1.45681277e-01 3.80184352e-01
-4.41829652e-01 6.77688569e-02 2.03487068e-01 2.33235136e-01
3.23707879e-01 5.76166138e-02 1.08481720e-01 -1.06000590e+00
-5.09981275e-01 -6.23072386e-01 -3.64593565e-01 -1.03504837e+00
-4.46289837e-01 -1.36627948e+00 -3.16498689e-02 -1.94187295e+00
7.18180239e-01 -4.18547630e-01 -2.64558852e-01 6.72970235e-01
-1.21108405e-01 5.12599885e-01 -7.36096129e-02 6.10080481e-01
-5.47328293e-01 2.13102564e-01 1.85495472e+00 -1.63334697e-01
1.87796503e-01 7.51909852e-01 -4.34879035e-01 8.41425002e-01
8.37175727e-01 -4.63257045e-01 -1.57690570e-01 -1.00827642e-01
-2.69274741e-01 4.83279049e-01 7.66374767e-01 -1.29641092e+00
3.03417861e-01 2.92833120e-01 8.07463944e-01 -1.21845961e+00
3.46728295e-01 -9.54033911e-01 -1.76429655e-02 1.39280379e+00
-9.47666466e-02 -2.32919842e-01 -8.14049840e-02 5.36104143e-01
-2.28424504e-01 -1.84574723e-01 8.41744542e-01 -6.85758591e-01
-2.57134527e-01 9.77168143e-01 -7.41089344e-01 -2.27725267e-01
1.27036631e+00 -1.30498156e-01 -7.59557337e-02 -1.75093338e-02
-8.63293707e-01 3.65599990e-01 3.16501893e-02 -1.61508217e-01
6.36216521e-01 -1.42579925e+00 -1.03788507e+00 1.82015434e-01
1.47531554e-01 4.66976553e-01 1.86144322e-01 1.51893830e+00
-8.19441199e-01 8.70549321e-01 2.24244848e-01 -1.06811082e+00
-1.25974953e+00 1.87075227e-01 9.73574758e-01 -9.47388530e-01
-7.92792618e-01 1.12396324e+00 5.98872244e-01 -7.47521937e-01
4.07808691e-01 -7.51970589e-01 9.50657390e-03 -3.16311836e-01
8.56503546e-02 6.84348419e-02 3.90962601e-01 -7.52719119e-02
-3.03683192e-01 2.66879350e-01 -2.63885379e-01 3.75777543e-01
9.90102887e-01 3.08209062e-01 -1.87765256e-01 -1.30420737e-03
1.01708186e+00 -4.67569351e-01 -8.74157667e-01 -4.94572788e-01
-1.25937060e-01 1.04024045e-01 1.41701087e-01 -8.83006394e-01
-1.39149320e+00 6.77739680e-01 7.22560823e-01 1.50682122e-01
1.06750762e+00 4.80186909e-01 8.26453924e-01 5.60923576e-01
-1.03151597e-01 -2.34515220e-01 3.00053805e-01 5.62937260e-01
7.24764764e-01 -1.58994269e+00 -1.70433987e-02 -7.35311806e-01
-7.00243294e-01 1.24362719e+00 9.45953667e-01 -1.47177741e-01
1.24185431e+00 3.54092062e-01 4.74418253e-02 -5.95428228e-01
-1.04089057e+00 -4.73027825e-01 5.35911918e-01 2.56172270e-01
5.52837372e-01 3.89619797e-01 2.36335829e-01 7.49740362e-01
5.01692183e-02 1.25459969e-01 2.67410666e-01 8.80333483e-01
-7.55535781e-01 -6.36443853e-01 -3.19172442e-01 1.23698437e+00
-4.02860135e-01 -2.22286925e-01 -3.64909232e-01 1.32945943e+00
3.00032079e-01 3.24173868e-01 1.08973369e-01 -2.27146342e-01
1.24113962e-01 -2.84721315e-01 1.14842810e-01 -1.01672280e+00
-9.72631514e-01 3.08730483e-01 -3.14402804e-02 -3.94187301e-01
-1.10562228e-01 -4.21272069e-01 -1.35502315e+00 9.94487703e-02
-3.07438254e-01 1.46381721e-01 3.06468487e-01 6.52112544e-01
1.84578121e-01 9.82802212e-01 7.41506517e-01 -6.00364685e-01
-6.77993834e-01 -1.03518987e+00 -6.24182999e-01 -4.30773050e-01
1.42117247e-01 -2.73964286e-01 -3.03581119e-01 -2.28315577e-01]
|
[15.414968490600586, -2.116024971008301]
|
224e8b45-f4c0-4444-bc96-59cacd30cebe
|
videoglue-video-general-understanding
|
2307.03166
| null |
https://arxiv.org/abs/2307.03166v1
|
https://arxiv.org/pdf/2307.03166v1.pdf
|
VideoGLUE: Video General Understanding Evaluation of Foundation Models
|
We evaluate existing foundation models video understanding capabilities using a carefully designed experiment protocol consisting of three hallmark tasks (action recognition, temporal localization, and spatiotemporal localization), eight datasets well received by the community, and four adaptation methods tailoring a foundation model (FM) for a downstream task. Moreover, we propose a scalar VideoGLUE score (VGS) to measure an FMs efficacy and efficiency when adapting to general video understanding tasks. Our main findings are as follows. First, task-specialized models significantly outperform the six FMs studied in this work, in sharp contrast to what FMs have achieved in natural language and image understanding. Second,video-native FMs, whose pretraining data contains the video modality, are generally better than image-native FMs in classifying motion-rich videos, localizing actions in time, and understanding a video of more than one action. Third, the video-native FMs can perform well on video tasks under light adaptations to downstream tasks(e.g., freezing the FM backbones), while image-native FMs win in full end-to-end finetuning. The first two observations reveal the need and tremendous opportunities to conduct research on video-focused FMs, and the last confirms that both tasks and adaptation methods matter when it comes to the evaluation of FMs.
|
['Boqing Gong', 'Ting Liu', 'Ming-Hsuan Yang', 'Hartwig Adam', 'Florian Schroff', 'Huisheng Wang', 'Mikhail Sirotenko', 'Luke Friedman', 'Tobias Weyand', 'Menglin Jia', 'Xuan Yang', 'Lu Jiang', 'Yin Cui', 'Hao Zhou', 'Long Zhao', 'Nitesh Bharadwaj Gundavarapu', 'Liangzhe Yuan']
|
2023-07-06
| null | null | null | null |
['temporal-localization', 'action-recognition-in-videos', 'video-understanding']
|
['computer-vision', 'computer-vision', 'computer-vision']
|
[ 4.03845966e-01 -3.12382430e-01 -5.42609274e-01 -3.34826440e-01
-6.80129945e-01 -5.70611298e-01 4.41002876e-01 -6.22916341e-01
-5.22935748e-01 4.72689897e-01 5.20006239e-01 -1.62893966e-01
2.52671149e-02 -1.36514768e-01 -1.14117408e+00 -4.81396407e-01
-2.70809621e-01 -1.06994053e-02 3.32320869e-01 -8.11024383e-02
2.92841375e-01 1.20885514e-01 -1.55692732e+00 7.93293715e-01
6.17026269e-01 1.12877584e+00 5.13000488e-01 9.71879005e-01
5.36788881e-01 1.28309262e+00 -3.90438557e-01 -2.47190341e-01
2.18714908e-01 -3.28451663e-01 -9.62522030e-01 2.58256316e-01
9.45741773e-01 -8.70268166e-01 -7.80930400e-01 5.85340858e-01
2.47941628e-01 4.92222041e-01 3.54398757e-01 -1.44384873e+00
-8.90682340e-01 4.55649167e-01 -2.27793247e-01 6.78378105e-01
5.75468183e-01 7.47094274e-01 8.59803498e-01 -8.73324573e-01
6.50242627e-01 1.30741906e+00 6.90962851e-01 7.83874452e-01
-7.80843318e-01 -6.75467312e-01 5.58207452e-01 6.38702989e-01
-1.15130186e+00 -8.86590421e-01 2.30065316e-01 -4.47217971e-01
1.22143126e+00 9.02912095e-02 4.97836649e-01 1.66494095e+00
2.74151951e-01 1.09475338e+00 6.82020962e-01 -1.05990479e-02
1.25047728e-01 -1.31974459e-01 -1.12725109e-01 5.52967191e-01
-1.36666253e-01 1.66476250e-01 -1.05363655e+00 3.84613901e-01
8.90859187e-01 4.30682786e-02 -5.98785877e-01 -1.12865992e-01
-1.53570199e+00 3.96621019e-01 9.83881280e-02 2.58194000e-01
-4.18278188e-01 5.30178070e-01 7.75028169e-01 5.68640769e-01
3.45197797e-01 3.25825304e-01 -7.15866923e-01 -6.92502439e-01
-9.53565180e-01 7.81222656e-02 4.16436076e-01 1.04815137e+00
6.60616338e-01 2.36258745e-01 -4.09009457e-01 3.06138098e-01
-1.34893790e-01 4.83793914e-01 6.68701589e-01 -1.68316102e+00
6.41417563e-01 9.27594006e-02 -1.37403549e-03 -8.57400179e-01
-2.86086172e-01 -1.88871384e-01 -5.16240895e-01 -1.97181806e-01
2.08703682e-01 -4.58685718e-02 -8.47468138e-01 2.10613775e+00
-1.72933370e-01 6.60299659e-01 1.66715577e-01 9.91328716e-01
7.02087343e-01 5.72507560e-01 3.92825902e-01 -2.01832905e-01
1.18860269e+00 -1.43623841e+00 -5.48154950e-01 -4.19917405e-01
8.67855191e-01 -3.46365422e-01 1.46824634e+00 3.37738752e-01
-1.07146907e+00 -1.08761883e+00 -7.87787020e-01 1.37588172e-03
-7.58028850e-02 4.94650453e-02 8.67407441e-01 3.48341465e-01
-1.46208823e+00 4.53498989e-01 -9.38665509e-01 -6.91715181e-01
4.16473508e-01 1.93228230e-01 -6.79553092e-01 -2.75189549e-01
-1.12836981e+00 6.74909711e-01 4.84959215e-01 -1.93354726e-01
-1.65590036e+00 -8.63479733e-01 -7.67554164e-01 1.99734673e-01
5.49903035e-01 -1.06969357e+00 1.31280017e+00 -1.55006242e+00
-1.37142301e+00 7.00484037e-01 -2.34325528e-01 -7.07150161e-01
4.72975224e-01 -4.65884119e-01 -5.39105654e-01 7.64773428e-01
3.55495095e-01 1.00774038e+00 1.17369878e+00 -8.93974185e-01
-8.41413200e-01 -1.96389720e-01 5.86755455e-01 3.82918626e-01
-5.94952285e-01 -1.54814452e-01 -6.52318239e-01 -7.22109854e-01
-3.77292126e-01 -8.43539655e-01 5.07924855e-02 -1.41270250e-01
5.04014967e-03 -4.01125802e-03 9.96324718e-01 -7.86132216e-01
1.35712707e+00 -2.36512232e+00 2.96539426e-01 -3.53703886e-01
2.15681568e-01 3.61514568e-01 -5.38315117e-01 3.24485064e-01
-2.62226492e-01 1.90686882e-01 2.07481146e-01 -1.89284623e-01
-1.46061718e-01 1.30785882e-01 -1.84008077e-01 2.33092755e-01
9.65558141e-02 1.14487326e+00 -8.93697381e-01 -3.47984403e-01
1.98638424e-01 1.79693684e-01 -8.65369022e-01 2.29740098e-01
-3.00046325e-01 6.28786266e-01 -4.72861201e-01 7.49705076e-01
2.10694939e-01 -5.12989044e-01 5.02646975e-02 -4.69227761e-01
2.52436638e-01 -2.40261972e-01 -7.11150885e-01 1.93205988e+00
-4.51521158e-01 8.18473101e-01 -6.87987730e-02 -1.10374820e+00
8.91589671e-02 5.12102306e-01 5.81982255e-01 -8.87810051e-01
-3.40737849e-01 -2.09992081e-01 -2.12936938e-01 -1.01802337e+00
5.28307855e-01 2.35064611e-01 8.98402855e-02 3.64932060e-01
3.72780144e-01 4.42088008e-01 3.98752958e-01 5.96695244e-01
1.39137685e+00 4.86939490e-01 7.25359991e-02 -7.31082773e-03
5.54916561e-01 -4.03600000e-02 3.80365491e-01 9.54106331e-01
-5.29959440e-01 4.97401237e-01 2.78545290e-01 -4.64143783e-01
-7.80353189e-01 -9.39369857e-01 4.06204969e-01 1.80546355e+00
2.40103289e-01 -6.71646059e-01 -9.57737148e-01 -8.44851494e-01
-1.51329443e-01 5.49148023e-01 -7.92674243e-01 -4.89116907e-01
-5.92831850e-01 -1.94629744e-01 6.78985357e-01 1.02732384e+00
8.11595678e-01 -1.02182829e+00 -6.83088243e-01 8.13630968e-03
-7.85354257e-01 -1.61194289e+00 -9.12757218e-01 -2.79242396e-01
-8.32785726e-01 -1.16638637e+00 -6.15249932e-01 -5.78705430e-01
3.30571681e-01 8.35461497e-01 1.27105582e+00 7.63951391e-02
1.41996682e-01 1.23601425e+00 -6.68974042e-01 1.83442935e-01
-2.61350155e-01 -3.24887559e-02 2.22476438e-01 1.06364347e-01
3.48345786e-01 -4.91275936e-01 -8.09113920e-01 6.20755732e-01
-1.07851589e+00 9.34596136e-02 7.70850599e-01 6.31211400e-01
3.60322952e-01 -2.44678423e-01 6.00445271e-01 -5.49322963e-01
2.46649787e-01 -5.14657855e-01 5.15171327e-02 5.13990998e-01
-3.16753507e-01 -2.26536781e-01 6.88019454e-01 -6.26058817e-01
-1.07630074e+00 -2.00140879e-01 5.09067476e-02 -9.68804479e-01
-3.45004827e-01 3.77376497e-01 -7.92536512e-02 -2.38516092e-01
6.51767075e-01 4.72429931e-01 -3.31437588e-02 -1.56376123e-01
3.65668476e-01 2.78766990e-01 7.60622084e-01 -6.86130881e-01
5.13411403e-01 5.46935141e-01 -4.41703260e-01 -9.95148480e-01
-8.02288949e-01 -5.67042649e-01 -5.27073741e-01 -5.37924588e-01
1.15034020e+00 -1.31556857e+00 -7.85392106e-01 5.04961312e-01
-9.66232121e-01 -6.51880801e-01 -1.85073048e-01 5.16097426e-01
-1.17650199e+00 6.33919537e-01 -6.70353413e-01 -2.57925838e-01
5.82160577e-02 -1.26476276e+00 1.21378529e+00 4.98220362e-02
-2.07607642e-01 -1.08335197e+00 -2.54514456e-01 8.13292086e-01
5.26320219e-01 -1.08122945e-01 6.09631479e-01 -3.08951706e-01
-7.67530560e-01 1.08134784e-01 -3.49202216e-01 5.81149757e-01
-1.80098623e-01 -1.67482436e-01 -8.54839861e-01 -5.25220573e-01
4.03024517e-02 -6.96697652e-01 1.16530228e+00 6.89288259e-01
1.33951199e+00 -2.36382782e-01 -2.56501198e-01 1.01424825e+00
1.09028769e+00 2.56960630e-01 7.62866080e-01 4.37305510e-01
7.07236350e-01 1.11787945e-01 8.11956823e-01 3.94774288e-01
4.72594023e-01 7.95103073e-01 3.69381547e-01 1.55148402e-01
-3.97989661e-01 -3.59816879e-01 1.17517436e+00 4.82240140e-01
-4.85488713e-01 -4.88226235e-01 -5.23715079e-01 3.19402039e-01
-2.13836432e+00 -1.22114646e+00 3.49278450e-01 1.89024496e+00
3.12406510e-01 -5.11604100e-02 2.37477019e-01 -4.69393611e-01
4.15772349e-01 4.05952841e-01 -7.51769960e-01 -5.94634600e-02
-2.66045094e-01 -1.39344111e-01 4.65941846e-01 1.46710545e-01
-1.26492882e+00 1.30931067e+00 6.78923178e+00 9.02024448e-01
-1.03134441e+00 1.98951840e-01 6.83242738e-01 -1.48370028e-01
2.07323339e-02 3.30592459e-03 -5.72413921e-01 5.58898807e-01
1.22363293e+00 -5.88655919e-02 5.69564104e-01 7.58674681e-01
5.60162961e-01 -1.33364331e-02 -1.40337372e+00 9.70357537e-01
3.28014135e-01 -1.35858941e+00 5.25777042e-01 -2.15064868e-01
6.95492744e-01 3.20545100e-02 3.55960250e-01 6.18178964e-01
-1.39353842e-01 -1.07193208e+00 8.44626904e-01 5.53746700e-01
1.04744864e+00 -2.03726396e-01 5.66265702e-01 1.99440449e-01
-1.20507550e+00 -3.93817186e-01 -1.88845634e-01 -1.17062949e-01
3.65277022e-01 -1.30498782e-01 -2.34483257e-01 4.14668769e-01
9.69955385e-01 1.10318029e+00 -6.91884100e-01 5.08790851e-01
4.09864485e-02 7.65156329e-01 8.51226151e-02 6.12717211e-01
5.53988516e-01 1.20232210e-01 4.88180399e-01 1.34512222e+00
1.90822810e-01 7.32746720e-02 3.45441669e-01 3.13456684e-01
-7.96439312e-03 -2.06910849e-01 -5.64680755e-01 -4.15431321e-01
1.70093179e-01 7.69824147e-01 -4.28758144e-01 -6.07195199e-01
-6.18371248e-01 1.30276108e+00 1.34078473e-01 9.00411189e-01
-1.18775392e+00 2.02578247e-01 8.57451618e-01 1.54073611e-01
4.53658998e-01 -2.22264305e-01 2.96262383e-01 -1.55144155e+00
4.09405045e-02 -1.24636853e+00 4.71774846e-01 -1.06569648e+00
-9.59628761e-01 4.79722440e-01 1.67486116e-01 -1.19944406e+00
-4.68741506e-01 -7.55307734e-01 -3.67189646e-01 7.41433203e-02
-1.32698011e+00 -1.24036789e+00 -5.03590167e-01 1.03194952e+00
1.13540292e+00 -2.66696870e-01 5.79472542e-01 4.71366584e-01
-5.63410282e-01 7.30663955e-01 -1.56605944e-01 6.74940497e-02
1.02310038e+00 -6.99528635e-01 2.63179302e-01 1.02101314e+00
2.08374754e-01 5.26595652e-01 5.52991390e-01 -4.58127290e-01
-1.57655454e+00 -1.05912280e+00 3.53358299e-01 -6.08788252e-01
6.76060975e-01 -2.41826437e-02 -6.62884235e-01 1.16692412e+00
2.12700322e-01 2.39711367e-02 4.23144192e-01 -8.05167407e-02
-4.89694715e-01 5.72542697e-02 -7.44446397e-01 5.39597034e-01
1.66214836e+00 -6.57739520e-01 -3.33560884e-01 5.54833829e-01
9.42025661e-01 -2.83322871e-01 -8.24831426e-01 4.84115273e-01
6.80265546e-01 -1.34150934e+00 1.18041313e+00 -1.16027117e+00
7.11936593e-01 -7.28100613e-02 -4.77501690e-01 -1.07620120e+00
-4.90261763e-01 -5.87084353e-01 -3.53705198e-01 7.81426251e-01
1.29945129e-01 -2.95978963e-01 7.94049919e-01 4.54438329e-01
-3.76711756e-01 -4.87910956e-01 -6.26337409e-01 -1.04791152e+00
-2.71080106e-01 -5.90264976e-01 8.00157562e-02 8.87709200e-01
-2.01960623e-01 2.03885540e-01 -6.22308016e-01 -4.57314104e-02
3.20340693e-01 -1.72746673e-01 9.27962542e-01 -5.22461355e-01
-5.25258601e-01 -3.83810729e-01 -5.97314775e-01 -1.54508376e+00
4.61096734e-01 -5.37659049e-01 -2.26433560e-01 -1.17879105e+00
4.06198472e-01 2.18989030e-01 -4.11166817e-01 4.68866974e-01
-2.21579120e-01 3.38247716e-01 4.80818838e-01 2.79981434e-01
-1.33767986e+00 4.00426537e-01 1.22025096e+00 -1.89163461e-01
8.43499750e-02 -2.34554317e-02 -7.55245388e-01 8.64391506e-01
4.55991030e-01 9.88192186e-02 -7.91262567e-01 -7.22292960e-01
-1.30411863e-01 2.44043171e-01 4.93613422e-01 -1.15438604e+00
2.23139480e-01 -2.96575069e-01 2.98397511e-01 -6.85641617e-02
3.26240271e-01 -7.31585979e-01 6.31243736e-02 2.36067474e-01
-3.98989350e-01 2.16594324e-01 2.68983483e-01 8.87625754e-01
-5.01131654e-01 1.47805959e-01 5.88450730e-01 -3.68192583e-01
-1.80500829e+00 4.96092945e-01 -4.22414571e-01 3.31690550e-01
1.16093242e+00 -7.47753024e-01 -3.78462672e-01 -8.77373338e-01
-9.07154858e-01 2.55450845e-01 5.35567582e-01 6.86180592e-01
7.58614600e-01 -1.03103447e+00 -6.90869272e-01 1.51813775e-01
2.70624101e-01 -6.46286190e-01 8.45655441e-01 1.09958136e+00
-4.02390420e-01 5.79578161e-01 -3.67383718e-01 -8.18261147e-01
-1.04441881e+00 6.68150365e-01 1.69584766e-01 -7.05513954e-02
-6.17360175e-01 1.07189810e+00 9.58365381e-01 1.51408315e-01
3.90133202e-01 -2.60718077e-01 3.99541184e-02 -1.83399305e-01
6.87193990e-01 3.66503417e-01 -1.89226314e-01 -6.92120671e-01
-3.82493734e-01 5.77380002e-01 -1.56816706e-01 1.41168922e-01
1.12402701e+00 -4.70465869e-01 3.57219905e-01 1.76263511e-01
1.17290974e+00 -4.95121390e-01 -1.69878244e+00 -5.08467108e-02
-2.57714212e-01 -5.03872991e-01 -1.57147095e-01 -8.35904062e-01
-1.13844967e+00 7.78988600e-01 4.21117961e-01 -2.31004715e-01
1.50843120e+00 -4.30364814e-03 7.47437894e-01 3.93036842e-01
3.20643932e-01 -1.12362182e+00 5.38246155e-01 5.74570417e-01
6.96497381e-01 -1.22900796e+00 -2.32811838e-01 -7.25639984e-02
-1.11651659e+00 9.94183064e-01 9.48806882e-01 1.22914270e-01
3.11970979e-01 1.81597006e-02 -1.73427612e-01 -3.53627242e-02
-1.17712677e+00 -1.37609348e-01 2.08040535e-01 6.69772983e-01
2.15211883e-01 -2.27147266e-01 1.62618384e-01 4.52774167e-01
3.07153404e-01 4.03858781e-01 3.09902936e-01 6.05813801e-01
-3.55305761e-01 -4.77093548e-01 1.05731590e-02 3.07139993e-01
-3.93065274e-01 -1.27488732e-01 -1.50713205e-01 9.60468531e-01
3.56247425e-02 9.63664711e-01 2.18836535e-02 -7.26831794e-01
1.98526740e-01 2.51804627e-02 5.07962227e-01 -4.08542454e-01
-3.51959229e-01 -2.73547530e-01 1.15988255e-02 -1.38823879e+00
-6.83488429e-01 -6.21283889e-01 -8.46517742e-01 -4.09989834e-01
4.22886424e-02 -3.72890122e-02 2.20241502e-01 1.08562315e+00
5.75595021e-01 4.98417467e-01 3.46369565e-01 -1.06867766e+00
-4.59552497e-01 -7.67617822e-01 -3.48155409e-01 7.15891540e-01
3.77470613e-01 -6.41639650e-01 -4.34259325e-01 5.39844632e-01]
|
[9.276544570922852, 0.7750512361526489]
|
6ce8a106-a9e5-4b61-8550-ea2938b43ea1
|
rpg-learning-recursive-point-cloud-generation
|
2105.14322
| null |
https://arxiv.org/abs/2105.14322v1
|
https://arxiv.org/pdf/2105.14322v1.pdf
|
RPG: Learning Recursive Point Cloud Generation
|
In this paper we propose a novel point cloud generator that is able to reconstruct and generate 3D point clouds composed of semantic parts. Given a latent representation of the target 3D model, the generation starts from a single point and gets expanded recursively to produce the high-resolution point cloud via a sequence of point expansion stages. During the recursive procedure of generation, we not only obtain the coarse-to-fine point clouds for the target 3D model from every expansion stage, but also unsupervisedly discover the semantic segmentation of the target model according to the hierarchical/parent-child relation between the points across expansion stages. Moreover, the expansion modules and other elements used in our recursive generator are mostly sharing weights thus making the overall framework light and efficient. Extensive experiments are conducted to demonstrate that our proposed point cloud generator has comparable or even superior performance on both generation and reconstruction tasks in comparison to various baselines, as well as provides the consistent co-segmentation among 3D instances of the same object class.
|
['Wei-Chen Chiu', 'Li-Heng Wang', 'Chen-Yi Chiu', 'Yu-Liang Kuo', 'Hui-Yu Huang', 'Wei-Jan Ko']
|
2021-05-29
| null | null | null | null |
['point-cloud-generation']
|
['computer-vision']
|
[ 1.71520725e-01 2.97309577e-01 2.41752028e-01 -2.31824443e-01
-8.67148876e-01 -6.02438509e-01 7.70712793e-01 1.38956144e-01
3.25239569e-01 1.59443870e-01 -2.76955426e-01 4.64907512e-02
8.29988569e-02 -1.22109544e+00 -7.81375945e-01 -4.76845145e-01
1.75152019e-01 1.30595660e+00 5.48586130e-01 -1.60044245e-02
3.70818406e-01 1.08114254e+00 -1.85245693e+00 7.05134030e-03
1.12029219e+00 1.07714248e+00 6.28988445e-01 2.54811943e-01
-5.80787063e-01 -1.50466964e-01 -4.38746423e-01 -4.28053826e-01
5.43139696e-01 -1.57530848e-02 -8.40861261e-01 7.74814665e-01
4.43718880e-01 -1.87087283e-01 2.34736294e-01 1.04916978e+00
1.60685495e-01 2.95198150e-02 7.99544692e-01 -1.26212835e+00
-1.87837660e-01 2.67102718e-01 -8.27665389e-01 -6.17980003e-01
2.97163129e-01 -1.04581736e-01 9.66190398e-01 -1.27756119e+00
7.34733641e-01 1.35062528e+00 5.87225556e-01 3.18003774e-01
-1.17266250e+00 -6.49317324e-01 2.26082131e-01 -2.58756667e-01
-1.42879438e+00 -2.21554656e-02 1.01032317e+00 -4.19862896e-01
7.87372947e-01 1.00824013e-01 9.95754838e-01 6.00749254e-01
-2.38631859e-01 7.01746345e-01 5.87954283e-01 -2.84534484e-01
3.58468890e-01 -1.23227939e-01 -1.57983154e-01 5.80466449e-01
1.97661966e-01 -1.88879669e-01 -3.08469296e-01 -5.12992442e-01
1.23151505e+00 1.37506142e-01 1.49970189e-01 -7.82869041e-01
-1.26687026e+00 6.86325014e-01 6.18986726e-01 4.09437381e-02
-6.30465686e-01 1.01373158e-01 -1.00758895e-01 -2.99446613e-01
8.02989960e-01 2.47203171e-01 -4.07741398e-01 1.89596310e-01
-1.18939626e+00 6.49791718e-01 4.45166081e-01 1.57699823e+00
1.20026982e+00 -2.64319777e-01 6.67119846e-02 7.34182715e-01
5.20000458e-01 5.64157128e-01 -2.48250812e-02 -9.94850338e-01
6.50183558e-01 1.10202312e+00 2.00020775e-01 -9.54893649e-01
-1.98542029e-01 -4.83719707e-01 -7.84242094e-01 2.80945599e-01
-1.84538230e-01 3.23669195e-01 -1.26742315e+00 1.18563771e+00
8.39501202e-01 4.49156582e-01 -1.53816029e-01 6.12662494e-01
8.85347962e-01 6.58923924e-01 -3.07295471e-02 1.42194271e-01
1.30156636e+00 -9.18300331e-01 -6.11631684e-02 -2.97946632e-01
2.63313800e-01 -7.93746948e-01 7.22538888e-01 -5.05714603e-02
-1.37219632e+00 -8.13816011e-01 -7.97652900e-01 -2.15211108e-01
1.34392828e-01 7.79419839e-02 5.93195796e-01 1.20525576e-01
-8.89871240e-01 6.76828980e-01 -9.81936812e-01 -1.27787739e-01
7.81220317e-01 6.94565922e-02 -2.27844283e-01 -2.14291993e-03
-4.55996484e-01 4.22597855e-01 5.32835841e-01 -2.59921774e-02
-8.78662050e-01 -9.76955712e-01 -7.32526422e-01 4.53342497e-02
8.79020095e-02 -1.33486331e+00 1.17660761e+00 -4.15159076e-01
-1.08578801e+00 1.17663574e+00 -4.31587070e-01 -2.00935081e-01
5.51824808e-01 -1.81996286e-01 2.12600410e-01 2.36605689e-01
5.54588020e-01 1.23667514e+00 8.98386002e-01 -1.82850969e+00
-1.05671108e+00 -5.83492875e-01 -1.36754870e-01 4.58822399e-01
3.96005660e-01 -4.37743545e-01 -1.05013692e+00 -5.35654306e-01
1.01845837e+00 -9.13052142e-01 -5.17556667e-01 -4.57902700e-02
-6.92342401e-01 -4.29575652e-01 9.19375002e-01 -2.80898690e-01
6.18615389e-01 -2.05062509e+00 1.74241975e-01 5.08109689e-01
2.30545372e-01 -3.15830290e-01 5.40073402e-03 3.42636257e-01
-5.07462360e-02 2.04880759e-01 -7.03009486e-01 -8.10474098e-01
-6.35304525e-02 1.33954853e-01 -5.13983011e-01 1.94865048e-01
2.86645710e-01 9.58714724e-01 -9.17475879e-01 -6.43407464e-01
4.57957655e-01 4.59059298e-01 -5.71869671e-01 1.94451049e-01
-5.64434588e-01 6.40748203e-01 -8.39925528e-01 8.43523026e-01
1.06519330e+00 -4.33463633e-01 -2.68589109e-01 -2.19586313e-01
9.57860705e-03 1.96644664e-01 -1.15847600e+00 2.22050357e+00
-2.78456986e-01 -8.32326859e-02 -1.29615039e-01 -5.17563045e-01
1.28653455e+00 1.90651491e-01 7.80339658e-01 -1.03350639e-01
-1.93106666e-01 3.21528316e-01 -5.11382759e-01 3.54314186e-02
7.49819636e-01 -2.25217402e-01 1.15342136e-03 3.60536575e-01
1.36604169e-02 -1.02093399e+00 -1.12894222e-01 2.41809890e-01
5.43702900e-01 5.97044289e-01 1.96442232e-02 5.61179146e-02
4.61333245e-01 4.59980547e-01 3.48644942e-01 4.92317557e-01
5.25354862e-01 1.12240934e+00 2.04459995e-01 -1.83806434e-01
-1.39275801e+00 -1.21993780e+00 -2.74451435e-01 1.45145372e-01
6.13929152e-01 -4.18903202e-01 -7.25632906e-01 -6.06640160e-01
1.25207335e-01 8.04769278e-01 -3.59722227e-01 1.23823777e-01
-4.53364968e-01 -4.53426331e-01 -4.66374867e-03 4.62302387e-01
4.69936073e-01 -1.04285228e+00 -5.37734389e-01 2.01555043e-01
-2.52392590e-01 -1.19368434e+00 -1.64482057e-01 -1.78294748e-01
-1.54388082e+00 -1.01477194e+00 -6.90963566e-01 -9.23773170e-01
1.17979622e+00 5.26490569e-01 1.39855695e+00 2.62775779e-01
1.46563649e-01 1.92446932e-01 -4.36278909e-01 -4.10895526e-01
-3.36062342e-01 2.94999450e-01 -2.56255239e-01 -1.22132532e-01
1.52934596e-01 -8.46483409e-01 -5.95111549e-01 2.58915991e-01
-8.33171904e-01 5.60042024e-01 5.36588967e-01 2.68746227e-01
1.39182985e+00 3.60311121e-01 5.52237555e-02 -8.73700261e-01
2.15304419e-01 -4.54214215e-01 -7.27886915e-01 -5.57110943e-02
-2.58451283e-01 -2.92292181e-02 1.54523388e-01 3.21293660e-02
-1.07333827e+00 3.74004930e-01 -3.89469385e-01 -8.78334761e-01
-4.30149078e-01 1.35285646e-01 -3.11114818e-01 1.94336697e-01
3.73911560e-01 3.05709124e-01 -2.99766421e-01 -8.87431324e-01
5.66618025e-01 2.90360153e-01 6.01396203e-01 -6.68736279e-01
1.38233733e+00 8.69657159e-01 2.61792839e-01 -6.31949902e-01
-7.70365894e-01 -5.18357873e-01 -1.21338367e+00 -2.94935219e-02
9.31658804e-01 -1.03436625e+00 -2.42104411e-01 4.39414889e-01
-1.52909672e+00 4.70720679e-02 -6.30434215e-01 6.05654009e-02
-7.69775331e-01 2.88210183e-01 -3.02036166e-01 -6.02988482e-01
-4.89215851e-01 -1.20590103e+00 1.94252896e+00 -1.11954398e-02
-6.04278930e-02 -7.09371328e-01 -2.09029634e-02 2.72036493e-01
-3.17778707e-01 4.90771323e-01 9.99611914e-01 -2.57270396e-01
-1.14006531e+00 -3.13685894e-01 -1.80232882e-01 1.96340699e-02
1.68729439e-01 1.40405938e-01 -8.32879245e-01 -5.42961508e-02
1.32767987e-02 1.87950954e-01 4.80774373e-01 3.91291171e-01
1.26179004e+00 3.73954810e-02 -6.76699579e-01 8.21938157e-01
1.46488452e+00 7.36810919e-03 6.36523426e-01 2.78850108e-01
9.06559289e-01 5.79875648e-01 7.10968792e-01 3.18228155e-01
5.03106892e-01 6.97329104e-01 7.52286673e-01 -1.83272615e-01
-4.88767661e-02 -7.85716534e-01 -2.62633890e-01 7.07562208e-01
-2.03813300e-01 1.47427991e-01 -1.03359687e+00 5.45380831e-01
-1.60800266e+00 -7.42424190e-01 -4.77245241e-01 2.17014980e+00
4.60789710e-01 1.28771454e-01 3.45215090e-02 5.84196746e-02
7.00940371e-01 2.83477958e-02 -5.70122659e-01 1.53445244e-01
9.62313935e-02 4.44804788e-01 2.78423697e-01 3.49989712e-01
-7.98400879e-01 1.19572711e+00 6.00396538e+00 9.11128819e-01
-7.20643997e-01 -1.15383258e-02 4.63035494e-01 1.25853270e-01
-6.01980448e-01 2.76904881e-01 -9.23574805e-01 2.15876877e-01
2.53413439e-01 -1.83523580e-01 -4.78089303e-02 1.10828793e+00
1.05124570e-01 2.11719442e-02 -1.04460692e+00 1.10617876e+00
-1.64259508e-01 -1.46698105e+00 4.91724551e-01 1.99508652e-01
1.09710574e+00 6.33386672e-02 -2.20098436e-01 -1.92126602e-01
2.25870311e-01 -7.44700849e-01 1.13079274e+00 4.40596014e-01
7.93621540e-01 -9.48169231e-01 3.89185756e-01 7.55720377e-01
-1.41986728e+00 4.27721471e-01 -6.39114141e-01 3.67805392e-01
4.48189348e-01 8.03107619e-01 -9.91425276e-01 1.00452435e+00
7.66680300e-01 7.35473812e-01 -5.67326307e-01 1.13070440e+00
-3.99456024e-01 1.18596099e-01 -3.96736026e-01 6.11565650e-01
2.30848491e-01 -6.87325239e-01 7.68258214e-01 5.96679747e-01
7.75112629e-01 5.87533005e-02 2.11815909e-01 1.39316988e+00
-8.23152363e-02 -4.66027372e-02 -5.26678801e-01 3.77459556e-01
7.89934516e-01 1.35115361e+00 -1.11933315e+00 -4.47531819e-01
-1.13467768e-01 9.50572670e-01 2.49391884e-01 1.30959228e-01
-6.55802667e-01 -9.43868607e-02 6.04926467e-01 3.56732130e-01
5.05225420e-01 -4.46500242e-01 -7.11887956e-01 -9.13743258e-01
2.94417650e-01 -2.78792918e-01 1.45656348e-03 -1.09113348e+00
-1.19425738e+00 7.84442902e-01 2.98527002e-01 -1.71940279e+00
-2.59124756e-01 -2.20695138e-02 -6.10331476e-01 1.19132900e+00
-1.40877151e+00 -1.42959070e+00 -5.62922001e-01 3.87331933e-01
7.72732854e-01 1.72972873e-01 6.44410849e-01 -9.15687010e-02
-9.32871029e-02 -2.50795390e-02 -2.90578067e-01 -1.75427780e-01
-7.52923787e-02 -1.19421053e+00 1.01466537e+00 7.95932472e-01
2.59927630e-01 4.91630673e-01 3.21733087e-01 -1.03661227e+00
-8.28780234e-01 -1.34778202e+00 7.17767119e-01 -6.39063179e-01
1.39900967e-01 -4.64327097e-01 -9.40203846e-01 5.68536162e-01
-3.78712088e-01 -3.49458337e-01 1.18391402e-01 -1.22963704e-01
-7.84578025e-02 2.53448874e-01 -1.32775390e+00 4.82400805e-01
1.35864604e+00 -2.48447224e-01 -7.23097742e-01 4.77909863e-01
8.94955754e-01 -9.17694330e-01 -7.72729099e-01 5.16655862e-01
1.33352876e-01 -1.09916604e+00 1.38897121e+00 -2.32597545e-01
7.09072411e-01 -5.64733684e-01 -1.54833077e-02 -1.16499615e+00
-4.09566194e-01 -3.07527900e-01 -9.12293047e-02 1.28256810e+00
3.30328703e-01 -3.46171647e-01 1.16316128e+00 3.28213006e-01
-5.52684963e-01 -7.56275475e-01 -8.49070311e-01 -6.23815536e-01
2.80683953e-02 -8.09373319e-01 1.30976915e+00 5.98280907e-01
-7.59233117e-01 2.70573348e-01 2.11341634e-01 6.33695424e-01
1.05481052e+00 8.23064268e-01 1.12302887e+00 -1.60654998e+00
9.85755250e-02 -3.55532229e-01 -3.93015772e-01 -1.47220612e+00
4.42336686e-02 -1.17970586e+00 1.59554463e-02 -1.96691167e+00
4.32721414e-02 -1.04402101e+00 4.56095546e-01 2.46080786e-01
-1.73591495e-01 3.14439684e-01 1.53897941e-01 7.78785348e-01
-1.25422806e-01 6.60909355e-01 1.55953920e+00 7.72796152e-03
-4.24115181e-01 4.80469137e-01 -7.34561324e-01 8.94041538e-01
4.21830863e-01 -5.27498722e-01 -4.25244629e-01 -7.08144248e-01
3.48502547e-02 1.01975098e-01 5.42087734e-01 -1.11467409e+00
-2.26769354e-02 -9.32538956e-02 5.68270624e-01 -1.51508319e+00
6.39225185e-01 -1.11407053e+00 6.64950192e-01 8.62793624e-02
1.21802658e-01 -4.46599014e-02 3.40917632e-02 5.72682858e-01
-1.16777651e-01 -3.53864074e-01 6.72537446e-01 -3.91796231e-01
-4.34814692e-01 9.69422579e-01 4.31338489e-01 -3.37773204e-01
1.16524434e+00 -7.57856965e-01 2.97561944e-01 3.73924598e-02
-7.63215661e-01 3.09473425e-01 1.05712759e+00 4.91417766e-01
8.35408926e-01 -1.51232672e+00 -7.69475698e-01 3.50469351e-01
2.63371646e-01 1.28883481e+00 2.93669611e-01 2.80732930e-01
-7.03532577e-01 3.37575227e-01 7.41618574e-02 -1.09594190e+00
-9.56299543e-01 3.45714688e-01 1.97592467e-01 -5.28261214e-02
-1.23981786e+00 8.24960172e-01 4.88245666e-01 -6.00088179e-01
-9.36012045e-02 -4.60768342e-01 7.20303133e-02 -9.27567706e-02
1.29899323e-01 2.38517195e-01 1.96155965e-01 -9.21143472e-01
-2.55876988e-01 9.70564425e-01 2.33093381e-01 -1.34178221e-01
1.49112535e+00 -4.42980900e-02 -1.90796852e-01 2.76054919e-01
9.00656879e-01 -3.11968047e-02 -1.25711739e+00 -1.84846416e-01
-1.43463671e-01 -8.20835173e-01 -1.87380493e-01 -2.48746380e-01
-1.06305850e+00 7.57495344e-01 1.02273025e-01 2.39196345e-01
9.09038365e-01 6.10651195e-01 8.88341904e-01 -1.47549972e-01
8.34548175e-01 -6.20569408e-01 -1.94473699e-01 3.18005651e-01
1.06064188e+00 -7.67464042e-01 1.48880839e-01 -1.12697923e+00
-4.20148253e-01 9.44461524e-01 4.94182169e-01 -3.33967060e-01
4.74506140e-01 -1.60642639e-01 -3.68643790e-01 -6.79570913e-01
-4.28819656e-01 -1.55234218e-01 4.23212230e-01 9.24279332e-01
-5.18682823e-02 1.31160347e-02 1.57678798e-01 1.79789737e-01
-7.18670070e-01 -1.53469190e-01 3.59032042e-02 7.37649560e-01
-4.85885203e-01 -1.13478017e+00 -5.83419204e-01 4.24153894e-01
2.36000270e-01 1.17514022e-01 -3.50861281e-01 6.60599530e-01
2.30575144e-01 6.55151308e-01 6.22161269e-01 -2.13308796e-01
6.18520439e-01 -7.00481534e-02 3.73805404e-01 -1.03794980e+00
-2.73590595e-01 2.96537280e-01 -3.43292773e-01 -5.95776200e-01
-4.68691856e-01 -8.29540789e-01 -1.59550750e+00 -1.09923154e-01
-4.10986871e-01 2.33583331e-01 7.10014939e-01 7.29608834e-01
6.29341364e-01 3.28839540e-01 8.39564562e-01 -1.46958387e+00
-1.08444288e-01 -8.66670072e-01 -5.86191475e-01 5.82477093e-01
-2.18067050e-01 -8.19150925e-01 -2.87480921e-01 1.53359830e-01]
|
[8.402247428894043, -3.508168935775757]
|
6e1d7183-c0d8-4da4-887f-5767c71304aa
|
automatic-sleep-staging-recent-development
|
2111.08446
| null |
https://arxiv.org/abs/2111.08446v3
|
https://arxiv.org/pdf/2111.08446v3.pdf
|
Automatic Sleep Staging of EEG Signals: Recent Development, Challenges, and Future Directions
|
Modern deep learning holds a great potential to transform clinical practice on human sleep. Teaching a machine to carry out routine tasks would be a tremendous reduction in workload for clinicians. Sleep staging, a fundamental step in sleep practice, is a suitable task for this and will be the focus in this article. Recently, automatic sleep staging systems have been trained to mimic manual scoring, leading to similar performance to human sleep experts, at least on scoring of healthy subjects. Despite tremendous progress, we have not seen automatic sleep scoring adopted widely in clinical environments. This review aims to give a shared view of the authors on the most recent state-of-the-art development in automatic sleep staging, the challenges that still need to be addressed, and the future directions for automatic sleep scoring to achieve clinical value.
|
['Kaare Mikkelsen', 'Huy Phan']
|
2021-11-03
| null | null | null | null |
['sleep-staging']
|
['medical']
|
[ 5.72495237e-02 9.93633270e-02 -2.50962138e-01 -6.88616037e-01
-5.58482230e-01 -1.58066183e-01 -2.16927320e-01 1.93899766e-01
-6.97656572e-01 6.05494797e-01 1.27845287e-01 -5.34870446e-01
1.07604675e-01 -9.54818912e-03 4.68018264e-01 -7.88740814e-01
-7.76055874e-03 6.08062267e-01 1.21188276e-01 5.26535176e-02
2.15457723e-01 3.43736321e-01 -1.22442198e+00 2.51323611e-01
5.46495438e-01 5.95278680e-01 2.38858670e-01 8.84139240e-01
-6.93307351e-03 7.51244903e-01 -9.23967004e-01 -3.66491139e-01
-1.70524254e-01 -7.42897391e-01 -9.61877346e-01 -1.07370548e-01
3.33503336e-01 -1.66807517e-01 1.39060155e-01 6.56917453e-01
9.52305675e-01 2.13416740e-01 1.90649047e-01 -6.56249344e-01
-2.36580744e-01 -3.12926695e-02 -8.39746426e-05 1.21284950e+00
2.69951314e-01 1.90381587e-01 7.05309451e-01 -4.77760136e-01
5.13297580e-02 2.96593875e-01 8.79114449e-01 1.09105718e+00
-1.09233165e+00 -6.53758466e-01 -5.08237600e-01 5.27653456e-01
-1.26965368e+00 -7.03072667e-01 3.10119450e-01 -1.27532095e-01
1.39634371e+00 3.33619863e-01 1.41591918e+00 6.01855814e-01
8.68130267e-01 4.51201439e-01 9.69740152e-01 -3.14412326e-01
3.29965889e-01 1.22351430e-01 1.14739887e-01 8.74160767e-01
2.90143192e-01 -3.51727992e-01 -7.32232869e-01 1.49216905e-01
3.57036203e-01 3.18390369e-01 4.99660224e-02 1.02700330e-01
-1.01907194e+00 8.61142874e-01 2.91921824e-01 6.18026912e-01
-1.32033139e-01 2.99819838e-02 5.54025054e-01 7.80676007e-02
3.87117326e-01 7.01934159e-01 -3.35211217e-01 -7.13775158e-01
-1.87789392e+00 -2.14041114e-01 8.23362529e-01 2.28004843e-01
6.15659133e-02 -6.00998849e-02 -1.88597262e-01 7.79633403e-01
3.29534352e-01 4.48070377e-01 1.02411819e+00 -1.10118520e+00
-2.86462400e-02 6.10221267e-01 4.43468504e-02 -5.31135440e-01
-1.07826316e+00 -5.04651725e-01 -8.51067543e-01 1.56324804e-01
1.56189471e-01 4.44407240e-02 -6.45974696e-01 1.20388007e+00
-1.67169143e-02 -1.65721834e-01 -1.21344373e-01 7.70911992e-01
1.03414941e+00 1.00496456e-01 1.92491099e-01 -3.80107105e-01
1.63965225e+00 -1.16487455e+00 -8.81825387e-01 -6.10030770e-01
6.39195204e-01 -9.70254064e-01 1.03832436e+00 6.64795101e-01
-1.38256371e+00 -3.88796449e-01 -1.18566382e+00 -3.49512279e-01
-1.84967786e-01 1.38691217e-01 7.35040188e-01 1.08484471e+00
-1.39038706e+00 6.35814548e-01 -1.75755429e+00 -8.80013466e-01
6.15461469e-01 9.02065992e-01 -2.26986676e-01 3.67282927e-01
-8.66253257e-01 1.41497111e+00 -1.69788584e-01 2.38992766e-01
-6.84739530e-01 -5.11496603e-01 -6.53354347e-01 1.71891555e-01
1.55422147e-02 -1.15379715e+00 1.83622217e+00 -4.24385160e-01
-1.46492136e+00 1.62305212e+00 -5.72730839e-01 -5.36438167e-01
8.55461434e-02 -1.41054004e-01 -4.77241904e-01 2.44990051e-01
1.75639987e-01 4.62262511e-01 4.06859577e-01 -1.55490950e-01
-5.52607596e-01 -2.86370635e-01 -1.56740740e-01 1.91854656e-01
-1.77059114e-01 3.15390676e-01 -1.01191863e-01 -1.50168195e-01
-1.18980795e-01 -1.09650576e+00 -3.93564403e-01 1.49038643e-01
-2.14091260e-02 -3.34241122e-01 2.01590732e-01 -3.45272005e-01
1.44262135e+00 -1.98826575e+00 -3.21995467e-01 -3.60174000e-01
7.32122600e-01 5.77959061e-01 5.46125948e-01 3.59786689e-01
3.37728374e-02 -1.94802321e-02 5.21881022e-02 -9.49198663e-01
-3.14126402e-01 3.48236144e-01 3.06198955e-01 8.68817568e-01
-2.12316543e-01 8.91664147e-01 -9.56284702e-01 -7.00767875e-01
3.74525636e-01 3.38265181e-01 -4.23120677e-01 3.31386328e-01
6.49026215e-01 3.51141304e-01 -9.99829769e-02 4.21678871e-01
-2.08890773e-02 -8.46400559e-01 1.05793856e-01 3.00513148e-01
-1.85914919e-01 7.76681066e-01 -3.87501746e-01 1.82654393e+00
-4.96209055e-01 6.72740996e-01 6.10977804e-05 -6.60024643e-01
6.57280087e-01 4.17549431e-01 5.33383846e-01 -3.61828715e-01
3.13684970e-01 1.89015478e-01 5.60345888e-01 -4.22262698e-01
1.95813626e-01 -1.04202509e+00 1.32008329e-01 6.11464322e-01
2.09886044e-01 -2.07124561e-01 3.65305185e-01 3.88704129e-02
1.36603808e+00 -6.39787436e-01 9.03390050e-01 -3.13290626e-01
5.57336271e-01 1.38127923e-01 4.67260927e-01 5.84904969e-01
-8.58249545e-01 6.28100216e-01 2.37430856e-01 -8.40640187e-01
-5.78886211e-01 -9.70087051e-01 -9.66366008e-03 1.00140536e+00
-2.14209348e-01 -6.29487753e-01 -7.44584382e-01 -5.43889403e-01
-5.92824340e-01 4.80320513e-01 -8.56669486e-01 -3.40822875e-01
-3.62925947e-01 -9.65181947e-01 4.57592845e-01 6.69543087e-01
1.26211196e-01 -1.36991167e+00 -1.30642271e+00 2.27340609e-01
-1.98155239e-01 -8.85012090e-01 -4.97269243e-01 7.85129189e-01
-1.12902057e+00 -1.16840661e+00 -8.24293137e-01 -8.11249912e-01
5.57162046e-01 5.23700237e-01 1.31215203e+00 6.17263198e-01
-5.28149545e-01 1.58526123e-01 -1.53281689e-01 -8.01672518e-01
-3.17165017e-01 5.10287702e-01 3.20678890e-01 -6.82254732e-01
9.80958581e-01 -6.60301387e-01 -1.31152010e+00 2.57006586e-01
-4.19191182e-01 7.41887614e-02 8.26990247e-01 7.75140584e-01
5.60069978e-01 -3.97172213e-01 1.55876726e-01 -8.89555037e-01
5.71379960e-01 -3.02289128e-01 -8.05294961e-02 -1.02311946e-01
-9.70419168e-01 -2.37468526e-01 7.18278825e-01 7.00678304e-02
-5.63869596e-01 -1.61374345e-01 -4.46390599e-01 -3.99741903e-02
-1.63771883e-01 3.30097288e-01 6.30587041e-01 2.83156708e-02
7.76743948e-01 2.22667620e-01 3.99447620e-01 -4.12952244e-01
-3.01370829e-01 5.43023765e-01 4.06591088e-01 1.25303686e-01
3.76419514e-01 7.43113041e-01 1.76310048e-01 -8.08422506e-01
-1.54039526e+00 -1.12674940e+00 -7.90126443e-01 5.62068671e-02
1.22183049e+00 -6.23557091e-01 -6.08033538e-01 -3.65547203e-02
-6.12135112e-01 -7.14850247e-01 -4.24917400e-01 3.65031898e-01
-6.11746371e-01 4.56096202e-01 -5.17630279e-01 -3.97058308e-01
-1.12106943e+00 -1.29987001e+00 1.04376411e+00 5.06586611e-01
-1.00804925e+00 -1.33566391e+00 7.38114953e-01 6.78140879e-01
6.17085218e-01 -3.23057711e-01 5.67322254e-01 -8.03951502e-01
2.75211278e-02 -2.42386132e-01 2.59022236e-01 4.28906798e-01
4.62545961e-01 -1.97001487e-01 -9.84724581e-01 -2.38164157e-01
4.59047437e-01 -2.18518794e-01 4.65482503e-01 7.15800345e-01
1.00670302e+00 2.26688817e-01 -3.59818339e-01 7.87245572e-01
9.65740561e-01 1.21946201e-01 5.14915049e-01 3.36080223e-01
4.77618873e-02 2.88251657e-02 3.71113479e-01 2.50673831e-01
3.30770075e-01 5.14166653e-01 -6.16584122e-02 -4.30173457e-01
-3.35232943e-01 3.97180736e-01 1.73185438e-01 1.26504087e+00
-2.10386112e-01 1.93070501e-01 -1.07835698e+00 3.16191822e-01
-1.23396301e+00 -8.99328709e-01 -7.75175914e-02 1.82480860e+00
6.80100620e-01 3.30391079e-01 4.06992644e-01 1.97667137e-01
1.69467703e-01 8.56966432e-03 -4.99483138e-01 -8.44345391e-01
5.42913616e-01 8.23802114e-01 1.77072659e-01 7.54639730e-02
-7.99416244e-01 6.83740258e-01 7.68096924e+00 3.03042680e-01
-1.40747905e+00 5.61921716e-01 4.05768096e-01 -6.62102997e-01
6.42349541e-01 -3.79786491e-01 -6.65145040e-01 6.77169979e-01
1.55449605e+00 -1.16314910e-01 2.76782304e-01 1.09743643e+00
8.28173697e-01 -4.29077625e-01 -9.78460848e-01 1.11488283e+00
2.82152295e-01 -1.23499787e+00 -6.12022340e-01 1.45527005e-01
4.41670179e-01 3.34105372e-01 9.70763490e-02 2.93320656e-01
-3.75703573e-02 -1.07001758e+00 1.75380200e-01 3.06297064e-01
1.01343739e+00 -2.87934572e-01 1.32430506e+00 1.83018334e-02
-9.66158926e-01 1.95434973e-01 -4.30710137e-01 -3.50478649e-01
1.91417158e-01 3.47436577e-01 -1.12181962e+00 -1.73145831e-01
8.47844541e-01 7.36222267e-01 -8.99669409e-01 1.47209251e+00
-4.21867192e-01 8.13291132e-01 -5.61749376e-02 -2.85650820e-01
1.27056092e-01 1.23361103e-01 -2.76744459e-02 1.17049181e+00
1.38939127e-01 1.34065658e-01 -5.37577234e-02 3.15338105e-01
2.72915572e-01 8.79253820e-02 -3.65176871e-02 -1.10688079e-02
-8.97509139e-03 1.62644887e+00 -1.27874219e+00 -2.98613936e-01
-7.58877099e-01 1.05015624e+00 2.43795902e-01 -4.12442207e-01
-6.44642234e-01 -2.02854082e-01 6.92403674e-01 3.85512859e-01
-1.46479547e-01 -6.81632310e-02 -6.08259737e-01 -9.32121933e-01
-3.82652402e-01 -7.02957332e-01 5.89764833e-01 -7.03217804e-01
-1.03386033e+00 4.02881712e-01 -3.23297769e-01 -1.23568666e+00
-4.64859456e-02 -5.00967979e-01 -1.04642308e+00 6.35212839e-01
-1.12808716e+00 -7.13502347e-01 -6.11264884e-01 3.79857987e-01
8.53181303e-01 -2.86078826e-03 1.18877017e+00 2.57041544e-01
-6.70385599e-01 6.28796160e-01 -1.33038322e-02 -1.19900912e-01
9.30098712e-01 -1.59625483e+00 2.77718216e-01 4.63035971e-01
2.25384876e-01 1.19122446e+00 8.20796430e-01 -2.46171877e-01
-9.11113083e-01 -6.63906932e-01 1.18602931e+00 -9.54503298e-01
4.87784773e-01 -2.54004840e-02 -5.83020091e-01 4.73081261e-01
1.65595740e-01 -3.20110619e-01 1.88754058e+00 4.39106613e-01
5.69016576e-01 -3.02437633e-01 -1.02651238e+00 4.49159831e-01
6.31303489e-01 -4.43318665e-01 -9.65681255e-01 4.23482567e-01
-8.77528042e-02 -5.74367225e-01 -7.13923275e-01 -1.79465473e-01
7.66597629e-01 -1.27748060e+00 6.53984725e-01 -3.79483730e-01
2.61601478e-01 -1.41490772e-01 7.21719980e-01 -1.05992508e+00
-4.87944245e-01 -7.05698907e-01 -3.61060612e-02 4.69429016e-01
2.56024688e-01 -3.26663643e-01 1.44714487e+00 7.85501003e-01
-7.51183033e-01 -1.26423311e+00 -1.09922528e+00 -4.90491539e-01
-7.86537305e-02 -1.55609101e-01 1.14363186e-01 3.49665970e-01
5.67883074e-01 6.24374509e-01 -1.31052628e-01 -3.28071177e-01
8.26300830e-02 -1.34654537e-01 6.27228439e-01 -1.24504924e+00
-8.22059736e-02 -6.40189648e-01 -6.46627426e-01 -5.94453037e-01
-2.48320013e-01 -9.90896106e-01 -9.07924306e-03 -1.94467747e+00
7.14052379e-01 1.05054267e-01 -6.38159573e-01 6.61166012e-01
-2.56512135e-01 8.32881033e-01 -1.18020721e-01 3.57974946e-01
-1.06911945e+00 1.24403074e-01 1.21267176e+00 1.78105950e-01
-3.29484820e-01 4.72709984e-01 -1.20164931e+00 7.36794114e-01
1.10367560e+00 -8.50399077e-01 -4.62973773e-01 3.51483785e-02
3.52653414e-01 -2.11262479e-01 -2.60298997e-01 -1.42800212e+00
3.40371788e-01 7.18921497e-02 5.05232036e-01 -4.33898628e-01
6.01477802e-01 -6.04021490e-01 1.98411345e-02 9.42350447e-01
-2.46492513e-02 4.65446174e-01 2.07968876e-01 3.36893946e-02
-4.73388880e-02 -3.77325088e-01 1.06157553e+00 -4.32721525e-01
-2.84706056e-01 6.35893047e-02 -1.05017877e+00 2.16045380e-01
6.84784889e-01 -5.92249513e-01 -6.50031567e-02 -3.26778948e-01
-1.20849729e+00 1.12206951e-01 4.93593574e-01 -1.06352143e-01
3.31562668e-01 -7.50036597e-01 -9.59692970e-02 5.20906933e-02
6.33726567e-02 -1.82110757e-01 5.33580005e-01 1.60642064e+00
-8.63374531e-01 8.82017672e-01 -3.11308652e-01 -5.09712696e-01
-1.68466675e+00 3.87676179e-01 4.41599965e-01 -8.08028340e-01
-6.83835804e-01 8.51231277e-01 -9.28246751e-02 1.51040286e-01
1.15348019e-01 -5.50927341e-01 -2.67117620e-01 -7.80593902e-02
8.16705644e-01 4.28271413e-01 6.19911790e-01 -2.45272771e-01
-7.15135574e-01 2.68113077e-01 -2.08463833e-01 3.27059895e-01
1.26194537e+00 -2.53887236e-01 -3.80955823e-02 6.02976859e-01
8.59407067e-01 -4.01007608e-02 -5.24556696e-01 3.41975838e-01
-1.49386004e-01 -9.31913033e-02 1.53960019e-01 -1.01430321e+00
-8.89016330e-01 1.35800505e+00 9.92713153e-01 1.69455975e-01
1.07231545e+00 1.79450944e-01 1.08436835e+00 3.69317979e-01
2.62750179e-01 -9.64565635e-01 2.38037467e-01 1.34318098e-01
2.75403768e-01 -1.28575563e+00 4.07187462e-01 1.64867446e-01
-6.25378966e-01 1.19535077e+00 3.95260513e-01 1.93639956e-02
7.64153421e-01 3.51314619e-02 5.37666142e-01 -7.25146413e-01
-6.16597354e-01 -2.27235481e-01 3.49112034e-01 4.62375104e-01
8.18506956e-01 6.11589029e-02 -4.68755156e-01 5.74123263e-01
-6.78598404e-01 4.56436843e-01 5.92688501e-01 8.69218647e-01
-5.97754240e-01 -1.19038105e+00 -8.07598457e-02 1.06608474e+00
-1.08004165e+00 -2.67799377e-01 -2.70999789e-01 4.95038331e-01
1.97878823e-01 1.03260851e+00 -1.32123783e-01 -2.27878228e-01
1.51423737e-01 3.90799999e-01 4.67952907e-01 -1.43356252e+00
-8.75483751e-01 -7.02160075e-02 -1.43584296e-01 -6.53931737e-01
-4.93638813e-01 -7.56677389e-01 -1.17668366e+00 -3.61092389e-01
-1.41275749e-01 3.68089050e-01 4.38091666e-01 1.02999759e+00
2.06100032e-01 6.11130774e-01 -1.21288821e-01 -8.66653621e-01
-2.85493970e-01 -9.78512943e-01 -9.05335128e-01 -8.09227899e-02
4.19796765e-01 -8.07363868e-01 -4.69035029e-01 1.53356493e-01]
|
[13.523439407348633, 3.503833055496216]
|
c365cff9-9f64-4dce-8520-2187d2604792
|
pneumococcus-and-the-stress-gradient
|
2205.12629
| null |
https://arxiv.org/abs/2205.12629v1
|
https://arxiv.org/pdf/2205.12629v1.pdf
|
Pneumococcus and the stress-gradient hypothesis: a trade-off links $R_0$ and susceptibility to co-colonization across countries
|
Modern molecular technologies have revolutionized our understanding of bacterial epidemiology, but reported data across different settings remain under-integrated in common theoretical frameworks. Pneumococcus serotype co-colonization, caused by the polymorphic bacteria Streptococcus pneumoniae, has been increasingly investigated in recent years. While the global genomic diversity and serotype distribution of S. pneumoniae are well-characterized, there is limited information on how co-colonization patterns vary globally, critical for understanding bacterial evolution and dynamics. Gathering a rich dataset of cross-sectional pneumococcal colonization studies in the literature, we quantified patterns of transmission intensity and co-colonization prevalence in children populations across 17 geographic locations. Fitting these data to an SIS model with co-colonization under the assumption of similarity among interacting strains, our analysis reveals strong patterns of negative co-variation between transmission intensity ($R_0$) and susceptibility to co-colonization ($k$). In support of the stress-gradient hypothesis in ecology (SGH), pneumococcus serotypes appear to compete more in high-transmission settings and less in low-transmission settings, a trade-off which ultimately leads to a conserved ratio of single to co-colonization $\mu=1/(R_0-1)k$. Within our mathematical model, such conservation suggests preservation of 'stability-diversity-complexity' regimes in multi-strain coexistence. We find no major study differences in serotype composition, pointing to underlying adaptation of the same set of serotypes across environments. Our work highlights that understanding pneumococcus transmission patterns from global epidemiological data can benefit from simple analytical approaches that account for quasi-neutrality among strains, co-colonization, as well as variable environmental adaptation.
|
['Erida Gjini', 'Ermanda Dekaj']
|
2022-05-25
| null | null | null | null |
['epidemiology']
|
['medical']
|
[ 3.99096906e-01 -4.77828145e-01 1.23452730e-01 1.62350565e-01
2.25296065e-01 -7.24822521e-01 2.23339856e-01 5.70339441e-01
-4.55464900e-01 8.07612121e-01 2.56952971e-01 -6.29497588e-01
-7.68962204e-01 -6.65877223e-01 -9.56512809e-01 -1.16861284e+00
-6.76519573e-01 3.42467934e-01 2.04783961e-01 -9.72075760e-02
-1.96244627e-01 1.51397422e-01 -1.38756037e+00 7.80915320e-02
9.20890868e-01 -2.01633558e-01 7.49788523e-01 1.09560847e+00
8.56409818e-02 6.77855834e-02 -1.34713918e-01 -9.92911160e-02
1.69080555e-01 -3.40533584e-01 -1.30850658e-01 -8.50335896e-01
-2.49631479e-01 -1.36393085e-01 5.79410680e-02 8.14041555e-01
4.82899308e-01 -4.82667595e-01 7.82751560e-01 -6.25997603e-01
-4.91174251e-01 3.49767923e-01 -3.66656840e-01 2.49328017e-01
1.58664778e-01 2.28517011e-01 7.50125766e-01 -6.98286071e-02
8.36188257e-01 1.24253678e+00 8.45308423e-01 3.54671568e-01
-1.35121930e+00 -1.21110231e-01 -1.96094830e-02 -4.52270284e-02
-1.19217885e+00 2.85547733e-01 3.08965206e-01 -1.04636765e+00
1.22938049e+00 4.83503222e-01 1.29934907e+00 1.11044538e+00
7.03591585e-01 2.52436269e-02 1.01966333e+00 -8.90200660e-02
1.60427004e-01 -2.97139101e-02 6.60037845e-02 3.00431311e-01
1.12843144e+00 4.09726918e-01 -1.73078537e-01 -7.45330095e-01
8.61004531e-01 4.86347616e-01 -4.29826081e-01 -3.61855209e-01
-9.90923762e-01 5.51624775e-01 8.13788474e-02 2.92380840e-01
-4.24632370e-01 -6.03926741e-02 1.21768720e-01 2.04148903e-01
3.89059633e-02 5.47501326e-01 -9.47110474e-01 -2.65516311e-01
1.15340129e-02 3.20821434e-01 5.38058043e-01 4.76395756e-01
6.05719626e-01 -3.85369509e-01 4.94965345e-01 8.35339785e-01
4.35878247e-01 1.07187843e+00 1.60176158e-01 -4.66925859e-01
-1.24040549e-03 3.94675553e-01 3.34395856e-01 -1.01331651e+00
-4.93980110e-01 -5.40684462e-01 -7.09879518e-01 -2.85485268e-01
6.66348755e-01 -2.84111619e-01 1.96743221e-03 2.41126084e+00
3.83357018e-01 2.05744252e-01 -9.05505046e-02 4.03557718e-01
-1.88784689e-01 4.60841984e-01 2.75544077e-01 -5.83932579e-01
1.42888927e+00 -3.55966389e-01 -1.14082638e-02 9.19195712e-02
4.64000493e-01 -5.31492412e-01 6.60176933e-01 -1.51103102e-02
-5.61419189e-01 -1.43009678e-01 -6.71601892e-01 7.99781442e-01
-2.49112010e-01 -8.07077467e-01 2.16604918e-01 9.20454800e-01
-1.14377666e+00 6.08228505e-01 -1.13649035e+00 -1.08253098e+00
1.11894935e-01 2.06989348e-01 -1.28212431e-03 -3.21131572e-02
-8.27090979e-01 6.09508753e-01 -5.09029701e-02 -2.64259428e-01
-4.51600730e-01 -1.08957982e+00 7.33350664e-02 -1.18236773e-01
1.26805007e-01 -1.10778022e+00 4.50669497e-01 -3.57887059e-01
-1.16342354e+00 4.49332029e-01 -1.05135664e-01 2.85578556e-02
2.68270671e-01 -4.63611513e-01 -2.91077346e-01 2.13988110e-01
-6.15748018e-02 -1.64871104e-02 8.08715969e-02 -1.19219232e+00
-3.36837202e-01 -6.03276491e-01 -3.17258030e-01 -5.24055511e-02
8.60090256e-02 3.23071539e-01 3.62591743e-01 -4.72545624e-01
2.33156960e-02 -1.20916164e+00 -5.06999552e-01 -4.20614153e-01
2.08128437e-01 1.04396462e-01 3.00374269e-01 -4.50866282e-01
9.81139779e-01 -2.02501440e+00 1.97914094e-01 1.48543477e-01
2.41769612e-01 2.25100830e-01 -2.30011120e-01 9.73749518e-01
-1.89317856e-02 3.65774900e-01 -4.39053297e-01 6.66611612e-01
-1.36867717e-01 1.56488463e-01 2.02755108e-02 7.69543707e-01
3.34102422e-01 4.74496394e-01 -1.37894917e+00 2.25326102e-02
2.41994664e-01 9.65835452e-01 -1.20402384e+00 3.99327308e-01
-2.30478063e-01 5.67189932e-01 -5.74689150e-01 4.92489398e-01
8.24254870e-01 -5.25660694e-01 8.82351875e-01 4.33322787e-01
-5.67256093e-01 2.45281458e-02 -8.94581199e-01 7.79791236e-01
7.73064718e-02 1.89130485e-01 1.71349257e-01 -7.76612997e-01
4.62032259e-01 1.96899042e-01 5.92101693e-01 -1.56254426e-01
6.86006099e-02 6.94410384e-01 6.58381224e-01 -6.22722328e-01
2.27923077e-02 -1.95822790e-01 5.07613420e-01 3.59916478e-01
-4.87506151e-01 3.15531850e-01 -1.29587710e-01 -2.55496353e-01
1.31013644e+00 5.24683036e-02 3.30122232e-01 -8.29212010e-01
3.25031698e-01 1.68996811e-01 5.08045733e-01 1.02592373e+00
-2.93507248e-01 4.04212326e-01 6.59116268e-01 3.20590138e-02
-1.15205991e+00 -1.42775309e+00 -7.86658347e-01 9.15840268e-01
9.13430825e-02 -2.99625807e-02 -7.36954451e-01 5.52284360e-01
7.69854337e-02 6.73271939e-02 -4.73624796e-01 -6.51245788e-02
-1.01429260e+00 -1.45038164e+00 5.80696404e-01 -6.50642291e-02
2.07205489e-02 -7.42085338e-01 -1.24849832e+00 5.02998769e-01
9.15334895e-02 -3.25068533e-01 -1.82754137e-02 1.83973104e-01
-7.67235816e-01 -1.63883114e+00 -7.61065960e-01 -4.10262108e-01
6.12458408e-01 2.02132165e-01 8.79175127e-01 2.32916519e-01
-7.72132337e-01 5.59695661e-01 -2.67698437e-01 -2.30990455e-01
-6.12577617e-01 -3.14667910e-01 3.92209142e-01 -5.65243065e-01
4.18323010e-01 -5.68228841e-01 -1.17886317e+00 5.38906217e-01
-6.52434289e-01 -5.33111274e-01 3.99015963e-01 8.25314999e-01
3.46291572e-01 -2.92457432e-01 4.81734544e-01 -3.40579957e-01
1.84378549e-01 -1.09978306e+00 -5.83836734e-01 3.86870980e-01
-5.14147758e-01 -2.28879169e-01 5.06823003e-01 -5.30140698e-01
-1.16879451e+00 -6.05904222e-01 9.38744470e-02 5.65951288e-01
-1.05482996e-01 3.53604585e-01 2.01013923e-01 4.07933861e-01
4.26592976e-01 9.03680474e-02 1.53319910e-01 -5.24987280e-01
-6.92408979e-02 5.25233746e-01 5.43929897e-02 -7.50851512e-01
2.72959262e-01 4.21108097e-01 1.34682551e-01 -1.29773319e+00
1.17726950e-02 -4.25845832e-01 -2.95201838e-01 -1.62147954e-01
1.27136111e+00 -7.89492905e-01 -1.05007982e+00 8.72048020e-01
-8.88845742e-01 -5.75939119e-01 -1.34997899e-02 9.33281779e-01
-4.38561052e-01 2.98318446e-01 -6.48478329e-01 -1.04224062e+00
1.84527770e-01 -1.21600437e+00 7.40183651e-01 -7.36860856e-02
-1.38144180e-01 -1.25839221e+00 1.14274812e+00 1.33184567e-01
7.26146519e-01 4.43490446e-01 1.26484001e+00 -5.21613777e-01
-7.64241755e-01 4.81337249e-01 6.05581235e-03 -2.23024208e-02
4.12229836e-01 5.20553887e-01 -3.33386660e-01 -5.48577070e-01
1.02353305e-01 3.65051717e-01 5.38140595e-01 8.51923704e-01
2.57604092e-01 -2.04823062e-01 -6.52952790e-01 6.42433465e-01
1.71258366e+00 6.30643606e-01 2.45513633e-01 2.81773627e-01
4.48233336e-01 1.11372495e+00 4.59940545e-03 4.60609645e-01
2.28736550e-01 5.39569020e-01 2.68689240e-03 3.58129352e-01
3.11648548e-01 -4.55440842e-02 2.27167323e-01 1.18682444e+00
-3.89679223e-01 -2.50822634e-01 -1.31909287e+00 3.97296041e-01
-1.67024088e+00 -1.13625813e+00 -3.86002243e-01 2.43821216e+00
8.85051787e-01 -2.73438781e-01 3.40693831e-01 -6.29951179e-01
6.10883296e-01 1.70157906e-02 -4.74201441e-01 -2.99264193e-01
-6.34112775e-01 2.16591224e-01 6.07183754e-01 6.09455943e-01
-5.19532382e-01 6.42040521e-02 7.86807394e+00 -1.05289668e-01
-1.22315741e+00 -9.68786776e-02 3.59328479e-01 -1.06663868e-01
-6.46076024e-01 -5.63675128e-02 -3.56435776e-01 5.01326621e-01
1.18968904e+00 -5.82026169e-02 4.15408909e-01 3.37750286e-01
2.13828355e-01 -4.53373194e-02 -8.64806235e-01 2.85641342e-01
-2.39306927e-01 -1.16174006e+00 -5.14987171e-01 2.42879674e-01
8.25998187e-01 5.09380579e-01 -1.01845991e-02 -2.34126344e-01
6.32575929e-01 -7.36261070e-01 2.67645329e-01 4.32841420e-01
3.45894903e-01 -5.14064908e-01 6.13428891e-01 1.48518965e-01
-1.22263288e+00 -3.95238101e-02 -3.85847121e-01 -2.88773090e-01
3.38525951e-01 7.21940160e-01 -7.46717930e-01 3.09855759e-01
8.15691650e-01 4.33044881e-01 -2.25989953e-01 7.52583206e-01
2.04387695e-01 4.09718484e-01 -5.02369583e-01 -3.32781523e-01
-2.13444546e-01 -5.53005993e-01 7.14779973e-01 1.21205640e+00
3.83047730e-01 2.64720172e-01 -4.06062067e-01 7.19388306e-01
8.91014993e-01 1.52908117e-01 -5.27599394e-01 -2.54592657e-01
7.60001302e-01 3.41653585e-01 -6.38395429e-01 -2.62020916e-01
-4.22195286e-01 2.93712556e-01 3.95427123e-02 4.49202985e-01
-3.36894512e-01 6.99134767e-02 1.36106098e+00 1.84383586e-01
4.03729796e-01 -3.93100530e-01 1.68447256e-01 -9.01108086e-01
-2.83782393e-01 -6.62566841e-01 2.98393935e-01 -1.37410164e-01
-1.28557110e+00 2.87971050e-01 3.64876628e-01 -3.99399668e-01
-2.35573605e-01 -9.05388713e-01 -4.52200383e-01 7.82670856e-01
-1.13595939e+00 -3.70702952e-01 1.33073956e-01 2.22693756e-01
-6.41805381e-02 1.24984361e-01 5.99375904e-01 2.67799497e-01
-7.36389101e-01 1.94829643e-01 1.01278746e+00 -6.75975144e-01
3.94108623e-01 -1.03097987e+00 5.68958484e-02 7.16149986e-01
-7.78057933e-01 1.51957917e+00 8.08891177e-01 -1.21409643e+00
-1.54436100e+00 -1.04560244e+00 5.39676905e-01 -6.10355556e-01
9.71488833e-01 -3.16814572e-01 -8.59578371e-01 4.39120382e-01
1.99101657e-01 -7.33994484e-01 1.13908768e+00 -6.44763336e-02
-5.33497691e-01 3.61193508e-01 -1.09894848e+00 9.62097883e-01
1.17687643e+00 -3.06950510e-01 -3.08333039e-01 1.96591541e-01
1.03613973e+00 4.01451379e-01 -1.09938085e+00 5.16798973e-01
9.60764170e-01 -1.19780409e+00 1.30242682e+00 -6.63086057e-01
1.87914982e-01 -3.05719137e-01 -5.22960901e-01 -1.01802099e+00
-4.94793057e-01 -3.07569385e-01 7.95346916e-01 6.64655268e-01
2.36081213e-01 -1.33310890e+00 1.23970196e-01 5.95708080e-02
-1.23105608e-01 -6.78754747e-01 -1.09790146e+00 -7.96977520e-01
5.70096016e-01 3.80776107e-01 7.57993519e-01 9.11940753e-01
-4.09389734e-02 -3.97967279e-01 -6.07807823e-02 3.49908501e-01
6.80741906e-01 -3.42273057e-01 6.28598332e-01 -1.08174133e+00
-8.82759035e-01 -6.86592937e-01 -5.73075861e-02 -4.14184391e-01
-3.82047981e-01 -5.39466500e-01 2.92253550e-02 -1.19757533e+00
6.25172496e-01 -6.43330216e-01 -4.64155674e-01 -2.33144507e-01
-4.38815713e-01 -4.44476992e-01 -1.68928042e-01 4.37579095e-01
-8.80311430e-02 5.30285574e-02 9.42557633e-01 4.89921719e-01
-4.96861577e-01 -3.82506192e-01 -4.36488360e-01 5.31255245e-01
7.60021031e-01 -3.90940905e-01 -4.95181352e-01 -2.59672761e-01
6.07592463e-01 4.48373817e-02 2.20979288e-01 -6.40039802e-01
-5.40654242e-01 -8.38800728e-01 -1.96473137e-01 -3.77196312e-01
-1.16917379e-01 -8.69156420e-01 1.09785450e+00 1.61713219e+00
2.27679819e-01 2.15075374e-01 6.34387359e-02 6.74980164e-01
4.14262027e-01 2.96321541e-01 6.00810647e-01 -9.46410149e-02
1.88330099e-01 -1.08035933e-02 -7.79569149e-01 7.24534467e-02
8.76591325e-01 -1.47377104e-01 -9.19606209e-01 5.07908762e-01
-2.03694344e-01 -2.99696892e-01 1.03887987e+00 2.98392385e-01
-3.92351970e-02 -9.03859854e-01 -8.38031232e-01 1.83616310e-01
-6.94510788e-02 -4.80699986e-01 7.47825623e-01 1.05313420e+00
-1.29749751e+00 5.93203843e-01 -9.88235921e-02 -9.76007462e-01
-1.04868674e+00 9.09936607e-01 5.28530717e-01 -1.58859000e-01
-1.96730673e-01 7.79186964e-01 5.24999797e-01 -5.36503613e-01
-2.02349290e-01 -1.37006506e-01 6.69521093e-02 -1.56114474e-01
4.34895396e-01 7.22740173e-01 -6.21069670e-01 -3.43014717e-01
-9.03793454e-01 7.61934519e-01 1.67484418e-01 2.11805731e-01
1.42256570e+00 -1.23353779e-01 -6.74520969e-01 6.30874574e-01
1.01747322e+00 1.42489180e-01 -1.08262455e+00 9.74515006e-02
-1.44795179e-01 -3.59334856e-01 -1.08860779e+00 -4.52422142e-01
-2.52103716e-01 2.34344453e-01 7.87023723e-01 1.52005196e-01
7.66660750e-01 2.45419741e-01 2.18557402e-01 3.06572944e-01
3.86692256e-01 -6.83264673e-01 1.36517182e-01 3.17176163e-01
7.01406598e-01 -7.63618588e-01 -3.11638325e-01 -4.83145177e-01
2.21511364e-01 2.13511392e-01 3.93579036e-01 -2.39309072e-01
9.83035743e-01 1.67417273e-01 -1.94025952e-02 -1.55533895e-01
-9.93225634e-01 3.99096668e-01 -2.43989334e-01 7.31177747e-01
6.10706806e-01 7.67788529e-01 -9.07767594e-01 3.41488928e-01
-3.87398988e-01 -4.69939679e-01 5.26230812e-01 8.79678488e-01
-7.20406532e-01 -1.36895287e+00 -7.39533365e-01 5.31374276e-01
-3.69730681e-01 -2.37411082e-01 -2.03853965e-01 6.49525106e-01
1.08290911e-01 9.16073740e-01 3.83816093e-01 4.11864668e-02
1.02034114e-01 1.38731763e-01 5.16231596e-01 -3.92243445e-01
-8.60751033e-01 3.99311453e-01 -6.31456599e-02 -8.49701315e-02
-6.20122194e-01 -1.27302372e+00 -1.04161346e+00 -7.10149348e-01
-1.94478147e-02 5.02109602e-02 5.07271826e-01 3.78663927e-01
5.43258369e-01 6.44406259e-01 4.51865375e-01 -4.00410444e-02
-4.13524806e-01 -5.66430688e-01 -7.78092384e-01 3.49299371e-01
3.70861262e-01 -4.88707930e-01 -8.00566971e-01 -8.98595825e-02]
|
[5.663492679595947, 4.534782409667969]
|
fdfce4e5-d75c-42ac-9188-7c4ff0ea6bab
|
reliability-hierarchical-memory-network-for
|
2303.14384
| null |
https://arxiv.org/abs/2303.14384v1
|
https://arxiv.org/pdf/2303.14384v1.pdf
|
Reliability-Hierarchical Memory Network for Scribble-Supervised Video Object Segmentation
|
This paper aims to solve the video object segmentation (VOS) task in a scribble-supervised manner, in which VOS models are not only trained by the sparse scribble annotations but also initialized with the sparse target scribbles for inference. Thus, the annotation burdens for both training and initialization can be substantially lightened. The difficulties of scribble-supervised VOS lie in two aspects. On the one hand, it requires the powerful ability to learn from the sparse scribble annotations during training. On the other hand, it demands strong reasoning capability during inference given only a sparse initial target scribble. In this work, we propose a Reliability-Hierarchical Memory Network (RHMNet) to predict the target mask in a step-wise expanding strategy w.r.t. the memory reliability level. To be specific, RHMNet first only uses the memory in the high-reliability level to locate the region with high reliability belonging to the target, which is highly similar to the initial target scribble. Then it expands the located high-reliability region to the entire target conditioned on the region itself and the memories in all reliability levels. Besides, we propose a scribble-supervised learning mechanism to facilitate the learning of our model to predict dense results. It mines the pixel-level relation within the single frame and the frame-level relation within the sequence to take full advantage of the scribble annotations in sequence training samples. The favorable performance on two popular benchmarks demonstrates that our method is promising.
|
['Zhenyu He', 'YaoWei Wang', 'Hongpeng Wang', 'Wenjie Pei', 'Kaige Mao', 'Zikun Zhou']
|
2023-03-25
| null | null | null | null |
['video-object-segmentation', 'video-semantic-segmentation']
|
['computer-vision', 'computer-vision']
|
[ 7.28411973e-02 2.92556822e-01 -3.13619912e-01 -4.63764250e-01
-6.95537388e-01 -1.26950666e-01 -2.24701618e-03 -1.96981892e-01
-1.42440706e-01 5.70285380e-01 -8.41376558e-02 1.79388866e-01
6.50161207e-02 -8.15493643e-01 -9.48573351e-01 -8.33199382e-01
2.95827925e-01 6.97920799e-01 7.58515358e-01 1.87154729e-02
4.50428426e-02 3.08771163e-01 -1.50187445e+00 4.14175302e-01
1.08578205e+00 1.36318493e+00 7.50289381e-01 3.00861031e-01
-1.74149007e-01 1.02996588e+00 -4.88373935e-01 4.87788655e-02
1.29173040e-01 -2.94540048e-01 -7.64314532e-01 4.61410284e-01
3.25353742e-01 -3.58530372e-01 -1.18058830e-01 1.03630900e+00
1.01367116e-01 1.96159080e-01 5.49607575e-01 -9.72414672e-01
-1.05743922e-01 5.71327627e-01 -1.12027097e+00 1.40384153e-01
3.28032067e-03 1.57408267e-01 9.71381843e-01 -1.07903373e+00
6.61641181e-01 1.11834443e+00 4.15373862e-01 3.84092122e-01
-7.38974988e-01 -7.55365968e-01 4.22617137e-01 2.00635836e-01
-1.46649396e+00 -2.66897589e-01 8.98609817e-01 -5.84655046e-01
5.32095850e-01 -3.29466350e-02 7.60050297e-01 6.74809217e-01
-4.21962701e-03 1.07967854e+00 1.03528142e+00 -1.21168181e-01
2.74627745e-01 -8.34535956e-02 3.02750736e-01 1.10857975e+00
3.64166591e-03 -7.34754577e-02 -5.67100883e-01 2.64143586e-01
1.04261124e+00 -1.21516380e-02 -3.98219019e-01 -1.85293466e-01
-9.84719992e-01 6.68526828e-01 5.54384410e-01 2.58038014e-01
-3.40419292e-01 1.93629250e-01 2.13544548e-01 -2.64966428e-01
4.28433239e-01 2.57229414e-02 -6.04785800e-01 1.54307038e-01
-1.30302966e+00 -2.65008897e-01 5.06580889e-01 9.37884271e-01
1.14294398e+00 1.82961166e-01 -2.64820069e-01 9.18447435e-01
4.95129734e-01 2.38448068e-01 1.74551666e-01 -1.06565142e+00
5.31203568e-01 6.30339861e-01 -1.76964507e-01 -8.57515574e-01
-2.56400853e-01 -5.39246857e-01 -1.12499654e+00 2.72993684e-01
2.35260457e-01 -6.85177818e-02 -1.30048776e+00 1.50374401e+00
5.19931138e-01 6.13607824e-01 -6.33730963e-02 1.09476268e+00
1.17585444e+00 1.07579589e+00 -1.70342568e-02 -4.73182023e-01
1.25202060e+00 -1.36610889e+00 -6.36421621e-01 -2.56953627e-01
3.02546710e-01 -4.92229789e-01 9.94350016e-01 4.14010555e-01
-1.17492151e+00 -9.13759947e-01 -9.94206190e-01 -6.14511222e-02
1.96147203e-01 2.63112694e-01 4.96129751e-01 2.07736343e-01
-9.66735661e-01 3.38251323e-01 -7.91417837e-01 8.03652927e-02
7.76230097e-01 3.45056862e-01 9.82155558e-03 -1.51706800e-01
-9.19102609e-01 5.17866433e-01 7.01781631e-01 2.64858156e-01
-1.38906312e+00 -6.20749056e-01 -9.54909563e-01 2.18477547e-01
7.82141149e-01 -6.17995739e-01 8.88419867e-01 -1.14220393e+00
-1.47207177e+00 7.15175748e-01 -3.61698359e-01 -3.09625357e-01
3.49724025e-01 -3.40331569e-02 9.08552408e-02 3.10697317e-01
1.94977447e-01 9.57219660e-01 1.13957846e+00 -1.63504231e+00
-8.20164382e-01 -1.68954074e-01 -8.08534846e-02 2.99574941e-01
-1.53857633e-01 -2.03743741e-01 -1.05030286e+00 -8.28018546e-01
3.84904236e-01 -6.87762380e-01 -2.62878329e-01 -2.71698207e-01
-4.38940465e-01 -2.50776321e-01 8.68697107e-01 -7.05960751e-01
1.23894072e+00 -2.06377006e+00 4.25490201e-01 4.40156788e-01
3.09161991e-01 1.91136420e-01 4.44109887e-02 -3.32789809e-01
8.98200274e-02 -6.96383789e-02 -4.08502609e-01 -3.98082465e-01
-5.04073381e-01 4.39857513e-01 -2.64890909e-01 1.48535043e-01
2.35495150e-01 9.81202722e-01 -9.14563239e-01 -1.01893485e+00
2.32736424e-01 1.80157378e-01 -4.09259945e-01 5.26589215e-01
-4.58820403e-01 6.79398537e-01 -4.96225238e-01 7.45966971e-01
5.83797395e-01 -5.11983097e-01 -4.44482639e-02 -5.41858494e-01
-1.07047670e-02 -1.25708342e-01 -1.06447291e+00 1.86793208e+00
-1.98244378e-01 2.80141354e-01 1.26053348e-01 -9.73508656e-01
1.07207394e+00 2.16845989e-01 5.66627026e-01 -4.21114355e-01
1.02232538e-01 9.88120437e-02 -1.26200259e-01 -5.16984820e-01
2.35045865e-01 -8.30552801e-02 2.08652213e-01 2.09794030e-01
1.00927196e-01 -1.76583737e-01 1.00392602e-01 2.99268872e-01
8.49027872e-01 3.91277581e-01 1.41338363e-01 -2.16138557e-01
6.16958678e-01 1.14853447e-02 1.06464648e+00 4.88249749e-01
2.39627697e-02 7.49985158e-01 4.21063781e-01 -3.14913005e-01
-8.13257456e-01 -9.14758205e-01 -1.63379367e-02 9.06678498e-01
5.34175932e-01 -3.28734070e-01 -8.59980643e-01 -8.92199576e-01
-4.28144604e-01 2.71070510e-01 -6.18075252e-01 1.13494070e-02
-6.65932953e-01 -5.78720331e-01 2.24060744e-01 6.97436094e-01
9.39826488e-01 -1.25391257e+00 -4.36636984e-01 1.00864228e-02
-2.57959932e-01 -1.20807397e+00 -5.54838061e-01 8.75066817e-02
-1.03897047e+00 -9.64044631e-01 -7.01781809e-01 -1.00779891e+00
9.43436384e-01 1.12872608e-01 1.16357994e+00 4.88627970e-01
1.00643001e-01 8.84187818e-02 -4.00070429e-01 -2.82150097e-02
-1.65281579e-01 2.18611322e-02 -3.37532401e-01 1.38638735e-01
-1.10393047e-01 -4.81771439e-01 -5.51938534e-01 5.30573726e-01
-7.69632399e-01 5.35470247e-01 7.60313153e-01 8.92838776e-01
1.08274174e+00 5.08854687e-01 6.36760533e-01 -8.68830025e-01
-1.00530997e-01 -4.68703866e-01 -6.07908785e-01 3.60560566e-01
-4.07970160e-01 -6.88597262e-02 2.76748866e-01 -4.57798600e-01
-1.26632547e+00 4.15441573e-01 -2.74622083e-01 -7.05937326e-01
-5.09381332e-02 5.15093505e-01 -5.13452172e-01 1.22940898e-01
2.59580284e-01 1.78031117e-01 -2.67827898e-01 -2.91157216e-01
1.80767715e-01 3.26179385e-01 6.25291348e-01 -8.49566102e-01
6.79805636e-01 5.44843554e-01 -1.19111113e-01 -4.91585881e-01
-1.22633934e+00 -3.91262442e-01 -6.58807695e-01 -5.58197856e-01
1.12351394e+00 -9.43836868e-01 -6.19821072e-01 4.91172075e-01
-1.18326020e+00 -5.97167075e-01 -4.53537077e-01 2.25515783e-01
-4.40102935e-01 4.10500288e-01 -8.62051308e-01 -7.86207199e-01
-3.47460330e-01 -1.32584977e+00 1.14786518e+00 4.36020941e-01
4.93715294e-02 -7.27223098e-01 -4.00276721e-01 6.78264439e-01
-2.33778656e-01 1.16587497e-01 7.11555123e-01 -2.27207780e-01
-1.09658122e+00 2.78110534e-01 -4.91685778e-01 5.08009911e-01
-4.48151939e-02 -7.06882030e-02 -8.94055128e-01 -7.49687031e-02
1.98647380e-01 -3.60939443e-01 1.12802076e+00 6.77055359e-01
1.41827536e+00 -7.11906403e-02 -1.88850597e-01 7.41586804e-01
1.12546098e+00 2.69996017e-01 7.32604861e-01 1.59114867e-01
1.16102087e+00 5.65448046e-01 1.06702054e+00 2.66353041e-01
4.94145334e-01 6.04606628e-01 4.33253258e-01 -3.55312735e-01
-3.65079314e-01 -3.56000006e-01 3.27716976e-01 9.07029510e-01
-1.14742510e-01 -5.81321977e-02 -7.91163981e-01 4.93338525e-01
-2.10889292e+00 -7.11145222e-01 -2.27217764e-01 2.00580621e+00
1.08820772e+00 2.39117369e-01 -5.12016937e-02 1.35838715e-02
8.36020052e-01 2.54440367e-01 -6.31337762e-01 7.83083290e-02
-1.36979967e-01 6.11834265e-02 1.68823183e-01 5.58530867e-01
-8.94849896e-01 1.18533468e+00 5.13855171e+00 1.24952900e+00
-7.61280119e-01 3.21095347e-01 1.20672464e+00 1.18099540e-01
-1.74809143e-01 8.88194889e-02 -9.93391156e-01 5.84317684e-01
2.48787105e-01 5.09799898e-01 2.58515239e-01 8.36869240e-01
3.00456345e-01 -5.76687455e-01 -1.09496462e+00 1.05334270e+00
2.00899884e-01 -1.38450789e+00 -2.98800576e-03 -2.82450736e-01
1.12288678e+00 -2.96155542e-01 -1.35303810e-01 1.77204445e-01
2.29955003e-01 -1.04129517e+00 8.19699407e-01 5.09489477e-01
7.62715280e-01 -5.98797321e-01 7.83943117e-01 5.86829484e-01
-1.49218261e+00 6.85781986e-02 -4.27567959e-01 1.19662859e-01
3.99558157e-01 1.02797282e+00 -5.83920181e-01 5.33118069e-01
7.67677307e-01 8.68062794e-01 -2.72004902e-01 8.89185846e-01
-5.26659489e-01 7.84084737e-01 -2.38998026e-01 5.35096765e-01
1.67951226e-01 -3.54223073e-01 2.50282943e-01 1.14122236e+00
2.69542664e-01 3.77373785e-01 6.37860775e-01 8.76201272e-01
-1.51581302e-01 -2.87226290e-02 -5.85779510e-02 4.26888973e-01
3.77937376e-01 1.29099894e+00 -1.00791323e+00 -6.97033525e-01
-1.93127543e-01 9.75342274e-01 3.35769117e-01 5.28525174e-01
-1.05442452e+00 2.07516342e-01 -5.75377643e-02 1.37271151e-01
3.43802929e-01 -1.66285396e-01 -7.36564636e-01 -1.06086183e+00
3.01952641e-02 -6.18284285e-01 5.04767537e-01 -9.82428372e-01
-1.13808942e+00 6.62588060e-01 -1.56503730e-02 -1.10735250e+00
2.14931667e-01 -1.57373011e-01 -5.76972127e-01 7.98981488e-01
-1.54108047e+00 -1.22076499e+00 -6.13380313e-01 7.27038026e-01
9.70866382e-01 -3.04139648e-02 3.21974665e-01 3.65491956e-01
-7.83568084e-01 2.87431777e-01 -5.56766689e-01 2.75747925e-01
2.92149484e-01 -9.49650109e-01 -1.82930097e-01 8.84337485e-01
2.12553903e-01 3.55938941e-01 2.55900085e-01 -1.02254510e+00
-8.20582926e-01 -1.18949926e+00 3.92080039e-01 -1.93212435e-01
4.44368869e-01 -1.44049153e-01 -1.05199838e+00 6.20675385e-01
-5.45388795e-02 2.36483529e-01 1.71587914e-01 -4.12367918e-02
-2.28498399e-01 -2.82814384e-01 -8.67612660e-01 3.58567715e-01
9.73448217e-01 -2.18686357e-01 -6.30335271e-01 3.47756445e-01
7.90598571e-01 -7.01519370e-01 -8.05848718e-01 8.17785978e-01
1.99948132e-01 -9.00580227e-01 8.26766133e-01 -9.49353874e-02
5.85052371e-01 -7.88334310e-01 7.81345516e-02 -8.91026855e-01
-1.40981168e-01 -3.28610539e-01 -4.92515832e-01 1.54134297e+00
3.10958862e-01 -1.41853437e-01 1.02691448e+00 4.82282728e-01
-4.25782710e-01 -1.12708092e+00 -9.08120573e-01 -5.00720263e-01
-3.26072812e-01 -4.79223430e-01 3.31648260e-01 6.61022305e-01
-4.29628342e-01 5.07273197e-01 -5.70409477e-01 3.88743848e-01
5.25668204e-01 3.31704646e-01 7.05183446e-01 -1.12074471e+00
-5.59069455e-01 -6.93826973e-02 -9.50702131e-02 -1.70301902e+00
2.17734635e-01 -6.75841987e-01 4.82680142e-01 -1.63327944e+00
5.96947908e-01 -9.95742857e-01 -1.76914215e-01 5.54055095e-01
-4.55734342e-01 2.17938825e-01 1.31869063e-01 5.43532014e-01
-9.16011631e-01 5.88877201e-01 1.56166279e+00 -2.88685232e-01
-3.77016187e-01 -5.08023705e-03 -4.42306310e-01 1.07233751e+00
6.24015868e-01 -4.59120095e-01 -6.38563931e-01 -3.93043220e-01
6.74900273e-03 1.86788127e-01 3.48147124e-01 -1.02330983e+00
3.25908393e-01 -3.62854540e-01 4.01344001e-01 -9.65763211e-01
4.56625283e-01 -9.07081604e-01 1.27881661e-01 2.09290624e-01
-7.85425156e-02 -2.90637881e-01 -6.44057170e-02 6.15213633e-01
-2.43497208e-01 -4.80673581e-01 8.94813776e-01 -2.56231576e-01
-1.02274549e+00 5.56030869e-01 1.49191460e-02 7.84336701e-02
1.20637071e+00 -4.02242839e-01 -1.09085158e-01 -1.63320139e-01
-6.98125422e-01 4.58927661e-01 4.06022400e-01 8.32702667e-02
9.03202653e-01 -9.94968712e-01 -4.20286417e-01 2.05843464e-01
-1.08924046e-01 8.55073094e-01 5.18543243e-01 8.94112825e-01
-3.49014312e-01 -1.12766527e-01 1.56135894e-02 -9.28607225e-01
-1.18351173e+00 6.92742527e-01 2.60498941e-01 -3.42462391e-01
-6.85014069e-01 1.14847386e+00 8.06428075e-01 5.98499514e-02
4.12350476e-01 -3.36625785e-01 -2.95102835e-01 -1.68226659e-01
3.29373986e-01 1.17551818e-01 -2.26306736e-01 -7.52357483e-01
-9.44991335e-02 8.36141825e-01 1.66430883e-02 3.64489574e-03
1.14069164e+00 -2.59696186e-01 -3.46536189e-01 4.72762764e-01
8.68488193e-01 -1.01769254e-01 -1.67614985e+00 -1.69527546e-01
-3.23359482e-02 -4.34314311e-01 6.60007522e-02 -7.78735220e-01
-1.64757848e+00 7.49225318e-01 3.21546018e-01 -4.01801705e-01
1.31587458e+00 2.57811606e-01 9.99335170e-01 9.39477682e-02
4.33916748e-01 -1.27988958e+00 4.37501878e-01 4.47433829e-01
6.18432045e-01 -1.22968662e+00 4.56257910e-02 -9.36773241e-01
-8.24608505e-01 8.91578019e-01 9.64734614e-01 -8.12177286e-02
6.33592784e-01 4.45563257e-01 -1.53544411e-01 -4.02057648e-01
-4.38876420e-01 -3.54574829e-01 4.70598280e-01 5.37959874e-01
6.57947734e-02 -1.32853046e-01 -1.55880973e-01 7.23623097e-01
2.23459870e-01 1.18245728e-01 1.80938065e-01 4.81180698e-01
-7.57619262e-01 -8.02911401e-01 -4.95098501e-01 5.36466241e-01
-1.43036366e-01 -1.21622853e-01 3.11382059e-02 2.48665348e-01
5.87008834e-01 9.83791649e-01 -9.27911028e-02 -4.31256652e-01
-2.37766709e-02 -2.00216159e-01 2.56822348e-01 -8.99987102e-01
-3.97296071e-01 3.68395358e-01 -2.63645314e-02 -5.41699886e-01
-5.79979420e-01 -4.17925090e-01 -1.96119201e+00 -8.05600733e-03
-4.50401783e-01 1.94393590e-01 2.43500352e-01 1.26738167e+00
1.16091833e-01 8.18812013e-01 4.26251292e-01 -9.10278499e-01
4.33908664e-02 -7.09440053e-01 -5.85224748e-01 4.12075520e-01
-5.81380464e-02 -7.33786047e-01 7.12078735e-02 2.83883840e-01]
|
[9.567122459411621, 0.003120080102235079]
|
8aa839f2-736f-4d38-8dbe-42582245e6a5
|
entity-extraction-from-wikipedia-list-pages
|
2003.05146
| null |
https://arxiv.org/abs/2003.05146v1
|
https://arxiv.org/pdf/2003.05146v1.pdf
|
Entity Extraction from Wikipedia List Pages
|
When it comes to factual knowledge about a wide range of domains, Wikipedia is often the prime source of information on the web. DBpedia and YAGO, as large cross-domain knowledge graphs, encode a subset of that knowledge by creating an entity for each page in Wikipedia, and connecting them through edges. It is well known, however, that Wikipedia-based knowledge graphs are far from complete. Especially, as Wikipedia's policies permit pages about subjects only if they have a certain popularity, such graphs tend to lack information about less well-known entities. Information about these entities is oftentimes available in the encyclopedia, but not represented as an individual page. In this paper, we present a two-phased approach for the extraction of entities from Wikipedia's list pages, which have proven to serve as a valuable source of information. In the first phase, we build a large taxonomy from categories and list pages with DBpedia as a backbone. With distant supervision, we extract training data for the identification of new entities in list pages that we use in the second phase to train a classification model. With this approach we extract over 700k new entities and extend DBpedia with 7.5M new type statements and 3.8M new facts of high precision.
|
['Heiko Paulheim', 'Nicolas Heist']
|
2020-03-11
| null | null | null | null |
['entity-extraction']
|
['natural-language-processing']
|
[-6.59639955e-01 6.41202450e-01 -6.99398398e-01 -4.95081842e-02
-3.15592676e-01 -9.68795419e-01 7.53063083e-01 9.01767015e-01
-5.12769818e-01 1.51193762e+00 2.27174118e-01 -2.02455148e-01
-3.81309241e-01 -1.45931256e+00 -9.93390441e-01 3.65633070e-02
-3.60435963e-01 8.97964180e-01 8.09040248e-01 -5.15897632e-01
1.25772953e-02 1.32715330e-01 -1.46720827e+00 3.43994617e-01
1.30417478e+00 7.57861376e-01 7.73482248e-02 -9.30386409e-02
-7.38904595e-01 8.25188637e-01 -6.13387227e-01 -1.03143692e+00
1.67918466e-02 2.45556369e-01 -1.29183483e+00 -3.03120404e-01
4.66943264e-01 4.49494198e-02 -3.68787080e-01 1.15280366e+00
2.58797854e-02 -1.72692046e-01 5.31616807e-01 -1.23927629e+00
-5.19980729e-01 1.01319778e+00 -2.78900921e-01 1.57824352e-01
6.41402900e-01 -5.21507144e-01 1.28326011e+00 -7.17482924e-01
1.31346405e+00 8.41984510e-01 6.32490635e-01 2.43550792e-01
-8.58316839e-01 -4.36515808e-01 3.97357624e-03 3.85648161e-01
-1.31400299e+00 -1.04910471e-01 4.10349995e-01 -5.52788317e-01
9.77590024e-01 1.48129836e-01 6.96298242e-01 8.15046787e-01
-2.18133941e-01 3.72531056e-01 9.21501696e-01 -5.12403965e-01
2.10283305e-02 8.47472072e-01 4.73586470e-01 7.11272001e-01
1.23942435e+00 -2.79613525e-01 -4.36085373e-01 -1.63755074e-01
3.39181244e-01 -2.37179846e-01 -4.08835471e-01 -6.88655078e-01
-9.29250062e-01 6.15292430e-01 6.05875194e-01 5.07986844e-01
-5.64163685e-01 -4.34953332e-01 4.25389081e-01 3.09191406e-01
3.97241592e-01 8.80711734e-01 -7.36680865e-01 1.17606424e-01
-5.02185822e-01 4.36259121e-01 1.46821988e+00 1.33892071e+00
1.08942044e+00 -7.13843286e-01 4.48338002e-01 8.46611083e-01
1.07350476e-01 3.84880126e-01 4.22283232e-01 -5.31772077e-01
8.27528834e-01 1.14734447e+00 3.48094791e-01 -1.21002603e+00
-4.10774469e-01 -4.48957175e-01 -4.78124738e-01 -3.37375432e-01
4.70260412e-01 -1.48244008e-01 -6.96643412e-01 1.50382340e+00
6.19744897e-01 -2.73091644e-01 2.09795490e-01 5.43204546e-01
1.15954781e+00 3.68619651e-01 1.35537088e-01 -1.25114381e-01
1.41090393e+00 -6.78173959e-01 -6.30846441e-01 -1.30780920e-01
6.97816491e-01 -2.48286754e-01 4.02203590e-01 2.14806572e-01
-6.73486829e-01 -1.76586255e-01 -8.52066755e-01 -3.27292159e-02
-1.32599592e+00 -2.86508858e-01 8.18442345e-01 3.04291129e-01
-6.37267292e-01 7.03064442e-01 -4.29377288e-01 -6.95112288e-01
2.59368718e-01 9.30414652e-04 -7.79589236e-01 -3.38886738e-01
-1.92335021e+00 1.39559722e+00 1.14132118e+00 -4.60238427e-01
-3.85142088e-01 -8.02274883e-01 -8.77951324e-01 1.35145904e-02
8.95863593e-01 -6.10473752e-01 6.95401013e-01 -5.02472341e-01
-5.89627028e-01 1.00973725e+00 2.10397050e-01 -7.32081890e-01
1.18555024e-01 -2.29405649e-02 -9.38448668e-01 1.31279364e-01
5.51598489e-01 2.65247792e-01 2.69261062e-01 -1.25637054e+00
-1.12775481e+00 -4.49295193e-01 7.68177450e-01 1.33821353e-01
-5.29516816e-01 -3.20533752e-01 -6.70739591e-01 -1.88351691e-01
-1.48022503e-01 -5.04925489e-01 7.80174062e-02 -6.20422125e-01
-5.67239165e-01 -3.61069411e-01 2.70044833e-01 -9.63235557e-01
1.71378767e+00 -1.74119151e+00 7.63494968e-02 4.29199129e-01
6.83061838e-01 2.06052959e-01 3.57948840e-01 6.75324619e-01
1.88651532e-01 2.97727138e-01 2.77435575e-02 5.79376221e-01
1.07045710e-01 3.81825417e-01 -1.28165200e-01 -2.21605718e-01
-1.37471303e-01 7.45134354e-01 -1.22908044e+00 -7.16395378e-01
-3.07867408e-01 -1.85450204e-02 -4.80100304e-01 -1.72042370e-01
-5.96984267e-01 -1.69375047e-01 -6.44158304e-01 5.50760269e-01
4.86095220e-01 -4.10979629e-01 7.40314245e-01 -3.77949804e-01
-9.56807435e-02 6.88774049e-01 -1.07616830e+00 1.51049733e+00
-3.30296218e-01 4.25562650e-01 -2.50672042e-01 -9.77458775e-01
6.55522108e-01 3.32281649e-01 5.31347573e-01 -4.42563742e-01
-2.24163771e-01 4.30947453e-01 -2.71478802e-01 -5.01486301e-01
7.97567785e-01 1.47972956e-01 -2.74999022e-01 -1.00301839e-02
5.46582162e-01 1.50749505e-01 1.00921416e+00 6.10778213e-01
1.23948002e+00 1.71562936e-02 7.27036357e-01 -1.50887877e-01
5.13672352e-01 8.01041603e-01 5.18403172e-01 4.29449409e-01
3.82410049e-01 -1.66262299e-01 6.42925739e-01 -3.99249703e-01
-1.14317513e+00 -1.00062716e+00 -3.52628499e-01 5.70127666e-01
9.04044062e-02 -8.29161286e-01 -5.09423614e-01 -1.08345127e+00
4.42805022e-01 7.02214897e-01 -4.81019408e-01 8.75913352e-02
-2.91303068e-01 -5.33648908e-01 3.30396175e-01 1.07917167e-01
6.13175035e-01 -7.05416918e-01 -9.86400098e-02 2.31285542e-01
-4.42330182e-01 -1.32271421e+00 2.84131497e-01 2.66306251e-01
-5.92470109e-01 -1.63919377e+00 -4.22700703e-01 -8.37604821e-01
7.20747292e-01 -7.25023448e-02 1.58568585e+00 3.37531604e-02
4.70768027e-02 2.24042118e-01 -5.70197701e-01 -4.95312005e-01
-4.36338246e-01 5.11673152e-01 1.31570011e-01 -5.43361425e-01
8.29586148e-01 -5.76461434e-01 -7.71849602e-02 1.11376740e-01
-8.75013471e-01 -2.36821324e-01 4.80362266e-01 6.50393844e-01
2.92611629e-01 5.43882608e-01 7.60096610e-01 -1.49873555e+00
4.63487029e-01 -1.10897040e+00 -6.09527171e-01 5.54693401e-01
-8.01062047e-01 1.25354022e-01 6.23225391e-01 5.17926030e-02
-1.00134385e+00 -2.59528011e-01 1.95420012e-01 5.83720468e-02
-1.21415131e-01 1.33014023e+00 -3.34570110e-01 1.79967150e-01
9.74472702e-01 -1.33599475e-01 -4.11421448e-01 -6.38703287e-01
5.79642475e-01 6.62358642e-01 4.53167766e-01 -7.56568134e-01
1.00620425e+00 1.42762512e-01 -1.92233458e-01 -7.07258642e-01
-1.20070589e+00 -8.03163171e-01 -8.73468637e-01 2.52045784e-03
6.22731984e-01 -1.06553090e+00 -3.50063831e-01 1.18508950e-01
-9.33531702e-01 8.15171376e-02 -3.37063104e-01 3.53245944e-01
2.34991666e-02 2.68682599e-01 -2.38485113e-01 -2.42111862e-01
-1.16777793e-02 -6.19559772e-02 1.87875643e-01 1.61963403e-01
-5.82346991e-02 -1.31852233e+00 1.74302176e-01 2.96391040e-01
8.31785277e-02 3.80993664e-01 1.10492492e+00 -1.31504595e+00
-6.61820650e-01 -4.27029222e-01 -1.50854394e-01 1.67898923e-01
2.33139783e-01 -2.64730573e-01 -4.05279070e-01 1.20357946e-02
-6.81064785e-01 -1.54667318e-01 6.08195841e-01 -2.76329815e-01
5.69277108e-01 -5.59618115e-01 -7.74955332e-01 1.90831974e-01
1.70005929e+00 -2.12631710e-02 4.94863510e-01 9.67292607e-01
6.15387678e-01 6.59603596e-01 6.86375201e-01 3.50759506e-01
7.71931708e-01 5.92500150e-01 2.36701384e-01 2.95769095e-01
-1.27582932e-02 -5.93078852e-01 -8.52686390e-02 8.20450902e-01
-3.34297061e-01 -7.50187263e-02 -1.09848022e+00 8.42036247e-01
-1.56199825e+00 -1.18115842e+00 -1.30390614e-01 2.44964075e+00
1.44733608e+00 3.73323202e-01 1.91598177e-01 -6.54250458e-02
7.32009053e-01 -1.64788082e-01 -2.56721616e-01 2.62344778e-01
-2.18426496e-01 5.17045185e-02 8.13761175e-01 3.79751831e-01
-1.10263884e+00 1.02978849e+00 5.22966337e+00 5.54778934e-01
-5.84962726e-01 2.90386043e-02 -1.66001871e-01 2.27882743e-01
-4.18251514e-01 3.37123543e-01 -1.20092046e+00 5.69674253e-01
8.72715056e-01 -7.88633466e-01 1.93387464e-01 1.07340527e+00
-5.42029321e-01 -2.65623063e-01 -9.31990325e-01 5.28151572e-01
-3.41147408e-02 -1.45813239e+00 1.29434913e-01 2.23515883e-01
7.33126998e-01 3.86534147e-02 -6.15539193e-01 6.07323587e-01
8.09234500e-01 -6.24636650e-01 4.80243921e-01 4.72111493e-01
7.54938245e-01 -6.62569344e-01 8.65574896e-01 3.95408750e-01
-1.07263064e+00 5.22974692e-02 -5.95596969e-01 2.23640233e-01
2.95127314e-02 1.07501161e+00 -9.79380667e-01 1.12742150e+00
8.76155019e-01 9.26959753e-01 -6.61270201e-01 1.22880983e+00
-5.56398571e-01 2.55697608e-01 -2.79193580e-01 -1.22615740e-01
3.90292867e-03 -5.07273898e-02 5.83158016e-01 1.08513892e+00
2.36290708e-01 8.83587077e-02 8.90698209e-02 4.99364913e-01
-5.72252691e-01 1.23880863e-01 -9.00859594e-01 -1.68902323e-01
9.03008461e-01 1.18685555e+00 -3.80582482e-01 -6.48849726e-01
-6.71449959e-01 6.22132897e-01 6.66387022e-01 2.35766247e-01
-6.36206388e-01 -9.22809124e-01 4.89547759e-01 3.69760096e-01
2.38972664e-01 -1.57836691e-01 3.59357148e-01 -1.48164451e+00
2.14237124e-01 -7.40843415e-01 7.51734078e-01 -6.08010709e-01
-1.39849544e+00 4.88344938e-01 2.04477385e-01 -1.10481679e+00
-3.66385579e-01 -7.37617493e-01 1.21503547e-01 7.06061006e-01
-1.86392057e+00 -1.02469838e+00 -3.28699261e-01 5.43041468e-01
-1.21291459e-01 -1.00929089e-01 8.47636759e-01 6.25166774e-01
-3.03802609e-01 2.29966685e-01 3.57622117e-01 6.35169983e-01
8.61665308e-01 -1.61241591e+00 3.54835749e-01 5.45377016e-01
3.22611213e-01 8.47397745e-01 7.29936957e-01 -1.14105082e+00
-1.18906879e+00 -1.04378986e+00 1.34017122e+00 -8.87411535e-01
1.05423796e+00 -1.48570061e-01 -1.12196386e+00 9.90547121e-01
6.62715957e-02 -1.35139450e-01 5.51456451e-01 6.73381507e-01
-6.04807794e-01 -1.35362893e-01 -9.99001682e-01 1.89011008e-01
1.07077670e+00 -3.53501230e-01 -1.29930532e+00 4.50980723e-01
6.79705918e-01 -5.79516530e-01 -1.36875033e+00 9.72489491e-02
2.17659220e-01 -5.02453804e-01 8.94985378e-01 -7.83789396e-01
5.27791142e-01 -3.09059143e-01 3.49998437e-02 -1.73752189e+00
-2.35316709e-01 -1.06348149e-01 -4.01249975e-01 1.50973272e+00
1.06778646e+00 -8.17058921e-01 6.72782183e-01 5.46348274e-01
-5.22752739e-02 -3.05336952e-01 -6.06168211e-01 -1.14825034e+00
5.51823191e-02 5.54530732e-02 6.07977331e-01 1.41421843e+00
6.99777186e-01 5.35355508e-01 -1.75021980e-02 2.55818009e-01
6.21853709e-01 2.08004758e-01 7.54281878e-01 -1.84967279e+00
4.97963354e-02 -9.57231447e-02 -6.27813280e-01 -6.26174569e-01
2.38903854e-02 -1.16521561e+00 -5.05337119e-01 -1.96527970e+00
3.08653980e-01 -6.59720778e-01 -1.21874370e-01 5.45753539e-01
1.61130950e-02 -2.99942970e-01 -2.14624166e-01 1.63426057e-01
-7.11862147e-01 2.71454640e-02 8.82884920e-01 -5.66984862e-02
-4.36335877e-02 -3.29491854e-01 -8.13793659e-01 7.69458532e-01
5.31475782e-01 -4.77415144e-01 -2.48723999e-01 -2.18777478e-01
7.97429264e-01 -2.56396800e-01 2.23467663e-01 -9.97443616e-01
4.23683405e-01 -1.84316918e-01 1.68629080e-01 -3.66974384e-01
5.03477193e-02 -1.05675089e+00 1.45228744e-01 1.16999231e-01
-4.16716114e-02 -3.57611597e-01 1.37483820e-01 6.13196313e-01
-6.74703479e-01 -3.86412740e-01 2.92996287e-01 -4.75109488e-01
-1.20925164e+00 2.28758261e-01 1.99885219e-01 6.26255512e-01
1.04508185e+00 3.76576930e-02 -7.28277147e-01 2.94434223e-02
-8.57203603e-01 4.18165773e-01 6.17762327e-01 4.29018408e-01
1.01851717e-01 -1.28054404e+00 -5.32034218e-01 -3.79036725e-01
4.32965070e-01 1.37189910e-01 5.38254157e-03 5.02802670e-01
-4.79160339e-01 7.16013610e-01 -4.44290459e-01 -9.05011818e-02
-9.57968414e-01 9.43270326e-01 1.01571165e-01 -6.03024244e-01
-5.73709786e-01 3.73295188e-01 -2.24637672e-01 -5.19830465e-01
4.04736027e-02 8.03130027e-03 -7.86353290e-01 5.27610838e-01
4.16905731e-01 1.98423997e-01 3.19202393e-01 -3.39365602e-01
-3.53459448e-01 1.82445407e-01 -2.37274557e-01 1.70417532e-01
1.58361912e+00 -2.43022308e-01 -4.22530830e-01 4.14909184e-01
7.50528753e-01 4.27918941e-01 -4.41288620e-01 -4.85678017e-01
7.16466129e-01 -4.06121582e-01 -4.43675190e-01 -1.09290516e+00
-7.88562655e-01 2.92725086e-01 -2.80897915e-01 7.33363807e-01
8.23666453e-01 3.60432237e-01 5.54515898e-01 6.00297391e-01
9.24703479e-01 -1.25045216e+00 -4.64928448e-01 7.09638476e-01
7.25750685e-01 -1.20462716e+00 1.80508897e-01 -8.03182304e-01
-4.80491787e-01 9.18372571e-01 8.06606829e-01 2.78047353e-01
7.91013300e-01 4.57272679e-03 -2.39980564e-01 -4.89834100e-01
-8.25675964e-01 -5.10902703e-01 3.09672266e-01 7.39164710e-01
3.87920827e-01 6.70185089e-02 -5.60549557e-01 7.83740461e-01
-3.47118706e-01 5.05588725e-02 5.87020338e-01 8.82480085e-01
-7.54419982e-01 -1.12526822e+00 -7.64869303e-02 7.68525481e-01
-5.64387977e-01 -3.20890754e-01 -4.03598160e-01 1.19126737e+00
3.05129170e-01 5.97564757e-01 -2.92551368e-02 -2.71552801e-01
5.41856647e-01 2.73328483e-01 2.52608299e-01 -8.99464190e-01
-3.01590919e-01 -7.91075528e-01 8.31288934e-01 -2.23070607e-01
-4.22473401e-01 -3.46175373e-01 -1.24783802e+00 -4.26340997e-01
-2.68769920e-01 8.03894877e-01 5.90139210e-01 8.27553451e-01
4.09698904e-01 3.73082608e-01 2.04652742e-01 6.75736368e-02
-2.05361903e-01 -7.77916729e-01 -7.94034839e-01 6.58476114e-01
-1.43353701e-01 -8.56783152e-01 -1.15628310e-01 8.09917301e-02]
|
[9.313729286193848, 8.354809761047363]
|
95b4e294-1923-42f2-bef3-7ba7136e135c
|
wide-baseline-stereo-matching-with-convex-1
| null | null |
http://openaccess.thecvf.com/content_iccv_2015/html/Galun_Wide_Baseline_Stereo_ICCV_2015_paper.html
|
http://openaccess.thecvf.com/content_iccv_2015/papers/Galun_Wide_Baseline_Stereo_ICCV_2015_paper.pdf
|
Wide Baseline Stereo Matching With Convex Bounded Distortion Constraints
|
Finding correspondences in wide baseline setups is a challenging problem. Existing approaches have focused largely on developing better feature descriptors for correspondence and on accurate recovery of epipolar line constraints. This paper focuses on the challenging problem of finding correspondences once approximate epipolar constraints are given. We introduce a novel method that integrates a deformation model. Specifically, we formulate the problem as finding the largest number of corresponding points related by a bounded distortion map that obeys the given epipolar constraints. We show that, while the set of bounded distortion maps is not convex, the subset of maps that obey the epipolar line constraints is convex, allowing us to introduce an efficient algorithm for matching. We further utilize a robust cost function for matching and employ majorization-minimization for its optimization. Our experiments indicate that our method finds significantly more accurate maps than existing approaches.
|
['Tal Amir', 'Yaron Lipman', 'Meirav Galun', 'Tal Hassner', 'Ronen Basri']
|
2015-12-01
| null | null | null |
iccv-2015-12
|
['stereo-matching']
|
['computer-vision']
|
[-2.28893962e-02 -9.96737257e-02 -1.67305365e-01 -3.40827495e-01
-9.33618963e-01 -8.08804989e-01 5.43555915e-01 1.62212938e-01
-3.31149817e-01 4.46518272e-01 2.12797523e-01 8.07916820e-02
-2.65008330e-01 -6.10847950e-01 -9.55955446e-01 -3.77559513e-01
8.04189816e-02 6.07775688e-01 2.04738110e-01 -2.84072697e-01
6.96372330e-01 8.33253324e-01 -1.26659024e+00 -5.32747805e-01
6.05647206e-01 5.51960647e-01 -9.32331383e-03 6.85178280e-01
4.60071683e-01 -4.84089293e-02 -4.62375432e-01 -3.71529818e-01
8.52109432e-01 -3.19411218e-01 -8.86239409e-01 3.22712660e-01
1.11928344e+00 -4.53572154e-01 -4.78787959e-01 1.05360937e+00
4.64714229e-01 3.50111365e-01 5.84375322e-01 -1.51933157e+00
-2.72260517e-01 -1.30014062e-01 -6.72217488e-01 1.10297188e-01
6.60680234e-01 -3.87071341e-01 1.31314385e+00 -1.09725416e+00
8.42099965e-01 9.62992430e-01 7.87137449e-01 1.15901902e-01
-1.34839594e+00 -4.22036529e-01 -1.54132664e-01 -5.22065535e-02
-1.78053308e+00 -7.02408016e-01 6.60498798e-01 -5.38145959e-01
8.35471213e-01 2.93769926e-01 5.41200817e-01 3.40258807e-01
5.91007173e-02 2.69195884e-01 5.38611710e-01 -6.97123170e-01
8.56676921e-02 -4.26084884e-02 2.26077214e-01 6.83188260e-01
4.89865005e-01 3.40106636e-02 -5.61296642e-01 -3.06856841e-01
1.15896189e+00 -6.20427392e-02 -3.19349498e-01 -8.31229270e-01
-1.37486720e+00 9.46543157e-01 3.98826540e-01 1.28118157e-01
-1.98364303e-01 1.96341962e-01 -1.53768227e-01 2.67186105e-01
3.69023800e-01 7.07713485e-01 -1.51118800e-01 6.01524636e-02
-1.02585971e+00 6.20306432e-01 9.43790436e-01 1.46067715e+00
1.27923477e+00 -5.83212554e-01 4.09230202e-01 7.11143076e-01
5.22594564e-02 5.84624887e-01 -2.56016076e-01 -1.29206717e+00
5.77554226e-01 4.26890850e-01 3.37436527e-01 -1.54260492e+00
-3.61459523e-01 -8.96457285e-02 -4.33043987e-01 2.27242559e-02
6.85595691e-01 -1.40273925e-02 -4.21401143e-01 1.69170809e+00
3.70432824e-01 9.66569334e-02 -1.96652468e-02 8.39686215e-01
2.21994132e-01 3.85925472e-01 -8.22501183e-01 -1.12233713e-01
1.03938198e+00 -7.29899466e-01 -3.99352521e-01 -1.75458893e-01
5.98887026e-01 -1.19639826e+00 5.53943396e-01 -1.23464905e-01
-1.32683921e+00 -2.77208418e-01 -1.25835061e+00 -2.86292315e-01
1.93466961e-01 9.63086486e-02 2.95576185e-01 2.53555804e-01
-1.30169094e+00 6.14718139e-01 -7.89805889e-01 -7.00947642e-01
-7.23049343e-02 7.55824089e-01 -7.40317225e-01 3.48170586e-02
-5.44528186e-01 1.08745289e+00 -3.37723196e-02 2.36560125e-02
-1.87919497e-01 -5.75806916e-01 -9.97473001e-01 -1.80148572e-01
3.26211482e-01 -9.36751485e-01 1.11277092e+00 -4.68010813e-01
-1.16609323e+00 1.06936777e+00 -5.81088305e-01 -2.74615079e-01
6.01590931e-01 -1.80557892e-01 3.42579007e-01 3.27213407e-01
2.85959870e-01 5.52263260e-01 4.12362248e-01 -1.25479639e+00
-5.83533943e-01 -3.96351844e-01 3.92375320e-01 4.86721069e-01
-2.65606761e-01 2.05489159e-01 -8.85141730e-01 -4.45375741e-01
7.13739634e-01 -1.33635402e+00 -3.43497992e-01 2.75840968e-01
-4.90874708e-01 -3.34718972e-02 4.05064255e-01 -4.49312687e-01
1.05136359e+00 -1.84640634e+00 3.27285677e-01 7.41961837e-01
2.69855559e-01 -2.48106718e-01 5.25647365e-02 6.47412419e-01
1.35952845e-01 -8.65862742e-02 -1.72216684e-01 -4.62024301e-01
-7.87811819e-03 1.55247882e-01 -3.06214452e-01 9.99976814e-01
1.06819365e-02 6.12318933e-01 -5.99941850e-01 -4.47749704e-01
2.11721450e-01 3.17182750e-01 -7.77670503e-01 6.44474253e-02
2.29516715e-01 2.99712867e-01 -4.06239629e-01 4.35017556e-01
1.01316893e+00 6.02947958e-02 -1.37370512e-01 -3.75499278e-01
-4.08486009e-01 6.17706515e-02 -1.62460423e+00 1.86942244e+00
-1.38478920e-01 8.68970513e-01 5.74117824e-02 -8.71764064e-01
1.14758122e+00 -3.80937867e-02 9.38350260e-01 -1.80178285e-01
2.50284076e-01 4.59342360e-01 -2.14207515e-01 -2.56292317e-02
7.56735504e-01 -2.06445336e-01 3.27837057e-02 4.48396236e-01
-3.56078863e-01 -2.42275029e-01 1.64364815e-01 1.30553678e-01
1.01624298e+00 9.53540429e-02 3.74428332e-01 -4.08330023e-01
3.47273260e-01 3.97713155e-01 6.49680376e-01 5.34515083e-01
-2.02471800e-02 9.88926411e-01 2.12166563e-01 -3.36903602e-01
-1.42124915e+00 -8.30558836e-01 -4.20705378e-01 3.68307114e-01
6.93460882e-01 -4.55158532e-01 -7.84183681e-01 -8.52369666e-02
2.58780599e-01 -4.12650593e-02 -3.83572519e-01 2.70115197e-01
-9.52516675e-01 -4.53834683e-01 2.67321616e-01 5.65213323e-01
4.13874835e-01 -1.94490194e-01 -5.71126223e-01 1.02212861e-01
-2.65684217e-01 -1.08572042e+00 -1.03016484e+00 -2.69482642e-01
-8.52311552e-01 -1.37927759e+00 -8.79516602e-01 -1.03434265e+00
1.16233420e+00 8.22253108e-01 9.04840887e-01 1.90224215e-01
-2.91462958e-01 5.35564840e-01 -1.36590749e-01 -1.09200597e-01
2.43727230e-02 5.74443936e-02 1.97236523e-01 -1.23271413e-01
4.35855061e-01 -5.12302220e-01 -5.30832350e-01 8.56083691e-01
-4.57257479e-01 -7.53690302e-02 1.36724919e-01 6.30623877e-01
9.84782875e-01 -2.75846541e-01 2.64540967e-02 -4.25275117e-01
2.93485254e-01 -1.25336453e-01 -8.84781539e-01 2.70408183e-01
-3.54924649e-01 1.51209027e-01 2.37126052e-01 -1.41495049e-01
-5.14443934e-01 4.83704060e-01 2.32637718e-01 -4.63186711e-01
2.56650180e-01 2.26033673e-01 -1.25958800e-01 -7.54277229e-01
6.74788654e-01 -1.03869168e-02 6.17398433e-02 -3.12149793e-01
2.71555156e-01 4.39610451e-01 7.81961024e-01 -8.13905299e-01
1.31534588e+00 8.31600010e-01 3.76923412e-01 -9.32343781e-01
-5.87045193e-01 -9.86377180e-01 -1.26093507e+00 -7.98227787e-02
5.44118762e-01 -8.64546835e-01 -6.20010793e-01 7.30599239e-02
-1.30630004e+00 1.90310255e-01 -1.22606017e-01 8.21062803e-01
-1.05815613e+00 6.60441935e-01 -2.14867830e-01 -3.78976196e-01
-1.70633391e-01 -1.22868097e+00 1.17328453e+00 4.86869067e-02
-4.60521132e-01 -9.96573627e-01 6.14479959e-01 2.14691937e-01
4.92870547e-02 4.89361256e-01 3.77837300e-01 -3.94842356e-01
-9.12346184e-01 -5.28618634e-01 -2.64558345e-01 -1.91029847e-01
1.17474720e-01 1.31527066e-01 -4.50093478e-01 -4.87551749e-01
-5.98652363e-02 1.54720470e-01 4.26819861e-01 5.44317901e-01
3.26313913e-01 -1.11878924e-01 -5.47449529e-01 1.20395076e+00
1.54123878e+00 2.95600854e-03 3.87689739e-01 6.90009296e-01
5.94307721e-01 7.11333752e-01 7.53112376e-01 2.87490755e-01
6.69582248e-01 1.12093163e+00 1.00470103e-01 -1.80724710e-01
8.74141008e-02 -2.48861894e-01 -1.05849303e-01 7.74311721e-01
-1.99539408e-01 1.48699693e-02 -8.80622804e-01 7.22101569e-01
-2.16627312e+00 -8.18208277e-01 -2.62368292e-01 2.55247450e+00
5.47638237e-01 -2.78663188e-01 2.36447290e-01 -7.87291527e-02
8.86771917e-01 -1.61385581e-01 -4.46210653e-01 -1.86203867e-01
-2.77571201e-01 1.13917023e-01 7.21054196e-01 1.00072217e+00
-1.13513267e+00 8.07951808e-01 7.03484201e+00 3.01824734e-02
-6.16064489e-01 -3.27314019e-01 -1.71859711e-01 1.35424450e-01
-2.89338261e-01 4.18625981e-01 -1.09718585e+00 -2.75280997e-02
2.49348342e-01 -7.51527607e-01 2.43287414e-01 7.33560681e-01
2.67873019e-01 -2.49567181e-01 -1.36697757e+00 9.44097281e-01
3.29571217e-01 -1.41803551e+00 -1.91787794e-01 3.81150991e-01
9.65248883e-01 -8.27851295e-02 -1.31681725e-01 -5.86119235e-01
2.65265584e-01 -6.14674568e-01 5.47576368e-01 3.13989073e-01
6.81599081e-01 -7.33308911e-01 3.97017300e-01 1.36146858e-01
-1.42793977e+00 5.32456696e-01 -7.83796132e-01 -1.09910071e-01
1.18435137e-01 1.57467544e-01 -7.20370352e-01 7.39582419e-01
2.26840585e-01 9.98229504e-01 -2.77299434e-01 1.45881307e+00
1.26943514e-01 -1.99413761e-01 -7.24458814e-01 3.48208815e-01
-3.47051509e-02 -5.43332040e-01 6.82659686e-01 1.06112087e+00
5.01012743e-01 2.47769251e-01 2.82446384e-01 5.91470301e-01
-7.38514289e-02 3.88719201e-01 -8.66126716e-01 5.42994082e-01
6.56887650e-01 9.03124869e-01 -7.24030912e-01 -2.86725396e-03
-6.21147394e-01 1.03710997e+00 3.83460075e-01 2.28202388e-01
-4.95616198e-01 -5.69745839e-01 8.89513135e-01 1.33553833e-01
6.45177662e-02 -5.66756070e-01 -3.97214055e-01 -1.36216724e+00
4.27954316e-01 -5.40022671e-01 3.53172988e-01 -5.25656343e-01
-9.81418550e-01 4.27248001e-01 2.60926515e-01 -1.39937329e+00
-2.49035299e-01 -3.70940953e-01 -7.97072470e-01 9.65126812e-01
-1.47092247e+00 -1.02584672e+00 -4.43262398e-01 7.77204692e-01
3.11625689e-01 1.65899783e-01 6.60741985e-01 2.43743762e-01
-4.74382401e-01 6.40342593e-01 1.79558739e-01 2.56569907e-02
8.63165557e-01 -1.14104855e+00 3.27121347e-01 1.08758450e+00
1.12739585e-01 9.16150093e-01 8.80015850e-01 -4.98400927e-01
-1.57527530e+00 -7.46310830e-01 1.24076176e+00 -4.63310182e-01
6.65917099e-01 -1.39445722e-01 -5.77260256e-01 1.19954717e+00
-7.91297927e-02 -7.18792453e-02 4.95459676e-01 -1.59756932e-02
-3.72628540e-01 1.56074930e-02 -1.02088749e+00 4.80645508e-01
1.16165221e+00 -4.45303053e-01 -5.83445549e-01 6.32240176e-01
3.39524776e-01 -7.49380887e-01 -9.49587166e-01 2.09888428e-01
5.99382937e-01 -6.94721520e-01 1.24658155e+00 -4.28536087e-01
4.13941033e-02 -5.31681240e-01 -4.61835951e-01 -1.14786184e+00
-2.33311757e-01 -9.55840409e-01 4.43609923e-01 1.02837515e+00
2.23459303e-01 -5.92582583e-01 1.08582091e+00 8.44741464e-01
-1.14100859e-01 -5.31267524e-01 -9.25513983e-01 -9.15006697e-01
6.62093908e-02 1.95037797e-02 3.33970398e-01 9.74138975e-01
1.65322319e-01 9.90876779e-02 -4.03833538e-01 4.29319143e-01
8.64439428e-01 2.80698031e-01 1.26549733e+00 -1.28651392e+00
-1.19784437e-01 -2.56075799e-01 -7.43910372e-01 -1.37063003e+00
8.95834342e-02 -8.33076537e-01 1.79314956e-01 -1.48798287e+00
3.13761652e-01 -6.40698731e-01 3.88446480e-01 1.49089217e-01
1.49033666e-02 2.43862033e-01 1.85595885e-01 5.88537812e-01
-2.07855254e-01 1.76470980e-01 1.07046878e+00 2.32170448e-01
-3.38644475e-01 1.63978592e-01 -5.87406278e-01 7.21616447e-01
6.07813835e-01 -4.00282711e-01 -2.78511822e-01 -6.48778379e-01
1.94242850e-01 8.89872946e-03 1.70026883e-01 -8.04003775e-01
7.18823910e-01 -2.66924977e-01 3.07893418e-02 -7.33398139e-01
5.19357383e-01 -8.68626416e-01 3.46656084e-01 2.40453824e-01
-5.02481647e-02 5.40481448e-01 -9.04950593e-03 3.69197279e-01
-2.51126558e-01 -3.76208246e-01 7.19233692e-01 1.32604226e-01
-3.77625197e-01 5.01588643e-01 1.75095350e-01 7.94005916e-02
1.15303409e+00 -5.89355826e-01 -2.27086097e-01 -5.64831555e-01
-5.78396857e-01 3.26271594e-01 1.13983381e+00 1.41115695e-01
6.41557038e-01 -1.52146733e+00 -8.77886295e-01 1.37543753e-01
3.35719675e-01 2.09846377e-01 -2.79022574e-01 9.14754927e-01
-9.22007203e-01 4.27339852e-01 -6.21133558e-02 -7.27044702e-01
-1.53887832e+00 1.61577106e-01 5.21137893e-01 2.76550166e-02
-7.27217615e-01 8.52165818e-01 2.40974456e-01 -2.94083863e-01
1.40324295e-01 3.18642929e-02 2.51651376e-01 -4.62907627e-02
3.15395892e-01 6.37444198e-01 8.27439129e-02 -1.15876162e+00
-4.71330643e-01 1.52172589e+00 -7.98871964e-02 -2.22415805e-01
1.31004786e+00 -2.74466723e-01 -1.43293947e-01 -1.08911000e-01
1.58290076e+00 2.22850248e-01 -1.15898013e+00 -4.74804670e-01
4.23224829e-02 -9.69461143e-01 -2.39644647e-01 5.77691831e-02
-8.01375151e-01 5.89313090e-01 4.96212058e-02 -1.99933514e-01
8.60186756e-01 4.33534719e-02 6.71318054e-01 5.01122653e-01
5.26688337e-01 -8.81621599e-01 -5.19986115e-02 4.72385705e-01
9.62945879e-01 -1.10887408e+00 3.17857534e-01 -8.33859861e-01
-9.68733430e-02 1.32541692e+00 4.22177672e-01 -5.87119043e-01
5.67700028e-01 2.76849419e-01 -1.25078596e-02 -2.53540665e-01
-1.78926885e-01 -2.15489969e-01 3.52809966e-01 4.59010452e-01
2.75631428e-01 -2.01882318e-01 -4.47472930e-01 -2.42483199e-01
-7.36086845e-01 -3.69716704e-01 6.63484275e-01 1.05402768e+00
-6.11915112e-01 -1.32257867e+00 -6.53927445e-01 7.70536885e-02
-6.43597683e-03 1.25077948e-01 -5.66008449e-01 8.16926062e-01
-2.08665386e-01 9.69654381e-01 3.19231719e-01 -3.98184597e-01
6.67270482e-01 -3.32249463e-01 8.29641759e-01 -6.14464700e-01
-1.73409134e-01 1.70589998e-01 7.12773344e-03 -5.40839374e-01
-5.90099156e-01 -1.05248022e+00 -1.19995558e+00 -6.10815704e-01
-7.03442395e-01 7.43507594e-02 6.80427730e-01 8.47331583e-01
1.84343845e-01 -2.48913661e-01 1.05001616e+00 -9.06667292e-01
-5.47258198e-01 -4.65033948e-01 -3.90246034e-01 3.73930871e-01
4.68531638e-01 -6.95430458e-01 -4.80214089e-01 1.55898809e-01]
|
[7.913965225219727, -2.335557222366333]
|
e490c3de-f2f0-497b-98ce-d360e50cfd9b
|
neural-network-hardware-co-design-for
|
1811.02187
| null |
http://arxiv.org/abs/1811.02187v2
|
http://arxiv.org/pdf/1811.02187v2.pdf
|
Neural Network-Hardware Co-design for Scalable RRAM-based BNN Accelerators
|
Recently, RRAM-based Binary Neural Network (BNN) hardware has been gaining
interests as it requires 1-bit sense-amp only and eliminates the need for
high-resolution ADC and DAC. However, RRAM-based BNN hardware still requires
high-resolution ADC for partial sum calculation to implement large-scale neural
network using multiple memory arrays. We propose a neural network-hardware
co-design approach to split input to fit each split network on a RRAM array so
that the reconstructed BNNs calculate 1-bit output neuron in each array. As a
result, ADC can be completely eliminated from the design even for large-scale
neural network. Simulation results show that the proposed network
reconstruction and retraining recovers the inference accuracy of the original
BNN. The accuracy loss of the proposed scheme in the CIFAR-10 testcase was less
than 1.1% compared to the original network. The code for training and running
proposed BNN models is available at:
https://github.com/YulhwaKim/RRAMScalable_BNN.
|
['Jae-Joon Kim', 'HyungJun Kim', 'Yulhwa Kim']
|
2018-11-06
| null | null | null | null |
['neural-network-simulation']
|
['computer-code']
|
[ 3.52846861e-01 -8.33637044e-02 -1.99423432e-01 -5.16283214e-01
-2.35115349e-01 -3.46546441e-01 -1.50831556e-02 -2.25802332e-01
-6.50026798e-01 8.76309752e-01 -4.79069024e-01 -4.44610864e-01
-8.72316211e-03 -6.67284548e-01 -7.63654113e-01 -6.77687645e-01
3.38881403e-01 1.49140032e-02 2.95264363e-01 2.09314361e-01
9.70760062e-02 8.00091982e-01 -1.12108600e+00 1.96954176e-01
4.90468293e-01 1.55821121e+00 3.50339174e-01 6.91581547e-01
2.21048579e-01 1.09390843e+00 -9.17093873e-01 -1.37923121e-01
4.37718958e-01 -3.63094032e-01 -3.11153494e-02 -9.58615959e-01
6.17445648e-01 -7.68222928e-01 -7.74280369e-01 1.21095693e+00
7.01719224e-01 -7.24993274e-02 5.26797891e-01 -1.06064105e+00
-4.47252572e-01 1.17468965e+00 -3.87681246e-01 3.76919478e-01
-5.09061277e-01 -1.75035521e-02 5.38581252e-01 -5.52029550e-01
5.93020115e-03 9.98951733e-01 1.02370596e+00 5.15207171e-01
-1.20221531e+00 -1.45147955e+00 -4.44357961e-01 2.96847403e-01
-1.63344288e+00 -6.88001692e-01 8.37430775e-01 1.01860948e-01
1.23430037e+00 9.58860293e-02 8.75786304e-01 9.03499007e-01
3.17788422e-01 4.87794638e-01 9.70107675e-01 -2.49653369e-01
3.40641022e-01 -8.33113566e-02 6.34026289e-01 5.72063804e-01
9.00151849e-01 -6.05927482e-02 -2.81176060e-01 3.12739871e-02
1.21960127e+00 2.79938608e-01 -3.09931248e-01 3.41963172e-01
-6.03088319e-01 4.72434312e-01 9.31151509e-01 2.67141372e-01
-4.03278321e-01 8.94416630e-01 1.20920777e-01 2.51469851e-01
-4.81233805e-01 1.06813096e-01 -2.13486269e-01 7.35607818e-02
-1.19973993e+00 1.11552933e-02 6.68512404e-01 7.44932652e-01
3.68977875e-01 1.02789152e+00 4.27066118e-01 8.78738999e-01
4.12871867e-01 8.82203281e-01 9.58229780e-01 -9.41087842e-01
3.64831626e-01 2.10141197e-01 -3.67645174e-01 -9.39728916e-01
-4.06482965e-01 -7.04823554e-01 -1.45335829e+00 2.81867921e-01
4.44764912e-01 -2.84100831e-01 -1.17689800e+00 1.62083519e+00
-3.16747367e-01 4.18534338e-01 2.83390909e-01 7.70326495e-01
8.35137963e-01 1.03456068e+00 -3.69727671e-01 -5.33827916e-02
1.26907253e+00 -8.13348532e-01 -8.62010360e-01 -5.54009140e-01
-1.21456504e-01 -2.98225373e-01 4.86162573e-01 4.33784217e-01
-1.26506031e+00 -8.65336001e-01 -1.76424205e+00 -1.54649124e-01
-2.17174381e-01 4.01876390e-01 4.28561807e-01 8.62694502e-01
-1.01162088e+00 5.85644662e-01 -8.58608425e-01 1.50195539e-01
7.15779185e-01 8.60228360e-01 1.65817142e-01 1.61665142e-01
-1.27498972e+00 8.41256201e-01 7.74160206e-01 4.92851049e-01
-6.51193798e-01 -3.92893970e-01 -3.20860416e-01 3.15687984e-01
-3.08773935e-01 -3.47218275e-01 1.48469722e+00 -1.12407088e+00
-1.59009433e+00 -3.60556878e-02 6.96252286e-02 -1.17789710e+00
-1.51078761e-01 6.12021759e-02 -8.04048300e-01 2.05722824e-01
-7.56699443e-01 8.54106843e-01 9.25175428e-01 -7.11755455e-01
-4.79298890e-01 -2.07547933e-01 -5.72614431e-01 -2.31170446e-01
-3.04025143e-01 -5.46609581e-01 1.79277003e-01 -7.86263466e-01
3.31280112e-01 -6.44814909e-01 -2.37933565e-02 1.32497549e-01
-1.79550454e-01 2.82406121e-01 9.07818019e-01 -5.81399441e-01
1.35083520e+00 -2.04438138e+00 -6.45295620e-01 3.83378595e-01
-3.02522965e-02 7.16769576e-01 2.41978727e-02 -2.92142898e-01
5.66089526e-02 -2.42402062e-01 -4.89107221e-02 1.01139933e-01
-3.09612334e-01 1.73965335e-01 -5.67917168e-01 2.93875873e-01
7.21885189e-02 9.24901366e-01 -1.75905079e-01 -1.05439395e-01
-1.22213453e-01 7.33714342e-01 -2.25246847e-01 -3.09662730e-01
1.50270894e-01 -2.50488847e-01 2.75007989e-02 6.21075451e-01
7.70442426e-01 -2.37599686e-01 3.72998774e-01 -7.22114444e-01
4.50065127e-03 3.68721575e-01 -1.15375590e+00 1.35854816e+00
-8.14518332e-02 1.04087842e+00 9.37192366e-02 -1.01446080e+00
1.30317771e+00 4.54029948e-01 -2.98520654e-01 -1.03021157e+00
5.75937569e-01 5.90329468e-01 5.86013913e-01 3.46798509e-01
3.04897785e-01 -7.51751736e-02 -2.47166101e-02 1.40317261e-01
2.34448627e-01 6.50497153e-02 -3.44103038e-01 -1.76683679e-01
8.07820380e-01 -4.55118209e-01 4.57143784e-01 -6.03537448e-02
2.51463741e-01 -1.79186255e-01 6.10174954e-01 8.32339764e-01
-1.50149375e-01 1.01551801e-01 -9.50353802e-04 -3.91739160e-01
-1.14353228e+00 -1.04373348e+00 -4.47574675e-01 4.54199463e-01
-8.61347169e-02 3.89605582e-01 -6.68209255e-01 2.69519120e-01
-7.56510794e-02 6.94265962e-01 5.14712445e-02 -1.71678647e-01
-8.39293480e-01 -5.13386011e-01 1.48871815e+00 6.36034310e-01
9.26120162e-01 -9.11157250e-01 -1.14218223e+00 4.00561094e-01
2.94347763e-01 -9.98549998e-01 -1.31803276e-02 6.98995888e-01
-1.49476230e+00 -5.22088468e-01 -6.60465240e-01 -1.07868659e+00
5.49977779e-01 -1.87669083e-01 3.70612204e-01 -2.84192353e-01
-1.95817620e-01 -3.99374515e-01 3.44048321e-01 -4.40508306e-01
-4.42369021e-02 -5.73804341e-02 2.53149450e-01 -3.30934584e-01
4.50980246e-01 -9.00684297e-01 -5.58564603e-01 -6.37312010e-02
-4.83703971e-01 5.74746132e-02 9.40174222e-01 7.80081689e-01
5.98393917e-01 1.83756396e-01 1.19427240e+00 -4.50678319e-01
4.81676757e-01 -9.03351083e-02 -9.88683641e-01 -9.32051763e-02
-6.45026445e-01 7.93534890e-02 1.08990467e+00 -9.34805155e-01
-6.22814119e-01 7.11029321e-02 -2.25467861e-01 -4.95002955e-01
1.89712599e-01 2.49918789e-01 -2.89447568e-02 -1.60841137e-01
5.25885820e-01 3.20127457e-01 7.27521405e-02 -4.23290700e-01
-1.19874105e-02 9.64006245e-01 8.89155447e-01 5.21473959e-02
5.32588184e-01 -1.83451995e-02 2.78325770e-02 -6.18369579e-01
-9.41896960e-02 3.27268034e-01 -4.01793234e-02 2.98781563e-02
2.45824456e-01 -1.25826275e+00 -9.05165255e-01 8.10857058e-01
-1.18098426e+00 -4.55545872e-01 -8.42413828e-02 6.32507324e-01
4.14614379e-02 -2.27385551e-01 -1.17572260e+00 -9.12976623e-01
-9.87076163e-01 -9.31052446e-01 -1.71569511e-01 7.35717773e-01
-2.13271633e-01 -4.64311749e-01 -3.76377791e-01 -7.54870176e-02
7.54776418e-01 -5.93043342e-02 7.90626049e-01 -6.85757995e-01
-6.45593345e-01 -4.21705753e-01 -4.13276255e-01 5.38648129e-01
-2.42066294e-01 -2.63148397e-01 -1.23774886e+00 -2.41222069e-01
4.57022130e-01 -2.41200119e-01 9.61648166e-01 5.21007240e-01
1.03521347e+00 -7.90954888e-01 -2.30271682e-01 4.52433556e-01
1.83877933e+00 8.07690144e-01 8.50761235e-01 -5.02729937e-02
6.07385397e-01 -3.50922465e-01 -6.24611005e-02 2.17407435e-01
4.50104959e-02 3.19389313e-01 3.40542823e-01 2.18999848e-01
-2.12828770e-01 -2.14968115e-01 3.88558477e-01 1.23904920e+00
1.64837644e-01 -3.17578793e-01 -7.56072044e-01 2.25462958e-01
-1.36171520e+00 -1.07199073e+00 -1.60793513e-01 1.99110603e+00
1.11812019e+00 4.87598717e-01 -4.08358276e-01 5.23150265e-01
6.68443322e-01 1.22799119e-02 -1.00574911e+00 -7.59990215e-01
-2.35706717e-01 8.01753223e-01 9.74640191e-01 3.82422596e-01
-6.38102055e-01 4.99472439e-01 5.61268234e+00 9.80345309e-01
-1.57366455e+00 2.11202726e-01 5.70841968e-01 -3.35358232e-01
2.05945566e-01 -3.85574609e-01 -1.11186612e+00 6.23785555e-01
1.67048311e+00 2.08816245e-01 5.46771944e-01 9.54476953e-01
-9.45528895e-02 -6.61783367e-02 -8.68604600e-01 1.24253619e+00
-2.15713814e-01 -1.27373183e+00 7.57447183e-02 -1.46770999e-01
5.92627943e-01 9.71144587e-02 1.15766123e-01 3.52899462e-01
-4.05504890e-02 -1.14953554e+00 9.54919457e-01 5.56362927e-01
1.03814793e+00 -8.63973022e-01 8.62086058e-01 3.05988014e-01
-1.15130270e+00 -5.33071756e-01 -7.74807513e-01 -4.19941515e-01
-1.22685730e-01 8.32224429e-01 -8.76332939e-01 -1.68142319e-01
5.22243023e-01 2.29187563e-01 -3.74040157e-01 9.09030318e-01
1.50081411e-01 8.16637635e-01 -5.92521727e-01 -3.78749698e-01
6.30389247e-03 1.98087394e-02 8.26527923e-02 1.08356559e+00
7.25253105e-01 3.42968225e-01 -5.46382785e-01 7.97789276e-01
-5.16878605e-01 -5.46866357e-01 -3.62346232e-01 3.43521498e-03
1.23464525e+00 1.12199998e+00 -5.60299516e-01 -5.60935855e-01
-4.26281504e-02 6.69680834e-01 1.95109800e-01 2.11951151e-01
-9.53830779e-01 -9.89965379e-01 2.43613392e-01 -3.11674699e-02
5.85683465e-01 -1.03197910e-01 -7.07303226e-01 -5.51910818e-01
-1.57123595e-01 -8.44291270e-01 -6.89581856e-02 -9.18280423e-01
-5.05933821e-01 7.86064148e-01 -5.48764169e-01 -1.07774603e+00
-3.64383191e-01 -8.06927025e-01 -2.53185630e-01 9.04074013e-01
-1.16961956e+00 -5.20708680e-01 -3.55457664e-01 3.98548543e-01
2.17653349e-01 -3.44636202e-01 8.75322700e-01 5.07764459e-01
-6.91165209e-01 1.02411282e+00 2.50913650e-01 3.81700903e-01
3.53580326e-01 -5.28258741e-01 2.82230407e-01 1.03373373e+00
-1.69509903e-01 7.38687158e-01 3.56716990e-01 -4.95009989e-01
-1.50218117e+00 -1.13302386e+00 4.56461817e-01 5.06503224e-01
3.66181105e-01 -4.26394790e-01 -9.90180969e-01 6.80362701e-01
3.57901692e-01 9.44474638e-02 4.33748543e-01 -8.18514764e-01
-4.72230971e-01 -8.93434167e-01 -1.41121864e+00 6.50511503e-01
4.90895420e-01 -3.22114766e-01 -5.15976310e-01 -5.12917280e-01
4.66349453e-01 -4.23322856e-01 -6.88999236e-01 3.87544125e-01
1.03345811e+00 -7.37282753e-01 9.32141304e-01 3.23420644e-01
1.94591954e-01 -5.80968738e-01 -5.60622811e-01 -7.61456847e-01
-3.15632850e-01 -1.32069290e-01 -5.88838756e-01 9.00056660e-01
4.84021932e-01 -1.03028500e+00 6.65265739e-01 2.68086642e-01
6.73372820e-02 -6.79690301e-01 -1.13090932e+00 -6.53352439e-01
-2.14059904e-01 -2.02159122e-01 4.05147403e-01 5.26901782e-01
-2.22950071e-01 3.84687901e-01 -2.80184299e-01 3.12146157e-01
7.51105964e-01 -2.66767383e-01 9.25601646e-02 -1.12882662e+00
-4.35999185e-01 -6.33335590e-01 -5.22731721e-01 -1.26633811e+00
-3.23693722e-01 -7.31812775e-01 -1.74148730e-03 -1.18161118e+00
-2.91508198e-01 -5.41425347e-01 -6.05753124e-01 5.93141437e-01
5.20410120e-01 6.46752775e-01 1.74391896e-01 2.08668619e-01
2.01543421e-02 -9.58317239e-03 6.89762473e-01 -1.30021110e-01
-4.23657596e-02 -1.60917073e-01 -4.99240190e-01 5.84259868e-01
1.39303339e+00 -5.89275956e-01 -5.73820293e-01 -4.31834549e-01
-1.07752495e-01 2.88990915e-01 5.54328442e-01 -1.89981902e+00
7.77735054e-01 3.41017991e-01 1.12652218e+00 -8.39924872e-01
5.98043621e-01 -1.08566463e+00 7.42114425e-01 1.04120994e+00
-2.10482553e-01 8.30531716e-02 5.05447924e-01 1.25427157e-01
-5.29029369e-02 -6.28214061e-01 1.09038818e+00 1.49256855e-01
-3.95369679e-01 -1.18239671e-01 -4.87852007e-01 -3.08039159e-01
5.67787170e-01 -6.22299135e-01 -6.88498914e-01 -7.30213225e-02
-2.61061370e-01 -2.53022015e-01 -1.86420679e-02 -1.43761694e-01
9.13401365e-01 -1.36054540e+00 -1.62720278e-01 5.30083597e-01
-8.04457486e-01 8.45470503e-02 3.02493989e-01 5.60147583e-01
-7.04297960e-01 8.13344240e-01 -6.69195950e-01 -3.29079360e-01
-1.05872381e+00 2.83196978e-02 6.00635827e-01 -8.05742368e-02
-3.33596259e-01 8.02529275e-01 -6.02370203e-01 9.51064602e-02
5.12080073e-01 -8.00961554e-01 7.44681805e-02 -1.09631807e-01
7.46342480e-01 4.80946362e-01 -4.35173139e-02 -1.20388687e-01
-3.74964267e-01 5.18289208e-01 -1.35471612e-01 -3.07373196e-01
1.22736871e+00 2.29292274e-01 -2.38516591e-02 6.48798585e-01
1.21582782e+00 -4.48023945e-01 -1.07013917e+00 -1.67656258e-01
-3.38889807e-01 2.32562885e-01 2.40219116e-01 -9.34583306e-01
-1.28160203e+00 1.01369131e+00 1.03569210e+00 -2.71837533e-01
1.46580946e+00 -8.37426960e-01 1.00021303e+00 8.32577586e-01
3.32569122e-01 -9.81027782e-01 -2.20010474e-01 5.78473270e-01
6.25235081e-01 -5.31101167e-01 1.26268879e-01 4.78492498e-01
-2.10356310e-01 1.36522496e+00 6.74110293e-01 -5.16473889e-01
6.05885327e-01 6.94540679e-01 1.52168293e-02 4.24803376e-01
-7.18432486e-01 6.33321166e-01 -1.83189392e-01 3.09490174e-01
-2.37764493e-02 7.20576793e-02 3.90959643e-02 1.06175005e+00
-4.37834501e-01 4.06781346e-01 7.58819878e-01 8.65855575e-01
-8.05801511e-01 -5.02952099e-01 -4.07612771e-01 8.62659335e-01
-5.36284864e-01 -3.39485347e-01 2.75810331e-01 4.64072913e-01
-3.62595506e-02 5.05614996e-01 5.30326366e-01 -5.78369915e-01
4.03018966e-02 2.80179739e-01 5.22846997e-01 1.06203884e-01
-5.65915048e-01 1.59195095e-01 -1.20737746e-01 -1.87273785e-01
3.33583727e-02 -4.93216850e-02 -1.73368919e+00 -4.39003140e-01
-4.41807657e-01 -1.44621626e-01 9.84271109e-01 3.20566177e-01
3.13182235e-01 8.34185123e-01 2.37268060e-01 -6.64565623e-01
-7.91344225e-01 -1.09264660e+00 -5.88765085e-01 -4.56737250e-01
5.20036101e-01 -7.36665577e-02 -2.21004352e-01 -9.69076827e-02]
|
[8.37532901763916, 2.743797540664673]
|
fc5d801b-7128-4cb8-90cc-9b852ebdffdd
|
spder-semiperiodic-damping-enabled-object
|
2306.15242
| null |
https://arxiv.org/abs/2306.15242v1
|
https://arxiv.org/pdf/2306.15242v1.pdf
|
SPDER: Semiperiodic Damping-Enabled Object Representation
|
We present a neural network architecture designed to naturally learn a positional embedding and overcome the spectral bias towards lower frequencies faced by conventional implicit neural representation networks. Our proposed architecture, SPDER, is a simple MLP that uses an activation function composed of a sinusoidal multiplied by a sublinear function, called the damping function. The sinusoidal enables the network to automatically learn the positional embedding of an input coordinate while the damping passes on the actual coordinate value by preventing it from being projected down to within a finite range of values. Our results indicate that SPDERs speed up training by 10x and converge to losses 1,500-50,000x lower than that of the state-of-the-art for image representation. SPDER is also state-of-the-art in audio representation. The superior representation capability allows SPDER to also excel on multiple downstream tasks such as image super-resolution and video frame interpolation. We provide intuition as to why SPDER significantly improves fitting compared to that of other INR methods while requiring no hyperparameter tuning or preprocessing.
|
['Chawin Sitawarin', 'Kathan Shah']
|
2023-06-27
| null | null | null | null |
['image-super-resolution', 'super-resolution', 'video-frame-interpolation']
|
['computer-vision', 'computer-vision', 'computer-vision']
|
[ 3.42033505e-01 6.20256782e-01 -3.22650671e-01 -2.13356853e-01
-6.84747517e-01 -4.76049520e-02 4.25419360e-01 -2.81321198e-01
-3.16613257e-01 7.17179894e-01 1.49626762e-01 -8.83532986e-02
1.20904483e-01 -7.18869388e-01 -9.10118580e-01 -8.27391982e-01
-1.19554311e-01 8.07286277e-02 1.09363627e-03 -1.87742025e-01
2.14160681e-01 5.05327463e-01 -1.52875268e+00 4.52383459e-01
5.12035191e-01 1.12643182e+00 7.87680298e-02 7.01110244e-01
-9.02226940e-02 9.18019593e-01 -6.08604014e-01 5.06600328e-02
1.80813953e-01 -1.75651520e-01 -6.37230277e-01 -4.35316324e-01
4.66799647e-01 -3.70948792e-01 -4.25913244e-01 9.32815492e-01
5.85202754e-01 1.79444909e-01 7.16927052e-01 -8.56997132e-01
-1.09810936e+00 6.34924054e-01 -6.02885842e-01 1.57822981e-01
1.74408540e-01 -1.40423864e-01 8.05606842e-01 -9.80674863e-01
3.34038258e-01 1.11138594e+00 1.23602569e+00 7.25916088e-01
-1.56635308e+00 -6.09051108e-01 8.74178857e-02 -2.27918342e-01
-1.51939571e+00 -6.22087836e-01 8.80464554e-01 -3.29085499e-01
1.11761582e+00 2.60011762e-01 5.24491131e-01 8.90971661e-01
-2.12367088e-01 4.42139924e-01 5.87372363e-01 -6.75888419e-01
3.21068913e-02 3.22892427e-01 3.20655778e-02 6.72161043e-01
-2.05321103e-01 1.05208300e-01 -6.74082637e-01 -1.80773735e-01
1.56084013e+00 -3.78216088e-01 -5.29030561e-01 -8.56613070e-02
-9.50420380e-01 9.17364955e-01 6.47222519e-01 2.74989307e-01
-3.74752998e-01 8.14917624e-01 3.91021520e-01 3.60072881e-01
4.44479704e-01 5.84369183e-01 -4.28573102e-01 -1.80551589e-01
-9.46899831e-01 1.15198180e-01 5.34741998e-01 6.40215278e-01
6.00573480e-01 7.46798515e-01 1.59274012e-01 1.00431001e+00
1.24494754e-01 1.11250788e-01 7.52457440e-01 -1.44643021e+00
3.73993844e-01 2.42099926e-01 5.26760109e-02 -1.13469648e+00
-2.08926082e-01 -4.92153555e-01 -1.00019085e+00 6.00063205e-01
4.98590887e-01 -1.74659073e-01 -7.07827210e-01 1.96523678e+00
-6.26805946e-02 4.77554947e-01 -4.45167385e-02 9.70672965e-01
8.59827995e-01 1.04162216e+00 -1.32270187e-01 -5.77615537e-02
1.11254501e+00 -7.31544435e-01 -4.88729089e-01 3.82065736e-02
3.47295254e-01 -4.33216006e-01 1.28949213e+00 4.65991080e-01
-1.30340314e+00 -7.95462072e-01 -1.14183819e+00 -2.19609678e-01
-1.41760781e-01 2.82952189e-01 6.83895409e-01 4.49492007e-01
-1.30183554e+00 9.18738127e-01 -7.18530416e-01 1.43463194e-01
3.62389237e-01 5.19180357e-01 -3.35198224e-01 6.06748819e-01
-1.20441520e+00 8.19144845e-01 3.10599595e-01 4.68858816e-02
-3.57398152e-01 -1.24272835e+00 -7.90634394e-01 2.30566546e-01
-1.88720718e-01 -7.98229694e-01 1.09138143e+00 -1.40357661e+00
-1.98426223e+00 7.15694010e-01 -1.26044065e-01 -8.71732652e-01
3.81381810e-01 -4.53118294e-01 -2.74113297e-01 3.08754891e-01
-3.42744082e-01 9.33872342e-01 1.26830804e+00 -1.12285447e+00
-5.55603802e-02 1.65483654e-01 7.35891052e-03 1.58489212e-01
-4.12557989e-01 -1.40140980e-01 -1.55540034e-01 -1.02547431e+00
1.84610873e-01 -6.63188279e-01 -1.35410041e-01 5.51175833e-01
-1.62699580e-01 -5.45641892e-02 8.62961471e-01 -5.99773288e-01
1.23929954e+00 -2.14501715e+00 1.62848756e-01 7.73863196e-02
1.88029036e-01 3.30464154e-01 -1.48123667e-01 1.96810469e-01
-5.73378623e-01 1.23159237e-01 -2.50585407e-01 -3.77334148e-01
-1.91618279e-01 1.45622060e-01 -7.84095407e-01 4.74614412e-01
3.00116122e-01 5.86757481e-01 -6.45077050e-01 -3.63915563e-02
1.02031514e-01 1.21654487e+00 -7.95261919e-01 8.24512728e-03
-1.01625688e-01 2.74463117e-01 -1.20455422e-01 1.86293408e-01
5.60968518e-01 -3.74252319e-01 -1.44593507e-01 -3.68094623e-01
-7.89131150e-02 5.10720074e-01 -1.17743790e+00 1.61323142e+00
-5.72109580e-01 1.04942679e+00 4.48968969e-02 -1.20267022e+00
1.07785225e+00 5.52879155e-01 4.05450433e-01 -4.02821958e-01
-7.48865530e-02 3.08495797e-02 -2.66256183e-01 -2.85636395e-01
4.54004973e-01 -1.62905842e-01 4.95250881e-01 4.53634381e-01
-4.49323803e-02 1.54558048e-01 -4.44128305e-01 -1.31302267e-01
6.76801085e-01 2.52082676e-01 5.07531017e-02 -2.22570240e-01
5.14145136e-01 -4.56075788e-01 4.14871663e-01 5.40220141e-01
1.43602714e-01 8.74426603e-01 5.55313706e-01 -6.37351871e-01
-1.28412712e+00 -9.25782502e-01 -3.80180985e-01 1.19020975e+00
-2.27177262e-01 -2.91027904e-01 -7.69249558e-01 3.31170037e-02
5.40037751e-02 5.21206379e-01 -8.09780002e-01 -1.14039533e-01
-9.23103392e-01 -5.86427331e-01 9.51175809e-01 7.43917823e-01
5.35605967e-01 -1.03175056e+00 -5.37514806e-01 2.50925630e-01
-4.50187773e-02 -7.01043189e-01 -4.13927019e-01 2.73011595e-01
-1.12367332e+00 -5.60224116e-01 -8.43363047e-01 -8.68323684e-01
7.59145737e-01 -2.57891536e-01 1.06896698e+00 2.87915822e-02
-1.04730897e-01 -5.13000693e-03 1.41984835e-01 -2.33566061e-01
-3.76914054e-01 1.16820194e-01 1.05883688e-01 -6.95660487e-02
-1.54402796e-02 -1.02029955e+00 -7.01679230e-01 7.80413523e-02
-5.92825651e-01 1.58217549e-01 3.29891115e-01 9.94133353e-01
7.16968417e-01 -1.31822482e-01 7.97676563e-01 -8.37871075e-01
6.15150571e-01 -2.76168823e-01 -4.30403054e-01 -3.65882486e-01
-4.01167154e-01 2.82306463e-01 8.42129588e-01 -1.02922428e+00
-8.31250012e-01 2.79840343e-02 -2.85334259e-01 -6.98244512e-01
1.70392901e-01 3.81664395e-01 3.99742991e-01 -4.22103018e-01
9.61375713e-01 2.34812096e-01 2.53066719e-01 -6.25980914e-01
5.28516948e-01 3.58133405e-01 9.05589998e-01 -5.64520419e-01
6.12422168e-01 3.18637460e-01 -6.77991137e-02 -1.04758894e+00
-5.47881544e-01 -8.72878283e-02 -3.99341673e-01 9.96713415e-02
4.06771779e-01 -9.54772830e-01 -7.67356753e-01 2.20628589e-01
-1.16857505e+00 -4.32232082e-01 -7.29792833e-01 2.83358395e-01
-7.15187490e-01 -3.06180604e-02 -9.95851636e-01 -8.63143504e-01
-5.48516452e-01 -7.78808892e-01 7.33200729e-01 1.27587065e-01
-4.53356296e-01 -9.21445727e-01 7.14822236e-05 -1.32854357e-01
7.09056556e-01 2.85620540e-01 9.18012917e-01 -4.22562994e-02
-3.70509803e-01 -1.06556483e-01 -1.89959407e-01 5.66686809e-01
-1.73582032e-01 1.33575127e-01 -1.26140535e+00 -2.30203107e-01
1.11581564e-01 -3.25324982e-01 1.00605345e+00 6.81822419e-01
1.53079879e+00 -7.08086371e-01 -8.26137513e-02 1.09851766e+00
1.29232812e+00 -7.71822110e-02 8.41050744e-01 4.74345326e-01
3.69984180e-01 1.97249636e-01 -6.75217360e-02 3.37079138e-01
4.97419536e-02 6.05370343e-01 4.05553907e-01 -3.21329325e-01
-1.54405043e-01 -3.46775919e-01 2.38283262e-01 8.16088200e-01
-3.88858557e-01 3.38147372e-01 -4.43123400e-01 3.87429863e-01
-1.66704786e+00 -1.02936816e+00 1.61868259e-01 2.17602777e+00
1.13234329e+00 2.69554377e-01 6.68144301e-02 4.88887817e-01
5.64993978e-01 2.76244730e-01 -5.82662880e-01 -6.59541368e-01
-1.09721392e-01 3.37531239e-01 4.53738600e-01 6.38340712e-01
-9.74617124e-01 6.30096853e-01 7.50313568e+00 7.44949758e-01
-1.54698789e+00 -2.17863545e-01 7.58809090e-01 -1.22502282e-01
-2.04551756e-01 -4.68474627e-01 -5.03471851e-01 4.10268396e-01
9.55044806e-01 -5.83513267e-02 6.43105388e-01 9.46174085e-01
1.60829723e-01 3.21224809e-01 -1.11746526e+00 1.09090006e+00
-1.15656443e-01 -1.73735797e+00 1.73514232e-01 -1.58130303e-01
5.78421831e-01 -8.11254531e-02 5.42283058e-01 2.39631072e-01
-2.22293139e-01 -1.59726107e+00 7.16157436e-01 7.63337195e-01
1.20357311e+00 -7.63019443e-01 3.74427497e-01 3.12193960e-01
-1.15930557e+00 -1.70477837e-01 -6.36535048e-01 -3.14712882e-01
-6.81920424e-02 4.11327302e-01 -7.82238126e-01 3.60531500e-03
5.83349943e-01 4.76214528e-01 -8.34968910e-02 8.60188365e-01
-8.54175687e-02 7.84904957e-01 -5.45948863e-01 2.45136425e-01
1.90579757e-01 -1.71789289e-01 5.26367128e-01 1.28750718e+00
4.00990307e-01 2.51703143e-01 -3.48747760e-01 1.04049075e+00
-2.11094871e-01 -1.94523662e-01 -5.29271901e-01 3.87574695e-02
6.48575783e-01 7.91356146e-01 -8.74069408e-02 -2.61140734e-01
-9.38044190e-02 7.04900444e-01 4.26492095e-01 6.60139382e-01
-6.92061961e-01 -7.42373168e-01 7.23639667e-01 3.76669705e-01
4.81340975e-01 -9.99708846e-02 -5.96349061e-01 -8.36347520e-01
-6.30938038e-02 -6.75709605e-01 6.10890724e-02 -1.05070877e+00
-9.62336957e-01 8.60185146e-01 -2.02178374e-01 -1.11646378e+00
-5.97339809e-01 -6.76249981e-01 -6.56679571e-01 1.22607052e+00
-1.41269565e+00 -8.47316682e-01 -4.11225529e-03 4.52682287e-01
2.94685811e-01 -2.66747683e-01 1.19059861e+00 3.23271930e-01
-2.76374012e-01 8.40097845e-01 -1.90788619e-02 2.92688996e-01
3.73829931e-01 -1.04712927e+00 2.87891805e-01 2.74474233e-01
6.62668645e-02 8.72872412e-01 8.94379437e-01 2.68259738e-02
-1.05013752e+00 -8.83912742e-01 5.75502872e-01 -1.66477367e-01
5.41314840e-01 -1.26663163e-01 -1.30018997e+00 7.87736177e-01
4.31154929e-02 3.02746296e-01 5.29036462e-01 4.12812196e-02
-5.94068825e-01 -3.08855534e-01 -1.24931037e+00 5.42721093e-01
6.99804962e-01 -8.12450051e-01 -6.22136295e-01 -8.23944286e-02
8.81902397e-01 -6.01147532e-01 -8.56823564e-01 1.69320792e-01
7.37959862e-01 -8.61295998e-01 1.49163055e+00 -4.90706742e-01
6.39951229e-01 -1.64758235e-01 -1.17791481e-01 -1.19497228e+00
-4.73030865e-01 -9.11475658e-01 -6.00215077e-01 1.06902921e+00
3.21998447e-01 -8.43813956e-01 9.95674551e-01 4.53590274e-01
-5.30282445e-02 -1.02696562e+00 -1.15714264e+00 -4.89314020e-01
2.94392794e-01 -4.01012719e-01 4.75896329e-01 1.04170561e+00
5.39150974e-03 2.05299601e-01 -4.11524206e-01 2.15597466e-01
5.86099565e-01 -2.77121425e-01 3.19375753e-01 -1.15361989e+00
-4.57417399e-01 -5.88263512e-01 -3.78679454e-01 -1.47648954e+00
5.21570854e-02 -7.86797762e-01 -3.30391437e-01 -1.05665767e+00
-3.22075486e-01 -6.35995865e-01 -3.90382469e-01 5.36337733e-01
3.39256316e-01 5.58757067e-01 1.69115171e-01 1.99935153e-01
1.02013581e-01 4.99079764e-01 1.23205566e+00 -3.38934958e-02
-4.31648552e-01 -1.43537462e-01 -8.48592341e-01 9.53048050e-01
7.40070701e-01 -2.71180511e-01 -4.59823012e-01 -6.31744802e-01
1.84547439e-01 2.13628083e-01 3.53488982e-01 -1.04040098e+00
1.66216359e-01 1.43757775e-01 6.82894647e-01 -4.19087946e-01
8.70790780e-01 -6.18885994e-01 3.45773578e-01 3.35678786e-01
-7.02076614e-01 -2.76738126e-02 3.54218811e-01 2.58326620e-01
-1.59351900e-01 -3.98306608e-01 9.56773520e-01 5.44689856e-02
-2.26489946e-01 7.64021724e-02 -4.12857354e-01 -9.17874053e-02
4.65855926e-01 -5.12419939e-01 -2.19838232e-01 -6.00056529e-01
-7.82810628e-01 -2.91953892e-01 1.54067189e-01 -6.55880943e-02
7.16442585e-01 -1.57558239e+00 -6.40896142e-01 4.73079622e-01
-4.76271421e-01 1.67492151e-01 -1.08274490e-01 4.74209517e-01
-7.19225049e-01 4.03276503e-01 -3.00450809e-02 -5.69323421e-01
-1.05808187e+00 1.84317306e-01 6.19670331e-01 -3.70051302e-02
-1.14190567e+00 1.18160355e+00 1.65912092e-01 -3.21370244e-01
5.65179586e-01 -4.93666798e-01 -2.01874837e-01 -1.02405675e-01
7.54201949e-01 3.36748332e-01 -2.19016686e-01 -5.14416337e-01
-2.94261500e-02 6.62361979e-01 2.30117247e-01 -1.21681578e-01
1.55539191e+00 1.35952771e-01 -2.21237093e-02 5.59707046e-01
1.38433051e+00 -1.38632134e-01 -1.57528484e+00 -1.93593770e-01
-5.35499036e-01 -3.39082003e-01 2.38652095e-01 -7.30129600e-01
-1.08842850e+00 8.66937220e-01 5.79720557e-01 3.84790301e-01
1.07951915e+00 -2.54400998e-01 5.57757199e-01 4.76389468e-01
4.27912883e-02 -9.53340650e-01 1.89220890e-01 5.46086192e-01
1.22284091e+00 -8.15111160e-01 2.15993114e-02 -3.25784504e-01
-4.78281170e-01 1.42914295e+00 2.59807765e-01 -6.39566243e-01
6.49125993e-01 4.94873583e-01 6.23946115e-02 2.09549308e-01
-8.28610301e-01 4.47783560e-01 3.90179276e-01 7.81907201e-01
7.04594195e-01 -2.18627930e-01 2.61813343e-01 5.30510664e-01
-5.77660620e-01 1.90837950e-01 3.09724629e-01 4.59433496e-01
-5.07665992e-01 -6.38813317e-01 -4.42164868e-01 3.07984799e-01
-5.16861856e-01 -2.59055704e-01 2.42519304e-01 7.21456587e-01
-3.05703320e-02 3.14988315e-01 4.92032468e-01 -8.85658115e-02
2.21454680e-01 2.07918119e-02 4.91251022e-01 -2.89420813e-01
-4.34829503e-01 -2.45889090e-02 1.25597537e-01 -3.68058771e-01
-2.29657933e-01 -2.86246657e-01 -1.53594971e+00 -3.48047078e-01
-2.28424951e-01 9.24798697e-02 6.03615999e-01 4.61009353e-01
2.11501852e-01 6.51408195e-01 5.08898556e-01 -1.29171705e+00
-6.18994892e-01 -8.59403253e-01 -3.60389382e-01 1.14915758e-01
9.04439390e-01 -5.24731696e-01 -5.45601964e-01 2.68010944e-01]
|
[11.208733558654785, -1.7655855417251587]
|
a1359bf4-f3fe-4546-8b48-4db036079f19
|
comparing-unsupervised-word-translation
| null | null |
http://papers.nips.cc/paper/8836-comparing-unsupervised-word-translation-methods-step-by-step
|
http://papers.nips.cc/paper/8836-comparing-unsupervised-word-translation-methods-step-by-step.pdf
|
Comparing Unsupervised Word Translation Methods Step by Step
|
Cross-lingual word vector space alignment is the task of mapping the vocabularies of two languages into a shared semantic space, which can be used for dictionary induction, unsupervised machine translation, and transfer learning. In the unsupervised regime, an initial seed dictionary is learned in the absence of any known correspondences between words, through {\bf distribution matching}, and the seed dictionary is then used to supervise the induction of the final alignment in what is typically referred to as a (possibly iterative) {\bf refinement} step. We focus on the first step and compare distribution matching techniques in the context of language pairs for which mixed training stability and evaluation scores have been reported. We show that, surprisingly, when looking at this initial step in isolation, vanilla GANs are superior to more recent methods, both in terms of precision and robustness. The improvements reported by more recent methods thus stem from the refinement techniques, and we show that we can obtain state-of-the-art performance combining vanilla GANs with such refinement techniques.
|
['Anders Søgaard', 'Yova Kementchedjhieva', 'Mareike Hartmann']
|
2019-12-01
| null | null | null |
neurips-2019-12
|
['unsupervised-machine-translation']
|
['natural-language-processing']
|
[ 3.40648264e-01 -3.39605473e-02 -3.72864544e-01 -1.85890377e-01
-1.02539408e+00 -9.26208615e-01 1.14850652e+00 3.96566540e-01
-7.24987864e-01 8.37790668e-01 3.48731697e-01 -4.37560529e-01
2.14783907e-01 -6.23590529e-01 -6.02352977e-01 -9.45806980e-01
2.53450096e-01 1.09306192e+00 -1.16372578e-01 -4.24206465e-01
-6.34523109e-02 1.83564723e-01 -1.24702215e+00 -2.10203722e-01
6.85642362e-01 5.70000947e-01 1.02755971e-01 4.31191772e-01
-3.41070086e-01 3.78940105e-01 -1.86251000e-01 -6.62293851e-01
4.05728668e-01 -7.81562567e-01 -8.06137800e-01 -7.15541169e-02
5.17297983e-01 7.38507435e-02 2.43966747e-02 1.33790600e+00
5.10848224e-01 1.14308313e-01 8.15236866e-01 -8.24989974e-01
-6.04468465e-01 6.08042181e-01 -3.05476099e-01 8.81259590e-02
4.36434567e-01 -2.20613465e-01 1.46859682e+00 -8.98005486e-01
8.72003913e-01 8.09783220e-01 4.13156152e-01 6.77113414e-01
-1.62976873e+00 -6.08958244e-01 4.36669514e-02 -1.05576050e-02
-1.29842269e+00 -6.40219450e-01 7.20410645e-01 -6.47687614e-01
9.64910090e-01 -1.13962129e-01 3.57928753e-01 1.13792777e+00
8.33850726e-03 5.84017992e-01 1.16671443e+00 -9.55283880e-01
3.35054606e-01 3.18608671e-01 1.08938709e-01 6.21377468e-01
2.62424439e-01 1.08220421e-01 -4.96043980e-01 -1.81526780e-01
4.75145310e-01 -4.92498666e-01 -1.85106471e-01 -8.09999585e-01
-1.42681825e+00 1.10558891e+00 1.95610791e-01 6.92586839e-01
-3.06700379e-01 -1.25807328e-02 4.63749319e-01 7.31428504e-01
5.16710162e-01 5.11238694e-01 -4.38379765e-01 -5.32979071e-02
-1.14887619e+00 1.19826689e-01 8.66286039e-01 1.05210316e+00
1.00650239e+00 9.04300883e-02 -4.91715223e-03 8.62644970e-01
1.69557333e-01 6.16104126e-01 5.10783732e-01 -3.72274846e-01
5.91268420e-01 2.00845659e-01 2.64859181e-02 -5.40427804e-01
-4.89866398e-02 -3.10680985e-01 -8.98921728e-01 -1.61967546e-01
5.68100691e-01 -5.32528087e-02 -9.55518723e-01 2.09529734e+00
3.48411679e-01 2.18162209e-01 2.29849771e-01 6.47235692e-01
4.41618592e-01 4.00433362e-01 3.57252769e-02 -4.12781119e-01
1.31563520e+00 -7.93726087e-01 -5.83393812e-01 -2.32063040e-01
8.67423236e-01 -8.66134346e-01 9.45852995e-01 2.39916116e-01
-1.02045751e+00 -4.49972540e-01 -1.14162350e+00 -9.93947834e-02
-4.36416119e-01 -1.03490673e-01 4.30586457e-01 6.85657322e-01
-1.27082932e+00 5.48229754e-01 -7.36858249e-01 -6.55969322e-01
9.77119282e-02 4.26720887e-01 -5.65300882e-01 -9.20638517e-02
-1.29448938e+00 1.06906044e+00 2.13119477e-01 -3.21601212e-01
-6.88926935e-01 -4.42977399e-01 -1.09461594e+00 -2.81204939e-01
1.61052674e-01 -7.53447771e-01 1.02614260e+00 -1.01961672e+00
-1.47074747e+00 1.32778263e+00 -4.26150084e-01 -6.18761599e-01
3.13569397e-01 1.59722373e-01 -1.36829168e-01 -2.73198277e-01
1.02284007e-01 4.70108032e-01 7.82594502e-01 -1.14978313e+00
-4.04854923e-01 -2.81802505e-01 1.36873014e-02 2.03644067e-01
-6.15645200e-02 1.10203914e-01 -4.77181375e-01 -7.85198510e-01
-5.65052032e-02 -1.18706465e+00 -2.26018772e-01 -5.84399760e-01
-1.52263463e-01 -2.97159106e-01 1.96064919e-01 -7.59106159e-01
1.12102079e+00 -1.86832273e+00 8.57350111e-01 4.94820744e-01
4.32280935e-02 1.80886880e-01 -2.47836649e-01 4.16403562e-01
-1.23545453e-01 -6.55256435e-02 -4.89616185e-01 -7.17326641e-01
1.44074619e-01 3.63276094e-01 -3.08706164e-01 7.43998051e-01
2.74831317e-02 9.48989868e-01 -9.90204096e-01 -3.84200573e-01
1.82296202e-01 3.36896420e-01 -6.97496772e-01 2.74987638e-01
-2.77269214e-01 8.12594771e-01 -1.77616462e-01 3.02024961e-01
2.24343866e-01 1.09650297e-02 4.66720730e-01 2.90215760e-02
9.54398215e-02 7.33449757e-01 -1.11666906e+00 2.15612626e+00
-6.38254821e-01 5.73810995e-01 -1.15903489e-01 -1.32769918e+00
9.28045750e-01 4.44879740e-01 4.35301214e-01 -7.10570693e-01
2.46221796e-01 4.96677876e-01 1.22766502e-01 1.41744345e-01
4.64488089e-01 -4.81468260e-01 -2.76722103e-01 5.80612779e-01
5.52565932e-01 -2.57432044e-01 2.55318999e-01 4.92645279e-02
8.08062434e-01 1.60317361e-01 6.21200979e-01 -5.46503782e-01
6.41546786e-01 -2.52734125e-02 1.80192530e-01 5.98370373e-01
1.46842226e-01 4.79241580e-01 2.64418691e-01 -1.22104533e-01
-1.49365985e+00 -1.04538751e+00 -1.47428855e-01 9.36727643e-01
-8.52524191e-02 -4.92770612e-01 -7.05065310e-01 -6.91799939e-01
-1.87593222e-01 7.92137384e-01 -6.14279151e-01 -3.06485500e-03
-7.18698680e-01 -6.38369441e-01 4.68469173e-01 3.92605126e-01
1.35198347e-02 -1.02061355e+00 -5.94142564e-02 2.61949092e-01
1.15453331e-02 -1.13364780e+00 -6.35295272e-01 4.42420155e-01
-7.31242657e-01 -7.12357461e-01 -7.91876078e-01 -9.04742360e-01
5.10133266e-01 -2.57330716e-01 1.40992200e+00 7.79550448e-02
1.39516309e-01 2.83382624e-01 -3.25909555e-01 -1.76566601e-01
-7.13186085e-01 3.71581972e-01 4.33181107e-01 -9.02105402e-03
6.26861691e-01 -7.38923371e-01 -2.63215005e-01 -2.94507612e-02
-8.34012151e-01 -1.76238254e-01 4.63095248e-01 1.09253693e+00
6.82364047e-01 -4.29572701e-01 3.54898185e-01 -1.18537617e+00
4.44950759e-01 -3.97283137e-01 -7.99628556e-01 1.96089059e-01
-7.16585636e-01 5.28757453e-01 6.18684053e-01 -3.75831753e-01
-5.07263362e-01 -2.99027618e-02 -2.52938539e-01 -2.96525657e-01
-1.59824118e-01 5.46593785e-01 -2.27052525e-01 -5.85852005e-03
5.48774004e-01 3.71552855e-01 -3.33193950e-02 -6.21400654e-01
6.68682635e-01 6.98245406e-01 4.74740922e-01 -6.94019735e-01
1.07505846e+00 1.78314611e-01 -2.39134386e-01 -6.43695533e-01
-7.15044200e-01 -6.24367356e-01 -1.06156278e+00 3.69986117e-01
8.98199201e-01 -1.07654917e+00 -8.17306340e-02 2.18078911e-01
-1.14822865e+00 -4.37134385e-01 -4.73455667e-01 4.25965369e-01
-7.39723861e-01 1.82060316e-01 -3.64325523e-01 -4.70800787e-01
-2.34626934e-01 -1.19273365e+00 1.14667010e+00 -1.67911932e-01
-4.66552526e-01 -1.49097025e+00 5.20359457e-01 1.47035077e-01
3.12613398e-01 -1.67216897e-01 1.20455849e+00 -1.12182856e+00
-3.52335900e-01 -9.94568318e-02 1.26462221e-01 5.01742244e-01
2.51849085e-01 -6.11313164e-01 -8.45039487e-01 -5.91781437e-01
-7.31368512e-02 -1.82797134e-01 6.27290606e-01 -4.49598283e-02
4.45580721e-01 -2.04128876e-01 -1.89935207e-01 6.54268086e-01
1.54589403e+00 5.07723764e-02 5.62161088e-01 3.23280036e-01
7.93802142e-01 3.54423821e-01 2.72770643e-01 -1.00384392e-01
3.43991011e-01 1.09703994e+00 -9.40682217e-02 -2.04002276e-01
-2.49270454e-01 -3.02189708e-01 3.48196596e-01 1.47028255e+00
-1.88809671e-02 -7.92231485e-02 -9.19332981e-01 9.91025925e-01
-1.63637638e+00 -8.11553240e-01 3.36711705e-01 2.59777141e+00
1.14675128e+00 -2.02776864e-02 7.59924352e-02 3.12791467e-02
4.77717519e-01 2.33124718e-01 -1.82853401e-01 -4.71218437e-01
-2.42721662e-01 8.74825478e-01 6.53945327e-01 9.48017895e-01
-9.09423470e-01 1.26699972e+00 6.18065691e+00 8.91929328e-01
-1.12866187e+00 5.12990773e-01 2.09969118e-01 1.41944334e-01
-7.41859496e-01 1.60505787e-01 -8.52360368e-01 3.30313742e-01
9.64299083e-01 -5.66209443e-02 7.78527021e-01 5.44836998e-01
-2.89463043e-01 1.39073893e-01 -1.40546596e+00 9.17643070e-01
1.77773863e-01 -1.18000531e+00 1.34044543e-01 1.93787754e-01
1.05558360e+00 3.09545577e-01 -1.48217633e-01 3.44463825e-01
6.43395364e-01 -1.05313408e+00 3.85704994e-01 1.82767034e-01
1.09186244e+00 -6.52739882e-01 7.59179473e-01 1.93888143e-01
-1.11419797e+00 5.78839898e-01 -3.08318913e-01 2.03064218e-01
1.75510481e-01 3.63402963e-01 -6.68502748e-01 8.03762972e-01
5.33360355e-02 6.97528243e-01 -1.76114902e-01 5.52695811e-01
-3.67475063e-01 7.27028906e-01 -3.92057449e-01 -5.83140738e-03
4.28610772e-01 -6.30074859e-01 6.40047073e-01 1.24119830e+00
2.72542149e-01 -2.89117455e-01 2.29125679e-01 6.60491824e-01
-1.88571751e-01 5.10542989e-01 -8.50051343e-01 4.49920259e-03
3.30454499e-01 9.42418396e-01 -4.89432275e-01 -3.80956173e-01
-6.76396370e-01 1.15840662e+00 6.21327996e-01 2.43683994e-01
-5.59462607e-01 7.50548788e-04 7.89940178e-01 -9.01995003e-02
5.97499132e-01 -4.10692126e-01 -3.02664906e-01 -1.44249558e+00
9.25657526e-02 -1.10542703e+00 3.17558080e-01 -2.13031188e-01
-1.29946148e+00 7.74757981e-01 3.17153558e-02 -9.99976814e-01
-7.11204290e-01 -6.07333362e-01 -3.82647842e-01 1.19361854e+00
-1.44051075e+00 -1.12287819e+00 2.88752675e-01 6.56749845e-01
5.22110522e-01 -5.42908788e-01 1.23791385e+00 3.16342771e-01
-2.08397090e-01 8.56166959e-01 4.14956361e-01 1.81122929e-01
6.78955376e-01 -1.44086003e+00 6.11660898e-01 1.03378522e+00
8.92969191e-01 5.75672746e-01 8.79673541e-01 -5.58336377e-01
-1.26279879e+00 -9.78966177e-01 1.34458613e+00 -5.78242183e-01
9.27605450e-01 -7.52232075e-01 -7.35827267e-01 9.04478788e-01
3.79938811e-01 -1.85125530e-01 7.77432323e-01 3.37524951e-01
-4.89428699e-01 6.77339360e-02 -7.93078423e-01 6.39006793e-01
9.55887139e-01 -8.24875951e-01 -7.90707111e-01 3.62166405e-01
6.10560477e-01 -3.94064397e-01 -7.80528486e-01 1.56490758e-01
4.26233381e-01 -5.05210161e-01 7.69606471e-01 -8.03837955e-01
1.26537412e-01 -2.75639653e-01 -5.26604772e-01 -1.41674542e+00
-1.72459975e-01 -7.93336570e-01 1.61551952e-01 1.29508102e+00
6.22537673e-01 -6.97002172e-01 5.75885296e-01 7.72111863e-02
1.54451758e-01 -5.18956125e-01 -1.26436627e+00 -7.44355261e-01
5.14798164e-01 -4.51848358e-01 5.19183874e-01 1.15957570e+00
-1.16684832e-01 7.30374753e-01 -4.87526447e-01 -5.14976680e-02
6.76520944e-01 1.01182491e-01 9.34711158e-01 -1.11694860e+00
-5.12334704e-01 -5.22073328e-01 -5.17331541e-01 -1.24258280e+00
4.86427933e-01 -1.29615366e+00 5.05516343e-02 -1.12468362e+00
6.60846382e-02 -6.26425028e-01 -3.70211005e-01 3.64528269e-01
-3.43706995e-01 4.42901462e-01 4.93365526e-02 3.31623524e-01
-2.07153499e-01 4.41975981e-01 7.97468424e-01 -1.77200377e-01
-1.26056150e-01 -1.48137659e-01 -6.86227143e-01 3.48757356e-01
4.64754075e-01 -5.04935741e-01 -3.23100388e-01 -4.71692830e-01
1.46252781e-01 -2.03926235e-01 -1.34128006e-02 -6.56606495e-01
2.34529004e-01 2.26657046e-03 -1.62811860e-01 -1.22539356e-01
9.92099047e-02 -7.32689798e-01 2.03326449e-01 2.11119771e-01
-3.68507713e-01 1.69548884e-01 1.22906633e-01 4.15798962e-01
-2.68629551e-01 -3.04068744e-01 7.92319894e-01 7.78971389e-02
-4.47121620e-01 3.38865757e-01 -1.80060461e-01 1.77995831e-01
7.35312521e-01 2.85780504e-02 2.74529964e-01 -3.86829168e-01
-7.20903218e-01 -8.71702135e-02 7.16863453e-01 4.47128326e-01
-1.54105425e-02 -1.51559913e+00 -9.45875466e-01 3.88908446e-01
3.48815292e-01 2.45800824e-03 -3.29382300e-01 9.25832272e-01
-2.88929015e-01 4.82463628e-01 7.63486847e-02 -6.42140567e-01
-1.05016696e+00 6.83375537e-01 3.17380279e-01 -7.35531747e-01
-4.96433049e-01 8.14698577e-01 2.88603663e-01 -6.07301414e-01
1.50075510e-01 2.82744970e-02 4.32117246e-02 1.29899114e-01
2.37353355e-01 -2.64773667e-01 4.42381710e-01 -9.97070909e-01
-5.19999921e-01 7.67617702e-01 -2.11359009e-01 -3.50977570e-01
1.16193676e+00 -1.05513610e-01 -1.96510255e-01 6.44222498e-01
1.19961894e+00 4.27155644e-01 -6.17254496e-01 -7.47788548e-01
1.09392680e-01 -1.00286230e-01 -1.28011391e-01 -4.41360265e-01
-9.94719088e-01 7.11668909e-01 4.24614012e-01 -3.85519154e-02
9.84045863e-01 1.41631901e-01 7.89033353e-01 1.51853159e-01
5.68595290e-01 -6.50282502e-01 -5.05259037e-01 6.49278343e-01
5.14572561e-01 -1.21908545e+00 -1.88290417e-01 -1.78244561e-01
-3.90528947e-01 7.95250654e-01 4.85052504e-02 -4.24468219e-01
5.53507090e-01 3.43873262e-01 1.24837078e-01 4.51530591e-02
-5.66062331e-01 -5.09474814e-01 6.82421029e-01 5.89077711e-01
5.90806186e-01 2.84884311e-02 -4.68998104e-01 3.13073367e-01
-4.46829885e-01 -3.97440821e-01 2.06706189e-02 6.04888678e-01
-8.31983984e-02 -1.82550287e+00 -4.48287092e-03 2.84397960e-01
-4.28649902e-01 -6.62872374e-01 -3.55933934e-01 8.78290653e-01
6.20624647e-02 5.96357048e-01 1.09241724e-01 -3.00580114e-01
1.91991672e-01 4.94244307e-01 7.70197570e-01 -8.17650378e-01
-6.79077029e-01 1.35831675e-02 1.70794412e-01 -3.15214902e-01
-5.74055612e-01 -8.02636385e-01 -8.05531263e-01 -1.25252500e-01
-4.61512864e-01 4.84867781e-01 7.06617594e-01 1.39476418e+00
8.37881342e-02 8.74880552e-02 4.23391551e-01 -6.85136080e-01
-4.41234469e-01 -9.72721696e-01 -3.55791509e-01 7.48255432e-01
4.03892696e-01 -5.71277916e-01 -3.08325738e-01 1.83991566e-01]
|
[11.110671997070312, 10.100927352905273]
|
054fd934-ca74-4b2f-a655-99032d868fe5
|
unsupervised-source-separation-by-steering
|
2110.13071
| null |
https://arxiv.org/abs/2110.13071v1
|
https://arxiv.org/pdf/2110.13071v1.pdf
|
Unsupervised Source Separation By Steering Pretrained Music Models
|
We showcase an unsupervised method that repurposes deep models trained for music generation and music tagging for audio source separation, without any retraining. An audio generation model is conditioned on an input mixture, producing a latent encoding of the audio used to generate audio. This generated audio is fed to a pretrained music tagger that creates source labels. The cross-entropy loss between the tag distribution for the generated audio and a predefined distribution for an isolated source is used to guide gradient ascent in the (unchanging) latent space of the generative model. This system does not update the weights of the generative model or the tagger, and only relies on moving through the generative model's latent space to produce separated sources. We use OpenAI's Jukebox as the pretrained generative model, and we couple it with four kinds of pretrained music taggers (two architectures and two tagging datasets). Experimental results on two source separation datasets, show this approach can produce separation estimates for a wider variety of sources than any tested supervised or unsupervised system. This work points to the vast and heretofore untapped potential of large pretrained music models for audio-to-audio tasks like source separation.
|
['Bryan Pardo', 'Prem Seetharaman', "Patrick O'Reilly", 'Ethan Manilow']
|
2021-10-25
| null | null | null | null |
['audio-generation', 'audio-source-separation', 'music-generation', 'music-generation']
|
['audio', 'audio', 'audio', 'music']
|
[ 5.01743495e-01 3.11670870e-01 9.82731804e-02 -5.80147952e-02
-1.27943003e+00 -8.58208358e-01 4.44386631e-01 -3.85273427e-01
-2.12341100e-02 4.54545587e-01 6.62924290e-01 1.48380861e-01
-2.26443917e-01 -4.56695914e-01 -4.84308869e-01 -1.01240909e+00
-2.84486175e-01 8.22789073e-01 1.41176239e-01 6.82867924e-03
-7.55584016e-02 6.99722245e-02 -1.61991298e+00 4.97026056e-01
5.46740472e-01 7.54575431e-01 9.61637422e-02 1.23450184e+00
-1.72714725e-01 6.90274000e-01 -8.93543780e-01 -1.68239772e-01
2.22671807e-01 -8.82996440e-01 -7.80990899e-01 -5.95778078e-02
3.33180457e-01 -2.19080057e-02 -2.05417350e-01 9.61317122e-01
8.03938389e-01 1.70975268e-01 8.40700269e-01 -1.22832704e+00
-3.94217849e-01 1.67910004e+00 -1.07255399e-01 5.37346117e-02
2.35943660e-01 -1.93289176e-01 1.07515049e+00 -5.66808224e-01
3.35591227e-01 1.13796270e+00 8.36126626e-01 6.02577507e-01
-1.45768309e+00 -9.05034363e-01 -2.09710121e-01 -2.63865948e-01
-1.30676675e+00 -1.02405703e+00 1.01943719e+00 -6.08337045e-01
6.47552013e-01 3.43306422e-01 6.27043128e-01 1.22834337e+00
-1.06812596e-01 7.16416061e-01 5.54762959e-01 -6.15352154e-01
3.81010860e-01 1.77592799e-01 -1.91501901e-01 2.63279051e-01
-4.52969849e-01 1.78556651e-01 -9.96802688e-01 -5.66653132e-01
6.49580956e-01 -4.47211534e-01 -2.42023736e-01 -2.41757289e-01
-1.31018543e+00 6.64330423e-01 2.33472914e-01 4.41649228e-01
-9.82563123e-02 4.02622104e-01 1.60380930e-01 2.51323730e-01
4.17051941e-01 5.09174049e-01 -1.99418962e-01 -2.41043255e-01
-1.60086906e+00 1.27592891e-01 9.35557127e-01 7.72483289e-01
4.99966145e-01 6.02160156e-01 -1.57754734e-01 1.00580561e+00
5.95230460e-01 5.52327335e-01 8.53841603e-01 -8.90614331e-01
2.88778543e-01 -6.13153391e-02 -7.71788135e-02 -6.89885497e-01
-1.17570557e-01 -8.53467762e-01 -6.76407814e-01 1.46775931e-01
3.38139027e-01 -2.94620454e-01 -1.07931757e+00 1.89534104e+00
1.29532844e-01 6.72587335e-01 2.70154495e-02 8.13763201e-01
6.74557745e-01 8.26255143e-01 -1.48139238e-01 -1.55602664e-01
8.23672473e-01 -9.19484615e-01 -5.12622595e-01 -2.81795055e-01
1.07637264e-01 -1.13477349e+00 7.02311337e-01 6.09968185e-01
-1.36772740e+00 -8.48968089e-01 -1.00323534e+00 3.16554934e-01
-1.45710349e-01 3.10986966e-01 4.29741442e-01 5.71556151e-01
-1.14790702e+00 7.48062849e-01 -8.20100605e-01 -9.30527449e-02
2.00765759e-01 3.92307699e-01 -2.52589229e-02 6.62699282e-01
-1.23285127e+00 1.19426169e-01 6.27931356e-01 -2.54420247e-02
-1.44462776e+00 -6.81175947e-01 -5.62845945e-01 1.54459238e-01
-1.09092616e-01 -7.24328339e-01 1.26942837e+00 -1.37276876e+00
-1.87063706e+00 5.85223496e-01 1.27846360e-01 -6.17825985e-01
2.30252102e-01 -2.75188535e-01 -6.48553252e-01 9.87069961e-03
1.52400553e-01 7.67535031e-01 1.44265735e+00 -1.34249413e+00
-5.50517261e-01 6.50998726e-02 -5.56567907e-01 1.75862670e-01
-2.45399073e-01 1.47035345e-01 -2.45513663e-01 -1.07419741e+00
2.77611881e-01 -1.07469368e+00 9.23019871e-02 -9.02253807e-01
-8.16547394e-01 1.40224501e-01 3.33890676e-01 -6.52027905e-01
1.41187298e+00 -2.45506430e+00 4.67448801e-01 4.02001053e-01
-5.13237193e-02 4.74018753e-02 -3.19962978e-01 3.27107161e-01
-3.65436971e-01 -8.49913880e-02 -4.19984341e-01 -5.57794034e-01
3.30121696e-01 -1.64413020e-01 -1.09551811e+00 1.36099011e-01
-9.35053006e-02 5.03474832e-01 -9.68327701e-01 -2.22805992e-01
-2.31906310e-01 7.09529221e-01 -7.07434356e-01 3.39633137e-01
-2.43335932e-01 4.65143710e-01 4.88419235e-02 1.33828312e-01
4.57389086e-01 2.54963458e-01 8.73308405e-02 7.72682950e-02
1.27396390e-01 8.58791173e-01 -1.57518697e+00 2.09845638e+00
-1.43538252e-01 6.48759305e-01 3.52735743e-02 -5.21916211e-01
9.79558885e-01 7.39485085e-01 5.64270735e-01 1.25814572e-01
8.87301117e-02 3.54948133e-01 2.00651005e-01 -1.63970470e-01
3.75032425e-01 -3.95364076e-01 -2.67915249e-01 7.73708999e-01
7.33735859e-01 -2.08757773e-01 5.88709489e-02 3.25737774e-01
8.72326255e-01 3.94466013e-01 -3.35271716e-01 -4.87539805e-02
1.89628482e-01 -3.04164320e-01 3.86879563e-01 7.21424639e-01
3.14605206e-01 8.96086395e-01 1.52593195e-01 7.21965432e-02
-8.37519944e-01 -1.54791391e+00 9.08008777e-03 1.60035992e+00
-4.21553373e-01 -6.81857109e-01 -8.30743790e-01 -4.55325842e-01
-1.68193534e-01 6.41803205e-01 -3.49690408e-01 -3.90794247e-01
-4.57031131e-01 -6.65894449e-01 1.10929537e+00 3.96516263e-01
-2.57642046e-02 -1.25757909e+00 -1.71362489e-01 4.69729692e-01
-3.03977698e-01 -3.59789252e-01 -5.62594116e-01 5.71902275e-01
-8.01488578e-01 -6.13460362e-01 -5.33038497e-01 -9.01796043e-01
1.86196625e-01 -2.81805813e-01 1.21354163e+00 -4.96790051e-01
3.87452962e-03 2.51382738e-01 -1.53811589e-01 -5.61987877e-01
-8.90124142e-01 4.03971404e-01 2.61994928e-01 2.51192510e-01
1.66800022e-01 -1.17597044e+00 -2.41885886e-01 1.15084127e-01
-8.83457005e-01 -2.20278446e-02 2.48638541e-01 3.94581348e-01
4.73065555e-01 3.46503258e-01 8.92440081e-01 -6.63616776e-01
5.77166498e-01 -5.68838298e-01 -2.87057132e-01 -1.73421130e-01
-2.44772270e-01 1.81015387e-01 4.53927517e-01 -7.54456043e-01
-9.75912571e-01 2.52235681e-01 -1.85100272e-01 -5.01999319e-01
-2.99187064e-01 2.60777533e-01 -3.31496865e-01 6.88261330e-01
8.23512971e-01 1.92175016e-01 -5.04172444e-01 -8.93758297e-01
6.17377639e-01 7.83011734e-01 1.03188229e+00 -3.58742952e-01
1.04596245e+00 2.72613525e-01 -4.44458514e-01 -4.74433303e-01
-7.57332027e-01 -4.48851138e-01 -8.96289706e-01 -6.09013215e-02
9.05389905e-01 -9.38239098e-01 -1.06795698e-01 4.93187934e-01
-1.05162978e+00 -5.15231550e-01 -8.24623644e-01 6.75391197e-01
-6.75567389e-01 -2.45466053e-01 -5.00281513e-01 -1.03959012e+00
-2.52574265e-01 -5.57812512e-01 1.22419775e+00 1.08937249e-01
-5.24185061e-01 -1.00554776e+00 7.63487399e-01 -3.73290591e-02
3.52210790e-01 -9.44250599e-02 6.85132325e-01 -9.18537557e-01
-4.04457599e-01 -1.29839167e-01 4.82643098e-01 4.49233294e-01
2.33147085e-01 8.03455114e-02 -1.70300770e+00 -9.93993953e-02
3.33701260e-02 -2.68883228e-01 1.30088794e+00 3.67077202e-01
6.87518239e-01 -3.69069129e-01 -2.82937706e-01 8.16179514e-01
8.39845598e-01 2.51428276e-01 6.16072536e-01 -6.07708357e-02
8.42145085e-01 2.49062359e-01 -1.16592392e-01 3.01816761e-01
-3.94624770e-02 6.06408656e-01 8.34931526e-03 -1.38260573e-02
-4.45488006e-01 -6.28532350e-01 9.18696880e-01 1.35857224e+00
-5.70417345e-02 -2.78825730e-01 -6.55610025e-01 6.38166487e-01
-1.62869012e+00 -1.37743437e+00 1.63009688e-01 2.35212874e+00
1.30307686e+00 1.78377837e-01 4.29006189e-01 5.76910853e-01
6.13439560e-01 1.75122172e-01 -3.16124678e-01 -1.62160739e-01
-5.96094169e-02 4.87904847e-01 -1.21158957e-01 6.71698928e-01
-1.34441984e+00 8.65588129e-01 6.93088770e+00 1.05414474e+00
-1.09906435e+00 1.94880664e-01 1.39886755e-02 -4.78004366e-01
-4.38924342e-01 4.85267900e-02 -6.22849822e-01 5.43498993e-01
1.34204876e+00 -3.52514014e-02 7.01330006e-01 5.63250422e-01
-1.17048912e-01 2.31700718e-01 -1.31000125e+00 8.69075954e-01
1.47255301e-01 -9.96700287e-01 1.07925273e-01 6.57141805e-02
7.38940418e-01 1.14636928e-01 2.48990208e-01 3.38902503e-01
6.38553441e-01 -9.79972482e-01 1.17407823e+00 7.04911292e-01
6.18028104e-01 -8.46091926e-01 3.29687357e-01 3.30978781e-01
-1.19654477e+00 8.12624171e-02 -3.38704556e-01 1.85140088e-01
1.40716717e-01 6.50209665e-01 -1.08814049e+00 2.68785268e-01
5.09118617e-01 4.99399811e-01 -4.24849033e-01 1.08792758e+00
-3.42775136e-01 1.22841525e+00 -4.07631963e-01 7.53510892e-01
-4.49230783e-02 -7.89264441e-02 1.05271804e+00 1.45494998e+00
5.59032202e-01 -5.50965369e-01 1.46783993e-01 9.44199800e-01
-7.82507882e-02 -6.57150894e-02 -3.85462046e-01 -2.52816230e-01
4.25887614e-01 1.22268510e+00 -8.86066556e-01 -4.88491923e-01
4.58322406e-01 1.03303659e+00 -8.62049535e-02 4.91902739e-01
-9.65895414e-01 -4.56794083e-01 3.73870850e-01 2.13823050e-01
3.15174818e-01 7.84796551e-02 2.48013693e-03 -1.10083354e+00
-5.32595098e-01 -7.40019858e-01 5.34733295e-01 -1.05631256e+00
-1.24329185e+00 9.89461005e-01 -9.98964608e-02 -1.42368340e+00
-9.13433909e-01 -1.92986000e-02 -8.44592392e-01 1.15143478e+00
-9.54947293e-01 -9.13830519e-01 1.52840778e-01 5.18393934e-01
3.91061991e-01 -6.55357718e-01 1.16022575e+00 2.10732922e-01
-2.39950299e-01 4.82296765e-01 1.19079717e-01 2.29087010e-01
1.00212598e+00 -1.51242673e+00 3.43536943e-01 6.80934489e-01
1.04568410e+00 5.32086551e-01 6.49115682e-01 -5.30436516e-01
-7.16678858e-01 -1.22834992e+00 8.66261184e-01 -8.28927398e-01
7.17712820e-01 -6.83265269e-01 -8.70363176e-01 7.67069459e-01
4.86625165e-01 -5.41432202e-01 1.22168469e+00 1.91321239e-01
-4.61172223e-01 -2.03098089e-01 -6.18692636e-01 3.22376192e-01
9.93213058e-01 -8.46869707e-01 -7.62018085e-01 1.00183748e-01
6.38597548e-01 -2.18029156e-01 -5.45065761e-01 1.08277937e-02
6.64745092e-01 -9.60755765e-01 1.06902337e+00 -6.70008183e-01
3.38380218e-01 -4.41866398e-01 -7.13848546e-02 -1.57670116e+00
-5.77196836e-01 -1.12140548e+00 -2.04840139e-01 1.90172315e+00
6.57901645e-01 -1.45037651e-01 5.10792315e-01 -4.74042706e-02
-2.44555891e-01 -4.41688411e-02 -7.17152536e-01 -6.51164412e-01
-3.67427547e-03 -7.36037314e-01 7.45314360e-01 8.78163457e-01
5.60786389e-02 6.73085809e-01 -3.65061253e-01 2.70326853e-01
5.85768342e-01 3.38277131e-01 7.63395667e-01 -1.37641764e+00
-9.59034503e-01 -5.63873053e-01 -2.49942824e-01 -7.82853246e-01
2.01330379e-01 -1.44521701e+00 4.37301636e-01 -1.35710037e+00
5.18089952e-03 -6.25752628e-01 -6.86200321e-01 5.17427921e-01
1.48328091e-03 6.68425262e-01 2.29213089e-01 4.86459404e-01
-5.24337709e-01 4.38579351e-01 5.73002696e-01 -2.72912771e-01
-6.33692145e-01 3.34393799e-01 -7.50606298e-01 1.03854263e+00
7.80390620e-01 -8.79221499e-01 -6.79836392e-01 -4.09494668e-01
4.04998988e-01 -8.43044296e-02 4.05366987e-01 -1.34079313e+00
1.21494323e-01 1.88446015e-01 5.17156065e-01 -3.80022883e-01
3.78508478e-01 -5.80572426e-01 6.72533453e-01 -1.05488531e-01
-6.51246071e-01 -5.84712148e-01 2.53333062e-01 3.49784672e-01
-2.53999501e-01 -3.32411975e-01 5.72309792e-01 1.60036102e-01
1.56100485e-02 2.14318279e-02 -3.89677018e-01 2.43224680e-01
2.12198943e-01 -3.68328951e-02 1.64842248e-01 -5.52444398e-01
-1.34989631e+00 -4.20251071e-01 1.01628534e-01 4.21948224e-01
2.49798521e-01 -1.79897392e+00 -7.64514625e-01 6.49499834e-01
-2.65795410e-01 -1.55996248e-01 9.68398824e-02 5.19594491e-01
1.17043458e-01 1.15533091e-01 5.54126017e-02 -5.98488152e-01
-9.29689884e-01 4.71519530e-01 3.22574437e-01 -1.38426170e-01
-3.90628576e-01 1.23548698e+00 2.31509149e-01 -3.90567958e-01
4.47386116e-01 -3.39680284e-01 -3.85896191e-02 4.39717621e-01
5.43828368e-01 3.49440396e-01 -5.67408502e-02 -7.88743258e-01
-2.80743182e-01 3.93015742e-01 4.07436430e-01 -8.06483626e-01
1.35254383e+00 1.93936341e-02 -8.24678540e-02 1.09353578e+00
9.15054500e-01 7.86094189e-01 -1.18711305e+00 -6.44579902e-02
-1.25044227e-01 -1.54357538e-01 -1.63244127e-04 -9.24237072e-01
-9.07227278e-01 9.25402880e-01 5.68831503e-01 6.44840777e-01
1.14889193e+00 2.35717237e-01 6.13409162e-01 -5.98477088e-02
1.87563762e-01 -1.07805145e+00 -2.47460306e-02 4.42862302e-01
8.69407058e-01 -5.61527193e-01 -5.32433808e-01 3.44573520e-02
-5.07959306e-01 9.32395220e-01 3.07777827e-03 -2.46560633e-01
7.32168078e-01 7.21846879e-01 3.41192812e-01 7.43223587e-03
-6.13181293e-01 -2.51678258e-01 7.01883912e-01 6.56667233e-01
6.14010334e-01 1.01633906e-01 6.92023158e-01 1.04264724e+00
-9.81359482e-01 -1.22226648e-01 2.37602338e-01 3.84733349e-01
-5.63482463e-01 -1.19946921e+00 -7.16008425e-01 1.29804581e-01
-3.61382514e-01 -3.32844079e-01 -6.80749059e-01 1.60918590e-02
6.41381800e-01 9.25947428e-01 2.57824957e-01 -4.58415776e-01
3.99930077e-03 7.98367083e-01 4.39525276e-01 -8.42450380e-01
-6.17469490e-01 1.05586076e+00 -4.50372174e-02 -2.14027062e-01
-5.51791072e-01 -6.30625486e-01 -1.34386575e+00 2.00687990e-01
-3.81769240e-01 6.20199323e-01 4.01082307e-01 6.38332844e-01
2.87456274e-01 9.10574436e-01 6.35346234e-01 -1.15514719e+00
-1.61880746e-01 -1.25815117e+00 -8.86361480e-01 2.97512352e-01
3.29946071e-01 -2.68421620e-01 -5.81167400e-01 7.69232452e-01]
|
[15.526710510253906, 5.605590343475342]
|
938df489-b15b-4ebd-98de-415b114983c3
|
towards-early-prediction-of-human-ipsc
|
2305.14575
| null |
https://arxiv.org/abs/2305.14575v1
|
https://arxiv.org/pdf/2305.14575v1.pdf
|
Towards Early Prediction of Human iPSC Reprogramming Success
|
This paper presents advancements in automated early-stage prediction of the success of reprogramming human induced pluripotent stem cells (iPSCs) as a potential source for regenerative cell therapies.The minuscule success rate of iPSC-reprogramming of around $ 0.01% $ to $ 0.1% $ makes it labor-intensive, time-consuming, and exorbitantly expensive to generate a stable iPSC line. Since that requires culturing of millions of cells and intense biological scrutiny of multiple clones to identify a single optimal clone. The ability to reliably predict which cells are likely to establish as an optimal iPSC line at an early stage of pluripotency would therefore be ground-breaking in rendering this a practical and cost-effective approach to personalized medicine. Temporal information about changes in cellular appearance over time is crucial for predicting its future growth outcomes. In order to generate this data, we first performed continuous time-lapse imaging of iPSCs in culture using an ultra-high resolution microscope. We then annotated the locations and identities of cells in late-stage images where reliable manual identification is possible. Next, we propagated these labels backwards in time using a semi-automated tracking system to obtain labels for early stages of growth. Finally, we used this data to train deep neural networks to perform automatic cell segmentation and classification. Our code and data are available at https://github.com/abhineet123/ipsc_prediction.
|
['James Shapiro', 'Nilanjan Ray', 'Nidheesh Dadheech', 'Omar Mouhammed', 'Ila Jasra', 'Abhineet Singh']
|
2023-05-23
| null | null | null | null |
['cell-segmentation']
|
['medical']
|
[ 2.38144949e-01 -3.13892603e-01 3.78809571e-02 1.70959443e-01
-8.71611714e-01 -9.54958737e-01 2.86666393e-01 6.38874948e-01
-4.22462434e-01 1.04157460e+00 -3.05995554e-01 -3.21247488e-01
3.92467737e-01 -8.13712716e-01 -5.74258089e-01 -9.43022728e-01
1.18913449e-01 8.76249671e-01 1.31550720e-02 2.04038039e-01
3.80589724e-01 7.91537106e-01 -1.12640381e+00 6.71983138e-02
7.13640869e-01 7.87227988e-01 3.79990220e-01 9.53630686e-01
-7.51913041e-02 -1.15067877e-01 -2.19922751e-01 -1.04336962e-01
1.27599269e-01 -4.46827918e-01 -5.73656976e-01 -2.27110788e-01
-1.64146036e-01 -1.64027676e-01 1.94043562e-01 4.94193166e-01
6.71511292e-01 -4.12658125e-01 7.63605773e-01 -8.67853522e-01
-2.57868439e-01 2.93406509e-02 -2.03884661e-01 2.98474908e-01
2.77279079e-01 3.09696317e-01 5.15348434e-01 -8.92663181e-01
1.04208505e+00 2.04005122e-01 5.71949899e-01 8.47967029e-01
-1.46031654e+00 -5.86970568e-01 -5.61225295e-01 -3.37495655e-01
-1.28833461e+00 -7.49532998e-01 1.25846118e-01 -1.02598703e+00
6.17687047e-01 1.47484377e-01 1.11260557e+00 7.32646167e-01
4.58708376e-01 2.82257169e-01 9.94303882e-01 -4.12850529e-01
3.91727358e-01 6.81763748e-03 -3.76520216e-01 8.59672248e-01
1.37687996e-01 -1.66461300e-02 -4.89574254e-01 -2.30494607e-02
1.16394103e+00 1.41102998e-02 -3.12377036e-01 3.79988968e-01
-1.28408754e+00 3.98852021e-01 1.09888874e-02 7.36583412e-01
-3.06756824e-01 3.49618107e-01 -6.35849014e-02 -2.62434840e-01
1.94464326e-01 3.83620143e-01 -5.01368344e-01 -4.17123228e-01
-8.38016391e-01 -1.56570688e-01 4.26880449e-01 6.48085356e-01
5.76420069e-01 -2.15303034e-01 1.47692978e-01 6.58387065e-01
-5.79419807e-02 6.89100027e-02 4.56986964e-01 -1.41072047e+00
-3.89507651e-01 7.95557439e-01 1.56904757e-01 -7.05513477e-01
-5.34599304e-01 -4.56354380e-01 -8.05844903e-01 3.10079068e-01
8.55045855e-01 -2.17459574e-01 -9.67021585e-01 1.38355863e+00
2.83551306e-01 1.86382726e-01 -2.94883966e-01 5.43405235e-01
2.85813451e-01 8.58021617e-01 7.14135394e-02 -4.12499189e-01
1.19608545e+00 1.94943324e-02 -3.57434273e-01 4.97074202e-02
9.27037358e-01 -6.28913701e-01 7.69187808e-01 2.75442421e-01
-1.04733539e+00 4.01412472e-02 -9.65323031e-01 1.62979275e-01
-6.14577174e-01 7.73146451e-02 7.57915020e-01 3.77379268e-01
-1.13731074e+00 1.08249652e+00 -9.20488238e-01 -7.69768357e-01
1.12572432e+00 6.77880347e-01 -6.00937605e-01 8.43204707e-02
-2.30951607e-01 8.20437670e-01 3.23130101e-01 -1.58179909e-01
-8.49498451e-01 -1.03294480e+00 -3.15216154e-01 -1.27532691e-01
-3.99009407e-01 -1.00405920e+00 8.97445858e-01 -6.36034787e-01
-1.41447401e+00 1.27364349e+00 -4.51197207e-01 -4.00219299e-02
1.90333053e-01 3.73376012e-01 -5.92844142e-03 2.75217444e-01
2.74839252e-01 9.69848454e-01 -8.21318701e-02 -1.27181244e+00
-1.00202596e+00 -4.86370444e-01 -6.10153139e-01 -2.45154709e-01
6.88769743e-02 1.63951010e-01 -6.34467542e-01 -5.27111173e-01
6.00679554e-02 -1.16833556e+00 -2.05834448e-01 2.34325171e-01
-9.66681540e-02 1.95070475e-01 5.05731165e-01 -7.49924660e-01
7.20346987e-01 -1.78121936e+00 4.59938161e-02 -3.83026153e-02
1.11595176e-01 2.59482473e-01 -7.71019980e-02 3.38962704e-01
3.41066509e-01 7.45613754e-01 -9.92185622e-02 -7.45807812e-02
-6.83690608e-01 -2.26245642e-01 2.96243846e-01 5.08834362e-01
1.08215347e-01 1.00026786e+00 -8.69403362e-01 -7.67792523e-01
1.27814516e-01 6.94048584e-01 -1.59452587e-01 1.48009658e-01
-1.41073391e-01 1.14397693e+00 -1.72851816e-01 1.24564409e+00
1.12142906e-01 -3.77137363e-01 -5.27732335e-02 1.71873838e-01
-2.94258922e-01 -2.37528339e-01 -2.38549501e-01 1.46863973e+00
-3.09981227e-01 5.95553517e-01 -4.22973111e-02 -6.72657192e-01
5.59521079e-01 4.68316853e-01 9.63275969e-01 -4.06804442e-01
5.12619257e-01 6.03942335e-01 -3.99021715e-01 -7.26502463e-02
-1.49640277e-01 -9.86927092e-01 1.19695738e-02 2.72336006e-01
2.93866575e-01 -1.45178795e-01 5.42376101e-01 -3.01127862e-02
9.46807981e-01 2.42366493e-02 -1.62656471e-01 -2.53179818e-01
1.39290497e-01 2.80308872e-01 9.87367988e-01 1.81487158e-01
-3.31120223e-01 1.22667861e+00 3.05149764e-01 -4.85576630e-01
-1.16341054e+00 -8.94248188e-01 -1.99827701e-01 5.69554627e-01
-1.07873261e-01 1.14014775e-01 -5.59273839e-01 -3.65764618e-01
-8.16477090e-02 6.42161965e-01 -5.82850039e-01 1.46422490e-01
-3.66370231e-01 -9.46443915e-01 5.24687767e-01 3.61887157e-01
6.57287762e-02 -7.12432146e-01 -2.11020052e-01 4.58111048e-01
-1.82712376e-01 -8.33992660e-01 -1.70335203e-01 1.63102150e-01
-8.87308836e-01 -1.14344609e+00 -1.16764247e+00 -1.08492494e+00
1.09435451e+00 -2.83831328e-01 7.04463422e-01 2.99986094e-01
-5.48534155e-01 1.67358011e-01 -1.43933579e-01 -5.50012946e-01
-5.78988850e-01 -1.82437494e-01 1.47324875e-01 -2.56328970e-01
-7.29124574e-03 -7.47365594e-01 -9.14099038e-01 2.77279258e-01
-5.20515621e-01 3.40182245e-01 3.95324618e-01 5.33401251e-01
1.14098024e+00 -1.49082512e-01 8.64626050e-01 -7.71685183e-01
1.96721792e-01 -2.60438770e-01 -7.78964698e-01 3.30070615e-01
-4.68305051e-01 -4.42171872e-01 8.11372042e-01 -1.55443802e-01
-8.11499536e-01 3.90104115e-01 -2.78443098e-01 -7.47051015e-02
-3.62655491e-01 4.26789075e-01 3.30622435e-01 -2.04017892e-01
4.24226373e-01 3.95089656e-01 1.17318004e-01 -3.35824378e-02
-2.75067925e-01 6.45698905e-01 5.36396086e-01 -4.52048391e-01
5.62287152e-01 7.01500833e-01 2.74495900e-01 -6.54811025e-01
-5.85374415e-01 -3.95031214e-01 -6.32921636e-01 -5.68652213e-01
9.75298166e-01 -7.29343534e-01 -9.50581372e-01 6.31360471e-01
-9.78750885e-01 -7.74313629e-01 1.61264967e-02 3.72389376e-01
-6.82090580e-01 -1.18828714e-01 -1.03282166e+00 -4.93073225e-01
-4.88753498e-01 -8.56831491e-01 1.01907873e+00 5.77789187e-01
-4.48380500e-01 -9.30275321e-01 4.02436882e-01 5.94180703e-01
3.61890614e-01 7.86947966e-01 1.07169402e+00 -2.06540793e-01
-6.80463552e-01 -5.57707846e-01 1.07360505e-01 -2.78481007e-01
3.65114748e-01 4.99832988e-01 -7.01903224e-01 -2.03493610e-01
-4.11755592e-01 -1.82399049e-01 5.30127287e-01 6.31132960e-01
8.00803721e-01 -3.43526751e-02 -8.86016130e-01 9.02232230e-01
1.74984479e+00 6.67354882e-01 5.52229464e-01 4.23415571e-01
4.10793692e-01 4.38442022e-01 5.67762136e-01 2.36553386e-01
8.42741728e-02 2.62196451e-01 -1.89591236e-02 -7.05356523e-02
1.65430817e-03 -1.68753862e-01 -7.88492784e-02 6.81050003e-01
-5.39472401e-01 -1.78019658e-01 -1.23331845e+00 8.00466418e-01
-1.19878495e+00 -6.79160655e-01 3.64077948e-02 2.35001826e+00
8.87486875e-01 -1.56632811e-01 1.77872315e-01 -1.79577902e-01
9.41698372e-01 -8.68562162e-01 -5.37339091e-01 -2.15691060e-01
-2.90836066e-01 8.10734779e-02 4.30570096e-01 3.71484220e-01
-7.38434792e-01 8.65678251e-01 6.46691751e+00 4.49457318e-01
-1.40294957e+00 -2.94910192e-01 1.38756657e+00 -3.44557106e-01
-1.12140216e-01 3.07961881e-01 -7.65267670e-01 7.35148728e-01
7.53656030e-01 -4.59602892e-01 3.09844941e-01 3.87898134e-03
3.67319703e-01 -4.84464705e-01 -1.20712852e+00 8.55084598e-01
-4.21864480e-01 -2.09448671e+00 -1.81815863e-01 3.95624995e-01
6.53512239e-01 1.11797370e-01 -3.41887653e-01 -1.41862094e-01
2.74238765e-01 -1.04740787e+00 4.26821291e-01 6.15679085e-01
1.27500057e+00 -6.44920647e-01 8.10614407e-01 5.54601431e-01
-8.36093485e-01 1.85221508e-01 -2.14712530e-01 1.96599573e-01
5.38027644e-01 6.93982899e-01 -1.10407543e+00 -4.99710790e-04
6.13715172e-01 4.21059340e-01 -3.55631024e-01 1.19549417e+00
2.98209190e-01 4.49402422e-01 -6.18812144e-01 7.94407800e-02
-3.58751655e-01 -3.50400925e-01 3.04510355e-01 1.01344430e+00
1.10352647e+00 4.94729966e-01 -4.29192543e-01 7.57672310e-01
-7.51182958e-02 -4.19953056e-02 -6.12865150e-01 -5.74636996e-01
5.43284655e-01 1.33015370e+00 -1.32847607e+00 -9.24449489e-02
3.95326428e-02 9.19989824e-01 3.98287505e-01 2.03032181e-01
-5.39357543e-01 -2.22747549e-01 2.17436031e-01 4.30303186e-01
7.52108097e-02 -1.52643204e-01 -6.50292337e-01 -8.83888781e-01
-3.61513436e-01 -2.70963460e-01 5.11954367e-01 -7.65420318e-01
-8.52612019e-01 1.79274753e-01 -8.22634816e-01 -1.25098062e+00
2.36872584e-01 -5.15223205e-01 -8.65692437e-01 6.45304441e-01
-9.29510176e-01 -1.11183655e+00 -1.48869632e-02 -1.30644530e-01
2.10182682e-01 -2.34497059e-02 1.04267967e+00 1.66169941e-01
-9.27075922e-01 8.83557871e-02 1.26262680e-01 -4.23897766e-02
3.47546548e-01 -1.02628589e+00 -1.70987874e-01 6.23619676e-01
-9.24017727e-02 6.21212780e-01 7.29123294e-01 -8.67309391e-01
-1.37011087e+00 -1.07542992e+00 8.62062395e-01 -4.08709615e-01
2.56772250e-01 -1.33976877e-01 -5.31030238e-01 7.90001690e-01
-3.31143616e-03 -1.40649140e-01 1.21837115e+00 -2.31670529e-01
5.15137851e-01 -6.74262643e-02 -1.19446337e+00 8.98072302e-01
5.53917825e-01 1.51312235e-03 2.37099752e-01 4.07119662e-01
1.37131542e-01 -2.31374249e-01 -1.24528384e+00 3.42675060e-01
7.96093225e-01 -7.39569485e-01 5.07195175e-01 -1.57398403e-01
5.22442579e-01 -4.90752190e-01 1.99673727e-01 -9.76245522e-01
-3.84514928e-01 -7.72674561e-01 3.46230865e-01 1.44405866e+00
9.38252151e-01 -5.72744727e-01 1.30420601e+00 8.11238647e-01
-1.31691024e-01 -1.23805225e+00 -9.67499495e-01 -5.91316819e-01
5.23418903e-01 3.03803328e-02 1.74285680e-01 8.72970104e-01
4.32215720e-01 1.33871421e-01 1.95318997e-01 -4.41694371e-02
3.56589735e-01 4.19868939e-02 6.15216196e-01 -1.18532693e+00
1.59806117e-01 -6.00456834e-01 -3.07687432e-01 -3.96517605e-01
-1.97594568e-01 -9.92697001e-01 6.18097857e-02 -1.96565521e+00
2.17911258e-01 -7.34633803e-01 -2.04959720e-01 4.68243450e-01
-2.17402317e-02 5.33639073e-01 -8.75460058e-02 2.81851798e-01
-2.90560961e-01 1.23409361e-01 9.94603693e-01 3.32681200e-04
-2.70599276e-01 1.04955817e-03 -7.11564839e-01 6.28007770e-01
1.21182883e+00 -5.98756969e-01 2.32820213e-01 -3.97144228e-01
2.20237494e-01 4.56857055e-01 7.72067234e-02 -1.13744140e+00
4.63391125e-01 -2.62578934e-01 7.85571754e-01 -5.71615934e-01
3.84561777e-01 -5.96557796e-01 6.65063739e-01 5.90312123e-01
-1.58487875e-02 -4.51077729e-01 2.21698001e-01 2.31394231e-01
1.04978926e-01 -8.57156217e-02 9.63436544e-01 -3.91468376e-01
-8.98245201e-02 5.08869052e-01 -7.74784386e-01 -2.31908426e-01
1.44994366e+00 -7.52236187e-01 -3.26753050e-01 -1.95449591e-01
-8.71847808e-01 2.59688616e-01 1.34140432e+00 -2.54187077e-01
5.37626088e-01 -1.17308354e+00 -6.82640493e-01 -1.77518144e-01
6.26346562e-03 1.00385077e-01 3.53700012e-01 9.65059936e-01
-1.31189477e+00 2.67529964e-01 -3.75979304e-01 -6.89373791e-01
-1.14800942e+00 4.08679128e-01 3.07505816e-01 -1.23857722e-01
-2.44540557e-01 1.05320978e+00 -5.01033664e-02 -1.95489481e-01
-4.62133676e-01 2.34837160e-01 -8.12572241e-02 -1.48244053e-01
1.78232774e-01 2.66689122e-01 -1.73839137e-01 -7.79579699e-01
-4.14239407e-01 9.51225698e-01 1.36550024e-01 -1.58468038e-01
1.56876040e+00 4.00188528e-02 -3.16453099e-01 5.46915591e-01
1.27368331e+00 -4.06761765e-02 -1.28124666e+00 6.88982010e-01
-3.04235555e-02 -3.39713931e-01 -7.85406306e-02 -8.80738974e-01
-1.15379393e+00 6.43631995e-01 4.26360250e-01 -7.02100545e-02
1.01130497e+00 4.84620258e-02 1.02528691e+00 -6.89324066e-02
5.58101892e-01 -1.16213012e+00 -3.74592602e-01 1.88326433e-01
6.66037619e-01 -1.10929549e+00 -9.11782682e-02 -5.27689278e-01
-1.64505348e-01 1.04452944e+00 4.33836967e-01 1.22879341e-01
5.45292139e-01 4.50134993e-01 2.50826418e-01 -1.24477163e-01
-8.58014345e-01 1.76095486e-01 -4.29834932e-01 7.03650117e-01
5.30586720e-01 -2.94568855e-02 -4.32777703e-01 5.27145147e-01
-4.30257358e-02 6.15071237e-01 6.07466280e-01 1.06432486e+00
-4.76547509e-01 -1.12748361e+00 -1.07254535e-01 7.31002033e-01
-7.83262789e-01 2.56335735e-01 -4.93221879e-01 3.79538238e-01
2.14859292e-01 6.80536509e-01 1.72845572e-01 -7.72867128e-02
-9.73421410e-02 1.27991244e-01 6.15876794e-01 -5.00319421e-01
-3.48898739e-01 2.73041517e-01 -1.30628084e-03 -2.33041689e-01
-2.42540032e-01 -8.24616849e-01 -1.67158985e+00 -2.54717767e-01
-2.15554997e-01 -6.91069365e-02 8.23016942e-01 8.21770906e-01
7.45543480e-01 1.09353460e-01 4.21812296e-01 -9.46532488e-01
7.76113510e-01 -4.53673184e-01 -4.89233285e-01 -1.32880616e-03
-1.50974579e-02 -1.39758185e-01 -5.02412081e-01 7.39156306e-01]
|
[14.561763763427734, -3.1678571701049805]
|
ad6f2ce2-09c3-4b85-a7ec-6073eabd89d9
|
mplug-2-a-modularized-multi-modal-foundation
|
2302.00402
| null |
https://arxiv.org/abs/2302.00402v1
|
https://arxiv.org/pdf/2302.00402v1.pdf
|
mPLUG-2: A Modularized Multi-modal Foundation Model Across Text, Image and Video
|
Recent years have witnessed a big convergence of language, vision, and multi-modal pretraining. In this work, we present mPLUG-2, a new unified paradigm with modularized design for multi-modal pretraining, which can benefit from modality collaboration while addressing the problem of modality entanglement. In contrast to predominant paradigms of solely relying on sequence-to-sequence generation or encoder-based instance discrimination, mPLUG-2 introduces a multi-module composition network by sharing common universal modules for modality collaboration and disentangling different modality modules to deal with modality entanglement. It is flexible to select different modules for different understanding and generation tasks across all modalities including text, image, and video. Empirical study shows that mPLUG-2 achieves state-of-the-art or competitive results on a broad range of over 30 downstream tasks, spanning multi-modal tasks of image-text and video-text understanding and generation, and uni-modal tasks of text-only, image-only, and video-only understanding. Notably, mPLUG-2 shows new state-of-the-art results of 48.0 top-1 accuracy and 80.3 CIDEr on the challenging MSRVTT video QA and video caption tasks with a far smaller model size and data scale. It also demonstrates strong zero-shot transferability on vision-language and video-language tasks. Code and models will be released in https://github.com/alibaba/AliceMind.
|
['Jingren Zhou', 'Fei Huang', 'Songfang Huang', 'Ji Zhang', 'Guohai Xu', 'Wei Wang', 'Qi Qian', 'Bin Bi', 'Chenliang Li', 'Yuanhong Xu', 'Jiabo Ye', 'Yaya Shi', 'Ming Yan', 'Qinghao Ye', 'Haiyang Xu']
|
2023-02-01
| null | null | null | null |
['visual-grounding', 'action-classification', 'video-question-answering', 'video-retrieval']
|
['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision']
|
[ 3.60315502e-01 -2.46549193e-02 -5.65288901e-01 -2.06850126e-01
-1.34972179e+00 -8.07984173e-01 9.41555738e-01 -5.18225789e-01
-4.46046114e-01 5.02318859e-01 4.54068273e-01 -4.08932745e-01
2.92804271e-01 -3.07303369e-01 -1.03303504e+00 -6.24304771e-01
2.64341831e-01 3.73865724e-01 -1.12068936e-01 -3.15136969e-01
-1.32417157e-01 -2.83578634e-01 -1.39014351e+00 1.03786612e+00
7.06885576e-01 8.66178930e-01 4.20971543e-01 9.72074151e-01
1.48252919e-01 9.68709171e-01 1.65827662e-01 -6.67233884e-01
1.31227419e-01 -4.86342996e-01 -1.03351772e+00 7.00354576e-02
9.28288460e-01 -5.39893806e-01 -1.01092994e+00 8.66629243e-01
7.31179237e-01 -8.73189196e-02 6.16397500e-01 -1.64018750e+00
-1.09992182e+00 6.61045134e-01 -6.17031872e-01 8.07996839e-02
6.01281822e-01 8.28081012e-01 1.08632040e+00 -1.08018613e+00
8.24518859e-01 1.35745573e+00 2.85796970e-01 1.17837799e+00
-1.16301596e+00 -7.40943909e-01 -4.55865078e-02 5.52654088e-01
-1.03268886e+00 -8.08897734e-01 2.33377159e-01 -5.24413764e-01
1.06521118e+00 1.05789386e-01 2.13892877e-01 1.81364989e+00
5.66556230e-02 1.11081266e+00 1.29465532e+00 -3.00435692e-01
-3.38209510e-01 -9.65829790e-02 -1.22856475e-01 8.30722868e-01
-1.82535842e-01 2.52348274e-01 -1.07295120e+00 3.11996549e-01
6.42330289e-01 -3.31210613e-01 -5.42909980e-01 -2.32791021e-01
-1.76652884e+00 8.27678561e-01 2.88083673e-01 -3.09418608e-02
-2.09464710e-02 5.66824079e-01 7.18112528e-01 4.91200149e-01
1.66268907e-02 4.12556529e-02 -3.13057333e-01 -2.60967791e-01
-7.49075711e-01 -1.64747655e-01 5.35893142e-01 1.07257867e+00
5.02780616e-01 5.72447628e-02 -5.53963900e-01 6.20273829e-01
4.36251462e-01 9.36528206e-01 3.26358527e-01 -1.14341199e+00
9.23744917e-01 3.63060855e-03 -3.44720989e-01 -1.84786990e-01
-2.60884166e-01 3.27083096e-02 -1.00967002e+00 8.96978192e-03
2.64920771e-01 -1.22569814e-01 -1.39626062e+00 2.24013972e+00
-6.39551505e-02 2.86012888e-01 4.43882555e-01 1.07605159e+00
1.53393674e+00 7.93141842e-01 4.44424748e-01 8.23860317e-02
1.61109388e+00 -1.36895990e+00 -4.36961919e-01 -3.21106613e-01
5.30766189e-01 -7.15951145e-01 1.11313236e+00 2.87236303e-01
-1.24494994e+00 -6.81775093e-01 -6.25093579e-01 -4.55195129e-01
-2.61054277e-01 6.42673345e-03 7.33160973e-01 4.24953789e-01
-1.41398144e+00 6.87343404e-02 -5.37999511e-01 -4.31498379e-01
4.96827662e-01 2.05022603e-01 -8.87871981e-01 -6.70507371e-01
-1.43707669e+00 9.97277498e-01 6.03217185e-01 -1.38480917e-01
-1.48570502e+00 -7.89697230e-01 -1.12754869e+00 -1.58759609e-01
2.33079299e-01 -1.40499914e+00 1.22148824e+00 -1.04867566e+00
-1.32491612e+00 1.28787637e+00 -1.58379316e-01 -4.42243338e-01
4.58188623e-01 -1.60657108e-01 -5.89240551e-01 5.93006670e-01
1.06059283e-01 1.45294821e+00 1.02824521e+00 -1.22090673e+00
-4.63228106e-01 -1.73661690e-02 3.52073669e-01 5.62962711e-01
-3.38036306e-02 -5.87030537e-02 -8.40185881e-01 -2.84433275e-01
-1.98460385e-01 -9.54740465e-01 3.00565004e-01 -1.16634950e-01
-3.95751983e-01 -1.25204921e-01 5.32468438e-01 -7.91929722e-01
6.68574929e-01 -2.21737552e+00 5.29437602e-01 -5.21290183e-01
4.16268677e-01 -2.24256031e-02 -8.69483352e-01 6.21962428e-01
-1.62338063e-01 8.56541842e-02 -4.13567796e-02 -5.60240924e-01
2.33937860e-01 1.04634635e-01 -4.16771084e-01 3.25727969e-01
2.03968242e-01 1.41678250e+00 -8.79430294e-01 -6.25459194e-01
3.92896414e-01 4.35666710e-01 -4.81333077e-01 1.04112171e-01
-3.31143767e-01 6.78530514e-01 1.21942684e-02 9.26366270e-01
6.14753544e-01 -6.55000627e-01 -1.96406636e-02 -6.44139469e-01
2.49989957e-01 -2.35028379e-02 -5.83537638e-01 2.41842699e+00
-5.27843356e-01 8.95550072e-01 1.34488702e-01 -8.38087857e-01
-1.34106487e-01 7.10075259e-01 3.13500226e-01 -8.65070581e-01
1.49343818e-01 1.64820373e-01 -1.31594628e-04 -7.42137671e-01
5.66768229e-01 -2.18698502e-01 -3.99862856e-01 3.27094138e-01
9.75741625e-01 -5.39687648e-02 3.34140092e-01 8.33934844e-01
9.15515244e-01 2.25782335e-01 -3.45797464e-02 2.37092853e-01
3.42126638e-01 -9.38235298e-02 2.06386298e-02 1.05765700e+00
-3.02806765e-01 8.40658188e-01 2.24231437e-01 4.03144300e-01
-9.24149692e-01 -1.55759430e+00 1.03000812e-01 1.41331816e+00
3.97967905e-01 -3.99522305e-01 -3.73759329e-01 -6.06983781e-01
-1.57275021e-01 6.68786883e-01 -5.68698704e-01 -3.17764461e-01
-1.10566899e-01 -4.78807867e-01 9.87138808e-01 3.41757715e-01
6.94301009e-01 -9.47217524e-01 -4.48812135e-02 -2.71054000e-01
-9.40200448e-01 -1.74390364e+00 -6.16414189e-01 -2.02780962e-01
-5.92208445e-01 -9.74085629e-01 -7.88972139e-01 -7.83527076e-01
2.75296271e-01 4.55089897e-01 1.34437072e+00 -2.66820729e-01
-3.11599448e-02 1.07372713e+00 -4.43109274e-01 1.28414139e-01
-5.10426104e-01 -5.97864576e-02 -1.74409896e-01 -5.91084100e-02
-3.78237404e-02 -2.85486370e-01 -6.31226599e-01 9.52366590e-02
-1.05078566e+00 6.89611971e-01 8.10275555e-01 1.12943995e+00
2.29295313e-01 -7.42561758e-01 5.59373021e-01 -4.02369291e-01
2.94484466e-01 -6.05117261e-01 -2.81268097e-02 5.95165014e-01
-1.49932161e-01 -3.78852002e-02 3.30668896e-01 -5.85348427e-01
-1.16280377e+00 -1.20523907e-01 -2.78052762e-02 -8.02342653e-01
-3.16328406e-01 4.87125158e-01 -3.44576418e-01 -6.23324402e-02
4.51749086e-01 6.92398727e-01 5.53012379e-02 -7.27775767e-02
1.16054153e+00 5.62939167e-01 9.08457875e-01 -5.51220596e-01
7.41027951e-01 5.63764632e-01 -2.60458231e-01 -6.22710884e-01
-6.69803262e-01 -3.92465055e-01 -3.70711744e-01 -4.69902068e-01
1.33054292e+00 -1.60783327e+00 -7.78149426e-01 8.08659792e-01
-1.24065471e+00 -5.61783612e-01 -9.95439515e-02 5.14246345e-01
-7.17378616e-01 6.04971349e-01 -9.13189530e-01 -2.75632083e-01
-4.89933282e-01 -1.29400289e+00 1.23285484e+00 3.45473677e-01
2.99019545e-01 -1.00844431e+00 -2.52105277e-02 1.12937307e+00
2.95612007e-01 -2.75931537e-01 6.73028290e-01 -3.37593675e-01
-8.63245249e-01 2.87097186e-01 -6.15944028e-01 3.05501372e-01
-3.21743220e-01 -4.47785825e-01 -1.25379527e+00 -4.82593387e-01
-4.86169934e-01 -1.14491415e+00 1.32400513e+00 2.77977914e-01
8.51428211e-01 1.17133848e-01 -1.97831467e-01 9.37703967e-01
1.37745035e+00 -2.13562056e-01 8.67704511e-01 8.35196897e-02
9.91147697e-01 2.51716554e-01 4.72360015e-01 7.53897801e-02
8.45012069e-01 4.85969931e-01 5.78467131e-01 -1.19208492e-01
-6.45292580e-01 -3.52344722e-01 8.91779482e-01 7.89573669e-01
5.40118143e-02 -6.66413188e-01 -6.77573860e-01 5.50973237e-01
-1.91981113e+00 -1.26719546e+00 -1.55876279e-01 1.80247557e+00
8.49856079e-01 -3.54952246e-01 1.55114774e-02 -6.91632807e-01
6.60803318e-01 3.55444282e-01 -5.49536109e-01 -1.64017573e-01
-5.92389643e-01 3.66226472e-02 4.78136599e-01 3.58536452e-01
-1.16881192e+00 1.23940897e+00 6.21820164e+00 1.16899490e+00
-1.07180440e+00 5.56395411e-01 5.03184438e-01 -2.35183299e-01
-4.95672822e-01 3.68722975e-02 -6.24279082e-01 4.11349356e-01
9.39220309e-01 1.59857050e-01 5.51382780e-01 2.13172391e-01
-6.70870394e-02 -2.09317207e-01 -1.19312620e+00 1.35706913e+00
3.16194952e-01 -1.45917606e+00 3.75855654e-01 -1.44177690e-01
8.17456365e-01 6.71727180e-01 3.37745279e-01 4.70369309e-01
1.21858992e-01 -1.23958039e+00 8.09797287e-01 5.06302714e-01
1.35868299e+00 -1.42585143e-01 3.34723622e-01 1.48893058e-01
-1.07935202e+00 -2.12139450e-02 1.03625730e-01 3.19494754e-01
6.15262032e-01 1.41716003e-01 -3.74095947e-01 8.72016132e-01
5.53653896e-01 6.73630714e-01 -4.87570435e-01 7.96109378e-01
-2.01975480e-01 6.09808981e-01 -1.05613358e-01 5.02259672e-01
3.06803882e-01 1.33155107e-01 4.96640295e-01 1.48588276e+00
2.51246452e-01 -7.34879673e-02 9.68989506e-02 7.20722616e-01
-3.47600818e-01 -3.03881556e-01 -7.16614246e-01 -3.19162428e-01
3.16501826e-01 1.22771847e+00 -1.82920277e-01 -5.53489745e-01
-7.76281178e-01 1.43101740e+00 8.99220631e-02 6.07537210e-01
-1.25086439e+00 5.22186384e-02 5.34873247e-01 -3.79752547e-01
1.89019769e-01 -2.49816775e-01 -7.37758204e-02 -1.71207952e+00
-3.51141453e-01 -1.08602452e+00 6.53188229e-01 -1.29806089e+00
-1.45793724e+00 4.57635731e-01 9.01816115e-02 -1.16514897e+00
-3.41610193e-01 -8.22649062e-01 -4.13719118e-01 8.26093435e-01
-1.48453152e+00 -1.87513793e+00 -3.44444096e-01 1.21577692e+00
6.29870117e-01 -2.66543299e-01 6.16429508e-01 4.26630676e-01
-2.46482998e-01 8.27240646e-01 -2.90583167e-02 2.58972257e-01
1.03437757e+00 -9.81453836e-01 1.76519692e-01 1.05283809e+00
2.76700199e-01 2.21610546e-01 4.43803072e-01 -3.78649801e-01
-1.81820095e+00 -8.25413585e-01 6.12112224e-01 -6.34617686e-01
9.04924333e-01 -3.53937596e-01 -4.46788877e-01 8.05389702e-01
8.54503810e-01 -1.12386562e-01 6.24356270e-01 -8.15299973e-02
-9.75999773e-01 2.53293544e-01 -8.76110971e-01 6.97713017e-01
1.38390672e+00 -1.17306066e+00 -5.06527185e-01 4.96119201e-01
9.61290777e-01 -5.52940607e-01 -8.83096099e-01 4.14156646e-01
5.34214497e-01 -8.52093816e-01 1.16392100e+00 -6.83147073e-01
9.41257060e-01 -1.31818026e-01 -4.49593306e-01 -1.13122261e+00
-2.48151571e-01 -6.23073041e-01 -3.05376142e-01 1.11351621e+00
4.25959498e-01 -5.44656038e-01 3.32855225e-01 2.53489703e-01
-3.44032764e-01 -3.70641708e-01 -1.04212773e+00 -5.94076633e-01
1.18898451e-01 -6.62083507e-01 -1.02129988e-01 9.74911392e-01
2.50836089e-02 7.87507772e-01 -7.45341837e-01 7.52614737e-02
8.24964464e-01 1.16166927e-01 6.70641005e-01 -4.44201648e-01
-6.16609514e-01 -4.19599712e-01 -1.83575287e-01 -1.24961913e+00
3.53657484e-01 -1.26049471e+00 -3.68943904e-03 -1.72488773e+00
7.95280039e-01 1.58887893e-01 -2.31850460e-01 6.64081097e-01
-8.51985812e-02 6.53436959e-01 7.20667958e-01 2.32714400e-01
-1.12286425e+00 5.51476359e-01 1.55327594e+00 -3.27441245e-01
3.24768364e-01 -5.43350697e-01 -8.07569742e-01 2.17293605e-01
6.24661684e-01 -1.07890433e-02 -6.00784779e-01 -7.78543770e-01
1.04645133e-01 3.29170495e-01 9.43989933e-01 -7.25534678e-01
1.44381002e-01 -1.03436388e-01 1.07757814e-01 -4.60346639e-01
7.17492163e-01 -3.79532218e-01 1.89518407e-01 2.49139637e-01
-3.47995877e-01 -1.12470917e-01 3.58326763e-01 5.26247203e-01
-2.57645816e-01 1.12817995e-02 5.93645513e-01 -1.84704080e-01
-1.40441358e+00 3.59579384e-01 -2.52059996e-01 3.74772549e-01
7.53339410e-01 -6.99741319e-02 -1.14328527e+00 -7.09784985e-01
-8.42627108e-01 4.61099237e-01 3.68892759e-01 7.52871037e-01
7.58327067e-01 -1.34078932e+00 -1.00342560e+00 -1.34949476e-01
4.81580406e-01 -4.61303502e-01 9.43396270e-01 1.20381808e+00
-3.21616307e-02 4.90444362e-01 -3.39440614e-01 -9.67081785e-01
-1.32151556e+00 4.37849909e-01 2.66721755e-01 -1.75651819e-01
-3.08320135e-01 9.96140897e-01 6.51297987e-01 -4.27393228e-01
-1.80214182e-01 1.89132348e-01 3.05619776e-01 2.74632722e-02
3.98061752e-01 2.09071543e-02 -3.29246998e-01 -9.33653593e-01
-3.61803204e-01 4.44121778e-01 3.40596251e-02 -5.84660113e-01
7.61101961e-01 -3.93123001e-01 -2.14322619e-02 2.95183569e-01
1.24769533e+00 -5.22411048e-01 -1.37133121e+00 -1.42826587e-01
-6.51880801e-01 -1.54894054e-01 -7.96860904e-02 -1.34948730e+00
-1.06947613e+00 1.06906033e+00 7.26725042e-01 -2.59093344e-01
1.32076442e+00 5.64631641e-01 7.94002116e-01 1.64498389e-01
3.71212244e-01 -8.32293749e-01 4.09860164e-01 8.25511098e-01
8.60536337e-01 -1.63405490e+00 -3.65792423e-01 -2.42020369e-01
-1.08187962e+00 7.46941686e-01 7.84276426e-01 4.03173566e-01
1.96563557e-01 -1.00416273e-01 2.30603740e-02 -6.14792779e-02
-1.07036173e+00 -6.02555096e-01 6.37458265e-01 8.44997525e-01
3.61270517e-01 1.06684707e-01 1.23179682e-01 3.80568266e-01
1.97831884e-01 6.69134930e-02 3.28405619e-01 6.23061061e-01
-1.01073906e-01 -9.34283555e-01 -2.81825662e-01 1.95745990e-01
-2.36840263e-01 -6.33457959e-01 -4.92599607e-02 8.68866265e-01
1.06414847e-01 1.04360700e+00 -1.21166885e-01 -6.01766348e-01
-2.31168166e-01 2.45530963e-01 1.09869230e+00 -3.60129982e-01
-3.98356110e-01 7.40181357e-02 3.19254696e-01 -7.33465612e-01
-6.54247105e-01 -4.67170596e-01 -1.16922486e+00 -3.88192028e-01
-1.34874746e-01 -3.18557799e-01 4.85709220e-01 9.95475829e-01
6.27633333e-01 3.76242965e-01 2.47637078e-01 -1.00187516e+00
-3.51494819e-01 -9.69226778e-01 -1.74875870e-01 6.43915415e-01
3.97705138e-01 -5.28613091e-01 -2.91327208e-01 3.52074474e-01]
|
[10.865400314331055, 1.358234167098999]
|
a09d138f-8db3-4f38-a329-981024aee5d2
|
an-order-complexity-model-for-aesthetic
|
2301.05908
| null |
https://arxiv.org/abs/2301.05908v1
|
https://arxiv.org/pdf/2301.05908v1.pdf
|
An Order-Complexity Model for Aesthetic Quality Assessment of Symbolic Homophony Music Scores
|
Computational aesthetics evaluation has made great achievements in the field of visual arts, but the research work on music still needs to be explored. Although the existing work of music generation is very substantial, the quality of music score generated by AI is relatively poor compared with that created by human composers. The music scores created by AI are usually monotonous and devoid of emotion. Based on Birkhoff's aesthetic measure, this paper proposes an objective quantitative evaluation method for homophony music score aesthetic quality assessment. The main contributions of our work are as follows: first, we put forward a homophony music score aesthetic model to objectively evaluate the quality of music score as a baseline model; second, we put forward eight basic music features and four music aesthetic features.
|
['Shuai Cui', 'Yiqing Rong', 'Duo Xu', 'Jinyu Wang', 'Wu Zhou', 'Xin Jin']
|
2023-01-14
| null | null | null | null |
['music-generation', 'music-generation']
|
['audio', 'music']
|
[-3.36036161e-02 -2.51223207e-01 2.75865704e-01 9.56598744e-02
-5.05763113e-01 -5.13440549e-01 6.81643486e-02 -1.16791993e-01
1.30426377e-01 2.78154939e-01 3.40010613e-01 4.72532421e-01
-3.60021204e-01 -7.12686300e-01 1.33609250e-01 -4.96412009e-01
1.72795817e-01 3.11363310e-01 -1.79354951e-01 -4.43172187e-01
6.36365473e-01 1.54157996e-01 -1.64331269e+00 4.10899282e-01
7.72988379e-01 1.22898018e+00 -1.04810692e-01 6.25149071e-01
-1.35466740e-01 3.82991999e-01 -9.36326921e-01 -7.58683860e-01
3.51580024e-01 -1.11743963e+00 -7.22966433e-01 2.24265666e-03
-1.16142463e-02 1.19413018e-01 5.27258962e-02 1.25749850e+00
9.83658612e-01 8.83226544e-02 8.48701894e-01 -1.39565897e+00
-1.07449102e+00 7.73368418e-01 -4.42272663e-01 -4.18140978e-01
3.38177204e-01 2.33421817e-01 1.78615236e+00 -6.94594383e-01
3.43424797e-01 1.12985146e+00 6.65811956e-01 3.51683378e-01
-1.05506516e+00 -8.86642575e-01 -4.56577331e-01 4.20597196e-01
-1.32571959e+00 -9.21827033e-02 1.15058589e+00 -3.96138102e-01
3.93910319e-01 5.84410548e-01 1.68154037e+00 5.43568671e-01
-4.31359559e-02 8.97499323e-01 1.27564812e+00 -4.77265596e-01
2.95180947e-01 -4.77011763e-02 -4.78847623e-01 2.12240621e-01
-3.11376989e-01 1.66332126e-01 -7.17190921e-01 1.00185759e-01
9.94403780e-01 -5.28558791e-01 8.38125572e-02 -1.58010900e-01
-1.04807055e+00 7.14835823e-01 5.46052575e-01 7.90271461e-01
-1.82273075e-01 1.77677795e-01 4.30036038e-01 4.19469774e-01
2.01399013e-01 1.04933369e+00 -1.47470981e-01 -6.23239934e-01
-1.06007433e+00 1.74733862e-01 4.67359245e-01 6.59176826e-01
4.13143665e-01 2.35249966e-01 -2.43873283e-01 1.15941513e+00
1.97024047e-01 5.18006980e-01 4.78133470e-01 -1.04256392e+00
-2.85111248e-01 8.03997159e-01 -2.37513140e-01 -1.40834475e+00
-1.06761687e-01 -3.79650652e-01 -7.74368107e-01 7.65151858e-01
-4.24649939e-03 3.26050609e-01 -2.37420037e-01 1.32318938e+00
-2.41493121e-01 -2.85433650e-01 -1.86935350e-01 1.16935825e+00
1.22015893e+00 6.63196683e-01 -6.29834235e-02 -8.81860331e-02
1.27928948e+00 -1.13045001e+00 -9.08621311e-01 4.07354802e-01
5.61210401e-02 -1.51714396e+00 1.80019903e+00 7.95906305e-01
-1.47888803e+00 -7.35578001e-01 -1.07833326e+00 5.60810342e-02
1.31793261e-01 3.27886760e-01 7.98146844e-01 8.83045197e-01
-9.54611123e-01 7.95976281e-01 -1.01434000e-01 -1.36050239e-01
5.10213494e-01 1.58706382e-01 6.33218233e-03 6.80522025e-01
-9.24032569e-01 7.27049112e-01 4.21294779e-01 -3.25787775e-02
-4.75190967e-01 -5.36027372e-01 -2.63382226e-01 1.76529184e-01
-4.57818508e-02 -7.87647665e-01 1.36413479e+00 -1.61798823e+00
-2.11799479e+00 1.00059843e+00 3.78482789e-01 3.14355999e-01
2.84804642e-01 -4.32883799e-02 -7.32626319e-01 -2.36990731e-02
-1.44247301e-02 4.97471958e-01 4.36395764e-01 -1.45380771e+00
-3.56575459e-01 -1.82299651e-02 4.58967946e-02 4.18456644e-01
-6.36670113e-01 3.58184963e-01 -5.68894982e-01 -1.25857246e+00
1.24759443e-01 -8.04338515e-01 -2.56695915e-02 5.28409407e-02
-5.12929022e-01 -1.65803850e-01 1.93913326e-01 -3.38669807e-01
1.77552629e+00 -2.25187349e+00 1.10719889e-01 3.71909469e-01
-2.27229204e-02 6.62108883e-02 -3.29408407e-01 5.29345334e-01
-2.14392528e-01 1.77228302e-01 -3.61850262e-02 3.67682874e-02
3.90983969e-01 -2.89158523e-01 -1.23918816e-01 -2.57225662e-01
-2.80660719e-01 9.99857426e-01 -8.14263642e-01 -7.46777952e-01
1.36833787e-01 2.50018597e-01 -6.99564457e-01 1.18764155e-01
1.24700487e-01 4.41839606e-01 -4.86078560e-01 9.01790798e-01
3.59470993e-01 -9.93647203e-02 -1.06307514e-01 -4.34562296e-01
-2.17666283e-01 -4.69894372e-02 -1.40746212e+00 1.95661700e+00
-4.21828389e-01 5.42401910e-01 -3.56923461e-01 -1.90765426e-01
1.27936530e+00 4.68260586e-01 6.37478054e-01 -8.42760324e-01
4.72142190e-01 3.87643486e-01 8.33550766e-02 -4.36980009e-01
6.39344096e-01 -8.07376027e-01 -4.29013371e-02 5.73734999e-01
-4.09496486e-01 -7.90726125e-01 1.78249646e-02 -1.08573422e-01
6.94961667e-01 1.36692271e-01 4.60045636e-01 -5.48532531e-02
5.79279661e-01 2.86400206e-02 4.49248582e-01 3.59452155e-04
-1.54379889e-01 8.12455773e-01 3.50886941e-01 -4.63131875e-01
-1.13008678e+00 -1.35720968e+00 3.37332338e-02 1.07219803e+00
3.58549207e-01 -9.02957618e-01 -7.52943158e-01 -2.60015786e-01
-2.79362857e-01 3.82431656e-01 -4.00011778e-01 -1.86763376e-01
-2.18483493e-01 -5.74362099e-01 6.56094670e-01 4.48247939e-01
5.40268242e-01 -1.73680770e+00 -1.94856435e-01 1.32989669e-02
-4.37233567e-01 -2.18288600e-01 -7.43893445e-01 -5.74431956e-01
-8.62329900e-01 -9.93661582e-01 -8.66075456e-01 -9.02011812e-01
3.37566257e-01 -9.00851190e-02 1.44703877e+00 9.32281371e-03
-5.11943281e-01 1.77986190e-01 -5.34473240e-01 -6.39495432e-01
-2.35831201e-01 -2.63881296e-01 -1.42719179e-01 -7.73467645e-02
2.25641757e-01 -8.62568498e-01 -1.02031946e+00 5.94002306e-01
-8.81992638e-01 2.15266511e-01 8.41619253e-01 4.27873462e-01
8.88964176e-01 2.59079427e-01 5.98728359e-01 -2.99238056e-01
9.97578800e-01 3.35899144e-01 1.83089375e-02 -2.59390622e-02
-7.27623582e-01 -2.35087171e-01 3.82428318e-01 -3.82843822e-01
-8.05005670e-01 -8.35745931e-02 -2.97540098e-01 -1.82802781e-01
1.29741490e-01 2.80381262e-01 -3.11065078e-01 -2.02684045e-01
6.85099721e-01 -2.07282111e-01 -2.39777356e-01 -6.77728653e-01
4.31997567e-01 6.20499134e-01 5.65239668e-01 -6.96806312e-01
8.14188838e-01 2.76204169e-01 -1.26968399e-02 -5.51439762e-01
-5.65324068e-01 -3.98790568e-01 -4.38950777e-01 -7.73823440e-01
6.95813477e-01 -3.99794549e-01 -1.31668949e+00 2.92867929e-01
-8.74292195e-01 1.27686635e-02 -9.05181289e-01 4.56167012e-01
-8.98984432e-01 5.89628220e-01 -4.25304055e-01 -9.16169703e-01
-6.70321882e-01 -1.02540338e+00 1.03303778e+00 2.46569410e-01
-6.74327850e-01 -7.03602374e-01 5.43431044e-01 5.66645980e-01
2.71267325e-01 2.76681393e-01 9.51252401e-01 2.12349460e-01
-4.05244827e-01 -1.34557679e-01 -2.98975408e-01 4.16421115e-01
1.58934284e-03 1.75523907e-01 -9.94690716e-01 2.03521937e-01
-7.65025839e-02 -4.81312662e-01 5.96927464e-01 3.52424622e-01
1.08806932e+00 -7.32939318e-02 3.56816530e-01 5.73370218e-01
1.47640634e+00 3.76879871e-01 9.63068902e-01 6.66625082e-01
3.54556203e-01 4.63497877e-01 7.88321018e-01 7.27076292e-01
-7.50080720e-02 9.32580173e-01 3.32527727e-01 -1.94375262e-01
-4.00408357e-01 -4.27316338e-01 2.65463352e-01 1.44242561e+00
-7.96818316e-01 8.63462761e-02 -4.85949367e-01 3.57046127e-01
-1.67933226e+00 -9.84147608e-01 -2.41420031e-01 2.02299380e+00
8.80846858e-01 -2.56438047e-01 4.38836724e-01 8.19796562e-01
6.71285629e-01 -1.34039715e-01 -1.01523124e-01 -5.83329439e-01
-4.06210959e-01 6.51091337e-01 -2.00921640e-01 -2.20237952e-02
-8.97544146e-01 1.02077985e+00 6.94438219e+00 1.19085085e+00
-8.06669354e-01 -1.44716799e-01 1.47736222e-01 -3.77301544e-01
-4.06624436e-01 9.25580338e-02 2.01847360e-01 3.33936185e-01
2.22615987e-01 -3.15460086e-01 5.52233756e-01 7.68974066e-01
3.60629588e-01 2.68264681e-01 -6.94100499e-01 1.63556719e+00
8.29199851e-02 -8.57172251e-01 2.90140212e-01 -1.28824905e-01
9.29447114e-01 -8.52920353e-01 5.66126347e-01 -5.09365648e-02
-4.97599877e-03 -1.20552599e+00 8.03634167e-01 8.29909086e-01
1.04337358e+00 -1.08087957e+00 5.23837805e-01 -3.74365717e-01
-1.44085443e+00 2.01564327e-01 -3.05351108e-01 -4.40102071e-01
2.56756097e-01 4.54243630e-01 -1.51823074e-01 4.76033896e-01
7.02585995e-01 8.24811578e-01 -5.47075510e-01 1.66723514e+00
-3.54106814e-01 4.71114814e-01 2.06565470e-01 -2.52418965e-01
-3.95184010e-02 -3.67078513e-01 6.13034487e-01 9.09190297e-01
7.08463073e-01 1.53672434e-02 -1.96956560e-01 9.61882472e-01
1.42379254e-01 1.24523807e+00 -9.98568758e-02 -3.69019389e-01
-5.63412383e-02 1.44462156e+00 -7.97095001e-01 -3.39622758e-02
-1.38514861e-02 1.15634120e+00 -3.73508543e-01 -1.75065193e-02
-7.11928904e-01 -5.61707795e-01 4.74744350e-01 -1.70280769e-01
-2.52642661e-01 1.81342259e-01 -9.78885889e-01 -9.38535810e-01
-6.79154769e-02 -9.17564929e-01 2.91870266e-01 -1.26587522e+00
-1.55387962e+00 5.23672760e-01 -7.74942875e-01 -1.88735902e+00
3.94045085e-01 -5.29451609e-01 -1.02082360e+00 6.91996694e-01
-8.48075926e-01 -1.09250093e+00 -4.14686918e-01 3.63125294e-01
3.18143278e-01 -4.90953416e-01 1.15597379e+00 3.73817086e-01
-2.45335754e-02 7.40324140e-01 6.03240281e-02 1.01026092e-02
9.68913615e-01 -1.29374075e+00 1.04805678e-01 -4.13770713e-02
4.86822397e-01 2.26863667e-01 7.36792862e-01 -3.51833612e-01
-9.05228853e-01 -4.13534045e-01 9.71663058e-01 -1.73577219e-01
6.45039380e-01 3.67382616e-01 -3.17886025e-01 -2.04841256e-01
3.04108053e-01 -7.40924358e-01 1.26326239e+00 3.82109433e-01
-2.25796342e-01 -2.44081095e-01 -8.83213758e-01 8.23893368e-01
1.08998251e+00 -2.38560751e-01 -6.02129340e-01 6.31701499e-02
1.31395265e-01 2.05495924e-01 -1.10979497e+00 2.92866021e-01
1.22675526e+00 -1.27749884e+00 1.00137532e+00 -2.38195360e-01
8.12524736e-01 -5.66785038e-01 -4.03165706e-02 -1.21812391e+00
-5.88785231e-01 -7.68598020e-01 4.81757909e-01 1.19821870e+00
2.76894093e-01 6.97034895e-02 8.82106960e-01 -3.50950658e-02
-2.41497010e-01 -6.30839229e-01 -5.84486365e-01 -9.08610106e-01
-1.06967308e-01 -6.66790485e-01 7.55251884e-01 8.42144668e-01
3.44766051e-01 6.84048355e-01 -6.17732167e-01 -6.98311269e-01
4.49545741e-01 4.91734356e-01 9.04268503e-01 -1.57123935e+00
-6.40452504e-01 -1.28429878e+00 -7.46621907e-01 -4.42293227e-01
-3.85648489e-01 -1.08610511e+00 -9.85896885e-02 -1.69003630e+00
4.85334665e-01 -3.28940511e-01 -7.06289470e-01 3.16453218e-01
2.59312224e-02 1.06881821e+00 5.75860679e-01 3.98302585e-01
-5.55834472e-01 8.49964559e-01 2.07304597e+00 -2.36944184e-01
-4.29307103e-01 4.28634025e-02 -9.46849287e-01 9.72147703e-01
8.74246716e-01 -2.94108152e-01 -3.32889915e-01 -9.02232062e-03
9.40059423e-01 -1.03858046e-01 9.11927372e-02 -1.15494549e+00
-1.42450005e-01 -3.04994434e-01 2.94927061e-01 -4.47391987e-01
3.76834333e-01 -7.15174377e-01 3.93238753e-01 2.62086451e-01
-3.03474158e-01 1.21175908e-02 5.36470395e-03 1.52094543e-01
-5.51390767e-01 -2.63140172e-01 1.01995814e+00 -6.88443854e-02
-5.30000389e-01 9.03758183e-02 4.74561676e-02 1.34393141e-01
7.36065745e-01 -5.50381362e-01 3.01489443e-01 -4.82180685e-01
-8.67121160e-01 -4.20420349e-01 6.90758049e-01 4.19078022e-01
6.94753647e-01 -2.05633616e+00 -6.97241485e-01 -2.61701465e-01
4.45403069e-01 -9.24377263e-01 5.45508027e-01 5.80391467e-01
-8.48722577e-01 -8.91389418e-03 -6.85924232e-01 -2.92664375e-02
-1.50978565e+00 3.97548556e-01 -6.49604052e-02 5.92863886e-03
-5.19879520e-01 7.20675588e-01 4.53170687e-01 -1.42966956e-01
2.34822914e-01 1.28583342e-01 -5.17823935e-01 5.76404445e-02
4.17057753e-01 7.31863976e-01 -3.53893310e-01 -6.83907390e-01
2.06513070e-02 1.14281631e+00 7.06583261e-01 -4.22964364e-01
1.21606481e+00 1.47396773e-01 -3.93439293e-01 7.18440413e-01
7.13863075e-01 4.08628881e-01 -4.88798231e-01 3.07841241e-01
-3.13991338e-01 -8.03740859e-01 -1.94845110e-01 -1.26599336e+00
-1.20230067e+00 1.00865948e+00 7.11960018e-01 1.71887308e-01
1.70528185e+00 -1.33216545e-01 1.05302680e+00 5.36033809e-02
1.63899451e-01 -1.57341838e+00 7.07393169e-01 2.78107226e-01
1.31185496e+00 -6.75306737e-01 -1.53403878e-01 -3.00395161e-01
-1.10156393e+00 1.18355370e+00 3.37269247e-01 -1.38733998e-01
5.67315698e-01 -9.26958025e-02 4.21551704e-01 -2.12941766e-01
-6.06443286e-02 -4.28807110e-01 9.62404728e-01 4.42537427e-01
8.47732067e-01 2.22666532e-01 -8.12046647e-01 1.16365731e+00
-1.03525448e+00 8.29313844e-02 1.13529824e-01 2.86219954e-01
-6.79131091e-01 -1.31866276e+00 -3.94391865e-01 2.89060157e-02
-4.51945335e-01 -2.55953968e-01 -9.03527260e-01 3.91571820e-01
5.13330936e-01 9.87560153e-01 -3.35650891e-01 -8.02153766e-01
6.20130777e-01 -8.29411820e-02 5.67083359e-01 -3.92173231e-01
-9.36650038e-01 4.16192383e-01 -1.89102784e-01 -4.60458487e-01
-4.89764601e-01 -3.56524110e-01 -1.19046354e+00 -4.63544399e-01
-1.51870936e-01 3.40587735e-01 7.07131088e-01 4.11149651e-01
-3.78452279e-02 7.69993126e-01 6.26197278e-01 -6.72269642e-01
1.67020053e-01 -9.28181112e-01 -1.11953425e+00 6.99182153e-01
-5.58876872e-01 -6.15218699e-01 -2.24174082e-01 2.34815434e-01]
|
[16.023849487304688, 5.467034816741943]
|
d1a7b52c-a9f5-4763-9b3a-fbbf7c709c00
|
temporal-phenotyping-using-deep-predictive
|
2006.08600
| null |
https://arxiv.org/abs/2006.08600v1
|
https://arxiv.org/pdf/2006.08600v1.pdf
|
Temporal Phenotyping using Deep Predictive Clustering of Disease Progression
|
Due to the wider availability of modern electronic health records, patient care data is often being stored in the form of time-series. Clustering such time-series data is crucial for patient phenotyping, anticipating patients' prognoses by identifying "similar" patients, and designing treatment guidelines that are tailored to homogeneous patient subgroups. In this paper, we develop a deep learning approach for clustering time-series data, where each cluster comprises patients who share similar future outcomes of interest (e.g., adverse events, the onset of comorbidities). To encourage each cluster to have homogeneous future outcomes, the clustering is carried out by learning discrete representations that best describe the future outcome distribution based on novel loss functions. Experiments on two real-world datasets show that our model achieves superior clustering performance over state-of-the-art benchmarks and identifies meaningful clusters that can be translated into actionable information for clinical decision-making.
|
['Mihaela van der Schaar', 'Changhee Lee']
|
2020-06-15
| null |
https://proceedings.icml.cc/static/paper_files/icml/2020/1742-Paper.pdf
|
https://proceedings.icml.cc/static/paper_files/icml/2020/1742-Paper.pdf
|
icml-2020-1
|
['patient-phenotyping', 'time-series-clustering']
|
['medical', 'time-series']
|
[-1.15981981e-01 2.53849570e-03 -4.85126406e-01 -6.29939020e-01
-9.43843603e-01 -1.83142290e-01 1.35368139e-01 1.13750887e+00
-3.33434343e-02 5.43861747e-01 7.80442595e-01 -3.50291908e-01
-6.85720384e-01 -6.77227795e-01 -2.16372818e-01 -8.08971167e-01
-6.86604917e-01 9.78569388e-01 -5.23883343e-01 4.37618554e-01
-1.41966522e-01 3.29656065e-01 -1.07686365e+00 5.94221473e-01
8.37867916e-01 9.84263539e-01 -1.08272202e-01 2.37084180e-01
2.53334306e-02 7.64000595e-01 -4.18855548e-01 2.94163227e-02
1.07472628e-01 -5.32826304e-01 -4.82569158e-01 4.37300093e-02
-3.73572916e-01 1.52620822e-01 -3.56341660e-01 6.97784007e-01
6.54741347e-01 -6.93968609e-02 5.44653118e-01 -1.08281791e+00
-3.91835868e-01 7.13546753e-01 -3.65103096e-01 1.27751291e-01
-4.99062948e-02 6.58490881e-02 8.61520171e-01 -2.90448382e-03
4.03451741e-01 8.73805821e-01 8.39877009e-01 6.85888708e-01
-1.43759787e+00 -3.93015951e-01 1.46678030e-01 2.38622129e-01
-1.43381333e+00 -1.62280455e-01 8.06145072e-01 -5.13010681e-01
4.23978984e-01 5.13795793e-01 5.53162456e-01 1.09560943e+00
4.99018371e-01 4.65143830e-01 5.46301365e-01 1.17502205e-01
5.68646491e-01 -3.05979073e-01 1.34343132e-01 2.58303195e-01
-3.34577002e-02 -1.80871874e-01 6.22281097e-02 -5.68126857e-01
2.59181410e-01 9.04478669e-01 -1.90203384e-01 -2.67325163e-01
-1.49144661e+00 7.01045156e-01 5.16863048e-01 2.78701127e-01
-8.01271439e-01 9.98675823e-03 7.17746735e-01 1.60683379e-01
5.65285087e-01 4.34853673e-01 -6.41223729e-01 -2.02682335e-02
-9.06530976e-01 -5.35002053e-02 4.36667323e-01 6.15858734e-01
2.19155058e-01 -5.60732126e-01 -6.57872498e-01 7.02517629e-01
8.98543820e-02 1.30947635e-01 9.78563249e-01 -8.29166234e-01
4.53511447e-01 1.05195570e+00 7.99253508e-02 -7.33791709e-01
-1.02010608e+00 -4.91016716e-01 -1.27606821e+00 -4.98166472e-01
1.64368212e-01 -4.33724940e-01 -9.16739285e-01 1.70250738e+00
1.72748536e-01 6.01722658e-01 2.34872296e-01 5.96631825e-01
4.67015773e-01 6.13997638e-01 4.33593988e-01 -4.43466306e-01
1.20423353e+00 -1.29545301e-01 -5.04009068e-01 2.59297907e-01
1.06162870e+00 -1.95856929e-01 5.96186399e-01 2.45180666e-01
-7.46680081e-01 -2.60183394e-01 -3.32270324e-01 3.40742558e-01
-1.23853162e-01 1.70497909e-01 8.51359010e-01 3.29546481e-01
-6.51034355e-01 9.10807252e-01 -1.29423296e+00 -4.82662022e-01
8.12937140e-01 5.35646379e-01 -1.43843576e-01 8.93098675e-03
-1.00138462e+00 2.41729781e-01 3.78831208e-01 -1.28206730e-01
-7.29831040e-01 -1.30164886e+00 -6.23566985e-01 3.29335213e-01
-6.75506815e-02 -9.92998362e-01 8.29904318e-01 -6.11262262e-01
-9.73674595e-01 7.98903108e-01 -1.42411068e-01 -8.02671373e-01
4.10538286e-01 6.99804053e-02 -6.68394983e-01 -6.14233315e-02
7.06104785e-02 3.79861087e-01 5.03574789e-01 -8.10918391e-01
-7.69991577e-01 -7.06451893e-01 -6.81073546e-01 -1.08828554e-02
-5.16572535e-01 -5.01493998e-02 -1.75683960e-01 -6.82161629e-01
2.12748852e-02 -9.49912369e-01 -8.69965434e-01 -1.71460852e-01
-6.20634854e-01 -3.36337388e-01 7.61782289e-01 -5.75122893e-01
1.66993260e+00 -2.24377322e+00 2.57095605e-01 1.67469770e-01
7.65161514e-01 6.46580895e-03 1.94722757e-01 5.05808055e-01
-3.61948758e-01 4.93323281e-02 -2.59317905e-01 -3.21166992e-01
-3.41695130e-01 6.64574504e-02 -2.00386211e-01 6.36352956e-01
9.42223147e-02 8.29337597e-01 -8.68570387e-01 -4.08823013e-01
3.08041990e-01 3.10720146e-01 -6.80320263e-01 4.50762123e-01
-3.30415547e-01 9.62176621e-01 -8.11783314e-01 4.42764521e-01
2.12998062e-01 -9.47039127e-01 3.08239073e-01 -1.29831368e-02
1.59678847e-01 1.15518048e-01 -4.62191999e-01 1.46468806e+00
-3.35768908e-01 2.88089186e-01 -4.09462541e-01 -1.33524752e+00
7.88553178e-01 4.85123008e-01 1.44575679e+00 -3.90096575e-01
2.29229838e-01 -2.10950658e-01 1.76331043e-01 -5.22566020e-01
-2.82204181e-01 -1.56635240e-01 -3.30414593e-01 4.83293027e-01
-5.55002272e-01 4.76338178e-01 -9.12351310e-02 -1.12393297e-01
1.30987132e+00 -7.83484280e-01 2.08926424e-01 -8.27041194e-02
3.58565778e-01 1.31285772e-01 1.02104557e+00 2.81708568e-01
-1.83697239e-01 6.56005144e-01 5.91556907e-01 -8.23078811e-01
-1.11769140e+00 -1.12336433e+00 -3.53274733e-01 4.94036973e-01
-3.09192091e-01 -3.63786161e-01 -2.16585115e-01 -5.75393498e-01
3.27526748e-01 5.77847302e-01 -8.51045191e-01 -7.10808814e-01
-4.41874206e-01 -1.29539835e+00 1.69266641e-01 7.15645194e-01
-1.94150075e-01 -1.01188099e+00 -6.71060920e-01 5.96462846e-01
-1.96515143e-01 -5.52626252e-01 -4.56410795e-01 3.07954907e-01
-1.12976444e+00 -1.11609256e+00 -8.00785065e-01 -8.31456542e-01
7.39980638e-01 -2.35360399e-01 1.12129796e+00 -4.28580530e-02
-5.09427369e-01 1.42043799e-01 -3.63786936e-01 -3.58174026e-01
-2.95072079e-01 3.35319154e-02 4.55173962e-02 4.66962636e-01
6.85975730e-01 -5.72886884e-01 -1.20154417e+00 -5.26074879e-02
-6.82466269e-01 -1.18665345e-01 3.89550030e-01 7.32755244e-01
9.31281865e-01 2.40723893e-01 7.86760092e-01 -1.11110318e+00
5.40147126e-01 -1.19575620e+00 -2.37423897e-01 2.21473441e-01
-7.82401383e-01 -2.47614030e-02 1.05855596e+00 -4.58685905e-01
-5.91288447e-01 2.36330733e-01 1.73076034e-01 -6.27050221e-01
-4.19684440e-01 7.26819754e-01 -4.78413329e-02 6.99346125e-01
3.30485553e-01 2.66084045e-01 8.62016007e-02 -5.20885408e-01
1.09140188e-01 8.26636851e-01 3.52847278e-01 -2.96029955e-01
1.00769415e-01 6.88625097e-01 9.65484008e-02 -3.79245251e-01
-9.13498163e-01 -7.76994765e-01 -4.66426879e-01 1.39984757e-01
1.05330920e+00 -9.58969533e-01 -9.15439248e-01 8.83734301e-02
-5.50069451e-01 -1.46536633e-01 -4.19411957e-01 6.06720030e-01
-6.47842109e-01 2.57906504e-02 -5.81734180e-01 -3.33389878e-01
-4.35144424e-01 -9.83284473e-01 1.04022884e+00 1.15582407e-01
-6.72584295e-01 -1.31288183e+00 4.67037708e-01 9.47139338e-02
9.71323103e-02 7.03022778e-01 1.59793258e+00 -9.12528992e-01
-1.26210019e-01 -3.52363139e-01 -6.39128983e-02 -5.26807420e-02
5.44129908e-01 -1.47057891e-01 -4.54631031e-01 -4.27948236e-01
-1.32411465e-01 1.60411865e-01 8.31483066e-01 1.02846742e+00
2.00892782e+00 -3.50639790e-01 -9.34288442e-01 1.06053317e+00
1.17692089e+00 5.70538700e-01 5.35932660e-01 -1.28024323e-02
7.88667500e-01 6.92437649e-01 2.61985689e-01 9.75663543e-01
6.61881208e-01 3.75578105e-01 3.90186965e-01 -1.54400259e-01
2.73384392e-01 -6.47797361e-02 -1.80388808e-01 6.37471676e-01
3.26271027e-01 -1.68840691e-01 -1.22711384e+00 7.98383296e-01
-2.15587211e+00 -1.05255163e+00 -2.10100546e-01 2.33480644e+00
7.38817811e-01 -1.89223826e-01 3.41120064e-01 -8.96569341e-02
8.82861912e-01 -1.33556962e-01 -9.57370818e-01 -3.39313522e-02
1.36646060e-02 4.58539054e-02 4.61368978e-01 -1.29905313e-01
-1.26423764e+00 4.91695940e-01 6.07108784e+00 2.32748300e-01
-1.48459399e+00 -5.48060164e-02 1.47003329e+00 -2.95301646e-01
-1.32140160e-01 -3.29514593e-01 -3.83094370e-01 9.79764342e-01
1.24335575e+00 -5.50277114e-01 1.76023841e-02 4.53877360e-01
7.12310314e-01 4.87805247e-01 -1.33153284e+00 9.69323635e-01
-4.66923684e-01 -1.50323689e+00 2.04955954e-02 2.69142576e-02
7.55869269e-01 8.27922747e-02 2.61037827e-01 6.03707135e-02
6.39074802e-01 -1.10060322e+00 2.05321033e-02 7.90196598e-01
8.22638988e-01 -8.57510746e-01 6.48619711e-01 7.08397999e-02
-9.76895690e-01 -6.73897386e-01 -1.93578601e-01 2.17118248e-01
1.50184348e-01 8.80205333e-01 -9.19006109e-01 4.08694357e-01
6.49301112e-01 1.26901972e+00 -4.69897896e-01 1.29552639e+00
5.42290688e-01 8.85421097e-01 -9.12564322e-02 2.08563074e-01
2.22010493e-01 -2.63858736e-01 4.33496125e-02 8.83620441e-01
4.74531502e-01 2.27795392e-01 5.19272745e-01 5.61142743e-01
-5.71551435e-02 1.61104366e-01 -3.37855846e-01 -2.64016449e-01
4.41503853e-01 1.10389018e+00 -7.45167851e-01 -3.49542707e-01
-3.29873800e-01 6.88267589e-01 2.18319133e-01 1.49879649e-01
-7.31300831e-01 -7.40264207e-02 9.19017971e-01 3.76886696e-01
1.26222730e-01 1.75669357e-01 -5.86137176e-01 -1.07787859e+00
-2.37208694e-01 -6.14552796e-01 9.97818410e-01 -7.08194375e-02
-1.79396939e+00 5.43124199e-01 -4.50984776e-01 -1.68913472e+00
-3.35891545e-01 -1.54923812e-01 -9.91350830e-01 6.89288497e-01
-1.09095156e+00 -8.85658503e-01 -1.99073911e-01 7.65986323e-01
1.50572672e-01 -2.67970115e-01 9.17926133e-01 4.41186458e-01
-8.43474329e-01 5.17274261e-01 5.55190206e-01 2.24925682e-01
7.71372080e-01 -1.19875181e+00 1.44226566e-01 1.46869272e-01
-3.10143054e-01 4.96179998e-01 3.46097469e-01 -7.31782973e-01
-1.11050761e+00 -1.81430054e+00 9.02849793e-01 -3.51762384e-01
7.49533474e-01 -3.66325788e-02 -1.08559287e+00 5.37130713e-01
-4.55594987e-01 1.39823869e-01 1.37289107e+00 2.47204870e-01
-4.93914708e-02 -3.71094823e-01 -1.13290524e+00 5.25535166e-01
6.53668582e-01 -3.10997605e-01 -2.87305385e-01 7.46847510e-01
7.27482617e-01 1.03167677e-02 -1.30916524e+00 5.58797896e-01
2.43017733e-01 -6.20495498e-01 9.25528646e-01 -1.29238796e+00
4.51593637e-01 -1.23460948e-01 1.30674914e-01 -1.47657514e+00
-7.21028268e-01 -7.23209858e-01 -1.68278649e-01 8.31226408e-01
3.89341027e-01 -3.99406523e-01 9.77691293e-01 7.92490065e-01
-9.41753313e-02 -9.50852454e-01 -8.46686959e-01 -3.97975087e-01
1.60797521e-01 -3.53053510e-02 9.13219273e-01 1.31875384e+00
2.53184199e-01 -1.60076830e-03 -1.96329117e-01 2.53277957e-01
5.39402902e-01 5.35276175e-01 2.82705754e-01 -1.56336045e+00
-1.35607734e-01 -7.22989440e-01 -5.15010774e-01 -2.72965252e-01
1.00095876e-01 -1.03980029e+00 -3.32258046e-01 -1.56900668e+00
4.41069692e-01 -8.10760319e-01 -9.14522588e-01 3.59011948e-01
-4.62168396e-01 -3.13632071e-01 -2.70520598e-01 3.62022161e-01
-7.93960094e-01 4.70340490e-01 7.57119536e-01 -1.71724662e-01
-6.57334268e-01 4.61170763e-01 -7.59697318e-01 5.74945509e-01
9.47576344e-01 -5.41019440e-01 -3.31812859e-01 -1.58923417e-01
-1.18246175e-01 5.40908992e-01 1.87050968e-01 -1.09537637e+00
2.91776687e-01 -2.91259676e-01 5.17898202e-01 -5.53689241e-01
-8.34543556e-02 -9.33018744e-01 3.59436125e-01 7.81579316e-01
-6.96076095e-01 3.38894188e-01 -2.40946729e-02 8.73650730e-01
-3.32342744e-01 6.64837241e-01 7.60470033e-01 1.26689211e-01
-3.18004698e-01 8.20581198e-01 -3.23337227e-01 7.25859553e-02
1.49598098e+00 2.38516346e-01 -2.87519842e-02 -4.26454872e-01
-1.04858279e+00 7.01824367e-01 2.70582408e-01 5.06214917e-01
5.07513523e-01 -1.41743004e+00 -8.91203523e-01 6.12695590e-02
4.32146639e-01 1.32815661e-02 5.30968308e-01 9.43243206e-01
-3.27424079e-01 3.17761093e-01 -3.41096222e-02 -7.63795078e-01
-1.20229387e+00 1.08704603e+00 2.20422000e-01 -4.00439978e-01
-9.19620574e-01 4.10859823e-01 2.56987303e-01 -1.45679355e-01
3.81344557e-01 -3.36492449e-01 -2.32546180e-01 1.29060566e-01
6.18594229e-01 2.35303462e-01 1.03005432e-01 -3.97233874e-01
-3.32358360e-01 2.95530319e-01 -3.40215027e-01 5.56002676e-01
1.73973215e+00 9.78537370e-03 -1.46512523e-01 5.45132816e-01
1.43903577e+00 -5.47625482e-01 -1.24683988e+00 -1.87357619e-01
3.42317700e-01 -1.16074093e-01 -4.73108813e-02 -5.62427521e-01
-1.30022454e+00 8.24623585e-01 7.52187252e-01 2.29288802e-01
1.40521288e+00 3.73608887e-01 9.97307003e-01 1.40359923e-01
6.97495341e-02 -7.53690779e-01 -2.19345883e-01 1.24277011e-01
3.48464012e-01 -1.29014707e+00 -2.52414018e-01 1.85864232e-02
-6.93495274e-01 7.82122731e-01 2.53663450e-01 -2.22463697e-01
1.03061974e+00 6.90909103e-02 2.28662014e-01 -2.34464929e-01
-1.08621907e+00 1.05321646e-01 2.01111168e-01 5.57761014e-01
5.47379971e-01 6.53516591e-01 -1.14805669e-01 8.66790414e-01
-6.16975985e-02 -7.17341304e-02 1.72553390e-01 5.27498722e-01
-7.86562562e-02 -1.10772157e+00 -2.59884924e-01 1.16454089e+00
-5.74022353e-01 -6.45555183e-03 -2.67874688e-01 1.52831703e-01
4.65319194e-02 7.22465515e-01 4.38210934e-01 -3.61436278e-01
1.88496843e-01 7.69665986e-02 -1.77645646e-02 -7.15571344e-01
-5.90979457e-01 6.91915900e-02 -3.85711998e-01 -6.90764785e-01
-5.23886383e-02 -7.96161652e-01 -1.54125297e+00 -2.57253766e-01
4.54962939e-01 2.91483849e-01 1.57961860e-01 7.64767349e-01
1.05743027e+00 9.04817998e-01 1.00742877e+00 -3.36322397e-01
-5.09162903e-01 -5.65683007e-01 -4.89099741e-01 8.78764749e-01
4.01854903e-01 -2.41027370e-01 2.85365433e-02 2.73498118e-01]
|
[7.835720062255859, 6.15600061416626]
|
6065b42b-2ce9-4da3-86c5-f4395d8d5775
|
launching-a-robust-backdoor-attack-under
|
2304.10985
| null |
https://arxiv.org/abs/2304.10985v1
|
https://arxiv.org/pdf/2304.10985v1.pdf
|
Launching a Robust Backdoor Attack under Capability Constrained Scenarios
|
As deep neural networks continue to be used in critical domains, concerns over their security have emerged. Deep learning models are vulnerable to backdoor attacks due to the lack of transparency. A poisoned backdoor model may perform normally in routine environments, but exhibit malicious behavior when the input contains a trigger. Current research on backdoor attacks focuses on improving the stealthiness of triggers, and most approaches require strong attacker capabilities, such as knowledge of the model structure or control over the training process. These attacks are impractical since in most cases the attacker's capabilities are limited. Additionally, the issue of model robustness has not received adequate attention. For instance, model distillation is commonly used to streamline model size as the number of parameters grows exponentially, and most of previous backdoor attacks failed after model distillation; the image augmentation operations can destroy the trigger and thus disable the backdoor. This study explores the implementation of black-box backdoor attacks within capability constraints. An attacker can carry out such attacks by acting as either an image annotator or an image provider, without involvement in the training process or knowledge of the target model's structure. Through the design of a backdoor trigger, our attack remains effective after model distillation and image augmentation, making it more threatening and practical. Our experimental results demonstrate that our method achieves a high attack success rate in black-box scenarios and evades state-of-the-art backdoor defenses.
|
['Xiaolei Liu', 'Mingyong Yin', 'Kangyi Ding', 'Yixiao Xu', 'Ming Yi']
|
2023-04-21
| null | null | null | null |
['backdoor-attack', 'image-augmentation']
|
['adversarial', 'computer-vision']
|
[ 1.81430459e-01 1.70370981e-01 -2.46117264e-01 -1.49081443e-02
-3.60329330e-01 -1.47144437e+00 7.21130371e-01 -7.64216781e-02
-6.03421509e-01 7.81372413e-02 -5.61118543e-01 -9.80882764e-01
5.62394917e-01 -6.99479878e-01 -1.06383145e+00 -5.39396226e-01
-1.65298879e-01 1.02736214e-02 3.67656201e-01 -2.20865086e-02
2.47623816e-01 8.76078546e-01 -8.42946947e-01 9.64039266e-02
3.27537626e-01 7.15213716e-01 -4.40864265e-01 8.41173053e-01
2.17211291e-01 7.97570169e-01 -1.15783107e+00 -7.27383077e-01
6.83693767e-01 -7.09210411e-02 -7.24029839e-01 -4.12751794e-01
5.75844884e-01 -9.68847632e-01 -6.89874172e-01 1.40938270e+00
2.58834571e-01 -3.09005022e-01 3.96330580e-02 -1.84260499e+00
-3.52977931e-01 6.98080659e-01 -4.89601612e-01 2.60675848e-01
-1.70424744e-01 7.05949068e-01 6.06938958e-01 -3.76748800e-01
2.93466240e-01 1.03628075e+00 4.29393560e-01 1.01495528e+00
-1.39576280e+00 -1.40136182e+00 2.86737621e-01 -2.36833706e-01
-1.26668131e+00 -5.37169993e-01 5.94502866e-01 -4.57214922e-01
9.80202556e-01 6.08457744e-01 4.74772632e-01 1.63568819e+00
3.29858065e-01 4.87762600e-01 9.49246466e-01 -2.65477002e-01
3.24551612e-01 3.88502359e-01 2.37220421e-01 7.42661655e-01
6.29459679e-01 6.49361730e-01 -3.33627850e-01 -8.57295573e-01
8.09467256e-01 -1.95383817e-01 -2.81307936e-01 -2.84046799e-01
-5.62662125e-01 9.91200984e-01 4.02136445e-01 -8.62098783e-02
2.11467296e-01 5.42182028e-01 5.76503754e-01 3.40679973e-01
-5.41772172e-02 8.27140033e-01 -4.47698087e-01 1.67706124e-02
-7.77681649e-01 2.91730583e-01 8.82623374e-01 6.12705648e-01
4.10060048e-01 2.76543587e-01 3.51829112e-01 -1.93114877e-01
3.10169756e-01 3.41272861e-01 2.23384693e-01 -7.47707725e-01
5.64684212e-01 1.88768163e-01 3.37094292e-02 -1.09782326e+00
-2.01961752e-02 -3.43678325e-01 -2.46339232e-01 6.78684890e-01
5.10983109e-01 -2.89420903e-01 -1.17904484e+00 1.90313101e+00
3.79654795e-01 4.01005387e-01 8.63332972e-02 6.60962939e-01
3.28952312e-01 5.10322511e-01 1.83476910e-01 3.05303365e-01
1.45257604e+00 -7.40651786e-01 -4.06193554e-01 -4.56717849e-01
8.76435578e-01 -4.23815399e-01 1.09692407e+00 5.19271791e-01
-7.32747674e-01 -9.01968479e-02 -1.46258235e+00 2.82294899e-01
-4.79434878e-01 -6.29636526e-01 6.02840126e-01 1.44168794e+00
-6.53870583e-01 3.57687563e-01 -1.21824515e+00 1.64093152e-01
7.99275756e-01 6.40739441e-01 -5.96656978e-01 5.79110086e-02
-1.21532619e+00 7.82692611e-01 3.86578679e-01 -4.57596481e-02
-1.67888570e+00 -8.62845182e-01 -8.11585903e-01 1.64819524e-01
5.01118720e-01 -3.72698724e-01 1.39326918e+00 -6.62908852e-01
-1.15040421e+00 7.55964398e-01 5.97398400e-01 -8.15922558e-01
8.21442902e-01 -3.67775381e-01 -3.12249124e-01 3.04953665e-01
-3.21442723e-01 4.71281916e-01 1.39043820e+00 -1.38607812e+00
-8.11521113e-02 -2.38833621e-01 7.14660823e-01 -2.46761873e-01
-9.33857024e-01 3.10808331e-01 -2.02270463e-01 -6.28778279e-01
-3.48842025e-01 -1.18804705e+00 -2.72566468e-01 1.85443938e-01
-8.51820707e-01 5.88928044e-01 1.45295143e+00 -4.02936608e-01
1.14905918e+00 -2.39324880e+00 -6.06388211e-01 3.87868315e-01
4.40996796e-01 7.27170467e-01 -2.09412593e-02 3.42051387e-01
-2.29517072e-01 1.00368404e+00 -7.81118721e-02 -2.65254229e-01
6.33376241e-02 1.65169150e-01 -9.09849823e-01 7.61210620e-01
-1.63669929e-01 7.47754812e-01 -4.66513216e-01 -2.02433258e-01
4.74577360e-02 5.19204915e-01 -7.02285528e-01 1.98098764e-01
-1.68119073e-01 1.51752025e-01 -3.99430364e-01 4.80510324e-01
6.79151118e-01 -1.23011701e-01 7.67583698e-02 6.06728578e-03
2.01276705e-01 2.33671159e-01 -9.26453471e-01 8.16934228e-01
-2.74357110e-01 5.51620185e-01 2.24909022e-01 -4.42601502e-01
5.32595456e-01 4.25115258e-01 -5.40789813e-02 -1.42834291e-01
2.88489759e-01 5.05560488e-02 1.75844491e-01 -1.64830923e-01
1.86072320e-01 -7.02598393e-02 -9.10844728e-02 7.44476318e-01
-3.43666762e-01 -1.27928242e-01 -3.92460674e-01 4.69634384e-01
1.23259914e+00 -3.31436574e-01 -2.48773117e-02 6.81551397e-02
1.37305155e-01 7.06562400e-02 3.96074295e-01 1.10894692e+00
-2.67965943e-01 1.23213053e-01 7.26460159e-01 -4.36960816e-01
-9.07560706e-01 -9.12296176e-01 1.21870965e-01 7.90550590e-01
1.18665531e-01 -5.95970929e-01 -1.19524324e+00 -1.25664115e+00
6.96362108e-02 8.63893807e-01 -7.70959496e-01 -8.79367352e-01
-6.50044560e-01 -4.00151134e-01 1.45375478e+00 6.34104729e-01
6.81421161e-01 -5.42065561e-01 -1.02137220e+00 -5.80841973e-02
2.87603885e-01 -1.25597811e+00 -6.81258440e-01 2.24951118e-01
-8.08241010e-01 -1.24855196e+00 -8.82936642e-02 -2.35431343e-01
8.71635199e-01 2.89732516e-01 5.36554575e-01 4.80653703e-01
-3.44226092e-01 1.53260827e-01 5.87443858e-02 -6.32913470e-01
-8.38570893e-01 -9.69753880e-03 1.47613168e-01 -2.20986962e-01
8.88434798e-02 -4.06827152e-01 -3.55672836e-01 3.91397417e-01
-1.22378266e+00 -3.82605702e-01 1.12381175e-01 6.37150466e-01
-4.70607076e-03 4.84073848e-01 1.40941724e-01 -1.10691988e+00
6.66985154e-01 -1.83106184e-01 -1.13539827e+00 -5.97011298e-03
-5.46992600e-01 -1.88780189e-01 8.58904481e-01 -1.04167831e+00
-6.09096289e-01 -6.22789338e-02 6.35741949e-02 -9.76052403e-01
4.10438981e-03 3.16846102e-01 -4.95051026e-01 -5.56161940e-01
9.72186506e-01 -5.24809249e-02 8.85326881e-03 -3.83542597e-01
2.11166129e-01 2.05018997e-01 3.91965061e-01 -5.26901722e-01
1.36352849e+00 6.03282273e-01 -2.92444453e-02 -5.98948359e-01
-4.37979490e-01 2.29102254e-01 -6.94894558e-03 5.61725944e-02
6.28673911e-01 -7.05875635e-01 -9.47436094e-01 7.89153516e-01
-1.25184357e+00 -4.21835840e-01 8.74802247e-02 1.09540440e-01
-4.06490592e-03 4.80568677e-01 -7.33973563e-01 -6.97065055e-01
-2.62064427e-01 -1.59600043e+00 3.70272040e-01 9.62004364e-02
-1.15033038e-01 -8.24473679e-01 -3.77063423e-01 4.41243827e-01
2.78251082e-01 3.10682297e-01 1.08762515e+00 -1.19900346e+00
-7.28708148e-01 -7.75988996e-01 2.07069814e-01 4.06280607e-01
-1.76667973e-01 5.23158461e-02 -1.39790189e+00 -7.04184532e-01
3.47383291e-01 -3.83130342e-01 5.22489488e-01 -9.28091258e-02
1.31946158e+00 -9.49331701e-01 -4.42255050e-01 7.05278575e-01
1.15981865e+00 3.76385778e-01 5.35240173e-01 6.29548073e-01
8.34761202e-01 4.42557335e-01 1.27063170e-01 2.84070939e-01
-2.88361311e-01 4.33285117e-01 1.03550136e+00 -1.42351553e-01
5.24481595e-01 -5.06089330e-01 4.97954607e-01 -3.04314733e-01
5.04006505e-01 -2.85182446e-01 -1.10801148e+00 -3.33566517e-02
-1.41135716e+00 -7.45960951e-01 1.98066562e-01 2.47583389e+00
6.66318655e-01 6.77615583e-01 1.69899464e-02 5.81641234e-02
5.21307707e-01 1.97602734e-01 -8.64359319e-01 -6.57217741e-01
3.71911079e-01 -4.09473889e-02 1.03181994e+00 5.48369229e-01
-1.19287300e+00 1.29917753e+00 6.68388319e+00 6.67729378e-01
-1.31401610e+00 1.88774452e-01 6.63558185e-01 -2.87007928e-01
-1.80321336e-01 2.46459231e-01 -8.72394621e-01 2.83996522e-01
9.07037377e-01 -6.16372377e-02 3.91893953e-01 1.13103533e+00
-6.29146546e-02 2.17051923e-01 -1.19301617e+00 6.13990366e-01
-4.42782044e-02 -1.34736550e+00 -1.08748458e-01 5.78009427e-01
1.52684599e-01 -1.69499829e-01 4.00280774e-01 2.40003914e-01
6.34927928e-01 -1.41448772e+00 9.10562098e-01 -3.04494321e-01
7.21895874e-01 -1.13433635e+00 3.02520096e-01 5.82234323e-01
-7.97184825e-01 -3.16452235e-01 -4.87844506e-03 1.26982793e-01
-1.36159599e-01 -1.29010864e-02 -9.63458896e-01 -1.76156506e-01
6.86701059e-01 -2.70174026e-01 -5.99999964e-01 5.67929447e-01
-5.44934332e-01 1.03317165e+00 -5.70453465e-01 3.13846380e-01
4.66539502e-01 4.46803510e-01 6.43228829e-01 8.77243936e-01
-1.93305612e-01 -1.36083022e-01 2.83731997e-01 1.03108215e+00
-2.01846018e-01 -5.06818771e-01 -9.54365134e-01 -5.57435870e-01
8.06150913e-01 9.59289014e-01 -7.13312685e-01 -1.37555495e-01
-1.27731562e-01 8.68428349e-01 2.66255625e-02 3.34627092e-01
-1.24866498e+00 -2.10573316e-01 1.05662119e+00 1.99359179e-01
-7.12834357e-04 -4.85617787e-01 -3.64072233e-01 -9.23518598e-01
-2.67740116e-02 -1.45457149e+00 3.86540562e-01 -3.04147899e-01
-8.29494894e-01 6.02516115e-01 1.56983554e-01 -6.14954770e-01
-5.05390130e-02 -4.80656624e-01 -7.24473536e-01 7.67979503e-01
-1.01343358e+00 -1.29055703e+00 2.96922207e-01 8.51474762e-01
1.17775284e-01 -2.47095600e-01 9.02034342e-01 3.90777066e-02
-8.37023854e-01 1.25794232e+00 -3.72452945e-01 4.84404594e-01
3.64991903e-01 -7.50566602e-01 9.07972038e-01 1.33267176e+00
2.23143756e-01 1.26240826e+00 9.81538653e-01 -8.62103283e-01
-1.56785607e+00 -8.89608502e-01 1.45878315e-01 -5.35880625e-01
8.49404216e-01 -9.07672226e-01 -9.54887450e-01 9.85928833e-01
-1.66352779e-01 3.23610097e-01 7.27368653e-01 -2.13596538e-01
-1.05866396e+00 2.28729248e-01 -1.36915922e+00 9.76157784e-01
5.87486386e-01 -9.28599000e-01 -6.90440461e-02 3.23618174e-01
9.25218165e-01 -4.13603693e-01 -3.75275165e-01 8.24103951e-02
5.97739100e-01 -6.89125001e-01 1.03605866e+00 -9.94060218e-01
-5.60325496e-02 -1.12332411e-01 -1.07705377e-01 -7.72488892e-01
1.88681364e-01 -1.06083810e+00 -6.34300113e-01 1.07385099e+00
3.24181974e-01 -9.94374812e-01 1.03177261e+00 1.18474317e+00
2.08738431e-01 -4.96886164e-01 -9.52946842e-01 -1.04933250e+00
2.93341666e-01 -5.93522549e-01 7.25610375e-01 1.15743268e+00
-2.29874894e-01 -9.89186093e-02 -4.11001414e-01 9.37056839e-01
7.59259343e-01 -5.79000056e-01 9.99451280e-01 -7.57030547e-01
-6.23014450e-01 -3.11151505e-01 -3.50027919e-01 -6.95333242e-01
1.83863342e-01 -6.16223931e-01 -3.21605116e-01 -6.53540850e-01
-1.23720672e-02 -3.23755473e-01 -1.78679466e-01 1.02030134e+00
1.50728524e-01 2.63398409e-01 7.77083874e-01 3.94978195e-01
2.58448988e-01 5.26720397e-02 8.13918293e-01 -3.00951928e-01
-6.20426945e-02 1.64302792e-02 -8.32998395e-01 9.22212481e-01
9.01784301e-01 -9.37041402e-01 -7.96723545e-01 -2.63885647e-01
1.46846548e-01 -2.97914088e-01 7.79038429e-01 -7.91466415e-01
2.63617247e-01 -2.06191480e-01 7.06500709e-02 3.09010129e-02
4.79163736e-01 -1.17562819e+00 2.06304699e-01 9.68827188e-01
-2.77054965e-01 1.39279142e-01 6.14844859e-01 5.36119699e-01
2.71042377e-01 -4.15400237e-01 8.92684162e-01 -1.09561853e-01
-3.25727493e-01 3.53283226e-01 -5.40035844e-01 -1.54966265e-01
1.29496813e+00 -4.13885802e-01 -7.81217217e-01 -1.32056803e-01
-5.61483920e-01 -1.49770156e-02 7.33448803e-01 3.85538250e-01
5.98961592e-01 -7.02833295e-01 -8.66969451e-02 4.42052394e-01
-6.05544411e-02 -1.84909388e-01 -5.71261533e-02 3.97621363e-01
-6.30777538e-01 -1.77933425e-02 1.49205513e-02 -2.67577440e-01
-1.82335305e+00 9.84991372e-01 6.57683909e-01 -2.25346684e-01
-8.35308552e-01 1.01874113e+00 6.74037516e-01 -2.26636827e-01
4.41666335e-01 8.70544389e-02 3.52215171e-01 -2.91162908e-01
7.88837254e-01 5.20826578e-02 -6.17171712e-02 -2.25917384e-01
-4.71512318e-01 -1.51389718e-01 -7.70658314e-01 -3.13598722e-01
8.58843863e-01 4.23066318e-01 1.19964473e-01 -2.02677459e-01
1.14105117e+00 6.90119490e-02 -1.44819212e+00 2.11571857e-01
-3.40648115e-01 -7.23428667e-01 3.12132150e-01 -9.48170185e-01
-1.03997719e+00 9.97775912e-01 6.06076717e-01 3.11521411e-01
7.14653790e-01 -3.11471969e-01 8.75812411e-01 5.74544609e-01
4.83573258e-01 -6.84903741e-01 3.33460867e-01 2.68882215e-01
6.06737792e-01 -8.64327133e-01 -9.62692201e-02 -4.34616148e-01
-4.56440598e-01 8.20073545e-01 9.66566384e-01 5.14525659e-02
6.49061918e-01 8.65323365e-01 2.15312913e-01 -1.65562049e-01
-4.98312533e-01 8.13211024e-01 -2.92121798e-01 5.27195275e-01
-3.83531123e-01 -9.55501199e-02 4.30643618e-01 5.28201222e-01
-2.13073641e-01 -4.17984694e-01 7.77605772e-01 1.35258460e+00
-3.38584006e-01 -1.09376907e+00 -8.48016322e-01 3.28643285e-02
-8.31855893e-01 -2.21024975e-01 -7.05515623e-01 1.07978141e+00
-1.15622684e-01 9.90232289e-01 -3.37005675e-01 -6.81443453e-01
6.09183311e-02 7.37117305e-02 2.27510348e-01 -6.87567890e-01
-8.92348886e-01 -8.57918113e-02 6.08742721e-02 -7.46806443e-01
4.84777898e-01 -2.09751785e-01 -1.21429253e+00 -6.42217159e-01
-6.91354036e-01 -3.88722941e-02 8.85014951e-01 6.95966601e-01
3.51287782e-01 1.55754521e-01 7.03558147e-01 -5.23689091e-01
-1.13977337e+00 -3.40744495e-01 -3.83023411e-01 6.99196197e-03
4.81769115e-01 -5.42437732e-01 -5.39090455e-01 -3.50323319e-02]
|
[5.749709129333496, 7.639607906341553]
|
65b14ab6-f5fd-45bc-ab08-e25a85d295c1
|
unsupervised-person-re-identification-via
|
2004.03547
| null |
https://arxiv.org/abs/2004.03547v1
|
https://arxiv.org/pdf/2004.03547v1.pdf
|
Unsupervised Person Re-identification via Softened Similarity Learning
|
Person re-identification (re-ID) is an important topic in computer vision. This paper studies the unsupervised setting of re-ID, which does not require any labeled information and thus is freely deployed to new scenarios. There are very few studies under this setting, and one of the best approach till now used iterative clustering and classification, so that unlabeled images are clustered into pseudo classes for a classifier to get trained, and the updated features are used for clustering and so on. This approach suffers two problems, namely, the difficulty of determining the number of clusters, and the hard quantization loss in clustering. In this paper, we follow the iterative training mechanism but discard clustering, since it incurs loss from hard quantization, yet its only product, image-level similarity, can be easily replaced by pairwise computation and a softened classification task. With these improvements, our approach becomes more elegant and is more robust to hyper-parameter changes. Experiments on two image-based and video-based datasets demonstrate state-of-the-art performance under the unsupervised re-ID setting.
|
['Lingxi Xie', 'Yu Wu', 'Qi Tian', 'Yutian Lin', 'Chenggang Yan']
|
2020-04-07
|
unsupervised-person-re-identification-via-2
|
http://openaccess.thecvf.com/content_CVPR_2020/html/Lin_Unsupervised_Person_Re-Identification_via_Softened_Similarity_Learning_CVPR_2020_paper.html
|
http://openaccess.thecvf.com/content_CVPR_2020/papers/Lin_Unsupervised_Person_Re-Identification_via_Softened_Similarity_Learning_CVPR_2020_paper.pdf
|
cvpr-2020-6
|
['unsupervised-person-re-identification']
|
['computer-vision']
|
[ 1.42707795e-01 -3.16245317e-01 -2.69791912e-02 -4.58216906e-01
-5.02703905e-01 -4.64808285e-01 6.07179165e-01 2.74049729e-01
-8.89812469e-01 6.11252964e-01 -1.20126814e-01 2.27019817e-01
-8.32324848e-02 -4.96584326e-01 -3.97463351e-01 -1.04189038e+00
-8.02052580e-03 7.68851459e-01 2.98582435e-01 1.91446617e-01
2.22362995e-01 3.07376146e-01 -1.92830169e+00 -1.91418439e-01
8.44647825e-01 8.03328931e-01 1.63459718e-01 4.34884429e-01
1.23920090e-01 4.37009245e-01 -4.35155690e-01 -7.37854719e-01
4.22103614e-01 -3.25238407e-01 -8.27378929e-01 4.49878186e-01
4.68660772e-01 -1.90873787e-01 -1.46801338e-01 1.34025466e+00
7.15821862e-01 5.43740511e-01 6.06839776e-01 -1.27947867e+00
-5.16194999e-01 4.33209538e-01 -7.48475671e-01 2.22301763e-02
2.44424909e-01 -2.08391890e-01 8.63215387e-01 -8.15358341e-01
4.80729431e-01 1.29713464e+00 7.83078074e-01 6.27213061e-01
-1.39811909e+00 -6.96797490e-01 4.36576083e-02 5.93930483e-01
-1.76618731e+00 -3.39156151e-01 5.73016286e-01 -5.60398400e-01
4.21653241e-01 2.46043041e-01 3.51304978e-01 7.56243944e-01
-4.53271300e-01 4.76727605e-01 1.08947301e+00 -7.25022197e-01
3.41001689e-01 4.23813671e-01 2.52793461e-01 5.01911283e-01
2.61892378e-01 -9.46625769e-02 -1.39544606e-01 -1.08998671e-01
4.59361792e-01 2.86068082e-01 -2.66778439e-01 -5.94898880e-01
-1.02699387e+00 7.83339798e-01 2.85818636e-01 3.93849641e-01
-1.00281186e-01 -1.67416573e-01 3.66924316e-01 2.59063751e-01
3.56934041e-01 1.10901400e-01 -1.17398731e-01 -1.29748225e-01
-1.14390779e+00 2.79973000e-01 6.36548936e-01 7.66704023e-01
1.00253737e+00 -6.10133708e-01 -6.42075688e-02 1.16051877e+00
3.33334543e-02 4.26954240e-01 7.39892185e-01 -9.43315148e-01
2.55729675e-01 4.36376721e-01 8.65165517e-02 -1.11751127e+00
-4.22707349e-01 -2.72923291e-01 -1.00665951e+00 1.74218208e-01
5.69762647e-01 -1.19744223e-02 -9.17412996e-01 1.76314771e+00
3.26736182e-01 3.71273160e-01 -9.63996500e-02 9.88987267e-01
3.62486780e-01 2.18139231e-01 4.07930836e-02 -2.20241651e-01
1.38796890e+00 -9.24859941e-01 -5.69431484e-01 -1.24579594e-02
5.39489329e-01 -6.40530467e-01 6.44101083e-01 4.71683860e-01
-8.55728686e-01 -7.84682810e-01 -9.81838942e-01 7.82784745e-02
-3.96691054e-01 9.52327717e-03 3.86723846e-01 8.91656578e-01
-1.21202767e+00 6.89163566e-01 -6.15964055e-01 -7.86914945e-01
2.12605163e-01 6.54244900e-01 -5.23702502e-01 -3.15976948e-01
-1.06238234e+00 7.21001148e-01 5.43462276e-01 4.99969954e-03
-3.59543473e-01 -2.96226948e-01 -6.28214777e-01 5.32874092e-02
4.68338370e-01 -5.40139437e-01 9.63923573e-01 -1.13620067e+00
-1.41068292e+00 9.97490942e-01 -2.46600226e-01 -4.45744783e-01
5.58820844e-01 9.57810283e-02 -3.63439977e-01 1.63981512e-01
2.48112887e-01 6.41910613e-01 1.01970434e+00 -1.32902074e+00
-8.75095367e-01 -6.72692895e-01 -4.74825874e-02 3.54376823e-01
-6.58776104e-01 -7.45266154e-02 -9.71979082e-01 -4.99061018e-01
1.36699900e-01 -1.34031701e+00 -3.21206421e-01 -2.69249082e-01
-2.79510558e-01 -4.69989538e-01 5.74790478e-01 -4.57863092e-01
1.13513482e+00 -2.31147838e+00 2.19646528e-01 4.77541208e-01
3.36420894e-01 3.52251321e-01 1.04174964e-01 2.38562778e-01
-2.28547677e-01 1.53162822e-01 -3.68025154e-01 -7.57163107e-01
-5.53952605e-02 -7.45597901e-03 1.67208076e-01 5.62808275e-01
-2.53362060e-01 5.14665186e-01 -8.44728947e-01 -7.15606272e-01
3.81191760e-01 3.23207229e-01 -5.52612245e-01 1.01409435e-01
3.79979730e-01 4.92950112e-01 -9.04201791e-02 3.25963616e-01
9.42058444e-01 -1.77883238e-01 2.56011933e-01 -3.01430672e-01
-8.88085738e-02 -3.34762067e-01 -1.49946284e+00 1.55951071e+00
-1.35010302e-01 3.04563642e-01 -1.00089587e-01 -1.40900743e+00
6.43215656e-01 1.10036694e-01 7.12758780e-01 -4.43765849e-01
8.37126449e-02 4.50143889e-02 -1.25385419e-01 -3.31009686e-01
4.14176583e-01 2.94760428e-03 2.32009850e-02 4.74213988e-01
-5.11149429e-02 4.89666700e-01 3.75443578e-01 1.87218532e-01
6.86436832e-01 -1.18720241e-01 1.24244511e-01 -2.79031813e-01
7.23278463e-01 -1.81344137e-01 6.56436563e-01 8.13489974e-01
-2.97653198e-01 7.24717736e-01 9.14181303e-03 -1.39657989e-01
-1.02292681e+00 -8.97520125e-01 -3.22792023e-01 9.15680647e-01
4.29929167e-01 -4.36875165e-01 -1.12923563e+00 -6.31888568e-01
1.73272453e-02 3.29139173e-01 -5.03556907e-01 -1.22881785e-01
-4.66514587e-01 -9.69541132e-01 3.18692952e-01 2.96198696e-01
5.85987151e-01 -6.55601740e-01 -2.49979898e-01 2.37458989e-01
-3.58636022e-01 -1.15851176e+00 -6.52952671e-01 -5.50748296e-02
-7.87102580e-01 -1.07288563e+00 -1.12430513e+00 -9.78854597e-01
8.68663251e-01 6.40204549e-01 7.08282173e-01 2.69905597e-01
-3.58130604e-01 5.59951246e-01 -4.40803856e-01 -1.19650558e-01
-8.53978768e-02 -2.69029103e-02 4.60016668e-01 5.07278502e-01
6.13123596e-01 -4.59913284e-01 -5.97944200e-01 5.73580801e-01
-6.82518065e-01 -3.58216256e-01 3.36309075e-01 9.55589831e-01
6.39144778e-01 4.96000916e-01 4.02906656e-01 -8.73380244e-01
3.13133210e-01 -1.48754731e-01 -3.98268580e-01 2.60539114e-01
-8.00656915e-01 -4.07025777e-02 6.67313755e-01 -3.69878560e-01
-1.01187992e+00 2.14466348e-01 1.55467480e-01 -3.74038428e-01
-2.80383706e-01 1.69324905e-01 -3.32562238e-01 -2.26144791e-01
4.75535929e-01 1.64378524e-01 5.63041754e-02 -6.52928174e-01
4.19134080e-01 1.01958060e+00 7.36137390e-01 -3.13663274e-01
9.48903143e-01 7.36803174e-01 -2.10155174e-01 -9.93929684e-01
-5.38453400e-01 -9.18943524e-01 -1.08855653e+00 -6.58627823e-02
1.03603268e+00 -9.46348965e-01 -7.43026197e-01 6.46906376e-01
-8.41865063e-01 8.68276134e-02 -1.94368958e-01 5.75487018e-01
-3.91474932e-01 1.01540768e+00 -3.52045000e-01 -7.73912132e-01
-9.94940475e-02 -1.13585520e+00 7.42148221e-01 3.12935829e-01
1.52796507e-03 -8.74055624e-01 -2.39124689e-02 4.48920429e-01
9.91712138e-02 -1.13509409e-01 6.55312121e-01 -6.36812449e-01
-2.98919529e-01 -3.01761806e-01 -3.06665421e-01 4.52276617e-01
2.56211340e-01 -3.83744031e-01 -1.10215068e+00 -6.42673492e-01
-4.82773967e-02 -2.35023677e-01 9.00357664e-01 1.91780031e-01
1.34189856e+00 -1.23917833e-01 -5.24734318e-01 7.06145108e-01
1.30572438e+00 1.23296872e-01 5.71196318e-01 4.16543812e-01
9.30938363e-01 7.86363244e-01 6.31504118e-01 4.85300779e-01
6.26175582e-01 1.06597888e+00 1.24781363e-01 -1.61826909e-01
-7.59023726e-02 9.22163762e-03 1.25038460e-01 8.19372833e-01
-3.54965150e-01 -2.27094442e-01 -7.59114146e-01 3.42781484e-01
-2.04443765e+00 -1.19846880e+00 4.17516232e-02 2.62794280e+00
6.23010933e-01 -1.11069649e-01 3.95154446e-01 4.59457010e-01
1.19824898e+00 -4.00382578e-01 -4.38356161e-01 1.15379572e-01
2.70620622e-02 -7.15952367e-02 6.93700433e-01 3.45442444e-01
-1.44489849e+00 8.24979007e-01 5.67581081e+00 8.37741375e-01
-9.37053978e-01 1.33404121e-01 6.43551171e-01 2.08746210e-01
2.96324104e-01 -6.00183643e-02 -8.64683628e-01 7.43418038e-01
8.49858046e-01 -8.91812518e-02 5.96276343e-01 7.02568054e-01
8.07369649e-02 -2.34985575e-01 -1.26004350e+00 1.57788074e+00
4.06563073e-01 -6.41853273e-01 -1.33797288e-01 9.81395841e-02
5.61272085e-01 -2.39850447e-01 -3.76464101e-03 9.29609314e-02
2.66818851e-01 -7.24087894e-01 6.53704584e-01 3.13732147e-01
8.61355245e-01 -9.22211707e-01 9.14942503e-01 2.39435315e-01
-1.18623590e+00 -3.18075120e-01 -6.02304101e-01 2.29651034e-01
-6.11713976e-02 5.85816264e-01 -5.89508891e-01 6.51495457e-01
1.12231112e+00 8.88648987e-01 -8.31038415e-01 1.22118664e+00
2.04641461e-01 4.06916499e-01 -3.15549582e-01 2.22335786e-01
-1.35092318e-01 -3.80487502e-01 2.88308740e-01 1.19294059e+00
2.90399671e-01 6.22910541e-03 6.03811502e-01 2.17982471e-01
3.56839225e-02 1.07895039e-01 -2.54742950e-01 4.50180620e-01
5.36004305e-01 1.20466745e+00 -9.47557092e-01 -5.32827497e-01
-3.28353554e-01 1.52417898e+00 3.94262910e-01 2.53423005e-01
-6.94024563e-01 -3.80121112e-01 4.38991427e-01 1.43212840e-01
4.78178173e-01 -1.12214759e-01 1.99310295e-02 -1.29266489e+00
-3.12194265e-02 -7.34345973e-01 6.88283980e-01 -1.97390199e-01
-1.61216021e+00 4.08086151e-01 1.98383674e-01 -1.38951087e+00
-3.48785281e-01 -4.05500203e-01 -1.81918353e-01 6.57305777e-01
-1.47213233e+00 -9.71938550e-01 -4.38305467e-01 8.82904708e-01
2.86615878e-01 -1.09756522e-01 7.47649312e-01 8.34321737e-01
-7.24247396e-01 1.02072382e+00 5.34760177e-01 2.81912655e-01
1.12083375e+00 -1.37832642e+00 -2.73981709e-02 8.20161343e-01
7.89652020e-02 5.97389936e-01 5.51575124e-01 -4.09940004e-01
-1.00902891e+00 -1.11686456e+00 9.57643151e-01 -4.35961276e-01
3.67304057e-01 -5.27985334e-01 -8.39059353e-01 3.46330136e-01
-1.16824187e-01 -6.09501116e-02 8.94652605e-01 6.76068366e-02
-3.36277813e-01 -3.53764862e-01 -1.26468325e+00 4.87508506e-01
1.21129155e+00 -4.17608768e-01 -3.34366053e-01 2.87191927e-01
3.36685121e-01 6.23122156e-02 -9.42426920e-01 8.06836039e-03
4.09122318e-01 -1.01806200e+00 1.09780633e+00 -1.12967692e-01
-3.46340716e-01 -5.81039727e-01 -4.64371666e-02 -1.26883698e+00
-6.66746140e-01 -4.70971167e-01 3.80060077e-01 1.65277410e+00
1.49628833e-01 -6.56053782e-01 7.26721942e-01 6.29736841e-01
2.17490226e-01 -9.69657227e-02 -1.06142652e+00 -9.11317945e-01
-1.81594312e-01 -8.26568604e-02 5.04745960e-01 1.01510620e+00
-1.93641055e-02 2.67264158e-01 -4.81829911e-01 1.54205978e-01
1.09128392e+00 7.24863913e-03 8.45836639e-01 -1.44657409e+00
-2.88797021e-01 -2.46139839e-01 -7.90963650e-01 -9.62374926e-01
4.71877120e-02 -9.69059467e-01 1.58497915e-01 -1.13198900e+00
5.72231352e-01 -7.35207736e-01 -2.85428971e-01 3.38075787e-01
-3.73363703e-01 5.38948178e-01 4.20016468e-01 7.25923896e-01
-7.62643158e-01 3.64000291e-01 7.62396753e-01 -2.64624029e-01
-2.62941241e-01 1.04458211e-03 -5.74619830e-01 6.83112204e-01
6.11648321e-01 -5.10565996e-01 -4.33532983e-01 -2.55896688e-01
-3.92543465e-01 -3.98491174e-01 4.72946435e-01 -1.28354192e+00
4.27148372e-01 2.46071458e-01 4.53505635e-01 -3.63755733e-01
2.50580370e-01 -9.72893000e-01 1.25743970e-01 3.57511580e-01
-2.26471812e-01 3.11282836e-02 -3.01587552e-01 7.52322137e-01
-1.62280366e-01 -4.43176121e-01 1.06346655e+00 -1.70983061e-01
-7.65441000e-01 5.18129110e-01 -1.73406467e-01 -9.73264873e-02
1.17731631e+00 -5.23733079e-01 1.92945123e-01 -2.88378090e-01
-8.39352906e-01 2.75380820e-01 7.07942665e-01 3.48701060e-01
3.23458523e-01 -1.34404778e+00 -6.88202202e-01 1.88034832e-01
1.34680539e-01 -8.09890181e-02 2.54839689e-01 7.54772902e-01
-1.53976753e-01 2.15985760e-01 -1.29836621e-02 -8.80592108e-01
-1.37633514e+00 8.67494822e-01 8.79898369e-02 -2.91482687e-01
-7.75564075e-01 6.18127584e-01 1.88345656e-01 -3.23007345e-01
4.83319461e-01 2.48787060e-01 -4.51992184e-01 2.97181100e-01
7.46821880e-01 4.55186844e-01 1.65504396e-01 -9.63362992e-01
-3.66269082e-01 9.01290238e-01 -3.06840926e-01 -1.03849061e-01
9.85306442e-01 -5.29758990e-01 -1.28911525e-01 2.89734781e-01
1.26025987e+00 -2.48822957e-01 -1.14215589e+00 -3.65265429e-01
2.03453004e-01 -5.05378246e-01 -1.77131906e-01 -4.09343779e-01
-1.01210046e+00 8.16102087e-01 1.16048300e+00 1.58301294e-01
1.19197953e+00 -1.30502924e-01 8.70469511e-01 4.25151765e-01
4.91371334e-01 -1.38107777e+00 -4.27124910e-02 2.53538460e-01
2.86663383e-01 -1.36983812e+00 2.79159900e-02 -4.16721553e-01
-5.45132756e-01 8.20410013e-01 4.65469003e-01 5.09153353e-04
7.72508979e-01 -1.13175660e-01 7.97991604e-02 2.16691300e-01
-5.66524044e-02 -3.72513205e-01 8.75156298e-02 8.86118293e-01
8.04523379e-02 5.62940054e-02 -2.49902740e-01 4.81613368e-01
-1.18648067e-01 -4.56466861e-02 3.81515056e-01 6.86193943e-01
-3.69251668e-01 -1.36523342e+00 -6.02209747e-01 4.01632130e-01
-4.48582113e-01 3.86950523e-02 -2.01772317e-01 6.09733045e-01
5.42040586e-01 1.15254843e+00 1.26171842e-01 -5.07386148e-01
2.52060831e-01 -8.40817764e-02 4.40914333e-01 -5.33080041e-01
-4.14500237e-01 -4.53785323e-02 -2.82583684e-01 -4.58687186e-01
-8.68640900e-01 -8.82691562e-01 -1.15405130e+00 -2.62187630e-01
-5.62845170e-01 4.38367158e-01 5.75253665e-01 7.47997165e-01
2.73159832e-01 3.32386531e-02 9.17156279e-01 -1.05512154e+00
-5.91649890e-01 -7.30219662e-01 -7.23495603e-01 1.03641546e+00
2.51528859e-01 -8.77086639e-01 -4.62327003e-01 3.32659930e-01]
|
[14.7271728515625, 1.0541502237319946]
|
f3570c63-3d1f-47e5-a314-d815fdeac7e3
|
sequential-attention-based-network-for-noetic
|
1901.02609
| null |
https://arxiv.org/abs/1901.02609v3
|
https://arxiv.org/pdf/1901.02609v3.pdf
|
Sequential Attention-based Network for Noetic End-to-End Response Selection
|
The noetic end-to-end response selection challenge as one track in Dialog System Technology Challenges 7 (DSTC7) aims to push the state of the art of utterance classification for real world goal-oriented dialog systems, for which participants need to select the correct next utterances from a set of candidates for the multi-turn context. This paper describes our systems that are ranked the top on both datasets under this challenge, one focused and small (Advising) and the other more diverse and large (Ubuntu). Previous state-of-the-art models use hierarchy-based (utterance-level and token-level) neural networks to explicitly model the interactions among different turns' utterances for context modeling. In this paper, we investigate a sequential matching model based only on chain sequence for multi-turn response selection. Our results demonstrate that the potentials of sequential matching approaches have not yet been fully exploited in the past for multi-turn response selection. In addition to ranking the top in the challenge, the proposed model outperforms all previous models, including state-of-the-art hierarchy-based models, and achieves new state-of-the-art performances on two large-scale public multi-turn response selection benchmark datasets.
|
['Qian Chen', 'Wen Wang']
|
2019-01-09
| null | null | null | null |
['goal-oriented-dialog', 'conversational-response-selection']
|
['natural-language-processing', 'natural-language-processing']
|
[-2.30821818e-02 1.36453927e-01 -1.39338508e-01 -9.50130045e-01
-1.06467414e+00 -5.23971677e-01 7.61572838e-01 1.37314409e-01
-4.95478421e-01 7.44529605e-01 6.39318407e-01 -3.81358385e-01
6.00833148e-02 -2.29888678e-01 2.97147512e-01 -1.54409572e-01
2.42098376e-01 1.18609202e+00 5.36173642e-01 -1.32999504e+00
3.87340754e-01 1.32551258e-02 -1.37870049e+00 1.12428796e+00
5.76529026e-01 8.70633245e-01 2.03015938e-01 1.01416159e+00
-4.35748398e-01 1.09244525e+00 -6.34771883e-01 -3.76234889e-01
-1.60789430e-01 -5.76526880e-01 -1.60312212e+00 -2.79725254e-01
4.08992797e-01 -4.63510960e-01 -1.91715837e-01 3.65357995e-01
8.66373062e-01 5.75205624e-01 3.16439569e-01 -1.05520606e+00
-2.46699050e-01 9.56709921e-01 3.81000161e-01 -3.09239347e-02
1.02006066e+00 2.46887177e-01 1.32349014e+00 -9.83189225e-01
7.50765383e-01 1.81639862e+00 4.52366471e-01 1.14205933e+00
-1.18273067e+00 -1.68293849e-01 5.36850393e-01 3.30952406e-01
-9.05719578e-01 -7.80217290e-01 6.08226359e-01 -2.49010444e-01
1.60368955e+00 7.48664141e-01 9.44145722e-04 1.12841702e+00
-2.05147127e-03 1.09660196e+00 1.13177586e+00 -5.62309146e-01
1.53854951e-01 2.81084776e-01 6.91812158e-01 4.94978160e-01
-1.20811641e+00 -9.98087898e-02 -7.33619392e-01 -4.53933328e-01
-5.15472814e-02 -4.26387668e-01 -1.40597671e-01 1.62491828e-01
-1.04119706e+00 1.10349035e+00 1.83453143e-01 4.12338525e-01
-1.24704674e-01 -4.65762645e-01 6.67197466e-01 7.53587842e-01
5.97296655e-01 5.35966635e-01 -7.52536356e-01 -2.95195609e-01
-4.14087176e-01 8.03290129e-01 1.40481579e+00 9.87737179e-01
4.38186854e-01 -2.99226284e-01 -9.06590343e-01 1.31349134e+00
2.53079116e-01 -7.64355361e-02 5.42007387e-01 -9.02263641e-01
7.49393702e-01 7.76492417e-01 2.78261364e-01 -3.06925982e-01
-8.21620166e-01 1.83248103e-01 -3.41579050e-01 -2.42723331e-01
6.62175238e-01 -2.72517323e-01 -4.69551295e-01 1.65584362e+00
4.34990346e-01 -3.05779636e-01 1.87943548e-01 1.00829434e+00
1.47292411e+00 3.49483728e-01 2.81356871e-01 -2.64915675e-01
1.54332066e+00 -1.35428393e+00 -6.33011699e-01 -1.83500469e-01
8.85119438e-01 -8.57451200e-01 1.32227695e+00 3.04826766e-01
-1.08482826e+00 -5.51107943e-01 -7.54636586e-01 -2.44835883e-01
-4.94491994e-01 -6.12039790e-02 6.13031089e-01 5.93183815e-01
-1.23728108e+00 1.69096112e-01 -5.47133796e-02 -5.41011393e-01
-6.71096921e-01 5.52452266e-01 -2.07238957e-01 1.90949798e-01
-1.79691422e+00 1.29077041e+00 1.57511588e-02 -1.28682747e-01
-7.05963135e-01 -4.28888619e-01 -6.71318412e-01 3.58037911e-02
6.24102354e-01 -3.97528529e-01 2.11987257e+00 -3.96572143e-01
-2.26715446e+00 7.55004883e-01 -3.59585404e-01 -6.22934043e-01
3.54340345e-01 -2.55998582e-01 -3.27070981e-01 1.17983140e-01
-2.06373036e-01 7.36020148e-01 3.29389095e-01 -1.00256860e+00
-1.01715887e+00 -1.88581631e-01 4.75768000e-01 6.75131857e-01
-1.66784190e-02 5.65599918e-01 -1.29729912e-01 2.72256862e-02
-3.18683945e-02 -1.07983696e+00 -3.93925190e-01 -9.64487314e-01
-3.13544840e-01 -1.12205803e+00 7.16197073e-01 -3.72206599e-01
1.44092977e+00 -1.55054653e+00 2.42772669e-01 -4.55043256e-01
1.72659814e-01 2.27267057e-01 1.09048542e-02 1.06520891e+00
3.21055204e-01 8.97740871e-02 2.29171813e-01 -9.87147450e-01
3.25928152e-01 -1.00184239e-01 -3.38060498e-01 -7.14552701e-02
-2.55170256e-01 6.60543025e-01 -8.90721440e-01 -3.43512505e-01
4.69322413e-01 -9.84583795e-02 -4.12733704e-01 8.06341469e-01
-6.47479951e-01 6.14682138e-01 -4.54842806e-01 3.17448139e-01
1.78861842e-01 1.45760238e-01 5.46648979e-01 4.22741994e-02
-1.63948759e-01 1.09784508e+00 -7.08509624e-01 1.63276148e+00
-6.82037354e-01 2.58614510e-01 2.69301742e-01 -6.21781647e-01
9.50776875e-01 6.99226320e-01 2.00352266e-01 -6.73769534e-01
-1.11904452e-02 2.81005442e-01 1.67836145e-01 -5.09111166e-01
1.04496229e+00 -9.95891169e-03 -7.69827664e-01 4.68798131e-01
2.14749441e-01 -3.10165405e-01 4.61296178e-02 4.66920257e-01
9.33963418e-01 -1.01518780e-01 3.31340253e-01 -2.70771384e-01
1.04437029e+00 -1.03054293e-01 2.46767521e-01 1.13693035e+00
-3.66646409e-01 4.74115551e-01 3.23492527e-01 -5.16382396e-01
-4.12078112e-01 -4.26071167e-01 2.87009120e-01 2.22036004e+00
2.98136715e-02 -3.50610048e-01 -9.06265736e-01 -1.02801239e+00
-2.41217136e-01 9.25920486e-01 -2.87718266e-01 1.53226987e-01
-7.90117502e-01 -3.57741088e-01 7.20772505e-01 1.35687828e-01
3.95122260e-01 -1.32916009e+00 -4.28706765e-01 5.41344702e-01
-6.29545867e-01 -1.33569932e+00 -6.45340025e-01 1.87283128e-01
-2.83307910e-01 -8.62762451e-01 -2.31615439e-01 -7.27868736e-01
-2.06451908e-01 1.24693573e-01 1.53326714e+00 2.53507048e-01
1.73449352e-01 5.08000493e-01 -8.39193940e-01 -1.52496189e-01
-6.52035713e-01 4.34952766e-01 1.68274596e-01 1.17582930e-02
6.34464145e-01 2.26566028e-02 -3.77103746e-01 7.99540043e-01
-2.28134945e-01 -4.33409819e-03 -8.98161530e-02 1.09607208e+00
-1.73068643e-01 -6.48684382e-01 1.13756227e+00 -9.82302487e-01
1.26248550e+00 -2.89310127e-01 -1.26631320e-01 5.74168503e-01
-5.19409657e-01 -1.37214348e-01 6.45155191e-01 -2.48619467e-01
-1.48501432e+00 3.77352200e-02 -5.94582319e-01 2.94621110e-01
-4.97549564e-01 4.08325285e-01 -5.70850298e-02 6.28133863e-03
6.37560368e-01 3.90263237e-02 -2.45009392e-01 -6.29202247e-01
4.91945833e-01 1.08750582e+00 2.01935947e-01 -7.77278423e-01
-2.88932532e-01 -1.82083741e-01 -5.08632898e-01 -8.21539342e-01
-6.91016018e-01 -9.57461298e-01 -6.83734298e-01 -3.48686039e-01
1.02514327e+00 -6.31782770e-01 -1.21838343e+00 4.75329906e-01
-1.27911079e+00 -6.49654031e-01 2.75679976e-01 1.38415799e-01
-5.42739272e-01 2.22368494e-01 -1.10217881e+00 -1.39313769e+00
-7.29262292e-01 -1.46922386e+00 9.19445634e-01 2.26942465e-01
-5.62849283e-01 -9.41501617e-01 4.96564172e-02 7.87332952e-01
6.34051681e-01 -4.98845488e-01 8.69439781e-01 -1.51577306e+00
-2.24155292e-01 -1.54291824e-01 3.09193820e-01 -6.58276752e-02
-2.17919618e-01 -4.36564505e-01 -1.04096365e+00 -1.80465207e-01
3.15060526e-01 -9.45071399e-01 7.91140079e-01 1.07425846e-01
5.07770538e-01 -2.40162954e-01 -2.01120913e-01 -2.65058905e-01
7.19198465e-01 5.59247375e-01 2.91410238e-01 1.02206215e-01
2.86336988e-01 1.23845756e+00 1.04187405e+00 3.49725753e-01
1.11519098e+00 1.33093989e+00 3.64415735e-01 3.18808079e-01
1.87660262e-01 -7.82872513e-02 3.19558084e-01 7.12633312e-01
4.73492652e-01 -5.68852663e-01 -9.03071165e-01 5.44659853e-01
-2.23358631e+00 -7.84739435e-01 -2.25744873e-01 2.17661691e+00
1.01231313e+00 3.22354913e-01 5.80439687e-01 -3.43164325e-01
5.34331024e-01 2.52670914e-01 -1.70738652e-01 -9.26186204e-01
8.46158564e-02 1.02115475e-01 -2.61769861e-01 1.20757115e+00
-1.19300711e+00 1.43883610e+00 5.88736343e+00 6.51764512e-01
-8.96181405e-01 1.58125609e-01 8.03053498e-01 1.31780252e-01
1.57364290e-02 1.73176184e-01 -1.48085356e+00 1.91004902e-01
1.25749791e+00 1.24697641e-01 3.33684415e-01 8.34972262e-01
1.63938344e-01 -1.52959377e-01 -1.46053886e+00 4.12936360e-01
7.26801455e-02 -1.17587972e+00 -2.99176693e-01 -3.79560649e-01
1.44817650e-01 -6.85114488e-02 -1.85161456e-01 1.13337135e+00
6.50618732e-01 -1.07908916e+00 3.92840028e-01 2.78528154e-01
3.80088091e-01 -3.75491023e-01 9.88525867e-01 6.58954918e-01
-1.05559421e+00 -2.27838546e-01 -1.91783890e-01 -2.00116068e-01
4.76960033e-01 -1.16678432e-01 -1.37078667e+00 6.55473232e-01
4.20220912e-01 -9.96290520e-03 -3.18184346e-01 4.25123662e-01
1.04901634e-01 4.82129931e-01 -2.08392799e-01 -6.41107142e-01
5.20445228e-01 8.81342292e-02 6.27746999e-01 1.59016526e+00
-3.21157813e-01 5.17508090e-01 8.79695654e-01 2.50081658e-01
-6.18395619e-02 2.58333147e-01 -1.52745262e-01 4.08874571e-01
8.67084384e-01 1.17465758e+00 -1.94564104e-01 -3.93725276e-01
-2.41506338e-01 6.96444869e-01 3.09066474e-01 4.14685570e-02
-3.22157085e-01 -6.77069053e-02 5.01379669e-01 -3.49408120e-01
-1.51255965e-01 8.66483077e-02 -1.81682661e-01 -8.37343872e-01
-2.67970294e-01 -1.31898296e+00 5.89828968e-01 -1.59423545e-01
-1.37458694e+00 1.17317319e+00 1.89394787e-01 -9.62159336e-01
-9.07262266e-01 -4.55375999e-01 -7.20042884e-01 1.05126834e+00
-9.64343071e-01 -1.36749589e+00 2.42940597e-02 5.59648871e-01
1.47514379e+00 -4.24077898e-01 1.38990247e+00 1.18923523e-01
-3.27594221e-01 8.25872302e-01 -3.68764579e-01 -3.18341404e-02
1.11445093e+00 -1.57423317e+00 4.91572559e-01 9.26696062e-02
-2.39304692e-01 7.21096337e-01 8.96170199e-01 -4.05630201e-01
-1.13560617e+00 -3.13170403e-01 1.45125079e+00 -7.74291456e-01
4.68811512e-01 -5.93185484e-01 -8.71457279e-01 4.00480747e-01
5.80850005e-01 -6.11694694e-01 6.86014533e-01 5.93139052e-01
-1.62240908e-01 1.67843506e-01 -1.15363514e+00 6.60529971e-01
6.91335082e-01 -6.35311246e-01 -7.55726993e-01 5.29688597e-01
1.09470451e+00 -8.24114561e-01 -7.63753474e-01 3.76591265e-01
6.83786511e-01 -1.22853768e+00 9.60547805e-01 -9.67668831e-01
2.56205499e-01 2.91169584e-01 -3.56565982e-01 -1.25216317e+00
1.04204953e-01 -9.43026006e-01 -2.00869050e-02 1.36857831e+00
5.91175139e-01 -3.54198813e-01 7.92837381e-01 1.02720809e+00
-5.32688141e-01 -1.04764438e+00 -1.07832897e+00 -2.55262792e-01
2.27094844e-01 -2.25331262e-01 5.99375784e-01 6.83844507e-01
6.78305268e-01 1.17723823e+00 -7.54807711e-01 -4.16460752e-01
4.72990004e-03 2.26590008e-01 9.04867768e-01 -1.09919047e+00
-2.70254910e-01 -6.68799102e-01 1.56826794e-01 -1.76958156e+00
3.99368137e-01 -5.49620748e-01 4.01271582e-01 -1.41538537e+00
-3.04918408e-01 -4.58550572e-01 4.76092845e-02 2.86942750e-01
-4.72827971e-01 -4.69455242e-01 5.71573675e-01 7.38414899e-02
-9.62060690e-01 5.96406817e-01 6.19036853e-01 -8.49519223e-02
-6.32621408e-01 5.60866654e-01 -5.36157548e-01 5.05431116e-01
6.45597398e-01 -1.74316838e-01 -5.11977732e-01 -1.30078435e-01
-3.30014020e-01 8.93993616e-01 -1.70382336e-01 -5.60235322e-01
6.05770588e-01 -2.77836591e-01 -5.15976846e-01 -7.40003228e-01
8.30794454e-01 -4.61404771e-01 -5.82192361e-01 1.37252748e-01
-1.21073234e+00 -1.43659078e-02 -3.94606628e-02 3.39108944e-01
-1.63917899e-01 -4.58492339e-01 5.58436692e-01 -4.29716885e-01
-8.84394050e-01 -1.27108231e-01 -7.26491511e-01 5.24485335e-02
5.35525262e-01 2.72793680e-01 -7.25878119e-01 -9.67232108e-01
-9.34658468e-01 7.06917644e-01 -1.83447957e-01 8.40284944e-01
7.06423819e-01 -8.14290583e-01 -9.58993912e-01 -3.81019235e-01
2.34736785e-01 -4.54965293e-01 4.85275507e-01 7.61534095e-01
4.75073978e-03 9.17630374e-01 1.26784787e-01 -5.43907344e-01
-1.70365918e+00 2.24601433e-01 5.24670541e-01 -1.01406920e+00
-2.83309489e-01 1.26478660e+00 -6.81429207e-02 -1.25654745e+00
5.40785432e-01 6.12919740e-02 -7.44454563e-01 1.23453185e-01
6.63608193e-01 1.79569215e-01 3.06788117e-01 -6.64546669e-01
-4.19936657e-01 -2.71423399e-01 -5.46528041e-01 -5.35520136e-01
7.52786875e-01 -4.99331236e-01 -1.18163899e-01 5.87041378e-01
8.97899210e-01 -2.84078240e-01 -6.46419644e-01 -6.06995225e-01
4.49176699e-01 -1.53890625e-01 -3.81539941e-01 -1.29089499e+00
-1.69026062e-01 7.81217635e-01 4.94727492e-01 6.85072660e-01
6.72636747e-01 -2.12355986e-01 9.50137615e-01 8.64361227e-01
7.53068805e-01 -1.29577076e+00 2.03111887e-01 1.31643510e+00
1.11641192e+00 -1.58154321e+00 -2.98617333e-01 -4.91283804e-01
-1.33542275e+00 9.28313613e-01 1.18787873e+00 2.78178513e-01
4.59785521e-01 -1.26767039e-01 3.59468639e-01 -8.96165296e-02
-1.48500812e+00 -2.67798126e-01 2.46072918e-01 4.09076363e-01
9.59236324e-01 2.26073325e-01 -5.86747050e-01 7.21832156e-01
-2.13600099e-01 -5.41224003e-01 3.89475971e-01 7.95098424e-01
-5.52822411e-01 -1.54492652e+00 -6.59881160e-02 4.31396395e-01
-4.40327495e-01 -2.80395776e-01 -1.08850932e+00 4.97421414e-01
-4.98612255e-01 1.80534089e+00 -5.37499249e-01 -9.88937974e-01
6.09104395e-01 6.76806092e-01 -7.32091218e-02 -9.86256242e-01
-1.66988492e+00 -3.93373594e-02 9.93064404e-01 -4.52124894e-01
-2.28928834e-01 -4.09528613e-01 -9.97000933e-01 -2.23983213e-01
-3.90560567e-01 4.82015282e-01 4.26538974e-01 1.17919469e+00
2.34161139e-01 2.27287158e-01 8.22132885e-01 -1.05420732e+00
-1.16482973e+00 -1.49648929e+00 -4.15006168e-02 4.10658777e-01
1.58501193e-01 -6.13574982e-01 -1.90418184e-01 -4.35714960e-01]
|
[12.73274040222168, 7.878838539123535]
|
828eb89c-7534-44ac-a1ad-e9192c10d4c8
|
prompt-based-metric-learning-for-few-shot-ner
|
2211.04337
| null |
https://arxiv.org/abs/2211.04337v1
|
https://arxiv.org/pdf/2211.04337v1.pdf
|
Prompt-Based Metric Learning for Few-Shot NER
|
Few-shot named entity recognition (NER) targets generalizing to unseen labels and/or domains with few labeled examples. Existing metric learning methods compute token-level similarities between query and support sets, but are not able to fully incorporate label semantics into modeling. To address this issue, we propose a simple method to largely improve metric learning for NER: 1) multiple prompt schemas are designed to enhance label semantics; 2) we propose a novel architecture to effectively combine multiple prompt-based representations. Empirically, our method achieves new state-of-the-art (SOTA) results under 16 of the 18 considered settings, substantially outperforming the previous SOTA by an average of 8.84% and a maximum of 34.51% in relative gains of micro F1. Our code is available at https://github.com/AChen-qaq/ProML.
|
['Zhilin Yang', 'Yanan Zheng', 'Yanru Chen']
|
2022-11-08
| null | null | null | null |
['few-shot-ner']
|
['natural-language-processing']
|
[-5.76757342e-02 1.80967874e-03 -4.26849037e-01 -7.18440890e-01
-1.28344560e+00 -7.78088927e-01 7.11555004e-01 4.16423231e-01
-9.01458085e-01 7.71285176e-01 2.03860939e-01 -1.87538993e-02
-1.01455897e-01 -7.74847031e-01 -3.77636522e-01 -2.66323209e-01
1.74589634e-01 5.75796664e-01 3.14375192e-01 -2.00592816e-01
9.81944352e-02 3.03789139e-01 -1.33881056e+00 1.82750747e-01
8.34203541e-01 8.61256123e-01 -6.46021869e-03 5.33891976e-01
-5.34966886e-01 1.04444265e+00 -5.60247481e-01 -6.24359369e-01
2.15963647e-02 -3.22814703e-01 -9.98481035e-01 -4.78852361e-01
4.91559535e-01 -1.35466203e-01 -3.41677219e-01 9.48918700e-01
7.36013830e-01 5.36704957e-01 7.07599461e-01 -1.03263724e+00
-9.60960746e-01 7.83957779e-01 -2.61690944e-01 3.26985687e-01
1.81940883e-01 -2.49967590e-01 1.36555457e+00 -1.14608872e+00
6.52299047e-01 1.04411077e+00 5.39870739e-01 9.84755456e-01
-1.18781507e+00 -8.04769874e-01 -2.25141030e-02 3.58944297e-01
-1.52772677e+00 -5.82372189e-01 3.65252197e-01 -3.55189331e-02
1.15255237e+00 1.27378196e-01 -2.10432976e-01 1.19281507e+00
-5.65849602e-01 7.94168234e-01 1.12470579e+00 -5.49647152e-01
4.21102732e-01 1.55831963e-01 4.52110827e-01 5.62338412e-01
1.83690488e-01 2.93021705e-02 -3.86937439e-01 -1.69024557e-01
1.99919552e-01 -6.91624582e-02 6.76009282e-02 -1.44606262e-01
-1.17133558e+00 8.43707085e-01 4.46166903e-01 5.67876577e-01
-3.18774790e-01 1.78174809e-01 4.46943432e-01 3.43992293e-01
5.60570419e-01 8.56169462e-01 -7.39220440e-01 -3.57051671e-01
-9.01209176e-01 3.74884531e-02 8.34324718e-01 1.12751627e+00
9.30459857e-01 -8.50756168e-02 -3.70162457e-01 1.13754320e+00
-4.86705415e-02 5.91838658e-01 5.73484540e-01 -9.17470038e-01
4.39652681e-01 4.83193070e-01 4.05348122e-01 -4.08087075e-01
-5.55630863e-01 -3.44477773e-01 -3.24588418e-01 -4.78589207e-01
3.18089575e-01 -3.33685428e-01 -1.11735237e+00 1.92230725e+00
3.75398934e-01 4.63180214e-01 1.51988745e-01 6.91173792e-01
9.30863678e-01 4.87600088e-01 7.01861262e-01 -3.69078889e-02
1.44482088e+00 -8.85791242e-01 -7.14202225e-01 -2.36145273e-01
1.29791486e+00 -6.56950891e-01 1.24732339e+00 -4.93331775e-02
-6.80090964e-01 -3.25152606e-01 -8.61438394e-01 -2.04050094e-01
-7.30226099e-01 1.95382640e-01 4.74220246e-01 8.13113034e-01
-9.55697775e-01 5.91891170e-01 -6.85130417e-01 -7.87583649e-01
2.74394780e-01 7.67857358e-02 -3.28873515e-01 -3.78068715e-01
-1.46705651e+00 1.08549964e+00 5.10239124e-01 -3.98045838e-01
-9.06362474e-01 -9.01753545e-01 -9.09597516e-01 1.50656343e-01
5.16323924e-01 -4.33579028e-01 1.51039338e+00 -3.50663573e-01
-1.24215770e+00 1.09648120e+00 -2.33801574e-01 -3.31099659e-01
1.94271147e-01 -4.40668941e-01 -7.02555180e-01 -3.11393030e-02
3.12799156e-01 8.62932265e-01 -1.22841641e-01 -9.77092862e-01
-7.09121764e-01 -2.02392325e-01 1.46539018e-01 1.33460194e-01
-8.61982048e-01 4.49146420e-01 -8.13777223e-02 -6.13356769e-01
-1.73497424e-01 -6.35300338e-01 -2.87997663e-01 -3.60143125e-01
-3.32321048e-01 -6.58249974e-01 3.57936621e-01 -3.51024747e-01
1.27373874e+00 -1.89399207e+00 -2.29045421e-01 -2.08923250e-01
4.94161882e-02 5.16751468e-01 -4.78163928e-01 5.01899421e-01
5.80556989e-02 4.23445046e-01 -2.03377113e-01 -4.37354714e-01
3.57080787e-01 1.67366594e-01 -2.33615622e-01 1.84307620e-01
2.71198094e-01 9.33084249e-01 -1.29053128e+00 -4.40386266e-01
1.57812923e-01 2.80468613e-01 -1.01316042e-01 3.25169474e-01
-9.96691436e-02 8.36350471e-02 -5.36813736e-01 8.03951681e-01
4.91853923e-01 -2.76004910e-01 2.79960811e-01 -1.28353080e-02
-2.80239303e-02 6.83239698e-01 -1.24424958e+00 1.93465889e+00
-4.31934983e-01 2.62738436e-01 -4.40050423e-01 -8.45164120e-01
1.02328014e+00 4.51571047e-01 5.21940112e-01 -8.75814259e-01
1.83639020e-01 3.42960745e-01 -4.14413959e-01 -2.62503386e-01
4.77672428e-01 -3.47920448e-01 -3.43138754e-01 5.34357965e-01
5.19245982e-01 3.05662513e-01 3.68040293e-01 1.25138596e-01
1.45729256e+00 -3.25596593e-02 4.88448352e-01 -2.27447212e-01
2.74621367e-01 5.97097613e-02 9.37790155e-01 7.81366408e-01
-5.24848461e-01 3.88283283e-01 1.65775225e-01 -2.35701740e-01
-9.19205785e-01 -1.18381441e+00 -1.13222145e-01 1.69261229e+00
6.82104677e-02 -5.10695398e-01 -7.00602472e-01 -9.82929826e-01
-1.63847849e-01 1.21475720e+00 -6.47023678e-01 -1.78859219e-01
-4.67327297e-01 -4.63233054e-01 1.05952573e+00 8.08167756e-01
3.57855499e-01 -1.05294442e+00 -3.21724862e-01 2.12315992e-01
-2.06774592e-01 -1.29584682e+00 -4.53009158e-01 4.05115187e-01
-7.91459203e-01 -9.61578369e-01 -6.09169960e-01 -7.54495621e-01
9.95502397e-02 3.00860167e-01 1.20647347e+00 -2.09584966e-01
-2.07557634e-01 3.31042618e-01 -7.13701963e-01 -3.11172068e-01
-2.13909060e-01 3.86601865e-01 2.18512386e-01 -3.00899923e-01
1.09473848e+00 -3.89434129e-01 -3.17725509e-01 4.63094562e-01
-7.88364053e-01 -4.46927160e-01 4.31574106e-01 8.83234203e-01
5.18117845e-01 -4.27625299e-01 8.58060479e-01 -1.06796706e+00
4.98783082e-01 -8.44925880e-01 -2.27531523e-01 6.14431322e-01
-8.61556530e-01 1.24185212e-01 4.99900699e-01 -5.74669838e-01
-1.14238548e+00 -1.38414770e-01 -1.32195637e-01 -3.90598148e-01
-6.46886945e-01 3.64621401e-01 -1.31814674e-01 2.09139675e-01
1.08601797e+00 -1.10779084e-01 -7.31892943e-01 -7.58428752e-01
7.79551446e-01 8.55792820e-01 3.33665699e-01 -8.55243742e-01
5.99002719e-01 2.25235403e-01 -4.14286077e-01 -4.26958650e-01
-1.33258426e+00 -1.01197827e+00 -6.82973504e-01 9.79902148e-02
7.54846036e-01 -9.26319778e-01 -5.26512377e-02 8.52587353e-03
-9.17457223e-01 -3.40409935e-01 -4.12453026e-01 6.66134894e-01
-4.10102218e-01 1.97722837e-01 -8.52571964e-01 -6.05406344e-01
-5.14428139e-01 -5.49185455e-01 8.47298324e-01 4.01744962e-01
-2.10347131e-01 -1.01469731e+00 4.41168040e-01 3.13692212e-01
5.64483464e-01 1.03602238e-01 6.55336440e-01 -1.31862867e+00
7.89513439e-02 -2.12125093e-01 -3.42464387e-01 2.43715316e-01
2.08721504e-01 -4.42196637e-01 -1.25366151e+00 -2.48667970e-01
-2.10226193e-01 -7.33001888e-01 8.23408782e-01 -4.09751711e-03
6.74473047e-01 3.62135959e-03 -3.59834284e-01 3.15282315e-01
1.42343354e+00 -7.72576928e-02 3.82792026e-01 4.14847106e-01
6.46902025e-01 6.67053282e-01 9.95628059e-01 5.96850216e-01
5.06521881e-01 6.90876424e-01 1.66005686e-01 1.16760910e-01
-2.98201174e-01 -4.45061266e-01 3.93167526e-01 7.54735649e-01
2.20073000e-01 -3.53507072e-01 -1.15095949e+00 9.74401832e-01
-1.81032920e+00 -8.49327862e-01 5.25523238e-02 2.00419760e+00
9.73296702e-01 -7.14113340e-02 7.02100322e-02 -2.23747492e-01
1.02182090e+00 8.84537250e-02 -7.54523993e-01 -2.67132461e-01
-3.55120003e-02 5.08204103e-01 6.12217665e-01 2.38933802e-01
-1.29962635e+00 1.39330876e+00 5.71804905e+00 1.00869358e+00
-7.24607706e-01 5.65627396e-01 3.61491084e-01 -1.05663784e-01
-3.93252134e-01 8.27955008e-02 -9.74583745e-01 2.01673701e-01
1.55234826e+00 -3.60566497e-01 2.05485523e-01 9.95333493e-01
-1.78696692e-01 4.71241832e-01 -1.05505586e+00 7.48972833e-01
5.13071241e-03 -9.94372487e-01 -3.07946969e-02 -1.46298215e-01
7.42259383e-01 5.15890896e-01 -2.49609113e-01 7.90035605e-01
6.79188251e-01 -8.07729840e-01 6.08623505e-01 3.93404931e-01
1.03227472e+00 -7.02472746e-01 8.04353297e-01 3.99066150e-01
-1.21903455e+00 -1.64804533e-01 -6.62375152e-01 5.88740669e-02
1.26076564e-01 5.06408215e-01 -8.80663633e-01 5.34278631e-01
6.58090234e-01 5.85720241e-01 -6.09321356e-01 1.07537103e+00
-5.56523204e-01 9.88917232e-01 -3.26007396e-01 -2.18779758e-01
3.99263114e-01 2.79885262e-01 2.94668525e-01 1.60435474e+00
3.23998511e-01 1.73235819e-01 2.00579002e-01 7.00011432e-01
-5.23769140e-01 4.95308429e-01 -5.44338703e-01 -2.46537879e-01
1.10063982e+00 1.45117354e+00 -6.10539079e-01 -3.92160863e-01
-5.22903204e-01 9.46435273e-01 7.79560149e-01 3.23872954e-01
-9.64626074e-01 -7.28229105e-01 7.59246230e-01 -2.19819635e-01
3.19215298e-01 -4.75111380e-02 -1.54297441e-01 -1.34409702e+00
-1.39789701e-01 -4.59037691e-01 9.59396422e-01 -4.92720068e-01
-1.68628204e+00 5.22640169e-01 -2.12037981e-01 -9.77123201e-01
-2.29716882e-01 -5.85261881e-01 -4.16057259e-01 5.06073833e-01
-1.79451787e+00 -1.24612212e+00 -2.04735056e-01 5.39335132e-01
4.14036751e-01 -4.39607725e-02 1.32157505e+00 6.38306975e-01
-4.23384070e-01 9.69771087e-01 2.31922925e-01 4.21030462e-01
1.19311512e+00 -1.56485176e+00 4.91233736e-01 8.52559268e-01
6.52243257e-01 5.03787220e-01 4.32368785e-01 -4.03666556e-01
-8.78451526e-01 -1.17996013e+00 1.54978895e+00 -8.05037677e-01
7.48305261e-01 -4.08187777e-01 -8.19733620e-01 6.79835260e-01
6.48592934e-02 1.72396094e-01 1.14235079e+00 5.80267072e-01
-1.04036880e+00 -4.33401428e-02 -1.38466847e+00 3.13775331e-01
1.17197978e+00 -6.88984156e-01 -7.98905015e-01 4.05930251e-01
9.25318420e-01 -2.67367065e-02 -1.19712329e+00 4.33669716e-01
1.68936789e-01 -5.57466090e-01 8.33284140e-01 -9.15710270e-01
-3.60421613e-02 -2.40068600e-01 -6.17094338e-01 -1.18836927e+00
-5.18510461e-01 -2.41081506e-01 -1.88859761e-01 1.58319223e+00
5.72938204e-01 -4.13658679e-01 5.79797745e-01 7.42481947e-01
-1.64074704e-01 -5.50242245e-01 -9.33770120e-01 -1.07717645e+00
3.21756244e-01 -4.75323260e-01 6.29773498e-01 1.49229813e+00
1.78228915e-01 4.58841830e-01 -3.32878560e-01 1.21144667e-01
5.55851400e-01 -1.88648537e-01 2.15418279e-01 -1.22990322e+00
1.27679080e-01 -3.31785053e-01 -3.63424063e-01 -7.99583673e-01
3.65566820e-01 -1.05416822e+00 1.36672676e-01 -1.60565090e+00
4.09099221e-01 -5.84545135e-01 -1.13330090e+00 9.39006984e-01
-2.77684659e-01 2.75310934e-01 1.07707389e-01 1.65729627e-01
-1.13175070e+00 5.41434765e-01 5.41712582e-01 3.87172401e-02
3.33253853e-02 -2.85134763e-01 -8.46033931e-01 3.64964902e-01
1.14279735e+00 -9.14290786e-01 -3.40442985e-01 -5.93354523e-01
5.93304820e-02 -2.98714578e-01 -1.19509418e-02 -1.04264152e+00
3.66014928e-01 -2.07157031e-01 4.64040339e-02 -9.75654349e-02
1.33254647e-01 -3.85852128e-01 -2.73926526e-01 1.21788971e-01
-6.89665020e-01 -4.35853004e-02 7.67933354e-02 5.99397957e-01
-2.06277862e-01 -4.84527171e-01 7.86557853e-01 -2.03579087e-02
-1.10516977e+00 3.02058399e-01 -2.35584881e-02 6.65572941e-01
8.35891306e-01 2.48693049e-01 -5.57428300e-01 -1.20084882e-01
-5.42971432e-01 3.46577287e-01 3.84399623e-01 5.99684060e-01
3.72549444e-01 -1.50591600e+00 -8.63536239e-01 -3.68852526e-01
6.02528334e-01 -4.91015851e-01 2.59828389e-01 5.59565544e-01
-1.47067621e-01 5.42072594e-01 -2.99313348e-02 -1.79916978e-01
-9.45293128e-01 5.27624309e-01 2.66772479e-01 -4.30962324e-01
-3.15514952e-01 1.01609468e+00 -1.84265092e-01 -9.49906826e-01
2.27677703e-01 -7.04132169e-02 -2.74636328e-01 2.45400071e-01
6.74952984e-01 6.50301039e-01 1.00692242e-01 -5.10942519e-01
-5.28119564e-01 2.41533443e-01 -1.61527976e-01 -9.51377675e-02
1.34601569e+00 -7.11370111e-02 2.01727793e-01 5.51849723e-01
1.23016286e+00 -1.96115553e-01 -8.37738216e-01 -6.76411092e-01
7.87983060e-01 -2.02311620e-01 1.22194579e-02 -1.10215950e+00
-6.05461121e-01 7.90038586e-01 6.83940113e-01 -3.95496786e-02
9.48364735e-01 2.13836923e-01 9.45299864e-01 8.17532361e-01
4.77564365e-01 -1.21712232e+00 -2.12608930e-02 6.52774096e-01
2.22285256e-01 -1.35237134e+00 -3.53593618e-01 -7.74105638e-02
-7.64364660e-01 8.59290004e-01 7.06912875e-01 9.29926038e-02
5.67513466e-01 1.42092155e-02 2.52529591e-01 -2.02674165e-01
-8.50253761e-01 -6.66094959e-01 1.49821267e-01 5.19875944e-01
7.27362216e-01 2.17812687e-01 -3.72783512e-01 8.44288528e-01
1.32440746e-01 -1.48990467e-01 2.80873954e-01 8.79130006e-01
-7.55584121e-01 -1.23544061e+00 1.83207348e-01 3.35133284e-01
-5.46143651e-01 -3.44056666e-01 -3.80326420e-01 5.88172376e-01
-9.94905159e-02 1.23743451e+00 -2.23528579e-01 -4.71162289e-01
6.55992806e-01 4.96207893e-01 1.41986400e-01 -1.00193512e+00
-4.67152029e-01 -3.31248969e-01 3.97963971e-01 -7.00698495e-01
-5.11143982e-01 -5.93211055e-01 -1.37040520e+00 -2.37003475e-01
-4.86454844e-01 3.67291540e-01 5.10765731e-01 8.30947816e-01
5.03054976e-01 1.54046580e-01 6.20602012e-01 -1.27365112e-01
-7.99269438e-01 -1.30179822e+00 -6.79216206e-01 5.94836175e-01
-1.96130440e-01 -8.02374899e-01 -3.93520921e-01 -2.21806481e-01]
|
[9.676888465881348, 9.337662696838379]
|
7db2f3a0-1179-4a28-8786-e5f164d290fd
|
synthcity-facilitating-innovative-use-cases
|
2301.07573
| null |
https://arxiv.org/abs/2301.07573v1
|
https://arxiv.org/pdf/2301.07573v1.pdf
|
Synthcity: facilitating innovative use cases of synthetic data in different data modalities
|
Synthcity is an open-source software package for innovative use cases of synthetic data in ML fairness, privacy and augmentation across diverse tabular data modalities, including static data, regular and irregular time series, data with censoring, multi-source data, composite data, and more. Synthcity provides the practitioners with a single access point to cutting edge research and tools in synthetic data. It also offers the community a playground for rapid experimentation and prototyping, a one-stop-shop for SOTA benchmarks, and an opportunity for extending research impact. The library can be accessed on GitHub (https://github.com/vanderschaarlab/synthcity) and pip (https://pypi.org/project/synthcity/). We warmly invite the community to join the development effort by providing feedback, reporting bugs, and contributing code.
|
['Mihaela van der Schaar', 'Bogdan-Constantin Cebere', 'Zhaozhi Qian']
|
2023-01-18
| null | null | null | null |
['irregular-time-series']
|
['time-series']
|
[-5.91613233e-01 -1.38720855e-01 -4.59259540e-01 -4.98637706e-01
-1.16904724e+00 -8.01379204e-01 4.64801580e-01 2.87378460e-01
1.81559417e-02 7.78997481e-01 3.91730666e-01 -4.54115659e-01
4.54929955e-02 -6.32448196e-01 -4.36315298e-01 -3.53828818e-01
-3.88089031e-01 4.76651698e-01 -2.12431848e-01 -2.20084473e-01
9.76005271e-02 1.77387580e-01 -1.37953150e+00 2.49358058e-01
6.40440345e-01 6.41730547e-01 -3.66652697e-01 4.84189391e-01
7.98362866e-02 5.75771689e-01 -1.28251061e-01 -5.69512188e-01
4.31090474e-01 -1.44270420e-01 -4.50473905e-01 -3.20834309e-01
4.42987606e-02 -3.12633738e-02 -8.63051489e-02 7.94900596e-01
5.89841008e-01 -1.17157169e-01 1.06428713e-01 -1.69336879e+00
-5.12879193e-01 4.53574836e-01 -7.11794496e-01 2.70027220e-01
5.72475195e-01 5.79329848e-01 9.19211447e-01 -9.36619699e-01
7.75758088e-01 1.12378156e+00 8.55923653e-01 3.78448904e-01
-1.29224634e+00 -9.45284605e-01 -2.24691391e-01 -1.03523798e-01
-1.19776917e+00 -7.88866997e-01 4.43441123e-01 -5.55833280e-01
5.25386035e-01 7.46573627e-01 4.47772026e-01 1.67422807e+00
-2.99398482e-01 9.96438503e-01 1.15499079e+00 1.42497778e-01
2.63818145e-01 3.22496235e-01 4.05860096e-01 3.87345284e-01
6.87702298e-02 2.02410057e-01 -3.29429924e-01 -7.52525806e-01
5.18810511e-01 3.72606039e-01 -1.43000662e-01 -2.57555813e-01
-1.27796018e+00 7.24517345e-01 -2.04147384e-01 1.24950469e-01
-2.92745858e-01 -1.22912347e-01 6.79279983e-01 5.15528023e-01
6.53750718e-01 2.96239793e-01 -7.36925125e-01 -8.39780688e-01
-6.99470043e-01 6.71248436e-01 1.06006837e+00 9.83856201e-01
3.19962591e-01 6.35896698e-02 -4.55719605e-02 8.93254042e-01
6.17282577e-02 5.86269021e-01 2.35239431e-01 -1.16919219e+00
2.97584325e-01 4.02705729e-01 3.84715915e-01 -7.94184327e-01
-5.41706443e-01 -3.27599496e-01 -8.36459339e-01 -1.75306216e-01
6.52360201e-01 -4.80852157e-01 -4.95903939e-01 1.72144961e+00
4.91025656e-01 -1.72082335e-01 -9.84213501e-02 8.82721364e-01
7.62218237e-01 5.62231600e-01 -2.92525798e-01 -1.01502284e-01
1.29953241e+00 -6.48399353e-01 -7.00068057e-01 1.97631270e-01
6.18991077e-01 -7.76718318e-01 1.13560808e+00 4.61763054e-01
-1.09430802e+00 -7.36211017e-02 -5.89841247e-01 6.53439686e-02
-5.97812057e-01 -2.21140891e-01 9.99840260e-01 3.93062055e-01
-8.11342120e-01 5.25155246e-01 -1.23532259e+00 -8.10198009e-01
3.89937192e-01 -1.33602777e-02 -4.86986041e-01 -1.73208177e-01
-1.19216049e+00 4.59150851e-01 3.54085155e-02 -2.53999650e-01
-6.34303391e-01 -1.02870131e+00 -7.97556937e-01 -3.03158641e-01
4.43362892e-01 -5.01319170e-01 1.03826988e+00 -5.27198076e-01
-8.90875995e-01 6.65443897e-01 1.17065839e-01 -3.47495645e-01
8.07044506e-01 -1.43683985e-01 -6.36383414e-01 -4.67398793e-01
2.35798344e-01 5.79629093e-02 2.03771368e-01 -7.42760122e-01
-1.00608163e-01 -5.39884627e-01 -5.10942876e-01 -2.20830098e-01
5.28380014e-02 3.28112006e-01 -2.90678382e-01 -7.84350872e-01
-2.84048885e-01 -8.81856918e-01 -7.42725208e-02 -3.56449410e-02
-5.60228229e-01 -5.49594760e-02 7.95255542e-01 -6.73597038e-01
1.17214692e+00 -2.37525988e+00 -2.88189769e-01 9.16676223e-02
3.63537256e-04 1.16869435e-01 3.16558741e-02 1.24511278e+00
-3.03960353e-01 3.56031895e-01 -1.51039839e-01 -5.00655234e-01
2.80507296e-01 -2.21152887e-01 -3.15835983e-01 5.74033022e-01
4.83683497e-02 7.42186308e-01 -9.53545392e-01 7.12601691e-02
1.77619472e-01 2.44848624e-01 -4.57141370e-01 8.77942331e-03
-2.15280056e-01 6.33925200e-01 -3.16765964e-01 9.14678037e-01
8.73156846e-01 -4.97009277e-01 6.21305825e-03 4.90407407e-01
-4.90894347e-01 3.46509248e-01 -1.40505683e+00 1.44811845e+00
-3.09403479e-01 5.19480109e-01 3.00726503e-01 -4.23379242e-01
8.61653984e-01 3.25125098e-01 6.31980121e-01 -4.32592839e-01
7.61651667e-03 2.64417052e-01 -3.66312563e-01 -4.37220305e-01
3.99424911e-01 1.65874168e-01 -2.36967608e-01 8.15575838e-01
-3.42144042e-01 8.04027393e-02 2.82896519e-01 2.30004042e-01
1.12432134e+00 1.53954789e-01 1.75786987e-01 -1.49339512e-01
-5.72355092e-02 -3.15382588e-03 5.64189136e-01 4.80455160e-01
-6.59124330e-02 6.55865192e-01 7.87252069e-01 -5.48156381e-01
-1.29706109e+00 -1.16594732e+00 -5.00627637e-01 1.03779185e+00
-5.13817489e-01 -6.74954474e-01 -3.50481093e-01 -6.60380721e-02
3.66262048e-01 7.18810618e-01 -6.68785095e-01 1.68328583e-01
-9.75297168e-02 -7.75207579e-01 6.80451930e-01 2.95487851e-01
3.12764227e-01 -9.46837544e-01 -8.65760297e-02 -6.13547117e-02
-3.38853121e-01 -6.35873616e-01 -5.96805573e-01 1.10855721e-01
-5.88940084e-01 -9.01312888e-01 -4.29282218e-01 -2.40289420e-01
1.67142078e-01 5.28384931e-02 9.30519700e-01 -3.37915117e-04
-4.86263812e-01 3.29540282e-01 -3.53391469e-01 -6.05700672e-01
-3.07952702e-01 1.21953107e-01 1.84606045e-01 -1.67183116e-01
2.84037381e-01 -7.57640064e-01 -6.72158301e-01 5.23097277e-01
-9.05423105e-01 8.95991027e-02 -2.44711146e-01 7.40289390e-01
4.53634828e-01 -7.64811814e-01 9.58288074e-01 -1.02068055e+00
7.12198198e-01 -1.19231880e+00 -1.02704883e+00 -8.57112259e-02
-6.27282798e-01 -1.84010670e-01 4.71049041e-01 -6.85661435e-02
-4.89025146e-01 -2.32220829e-01 -2.36143827e-01 -4.05337155e-01
-2.22057283e-01 4.86764252e-01 1.49451151e-01 5.54032087e-01
8.26929450e-01 1.79969832e-01 4.50088322e-01 -8.09895754e-01
4.40092146e-01 8.86648297e-01 3.57729971e-01 -5.83074510e-01
2.87035435e-01 4.71952677e-01 -3.44975799e-01 -6.57062054e-01
-2.96770394e-01 -6.37458205e-01 -1.62743479e-01 1.95874244e-01
3.57634604e-01 -1.03964782e+00 -8.05019438e-01 6.18237257e-01
-6.24335349e-01 -6.64888859e-01 -3.40397120e-01 2.77587354e-01
-4.10071492e-01 1.62100449e-01 -5.49085498e-01 -7.87831664e-01
-2.73729950e-01 -8.07508111e-01 8.29868555e-01 1.57046229e-01
-5.43923557e-01 -1.03774023e+00 3.33631486e-01 4.74329025e-01
5.64254940e-01 8.03826988e-01 4.07073170e-01 -1.01356435e+00
-5.98176122e-01 -6.15480959e-01 -1.43018821e-02 8.40569809e-02
9.78470892e-02 6.01685941e-01 -9.76987422e-01 -4.23788875e-01
-2.91926712e-01 -5.53011298e-01 5.55471480e-01 2.50872105e-01
1.40716648e+00 -4.71196592e-01 -2.49670550e-01 8.49605024e-01
1.07102621e+00 -2.67747929e-03 4.79498863e-01 2.55047530e-01
2.30908677e-01 4.90022600e-01 6.51057839e-01 1.10347927e+00
8.43820155e-01 4.76160020e-01 3.72964293e-01 -2.56827772e-01
3.72158080e-01 -8.76905248e-02 2.42825061e-01 5.96738696e-01
1.60128653e-01 -2.07445771e-01 -1.29430199e+00 7.55295277e-01
-2.01487088e+00 -1.18301833e+00 -6.14593029e-01 2.79269433e+00
7.05666840e-01 -6.86014593e-02 5.53127348e-01 -1.15164496e-01
4.55889106e-01 2.42877796e-01 -5.37695646e-01 -5.09465873e-01
4.44763191e-02 -8.12665597e-02 5.22480428e-01 3.80681038e-01
-1.04440844e+00 6.33488357e-01 5.32006264e+00 8.31866443e-01
-1.21460855e+00 2.33237281e-01 7.72808015e-01 -3.93271804e-01
-4.25603420e-01 3.26695479e-02 -5.10606587e-01 7.37466812e-01
1.31271780e+00 -7.06991911e-01 6.97107315e-01 7.97135055e-01
4.97600943e-01 -1.62863150e-01 -1.00430202e+00 7.11607277e-01
-5.83834648e-01 -1.50565434e+00 -8.10337782e-01 2.94786066e-01
6.02962852e-01 9.09709871e-01 2.38576174e-01 2.94053912e-01
5.49882233e-01 -8.34739506e-01 3.67828190e-01 4.66381192e-01
8.11572134e-01 -5.92821240e-01 4.94607568e-01 4.15988803e-01
-7.52402008e-01 1.05521850e-01 1.53518748e-02 -2.74091840e-01
-1.69801563e-02 5.22528827e-01 -8.65503192e-01 6.72785997e-01
6.99672580e-01 8.23251426e-01 -6.07784212e-01 1.16635799e+00
4.72067356e-01 7.51115918e-01 -6.03589475e-01 1.72301829e-01
-1.45241007e-01 -1.64615944e-01 6.29704893e-01 1.14565372e+00
1.39184386e-01 -8.98436904e-02 1.51951984e-01 6.86933517e-01
1.68105066e-01 9.85844582e-02 -1.00059211e+00 -3.82749379e-01
8.52462947e-01 1.18597305e+00 -5.62180221e-01 1.74257392e-03
-4.83312160e-01 4.48035061e-01 -1.42338155e-02 3.73983145e-01
-7.08800077e-01 -5.67070782e-01 9.67612624e-01 3.38072896e-01
-1.02879316e-01 -2.18353823e-01 -5.60360312e-01 -1.17255950e+00
5.67242652e-02 -1.28903866e+00 7.36661494e-01 -8.39526236e-01
-1.33954692e+00 3.87286216e-01 2.11246498e-02 -1.26893091e+00
-6.14107549e-01 -1.30725279e-01 -5.65557361e-01 8.93877983e-01
-6.79588079e-01 -7.99361348e-01 -3.39303315e-01 5.89136899e-01
1.19316742e-01 -3.78649309e-02 8.48094881e-01 4.92177933e-01
-7.20182717e-01 7.44243383e-01 5.45204103e-01 9.72697064e-02
9.08904016e-01 -1.09463549e+00 8.07435513e-01 4.79899466e-01
-1.95332810e-01 1.03100860e+00 7.67590702e-01 -5.92616558e-01
-1.57800007e+00 -1.10101390e+00 6.05450869e-01 -9.73344564e-01
1.03444493e+00 -7.10841417e-01 -9.54367340e-01 1.08610904e+00
1.37714759e-01 -5.47377244e-02 9.62063313e-01 3.82203996e-01
-3.51463169e-01 -1.23413593e-01 -1.18302202e+00 5.88950217e-01
6.08264983e-01 -3.79816741e-01 1.30687892e-01 6.73615575e-01
7.50888169e-01 -4.19296175e-01 -1.21526539e+00 1.73311591e-01
7.65149355e-01 -1.11267841e+00 6.18186474e-01 -2.91368634e-01
3.70825291e-01 -7.18197227e-02 -1.12328688e-02 -9.76500392e-01
1.80790141e-01 -1.16530704e+00 1.86715782e-01 1.40861094e+00
8.03572774e-01 -1.05063546e+00 5.55239260e-01 6.91004992e-01
-4.05954868e-02 -8.19807351e-01 -7.01435149e-01 -7.83990383e-01
-3.96628119e-02 -8.77971172e-01 9.41935837e-01 1.27211523e+00
1.79856315e-01 -1.68342322e-01 -4.65031564e-01 -9.54432040e-02
5.92195094e-01 9.08177868e-02 1.25759566e+00 -9.66789246e-01
-3.10012311e-01 -2.11039633e-01 -1.05027787e-01 -3.67151439e-01
-5.40673971e-01 -1.14243448e+00 -5.09912848e-01 -1.31291950e+00
5.74545376e-02 -5.91989934e-01 -1.04813851e-01 7.94820428e-01
1.84362352e-01 2.60475159e-01 2.19206229e-01 3.01779389e-01
-5.95487237e-01 3.61997843e-01 1.11383760e+00 4.00340021e-01
-2.65648127e-01 4.34952885e-01 -9.21953559e-01 3.03797036e-01
1.16915143e+00 -7.04978585e-01 -1.31880939e-01 -2.86748670e-02
2.20797107e-01 5.14642715e-01 6.69490278e-01 -8.23913991e-01
-3.05897947e-02 -2.61114061e-01 1.04494050e-01 -5.38291514e-01
3.95540714e-01 -5.86500049e-01 7.01106489e-01 2.15622827e-01
-2.17023432e-01 2.96084732e-01 4.58119035e-01 2.68301666e-01
6.35639504e-02 3.60069275e-01 4.71163750e-01 1.38960378e-02
-4.78209322e-03 5.26076376e-01 5.41603379e-02 5.23745596e-01
9.03893113e-01 1.00797161e-01 -8.21230352e-01 -8.35828781e-01
-7.41083682e-01 8.28198671e-01 9.74203944e-01 7.94853806e-01
1.02133468e-01 -1.26089442e+00 -8.26695204e-01 6.23859346e-01
2.57174969e-01 -2.38208726e-01 8.28555152e-02 1.31097853e+00
-4.08364445e-01 1.68161064e-01 -1.53420106e-01 -4.53737020e-01
-1.10330606e+00 4.95402068e-01 3.74706127e-02 1.11020967e-01
-5.48777461e-01 7.83194244e-01 -2.26364762e-01 -8.56499791e-01
2.18251765e-01 -2.73497611e-01 5.61221540e-01 9.02896598e-02
5.24692059e-01 5.54960787e-01 -3.95259336e-02 -2.28417233e-01
-5.65273762e-01 -2.29406774e-01 9.26649198e-02 -2.21281826e-01
1.48645830e+00 6.79083169e-03 -2.24378601e-01 1.01816547e+00
1.15232360e+00 1.68789566e-01 -1.17264211e+00 8.21060911e-02
1.62874795e-02 -4.22395259e-01 -4.74257499e-01 -9.36463892e-01
-6.66205168e-01 4.82424796e-01 4.17097211e-01 7.28653312e-01
9.25617099e-01 7.11271465e-02 3.46566617e-01 -3.65393162e-02
5.62676966e-01 -7.10498154e-01 -2.00839579e-01 4.23508286e-01
1.24400759e+00 -1.32699466e+00 1.20373167e-01 3.65232080e-02
-8.78380418e-01 6.58348024e-01 2.83965617e-01 -4.71201614e-02
7.04995215e-01 5.33547938e-01 2.90257394e-01 -2.59552538e-01
-1.36697948e+00 1.12616770e-01 -2.40314141e-01 6.20700896e-01
6.58461809e-01 3.05994064e-01 -3.23009044e-01 7.65433490e-01
-3.78570795e-01 2.35600814e-01 7.81577051e-01 6.47754133e-01
1.70442596e-01 -1.44297576e+00 -4.22622085e-01 8.64067137e-01
-7.12582290e-01 -2.77869105e-01 -3.33046526e-01 1.09192789e+00
-3.11553061e-01 7.99604893e-01 8.90765563e-02 -2.89139718e-01
2.60499507e-01 2.24463999e-01 -2.50899762e-01 -4.80603158e-01
-8.12753201e-01 5.70694823e-03 4.13895071e-01 -7.47351408e-01
3.73427749e-01 -1.18478429e+00 -9.65024650e-01 -9.15241599e-01
9.26311165e-02 2.97421455e-01 9.11913455e-01 1.79928169e-01
8.94332409e-01 7.91495219e-02 6.74630463e-01 -5.26453376e-01
-3.82528126e-01 -1.02991056e+00 -5.43312669e-01 1.01661071e-01
4.79758054e-01 -3.69756341e-01 -4.03875023e-01 -2.83664256e-01]
|
[7.2228617668151855, 3.7023627758026123]
|
0df1e9d0-40f0-4ea5-83f0-da60a5ad62b3
|
lidarmultinet-towards-a-unified-multi-task
|
2209.09385
| null |
https://arxiv.org/abs/2209.09385v2
|
https://arxiv.org/pdf/2209.09385v2.pdf
|
LidarMultiNet: Towards a Unified Multi-Task Network for LiDAR Perception
|
LiDAR-based 3D object detection, semantic segmentation, and panoptic segmentation are usually implemented in specialized networks with distinctive architectures that are difficult to adapt to each other. This paper presents LidarMultiNet, a LiDAR-based multi-task network that unifies these three major LiDAR perception tasks. Among its many benefits, a multi-task network can reduce the overall cost by sharing weights and computation among multiple tasks. However, it typically underperforms compared to independently combined single-task models. The proposed LidarMultiNet aims to bridge the performance gap between the multi-task network and multiple single-task networks. At the core of LidarMultiNet is a strong 3D voxel-based encoder-decoder architecture with a Global Context Pooling (GCP) module extracting global contextual features from a LiDAR frame. Task-specific heads are added on top of the network to perform the three LiDAR perception tasks. More tasks can be implemented simply by adding new task-specific heads while introducing little additional cost. A second stage is also proposed to refine the first-stage segmentation and generate accurate panoptic segmentation results. LidarMultiNet is extensively tested on both Waymo Open Dataset and nuScenes dataset, demonstrating for the first time that major LiDAR perception tasks can be unified in a single strong network that is trained end-to-end and achieves state-of-the-art performance. Notably, LidarMultiNet reaches the official 1st place in the Waymo Open Dataset 3D semantic segmentation challenge 2022 with the highest mIoU and the best accuracy for most of the 22 classes on the test set, using only LiDAR points as input. It also sets the new state-of-the-art for a single model on the Waymo 3D object detection benchmark and three nuScenes benchmarks.
|
['Hassan Foroosh', 'Panqu Wang', 'Yu Wang', 'Yufei Xie', 'Weijia Chen', 'Zixiang Zhou', 'Dongqiangzi Ye']
|
2022-09-19
| null | null | null | null |
['panoptic-segmentation']
|
['computer-vision']
|
[ 3.44954759e-01 -1.09247841e-01 -4.63499390e-02 -6.34066939e-01
-1.15287971e+00 -4.09732521e-01 5.61866343e-01 -4.33287658e-02
-6.63773835e-01 2.83062607e-01 -4.47021216e-01 -3.26070279e-01
1.58724234e-01 -7.96923339e-01 -8.73805940e-01 -4.81124490e-01
1.37978584e-01 8.26292336e-01 8.61908615e-01 -5.20945229e-02
-4.16038260e-02 4.79154587e-01 -1.89201343e+00 2.60887563e-01
9.66201425e-01 1.35130274e+00 7.14912951e-01 6.70270503e-01
-3.43386024e-01 5.10881543e-02 -3.86435688e-01 -1.63222983e-01
6.25761688e-01 2.98531115e-01 -4.43583518e-01 3.27422470e-02
1.35903513e+00 -2.74246335e-01 1.58018872e-01 7.69326448e-01
8.06663513e-01 7.56661817e-02 6.08397305e-01 -1.23520172e+00
-3.38984519e-01 3.71574104e-01 -8.10627818e-01 1.09730922e-01
-2.71912247e-01 3.04026753e-01 1.20194113e+00 -1.15532005e+00
3.46054971e-01 1.56355166e+00 7.82227218e-01 3.27169716e-01
-1.45722115e+00 -9.08481836e-01 3.72778207e-01 -7.63041675e-02
-1.50723994e+00 -5.81336282e-02 5.90323865e-01 -3.92089218e-01
1.23093235e+00 -5.69815747e-02 3.83216947e-01 8.16245496e-01
1.51342988e-01 8.40718508e-01 1.04410887e+00 1.98484927e-01
1.29129635e-02 -7.19670057e-02 2.78155487e-02 7.41039872e-01
3.04380417e-01 1.73915312e-01 -4.25663114e-01 4.11524951e-01
3.88155907e-01 -1.74689844e-01 2.60041893e-01 -5.13548732e-01
-9.62754488e-01 9.25619841e-01 9.55123127e-01 2.31474079e-03
-1.56501934e-01 3.79188746e-01 3.88486326e-01 -7.46955648e-02
7.85387337e-01 -4.35100542e-03 -6.78008854e-01 4.07499284e-01
-1.13706315e+00 3.56219471e-01 4.39300358e-01 9.14465308e-01
1.12836921e+00 1.27286822e-01 -7.20302612e-02 8.88303459e-01
3.53634208e-01 8.42688322e-01 -1.03142232e-01 -7.91657329e-01
7.06076205e-01 6.06074691e-01 -2.56225646e-01 -3.89169365e-01
-5.65547824e-01 -4.92815554e-01 -5.09655833e-01 5.91048837e-01
2.11326748e-01 -2.22569890e-02 -1.43500948e+00 1.47308719e+00
4.40257192e-01 4.49510872e-01 -8.46285522e-02 8.52699637e-01
1.15372181e+00 7.73917139e-01 3.29859138e-01 5.17059267e-01
1.40397727e+00 -1.03449941e+00 -1.69646367e-02 -8.17637324e-01
3.17484111e-01 -8.42546761e-01 1.14864540e+00 1.23816364e-01
-8.40034306e-01 -1.10404658e+00 -1.26075423e+00 -5.97954810e-01
-8.79210234e-01 1.22815408e-01 6.60466194e-01 5.70031047e-01
-1.13758767e+00 2.84289330e-01 -6.55319214e-01 -2.95135826e-01
8.20802152e-01 3.67125541e-01 -7.48536661e-02 -1.29851520e-01
-8.22047114e-01 1.04402602e+00 5.59543014e-01 2.16913059e-01
-1.18979633e+00 -1.13561070e+00 -1.04586864e+00 -3.07666399e-02
6.46152139e-01 -9.21672761e-01 1.28117561e+00 -2.06618920e-01
-1.11581314e+00 1.14374387e+00 -9.30906087e-02 -6.58540130e-01
5.25630713e-01 -2.86348104e-01 -7.98131824e-02 7.15472251e-02
4.73279238e-01 1.63149095e+00 9.43655372e-01 -1.45509815e+00
-1.29041076e+00 -3.50504011e-01 -1.53143615e-01 2.31148556e-01
3.18749458e-01 -3.49469721e-01 -5.77719271e-01 -2.38410816e-01
1.10619597e-01 -8.44604850e-01 -3.33418578e-01 3.05176407e-01
-3.49399447e-01 -4.47249144e-01 1.38673067e+00 -6.10189028e-02
3.85148585e-01 -2.05934715e+00 -4.96652089e-02 -3.54662687e-01
1.19527094e-01 4.09930378e-01 -4.85239714e-01 -2.35980392e-01
1.89320464e-02 2.30899945e-01 -7.71244228e-01 -8.76350224e-01
1.35760099e-01 3.83617908e-01 -3.47440422e-01 1.57025501e-01
5.56176901e-01 1.19480622e+00 -6.77372158e-01 -5.11165619e-01
7.54710555e-01 5.93190789e-01 -4.45335120e-01 6.18810020e-02
-6.89974606e-01 3.16836506e-01 -2.75788128e-01 8.05559635e-01
1.20375955e+00 1.53083965e-01 -4.34891433e-01 -3.01848143e-01
-4.60775405e-01 4.22878683e-01 -1.18299901e+00 2.00700855e+00
-9.43083048e-01 7.33313143e-01 3.48913729e-01 -8.54734123e-01
1.04730332e+00 -1.92381665e-02 4.08211172e-01 -7.50767529e-01
8.05190951e-02 4.71944422e-01 -1.13455519e-01 -2.66640931e-01
5.99265754e-01 -1.73478395e-01 -3.76831591e-01 2.15646744e-01
3.01433295e-01 -1.06729114e+00 2.32515469e-01 -1.07552186e-01
4.68249679e-01 3.10755610e-01 -1.31187484e-01 -2.63374418e-01
5.22007167e-01 8.54990557e-02 4.56006676e-01 7.72995830e-01
-2.60211021e-01 7.21695900e-01 1.22985467e-01 -3.86770219e-01
-8.17398012e-01 -1.46636724e+00 -5.67500591e-01 1.18866920e+00
2.19789311e-01 -1.18016891e-01 -2.82728732e-01 -6.78255737e-01
5.32459199e-01 7.40014672e-01 -4.46283847e-01 2.34561712e-01
-5.06024361e-01 -7.99082458e-01 7.55484879e-01 4.89532083e-01
9.46385682e-01 -9.41548645e-01 -9.38853383e-01 2.21838847e-01
2.99098366e-03 -1.81314397e+00 -1.83187321e-01 5.51697850e-01
-8.18139672e-01 -9.63657320e-01 -4.69478488e-01 -7.73608148e-01
-9.57209209e-04 5.60622334e-01 1.11032033e+00 -4.04322475e-01
-4.51520711e-01 5.51535822e-02 4.48044576e-02 -8.50981414e-01
3.71925607e-02 5.83317339e-01 -4.12583679e-01 -1.91293001e-01
4.61308897e-01 -6.27672017e-01 -3.02984983e-01 2.89877862e-01
-7.95317948e-01 5.09316549e-02 6.53198600e-01 4.13612634e-01
8.22471082e-01 -1.98584110e-01 5.03571212e-01 -6.29782379e-01
1.73793003e-01 -2.44525999e-01 -1.00024843e+00 -1.55680120e-01
-4.84851122e-01 -2.54972488e-01 2.85525382e-01 9.95944161e-03
-1.01197743e+00 3.89409304e-01 -3.22303891e-01 -4.64694679e-01
-4.39699262e-01 1.98500946e-01 -1.43624097e-01 -1.96303710e-01
3.82407844e-01 9.15788859e-02 -3.38108867e-01 -7.41834342e-01
7.33913481e-01 5.54315031e-01 6.85226083e-01 -4.58712757e-01
1.03451383e+00 6.96210384e-01 3.51044387e-01 -1.06729841e+00
-1.33515275e+00 -5.76161981e-01 -1.05836701e+00 -7.91589767e-02
1.23507917e+00 -1.35307300e+00 -6.15940034e-01 6.37089729e-01
-1.22708881e+00 -5.25419533e-01 -3.09040397e-01 9.96955037e-02
-4.25379306e-01 1.09028660e-01 -8.81750286e-02 -6.09629154e-01
-3.69316429e-01 -1.37503481e+00 1.73640752e+00 1.52836055e-01
3.80481929e-01 -7.06872106e-01 -2.50610024e-01 5.35360873e-01
3.64401758e-01 2.74411470e-01 8.68249655e-01 -3.28101754e-01
-1.00940955e+00 1.03051193e-01 -8.78648698e-01 5.25913775e-01
-1.72658309e-01 2.90545411e-02 -1.37013471e+00 -2.33026482e-02
-3.24764311e-01 -4.90655005e-01 1.68564129e+00 5.32655776e-01
1.19762170e+00 6.35234177e-01 -3.08228910e-01 9.10203874e-01
1.52950454e+00 -1.02252036e-01 2.62384504e-01 3.34074609e-02
9.39307809e-01 5.63170552e-01 6.17185831e-01 -1.66985959e-01
8.74544442e-01 7.22868681e-01 9.22531784e-01 -2.96862006e-01
-5.03386855e-01 -2.24217683e-01 3.34705085e-01 3.25338691e-01
3.53593618e-01 -2.24149570e-01 -1.02547657e+00 7.40747392e-01
-1.67711353e+00 -5.06274343e-01 -5.35984695e-01 1.98160744e+00
3.89102519e-01 5.42256534e-01 -1.40224472e-02 -1.42888084e-01
4.33399826e-01 6.93204224e-01 -6.83152795e-01 -5.23913682e-01
-2.16285065e-01 7.39638090e-01 7.55549133e-01 6.99247062e-01
-1.52131653e+00 1.66578877e+00 5.41729546e+00 8.97196949e-01
-1.33944023e+00 4.62907821e-01 2.70501852e-01 -2.26828128e-01
3.48920375e-02 -7.88584948e-02 -1.34281456e+00 1.65477991e-01
8.27235103e-01 2.31756002e-01 -6.71522543e-02 8.74419630e-01
2.32349291e-01 -4.21149224e-01 -1.01827312e+00 7.96709538e-01
-1.13424480e-01 -1.29716325e+00 2.02181756e-01 -1.08808407e-03
6.84024751e-01 1.00791240e+00 2.92514805e-02 4.69177544e-01
1.29344985e-01 -1.00163579e+00 7.80931771e-01 -4.21020910e-02
9.28276896e-01 -6.93565011e-01 5.88236392e-01 2.97654212e-01
-1.82863498e+00 -5.83018363e-02 -5.80463469e-01 1.24743730e-01
4.75804448e-01 6.98609471e-01 -8.64089966e-01 5.76656699e-01
8.48450959e-01 6.65640831e-01 -7.07177401e-01 1.15544391e+00
-3.78372967e-01 3.08069050e-01 -7.35578597e-01 2.14429766e-01
8.42056036e-01 -2.09018305e-01 6.72041476e-01 1.47205102e+00
2.22829506e-01 -4.08298612e-01 6.08539283e-01 1.14696324e+00
-2.32840180e-01 -2.36955181e-01 -6.71001792e-01 4.48732346e-01
2.93915629e-01 1.46652055e+00 -6.91622615e-01 -3.41189265e-01
-3.61477435e-01 7.15583861e-01 9.24835652e-02 1.28921852e-01
-9.36750472e-01 -3.01259071e-01 9.12418783e-01 -5.84729463e-02
8.41078043e-01 -4.80869800e-01 -7.11108267e-01 -6.69525921e-01
-7.82479495e-02 -1.38773009e-01 2.88708925e-01 -8.57625306e-01
-1.09600770e+00 4.65724736e-01 2.04557210e-01 -8.59732568e-01
1.69424027e-01 -7.73952544e-01 -6.41063452e-01 1.01309013e+00
-2.37645864e+00 -1.67092144e+00 -5.68583190e-01 4.13290739e-01
9.54489410e-01 2.52607942e-01 4.62772876e-01 6.07567251e-01
-6.42357171e-01 1.45678177e-01 -6.38563216e-01 -1.69370979e-01
5.96865535e-01 -1.47270274e+00 8.40198755e-01 8.57812941e-01
3.99088450e-02 -2.50778198e-01 1.30963892e-01 -5.68362832e-01
-8.82960856e-01 -1.68497753e+00 1.14957452e+00 -5.12972772e-01
6.16656005e-01 -8.41987848e-01 -8.20563257e-01 5.73575914e-01
-9.62988380e-03 1.25290021e-01 1.23615623e-01 -1.36898547e-01
-3.59150440e-01 -4.82809782e-01 -1.09546137e+00 3.72623175e-01
1.23802185e+00 -6.02476835e-01 -6.13676429e-01 3.86621654e-01
1.15721130e+00 -5.83465040e-01 -5.45385242e-01 6.76373720e-01
2.50593066e-01 -1.01260269e+00 1.33829737e+00 -1.47379667e-01
4.74793911e-01 -6.19817853e-01 -4.35673535e-01 -1.04833138e+00
5.23938164e-02 -2.91744918e-01 1.88622624e-01 1.08111763e+00
6.76913857e-01 -5.97567260e-01 7.19454527e-01 -1.00654908e-01
-6.64678037e-01 -6.67286813e-01 -1.37549424e+00 -9.01262760e-01
2.25091785e-01 -9.94192600e-01 5.72058082e-01 4.10978526e-01
-1.19736278e+00 8.09530795e-01 5.78985959e-02 4.47126925e-01
8.75247896e-01 4.14261252e-01 8.97624195e-01 -1.52842057e+00
3.54263157e-01 -6.82150424e-01 -1.77813396e-01 -1.33477831e+00
2.09429890e-01 -1.40455043e+00 3.25571388e-01 -1.69684601e+00
-2.11366013e-01 -5.64740598e-01 -6.91700429e-02 5.93591511e-01
-6.83118030e-02 5.85405052e-01 4.77240443e-01 -2.02476755e-01
-3.55414510e-01 7.68839717e-01 1.31049728e+00 -3.20781142e-01
-2.88496733e-01 1.17503434e-01 -5.24393201e-01 8.02020848e-01
6.58369124e-01 -6.04145586e-01 -3.30386996e-01 -8.11693072e-01
-4.46981527e-02 -4.56812590e-01 8.59353960e-01 -1.25111139e+00
1.74522027e-01 1.33190423e-01 8.71966332e-02 -1.44067621e+00
7.35502124e-01 -8.75556409e-01 -4.15279001e-01 3.92844230e-01
1.47110134e-01 -1.88201860e-01 5.91450274e-01 4.31555212e-01
-7.71206245e-02 -2.80664600e-02 1.12632716e+00 -1.48002282e-01
-1.21859634e+00 5.62506676e-01 -1.42609462e-01 1.15953527e-01
9.27360177e-01 -4.41161990e-01 -2.63340890e-01 2.76999444e-01
-4.97001112e-01 7.10114777e-01 -1.20711625e-01 7.48754561e-01
6.23621821e-01 -8.60392272e-01 -8.29071403e-01 1.46599904e-01
1.80660889e-01 8.57120454e-01 8.24260265e-02 6.43917620e-01
-3.37617278e-01 6.01513267e-01 -8.44386145e-02 -1.29858911e+00
-9.57298815e-01 4.63687964e-02 5.56429088e-01 -1.96355686e-01
-5.74060142e-01 1.06196761e+00 2.57214397e-01 -9.34419930e-01
1.81655467e-01 -7.98366129e-01 7.52508193e-02 4.36599135e-01
-7.93937966e-02 1.99690714e-01 2.67890006e-01 -6.57647431e-01
-4.18726653e-01 1.07332098e+00 1.69096038e-01 1.29445791e-01
1.33562112e+00 -4.72627953e-02 5.72756119e-02 5.46240687e-01
1.10106814e+00 -3.75847697e-01 -1.51004398e+00 -2.70804465e-01
-1.20453730e-01 -3.23822051e-01 5.25872529e-01 -9.61295545e-01
-1.25123143e+00 1.48236370e+00 7.54410148e-01 -1.27185836e-01
9.17254388e-01 5.61832115e-02 1.07447600e+00 4.39472288e-01
1.85257256e-01 -1.04367375e+00 3.72720021e-03 8.94344568e-01
7.66840100e-01 -1.61197710e+00 1.82356697e-03 -7.21448243e-01
-4.54420477e-01 7.75129020e-01 9.29346025e-01 -2.24527970e-01
7.97524154e-01 3.00273210e-01 9.22736302e-02 -3.53231698e-01
-5.79167366e-01 -9.21595156e-01 4.71578777e-01 7.18272209e-01
1.82889123e-02 1.63462698e-01 1.55198067e-01 3.13643783e-01
-2.72405833e-01 -1.36631832e-01 2.10333429e-02 5.38031459e-01
-7.98628092e-01 -9.24448967e-01 -6.09558895e-02 3.72463137e-01
7.54390210e-02 -1.93755403e-01 -3.15483302e-01 1.10945666e+00
8.35161507e-01 8.31278682e-01 4.19539809e-01 -2.95728564e-01
5.41385651e-01 1.09720632e-01 3.12074393e-01 -8.47646773e-01
-7.49621749e-01 2.31775432e-03 6.62579834e-02 -6.22315347e-01
-4.06893671e-01 -5.25628686e-01 -1.17376554e+00 -1.00071162e-01
-1.55957386e-01 -4.89678562e-01 9.53450263e-01 9.54207182e-01
4.32592005e-01 6.37290120e-01 4.28102881e-01 -1.36780000e+00
-2.84760088e-01 -8.41190159e-01 -2.37188607e-01 -1.54945880e-01
4.30546999e-01 -8.24600875e-01 -4.54486534e-02 -3.37270707e-01]
|
[8.125592231750488, -2.7149081230163574]
|
e8ac1f98-ccf5-4f42-8e06-cf17350ef299
|
multi-modal-dense-video-captioning
|
2003.07758
| null |
https://arxiv.org/abs/2003.07758v2
|
https://arxiv.org/pdf/2003.07758v2.pdf
|
Multi-modal Dense Video Captioning
|
Dense video captioning is a task of localizing interesting events from an untrimmed video and producing textual description (captions) for each localized event. Most of the previous works in dense video captioning are solely based on visual information and completely ignore the audio track. However, audio, and speech, in particular, are vital cues for a human observer in understanding an environment. In this paper, we present a new dense video captioning approach that is able to utilize any number of modalities for event description. Specifically, we show how audio and speech modalities may improve a dense video captioning model. We apply automatic speech recognition (ASR) system to obtain a temporally aligned textual description of the speech (similar to subtitles) and treat it as a separate input alongside video frames and the corresponding audio track. We formulate the captioning task as a machine translation problem and utilize recently proposed Transformer architecture to convert multi-modal input data into textual descriptions. We demonstrate the performance of our model on ActivityNet Captions dataset. The ablation studies indicate a considerable contribution from audio and speech components suggesting that these modalities contain substantial complementary information to video frames. Furthermore, we provide an in-depth analysis of the ActivityNet Caption results by leveraging the category tags obtained from original YouTube videos. Code is publicly available: github.com/v-iashin/MDVC
|
['Esa Rahtu', 'Vladimir Iashin']
|
2020-03-17
| null | null | null | null |
['dense-video-captioning']
|
['computer-vision']
|
[ 5.11268020e-01 1.17102526e-01 -2.46706441e-01 -2.48835415e-01
-1.29717267e+00 -7.35728383e-01 7.56809115e-01 -1.90351292e-01
-1.10221155e-01 7.47084737e-01 8.91413569e-01 6.88814521e-02
4.45907921e-01 -8.88906270e-02 -1.12349427e+00 -5.25566280e-01
6.54485375e-02 2.00977504e-01 1.39088526e-01 6.97932616e-02
-7.06503689e-02 5.63064702e-02 -1.54758203e+00 9.35691714e-01
2.46894598e-01 9.95233893e-01 4.82453674e-01 9.07552183e-01
-6.87964335e-02 1.19550669e+00 -5.95801532e-01 -1.42517127e-02
-4.20925319e-02 -7.11570501e-01 -7.94338703e-01 3.17510754e-01
4.68252689e-01 -5.43801785e-01 -8.81997228e-01 7.98277080e-01
3.57164741e-01 1.38617635e-01 3.32164168e-01 -1.61754751e+00
-7.96546221e-01 6.73788071e-01 -3.77958536e-01 5.03771365e-01
9.67836380e-01 4.75254059e-02 8.72287273e-01 -9.80245590e-01
7.59700477e-01 9.97234404e-01 2.16065019e-01 6.59332871e-01
-8.80632937e-01 -5.81590295e-01 2.76294798e-01 6.54176295e-01
-1.41198623e+00 -7.83305526e-01 8.50989521e-01 -4.75085557e-01
9.12393689e-01 3.68121058e-01 6.79904759e-01 1.82936549e+00
-3.33068788e-01 1.00508201e+00 4.41405416e-01 -2.51288414e-01
6.57753572e-02 3.33757140e-02 -1.58889294e-01 2.50919968e-01
-2.86068469e-01 2.74699610e-02 -9.66345787e-01 1.87103394e-02
7.52600849e-01 9.18804407e-02 -6.74690962e-01 -1.81653991e-01
-1.76602733e+00 5.98831475e-01 2.74454474e-01 3.93208772e-01
-5.73818684e-01 4.81514633e-01 5.53898990e-01 3.41884308e-02
3.55243951e-01 7.25892261e-02 -1.55882925e-01 -5.35759032e-01
-1.04611135e+00 -2.62013555e-01 4.25741464e-01 1.27127123e+00
5.35690188e-01 1.63070887e-01 -5.34914851e-01 6.13603175e-01
2.48473451e-01 5.87074876e-01 5.43148994e-01 -1.05394995e+00
7.72942245e-01 4.97523360e-02 1.10431485e-01 -7.97133088e-01
5.14258631e-03 5.50451083e-03 -4.52324569e-01 -5.81180274e-01
4.72956970e-02 -1.74902245e-01 -1.09995973e+00 1.81797874e+00
-4.45979200e-02 8.99962187e-01 1.59125611e-01 1.13458109e+00
1.14585626e+00 1.23430419e+00 2.55008489e-01 -2.43385583e-01
1.45803905e+00 -1.03465486e+00 -9.77461278e-01 -3.83694917e-01
6.17488623e-02 -7.20013261e-01 8.19199085e-01 -6.62944987e-02
-1.06505311e+00 -4.90445584e-01 -8.22585762e-01 -1.36901110e-01
-1.20825566e-01 1.49749145e-01 3.06681246e-01 7.42233023e-02
-1.24175656e+00 -4.39139605e-02 -9.41783547e-01 -6.44512534e-01
1.72482491e-01 1.12029023e-01 -6.96320832e-01 -1.30815640e-01
-1.31970727e+00 5.48025846e-01 5.20273030e-01 3.39414477e-02
-1.46119833e+00 -4.01151508e-01 -1.27839231e+00 1.89581603e-01
3.43229502e-01 -5.41060507e-01 1.47101831e+00 -1.26821971e+00
-1.12478149e+00 5.14047563e-01 -6.35322809e-01 -5.27655363e-01
-2.42758542e-02 -3.93128321e-02 -5.30237079e-01 9.92832243e-01
1.07161954e-01 1.13355589e+00 1.01046693e+00 -1.24551189e+00
-5.75679004e-01 1.35446623e-01 2.03524426e-01 3.82140577e-01
-3.87324214e-01 2.22848579e-01 -7.42917836e-01 -8.14648926e-01
-2.49055222e-01 -8.40649784e-01 2.89009035e-01 -7.55030662e-02
-2.53046781e-01 2.81009346e-01 1.11182201e+00 -9.55085993e-01
1.15121758e+00 -2.40754461e+00 2.01775029e-01 -2.70308912e-01
1.31026551e-01 -2.39640959e-02 -4.26223010e-01 5.34123898e-01
-3.06068987e-01 7.98264742e-02 -1.49250001e-01 -4.47312355e-01
-1.37294620e-01 2.08328515e-01 -5.84473133e-01 3.42567116e-01
3.74744713e-01 1.05902481e+00 -1.04116416e+00 -6.69292092e-01
3.16380143e-01 9.29137349e-01 -4.09085870e-01 4.13756669e-01
-1.62142709e-01 5.15029907e-01 -3.40489149e-01 7.92255044e-01
1.93716452e-01 -4.33254570e-01 -1.49031326e-01 -5.02902806e-01
5.65816388e-02 1.35791376e-01 -6.64187551e-01 2.02604437e+00
-4.15104419e-01 1.18321812e+00 1.49661705e-01 -9.34308529e-01
5.15231252e-01 1.05431759e+00 6.90114796e-01 -5.79925776e-01
6.56068474e-02 -4.92383391e-02 -4.59697187e-01 -7.10706592e-01
4.95312423e-01 -3.74247902e-04 -2.57174641e-01 4.16200489e-01
3.75661165e-01 3.28489155e-01 2.13250041e-01 5.82843781e-01
1.26952589e+00 2.31497690e-01 2.16288254e-01 5.11000395e-01
2.73595303e-01 1.84367537e-01 2.16962695e-01 5.70972204e-01
-3.85091633e-01 1.34688723e+00 2.84148753e-01 6.29011020e-02
-1.24499333e+00 -1.15997458e+00 2.02621594e-01 1.10675132e+00
1.69640854e-01 -5.81277370e-01 -6.09419823e-01 -4.88095254e-01
-5.48035204e-01 6.23034596e-01 -6.50370657e-01 -6.46204054e-02
-5.51697433e-01 -1.82494760e-01 5.18219054e-01 6.65246546e-01
2.96083599e-01 -1.19827354e+00 -3.31487805e-01 8.19523484e-02
-1.00846875e+00 -1.70768559e+00 -9.34313476e-01 7.97557179e-03
-3.77684534e-01 -7.97642767e-01 -1.17010927e+00 -9.86569703e-01
5.60532570e-01 4.38180357e-01 9.46796894e-01 -2.70597398e-01
-7.21602514e-02 9.25096750e-01 -7.66557097e-01 -9.89406779e-02
-4.12510246e-01 -1.74462855e-01 -1.30933002e-02 2.86082149e-01
3.19745928e-01 -5.38495600e-01 -4.07651275e-01 2.62262195e-01
-9.54680085e-01 4.13603097e-01 5.11201382e-01 5.76492667e-01
5.50151885e-01 -5.94774783e-01 6.11991048e-01 -2.02634841e-01
3.07860851e-01 -9.19467449e-01 -1.50317710e-03 1.22099809e-01
2.28897557e-01 -1.85550570e-01 4.81251150e-01 -6.48742318e-01
-9.49236572e-01 4.96808439e-01 -7.63160363e-02 -1.12560606e+00
-3.59766454e-01 4.47524607e-01 -1.44285172e-01 2.89579213e-01
4.13597792e-01 5.63796878e-01 -5.88756008e-03 -2.49311984e-01
2.79594123e-01 9.99285936e-01 8.90213370e-01 -3.47230762e-01
7.18317032e-01 5.05901814e-01 -5.29518127e-01 -8.83359253e-01
-6.56832159e-01 -7.28220880e-01 -4.19309914e-01 -5.49302340e-01
1.05153644e+00 -1.39080620e+00 -4.67825025e-01 -5.04564419e-02
-1.44681823e+00 -6.18413351e-02 -2.14897931e-01 6.62547231e-01
-8.06230545e-01 4.01821822e-01 -5.76326549e-01 -4.23021346e-01
-1.62070230e-01 -1.13409102e+00 1.51437414e+00 1.34265304e-01
-2.94965446e-01 -7.33324945e-01 -1.53481960e-03 5.10913134e-01
2.81672060e-01 2.55360365e-01 2.76079178e-01 -7.92718232e-01
-6.52496755e-01 -2.14993507e-01 -2.73079395e-01 7.35648721e-02
1.51802808e-01 -1.68054670e-01 -1.16921306e+00 -1.51709870e-01
-1.85684681e-01 -2.89730549e-01 7.48981476e-01 3.90655100e-01
9.57368672e-01 -4.42312211e-01 -3.28146338e-01 4.51408058e-01
1.07818818e+00 4.06182587e-01 7.06805527e-01 2.14352459e-01
9.15777504e-01 3.77522290e-01 6.10209167e-01 4.96526003e-01
2.94382811e-01 7.82116413e-01 5.09995162e-01 -1.65447474e-01
-4.60524499e-01 -6.00508988e-01 9.15335298e-01 7.98264444e-01
1.59836560e-01 -5.77630281e-01 -7.41247714e-01 8.89053047e-01
-2.01255941e+00 -1.39793015e+00 2.30677485e-01 1.74864161e+00
6.36384904e-01 -2.78106153e-01 6.28545135e-02 -1.07137486e-01
1.20472968e+00 2.92176485e-01 -4.05018091e-01 -9.20375660e-02
-9.84837040e-02 -2.01896593e-01 1.88927293e-01 3.55105221e-01
-1.03762817e+00 7.40447879e-01 6.35105515e+00 6.35986209e-01
-1.18608737e+00 2.71177709e-01 3.39410752e-01 -5.05814135e-01
-3.76243412e-01 -1.89853489e-01 -4.03744131e-01 6.21569455e-01
1.49033451e+00 -3.83433342e-01 4.10293460e-01 6.25606060e-01
6.17110491e-01 2.95379129e-03 -1.44011974e+00 1.33828580e+00
4.74276364e-01 -1.50575280e+00 3.72101843e-01 -2.04107523e-01
4.39376652e-01 1.23598553e-01 -3.44785787e-02 2.46584401e-01
-3.48423928e-01 -9.79153275e-01 1.13185608e+00 2.71273524e-01
1.04212546e+00 -2.98687130e-01 5.52852929e-01 2.73929238e-02
-1.53889239e+00 -1.39094722e-02 -7.94554688e-03 9.49953347e-02
8.10097039e-01 5.83858900e-02 -1.09182179e+00 4.24802691e-01
7.66459942e-01 1.11740518e+00 -4.09129828e-01 1.06623900e+00
-1.36122257e-01 7.07193911e-01 -2.38132417e-01 3.02302241e-01
2.95881093e-01 2.52306074e-01 7.81686902e-01 1.32541358e+00
6.26339436e-01 1.23847857e-01 5.91623411e-02 6.80754781e-01
-2.38023087e-01 -1.25276431e-01 -6.93309069e-01 -4.62500572e-01
4.58377630e-01 1.03666556e+00 -5.95943034e-01 -4.39248711e-01
-7.36835122e-01 1.14333332e+00 -1.58130318e-01 5.79457700e-01
-1.26054204e+00 -1.34027138e-01 6.43741667e-01 1.03528328e-01
4.38539863e-01 -6.59478530e-02 3.52762491e-01 -1.44360578e+00
2.01028049e-01 -8.10927927e-01 3.93924832e-01 -1.51539099e+00
-7.92474389e-01 8.59524071e-01 2.79218048e-01 -1.69186032e+00
-6.49003804e-01 -1.62076131e-01 -4.20560122e-01 5.42716444e-01
-1.31477571e+00 -1.12810934e+00 -5.78218699e-01 9.18932438e-01
1.07732034e+00 -6.31515980e-02 5.87297261e-01 6.28758252e-01
-4.13439572e-01 2.84437954e-01 -6.08759597e-02 3.33962291e-01
8.92274499e-01 -8.31752777e-01 2.96391189e-01 1.07948923e+00
4.72355574e-01 2.17791483e-01 8.81703079e-01 -6.38514638e-01
-1.35415804e+00 -1.23685491e+00 8.31333637e-01 -4.38559860e-01
7.66458511e-01 -5.34193754e-01 -7.71770895e-01 1.09678602e+00
5.64326227e-01 8.96419734e-02 5.81948817e-01 -5.24063885e-01
-1.63653627e-01 2.51039952e-01 -8.11194003e-01 4.67297822e-01
1.01542175e+00 -9.94126558e-01 -8.63409221e-01 2.67292529e-01
1.13059986e+00 -3.94501060e-01 -5.74331999e-01 8.95543844e-02
3.34240705e-01 -4.67568845e-01 1.03319800e+00 -4.24103230e-01
6.75807476e-01 -5.09076178e-01 -2.84408599e-01 -1.08471942e+00
-5.57453595e-02 -7.69369662e-01 -3.23672980e-01 1.37049425e+00
4.12870586e-01 4.83729318e-02 6.34897053e-01 3.74260932e-01
-3.76015455e-01 -8.29508603e-02 -9.49966669e-01 -5.74120224e-01
-7.38019824e-01 -5.84436953e-01 2.14982510e-01 8.63345921e-01
4.06895697e-01 4.25017655e-01 -8.67299080e-01 3.16747934e-01
2.88985729e-01 2.16602348e-02 3.33110213e-01 -6.52183235e-01
-1.17834441e-01 2.05027372e-01 -4.46617782e-01 -1.16018486e+00
3.32061827e-01 -9.37099099e-01 3.94273520e-01 -1.73381090e+00
4.77862716e-01 4.37086612e-01 -1.24826767e-01 7.14422762e-01
2.12552220e-01 6.20238185e-01 3.15923303e-01 3.98744553e-01
-9.16319311e-01 6.09293103e-01 1.03532112e+00 -3.80852669e-01
-3.16205844e-02 -4.29524928e-01 -6.15531027e-01 3.98450404e-01
4.62272763e-01 -5.07908225e-01 -5.27362883e-01 -5.76611996e-01
-1.13798879e-01 5.95142305e-01 5.69426954e-01 -9.32177484e-01
3.43426108e-01 -8.67043883e-02 3.21745157e-01 -4.75252599e-01
8.07445943e-01 -9.16833699e-01 4.40662771e-01 -5.31985909e-02
-5.24255872e-01 9.57832858e-03 4.25129861e-01 7.10671067e-01
-8.00721288e-01 2.87023801e-02 3.37901086e-01 -1.03847168e-01
-1.12783015e+00 3.79961193e-01 -7.62518823e-01 -4.79400791e-02
1.18828845e+00 -3.29972804e-01 -5.02093554e-01 -9.08396125e-01
-1.12477136e+00 1.62571654e-01 3.33177119e-01 7.89064527e-01
8.36129427e-01 -1.58762515e+00 -7.61398733e-01 -1.20653287e-01
3.38245869e-01 -4.19346660e-01 3.78884524e-01 7.49316633e-01
-2.86663592e-01 7.22115517e-01 -3.64221603e-01 -6.54822350e-01
-1.35327661e+00 7.53426015e-01 -4.73905494e-03 3.78768981e-01
-6.93777800e-01 5.31308889e-01 5.07444680e-01 4.73619163e-01
4.30741280e-01 -1.97000101e-01 -3.48015040e-01 2.08650529e-01
8.47619593e-01 -4.89716977e-02 -2.15183869e-01 -1.08024728e+00
-5.03817677e-01 3.56827110e-01 9.12354514e-02 -3.56367290e-01
1.27583647e+00 -5.94475448e-01 3.76237839e-01 4.00095969e-01
1.49261212e+00 -2.72602081e-01 -1.35237706e+00 -1.99547056e-02
-3.68652731e-01 -3.59593928e-01 -2.69095041e-02 -5.97479641e-01
-1.00526667e+00 8.89891982e-01 4.42087233e-01 7.13534746e-03
1.21628213e+00 5.21480381e-01 8.80508602e-01 1.63261935e-01
1.92395031e-01 -7.23964095e-01 3.07570964e-01 3.60567957e-01
1.13856864e+00 -1.27769530e+00 -4.47770029e-01 -4.15433377e-01
-9.65086818e-01 1.04089212e+00 5.31150937e-01 3.53047311e-01
2.56186873e-01 2.82316297e-01 4.78461012e-02 2.67829839e-02
-9.75507259e-01 -4.01557416e-01 1.73982710e-01 7.44353712e-01
2.16898277e-01 -3.94344896e-01 2.55016059e-01 6.97805166e-01
4.00501490e-02 1.01776943e-01 9.31437016e-01 9.01921868e-01
-3.19728076e-01 -7.18373775e-01 -6.50871038e-01 7.99829066e-02
-4.40019190e-01 -3.31777006e-01 -3.65421087e-01 4.52252090e-01
-2.60269642e-01 1.09674931e+00 3.45418841e-01 -4.73298937e-01
1.05037808e-01 8.82761627e-02 2.36959115e-01 -7.70975351e-01
-2.52383977e-01 3.12797129e-01 1.22575328e-01 -7.27451801e-01
-6.38945580e-01 -6.49238169e-01 -1.26280808e+00 2.94796508e-02
1.91714197e-01 4.28970814e-01 5.78309059e-01 8.58490109e-01
4.96586055e-01 6.22819364e-01 3.72402668e-01 -1.28464997e+00
1.39996797e-01 -8.96328509e-01 -3.26085031e-01 4.46979582e-01
7.62957692e-01 -5.27285755e-01 -6.65768921e-01 8.14001918e-01]
|
[10.518889427185059, 0.8202311396598816]
|
d35832bf-49e9-4401-99cd-454e9558af8e
|
hierarchical-bayesian-inference-for-community
|
2301.07386
| null |
https://arxiv.org/abs/2301.07386v1
|
https://arxiv.org/pdf/2301.07386v1.pdf
|
Hierarchical Bayesian inference for community detection and connectivity of functional brain networks
|
Many functional magnetic resonance imaging (fMRI) studies rely on estimates of hierarchically organised brain networks whose segregation and integration reflect the dynamic transitions of latent cognitive states. However, most existing methods for estimating the community structure of networks from both individual and group-level analysis neglect the variability between subjects and lack validation. In this paper, we develop a new multilayer community detection method based on Bayesian latent block modelling. The method can robustly detect the group-level community structure of weighted functional networks that give rise to hidden brain states with an unknown number of communities and retain the variability of individual networks. For validation, we propose a new community structure-based multivariate Gaussian generative model convolved with haemodynamic response function to simulate synthetic fMRI signal. Our result shows that the inferred community memberships using hierarchical Bayesian analysis are consistent with the predefined node labels in the generative model. The method is also tested using real working memory task-fMRI data of 100 unrelated healthy subjects from the Human Connectome Project. The results show distinctive community structures and subtle connectivity patterns between 2-back, 0-back, and fixation conditions, which may reflect cognitive and behavioural states under working memory task conditions.
|
['Adeel Razi', 'Jonathan Keith', 'Leonardo Novelli', 'Nizhuan Wang', 'Lingbin Bian']
|
2023-01-18
| null | null | null | null |
['community-detection']
|
['graphs']
|
[ 1.73179463e-01 4.08099405e-02 3.43094736e-01 -3.76811206e-01
4.79676604e-01 -2.77950615e-01 5.60026944e-01 -1.74892634e-01
-2.32972309e-01 6.31228864e-01 3.69196385e-01 -1.50177136e-01
-5.18839598e-01 -7.20397949e-01 -8.51342976e-02 -8.39759469e-01
-7.78318048e-01 6.02942526e-01 4.44357336e-01 2.54125834e-01
1.81013554e-01 2.10224688e-01 -1.05068576e+00 5.88695586e-01
8.53367925e-01 1.51387155e-01 3.01140487e-01 4.64347303e-01
1.70350894e-01 5.32622516e-01 -3.93871665e-01 -1.25438878e-02
-8.93978216e-03 -5.30193269e-01 -5.45061827e-01 1.73713550e-01
-2.70063151e-02 3.50628840e-03 -3.94789398e-01 1.23975742e+00
4.09747988e-01 4.03528437e-02 9.24407959e-01 -1.17446554e+00
-4.33587343e-01 7.72279799e-01 -6.75899863e-01 5.28035402e-01
7.03452481e-03 1.04562230e-01 6.73156679e-01 -6.98798716e-01
7.41210938e-01 1.44799924e+00 6.89686358e-01 1.31121099e-01
-1.93244648e+00 -8.50149214e-01 1.64641336e-01 3.61498535e-01
-1.59464729e+00 -4.30722445e-01 6.73182249e-01 -1.22551036e+00
8.64841878e-01 -1.18535757e-01 1.02882755e+00 1.09821057e+00
6.97160006e-01 -1.98906258e-01 1.66482520e+00 -3.03581785e-02
2.72439390e-01 -1.87637508e-01 4.66279656e-01 8.70930076e-01
6.35670185e-01 8.79625045e-03 -4.76161063e-01 -7.37597525e-01
8.67609739e-01 -9.18390527e-02 -2.29130447e-01 -6.60118222e-01
-1.55220246e+00 9.25665736e-01 3.87935847e-01 6.25362754e-01
-4.53182101e-01 -1.37661204e-01 2.32624218e-01 8.88889730e-02
5.41602910e-01 -2.87211239e-01 1.83273837e-01 6.87960327e-01
-1.50102890e+00 -3.02218378e-01 6.67512059e-01 3.87222230e-01
6.83200002e-01 4.47266549e-02 -2.27036804e-01 5.82689464e-01
9.18737471e-01 1.84217677e-01 4.97112662e-01 -1.09587955e+00
-1.32516116e-01 3.63064140e-01 -3.27710509e-01 -1.49816811e+00
-5.84101617e-01 -6.32191598e-01 -1.43043280e+00 3.27281326e-01
6.76207691e-02 -6.59551844e-02 -4.32007045e-01 1.70884311e+00
1.45789102e-01 9.03963894e-02 -4.11347896e-01 6.64884746e-01
5.50957382e-01 1.02266639e-01 1.25822768e-01 -7.35124350e-01
1.31426370e+00 -4.82699007e-01 -8.65475118e-01 -1.07897192e-01
-4.51969579e-02 -1.48654759e-01 2.44776770e-01 1.36653200e-01
-9.79578674e-01 -6.32704020e-01 -9.37237024e-01 7.03339279e-01
1.80279836e-01 -3.96057844e-01 5.81309557e-01 7.40996897e-01
-1.36903489e+00 4.86870080e-01 -1.12890148e+00 -2.70560354e-01
4.23374444e-01 1.48813292e-01 -5.71139514e-01 -1.11560509e-01
-1.16011608e+00 6.43814266e-01 5.47867894e-01 4.28035110e-01
-1.13451362e+00 -1.88778728e-01 -4.38182443e-01 6.54358044e-02
-2.02401921e-01 -9.50993657e-01 1.82152882e-01 -9.06085193e-01
-8.89604211e-01 7.31142759e-01 -3.07916313e-01 -1.05053216e-01
3.24666589e-01 6.54843390e-01 -3.38209450e-01 4.59308892e-01
2.90571511e-01 7.02969909e-01 8.71199787e-01 -9.92274582e-01
4.94281709e-01 -4.59847063e-01 -7.36502290e-01 2.60700230e-02
-7.85833746e-02 1.28386885e-01 4.33822185e-01 -3.79317343e-01
9.12748516e-01 -6.84354603e-01 -2.62654513e-01 3.17041576e-03
-3.62401009e-01 1.05290897e-01 1.53047502e-01 -7.79508591e-01
1.09662879e+00 -2.20377541e+00 1.97086066e-01 6.95301592e-01
1.11858344e+00 -4.55773413e-01 1.83042157e-02 3.71376306e-01
-4.14637923e-01 3.17963630e-01 -3.88591975e-01 1.59688354e-01
-1.51264787e-01 -7.14272633e-02 3.19677413e-01 1.01369143e+00
-1.30276933e-01 7.94861913e-01 -8.53975475e-01 -6.74413025e-01
-4.29976642e-01 4.47000504e-01 -5.66770077e-01 -7.30452687e-02
4.57057089e-01 7.33121693e-01 3.83383851e-03 -2.22950354e-02
6.85178399e-01 -5.98833025e-01 1.03348029e+00 -2.97981054e-01
-6.76655322e-02 -7.49451146e-02 -9.89037275e-01 1.29064655e+00
7.62307644e-01 7.96070933e-01 4.73080575e-01 -1.18981218e+00
8.89436364e-01 4.95274842e-01 3.98469985e-01 -2.42817581e-01
-9.25275162e-02 -1.91887900e-01 1.17214036e+00 -3.40072930e-01
-4.16369200e-01 -4.12587136e-01 3.41575444e-01 7.94519007e-01
2.40324780e-01 4.55357850e-01 1.68267637e-01 5.31483829e-01
1.46312714e+00 -2.80703366e-01 2.75258482e-01 -8.36213887e-01
2.09936365e-01 -4.69130814e-01 6.56514525e-01 7.73561001e-01
-6.50796413e-01 2.01012969e-01 8.89507711e-01 1.75940164e-03
-1.07146025e+00 -1.40261352e+00 -2.55260348e-01 6.31678343e-01
-4.59611624e-01 -1.63371027e-01 -6.61723018e-01 1.23041168e-01
-2.76168197e-01 1.18013203e-01 -7.81558990e-01 -2.61168838e-01
-2.40707889e-01 -1.33931780e+00 4.71208841e-01 2.50732452e-02
4.71563905e-01 -1.07889676e+00 -5.29152036e-01 3.17089796e-01
-4.85253304e-01 -8.13692093e-01 -2.42975876e-01 -9.27747935e-02
-1.25910163e+00 -1.17998099e+00 -5.34279764e-01 -6.50678813e-01
8.06118786e-01 1.75439864e-01 8.43013406e-01 9.68839899e-02
-6.03353143e-01 2.36158550e-01 2.24326327e-02 3.66664261e-01
-4.13787603e-01 -3.17555696e-01 2.78189629e-01 2.51540095e-01
8.90149474e-02 -1.17287457e+00 -7.80710161e-01 4.16555434e-01
-6.61542058e-01 3.51672500e-01 5.57434320e-01 6.85271978e-01
1.96215793e-01 3.95145953e-01 5.35367966e-01 -5.72730422e-01
8.94265890e-01 -9.32495832e-01 -2.07767144e-01 1.73750997e-01
-7.02105105e-01 5.75935878e-02 -3.33230227e-01 -6.60471320e-01
-1.03872836e+00 -3.15858394e-01 5.70859909e-01 -1.85824916e-01
-1.75401255e-01 7.90718853e-01 -4.35730927e-02 2.43816480e-01
7.28847384e-01 3.15663993e-01 4.39371496e-01 -1.72098309e-01
1.26759216e-01 3.77186596e-01 1.58087954e-01 -3.58775318e-01
4.72318590e-01 6.81683481e-01 6.04891479e-02 -8.30705762e-01
-1.99379489e-01 -1.98417857e-01 -1.37270486e+00 -6.10110998e-01
8.93438697e-01 -9.74163830e-01 -6.80265486e-01 2.89056152e-01
-1.01468182e+00 -3.34644198e-01 4.02176112e-01 7.11145878e-01
-2.18259081e-01 9.48458850e-01 -8.38479638e-01 -9.39450681e-01
-2.03611001e-01 -8.49216461e-01 1.76109850e-01 -4.16241407e-01
-5.14291286e-01 -1.21411717e+00 4.66311157e-01 2.28347495e-01
5.26841104e-01 5.53661823e-01 1.14667130e+00 -1.16952039e-01
-3.78324926e-01 1.76896662e-01 -2.40019277e-01 2.01044418e-02
3.89001481e-02 5.67057952e-02 -6.38887227e-01 -4.25827652e-01
2.72012413e-01 2.73862988e-01 8.58996987e-01 6.56436741e-01
3.64044458e-01 -2.56959945e-01 -5.36459625e-01 1.56475201e-01
1.05495369e+00 -2.01366022e-02 7.43396521e-01 -2.91348547e-01
4.46762949e-01 1.20712173e+00 -4.08468962e-01 1.57489747e-01
1.04112156e-01 1.97342902e-01 6.78963065e-02 2.30614632e-01
-2.86178607e-02 2.48856574e-01 5.22204041e-01 1.41842604e+00
-2.19335407e-01 3.22872519e-01 -1.19894195e+00 4.27572042e-01
-1.91865265e+00 -1.48888087e+00 -7.47190535e-01 2.00134802e+00
8.73796880e-01 3.27630997e-01 1.96731389e-01 -2.06058398e-01
1.24081731e+00 -1.21845290e-01 -4.17189509e-01 3.06969106e-01
-2.75428683e-01 1.87465851e-03 -1.35637373e-01 4.96882617e-01
-5.35892904e-01 4.28078711e-01 7.50077105e+00 2.41578326e-01
-5.02450049e-01 6.63451374e-01 4.50684011e-01 -1.66908920e-01
-2.55038410e-01 1.34465128e-01 -1.24501295e-01 3.43923151e-01
1.23322499e+00 -2.05606490e-01 4.99517441e-01 2.59866454e-02
5.51800847e-01 -2.45260298e-01 -7.18802333e-01 6.51552618e-01
-6.71983287e-02 -8.69246364e-01 -3.70439589e-01 4.85796690e-01
5.28183699e-01 1.76883176e-01 -3.46613467e-01 -8.45044404e-02
3.31026018e-01 -1.00623417e+00 6.38707697e-01 1.27672184e+00
8.24139953e-01 -1.45672932e-01 4.48594332e-01 7.61464059e-01
-1.04170918e+00 5.20947687e-02 -5.06382346e-01 -3.43821496e-01
2.78539002e-01 9.70097482e-01 -6.90577209e-01 3.89832221e-02
6.83427095e-01 6.20134175e-01 -7.51300156e-01 1.10319912e+00
-1.83331110e-02 7.94480562e-01 -7.27046803e-02 3.27446222e-01
-3.46823275e-01 -4.06589895e-01 6.20893657e-01 9.78719234e-01
9.78770927e-02 -9.36307833e-02 -1.13572413e-03 1.65225744e+00
5.11843503e-01 -1.53711289e-01 -5.08024991e-01 -3.22443604e-01
3.07103992e-01 1.49748480e+00 -1.54259944e+00 -2.14737117e-01
-1.02234647e-01 7.18568921e-01 4.67217535e-01 4.45385665e-01
-4.21844214e-01 9.17170346e-02 8.08831230e-02 6.41149700e-01
7.21425563e-02 -7.42070675e-01 4.08389233e-02 -1.31878746e+00
-4.07041281e-01 -3.80818784e-01 7.21941963e-02 -8.96955848e-01
-1.77274060e+00 6.10975862e-01 5.52206635e-01 -5.12693882e-01
4.41502966e-02 -2.68382460e-01 -7.28490353e-01 1.25824153e+00
-7.09329247e-01 -6.40333176e-01 -3.95851374e-01 6.97739184e-01
-1.71823636e-01 -2.29088113e-01 8.89002681e-01 2.73624897e-01
-6.07396483e-01 -1.65945217e-01 -1.31585123e-02 3.54371279e-01
5.12931287e-01 -8.46737564e-01 8.52830335e-02 8.24559033e-01
-2.22285837e-01 1.25257087e+00 4.43956405e-01 -1.28873396e+00
-3.27228159e-01 -7.32840717e-01 5.99696338e-01 -3.15443337e-01
9.40631092e-01 -8.52768838e-01 -1.07947946e+00 5.45752883e-01
4.00690079e-01 -4.32502814e-02 1.07008004e+00 9.76521056e-03
-2.81322688e-01 4.46456909e-01 -9.68994558e-01 3.06965590e-01
1.24391258e+00 -5.93937814e-01 -6.98982000e-01 2.59129554e-01
2.75469244e-01 8.54081690e-01 -1.06056917e+00 2.85752386e-01
8.00646424e-01 -1.17707849e+00 7.99947500e-01 -5.26547194e-01
-1.39007911e-01 -3.95720840e-01 8.28104615e-02 -1.09555292e+00
-1.41617870e+00 -1.05147198e-01 -8.11761171e-02 9.01931763e-01
4.13380703e-03 -7.74579406e-01 4.48036760e-01 3.14338088e-01
3.22633475e-01 1.07300840e-02 -1.02164781e+00 -5.32620549e-01
6.05248325e-02 1.20051272e-01 -8.61851647e-02 1.29310155e+00
8.89721736e-02 6.93521976e-01 1.67028625e-02 -1.55901480e-02
1.30482721e+00 -2.78689802e-01 -5.09886816e-02 -1.89512277e+00
-3.93585324e-01 -6.04856908e-01 -4.46988821e-01 2.13641542e-04
3.99872452e-01 -1.12786150e+00 -1.20766625e-01 -1.61663234e+00
9.55598235e-01 -2.87467837e-02 -3.02467108e-01 2.70019799e-01
3.78691941e-04 1.17967889e-01 -1.85143679e-01 7.62491584e-01
-5.17169237e-01 3.74253631e-01 1.01957870e+00 -1.09954901e-01
1.38433143e-01 -3.29007387e-01 -5.12381554e-01 6.34137273e-01
7.45070517e-01 -7.77053177e-01 -4.24717247e-01 2.52026886e-01
5.30272841e-01 1.26414239e-01 7.21789896e-01 -9.77657795e-01
2.40664348e-01 7.87029341e-02 6.44305110e-01 -3.19809735e-01
1.23236582e-01 -5.44997334e-01 9.67865407e-01 9.61406529e-01
-7.57782012e-02 3.84369381e-02 -2.53545135e-01 8.65813553e-01
1.17221087e-01 -2.80956060e-01 7.03016877e-01 -5.55968344e-01
5.04634622e-03 1.11992277e-01 -1.10182989e+00 -3.09693933e-01
7.32883692e-01 -3.11727464e-01 -3.08348149e-01 -2.38227248e-01
-1.52336168e+00 -1.21114038e-01 7.52359480e-02 -1.49173766e-01
7.41386831e-01 -1.30915940e+00 -1.05424476e+00 2.64305830e-01
-4.07509714e-01 -7.55583882e-01 4.77219284e-01 1.49222827e+00
-4.02518839e-01 1.94657162e-01 -8.10026407e-01 -7.73544133e-01
-1.27122617e+00 4.06869948e-01 5.32661736e-01 -2.98669636e-01
-4.50555027e-01 4.63051558e-01 4.33959752e-01 -3.48859072e-01
-2.86890596e-01 2.03570321e-01 -5.33646047e-01 5.10785937e-01
4.02909189e-01 5.29711187e-01 -3.80385965e-01 -8.08414578e-01
-2.91861832e-01 -8.81084800e-02 2.81609327e-01 -3.74648064e-01
1.21731746e+00 -4.31820929e-01 -9.89670455e-01 9.07501101e-01
7.73452818e-01 -2.48617217e-01 -1.01554239e+00 -1.80023760e-01
1.02262512e-01 -7.46057928e-02 7.75095671e-02 -3.84849817e-01
-9.12028611e-01 1.17799032e+00 9.90382791e-01 4.19275880e-01
5.51358640e-01 -6.08571246e-02 -2.46932998e-01 6.12478834e-05
4.59813088e-01 -5.95037401e-01 2.45161176e-01 1.68683380e-01
8.01137209e-01 -7.23109841e-01 2.22783744e-01 -2.67265528e-01
-1.87015638e-01 9.01844680e-01 3.41347665e-01 -2.89713562e-01
1.24244452e+00 -4.40315865e-02 -6.36578619e-01 -8.53521347e-01
-8.71474385e-01 1.88537955e-01 3.52861732e-01 7.83622324e-01
6.13256395e-01 2.35962898e-01 -3.55085403e-01 6.11103833e-01
6.82231560e-02 -2.94185370e-01 5.07714748e-01 5.86600125e-01
-7.22568691e-01 -6.67312801e-01 -5.17319083e-01 6.58876240e-01
-2.15708822e-01 -3.88811678e-01 -3.39444518e-01 3.79178822e-01
1.27760276e-01 8.30878317e-01 1.82550281e-01 -1.35713592e-02
-2.21614808e-01 6.63004935e-01 7.05594599e-01 -1.00447476e+00
-2.48177484e-01 5.85043550e-01 -1.76287904e-01 -1.89652056e-01
-8.78793836e-01 -1.11185670e+00 -9.17344451e-01 -3.18115175e-01
-4.17720795e-01 2.19427168e-01 1.77086189e-01 7.35618353e-01
3.24157178e-01 7.74222255e-01 1.10877618e-01 -9.15755093e-01
1.27016068e-01 -1.65051556e+00 -9.62128758e-01 1.70886606e-01
-1.24589376e-01 -9.43117857e-01 -5.06448925e-01 3.43828797e-01]
|
[12.399659156799316, 3.4270384311676025]
|
30b3dd28-bc9b-4c17-9a1c-1d8b01ce8142
|
keyphrase-generation-with-fine-grained
|
2104.08799
| null |
https://arxiv.org/abs/2104.08799v2
|
https://arxiv.org/pdf/2104.08799v2.pdf
|
Keyphrase Generation with Fine-Grained Evaluation-Guided Reinforcement Learning
|
Aiming to generate a set of keyphrases, Keyphrase Generation (KG) is a classical task for capturing the central idea from a given document. Based on Seq2Seq models, the previous reinforcement learning framework on KG tasks utilizes the evaluation metrics to further improve the well-trained neural models. However, these KG evaluation metrics such as $F_1@5$ and $F_1@M$ are only aware of the exact correctness of predictions on phrase-level and ignore the semantic similarities between similar predictions and targets, which inhibits the model from learning deep linguistic patterns. In response to this problem, we propose a new fine-grained evaluation metric to improve the RL framework, which considers different granularities: token-level $F_1$ score, edit distance, duplication, and prediction quantities. On the whole, the new framework includes two reward functions: the fine-grained evaluation score and the vanilla $F_1$ score. This framework helps the model identifying some partial match phrases which can be further optimized as the exact match ones. Experiments on KG benchmarks show that our proposed training framework outperforms the previous RL training frameworks among all evaluation scores. In addition, our method can effectively ease the synonym problem and generate a higher quality prediction. The source code is available at \url{https://github.com/xuyige/FGRL4KG}.
|
['Qi Zhang', 'Xipeng Qiu', 'Jiacheng Ye', 'Yige Xu', 'Yichao Luo']
|
2021-04-18
| null |
https://aclanthology.org/2021.findings-emnlp.45
|
https://aclanthology.org/2021.findings-emnlp.45.pdf
|
findings-emnlp-2021-11
|
['keyphrase-generation']
|
['natural-language-processing']
|
[-1.87896147e-01 2.16761194e-02 -3.60142827e-01 -2.57496119e-01
-9.72224951e-01 -4.72038597e-01 4.75272387e-01 2.29175732e-01
-5.55854917e-01 1.01942301e+00 4.27947074e-01 -5.29293828e-02
-8.31000730e-02 -1.12826431e+00 -8.92116547e-01 -5.68989992e-01
2.18488753e-01 1.48597494e-01 2.59546731e-02 -4.77349490e-01
5.99907219e-01 -4.47907969e-02 -1.38191831e+00 3.48165542e-01
1.10978818e+00 9.93168414e-01 5.16983986e-01 3.72681350e-01
-5.23163974e-01 6.62128687e-01 -6.74186647e-01 -6.85496330e-01
1.57786906e-01 -5.39703727e-01 -5.49599528e-01 -6.78943872e-01
1.13151297e-01 -3.96198630e-01 -5.94934039e-02 1.21195138e+00
5.96232116e-01 2.57213444e-01 3.40710163e-01 -1.18254232e+00
-9.00288701e-01 1.35293341e+00 -3.38578641e-01 1.72393247e-01
4.52262074e-01 1.90035820e-01 1.68174231e+00 -9.14700806e-01
4.55402315e-01 1.05498934e+00 4.45247859e-01 4.48644847e-01
-7.51011610e-01 -9.46302176e-01 3.80967081e-01 3.95952016e-01
-1.31714082e+00 -1.15521038e-02 5.90118229e-01 -5.61902523e-02
1.01647699e+00 3.19661424e-02 4.27459061e-01 9.51720655e-01
1.64668739e-01 1.07190883e+00 9.05937850e-01 -3.55339646e-01
8.29316750e-02 1.21493964e-02 6.30462989e-02 6.20380521e-01
1.09178662e-01 1.61235973e-01 -6.45736516e-01 4.54079621e-02
4.42010731e-01 -3.02334633e-02 -4.06353205e-01 1.05946988e-01
-1.39268637e+00 9.17594254e-01 4.64689672e-01 3.90072972e-01
-4.11659062e-01 3.80465150e-01 3.25644165e-01 6.24442510e-02
2.26797312e-01 8.46468151e-01 -7.47649848e-01 -4.19802547e-01
-7.08664596e-01 5.44566572e-01 5.78859568e-01 9.64993894e-01
9.51238871e-01 -6.07884452e-02 -8.48548055e-01 7.29412735e-01
1.20211944e-01 5.01941383e-01 8.17708969e-01 -5.58388829e-01
7.19177842e-01 5.84850907e-01 1.23503037e-01 -1.01765001e+00
-1.91421241e-01 -6.69972420e-01 -7.77972937e-01 -3.98627669e-01
1.73283685e-02 -2.86057740e-01 -6.09031379e-01 2.06435251e+00
5.83821982e-02 2.17049941e-01 7.27131069e-02 7.13610590e-01
8.88788283e-01 9.89923716e-01 4.26726080e-02 -2.95294285e-01
1.03752291e+00 -1.19383013e+00 -8.28574419e-01 -1.28715232e-01
7.62122512e-01 -8.09095740e-01 1.30026329e+00 2.71715224e-01
-1.02763367e+00 -7.62048483e-01 -9.40404356e-01 1.78046882e-01
-4.70446378e-01 4.18775350e-01 5.14338374e-01 3.35857064e-01
-8.03292990e-01 8.22426736e-01 -1.93289444e-01 1.10682704e-01
1.44910827e-01 2.07819358e-01 9.53387320e-02 4.70186733e-02
-1.91945183e+00 7.13665307e-01 8.87222707e-01 1.11750722e-01
-5.39177477e-01 -7.70170867e-01 -5.87895572e-01 2.91956782e-01
6.59499705e-01 -6.30382538e-01 1.31130552e+00 -6.50093079e-01
-1.58409715e+00 3.24023962e-01 1.16023170e-02 -4.78993684e-01
3.22744519e-01 -4.86844599e-01 -4.32185411e-01 -1.11294538e-01
3.18516850e-01 8.15529346e-01 7.49798775e-01 -8.50834489e-01
-1.02249706e+00 6.58170357e-02 1.62477195e-01 4.85058457e-01
-3.04839969e-01 -2.71343082e-01 -2.60781199e-01 -1.01977050e+00
-4.50713694e-01 -5.52430153e-01 -9.85961482e-02 -6.75665736e-01
-5.24414599e-01 -5.12553334e-01 2.65174285e-02 -6.42468870e-01
1.66934371e+00 -1.88263595e+00 9.98260304e-02 2.17425272e-01
-4.88139540e-02 4.06510204e-01 -3.63206267e-01 6.68902516e-01
1.09079681e-01 4.78078425e-01 1.70612009e-03 5.61499260e-02
2.48379543e-01 -1.19731545e-01 -5.10737836e-01 -4.15104121e-01
2.01692700e-01 1.14511061e+00 -1.12706792e+00 -2.86916137e-01
-1.97251618e-01 -4.53208480e-03 -6.00692034e-01 4.09636497e-01
-6.20155036e-01 3.23957019e-02 -6.31036103e-01 2.60210454e-01
5.17811239e-01 -1.04028188e-01 -1.03690900e-01 -3.32164973e-01
-8.38868171e-02 5.53708613e-01 -1.24697900e+00 1.75632596e+00
-4.89646494e-01 -1.04563236e-01 -5.71446896e-01 -1.00404525e+00
1.20525050e+00 5.41423969e-02 2.81514198e-01 -8.74173462e-01
6.82013249e-03 2.64331073e-01 -3.66191976e-02 -1.34469911e-01
8.51724803e-01 2.06136946e-02 -2.58655697e-01 6.63045585e-01
5.13128098e-03 -5.79182478e-03 4.45229411e-01 3.31585288e-01
9.53446269e-01 3.27480823e-01 4.30060118e-01 3.45729291e-03
6.08866394e-01 -1.90783247e-01 6.59125745e-01 9.68683898e-01
7.75020346e-02 2.89257646e-01 6.02680922e-01 -2.48765841e-01
-8.76884758e-01 -8.45367849e-01 3.43741834e-01 1.21886563e+00
2.53428638e-01 -8.17103326e-01 -6.83911979e-01 -8.75478745e-01
-2.68063089e-03 1.00637293e+00 -5.87374449e-01 -4.99280185e-01
-5.63297987e-01 -6.73210025e-01 6.98192835e-01 4.26792026e-01
6.27526104e-01 -1.38473797e+00 -2.67631680e-01 3.57014298e-01
-4.75416392e-01 -8.74661863e-01 -7.26031959e-01 5.75330108e-02
-4.79899198e-01 -7.46732950e-01 -6.77951396e-01 -5.50871074e-01
3.39593112e-01 -2.64455695e-02 1.17270303e+00 1.05199970e-01
8.33471641e-02 3.36518995e-02 -8.54350328e-01 -4.73526329e-01
-3.66425335e-01 2.75457114e-01 -1.37063116e-01 -1.46280512e-01
4.96089518e-01 -3.23939264e-01 -6.66295350e-01 -1.92770772e-02
-8.01005304e-01 1.90271720e-01 9.86713409e-01 1.05247009e+00
7.77351499e-01 2.04289541e-01 9.25550461e-01 -8.42865884e-01
1.27166533e+00 -4.18508887e-01 -3.77842486e-01 4.84541893e-01
-9.31160271e-01 4.98580813e-01 1.04210615e+00 -3.32271099e-01
-9.41608429e-01 -5.26849210e-01 -2.65573502e-01 -2.35072821e-01
6.09077364e-02 7.98467219e-01 -1.07568197e-01 3.90624851e-01
4.31011617e-01 7.25219488e-01 -2.56054997e-01 -3.51094127e-01
5.43317497e-01 3.03537905e-01 2.50345111e-01 -7.30690062e-01
8.22624981e-01 -2.19138592e-01 -3.27012688e-01 -7.24703893e-02
-1.01233113e+00 -2.12241352e-01 -3.27198386e-01 7.54776448e-02
3.83956790e-01 -7.38037884e-01 -6.89076662e-01 2.68440127e-01
-1.22576714e+00 -1.23957686e-01 -4.20958906e-01 4.11841422e-01
-5.39361596e-01 4.70795304e-01 -4.03789103e-01 -4.42572445e-01
-8.60246539e-01 -1.13580871e+00 9.19708669e-01 3.98121864e-01
-1.17091671e-01 -5.99437654e-01 6.55575991e-02 2.29806062e-02
3.59624177e-01 -6.94034323e-02 1.14036131e+00 -8.79048228e-01
-7.01849639e-01 3.93048860e-02 -1.37824118e-01 5.44596136e-01
2.09865734e-01 -4.93448228e-02 -4.20279980e-01 -3.34622934e-02
-2.76949823e-01 -3.38608474e-01 9.92020309e-01 2.67297179e-01
1.23437202e+00 -6.67473495e-01 -8.84956419e-02 2.29850501e-01
1.39936221e+00 3.53742808e-01 6.06304407e-01 4.53803480e-01
5.94460547e-01 3.41754913e-01 1.07295704e+00 7.20179796e-01
3.68976325e-01 6.40869558e-01 2.91416347e-01 2.73836493e-01
9.03913453e-02 -7.63698280e-01 4.86661494e-01 8.07388067e-01
1.32642701e-01 -2.22220510e-01 -6.91379964e-01 3.28710258e-01
-1.91016006e+00 -1.07200193e+00 3.76020938e-01 2.11301064e+00
1.20917320e+00 4.37588930e-01 -1.62463188e-02 -3.60563397e-02
7.43226290e-01 1.87444612e-01 -5.06338000e-01 -5.47312498e-01
2.87397280e-02 4.89970535e-01 3.11430454e-01 3.73080909e-01
-7.71562636e-01 1.24224997e+00 4.85976267e+00 1.40446198e+00
-9.61633384e-01 -3.23502123e-01 5.47700584e-01 -8.98952037e-02
-6.90920234e-01 -5.17288260e-02 -1.30789948e+00 7.62767613e-01
6.90589309e-01 -6.54870450e-01 6.05225444e-01 8.26371014e-01
2.89421588e-01 3.11461091e-02 -1.01449335e+00 7.52616644e-01
-3.64170223e-02 -1.51907897e+00 5.49642563e-01 -2.57699311e-01
6.41606331e-01 -1.84862241e-01 1.84162319e-01 7.03181922e-01
4.79948163e-01 -9.61517632e-01 6.83854759e-01 6.98262572e-01
4.78029102e-01 -1.12639976e+00 8.41901064e-01 4.53316808e-01
-1.27377188e+00 -6.71536922e-02 -5.37324190e-01 1.27750337e-01
2.06857175e-02 7.31193066e-01 -1.01823068e+00 8.86974812e-01
6.31760538e-01 8.20808411e-01 -5.54086268e-01 7.47012377e-01
-6.55754626e-01 3.46688539e-01 4.95670848e-02 -6.28175199e-01
4.83360827e-01 -7.45448843e-02 4.27638054e-01 1.22890806e+00
5.46544969e-01 8.06288794e-02 1.77341640e-01 1.00450850e+00
-3.04242343e-01 5.29736817e-01 -2.55070895e-01 -1.64019525e-01
7.27745235e-01 1.34014714e+00 -3.55084419e-01 -3.19268346e-01
-1.64116576e-01 7.93522358e-01 5.33719063e-01 2.13844433e-01
-9.10392106e-01 -7.23224401e-01 6.00231409e-01 -2.64053702e-01
5.00276327e-01 1.46156013e-01 -5.28314225e-02 -1.12646723e+00
1.84268013e-01 -1.17676985e+00 3.17235410e-01 -7.18600035e-01
-1.22811043e+00 5.41942596e-01 -3.22329476e-02 -1.30963945e+00
-4.86837268e-01 -3.32183272e-01 -6.54880285e-01 9.35561717e-01
-1.69758940e+00 -9.39177632e-01 -6.20270818e-02 3.90353799e-01
5.56364238e-01 -2.79751927e-01 7.66819239e-01 5.19813299e-02
-5.31179786e-01 9.08492327e-01 1.90257832e-01 1.10323548e-01
7.93547630e-01 -1.32432818e+00 4.66577053e-01 8.55043173e-01
1.98188707e-01 7.35483587e-01 5.12357473e-01 -6.48452759e-01
-1.02657306e+00 -1.14564192e+00 9.57059324e-01 -3.40974666e-02
5.82761765e-01 -3.28628644e-02 -7.86687553e-01 3.27999502e-01
1.98531359e-01 -4.80441213e-01 6.26340151e-01 -4.81831320e-02
-3.22996497e-01 -3.26280594e-01 -7.89881527e-01 8.01539660e-01
9.03079867e-01 -2.86805272e-01 -5.41807175e-01 2.84457952e-01
1.16222262e+00 -2.80263633e-01 -9.35691595e-01 5.61739564e-01
4.55016136e-01 -9.90266800e-01 8.67421865e-01 -5.48119307e-01
8.73249412e-01 -1.68456942e-01 -4.83369716e-02 -1.61684251e+00
-4.34418470e-01 -4.89258766e-01 -1.64321825e-01 1.41834426e+00
6.50267184e-01 -4.89248961e-01 6.60413682e-01 2.16944560e-01
-1.14928037e-01 -1.17799926e+00 -5.61385453e-01 -7.09507227e-01
1.61943257e-01 -3.37680757e-01 1.04603589e+00 7.36123443e-01
2.82930005e-02 4.12743092e-01 -5.39640844e-01 -1.29294664e-01
3.87111694e-01 4.01111841e-01 6.90503955e-01 -7.86304712e-01
-5.42565584e-01 -7.94458687e-01 1.23959884e-01 -1.23725951e+00
1.87673226e-01 -1.11047685e+00 1.84649639e-02 -1.52005816e+00
3.58625688e-02 -4.90400195e-01 -6.44854546e-01 6.48705602e-01
-6.93131626e-01 -3.10956270e-01 2.36679614e-01 8.90453681e-02
-5.59945524e-01 8.49021792e-01 1.38200521e+00 -1.75328091e-01
-1.56350240e-01 1.86451171e-02 -9.83694494e-01 3.41184974e-01
1.16391563e+00 -4.72849101e-01 -3.68102998e-01 -3.32730889e-01
6.29182935e-01 -1.11412354e-01 1.46364689e-01 -7.25136042e-01
2.77551800e-01 -3.96324396e-01 2.80749947e-01 -6.81867719e-01
1.88617185e-02 -4.30425197e-01 -1.94075152e-01 5.37134588e-01
-8.10220063e-01 1.80022851e-01 1.49423957e-01 5.93454838e-01
-3.61975551e-01 -4.41629112e-01 3.46470624e-01 -4.36421424e-01
-9.54443216e-01 4.03902739e-01 2.64415964e-02 2.41101906e-01
8.37307930e-01 7.09264502e-02 -4.74428385e-01 -3.90815377e-01
-2.57581770e-01 3.84066939e-01 -7.21138492e-02 6.55792058e-01
7.11491823e-01 -1.45514035e+00 -9.15089071e-01 -1.28287394e-02
2.08617851e-01 2.53696963e-02 2.34128475e-01 3.99855435e-01
-1.62169397e-01 6.42454386e-01 -2.57145345e-01 -4.93179820e-02
-8.79212081e-01 5.01162469e-01 2.70049065e-01 -9.48835492e-01
-1.62139475e-01 9.90243793e-01 1.69973612e-01 -4.63027060e-01
2.08430320e-01 -4.87771839e-01 -4.18434143e-01 -3.23140696e-02
5.37006676e-01 2.36583397e-01 7.10012540e-02 -1.79236412e-01
-1.07523806e-01 4.08516407e-01 -5.51862061e-01 1.34744970e-02
1.08626711e+00 1.01723358e-01 -1.39084041e-01 2.16215074e-01
1.02442253e+00 1.17057040e-02 -7.79350996e-01 -5.18378913e-01
5.73192164e-02 -2.75506645e-01 -2.35312939e-01 -1.01068592e+00
-9.09801245e-01 7.84469485e-01 1.88354492e-01 2.33294517e-02
1.22059989e+00 -1.99071839e-01 9.86502349e-01 5.41490972e-01
2.13476017e-01 -1.29701948e+00 4.82973784e-01 7.07682490e-01
9.13549423e-01 -8.68868172e-01 -1.11489058e-01 -1.84818000e-01
-7.89698839e-01 1.03406525e+00 9.30471838e-01 1.03780546e-03
2.70261139e-01 -7.18349069e-02 -2.61232406e-01 2.44606078e-01
-9.39647436e-01 -1.32044032e-01 3.17158729e-01 2.44920492e-01
4.36557412e-01 5.72303012e-02 -8.19260836e-01 1.03396583e+00
-5.93899131e-01 -3.93258780e-02 3.55452955e-01 5.85150898e-01
-5.55901289e-01 -1.42994940e+00 7.87297487e-02 6.77232742e-01
-5.09198189e-01 -6.19375527e-01 -2.29516208e-01 4.70524728e-01
2.06151068e-01 6.40193820e-01 -2.57276535e-01 -7.41430819e-01
3.88707757e-01 6.27178624e-02 1.06783547e-01 -6.84039474e-01
-7.42809713e-01 -1.94309205e-01 -9.25631449e-02 -4.34421569e-01
-2.56668538e-01 -2.38664225e-01 -1.33248687e+00 -3.24223876e-01
-3.50466996e-01 5.40805697e-01 3.56755525e-01 8.77067327e-01
4.72362846e-01 5.81309438e-01 8.58750343e-01 -2.86126733e-01
-9.90941167e-01 -1.07177246e+00 -4.65355396e-01 3.96853566e-01
-4.68198061e-02 -3.86525154e-01 -8.06764439e-02 -2.52007097e-01]
|
[11.9387845993042, 8.81673526763916]
|
9890009b-e8fe-4572-beb8-1aa1bcfa75ab
|
revisiting-image-deblurring-with-an-efficient
|
2302.02234
| null |
https://arxiv.org/abs/2302.02234v1
|
https://arxiv.org/pdf/2302.02234v1.pdf
|
Revisiting Image Deblurring with an Efficient ConvNet
|
Image deblurring aims to recover the latent sharp image from its blurry counterpart and has a wide range of applications in computer vision. The Convolution Neural Networks (CNNs) have performed well in this domain for many years, and until recently an alternative network architecture, namely Transformer, has demonstrated even stronger performance. One can attribute its superiority to the multi-head self-attention (MHSA) mechanism, which offers a larger receptive field and better input content adaptability than CNNs. However, as MHSA demands high computational costs that grow quadratically with respect to the input resolution, it becomes impractical for high-resolution image deblurring tasks. In this work, we propose a unified lightweight CNN network that features a large effective receptive field (ERF) and demonstrates comparable or even better performance than Transformers while bearing less computational costs. Our key design is an efficient CNN block dubbed LaKD, equipped with a large kernel depth-wise convolution and spatial-channel mixing structure, attaining comparable or larger ERF than Transformers but with a smaller parameter scale. Specifically, we achieve +0.17dB / +0.43dB PSNR over the state-of-the-art Restormer on defocus / motion deblurring benchmark datasets with 32% fewer parameters and 39% fewer MACs. Extensive experiments demonstrate the superior performance of our network and the effectiveness of each module. Furthermore, we propose a compact and intuitive ERFMeter metric that quantitatively characterizes ERF, and shows a high correlation to the network performance. We hope this work can inspire the research community to further explore the pros and cons of CNN and Transformer architectures beyond image deblurring tasks.
|
['Bin Chen', 'Karol Myszkowski', 'Hans-Peter Seidel', 'Mojtaba Bemana', 'Lingyan Ruan']
|
2023-02-04
| null | null | null | null |
['deblurring']
|
['computer-vision']
|
[ 7.50834420e-02 -3.53111506e-01 -3.53071131e-02 -1.68826386e-01
-4.57084447e-01 -2.48795509e-01 3.82614553e-01 -6.12168849e-01
-4.72837090e-01 5.75779498e-01 4.42152441e-01 -3.16028744e-01
-1.16904236e-01 -4.39502507e-01 -6.78073823e-01 -1.07424366e+00
1.16898768e-01 -4.18350458e-01 2.60867357e-01 -5.65789752e-02
1.41545057e-01 4.50263143e-01 -1.36791146e+00 -1.41354606e-01
1.08171046e+00 1.25055313e+00 5.26019812e-01 4.76167768e-01
4.84420270e-01 8.92310858e-01 -3.46843213e-01 -2.75203615e-01
9.41472724e-02 -2.06787840e-01 -6.88530326e-01 -2.55435556e-01
7.03333318e-01 -8.16071272e-01 -8.41166496e-01 1.18819618e+00
7.32427955e-01 -3.73021816e-03 3.97515684e-01 -7.62036264e-01
-1.12201726e+00 4.92122263e-01 -7.91795254e-01 5.13645649e-01
-1.79108202e-01 2.41543084e-01 6.94179893e-01 -8.06531489e-01
1.14248477e-01 9.80618060e-01 7.12084413e-01 5.71022809e-01
-1.06202936e+00 -8.84115398e-01 -1.56685442e-01 5.18058002e-01
-1.36959088e+00 -4.72546160e-01 5.75722873e-01 -1.40690804e-02
9.20107543e-01 1.31155640e-01 1.94556400e-01 1.08446085e+00
2.86892802e-01 6.51160777e-01 1.16401398e+00 -3.91041413e-02
8.81375000e-02 -2.02929139e-01 -1.54896239e-02 4.99329746e-01
2.54714370e-01 2.47208685e-01 -4.52449143e-01 3.20607334e-01
1.17898297e+00 1.06527686e-01 -1.02843487e+00 -4.81212176e-02
-1.20609355e+00 5.31833231e-01 9.47517753e-01 3.71374220e-01
-3.24458748e-01 5.10916948e-01 1.91247165e-01 1.74546823e-01
4.67046946e-01 4.80841786e-01 -2.78358132e-01 -4.03710008e-02
-9.74760413e-01 -3.08993761e-03 2.69483387e-01 8.12170804e-01
5.66694438e-01 8.24710131e-02 -2.70396113e-01 8.94781649e-01
-1.46053368e-02 5.39179444e-01 6.28954709e-01 -1.00554860e+00
2.37659261e-01 2.12215737e-01 1.27366707e-01 -1.03769040e+00
-2.82367408e-01 -7.20272601e-01 -1.59571457e+00 6.07812218e-02
2.88705021e-01 1.41629413e-01 -7.15031385e-01 1.83839095e+00
-2.37477738e-02 3.58623534e-01 6.59913421e-02 1.38964057e+00
7.38375068e-01 6.01747453e-01 -2.45163471e-01 -1.56064272e-01
1.47347188e+00 -1.09023261e+00 -7.72196591e-01 -2.31492221e-01
1.31722629e-01 -8.33389342e-01 9.87583280e-01 2.85126209e-01
-1.12383437e+00 -5.90531349e-01 -1.10610116e+00 -3.51245046e-01
1.06834859e-01 1.91078797e-01 7.86757648e-01 6.09913945e-01
-1.35033417e+00 6.73238277e-01 -7.65280902e-01 -1.31005347e-01
5.30386508e-01 3.17382544e-01 -1.87321335e-01 -2.58365244e-01
-1.19199586e+00 8.19481134e-01 1.47920549e-01 4.22023773e-01
-9.27624702e-01 -8.87901664e-01 -7.46963024e-01 3.66562665e-01
1.26353666e-01 -8.24786484e-01 1.12035859e+00 -6.59614861e-01
-1.56439924e+00 4.03175563e-01 -9.73884836e-02 -5.94033957e-01
4.54663724e-01 -4.21049386e-01 -2.87057191e-01 2.56510735e-01
-1.61116123e-01 6.38089657e-01 1.26384640e+00 -9.12589371e-01
-3.96361440e-01 -2.36502871e-01 8.45641717e-02 6.71858862e-02
-6.44028366e-01 1.04722314e-01 -4.28972214e-01 -9.24537063e-01
1.08897537e-01 -6.76647425e-01 -3.35664190e-02 8.45578089e-02
-1.56557471e-01 6.85901800e-03 8.50275278e-01 -6.29851103e-01
1.38966215e+00 -2.23574471e+00 3.09699684e-01 -5.36511719e-01
7.14471698e-01 5.04269838e-01 -1.99206084e-01 4.18823063e-02
-2.50365406e-01 -4.48563173e-02 -2.05447167e-01 -2.31373280e-01
-2.20415622e-01 -2.00299844e-01 -5.09717345e-01 7.47839510e-01
8.17442015e-02 1.05691934e+00 -7.94874609e-01 -1.95614714e-02
2.51883626e-01 7.93859482e-01 -5.15885830e-01 3.14702153e-01
2.39873454e-01 3.21134329e-01 -2.79949427e-01 5.72440147e-01
1.21198297e+00 -6.03621244e-01 -1.92296833e-01 -7.33531058e-01
-3.61058742e-01 -7.17509389e-02 -6.61458969e-01 1.58223712e+00
-5.96528888e-01 1.05694616e+00 2.58475691e-01 -7.73987293e-01
7.50169158e-01 2.34735921e-01 1.42884031e-01 -7.45529234e-01
3.44094068e-01 4.05178457e-01 -1.40492365e-01 -6.12691760e-01
6.37040317e-01 -1.00892574e-01 3.56914192e-01 2.39551604e-01
-8.12210143e-02 5.32227978e-02 -2.45761707e-01 5.82807288e-02
1.11599326e+00 -2.17194781e-01 2.22197510e-02 -4.52179313e-01
5.88494897e-01 -6.57870770e-01 1.86332226e-01 6.66413367e-01
-2.61728972e-01 7.96218216e-01 2.27505654e-01 -4.57074940e-01
-1.00618398e+00 -8.40030134e-01 -3.50762576e-01 7.21592486e-01
6.90838456e-01 -3.75695229e-02 -8.04251850e-01 -2.07515776e-01
-3.27578872e-01 2.78204411e-01 -5.35048366e-01 -3.37963194e-01
-6.85830712e-01 -9.11794901e-01 7.01716065e-01 6.35675609e-01
1.28065264e+00 -7.27192461e-01 -6.13374174e-01 9.83674172e-03
-4.53361034e-01 -1.22840703e+00 -8.07891965e-01 -9.27634835e-02
-7.80888915e-01 -8.31633151e-01 -1.29334593e+00 -9.03538942e-01
4.59468007e-01 8.33255231e-01 8.40066552e-01 -1.19973645e-01
-7.65025541e-02 1.25266016e-02 -2.82650799e-01 2.13756233e-01
-7.62596503e-02 1.33276209e-01 1.31279230e-01 8.13153237e-02
-3.64231691e-02 -8.53170931e-01 -1.20919394e+00 5.23226261e-01
-1.28947210e+00 2.07290217e-01 9.06664193e-01 1.09556711e+00
6.69078529e-02 1.23825662e-01 4.92771894e-01 -1.73907116e-01
6.68095469e-01 -2.41636366e-01 -5.88859916e-01 5.98220527e-02
-7.23230660e-01 1.20425299e-01 6.95937872e-01 -5.02317250e-01
-1.16167104e+00 -5.39349794e-01 -2.53021065e-02 -7.21607327e-01
1.96930394e-02 2.87381768e-01 -1.65927876e-03 -4.18748260e-01
5.22930741e-01 4.56817180e-01 2.42176484e-02 -5.33998013e-01
3.38389486e-01 8.11521411e-01 8.72347236e-01 -2.17506528e-01
8.27305138e-01 5.26567101e-01 -1.05913371e-01 -7.34416962e-01
-6.99016631e-01 -3.04829419e-01 -9.59670991e-02 -9.74174365e-02
8.68885398e-01 -1.22010481e+00 -1.14128220e+00 1.17596269e+00
-1.26696837e+00 -2.42682248e-01 2.38297626e-01 5.90255022e-01
-3.30932826e-01 6.49103761e-01 -9.69064474e-01 -2.88669199e-01
-6.26663446e-01 -1.26288331e+00 8.54037166e-01 5.50757408e-01
2.93281734e-01 -7.39868045e-01 -3.90948862e-01 4.27193999e-01
1.13661206e+00 -2.06383571e-01 7.14686215e-01 1.34891018e-01
-9.63526845e-01 1.15491837e-01 -1.01014364e+00 6.88253284e-01
2.10062608e-01 -5.37920833e-01 -1.17122257e+00 -6.83863878e-01
2.47823641e-01 -2.39492953e-01 1.28243148e+00 6.94481432e-01
1.35310328e+00 -3.49195838e-01 -5.46405092e-02 1.12464702e+00
1.57649875e+00 -9.15290862e-02 9.78034019e-01 4.06397611e-01
7.56947577e-01 1.84628084e-01 1.60069942e-01 2.89813817e-01
2.78264165e-01 6.76657200e-01 6.96151674e-01 -4.06625330e-01
-5.08337855e-01 7.49865919e-02 3.74308467e-01 7.63724506e-01
-1.97033077e-01 -3.73116195e-01 -5.21160007e-01 4.78419602e-01
-1.69398427e+00 -8.61575902e-01 1.84459117e-04 2.07404160e+00
9.75012600e-01 -2.95285136e-01 -3.73977572e-01 -1.59305796e-01
6.39010608e-01 4.73950326e-01 -6.99671328e-01 9.15929079e-02
-3.67637962e-01 1.84706867e-01 7.37445176e-01 4.50083494e-01
-1.05992389e+00 8.93539906e-01 5.88914871e+00 1.08607507e+00
-1.35813403e+00 2.34619826e-02 8.31138492e-01 7.97201414e-03
-6.91889599e-02 -2.72812903e-01 -5.62957466e-01 5.07268131e-01
6.66648924e-01 -5.02151363e-02 7.53153026e-01 5.70516944e-01
2.51018375e-01 -8.09722468e-02 -8.21734250e-01 1.22469258e+00
6.45847395e-02 -1.47982812e+00 -2.04473466e-01 6.88775554e-02
8.14307213e-01 2.74091959e-01 4.28748757e-01 -9.44110453e-02
-1.81108594e-01 -1.15683949e+00 6.71314180e-01 2.87693053e-01
1.15682161e+00 -6.36825323e-01 9.53906417e-01 1.35635734e-01
-1.04756165e+00 -1.97079107e-01 -5.06138027e-01 -6.24839589e-03
1.02581484e-02 6.54787600e-01 -2.62872159e-01 5.41618705e-01
1.14104521e+00 9.22212958e-01 -3.97280097e-01 1.15624547e+00
-1.37637660e-01 5.14264345e-01 1.00247383e-01 1.60042062e-01
2.13979572e-01 -2.00768948e-01 5.48730552e-01 1.13359284e+00
5.19860923e-01 1.43111244e-01 -5.53009748e-01 1.01905358e+00
-2.56893009e-01 -3.93901139e-01 -1.35685861e-01 2.03470185e-01
4.39864695e-01 1.26295662e+00 -2.93340474e-01 -2.20422685e-01
-3.61266136e-01 1.28466618e+00 1.15438499e-01 6.42489851e-01
-1.09445286e+00 -6.01344705e-01 1.03995419e+00 -6.48422539e-02
5.02817988e-01 -1.83001027e-01 -1.38144135e-01 -1.36523676e+00
9.31462795e-02 -8.36156070e-01 -1.59677163e-01 -8.85604382e-01
-1.29844820e+00 9.12216067e-01 -2.93817610e-01 -1.27094400e+00
2.80403465e-01 -6.64299130e-01 -3.49601328e-01 1.13229764e+00
-2.05746865e+00 -1.01817107e+00 -8.41109157e-01 5.49354911e-01
4.68267381e-01 1.06339917e-01 4.05942053e-01 4.36049640e-01
-7.78678715e-01 5.94546378e-01 2.35931218e-01 6.45417050e-02
9.69526350e-01 -9.17940319e-01 4.09094006e-01 9.72080946e-01
-5.81746817e-01 7.72467852e-01 7.24484682e-01 -1.93750024e-01
-1.49664176e+00 -1.17016876e+00 6.12381816e-01 -3.37709785e-02
7.70326734e-01 -1.13323741e-02 -1.11249876e+00 3.07809532e-01
5.11309028e-01 1.25479564e-01 4.62528691e-02 -3.31725359e-01
-6.30066931e-01 -3.65094960e-01 -8.73053789e-01 5.42822719e-01
9.41480100e-01 -5.49350441e-01 -2.56685317e-01 -6.81204274e-02
8.97277474e-01 -5.78269482e-01 -7.43214905e-01 5.92070639e-01
5.46334803e-01 -1.43036914e+00 1.11358440e+00 4.54528853e-02
8.18545938e-01 -4.13065702e-01 4.43140743e-03 -1.21362412e+00
-7.22219646e-01 -8.11743259e-01 -2.12079450e-01 1.03037453e+00
-3.84482183e-02 -9.52998281e-01 4.71759111e-01 1.82292998e-01
-2.86719352e-01 -8.88158083e-01 -9.38533247e-01 -8.68354261e-01
1.26945242e-01 -1.12677529e-01 4.90392894e-01 7.82042384e-01
-3.79383087e-01 1.95717201e-01 -7.39131629e-01 3.98745537e-01
7.11901426e-01 8.11729133e-02 2.67808497e-01 -7.43534684e-01
-1.64152592e-01 -6.27352774e-01 -2.61072874e-01 -1.85429418e+00
1.14726974e-02 -4.75681603e-01 5.05988114e-02 -1.12042582e+00
4.96979028e-01 -3.10844123e-01 -3.01974505e-01 4.00340736e-01
-2.59411722e-01 5.53485870e-01 7.15252478e-03 4.90559876e-01
-3.41688275e-01 8.53338063e-01 1.53682768e+00 -2.45295197e-01
-1.98970959e-02 -1.76547647e-01 -8.17360103e-01 4.83416378e-01
6.44069672e-01 -8.30074623e-02 -2.85712481e-01 -9.52332437e-01
6.21353388e-02 5.28979627e-03 6.92763031e-01 -9.92231965e-01
4.43797112e-01 8.84366184e-02 2.97519833e-01 -2.66131580e-01
2.21462280e-01 -6.42780840e-01 1.34840667e-01 3.48433763e-01
-1.99529827e-01 -1.14407964e-01 1.43467739e-01 5.53608894e-01
-3.83470207e-01 8.87880772e-02 1.06738579e+00 1.42427310e-01
-5.89873672e-01 4.34212029e-01 -1.93812400e-01 -2.43737593e-01
6.51725173e-01 -1.72274053e-01 -8.35629225e-01 -5.04495978e-01
-1.18910283e-01 -8.26570541e-02 6.22146249e-01 4.93930608e-01
7.36512661e-01 -1.22073627e+00 -6.51598215e-01 2.84070820e-01
-1.92645907e-01 -6.90993667e-02 6.53530478e-01 1.24041927e+00
-5.93271315e-01 5.34353137e-01 -2.82929748e-01 -4.74429041e-01
-1.04537320e+00 4.39442754e-01 5.24223804e-01 4.98935692e-02
-6.57277644e-01 1.04412496e+00 6.03423536e-01 2.49149650e-01
2.07924053e-01 -4.75869536e-01 -3.70462239e-02 -2.69331783e-01
9.04264987e-01 6.22323632e-01 5.70585718e-03 -5.04388690e-01
-1.99727774e-01 6.44799232e-01 -1.62239924e-01 4.03844208e-01
1.38043010e+00 -4.23412621e-01 -4.16566521e-01 -3.65415603e-01
1.21265614e+00 -1.32623926e-01 -1.71602297e+00 -4.15384948e-01
-5.16312599e-01 -6.52044415e-01 6.05187833e-01 -6.25769973e-01
-1.46159053e+00 8.41506720e-01 7.60954320e-01 1.56443492e-01
1.69053042e+00 1.56007811e-01 1.03681648e+00 -7.32741924e-03
1.78271443e-01 -4.69245762e-01 3.28767836e-01 4.22754318e-01
9.89418507e-01 -1.20795822e+00 -7.21216425e-02 -4.15814698e-01
-3.49664658e-01 1.19668901e+00 5.40361643e-01 -8.47309306e-02
3.51587355e-01 3.24596882e-01 -3.58858183e-02 -2.35938430e-02
-5.28055668e-01 1.31020872e-02 2.92028040e-01 3.92695904e-01
3.60615790e-01 -9.86873433e-02 -1.12967104e-01 4.93109077e-01
-4.33505028e-02 1.51718915e-01 4.90758955e-01 2.58109719e-01
-4.91665512e-01 -6.41546786e-01 -1.88127711e-01 6.42064884e-02
-4.70836759e-01 -5.12486637e-01 2.53209591e-01 4.65622216e-01
1.14683338e-01 9.40166771e-01 2.09351294e-02 -3.78440112e-01
7.59436339e-02 -7.26438880e-01 5.10939896e-01 6.05762042e-02
-3.48770499e-01 1.23021483e-01 -4.50883508e-01 -7.74242640e-01
-5.67786157e-01 -2.16164812e-01 -7.76334226e-01 -6.52821958e-01
-5.32033920e-01 -2.28240833e-01 3.83608282e-01 6.62094355e-01
5.45459449e-01 4.91151601e-01 7.54091382e-01 -1.04241467e+00
-7.21343517e-01 -1.06152511e+00 -5.57217062e-01 -3.42411771e-02
8.22475493e-01 -4.32159096e-01 -6.02522731e-01 -2.75311563e-02]
|
[11.391231536865234, -2.46476674079895]
|
bbcb7d77-5f76-45d1-b7b5-5e0a33874b34
|
semi-supervised-vision-transformers-at-scale
|
2208.05688
| null |
https://arxiv.org/abs/2208.05688v1
|
https://arxiv.org/pdf/2208.05688v1.pdf
|
Semi-supervised Vision Transformers at Scale
|
We study semi-supervised learning (SSL) for vision transformers (ViT), an under-explored topic despite the wide adoption of the ViT architectures to different tasks. To tackle this problem, we propose a new SSL pipeline, consisting of first un/self-supervised pre-training, followed by supervised fine-tuning, and finally semi-supervised fine-tuning. At the semi-supervised fine-tuning stage, we adopt an exponential moving average (EMA)-Teacher framework instead of the popular FixMatch, since the former is more stable and delivers higher accuracy for semi-supervised vision transformers. In addition, we propose a probabilistic pseudo mixup mechanism to interpolate unlabeled samples and their pseudo labels for improved regularization, which is important for training ViTs with weak inductive bias. Our proposed method, dubbed Semi-ViT, achieves comparable or better performance than the CNN counterparts in the semi-supervised classification setting. Semi-ViT also enjoys the scalability benefits of ViTs that can be readily scaled up to large-size models with increasing accuracies. For example, Semi-ViT-Huge achieves an impressive 80% top-1 accuracy on ImageNet using only 1% labels, which is comparable with Inception-v4 using 100% ImageNet labels.
|
['Stefano Soatto', 'Zhuowen Tu', 'Rahul Bhotika', 'Davide Modolo', 'Manchen Wang', 'Paolo Favaro', 'Avinash Ravichandran', 'Zhaowei Cai']
|
2022-08-11
| null | null | null | null |
['semi-supervised-image-classification']
|
['computer-vision']
|
[ 2.05606312e-01 2.44109154e-01 -2.92350024e-01 -6.54673934e-01
-9.62707281e-01 -4.88077372e-01 7.89900839e-01 -1.97858706e-01
-4.37339813e-01 4.77836400e-01 3.68439080e-03 -3.55474353e-01
1.41885191e-01 -5.87362766e-01 -9.81421888e-01 -5.57510197e-01
6.17397785e-01 6.45067871e-01 5.26435137e-01 2.01259270e-01
-1.54822886e-01 7.14936107e-02 -1.53956532e+00 3.05975795e-01
1.13340414e+00 1.28826022e+00 3.10086995e-01 4.39258575e-01
-1.05797470e-01 1.27507377e+00 -1.80061162e-01 -7.27872610e-01
2.09898278e-01 -6.52017221e-02 -9.15771723e-01 1.54652402e-01
1.06555676e+00 -2.00534195e-01 -6.83429763e-02 9.41574633e-01
2.58266151e-01 -3.05317193e-01 7.61388838e-01 -1.32245421e+00
-7.81110764e-01 7.62910664e-01 -5.00610948e-01 -2.89324015e-01
-3.22851688e-01 5.16733348e-01 1.29739368e+00 -1.03476965e+00
2.92085409e-01 1.36723459e+00 9.59610045e-01 5.89393198e-01
-1.50569856e+00 -6.58299029e-01 1.32249326e-01 9.35403034e-02
-1.19179547e+00 -6.04819596e-01 5.40381432e-01 -4.75248128e-01
8.43356729e-01 -2.91759342e-01 3.69322956e-01 1.23464775e+00
-1.58515140e-01 1.03371096e+00 1.29421973e+00 -3.52948725e-01
1.35766163e-01 3.79128009e-01 1.90484524e-01 8.84836137e-01
-1.70509994e-01 3.38768959e-02 -4.05275941e-01 1.74070746e-01
6.67536914e-01 -3.85435447e-02 -6.93802722e-04 -4.01108265e-01
-1.23017097e+00 7.42685974e-01 6.96812809e-01 -9.67489276e-03
-2.93423355e-01 4.21680599e-01 5.43334901e-01 3.07273000e-01
5.86853445e-01 2.21664071e-01 -6.39481723e-01 1.01243414e-01
-1.25039041e+00 -1.73900083e-01 5.99326670e-01 1.07157993e+00
1.02012479e+00 2.72487134e-01 -4.52679634e-01 1.13135433e+00
4.34248894e-01 5.05521297e-01 3.20916444e-01 -1.01989961e+00
3.94612312e-01 5.28887570e-01 -2.35117868e-01 -1.44203573e-01
-1.62317768e-01 -7.57011890e-01 -9.71748769e-01 2.27907181e-01
5.30437231e-01 2.37557739e-01 -1.32072866e+00 1.79749382e+00
1.13600008e-01 3.31041336e-01 -6.22406155e-02 6.46215439e-01
9.33654130e-01 4.55166668e-01 3.91245872e-01 -2.89095435e-02
1.18251228e+00 -1.64870226e+00 -2.96068281e-01 -3.00242871e-01
4.86966938e-01 -6.51786864e-01 1.50360322e+00 4.52158153e-01
-9.76382613e-01 -8.01604927e-01 -8.63058746e-01 -4.54698622e-01
-1.31369352e-01 4.53360647e-01 5.44215918e-01 6.24727011e-01
-1.41099417e+00 5.68471372e-01 -9.39308941e-01 -3.12691420e-01
9.27869558e-01 3.02739918e-01 -1.93624884e-01 -1.51945382e-01
-7.98764169e-01 7.75154114e-01 6.81197569e-02 -3.62828597e-02
-1.45704305e+00 -1.04352880e+00 -9.76383865e-01 -1.17846094e-01
4.38475728e-01 -7.51161337e-01 1.46009409e+00 -1.16698885e+00
-1.73789883e+00 1.05856538e+00 -2.02291712e-01 -7.48827994e-01
6.58756614e-01 -5.13677038e-02 6.95458949e-02 -1.25092268e-01
1.61963731e-01 1.10051489e+00 1.36493170e+00 -1.10704398e+00
-4.67671305e-01 -1.89231187e-01 1.39724568e-01 -1.04795657e-01
-4.55997914e-01 -2.90078670e-01 -7.52953529e-01 -5.43324649e-01
-2.05519065e-01 -8.52052271e-01 -1.95601821e-01 1.50170773e-01
-4.69261795e-01 -5.44724286e-01 6.39113903e-01 -1.68123037e-01
7.80108333e-01 -2.11102414e+00 -9.65916514e-02 -1.15080200e-01
4.37550873e-01 5.28621912e-01 -3.39694381e-01 4.46566641e-02
-1.24481291e-01 -2.14159071e-01 -3.94725740e-01 -1.05268764e+00
1.25515297e-01 2.43276387e-01 -4.01633143e-01 2.50479817e-01
4.57316250e-01 1.22515452e+00 -1.00486445e+00 -5.55577755e-01
3.53040546e-01 4.57976758e-01 -5.31989574e-01 1.45005137e-01
-4.88444895e-01 4.80347812e-01 -2.42522940e-01 7.63220310e-01
6.09742522e-01 -5.81165731e-01 -4.22578044e-02 -4.97516602e-01
-9.08544138e-02 4.19080436e-01 -6.92610502e-01 1.64652193e+00
-7.47986674e-01 5.63162148e-01 -1.10214114e-01 -9.94209528e-01
9.10877764e-01 1.50226310e-01 2.88606226e-01 -6.75957620e-01
-1.21555105e-01 2.73164093e-01 -6.00424826e-01 -2.04977527e-01
3.13487679e-01 -1.19588614e-01 3.71595919e-01 3.50302249e-01
6.06333196e-01 -2.64414847e-01 3.31719778e-02 2.76230603e-01
8.92474949e-01 5.80164552e-01 2.19850168e-02 -2.68137425e-01
4.70536411e-01 -6.32941574e-02 7.04216182e-01 8.38125408e-01
-3.23186487e-01 7.12226272e-01 4.09204274e-01 -3.34643722e-01
-1.00020218e+00 -1.15254629e+00 -2.57977277e-01 1.21171105e+00
-1.10267408e-01 -4.97364819e-01 -8.99503231e-01 -1.13900983e+00
9.40331221e-02 4.58247334e-01 -5.38669825e-01 -7.69798905e-02
-1.84145615e-01 -5.92822492e-01 6.77282691e-01 7.03727424e-01
7.55245328e-01 -1.04337168e+00 -8.81915092e-02 6.25214800e-02
-1.36010408e-01 -1.27249491e+00 -4.40452307e-01 4.75616068e-01
-9.78410542e-01 -9.63788927e-01 -7.66471863e-01 -7.97008932e-01
6.41047478e-01 7.74035603e-02 1.45066309e+00 -2.04785094e-01
1.16986394e-01 3.82592052e-01 -1.96979031e-01 -3.60375881e-01
-4.36833769e-01 3.29542994e-01 2.22666878e-02 2.15518191e-01
2.08887771e-01 -4.97155070e-01 -6.60554409e-01 4.06213015e-01
-7.39444137e-01 1.95570916e-01 6.53218389e-01 9.78538275e-01
7.20341682e-01 -3.13743621e-01 5.45798957e-01 -1.17930567e+00
3.88715900e-02 -1.57394290e-01 -5.90272188e-01 4.86416399e-01
-9.55177248e-01 2.05118671e-01 8.22735131e-01 -4.86644894e-01
-1.05592740e+00 3.08746338e-01 -2.74229705e-01 -7.19888568e-01
-5.42973839e-02 1.50639161e-01 -1.05801851e-01 -2.71213055e-01
8.31771553e-01 2.07046911e-01 -4.35938984e-02 -4.57043111e-01
7.07479417e-01 7.58777320e-01 6.98436558e-01 -5.61304688e-01
8.59403789e-01 5.21629155e-01 -2.12248608e-01 -5.96018553e-01
-1.42271590e+00 -4.51528996e-01 -6.60550833e-01 -8.38559344e-02
7.97474444e-01 -1.28389823e+00 -5.23801386e-01 8.07001710e-01
-8.60230148e-01 -9.64357615e-01 -5.46824336e-01 2.61993229e-01
-6.08349144e-01 3.21070492e-01 -6.64814532e-01 -6.13760889e-01
-5.47044218e-01 -1.39027095e+00 1.22389257e+00 1.07663684e-01
1.73381835e-01 -1.12806201e+00 -1.15879461e-01 7.13853836e-01
4.54772234e-01 -1.90324068e-01 6.82745159e-01 -4.59707141e-01
-6.49650276e-01 2.50019312e-01 -5.79578459e-01 9.68272150e-01
-1.77736916e-02 8.86168610e-03 -1.60503507e+00 -3.42743874e-01
-3.41988236e-01 -1.11822426e+00 1.38424671e+00 4.43509638e-01
1.16335177e+00 -1.77356191e-02 -1.20440327e-01 8.59208167e-01
1.28957903e+00 -5.61219931e-01 6.81875169e-01 1.06369749e-01
9.87764001e-01 3.53390068e-01 5.42040765e-01 2.00437576e-01
8.94999564e-01 5.02831817e-01 5.92745602e-01 -2.77990729e-01
-4.92691934e-01 -4.53954488e-01 5.26264787e-01 8.75736654e-01
1.00820392e-01 1.05625361e-01 -7.16175616e-01 6.17600858e-01
-1.83692062e+00 -5.58290362e-01 -2.76297659e-01 2.24602628e+00
1.06372511e+00 3.55962783e-01 2.02057511e-01 2.93254964e-02
3.90392721e-01 1.89239204e-01 -8.22993755e-01 -3.64917926e-02
-8.60498324e-02 4.94547874e-01 6.24952137e-01 4.03000474e-01
-1.17361009e+00 1.47397363e+00 6.51025438e+00 1.19984829e+00
-1.19362485e+00 3.88556361e-01 7.87980437e-01 2.61567950e-01
-2.09700376e-01 1.45919055e-01 -9.36173439e-01 2.43856579e-01
9.52386975e-01 5.29792547e-01 4.44462359e-01 1.03250003e+00
1.24504685e-01 1.73453316e-02 -1.17455220e+00 1.01834548e+00
-3.36283147e-02 -1.29559839e+00 2.03932241e-01 -2.84659833e-01
8.36173892e-01 6.40024483e-01 1.83031216e-01 5.44595659e-01
5.65977752e-01 -7.32413471e-01 9.69082773e-01 3.82565588e-01
1.30809653e+00 -3.45422238e-01 3.97351027e-01 3.25211525e-01
-1.23641360e+00 -6.57215640e-02 -5.52829206e-01 1.94448292e-01
-1.43396974e-01 9.28836703e-01 -7.64362991e-01 4.17242795e-01
8.09564710e-01 1.25147271e+00 -8.36442888e-01 9.09617901e-01
-5.59608459e-01 1.14636481e+00 -2.72527188e-01 3.79023761e-01
3.76447141e-01 -1.30337626e-01 9.41955578e-03 1.21854115e+00
-9.71993357e-02 -5.28482795e-01 1.78351149e-01 9.44797277e-01
-3.53414774e-01 -1.59355745e-01 -2.52544016e-01 1.22884780e-01
3.35262746e-01 1.48164833e+00 -5.45405626e-01 -5.05477130e-01
-4.74505693e-01 8.92057359e-01 7.48301029e-01 3.88844937e-01
-8.41325164e-01 -1.12952635e-01 3.48509073e-01 5.08302487e-02
3.73479933e-01 7.81910345e-02 -3.84073883e-01 -1.26365995e+00
-1.63761124e-01 -8.17046881e-01 3.07647824e-01 -8.81912410e-01
-1.38187146e+00 7.21074879e-01 -2.71920234e-01 -1.13701034e+00
-9.54161398e-03 -6.77100182e-01 -6.12292111e-01 6.28169179e-01
-2.03446198e+00 -1.81112635e+00 -4.28490907e-01 7.79160738e-01
6.42677307e-01 -1.74926102e-01 6.01143122e-01 3.28219742e-01
-5.98317385e-01 8.34076583e-01 -5.27712964e-02 -2.19227243e-02
1.14429677e+00 -1.58245277e+00 5.71136951e-01 6.76713288e-01
3.45056027e-01 3.75430912e-01 3.10380042e-01 -3.52696091e-01
-9.96852398e-01 -1.65112817e+00 9.10543680e-01 -6.42358243e-01
7.69032598e-01 -4.99351114e-01 -7.83117533e-01 9.15046751e-01
1.96431026e-01 3.33657593e-01 3.00308853e-01 1.76244229e-01
-9.41580892e-01 -5.81755936e-01 -1.06874144e+00 3.89065027e-01
1.06604958e+00 -8.88402522e-01 -3.47251922e-01 4.45845127e-01
8.80500436e-01 -1.59701362e-01 -7.76964426e-01 4.10788685e-01
4.50697780e-01 -1.05306852e+00 9.95666683e-01 -1.58309847e-01
4.69010949e-01 -3.25601459e-01 7.94175826e-03 -1.22847092e+00
-4.48517501e-01 -7.28096724e-01 5.40197678e-02 1.46373796e+00
5.27706802e-01 -5.67666352e-01 7.46300161e-01 1.73190072e-01
-2.39899173e-01 -7.47588515e-01 -6.18155777e-01 -7.00849533e-01
7.27726817e-02 -8.33763659e-01 3.42361212e-01 8.52041900e-01
-4.84051526e-01 4.10197824e-01 -4.27916825e-01 -3.36252339e-03
1.06492627e+00 9.87610966e-03 8.51572871e-01 -1.16811275e+00
-4.24303830e-01 -2.30872452e-01 -1.17147744e-01 -1.35638428e+00
3.52173865e-01 -1.07910991e+00 1.89129427e-01 -1.35559249e+00
4.19799924e-01 -8.23121190e-01 -4.01670098e-01 9.12186682e-01
-2.56046385e-01 7.47566164e-01 -8.50786269e-02 4.25711513e-01
-8.85852695e-01 7.95057714e-01 1.19913542e+00 -3.41639549e-01
-4.69098501e-02 1.28040567e-01 -5.62386632e-01 7.93573797e-01
3.75217974e-01 -4.15288776e-01 -5.50341666e-01 -6.41222537e-01
-1.16779907e-02 -4.87846315e-01 5.57981849e-01 -1.01882482e+00
2.66774029e-01 6.31072819e-02 2.58447770e-02 -5.28402865e-01
2.09524572e-01 -5.78829110e-01 -2.00762615e-01 1.45497352e-01
-4.12608743e-01 -4.59019274e-01 -7.11269155e-02 4.72247213e-01
-1.85408622e-01 -1.40643850e-01 1.17626119e+00 -3.27922124e-03
-6.74550116e-01 5.91604710e-01 -3.74919660e-02 1.93490803e-01
7.27123082e-01 -3.99345048e-02 -3.86176795e-01 -2.23365515e-01
-5.88384748e-01 4.07962650e-01 3.43604654e-01 2.17599019e-01
4.53396082e-01 -1.12705362e+00 -6.25993431e-01 8.97068381e-02
4.83890533e-01 3.50343049e-01 4.06681038e-02 1.06518483e+00
-2.12465867e-01 2.40880847e-01 9.68757048e-02 -1.04344964e+00
-9.53045547e-01 3.29755962e-01 3.89552325e-01 -4.81816620e-01
-6.02331281e-01 1.00956821e+00 3.84374350e-01 -7.97160804e-01
4.50779766e-01 -6.62100494e-01 -8.76132101e-02 -1.02868117e-01
3.75542521e-01 1.56232610e-01 1.86887905e-01 -3.71888816e-01
-2.24230424e-01 5.43148279e-01 -1.81507632e-01 1.03952654e-01
1.39851403e+00 -5.72970659e-02 -9.68276337e-02 5.76009989e-01
9.54141319e-01 -2.76868105e-01 -1.82134533e+00 -7.58911014e-01
-2.13487651e-02 2.82952283e-02 2.54636377e-01 -9.03324842e-01
-1.24784517e+00 9.63867903e-01 3.51881862e-01 -2.11503863e-01
1.13183200e+00 1.92355439e-01 7.04957902e-01 4.64340806e-01
5.00918806e-01 -8.60249460e-01 1.82106793e-01 6.57074869e-01
5.72686434e-01 -1.55079019e+00 -2.51153618e-01 -3.83665621e-01
-8.58801603e-01 7.35378861e-01 6.77584291e-01 -1.36252284e-01
6.23427808e-01 2.43369967e-01 2.78555542e-01 1.92620289e-02
-9.69947815e-01 -4.02973175e-01 3.49888861e-01 6.87149823e-01
2.89948493e-01 -9.26541016e-02 3.53406072e-01 3.64386082e-01
5.88707179e-02 2.08141357e-01 1.37407675e-01 4.47541833e-01
-1.29301742e-01 -1.04145503e+00 -1.06547400e-01 5.45431554e-01
-2.42115155e-01 -3.81342441e-01 -2.92808563e-01 4.00298089e-01
2.25793317e-01 9.61711049e-01 -7.16788322e-02 -4.58636940e-01
2.90016681e-01 -1.71196442e-02 5.24380445e-01 -6.80433333e-01
-7.53585577e-01 2.40956880e-02 -1.38023689e-01 -7.13438392e-01
-7.13445842e-01 -4.37950641e-01 -8.94405961e-01 -2.44603660e-02
-4.48072791e-01 -7.78076947e-02 6.17285132e-01 1.10504651e+00
2.55777091e-01 4.67799932e-01 7.12483644e-01 -7.68492103e-01
-7.33251035e-01 -1.13389575e+00 -3.07625532e-01 3.39207858e-01
3.14458638e-01 -6.06141567e-01 -3.17900807e-01 2.94648528e-01]
|
[9.627009391784668, 1.6054438352584839]
|
62a9e89c-c106-42e0-94db-087713dfa294
|
question-answering-over-biological-knowledge
|
2210.06040
| null |
https://arxiv.org/abs/2210.06040v1
|
https://arxiv.org/pdf/2210.06040v1.pdf
|
Question Answering Over Biological Knowledge Graph via Amazon Alexa
|
Structured and unstructured data and facts about drugs, genes, protein, viruses, and their mechanism are spread across a huge number of scientific articles. These articles are a large-scale knowledge source and can have a huge impact on disseminating knowledge about the mechanisms of certain biological processes. A knowledge graph (KG) can be constructed by integrating such facts and data and be used for data integration, exploration, and federated queries. However, exploration and querying large-scale KGs is tedious for certain groups of users due to a lack of knowledge about underlying data assets or semantic technologies. A question-answering (QA) system allows the answer of natural language questions over KGs automatically using triples contained in a KG. Recently, the use and adaption of digital assistants are getting wider owing to their capability at enabling users to voice commands to control smart systems or devices. This paper is about using Amazon Alexa's voice-enabled interface for QA over KGs. As a proof-of-concept, we use the well-known DisgeNET KG, which contains knowledge covering 1.13 million gene-disease associations between 21,671 genes and 30,170 diseases, disorders, and clinical or abnormal human phenotypes. Our study shows how Alex could be of help to find facts about certain biological entities from large-scale knowledge bases.
|
['Stefan Decker', 'Mohamed Abdelwaheb', 'Prinon Das', 'Hussain Ali', 'Md. Rezaul Karim']
|
2022-10-12
| null | null | null | null |
['data-integration']
|
['knowledge-base']
|
[-5.83345830e-01 3.59693050e-01 -3.98261219e-01 -1.64505050e-01
-4.04834300e-01 -7.01862156e-01 3.16004395e-01 8.28845322e-01
-1.64273113e-01 1.19759142e+00 4.64470536e-01 -4.96495157e-01
-4.14662302e-01 -1.15428424e+00 -7.19439209e-01 -2.66744524e-01
-8.04342702e-02 7.33136654e-01 4.36255842e-01 -4.01788056e-01
3.30471396e-02 3.80612582e-01 -1.27614129e+00 2.51247585e-01
1.21752024e+00 7.29375422e-01 2.11605832e-01 2.60820895e-01
-8.01197886e-01 5.07490575e-01 -7.02952206e-01 -8.27786922e-01
-3.97939026e-01 -1.31059110e-01 -1.02356362e+00 -8.02442551e-01
-1.67815417e-01 2.77944863e-01 7.21727237e-02 1.11862266e+00
5.43940485e-01 -3.20646346e-01 -2.12840457e-02 -1.41235721e+00
-6.87337995e-01 5.31110346e-01 -1.95183437e-02 2.71154884e-02
8.25726151e-01 6.25618622e-02 1.04725420e+00 -6.11849070e-01
1.44373357e+00 1.21356130e+00 4.10212100e-01 3.00642788e-01
-6.95819020e-01 -2.75807649e-01 -3.86300504e-01 4.46684778e-01
-1.44413686e+00 -2.10763156e-01 7.95950368e-02 -4.02985603e-01
1.26007926e+00 4.65161443e-01 8.11041951e-01 5.61561823e-01
2.02883393e-01 2.74751842e-01 4.67314363e-01 -2.41790965e-01
4.35286760e-01 4.79703248e-01 1.41202018e-01 1.18057573e+00
7.47824490e-01 -5.62673748e-01 -8.93923759e-01 -9.14826035e-01
2.57492691e-01 -5.75638376e-02 -4.71055269e-01 -1.25334293e-01
-1.20808017e+00 6.09556258e-01 2.84750879e-01 3.51337850e-01
-6.19634390e-01 -2.71015048e-01 4.30547297e-01 2.64491200e-01
3.11990798e-01 8.56118977e-01 -8.75803709e-01 -2.52148479e-01
-2.70444602e-01 1.73983127e-01 1.38175535e+00 9.33761120e-01
7.74858415e-01 -6.40401423e-01 2.21349612e-01 6.99992061e-01
2.26036102e-01 2.93140858e-01 5.99419534e-01 -5.82621455e-01
2.07476661e-01 1.26271260e+00 2.58107305e-01 -1.11505926e+00
-5.10779738e-01 -1.49806505e-02 -5.26339173e-01 -5.17982543e-01
3.48867834e-01 -1.46341130e-01 -5.78925490e-01 1.52059019e+00
9.77192461e-01 -1.68912083e-01 1.81819573e-01 7.87454605e-01
1.34704745e+00 3.44633698e-01 3.26072782e-01 -1.47140339e-01
1.87128532e+00 -2.55400687e-01 -9.74238157e-01 3.21657717e-01
7.86846220e-01 -3.87623370e-01 8.25353384e-01 3.12860101e-01
-7.23805606e-01 6.53315336e-02 -6.46189988e-01 -2.11881474e-01
-1.37200749e+00 -3.51342052e-01 8.34841013e-01 4.10241932e-01
-9.34114933e-01 3.16799581e-01 -4.49456692e-01 -9.14375424e-01
6.90928102e-01 2.15592623e-01 -5.42913616e-01 -2.01247036e-01
-1.59025192e+00 1.06667292e+00 4.75831509e-01 -2.44929045e-01
-4.60502446e-01 -1.11171985e+00 -5.23365140e-01 1.00024618e-01
8.62343669e-01 -1.22463608e+00 3.68041486e-01 7.52553344e-02
-7.58853793e-01 8.33534598e-01 -3.10782701e-01 -2.93962300e-01
-1.94202706e-01 -1.35095075e-01 -9.70480382e-01 2.04767346e-01
3.52738619e-01 3.31108570e-01 4.32662480e-02 -4.65141296e-01
-4.64490622e-01 -7.55212307e-01 1.66217685e-01 -1.47655904e-01
-2.52701938e-01 2.07092002e-01 -5.70128381e-01 -1.58664316e-01
-2.92056769e-01 -3.64125371e-01 -8.70885402e-02 8.29720870e-02
-6.58830464e-01 -3.86932850e-01 7.77790010e-01 -8.62470269e-01
1.28144813e+00 -1.55181932e+00 8.51599053e-02 1.92212924e-01
4.79621649e-01 4.26086485e-01 1.60354987e-01 8.16858113e-01
2.36493900e-01 7.14469790e-01 1.89178243e-01 7.79939234e-01
-2.17125192e-01 3.57765704e-01 -1.12710230e-01 -5.13768708e-03
2.54976060e-02 1.22239912e+00 -9.47528422e-01 -4.31769818e-01
-2.72675484e-01 4.34876919e-01 -2.41123855e-01 -5.84821217e-02
-9.35888350e-01 1.20296963e-01 -9.32394981e-01 1.05317044e+00
2.49932587e-01 -6.62989616e-01 4.38371301e-01 -3.32471848e-01
-7.39530835e-04 3.19041908e-01 -7.25710571e-01 1.68829894e+00
-2.87221726e-02 1.42528698e-01 1.11811467e-01 -7.64488339e-01
7.03334153e-01 4.55361992e-01 5.00709057e-01 -3.30869257e-01
-3.06394696e-02 1.36160806e-01 -3.10286492e-01 -9.26167011e-01
1.91810638e-01 1.78326339e-01 6.04234152e-02 2.99893975e-01
2.82538659e-03 1.54935896e-01 2.92707473e-01 5.67099631e-01
1.43345964e+00 -1.26523942e-01 6.69001341e-01 -2.59383976e-01
4.19389993e-01 4.11491722e-01 5.21622896e-01 3.28395605e-01
3.09031755e-01 -3.24698836e-01 7.39994884e-01 -6.15863621e-01
-6.96172118e-01 -8.09290648e-01 -1.79592460e-01 6.71447277e-01
-1.04335412e-01 -8.92649174e-01 -4.18712139e-01 -5.24346054e-01
3.52430254e-01 4.69937950e-01 -5.09700239e-01 2.38496400e-02
1.57146081e-01 -8.08781505e-01 5.87611973e-01 -7.11487548e-04
4.47721571e-01 -8.69151056e-01 -5.00706732e-01 9.87823159e-02
-4.62785631e-01 -1.20614243e+00 -3.09577025e-02 -2.19120651e-01
-6.06299758e-01 -1.57217884e+00 -3.33594680e-01 -2.99765348e-01
5.72436631e-01 -1.92077026e-01 1.20711374e+00 1.48020133e-01
-7.12681115e-01 2.58037001e-01 -2.11381748e-01 -9.30431902e-01
-3.08066875e-01 -5.55877015e-02 1.52364179e-01 -4.28226829e-01
5.96936464e-01 -4.59029406e-01 -3.99608254e-01 3.74434859e-01
-9.69068587e-01 -5.73868863e-02 2.63814569e-01 4.48160082e-01
5.98558664e-01 6.37988374e-02 9.69108224e-01 -1.12243342e+00
8.82190645e-01 -1.11314368e+00 -6.81104183e-01 7.80391216e-01
-8.89083922e-01 2.47942746e-01 4.07022387e-01 3.44556831e-02
-7.91933298e-01 -2.99700081e-01 -1.07042290e-01 1.74814165e-02
-1.11641489e-01 1.15289807e+00 -5.42199910e-01 2.50776023e-01
5.82830846e-01 -1.17241055e-01 6.73390552e-02 -7.24987090e-01
6.92923129e-01 7.11594760e-01 3.26106906e-01 -4.67824787e-01
3.46810699e-01 3.74977440e-01 3.43403704e-02 -8.19486976e-01
-5.66534638e-01 -5.74816883e-01 -1.41344234e-01 6.14839941e-02
7.35956848e-01 -7.76629567e-01 -1.25825703e+00 6.77470416e-02
-9.20694590e-01 1.87438935e-01 -2.82878995e-01 1.83102205e-01
2.00302508e-02 1.28210023e-01 -3.65695775e-01 -3.42707962e-01
-5.67369342e-01 -5.33783317e-01 7.27853358e-01 3.01933765e-01
-3.50926608e-01 -1.00216925e+00 2.16977417e-01 4.78696495e-01
4.52546030e-01 2.59191364e-01 1.47764456e+00 -9.06370163e-01
-5.64566910e-01 -3.05893421e-01 -1.95021555e-01 -4.17091101e-01
4.93075937e-01 -1.25640497e-01 -5.14945745e-01 2.56808937e-01
-5.20786405e-01 -1.95579991e-01 3.64310712e-01 -4.94565368e-02
1.05562317e+00 -5.57089210e-01 -8.28303874e-01 1.69525847e-01
1.31577921e+00 1.49293214e-01 6.50451243e-01 1.86080888e-01
5.90717912e-01 6.47544801e-01 6.41360402e-01 4.99766558e-01
4.77549493e-01 6.53851092e-01 3.16732764e-01 1.16863392e-01
3.05990279e-01 -2.96813011e-01 -1.30848408e-01 4.21521097e-01
-2.44067892e-01 -2.64734983e-01 -1.14444172e+00 6.11437559e-01
-1.75340545e+00 -8.54656935e-01 -8.35445300e-02 2.11405253e+00
1.24927437e+00 -3.73112231e-01 6.66031763e-02 -5.49760163e-01
6.13924742e-01 -4.62703854e-01 -9.34914827e-01 -2.11937591e-01
-1.81036651e-01 1.66658238e-01 2.96706200e-01 2.51224309e-01
-3.38792801e-01 8.86366844e-01 6.09730911e+00 8.99129987e-01
-8.28026295e-01 -6.70003891e-02 3.09881806e-01 -2.53838524e-02
-6.25287652e-01 4.22219448e-02 -8.73083413e-01 3.34640026e-01
1.12400472e+00 -7.67745912e-01 3.35948855e-01 5.82340539e-01
2.61898994e-01 -4.16158348e-01 -7.45183468e-01 7.56682813e-01
-3.08819890e-01 -1.99434400e+00 1.20878719e-01 2.90681303e-01
3.59265625e-01 1.56244621e-01 -5.26784301e-01 -1.90205887e-01
8.01083505e-01 -8.98172140e-01 -1.25999954e-02 1.07486749e+00
8.65486324e-01 -5.10220408e-01 6.70579910e-01 3.15415263e-01
-9.02039170e-01 1.28330231e-01 -3.72290969e-01 2.98351973e-01
1.10195607e-01 1.24710214e+00 -1.30552793e+00 9.41048682e-01
8.56238306e-01 4.27490413e-01 -4.67888325e-01 1.04085016e+00
-1.47482887e-01 4.71070468e-01 -4.80393380e-01 -3.21771622e-01
-3.87633473e-01 -8.26769918e-02 4.07296002e-01 8.93902481e-01
3.45742345e-01 4.13605988e-01 -1.70499265e-01 9.32267189e-01
-2.90846229e-01 4.17688191e-01 -6.80202782e-01 -7.11355567e-01
6.67317748e-01 1.29368162e+00 -5.90578556e-01 -5.29415071e-01
-4.45210725e-01 5.93092978e-01 3.75939608e-01 3.98515344e-01
-5.13334751e-01 -5.11980832e-01 8.95057976e-01 2.47252285e-01
2.05538608e-02 1.16278328e-01 2.51555532e-01 -1.06682777e+00
-8.72057900e-02 -9.73636448e-01 7.88344741e-01 -1.06502521e+00
-1.25268781e+00 2.77148902e-01 -8.81518796e-02 -6.05678380e-01
-2.29989335e-01 -3.95738274e-01 -3.00898980e-02 8.68280411e-01
-1.10459101e+00 -8.13074470e-01 -1.03189565e-01 6.99076653e-01
-1.79706514e-01 -1.23673722e-01 1.17801809e+00 2.23211467e-01
-4.91730422e-01 -9.53728706e-02 9.40392166e-02 -8.45792666e-02
6.22532904e-01 -9.97788131e-01 2.90547609e-01 -2.90970858e-02
5.78777902e-02 1.12511718e+00 6.28730595e-01 -1.16052222e+00
-1.86748219e+00 -9.70995545e-01 1.12290657e+00 -6.61040664e-01
7.76629567e-01 -2.90612847e-01 -1.06738436e+00 4.34656113e-01
6.63901195e-02 -3.51136550e-02 1.13216937e+00 3.82332653e-01
-3.39103013e-01 -3.00688110e-02 -1.45326185e+00 5.90790987e-01
1.07436228e+00 -5.52609563e-01 -4.87356603e-01 5.23342252e-01
8.36049080e-01 -9.82179567e-02 -1.36667979e+00 1.20305844e-01
5.00811875e-01 -3.65640998e-01 9.19007540e-01 -1.17584479e+00
7.34585300e-02 -4.19404656e-01 4.15967628e-02 -1.27183151e+00
1.31680593e-01 -6.32407367e-01 -2.08193079e-01 1.09616387e+00
7.25287616e-01 -9.47467268e-01 4.90604401e-01 1.02969992e+00
1.53606892e-01 -8.48998129e-01 -7.97627747e-01 -4.56599921e-01
-3.82267028e-01 -2.30116859e-01 9.81353760e-01 1.23120499e+00
7.40774512e-01 3.82927001e-01 6.70852810e-02 2.35250831e-01
1.65577531e-01 2.11762354e-01 6.72107399e-01 -1.57951152e+00
-6.95313811e-02 2.31652800e-02 -5.36063671e-01 -3.32172692e-01
-4.65656400e-01 -9.37992156e-01 -6.56273782e-01 -1.95000172e+00
2.44851634e-01 -3.48604172e-01 6.30681738e-02 9.09237921e-01
-6.63652569e-02 -3.15107077e-01 -3.95484179e-01 1.76409721e-01
-7.69375980e-01 2.48104036e-01 1.01626992e+00 1.85495056e-02
-2.59597693e-02 -4.78976190e-01 -9.76083994e-01 6.68467939e-01
7.31959224e-01 -4.21886832e-01 -4.55274493e-01 1.02538764e-01
1.01438439e+00 1.96364313e-01 2.97151893e-01 -4.53138471e-01
5.61758697e-01 -2.71418661e-01 2.50646509e-02 -3.54573220e-01
1.17281958e-01 -6.51248395e-01 7.83492923e-01 4.71526116e-01
-1.03106685e-01 -8.93701687e-02 3.56050283e-01 7.02389419e-01
-2.07440376e-01 -5.83764240e-02 -6.82965154e-04 -4.66393471e-01
-5.65068007e-01 2.86135048e-01 -3.01901370e-01 3.13413352e-01
9.39775169e-01 1.19535647e-01 -9.78652298e-01 -3.98019254e-01
-8.69191766e-01 4.76095945e-01 2.75780946e-01 4.75096524e-01
5.87600052e-01 -8.62375498e-01 -4.36853439e-01 -5.13350517e-02
2.97747016e-01 -2.25754648e-01 1.39895976e-01 6.72159553e-01
-4.18991029e-01 8.41127634e-01 -5.48701473e-02 -8.21721330e-02
-1.38102949e+00 8.23168695e-01 3.94684225e-02 -2.11831644e-01
-3.84860456e-01 5.37222624e-01 -2.25183219e-01 -3.58727962e-01
1.63845662e-02 -1.34074837e-01 -1.88966915e-01 2.09603161e-01
7.31035173e-01 3.65018696e-01 5.31986281e-02 -8.89104754e-02
-6.89189136e-01 2.32718796e-01 -4.82381620e-02 1.81981415e-01
1.46349323e+00 -6.42093718e-02 -6.98792517e-01 3.94059420e-01
8.54626358e-01 3.21313947e-01 -4.76222150e-02 -2.15382010e-01
1.92484617e-01 -2.23865345e-01 -4.14690018e-01 -1.34268069e+00
-8.92272294e-01 3.10153097e-01 9.63058323e-02 3.19981754e-01
8.22574735e-01 6.26798093e-01 5.22483706e-01 6.94634497e-01
7.42900789e-01 -8.31840098e-01 -5.04419327e-01 1.38190001e-01
8.29943240e-01 -1.05671060e+00 1.29281223e-01 -7.11671770e-01
-4.07479525e-01 8.87014925e-01 2.78252006e-01 8.97560894e-01
8.29779565e-01 2.39629835e-01 -9.91118848e-02 -1.03468037e+00
-1.11166477e+00 -1.88846678e-01 1.62311941e-01 6.71264768e-01
3.48278612e-01 -4.55896743e-02 -5.20259321e-01 7.76101172e-01
-1.08373567e-01 5.84141433e-01 1.13190599e-01 8.27370286e-01
-5.64242482e-01 -1.30017233e+00 -1.84260219e-01 7.56138921e-01
-6.70154154e-01 -4.21354085e-01 -6.71653271e-01 5.05240142e-01
2.46372506e-01 9.35039461e-01 -3.91693860e-01 -1.28131956e-01
2.84968257e-01 4.22319144e-01 -1.12478146e-02 -5.14504373e-01
-2.96890497e-01 -5.70856929e-01 6.09553754e-01 -7.00383008e-01
-3.20102870e-01 -3.71230692e-01 -1.61082578e+00 -3.26554537e-01
-3.56870234e-01 6.43140197e-01 8.22492659e-01 8.34569812e-01
1.08856750e+00 2.24376723e-01 -1.41065434e-01 5.34340799e-01
1.50902107e-01 -5.85814238e-01 -6.19884312e-01 1.76421970e-01
-2.99002022e-01 -4.15309340e-01 2.51325399e-01 1.25318155e-01]
|
[8.732985496520996, 8.410316467285156]
|
bea9ddd2-fbcb-4365-933d-a0c0f7ba647b
|
applying-second-order-quantifier-elimination
|
2110.11108
| null |
https://arxiv.org/abs/2110.11108v1
|
https://arxiv.org/pdf/2110.11108v1.pdf
|
Applying Second-Order Quantifier Elimination in Inspecting Gödel's Ontological Proof
|
In recent years, G\"odel's ontological proof and variations of it were formalized and analyzed with automated tools in various ways. We supplement these analyses with a modeling in an automated environment based on first-order logic extended by predicate quantification. Formula macros are used to structure complex formulas and tasks. The analysis is presented as a generated type-set document where informal explanations are interspersed with pretty-printed formulas and outputs of reasoners for first-order theorem proving and second-order quantifier elimination. Previously unnoticed or obscured aspects and details of G\"odel's proof become apparent. Practical application possibilities of second-order quantifier elimination are shown and the encountered elimination tasks may serve as benchmarks.
|
['Christoph Wernhard']
|
2021-10-21
| null | null | null | null |
['automated-theorem-proving', 'automated-theorem-proving']
|
['miscellaneous', 'reasoning']
|
[ 3.30640048e-01 1.18597639e+00 1.16830021e-02 -2.71474153e-01
-1.59430593e-01 -8.37817788e-01 8.45068455e-01 3.36696297e-01
1.50433391e-01 1.00564718e+00 -1.26437038e-01 -1.19388831e+00
-7.36589968e-01 -7.93877602e-01 -5.83261073e-01 1.92628801e-01
-3.15443993e-01 6.24986768e-01 4.78451520e-01 -3.53943676e-01
3.67811918e-01 3.58207405e-01 -1.97421038e+00 5.64185381e-01
9.68546271e-01 9.35092270e-01 1.01256326e-01 9.48572397e-01
-6.25532448e-01 1.62433958e+00 -6.76654160e-01 -6.42579913e-01
-1.95259359e-02 -2.19153598e-01 -1.03424037e+00 -1.56510890e-01
2.91051030e-01 -1.73119113e-01 1.27599344e-01 1.41144598e+00
-6.05398238e-01 -4.08026338e-01 4.85383987e-01 -1.97033572e+00
-2.20470935e-01 1.05497313e+00 2.82759964e-01 -2.24127352e-01
1.06897044e+00 1.52440578e-01 1.03642893e+00 -3.70942533e-01
7.44969308e-01 1.64102960e+00 7.18729079e-01 4.82795000e-01
-1.31634438e+00 -4.79364246e-01 -2.08894685e-01 5.46545506e-01
-1.41139245e+00 -2.08203375e-01 6.57099485e-01 -7.52586782e-01
1.46627796e+00 9.76725578e-01 5.86554170e-01 2.71841377e-01
6.45080388e-01 5.50399184e-01 1.14087868e+00 -8.91247153e-01
9.37450752e-02 7.52916515e-01 5.71603179e-01 1.22280514e+00
7.20074356e-01 8.23988467e-02 -2.36887187e-01 -2.91616172e-01
5.56371868e-01 -3.18257272e-01 3.35037075e-02 -2.85125494e-01
-1.10145462e+00 6.90214038e-01 -2.68353701e-01 4.44187492e-01
-8.30948725e-02 3.37820709e-01 4.64226127e-01 2.99023151e-01
-1.91572890e-01 4.93959785e-01 -5.74153543e-01 -2.09838524e-01
-8.08573127e-01 8.51922870e-01 1.29506111e+00 1.43451500e+00
8.23829234e-01 -1.11758210e-01 -3.37430649e-02 -8.66069272e-02
5.61630130e-01 3.53712738e-01 1.56354941e-02 -1.38750899e+00
1.49258047e-01 1.24972296e+00 6.03165805e-01 -9.77226317e-01
-3.58841568e-01 -9.27626118e-02 -2.35491753e-01 6.83059752e-01
4.86937761e-01 9.94771570e-02 -3.43686074e-01 1.24694014e+00
2.32054681e-01 -3.44224960e-01 3.65327448e-01 4.56595302e-01
7.70895243e-01 3.55441928e-01 -1.50021780e-02 -4.58783329e-01
1.50545430e+00 -6.02659106e-01 -1.29044008e+00 3.49283874e-01
8.26294899e-01 -3.11883926e-01 1.07248533e+00 8.72081041e-01
-1.58229101e+00 -2.38620732e-02 -1.03775513e+00 -3.46090466e-01
-6.61084533e-01 -2.08253860e-01 1.12900150e+00 4.89666820e-01
-8.73207152e-01 6.60040557e-01 -5.85357904e-01 1.52232209e-02
3.61359656e-01 4.79645491e-01 -3.14230807e-02 1.18414998e-01
-1.20648193e+00 9.69004452e-01 6.25555098e-01 1.28853740e-02
-7.46309221e-01 -7.64677584e-01 -1.16839635e+00 1.31478071e-01
7.04827487e-01 -5.23635566e-01 1.57297301e+00 -7.11851597e-01
-1.41383088e+00 1.02958870e+00 -7.34619275e-02 -7.49496579e-01
8.58095169e-01 7.92906955e-02 -8.25083733e-01 1.21795379e-01
2.38945261e-01 -5.99695183e-02 3.23913991e-01 -1.07608008e+00
-7.59371996e-01 -2.66911477e-01 1.08684003e+00 -4.87810045e-01
5.90019286e-01 2.42263094e-01 5.03614247e-01 8.26705918e-02
-3.06588039e-02 -2.13151440e-01 1.47133470e-01 4.91413614e-03
-5.60317099e-01 -3.26078951e-01 6.12354934e-01 -3.46709758e-01
1.62759018e+00 -1.64053321e+00 -7.22908005e-02 1.66506484e-01
5.76983511e-01 -1.66511387e-01 5.30729055e-01 6.21822596e-01
-2.51273721e-01 4.01549220e-01 -3.66186388e-02 2.41307825e-01
1.00359416e+00 3.86064887e-01 -3.77435297e-01 8.66672471e-02
2.50869840e-01 7.05336154e-01 -1.05632985e+00 -9.14900601e-01
4.43047017e-01 -1.34624958e-01 -6.49038553e-01 -7.23199695e-02
-9.43648696e-01 -3.35369796e-01 -2.70140201e-01 9.30763483e-01
4.94432360e-01 -2.12206960e-01 4.45930779e-01 -9.08747017e-02
-5.36166131e-01 5.68289578e-01 -1.33339500e+00 1.27843392e+00
-4.62343931e-01 3.97710174e-01 1.03974730e-01 -4.68949676e-01
5.88957250e-01 3.80651474e-01 -2.69168615e-01 -3.40596795e-01
2.59056747e-01 3.55691403e-01 3.20298672e-02 -8.35499346e-01
4.73355412e-01 -5.62688589e-01 -1.00771502e-01 2.44079635e-01
-2.41186023e-01 -5.89735806e-01 8.51061702e-01 3.15130860e-01
1.21596658e+00 4.56547916e-01 8.59887958e-01 -6.80165470e-01
1.03568602e+00 6.61862731e-01 4.64712143e-01 6.58045709e-01
1.20550483e-01 -1.98849082e-01 1.11661148e+00 -8.05508137e-01
-8.32165778e-01 -8.42077315e-01 -2.01367646e-01 6.25351191e-01
-8.60169381e-02 -1.11834800e+00 -6.84465528e-01 -5.29836774e-01
1.23251468e-01 1.34116006e+00 -5.40016770e-01 1.62265107e-01
-1.25180706e-01 2.09136963e-01 7.77543485e-01 2.31130585e-01
2.54513204e-01 -9.58238602e-01 -1.21118379e+00 7.06843212e-02
7.40704089e-02 -9.78036642e-01 6.80749178e-01 3.31850231e-01
-6.99144363e-01 -1.70273900e+00 4.98720884e-01 -2.46400312e-01
6.49221778e-01 -5.52549362e-01 1.31844223e+00 6.25536978e-01
-3.13263088e-01 2.65683144e-01 -1.96705401e-01 -9.41937804e-01
-8.74825001e-01 -5.14424503e-01 -6.08763583e-02 -6.22590005e-01
3.35260391e-01 -5.25068700e-01 1.68476179e-01 -1.13257453e-01
-1.03730059e+00 2.51124293e-01 1.54523134e-01 5.04666686e-01
3.24863791e-01 6.34478867e-01 -6.03418872e-02 -1.31754267e+00
7.69619167e-01 7.34348148e-02 -1.21736574e+00 5.63524544e-01
-6.39478862e-01 4.27250236e-01 8.83694589e-01 -1.01083510e-01
-9.00170863e-01 -3.98456961e-01 3.87898773e-01 -1.53754607e-01
-8.44164938e-02 5.44157445e-01 -3.84890109e-01 1.68389708e-01
6.88071549e-01 -5.26784547e-02 -1.20076694e-01 5.25136590e-02
1.17594764e-01 4.41246182e-01 5.68937719e-01 -1.01749849e+00
1.01467085e+00 1.79771930e-01 6.47439003e-01 -2.05378801e-01
-5.80404103e-01 3.32548141e-01 -3.03387761e-01 -2.73648739e-01
4.78651226e-01 -2.81355560e-01 -1.50136304e+00 -2.71873951e-01
-1.35637701e+00 -7.71291137e-01 -7.73220897e-01 9.41740125e-02
-9.00547743e-01 2.73974657e-01 -1.28288478e-01 -1.68383157e+00
1.19478621e-01 -1.10533035e+00 8.63607228e-01 -1.22480720e-01
-7.74473071e-01 -1.01733291e+00 -1.71483144e-01 5.88034056e-02
-3.66912000e-02 3.09919268e-01 1.42396033e+00 -6.40463293e-01
-4.57526416e-01 -2.42156684e-01 -2.24277526e-01 2.04296991e-01
-7.12188855e-02 3.67980301e-01 -7.69747376e-01 5.24008632e-01
-1.89142600e-01 8.96599591e-02 -3.42942387e-01 -1.41130984e-01
9.55640852e-01 -9.86119151e-01 -2.42334425e-01 -6.68865889e-02
1.49760211e+00 1.32735670e-01 9.58328068e-01 6.62303984e-01
-1.25453435e-02 5.36310852e-01 7.72076547e-01 3.99165660e-01
3.58353436e-01 3.73907268e-01 6.87453389e-01 6.27766311e-01
8.58860463e-02 -1.49015799e-01 5.07724248e-02 1.14081837e-02
-5.12359738e-01 3.41132194e-01 -1.12923384e+00 2.65865594e-01
-1.80978012e+00 -1.32362425e+00 -7.52137125e-01 1.82825410e+00
1.02768576e+00 6.12358928e-01 4.94654961e-02 7.58044899e-01
4.41060036e-01 -3.04835349e-01 1.82225034e-01 -8.14014077e-01
8.84407610e-02 2.04865828e-01 1.94921553e-01 1.28173065e+00
-5.16536593e-01 7.85810351e-01 7.29930019e+00 5.48952818e-01
-4.01207268e-01 -2.88265914e-01 -4.24991429e-01 3.60713691e-01
-7.48271227e-01 6.23903811e-01 -5.39215624e-01 4.22621705e-02
7.54116952e-01 -5.45473099e-01 7.50006795e-01 8.98590505e-01
6.11395948e-02 -4.55531240e-01 -1.56135345e+00 6.56152666e-01
-2.53952384e-01 -1.42152441e+00 1.47800922e-01 -1.99015036e-01
1.46127492e-01 -8.82172048e-01 -5.21384954e-01 5.77683389e-01
3.54484320e-01 -9.11689281e-01 1.32421243e+00 8.29527378e-01
6.84167445e-01 -5.69648325e-01 7.65621781e-01 3.92531425e-01
-8.85508358e-01 -4.76395577e-01 -4.39080670e-02 -1.04126871e+00
-2.13222355e-01 5.78533351e-01 -9.40177083e-01 9.11535382e-01
2.02608645e-01 1.30618319e-01 -4.39447761e-01 5.73458016e-01
-6.84112608e-01 1.73593253e-01 -3.20200384e-01 -3.41660798e-01
-1.07770981e-02 -2.49099195e-01 6.27779186e-01 1.34895778e+00
2.54224762e-02 3.48920703e-01 -3.23470742e-01 1.38483500e+00
5.94827831e-01 -3.41803432e-01 -7.02130675e-01 -1.97456151e-01
2.27671847e-01 1.06125462e+00 -5.07958293e-01 -7.99915493e-01
-2.39303380e-01 3.58496338e-01 -1.24342941e-01 2.77012616e-01
-1.21561801e+00 -6.98792934e-01 3.79302800e-01 5.58687210e-01
-2.47921318e-01 -8.07713419e-02 -3.12467307e-01 -1.00101519e+00
-7.90921785e-03 -9.13620114e-01 2.35775322e-01 -1.33582938e+00
-7.67770350e-01 4.03650850e-01 6.43777788e-01 -7.59679139e-01
-6.09825194e-01 -1.02109468e+00 -3.49740505e-01 7.08428264e-01
-1.33492243e+00 -9.59998548e-01 -1.98449194e-01 4.90981340e-01
1.47063196e-01 -2.43089665e-02 1.22095656e+00 6.43466413e-02
-1.38517261e-01 1.39557794e-01 -8.42468917e-01 -2.88560838e-01
-2.52553165e-01 -1.62354517e+00 -4.28354383e-01 7.15255439e-01
-2.95331120e-01 8.93387139e-01 1.52297068e+00 -6.34561002e-01
-1.68434978e+00 -7.09950149e-01 1.50270593e+00 -5.05417883e-01
1.21265173e+00 -3.16515177e-01 -4.84458864e-01 1.31909502e+00
1.63830101e-01 -1.99775696e-01 2.83221424e-01 -1.82694197e-02
-3.65546405e-01 -2.14539781e-01 -1.53021526e+00 8.17463338e-01
1.03442264e+00 -6.90676033e-01 -1.28333247e+00 6.76380277e-01
4.84701902e-01 -4.63446915e-01 -7.08918333e-01 2.71734476e-01
7.59373724e-01 -1.09443533e+00 4.99341965e-01 -7.98341751e-01
4.90980387e-01 -6.87074006e-01 -4.90119189e-01 -2.72405118e-01
-6.61312044e-02 -1.08497345e+00 -3.00509065e-01 9.63140905e-01
5.26555777e-01 -7.96456516e-01 2.09201321e-01 1.11144674e+00
-2.61004001e-01 -4.74343508e-01 -6.90948904e-01 -8.59728992e-01
-2.85091311e-01 -9.46599603e-01 6.94144249e-01 8.33863616e-01
1.14605975e+00 2.50411928e-01 4.62412864e-01 1.54508695e-01
6.28317952e-01 2.01909602e-01 7.89422214e-01 -1.53912723e+00
-2.38490328e-01 -6.03530049e-01 -6.17618918e-01 -4.47137177e-01
4.78218585e-01 -1.02658772e+00 -1.66569129e-01 -1.70896018e+00
1.21438140e-02 -2.35802650e-01 2.47589514e-01 8.01510215e-01
6.21601105e-01 -1.80575892e-01 4.37456779e-02 -1.79468542e-01
-7.80745208e-01 -1.75751939e-01 9.98258173e-01 -3.40524435e-01
1.07554026e-01 -3.19921851e-01 -5.04700422e-01 1.17961454e+00
5.38860977e-01 -4.15108114e-01 -5.44551194e-01 2.77137309e-02
1.10785615e+00 3.40909094e-01 9.56373990e-01 -1.07774520e+00
1.30328760e-01 -5.33503652e-01 -4.07515198e-01 -5.87569714e-01
-5.23850285e-02 -1.32433712e+00 7.79185891e-01 4.66951787e-01
-2.69556016e-01 8.94177929e-02 4.11378175e-01 -1.52858540e-01
-3.99145596e-02 -7.10099339e-01 1.48063868e-01 -3.29787284e-01
-5.41372001e-01 -5.85434556e-01 -6.07040286e-01 -3.44834954e-01
1.12643623e+00 -3.28530729e-01 -3.98353428e-01 -1.56142682e-01
-1.03390217e+00 1.45341545e-01 5.48689604e-01 -3.21114123e-01
5.34277618e-01 -1.03487277e+00 -1.75312921e-01 1.30412420e-02
-5.30863069e-02 6.58629881e-03 9.96408910e-02 1.01548064e+00
-1.06213415e+00 7.26922929e-01 -2.60727376e-01 -2.94894695e-01
-1.20794952e+00 9.37072933e-01 4.67541933e-01 -5.60376704e-01
-5.06951630e-01 2.38524631e-01 5.73492423e-02 -2.68404484e-01
6.07407875e-02 -1.16052997e+00 4.02232148e-02 -5.04152119e-01
8.43415499e-01 3.87128085e-01 -1.12837218e-01 -1.08705662e-01
-8.91496062e-01 1.74662024e-01 3.42603028e-01 -1.46693364e-01
1.23177075e+00 -8.25289190e-02 -7.07107365e-01 9.24226880e-01
4.08344656e-01 2.23371372e-01 -6.08891845e-01 3.42099726e-01
1.01711556e-01 -7.87000209e-02 -3.88184339e-01 -1.19429183e+00
-6.95097968e-02 5.71100891e-01 -2.14680910e-01 9.15604353e-01
8.53652775e-01 6.47123754e-02 -2.02731509e-02 7.46137500e-01
6.72088444e-01 -9.82146204e-01 -7.53784239e-01 3.42675686e-01
1.11087525e+00 -6.32305801e-01 5.00526905e-01 -8.39322448e-01
-1.82621345e-01 1.47888076e+00 1.39044091e-01 -9.40754544e-03
1.92539170e-01 9.09254491e-01 -1.35071829e-01 -5.71937025e-01
-8.74978364e-01 -6.80572912e-02 -2.18719050e-01 6.40231192e-01
3.56191516e-01 7.94609860e-02 -6.15868092e-01 1.02486873e+00
-6.97440922e-01 6.37547374e-01 8.49645972e-01 1.22724819e+00
-4.00596678e-01 -1.01310825e+00 -7.55458474e-01 4.47515287e-02
-4.99223381e-01 -6.85895309e-02 -3.57811540e-01 1.66829515e+00
4.03015792e-01 1.05665052e+00 -3.23034912e-01 -1.38799131e-01
3.87479484e-01 3.64188284e-01 1.22729790e+00 -6.43767595e-01
-5.73189557e-01 -3.68634015e-01 7.73837447e-01 -5.81667542e-01
-4.48809594e-01 -4.37188059e-01 -1.83018327e+00 -4.34234828e-01
-1.59268185e-01 5.60139716e-01 4.99013245e-01 1.18787920e+00
-2.59267986e-01 8.30918670e-01 -3.43294263e-01 -3.94300640e-01
-4.75551188e-01 -6.56571567e-01 -8.30476284e-01 -4.26436365e-02
4.50345278e-01 -7.50686228e-01 -5.01682103e-01 2.72907972e-01]
|
[8.779589653015137, 6.82900333404541]
|
be1d8ad2-cda0-4cc4-88a4-7ecd1bf69a99
|
test-time-adaptation-vs-training-time
|
2212.06242
| null |
https://arxiv.org/abs/2212.06242v1
|
https://arxiv.org/pdf/2212.06242v1.pdf
|
Test-time Adaptation vs. Training-time Generalization: A Case Study in Human Instance Segmentation using Keypoints Estimation
|
We consider the problem of improving the human instance segmentation mask quality for a given test image using keypoints estimation. We compare two alternative approaches. The first approach is a test-time adaptation (TTA) method, where we allow test-time modification of the segmentation network's weights using a single unlabeled test image. In this approach, we do not assume test-time access to the labeled source dataset. More specifically, our TTA method consists of using the keypoints estimates as pseudo labels and backpropagating them to adjust the backbone weights. The second approach is a training-time generalization (TTG) method, where we permit offline access to the labeled source dataset but not the test-time modification of weights. Furthermore, we do not assume the availability of any images from or knowledge about the target domain. Our TTG method consists of augmenting the backbone features with those generated by the keypoints head and feeding the aggregate vector to the mask head. Through a comprehensive set of ablations, we evaluate both approaches and identify several factors limiting the TTA gains. In particular, we show that in the absence of a significant domain shift, TTA may hurt and TTG show only a small gain in performance, whereas for a large domain shift, TTA gains are smaller and dependent on the heuristics used, while TTG gains are larger and robust to architectural choices.
|
['Fatih Porikli', 'Hyojin Park', 'Debasmit Das', 'Kambiz Azarian']
|
2022-12-12
| null | null | null | null |
['human-instance-segmentation']
|
['computer-vision']
|
[ 4.43982750e-01 2.60654151e-01 -7.64196739e-02 -3.41989756e-01
-8.93229425e-01 -8.21813881e-01 5.18468797e-01 1.00524954e-01
-7.42455184e-01 6.67349756e-01 -3.25412095e-01 -4.21293914e-01
-1.73450679e-01 -5.03082097e-01 -8.63339424e-01 -7.23747909e-01
1.34786278e-01 5.81698060e-01 7.50561237e-01 2.17877571e-02
3.07085276e-01 5.91903389e-01 -1.24455655e+00 8.09017867e-02
8.71162176e-01 1.11933947e+00 1.68835104e-01 7.03916609e-01
-2.55431533e-02 1.02486439e-01 -8.33226264e-01 -2.53943652e-01
6.13143861e-01 -4.79519904e-01 -8.32110882e-01 2.88618505e-01
4.53599125e-01 -1.92932963e-01 2.80083232e-02 9.90732074e-01
3.91878754e-01 1.42013803e-01 4.99011904e-01 -1.23178589e+00
-9.51677710e-02 3.82374555e-01 -6.34646535e-01 2.63174772e-01
5.54656833e-02 2.92872399e-01 7.56009996e-01 -7.53225207e-01
8.57059479e-01 7.86517203e-01 5.20343184e-01 4.99782890e-01
-1.32638359e+00 -5.04696190e-01 4.44292665e-01 1.10774487e-03
-1.23182726e+00 -3.30519110e-01 5.96866071e-01 -3.98555934e-01
6.72697842e-01 3.22645664e-01 6.06331348e-01 5.68876088e-01
-1.09253436e-01 7.42955744e-01 1.19697857e+00 -6.37402594e-01
5.33913672e-01 2.54231781e-01 3.11751395e-01 7.62801349e-01
1.46529168e-01 6.15260517e-03 -3.62832099e-01 -8.32194388e-02
7.48953819e-01 -3.45335603e-01 -3.68378937e-01 -7.35947490e-01
-1.14665461e+00 5.88265896e-01 4.08945382e-01 3.34972471e-01
-4.32628483e-01 -2.26406138e-02 1.82028010e-01 4.92864728e-01
3.60770315e-01 8.48802805e-01 -6.27492607e-01 3.24417911e-02
-1.20308280e+00 2.31591552e-01 5.69791496e-01 7.92731047e-01
1.03275263e+00 -1.03132784e-01 -3.29363227e-01 5.55576503e-01
-8.31536353e-02 2.69809633e-01 7.15881467e-01 -1.02499270e+00
5.49031377e-01 5.26097655e-01 2.29853123e-01 -4.86164480e-01
-4.73356992e-01 -5.02407610e-01 3.00780404e-02 4.24376339e-01
9.24209416e-01 -4.08870906e-01 -1.49518478e+00 1.80906653e+00
4.19863462e-01 1.35371119e-01 -1.41351938e-01 6.38979793e-01
1.33779630e-01 2.95580178e-01 -3.24964255e-01 -1.88694566e-01
1.00095844e+00 -9.40471709e-01 -7.92469978e-02 -5.34672081e-01
8.27235043e-01 -7.08433509e-01 1.13576186e+00 5.54657996e-01
-1.04798353e+00 -5.22231579e-01 -1.32828832e+00 3.91768306e-01
-4.29416031e-01 -1.54870320e-02 2.69634336e-01 6.66269481e-01
-1.13731706e+00 7.93808281e-01 -8.32723618e-01 -4.97105628e-01
2.69507855e-01 6.59946263e-01 -2.74972588e-01 -2.43273720e-01
-6.68171525e-01 7.09284604e-01 5.08338213e-01 -1.50277495e-01
-7.25635469e-01 -7.20876038e-01 -6.18423760e-01 5.78573272e-02
5.93375683e-01 -3.13560128e-01 1.27233326e+00 -1.18907273e+00
-1.43817222e+00 7.36984849e-01 -1.23968311e-01 -3.10994357e-01
5.87993562e-01 -7.45403096e-02 -1.17545016e-01 8.26650262e-02
8.60011280e-02 1.02108276e+00 1.01094306e+00 -1.43607819e+00
-7.91641057e-01 -3.04266661e-01 1.60245359e-01 2.33867139e-01
-3.00051063e-01 -4.44153905e-01 -7.62617111e-01 -4.87915367e-01
2.82195121e-01 -1.29947007e+00 -2.22033277e-01 -2.18689330e-02
-5.25139809e-01 1.80996180e-01 7.66964912e-01 -4.22590911e-01
1.04510248e+00 -2.40974236e+00 1.30915120e-01 6.65018499e-01
9.21322778e-02 3.53326499e-01 -4.44678634e-01 4.53260541e-02
-1.74689829e-01 1.20710559e-01 -5.63397944e-01 -1.56276986e-01
-1.98745579e-01 2.28499785e-01 -1.88143983e-01 2.67894417e-01
1.74838766e-01 7.11301088e-01 -8.21554005e-01 -3.32538515e-01
1.80827886e-01 -9.85039771e-03 -7.50169933e-01 1.23676788e-02
-2.50675619e-01 4.06641781e-01 -4.88815129e-01 3.86565059e-01
5.34091532e-01 -2.07079098e-01 7.75797591e-02 -1.42625853e-01
-4.99780811e-02 7.49373659e-02 -1.41854966e+00 1.50763154e+00
-2.67841727e-01 5.07555366e-01 6.24042749e-03 -8.97542298e-01
5.85569084e-01 1.34073645e-01 4.13627893e-01 -6.38318598e-01
-4.86259349e-02 4.22054648e-01 3.17284197e-01 -1.24438435e-01
1.47085190e-01 -2.14553401e-02 1.42026410e-01 5.30008197e-01
7.63703212e-02 -1.63604349e-01 3.17044973e-01 1.46842405e-01
1.33584774e+00 8.93033221e-02 1.73061833e-01 -8.52001458e-02
2.69472778e-01 1.62587494e-01 5.72705746e-01 7.96770036e-01
-4.17811535e-02 6.64827049e-01 6.06779158e-01 -2.29214787e-01
-9.33924735e-01 -1.06103098e+00 -1.62310883e-01 9.47059989e-01
1.43232569e-01 -2.06781998e-01 -9.76532638e-01 -1.21805954e+00
-6.98649585e-02 6.82900369e-01 -7.97152281e-01 -9.70416442e-02
-6.93681240e-01 -6.67567015e-01 2.30517969e-01 5.39144576e-01
6.16270661e-01 -9.03709054e-01 -7.19070315e-01 -1.50812110e-02
6.48371354e-02 -9.79448199e-01 -7.91038036e-01 5.08970559e-01
-1.01306129e+00 -9.45157588e-01 -8.30499053e-01 -5.77120841e-01
1.07165551e+00 5.12384176e-02 7.53326774e-01 1.88108519e-01
-1.80070907e-01 4.20326918e-01 -3.01464885e-01 -2.74536520e-01
-2.68976241e-01 2.49573514e-01 -1.87159389e-01 -6.98162615e-02
-5.36863394e-02 -4.12055284e-01 -5.17649472e-01 4.41608667e-01
-9.42933679e-01 -2.76440203e-01 7.36063659e-01 7.49797761e-01
7.43507326e-01 -3.98334749e-02 4.51189280e-01 -1.16808617e+00
4.48439121e-01 -2.59993523e-01 -6.80781364e-01 2.63412088e-01
-8.43722343e-01 2.85111487e-01 5.07549226e-01 -6.89387202e-01
-7.31479466e-01 2.68137217e-01 4.23307791e-02 -5.61774015e-01
-1.15596414e-01 4.04565632e-01 -1.82258949e-01 -4.21406358e-01
7.95709372e-01 -9.62178502e-03 -1.76534206e-01 -3.68009627e-01
2.92758167e-01 2.32682094e-01 5.06904185e-01 -5.44568777e-01
9.60308015e-01 2.53607482e-01 -1.77661985e-01 -5.74380100e-01
-6.21969402e-01 -3.11418146e-01 -8.00352037e-01 -1.97379783e-01
8.68900001e-01 -3.26805413e-01 -4.11290266e-02 3.54615450e-01
-7.24960625e-01 -8.71021807e-01 -6.29811287e-01 4.87099975e-01
-5.92543006e-01 1.88002139e-01 -2.49036118e-01 -2.66108185e-01
5.26048765e-02 -1.45661974e+00 7.46087193e-01 1.54828921e-01
-2.50175625e-01 -7.59831607e-01 -3.08769852e-01 1.08422160e-01
1.91415116e-01 1.54850245e-01 1.14418387e+00 -1.07817602e+00
-5.94966054e-01 -3.98772478e-01 -8.81098136e-02 3.62599194e-01
2.30776057e-01 -1.60565078e-01 -9.60135937e-01 -4.96207058e-01
7.75557235e-02 -1.75500706e-01 6.78240061e-01 5.41094065e-01
1.07597077e+00 -5.23382910e-02 -4.68484432e-01 7.13919163e-01
1.19852042e+00 4.54056531e-01 4.25778300e-01 5.97558558e-01
6.65146470e-01 5.37127435e-01 7.26511836e-01 7.22257793e-02
9.86503512e-02 7.39701450e-01 1.87834918e-01 -3.66968006e-01
-2.92415649e-01 -1.13696590e-01 1.56449109e-01 4.29574668e-01
1.23544455e-01 -2.52879143e-01 -8.93197894e-01 5.79114139e-01
-1.61264884e+00 -4.89640862e-01 2.84294486e-01 2.61822844e+00
7.47294664e-01 5.88301539e-01 2.35307530e-01 1.88470766e-01
6.57283664e-01 -8.58755633e-02 -9.54573214e-01 -2.86592811e-01
2.78904140e-01 2.94544727e-01 7.55396664e-01 6.41129375e-01
-9.05723572e-01 9.54731703e-01 6.43151379e+00 7.17751145e-01
-1.26832616e+00 4.26850393e-02 6.63043678e-01 -2.03446552e-01
-2.22043514e-01 1.01625957e-01 -5.77465534e-01 3.37094784e-01
7.14333296e-01 3.98397967e-02 4.95339185e-01 6.20773375e-01
-1.47699192e-01 -4.08490777e-01 -1.31389940e+00 6.37026548e-01
-8.62127170e-02 -1.05251348e+00 -1.80596218e-01 2.43506148e-01
7.40568340e-01 4.79213372e-02 8.76948982e-02 2.98331410e-01
2.76801586e-01 -6.01554632e-01 8.12084436e-01 1.57191724e-01
7.38437116e-01 -6.73336029e-01 5.77121317e-01 3.05220902e-01
-8.83620620e-01 -1.20820269e-01 -8.24541524e-02 3.19103241e-01
-4.82909940e-02 4.32097882e-01 -9.94472921e-01 2.23991215e-01
6.00288749e-01 1.54542297e-01 -8.14128399e-01 1.25277114e+00
-2.95478344e-01 8.90914559e-01 -4.45267916e-01 3.74218196e-01
3.43043536e-01 6.62670354e-04 5.35740972e-01 9.45447028e-01
3.61737758e-01 -1.31833985e-01 3.91967714e-01 5.99010050e-01
-1.09414376e-01 -3.06002200e-02 -3.58059406e-01 1.91913590e-01
5.82729161e-01 1.16761923e+00 -1.24528766e+00 -3.92869025e-01
-2.84240901e-01 1.05067313e+00 2.25459158e-01 7.08774030e-01
-6.20026112e-01 -4.19806302e-01 3.67768407e-01 2.45898232e-01
6.17776275e-01 3.90080772e-02 -3.65144938e-01 -8.35125983e-01
1.86737508e-01 -8.96576762e-01 4.21224028e-01 -6.86081171e-01
-8.02808166e-01 5.81825197e-01 3.24612200e-01 -1.14616001e+00
-4.48174745e-01 -4.69209909e-01 -6.12744212e-01 8.89688730e-01
-1.35747170e+00 -7.79372990e-01 -5.90711869e-02 5.72838008e-01
5.69890976e-01 -1.38045149e-02 5.05254090e-01 7.22137466e-02
-5.16348004e-01 7.78532922e-01 -8.98535699e-02 1.96795305e-03
7.45007813e-01 -1.36271095e+00 4.26827699e-01 1.02965772e+00
2.24685118e-01 4.99595284e-01 7.98686326e-01 -6.35967433e-01
-8.49147081e-01 -9.59360898e-01 3.78609538e-01 -3.21864694e-01
3.95374805e-01 -3.90171200e-01 -1.09863460e+00 7.93012202e-01
-1.19691081e-02 2.56666899e-01 2.83136010e-01 6.46985099e-02
-1.40416935e-01 -2.22778514e-01 -1.29208934e+00 4.94587123e-01
6.73308074e-01 -2.68906564e-01 -4.22692567e-01 1.49651900e-01
6.08109534e-01 -4.11656857e-01 -6.28080130e-01 4.32838857e-01
3.52521777e-01 -6.95892751e-01 5.87076724e-01 -2.82736897e-01
-2.33141165e-02 -3.48891050e-01 -4.65614721e-02 -1.54412758e+00
-2.34277055e-01 -3.45286280e-01 2.77331352e-01 1.12738824e+00
8.45354676e-01 -8.10350776e-01 1.04954827e+00 8.07963729e-01
-2.42399216e-01 -8.64559948e-01 -1.01787579e+00 -8.61324966e-01
1.18501231e-01 -3.91024202e-01 5.58130860e-01 8.63741577e-01
-2.51933396e-01 8.51822272e-02 1.43578336e-01 2.95956731e-01
3.57344866e-01 2.03179978e-02 6.68858588e-01 -9.88780141e-01
-4.80327606e-01 -3.07161897e-01 -3.39511871e-01 -9.55022752e-01
-9.31318253e-02 -7.45563328e-01 3.35356116e-01 -1.23791087e+00
-7.43890479e-02 -5.50348878e-01 -4.44299906e-01 7.32779860e-01
-2.46774793e-01 2.55041301e-01 3.06270927e-01 2.34246701e-01
-2.84985393e-01 1.76588483e-02 1.23652196e+00 -1.17964961e-01
-5.31112552e-01 2.13267416e-01 -5.23564041e-01 8.18843901e-01
7.81168520e-01 -6.97307169e-01 -8.43144774e-01 -5.77067316e-01
-5.55797219e-02 -5.37121221e-02 2.62930155e-01 -1.20530808e+00
2.21280858e-01 -1.18453912e-01 5.51905155e-01 -1.83100909e-01
2.59047955e-01 -6.43682420e-01 -2.28038102e-01 4.70717221e-01
-2.46330097e-01 1.56205371e-01 4.45593625e-01 5.41307032e-01
-8.45100656e-02 -4.88396108e-01 9.29201961e-01 6.98610693e-02
-6.87042236e-01 8.12947005e-02 -3.39540929e-01 6.44359142e-02
1.08152676e+00 -5.84617615e-01 -1.65470362e-01 -1.83191717e-01
-8.60850632e-01 1.69375926e-01 6.63397789e-01 1.11968830e-01
3.75750810e-01 -9.95000124e-01 -3.49025667e-01 3.92248392e-01
2.91518550e-02 3.15386653e-02 -7.57226273e-02 6.95578516e-01
-1.90515026e-01 2.01162770e-01 -4.81972732e-02 -6.14015877e-01
-1.14351892e+00 6.46868408e-01 4.42518473e-01 -2.43148863e-01
-4.70323414e-01 9.83559251e-01 5.19284725e-01 -3.19384098e-01
2.47265339e-01 -4.80321318e-01 4.02540751e-02 -3.68117467e-02
2.42032751e-01 1.06483020e-01 4.05653685e-01 -2.81789094e-01
-2.65522003e-01 5.71064293e-01 -3.76418322e-01 -4.75879848e-01
1.13923037e+00 8.22290033e-02 2.79290795e-01 4.51734990e-01
9.48244810e-01 1.27533525e-01 -1.45808697e+00 -1.28302053e-01
1.59507275e-01 -5.24136364e-01 3.45306806e-02 -1.11974907e+00
-1.25982416e+00 6.26761258e-01 8.70703280e-01 6.27789348e-02
1.36871707e+00 5.52306212e-02 6.15079701e-01 4.06123936e-01
3.84422481e-01 -1.03299725e+00 1.24300830e-01 1.81764498e-01
6.09398305e-01 -1.02572536e+00 -5.12927286e-02 -3.97749454e-01
-6.39360845e-01 8.27734113e-01 7.91577816e-01 -4.12441865e-02
5.34201443e-01 8.73812139e-02 2.67054528e-01 -1.90506443e-01
-5.67074418e-01 -2.83369005e-01 3.73961627e-01 5.41601121e-01
1.07144592e-02 -1.96094602e-01 -1.33656487e-01 2.45926231e-01
-1.73887953e-01 -1.05244733e-01 3.02673846e-01 1.00380850e+00
-3.81437093e-01 -1.30216491e+00 -3.62106830e-01 6.54837906e-01
-2.25021958e-01 -2.09070109e-02 -4.78379726e-01 1.03921521e+00
3.83313358e-01 7.04566956e-01 1.27060682e-01 -4.65478599e-01
5.81356883e-01 3.66600335e-01 5.62182128e-01 -9.24705803e-01
-4.86363828e-01 2.09941000e-01 -2.77489834e-02 -5.40502965e-01
-8.64073336e-02 -7.06034720e-01 -1.34724295e+00 1.09358333e-01
-5.91255188e-01 1.96289033e-01 6.61356807e-01 1.12370837e+00
2.10277557e-01 5.35111308e-01 4.56480443e-01 -8.49762678e-01
-3.33452612e-01 -7.70686328e-01 -5.43452740e-01 3.62821847e-01
4.12399352e-01 -5.96185148e-01 -2.73696631e-01 3.66652524e-03]
|
[9.44581127166748, 1.238126277923584]
|
febffb69-bb6a-4c0f-969b-6a798dd3607e
|
homeomorphic-image-registration-via-conformal
|
2303.08113
| null |
https://arxiv.org/abs/2303.08113v2
|
https://arxiv.org/pdf/2303.08113v2.pdf
|
Learning Homeomorphic Image Registration via Conformal-Invariant Hyperelastic Regularisation
|
Deformable image registration is a fundamental task in medical image analysis and plays a crucial role in a wide range of clinical applications. Recently, deep learning-based approaches have been widely studied for deformable medical image registration and achieved promising results. However, existing deep learning image registration techniques do not theoretically guarantee topology-preserving transformations. This is a key property to preserve anatomical structures and achieve plausible transformations that can be used in real clinical settings. We propose a novel framework for deformable image registration. Firstly, we introduce a novel regulariser based on conformal-invariant properties in a nonlinear elasticity setting. Our regulariser enforces the deformation field to be smooth, invertible and orientation-preserving. More importantly, we strictly guarantee topology preservation yielding to a clinical meaningful registration. Secondly, we boost the performance of our regulariser through coordinate MLPs, where one can view the to-be-registered images as continuously differentiable entities. We demonstrate, through numerical and visual experiments, that our framework is able to outperform current techniques for image registration.
|
['Angelica I Aviles-Rivero', 'Carola-Bibiane Schönlieb', 'Jing Qin', 'Lihao Liu', 'Noémie Debroux', 'Jing Zou']
|
2023-03-14
| null | null | null | null |
['deformable-medical-image-registration', 'medical-image-registration']
|
['medical', 'medical']
|
[ 3.68465006e-01 2.37103313e-01 -8.83484632e-02 -4.48478341e-01
-7.36404061e-01 -6.37828648e-01 6.21715546e-01 4.42503273e-01
-4.94653255e-01 5.15647829e-01 1.62824765e-01 -9.80852172e-02
-6.26133442e-01 -7.51036584e-01 -7.41385937e-01 -8.85558367e-01
-2.10030973e-01 5.71338058e-01 2.55289137e-01 -4.01053876e-01
-1.21069299e-02 8.47703278e-01 -9.04017389e-01 -1.64679453e-01
8.71465445e-01 7.29481637e-01 -8.80170017e-02 3.58731389e-01
4.77119088e-01 2.34798968e-01 -7.43620051e-03 -2.93837667e-01
4.66346264e-01 -3.19197595e-01 -1.19564044e+00 -1.65307537e-01
4.89961118e-01 -1.22618698e-01 -2.45319128e-01 1.11145127e+00
5.71440399e-01 2.76621193e-01 6.14689827e-01 -8.75780940e-01
-6.96704566e-01 3.58027697e-01 -2.82988220e-01 9.68504995e-02
2.99063511e-02 -1.66483253e-01 8.04313242e-01 -3.99002999e-01
7.22736716e-01 9.26303923e-01 8.49781454e-01 4.46242124e-01
-1.40155196e+00 -1.16023831e-01 -1.88563243e-01 -8.11159536e-02
-1.23016703e+00 -2.79501706e-01 9.20130730e-01 -4.96057451e-01
3.99806619e-01 7.01086044e-01 5.54138303e-01 6.99230373e-01
7.26100147e-01 2.08716154e-01 1.04631019e+00 -2.15737313e-01
1.10633168e-02 -3.80365521e-01 -2.13088885e-01 6.54274106e-01
1.23650618e-01 6.02393923e-03 1.89550191e-01 -1.17258511e-01
1.28031182e+00 1.74206764e-01 -5.23303568e-01 -6.53545678e-01
-1.65495431e+00 8.26388478e-01 9.38595951e-01 7.79829383e-01
-4.27547336e-01 2.98749804e-01 1.75178587e-01 1.66808933e-01
5.52098811e-01 4.75823373e-01 -3.59961987e-02 2.50685781e-01
-6.87309027e-01 8.71827826e-02 3.81052822e-01 3.77225995e-01
3.11932534e-01 -2.39465073e-01 -3.53730440e-01 6.09905839e-01
3.23681086e-01 2.99861431e-01 6.29926562e-01 -9.05693114e-01
8.61010328e-02 5.01035571e-01 -1.59526616e-01 -1.26839912e+00
-6.70310140e-01 -3.83086711e-01 -1.44659400e+00 3.41049552e-01
4.95708704e-01 3.49666715e-01 -6.38268530e-01 1.59721649e+00
5.25251031e-01 3.80143911e-01 -1.13600187e-01 9.46670175e-01
5.85224807e-01 4.43415754e-02 1.53618857e-01 -1.48161784e-01
1.36393893e+00 -4.70838904e-01 -5.80423772e-01 2.67919451e-01
5.13732314e-01 -7.44303703e-01 8.36303353e-01 -2.14848351e-02
-1.36345673e+00 -2.88539678e-01 -8.56283784e-01 -1.83197200e-01
2.98560578e-02 -3.06812376e-01 4.45098251e-01 4.18157041e-01
-1.32132244e+00 9.20478046e-01 -1.32277238e+00 -1.86265156e-01
5.69905221e-01 7.86219001e-01 -7.85970688e-01 3.57770443e-01
-1.03105187e+00 9.86861229e-01 2.02343524e-01 4.12527978e-01
-4.24844950e-01 -7.87033975e-01 -8.89930308e-01 -1.33428559e-01
-4.35538366e-02 -9.70003009e-01 9.23958361e-01 -7.58480191e-01
-1.40868652e+00 1.39819419e+00 1.08272500e-01 -4.00661141e-01
1.01275253e+00 4.89851795e-02 -1.98972613e-01 1.75869823e-01
-1.53931513e-01 3.28192592e-01 6.34989917e-01 -1.22052860e+00
-1.39283866e-01 -3.87948424e-01 -1.71753168e-02 1.61779419e-01
-3.09707969e-01 -2.12838352e-02 -1.45382002e-01 -9.60858583e-01
3.36360544e-01 -1.06985438e+00 -5.82846463e-01 4.71750706e-01
-3.37871134e-01 4.57712496e-03 3.84390026e-01 -8.61432433e-01
7.63138950e-01 -1.93219948e+00 4.99460876e-01 5.25207937e-01
5.38908899e-01 2.34229535e-01 -3.54200415e-02 -8.51533562e-02
-2.34225124e-01 1.16809569e-01 -6.22624159e-01 -2.91640788e-01
-2.17467956e-02 1.62533760e-01 -4.26594168e-02 1.00873303e+00
1.09555259e-01 1.35222626e+00 -7.63463736e-01 -3.99185956e-01
3.44769627e-01 1.09576583e+00 -5.72065353e-01 2.26572305e-01
1.61067113e-01 1.24367177e+00 -5.41751623e-01 1.29152566e-01
6.07928991e-01 -7.19330236e-02 1.77134067e-01 -4.58601803e-01
-3.06496061e-02 -4.21282314e-02 -8.60550225e-01 1.84516060e+00
-4.74478364e-01 3.43587428e-01 2.05484107e-01 -1.36673117e+00
9.12005365e-01 4.29573089e-01 1.09817481e+00 -7.35451162e-01
4.20510650e-01 3.59882623e-01 -4.08698134e-02 -4.03477848e-01
1.79953694e-01 -4.18884575e-01 1.77775428e-01 3.58791083e-01
-3.00932616e-01 -1.51370764e-01 -3.80169421e-01 -3.34053755e-01
8.84390771e-01 -2.54778862e-02 2.57809758e-01 -7.08183169e-01
7.99000382e-01 -4.91548181e-01 3.14781904e-01 2.12677255e-01
-1.67572007e-01 8.05900753e-01 2.73239315e-01 -7.47811198e-01
-1.03986144e+00 -1.06190979e+00 -5.18391669e-01 5.85433841e-01
3.88740420e-01 4.49538440e-01 -9.15418029e-01 -5.15449405e-01
-1.99861482e-01 -3.17887105e-02 -7.66058266e-01 -2.32678592e-01
-1.11633074e+00 -7.58943498e-01 4.97628927e-01 4.23326373e-01
2.95889735e-01 -1.11291313e+00 -6.17925406e-01 3.38993996e-01
-1.50676116e-01 -1.08812380e+00 -7.20182180e-01 -9.86650586e-02
-1.09454954e+00 -1.11727262e+00 -1.10740280e+00 -1.14995193e+00
9.44589555e-01 -1.75124496e-01 9.52995658e-01 4.13786888e-01
-3.72880816e-01 4.04674470e-01 2.56970283e-02 2.13573396e-01
-8.47823381e-01 1.49317250e-01 6.95804283e-02 1.83827952e-01
-3.44376564e-01 -8.01564276e-01 -9.22216773e-01 4.16867375e-01
-1.19488895e+00 -1.42087460e-01 3.97595763e-01 6.19093597e-01
1.02294862e+00 -1.53228790e-01 2.93141901e-01 -8.58717918e-01
6.30082190e-01 -7.82642737e-02 -5.10716140e-01 4.39438730e-01
-3.34136158e-01 3.68493885e-01 3.45779061e-01 -3.90940815e-01
-7.01342165e-01 1.95045024e-01 -6.34977400e-01 -1.11346118e-01
-1.71433762e-01 3.79769623e-01 -1.87262353e-02 -8.19828510e-01
6.73794925e-01 6.82579577e-02 3.00212204e-01 -3.14995676e-01
2.49592185e-01 1.64451838e-01 7.79281497e-01 -6.92613542e-01
1.06436539e+00 8.87549877e-01 5.03227353e-01 -6.10697746e-01
-3.56072992e-01 -3.34966272e-01 -1.13721895e+00 -3.91505696e-02
1.01290607e+00 -4.71443444e-01 -1.08411765e+00 3.30116451e-01
-1.05404985e+00 -5.23023009e-01 -4.90063220e-01 4.72597063e-01
-7.65791595e-01 5.66735506e-01 -4.95186448e-01 -1.66991562e-01
-7.04820871e-01 -1.33696151e+00 1.13949478e+00 3.60109806e-02
-1.34960204e-01 -1.77288640e+00 3.60727042e-01 -9.36153904e-02
7.88616300e-01 9.50995028e-01 8.06978464e-01 -5.79904318e-01
-3.65170628e-01 -1.42418832e-01 -7.33033791e-02 7.66780302e-02
5.71067691e-01 -3.72268379e-01 -6.67841077e-01 -4.62044805e-01
3.16625416e-01 1.39187515e-01 6.02809429e-01 5.97076297e-01
1.20126927e+00 -3.28452915e-01 -3.25703233e-01 9.01414394e-01
1.30320013e+00 -1.19660705e-01 6.89564466e-01 2.36705646e-01
9.74765837e-01 6.96941555e-01 1.35586739e-01 4.64451872e-02
4.06721830e-01 1.03072345e+00 5.50705492e-01 -6.96621478e-01
-1.85397342e-01 1.82011634e-01 -4.73232605e-02 8.55997384e-01
-6.51603580e-01 2.27508157e-01 -8.57616067e-01 4.17827189e-01
-1.80990481e+00 -6.54389322e-01 -3.39174002e-01 2.43401098e+00
1.09139061e+00 -7.53115863e-02 -5.14892340e-02 1.12001710e-01
7.06683874e-01 -1.09899774e-01 -2.77199060e-01 -1.53126091e-01
-5.23987971e-03 5.39433599e-01 5.10626793e-01 7.66957223e-01
-1.33670545e+00 5.61444342e-01 5.90548229e+00 3.58102500e-01
-1.44780159e+00 2.35087886e-01 5.66999316e-01 5.36964178e-01
-6.16380930e-01 -3.84974569e-01 -2.07642496e-01 7.70337358e-02
4.31181043e-01 -2.11632445e-01 2.50196368e-01 2.46191353e-01
2.06807181e-01 4.51661617e-01 -1.19459033e+00 8.59851360e-01
-3.12771648e-02 -1.36886358e+00 -1.29928187e-01 2.87129015e-01
6.71876609e-01 -2.33613431e-01 2.66031295e-01 -4.52682793e-01
1.03589401e-01 -1.38780522e+00 4.17560995e-01 7.48460054e-01
9.34438884e-01 -5.09538531e-01 6.41002893e-01 2.62812525e-02
-1.15076518e+00 4.75321293e-01 -2.57527679e-01 4.54023182e-01
4.01212990e-01 4.74215537e-01 -5.67028463e-01 9.28744078e-01
3.85785520e-01 8.54151905e-01 -3.42533022e-01 1.15546250e+00
-5.96629903e-02 6.90838695e-02 -1.92025065e-01 5.63455105e-01
3.29779531e-03 -3.33767474e-01 4.59861428e-01 1.01405942e+00
3.21036100e-01 5.65942377e-02 2.01656103e-01 7.16868341e-01
-1.56908736e-01 2.79920548e-01 -4.38190222e-01 4.86960500e-01
-1.82535067e-01 1.35666370e+00 -1.12592411e+00 1.52984843e-01
1.08960457e-02 1.05129337e+00 1.19195595e-01 1.21668547e-01
-8.23693573e-01 1.95015408e-02 6.23065174e-01 3.15232724e-01
-3.20933461e-02 -3.31179142e-01 -1.48416728e-01 -1.15869069e+00
1.87555596e-01 -6.86832190e-01 3.05826306e-01 -1.04801640e-01
-1.30408740e+00 8.78944755e-01 -7.72083402e-02 -1.25631368e+00
1.80785973e-02 -5.44599414e-01 -5.54031312e-01 8.72097790e-01
-1.71582949e+00 -1.54814184e+00 -4.09718812e-01 8.22586894e-01
-1.68354216e-03 2.28977531e-01 9.81435359e-01 5.82777500e-01
1.09304292e-02 6.86074018e-01 -2.70338524e-02 1.87566504e-01
7.59886742e-01 -1.28794336e+00 5.19970536e-01 6.06897295e-01
1.71185896e-01 7.38864660e-01 5.15277386e-01 -4.70120132e-01
-1.27306652e+00 -1.11442673e+00 6.59347653e-01 -5.24609566e-01
5.75687170e-01 -4.95495275e-02 -1.21482348e+00 8.24866116e-01
5.82725257e-02 6.74932420e-01 4.53655928e-01 -4.90828753e-01
-6.86719790e-02 -1.03413858e-01 -1.36107743e+00 3.96457136e-01
9.61834669e-01 -3.85866255e-01 -3.56282830e-01 5.15167892e-01
5.35362303e-01 -9.06866014e-01 -1.53210187e+00 6.24522746e-01
7.22007334e-01 -7.03472733e-01 1.42389822e+00 -5.06105721e-01
2.22198561e-01 -1.32176727e-01 3.63556981e-01 -1.19786859e+00
-2.93018520e-01 -8.54517281e-01 2.07218155e-01 9.52870011e-01
-1.00375012e-01 -9.99567389e-01 4.34252858e-01 6.38809741e-01
-1.44017711e-01 -6.82656169e-01 -1.23513985e+00 -8.01411867e-01
6.86580777e-01 4.64165099e-02 5.62609911e-01 1.29311049e+00
-2.32288897e-01 -4.75957572e-01 -1.97877422e-01 1.40305892e-01
6.65234387e-01 -1.22190133e-01 2.99136013e-01 -1.45989931e+00
-1.38265565e-01 -7.58128524e-01 -6.13391042e-01 -7.26519525e-01
2.59613007e-01 -1.27742279e+00 1.14587083e-01 -1.52051747e+00
1.10481389e-01 -1.01478314e+00 -4.38321829e-01 5.74505985e-01
-8.16572309e-02 8.14690411e-01 -1.41938701e-01 4.54088390e-01
-2.55495548e-01 3.25368822e-01 1.57649720e+00 -7.64097646e-02
-2.61785716e-01 1.50250942e-01 -5.00407040e-01 5.90895355e-01
8.15626800e-01 -2.89268970e-01 -2.89061014e-02 -4.07132775e-01
1.96402818e-01 -1.45269379e-01 6.03368759e-01 -6.87583268e-01
1.61126822e-01 1.75076555e-02 6.03273660e-02 3.65491331e-01
-7.75488168e-02 -1.12212431e+00 5.06078780e-01 8.05410862e-01
-4.23787683e-01 4.45040315e-02 4.93431501e-02 1.24914363e-01
-1.50562719e-01 1.10575385e-01 9.95724440e-01 2.37737775e-01
-5.96570037e-02 7.21923232e-01 2.06985921e-01 1.69987902e-02
1.07063675e+00 -2.13092826e-02 6.43460900e-02 -1.20969161e-01
-8.83612156e-01 -1.74815997e-01 5.47206879e-01 4.22337532e-01
4.97753024e-01 -1.45655072e+00 -8.96688938e-01 2.15790257e-01
-1.29361242e-01 2.24402115e-01 1.96724534e-01 1.37941360e+00
-9.77701902e-01 1.94315806e-01 -3.56000662e-01 -8.20221364e-01
-1.31503165e+00 4.58812892e-01 7.16056705e-01 -5.11407077e-01
-9.44027185e-01 3.52732092e-01 3.28696638e-01 -5.30135274e-01
-1.53568434e-02 -6.84402406e-01 -3.65011483e-01 -3.62069994e-01
2.33799636e-01 7.72777498e-02 4.23861504e-01 -1.11686206e+00
-4.86024797e-01 1.18486345e+00 2.79963434e-01 1.19644441e-01
1.53932643e+00 8.16989243e-02 -4.24691856e-01 -6.60259183e-03
1.22443771e+00 8.64024907e-02 -1.24205756e+00 -3.05143565e-01
1.91257838e-02 -1.74218491e-01 -1.29844114e-01 -2.73140311e-01
-1.36063027e+00 7.91311622e-01 7.06942379e-01 4.63902533e-01
1.05686653e+00 1.05564542e-01 8.62664342e-01 5.65384654e-03
2.28251562e-01 -3.38462204e-01 -1.18711382e-01 9.44686309e-02
1.16840041e+00 -1.31270874e+00 6.78007156e-02 -5.92002273e-01
-2.35036075e-01 1.19199812e+00 -1.23892456e-01 -3.60927761e-01
6.82126820e-01 2.77340800e-01 1.79250881e-01 -1.40562162e-01
2.02126101e-01 7.11214021e-02 9.53259528e-01 5.61883748e-01
6.71178758e-01 3.35177928e-02 -4.88266945e-01 1.76991642e-01
-1.48888811e-01 -1.99639440e-01 3.08090877e-02 6.78250790e-01
-5.11123650e-02 -1.46321929e+00 -3.39286864e-01 -1.19298752e-02
-6.91607952e-01 1.71695679e-01 -4.57251184e-02 7.36091554e-01
-1.46896228e-01 3.19410533e-01 8.13694671e-02 1.40484989e-01
4.55952942e-01 -3.40913862e-01 8.10426235e-01 -3.75288129e-01
-8.54061604e-01 1.23941801e-01 -6.39086306e-01 -6.54136181e-01
-8.13603938e-01 -4.48196977e-01 -1.43366683e+00 -1.46856487e-01
-3.03353313e-02 -1.71537146e-01 6.19946480e-01 1.07750559e+00
1.96419671e-01 6.76629484e-01 6.12396181e-01 -9.42289412e-01
-3.75365704e-01 -4.05835629e-01 -3.44663382e-01 8.12328339e-01
5.33129871e-01 -4.50277776e-01 -5.84799843e-03 4.21876341e-01]
|
[14.038735389709473, -2.5914697647094727]
|
da1cc2d4-ef27-41f4-b9f4-3e4969fad55b
|
are-all-point-clouds-suitable-for-completion
|
2303.01804
| null |
https://arxiv.org/abs/2303.01804v1
|
https://arxiv.org/pdf/2303.01804v1.pdf
|
Are All Point Clouds Suitable for Completion? Weakly Supervised Quality Evaluation Network for Point Cloud Completion
|
In the practical application of point cloud completion tasks, real data quality is usually much worse than the CAD datasets used for training. A small amount of noisy data will usually significantly impact the overall system's accuracy. In this paper, we propose a quality evaluation network to score the point clouds and help judge the quality of the point cloud before applying the completion model. We believe our scoring method can help researchers select more appropriate point clouds for subsequent completion and reconstruction and avoid manual parameter adjustment. Moreover, our evaluation model is fast and straightforward and can be directly inserted into any model's training or use process to facilitate the automatic selection and post-processing of point clouds. We propose a complete dataset construction and model evaluation method based on ShapeNet. We verify our network using detection and flow estimation tasks on KITTI, a real-world dataset for autonomous driving. The experimental results show that our model can effectively distinguish the quality of point clouds and help in practical tasks.
|
['Shaojie Shen', 'Xiaozhi Chen', 'Peiliang Li', 'Jieqi Shi']
|
2023-03-03
| null | null | null | null |
['point-cloud-completion']
|
['computer-vision']
|
[-3.72506887e-01 -4.52006847e-01 -7.43470788e-02 -7.26184726e-01
-5.42256236e-01 -5.32125831e-01 2.18494490e-01 1.00475304e-01
-4.09747422e-01 2.39262313e-01 -4.98480618e-01 -3.99520725e-01
-1.13791965e-01 -1.10396409e+00 -6.55261397e-01 -2.98232198e-01
6.27629012e-02 8.26409936e-01 5.90576410e-01 -2.49250039e-01
4.78154868e-01 1.01979029e+00 -2.09520888e+00 -8.92269760e-02
1.16814792e+00 1.16814268e+00 5.16927779e-01 5.46389699e-01
-2.94156015e-01 3.73522758e-01 -7.96301723e-01 -2.14723080e-01
5.63908339e-01 4.17410970e-01 -2.74682403e-01 -2.03546602e-02
6.04196429e-01 -4.27590340e-01 -7.86208287e-02 1.04204118e+00
5.05454838e-01 9.52665135e-02 3.94514799e-01 -1.59466517e+00
-5.07437997e-02 1.04204640e-01 -2.65718877e-01 -9.47263092e-02
3.72034125e-02 4.74099040e-01 7.45651424e-01 -1.23280597e+00
4.11254257e-01 1.11860752e+00 8.36551309e-01 2.26197854e-01
-6.45014107e-01 -1.24465072e+00 1.66223794e-02 3.83921146e-01
-1.48900223e+00 -5.18356204e-01 9.91981924e-01 -5.98074734e-01
5.22066474e-01 2.57630259e-01 8.89150202e-01 3.24909657e-01
-3.88219208e-02 5.98582447e-01 7.15425491e-01 1.40138596e-01
3.64409655e-01 4.69255596e-02 7.32871145e-02 7.04560816e-01
4.55077916e-01 2.93938309e-01 -2.50205100e-01 -4.52147424e-02
1.01807857e+00 -3.02573927e-02 -6.70476630e-03 -6.11522436e-01
-1.19255137e+00 6.48146451e-01 6.69892013e-01 -8.50218758e-02
-2.34881744e-01 2.38720790e-01 2.47422189e-01 2.83832997e-01
2.52889216e-01 3.98643732e-01 -3.73240411e-01 -1.15607306e-01
-9.79439199e-01 3.60139430e-01 5.00785768e-01 1.34506178e+00
1.05793214e+00 -1.83762982e-01 1.28826782e-01 7.90245831e-01
2.94534236e-01 8.16198766e-01 2.12714430e-02 -1.22755027e+00
5.93252182e-01 8.68883193e-01 2.71050125e-01 -1.19780815e+00
-3.16502064e-01 -2.77156740e-01 -6.05147064e-01 8.65162671e-01
2.65538096e-01 1.74321935e-01 -9.06124592e-01 9.76055980e-01
4.48272705e-01 5.27464092e-01 -1.97314903e-01 1.10804415e+00
1.08564162e+00 4.46776956e-01 -1.87840134e-01 1.55196726e-01
1.07573307e+00 -5.15176475e-01 -5.81391037e-01 -1.20541371e-01
6.56563222e-01 -8.68951023e-01 1.16352594e+00 5.86957872e-01
-9.02283847e-01 -1.02287376e+00 -1.22581613e+00 -1.57713696e-01
-1.91825852e-01 5.49727261e-01 7.94719160e-01 3.46296281e-01
-8.78359437e-01 8.31438780e-01 -9.03387189e-01 5.11428714e-02
5.72337091e-01 4.86804664e-01 -3.99614722e-01 -1.57054916e-01
-8.11295569e-01 9.85248923e-01 1.93195358e-01 3.37750047e-01
-8.06363583e-01 -7.95173109e-01 -7.14635015e-01 -1.46302417e-01
2.73014992e-01 -7.38225102e-01 1.31748128e+00 -3.99926215e-01
-1.29218507e+00 6.58342302e-01 -1.53298855e-01 -2.30122805e-01
6.34649098e-01 -2.48140380e-01 -2.23480895e-01 8.88416320e-02
3.31935078e-01 8.11790884e-01 7.78482676e-01 -1.37188566e+00
-1.04703081e+00 -4.95968372e-01 -5.69878109e-02 2.23215565e-01
1.71908244e-01 -2.63214678e-01 -8.24208736e-01 -1.06990717e-01
4.90868628e-01 -8.53298306e-01 -3.91461164e-01 5.01020253e-01
-1.36494130e-01 -3.75032663e-01 1.07448876e+00 -2.60419577e-01
9.81704772e-01 -2.08016062e+00 -4.36980337e-01 5.07785678e-01
3.23046505e-01 2.77216613e-01 -8.09015632e-02 -7.10734800e-02
1.23195924e-01 -4.22616582e-03 -4.86174487e-02 -3.05081666e-01
-1.15010969e-01 3.81980568e-01 -1.73202857e-01 4.37184125e-01
2.76745707e-01 6.20425224e-01 -8.57434869e-01 -5.83522320e-01
6.78868115e-01 3.95086855e-01 -3.52826536e-01 2.30884477e-01
-6.06051355e-04 2.27309749e-01 -6.69434130e-01 6.97804809e-01
1.00891948e+00 -9.40166563e-02 -4.95473951e-01 -3.95059943e-01
-1.70456365e-01 3.39641035e-01 -1.54576862e+00 1.64461696e+00
-5.84825754e-01 8.16561341e-01 -4.26303186e-02 -4.10280496e-01
1.40061343e+00 -1.80520132e-01 5.49350441e-01 -6.16908193e-01
1.76484227e-01 2.12205440e-01 -7.61884218e-03 -4.45693970e-01
8.62854302e-01 2.27055624e-01 2.16793075e-01 8.92283171e-02
-4.77434039e-01 -8.36122811e-01 1.11310713e-01 3.46542299e-02
8.93086672e-01 1.02129675e-01 2.69717164e-02 -9.79599431e-02
3.21206123e-01 4.17681277e-01 7.55493164e-01 5.53635120e-01
-4.19078320e-01 7.44137764e-01 7.98189268e-02 -6.00976825e-01
-1.16736138e+00 -8.90436530e-01 -3.83931130e-01 6.27949715e-01
8.10686171e-01 -4.43980247e-01 -3.74594331e-01 -5.71365952e-01
3.77453148e-01 5.66271663e-01 -2.48093158e-01 -7.63947964e-02
-5.83262801e-01 -1.51511565e-01 3.43138456e-01 6.78232312e-01
4.94088203e-01 -9.63288069e-01 -5.82503676e-01 2.97850817e-02
-1.61935031e-01 -1.07405901e+00 -1.69820085e-01 -1.53361708e-01
-1.29381061e+00 -1.23290002e+00 2.10535564e-02 -6.26412213e-01
8.39570045e-01 7.88686752e-01 1.27559555e+00 5.88598847e-01
1.06641464e-01 3.31615470e-02 -2.45074615e-01 -7.30331838e-01
-2.58193880e-01 2.56433953e-02 3.20175402e-02 -4.93465126e-01
5.58611929e-01 -4.34744835e-01 -7.32609034e-01 8.49710584e-01
-6.84310675e-01 3.41852531e-02 4.17691022e-01 4.50934142e-01
7.91019261e-01 3.31228703e-01 1.58143729e-01 -4.81317669e-01
5.48331261e-01 -2.57732067e-02 -1.15548229e+00 -1.28746077e-01
-6.20622098e-01 -8.67630541e-02 3.82521152e-01 -1.37423843e-01
-6.55219257e-01 3.49038482e-01 -2.83883572e-01 -8.70766222e-01
-1.82057202e-01 2.71646440e-01 -1.99203208e-01 -3.62727106e-01
6.39470637e-01 -2.97045708e-01 1.12172626e-01 -4.97007430e-01
1.96713224e-01 6.50534093e-01 6.22592568e-01 -4.01891410e-01
1.14764786e+00 6.25732720e-01 4.14969288e-02 -6.43643737e-01
-3.82582366e-01 -7.41708338e-01 -9.52236056e-01 -4.72150922e-01
5.28812468e-01 -1.02454913e+00 -9.28419292e-01 2.68826842e-01
-1.33177567e+00 -5.13444431e-02 -1.56657547e-01 5.10444582e-01
-3.59372616e-01 2.77807534e-01 -2.04040930e-01 -7.00040042e-01
-2.25191548e-01 -1.48163664e+00 1.39153624e+00 1.08323231e-01
2.63352990e-01 -6.49062216e-01 -1.25897557e-01 2.85691023e-01
8.49144682e-02 -2.39939112e-02 2.93163747e-01 -9.88027453e-02
-1.06450021e+00 -4.44398314e-01 -3.73759925e-01 3.13892215e-01
-2.06353497e-02 5.03276765e-01 -1.09458041e+00 -1.21088490e-01
-1.04635894e-01 9.55968630e-03 7.56670594e-01 3.68257523e-01
1.40074956e+00 3.18371475e-01 -4.28701758e-01 7.78948545e-01
1.34108007e+00 1.12907238e-01 8.74529183e-01 3.84661734e-01
7.93231010e-01 6.14575267e-01 1.34819293e+00 3.15953255e-01
4.23248291e-01 6.05199635e-01 8.24641109e-01 -2.35225633e-01
2.97995061e-02 -3.71846646e-01 1.32587865e-01 9.55735207e-01
-1.83607548e-01 2.28980541e-01 -1.04450583e+00 4.19977784e-01
-1.74202681e+00 -7.72880912e-01 -5.23340166e-01 2.32185888e+00
4.42698926e-01 4.06448841e-01 -9.56551731e-02 2.20247805e-01
6.74381256e-01 -2.57288843e-01 -5.45142829e-01 3.31023186e-02
2.51840770e-01 1.20797545e-01 8.24252903e-01 4.55107838e-01
-1.02961755e+00 1.04529464e+00 6.61069822e+00 7.97478259e-01
-1.22640455e+00 -2.14420035e-01 2.20715672e-01 9.95703414e-02
-1.14137746e-01 5.66803627e-02 -9.52499628e-01 4.63105202e-01
5.89704394e-01 -9.60764140e-02 1.08795706e-02 1.21320844e+00
5.42643726e-01 -1.18049718e-01 -1.11166155e+00 1.34411371e+00
-4.23824400e-01 -1.40572000e+00 -9.44536924e-02 4.44670618e-02
3.63372087e-01 4.10565168e-01 -1.06193960e-01 1.26543850e-01
4.60612297e-01 -8.67094934e-01 7.80763924e-01 5.42462468e-01
9.04534042e-01 -8.54091644e-01 8.17704856e-01 5.56620121e-01
-1.32031238e+00 4.57261130e-03 -9.99100029e-01 -6.08822480e-02
6.85820878e-02 8.20536137e-01 -1.18021977e+00 4.85093832e-01
7.04240143e-01 8.05896342e-01 -8.21545064e-01 1.51708293e+00
-1.78553253e-01 4.59380507e-01 -5.57891190e-01 -4.08363752e-02
-1.00681089e-01 -3.56134325e-01 6.00132108e-01 8.56991768e-01
5.29185414e-01 -1.48525881e-03 3.23395371e-01 9.01505947e-01
1.58295155e-01 9.11994576e-02 -7.45469987e-01 5.07955432e-01
7.86140740e-01 1.36699295e+00 -6.19434893e-01 -4.15163338e-01
-8.53044167e-02 2.45948955e-01 1.67274520e-01 7.44728148e-02
-8.17594171e-01 -4.95957226e-01 9.38736975e-01 2.21781209e-01
1.32162467e-01 -5.38368106e-01 -7.01380134e-01 -9.94138181e-01
3.24352473e-01 -3.85017276e-01 -1.47149101e-01 -1.08567905e+00
-1.13704646e+00 7.28311658e-01 -8.68523642e-02 -2.04888892e+00
1.34075405e-02 -6.55886710e-01 -7.69333601e-01 7.31269479e-01
-1.75049412e+00 -9.12300766e-01 -9.50639784e-01 4.60873395e-01
4.26006585e-01 -1.38696447e-01 3.59759867e-01 4.86438543e-01
-4.32094902e-01 2.73829043e-01 -2.22345740e-01 1.21868111e-01
6.42692506e-01 -1.05484486e+00 6.45742714e-01 8.22176456e-01
-1.09013900e-01 5.29638112e-01 6.86965227e-01 -8.14280868e-01
-1.35595620e+00 -1.29383099e+00 4.85728800e-01 -6.59584284e-01
3.30756485e-01 -2.68218368e-01 -1.10244608e+00 2.42382824e-01
-4.58009034e-01 5.03350496e-02 1.83487415e-01 5.02361879e-02
-5.03210053e-02 -4.40771401e-01 -1.18213749e+00 2.42948547e-01
1.26436746e+00 -2.27956906e-01 -4.83912140e-01 2.16936424e-01
7.74496496e-01 -5.80315650e-01 -8.06963980e-01 7.15463281e-01
4.82827187e-01 -9.91244912e-01 9.82652128e-01 -1.67174935e-01
2.22577825e-01 -8.48811805e-01 6.59594834e-02 -1.29865432e+00
-4.02327001e-01 -1.54460788e-01 2.79627025e-01 7.49309719e-01
2.58262485e-01 -4.25807923e-01 1.11148560e+00 5.75903594e-01
-4.75310802e-01 -4.15188342e-01 -9.54216421e-01 -6.56079471e-01
-2.56095916e-01 -1.12618136e+00 1.23218846e+00 8.05140197e-01
-4.93099153e-01 1.57370746e-01 1.19041748e-01 5.06167114e-01
6.41778946e-01 1.45132542e-01 1.39054739e+00 -1.80534673e+00
2.45687053e-01 -3.40092212e-01 -7.59713471e-01 -1.17390478e+00
-8.86785910e-02 -7.83140182e-01 2.46726990e-01 -1.66925418e+00
-3.51188809e-01 -1.22086513e+00 7.46095106e-02 3.21520835e-01
-5.08336164e-02 2.15974003e-01 1.51076227e-01 5.78028798e-01
-5.34439087e-01 5.32478034e-01 1.60027242e+00 -1.35474131e-01
-4.29055840e-01 3.95898044e-01 -4.19488847e-01 7.61929512e-01
7.44776726e-01 -4.17051017e-01 -4.73419756e-01 -6.19335592e-01
3.11296374e-01 -1.01487981e-02 5.06573677e-01 -1.42431271e+00
4.09519285e-01 -2.77733713e-01 4.35628355e-01 -1.23724282e+00
3.36103499e-01 -1.28892231e+00 1.36019915e-01 1.73486724e-01
6.49843812e-02 2.66024530e-01 2.66144335e-01 2.69477248e-01
-2.47180387e-01 -2.42602289e-01 6.71894312e-01 1.06216103e-01
-8.40335488e-01 6.91219687e-01 2.73003399e-01 -5.10756910e-01
1.01673722e+00 -6.78789735e-01 -1.43043265e-01 -2.75463015e-01
-5.95229149e-01 6.27968848e-01 7.76199281e-01 4.92237628e-01
1.08316731e+00 -1.60934150e+00 -6.26280844e-01 5.14131367e-01
4.52942908e-01 6.21987224e-01 -4.57401425e-02 4.06631470e-01
-9.08531547e-01 -2.25312659e-03 -1.66483909e-01 -1.20960867e+00
-1.26623189e+00 3.31330419e-01 3.12971056e-01 1.91151232e-01
-6.94210231e-01 4.83173281e-01 -4.83539328e-03 -6.13828480e-01
2.39956245e-01 -9.20632899e-01 -3.34084690e-01 -2.21833929e-01
4.71437275e-01 5.43458462e-01 4.56978351e-01 -6.71954513e-01
-1.98237807e-01 7.60132551e-01 1.95561603e-01 1.19190007e-01
1.16461992e+00 4.04589549e-02 5.86810932e-02 2.63648838e-01
1.00436914e+00 -1.93751514e-01 -1.42752063e+00 -9.63539332e-02
-2.43545339e-01 -9.25023377e-01 2.62615174e-01 -4.31571215e-01
-1.23355901e+00 9.97539937e-01 8.27102125e-01 3.94969545e-02
9.30844665e-01 -1.86224520e-01 8.31886590e-01 5.95487595e-01
7.01171935e-01 -1.17134857e+00 -2.41138637e-01 3.23377460e-01
1.00544906e+00 -1.49073052e+00 3.89038697e-02 -7.14075923e-01
-5.16924322e-01 1.14580607e+00 8.65982771e-01 -3.33203167e-01
7.21880198e-01 3.59812528e-01 3.08082998e-01 -4.53208804e-01
-7.04117537e-01 -2.01796889e-01 5.06477118e-01 8.56374025e-01
-1.09873332e-01 1.92752466e-01 1.59275562e-01 1.52729243e-01
-7.76278973e-01 1.66004419e-01 4.81364459e-01 7.81938255e-01
-9.22264218e-01 -9.80285823e-01 -6.49738073e-01 6.61226332e-01
2.36680880e-01 2.54923344e-01 -1.94065109e-01 6.96848989e-01
2.72682875e-01 9.31727290e-01 4.36204493e-01 -7.07920790e-01
8.40745389e-01 -3.81025672e-01 1.95925936e-01 -4.77120817e-01
-4.05146241e-01 -1.46027148e-01 -1.23024501e-01 -8.26325595e-01
-2.77593851e-01 -5.34712851e-01 -1.38051879e+00 -5.85484922e-01
-6.62333965e-01 9.30374786e-02 1.10312784e+00 5.50220013e-01
5.91026366e-01 3.64455104e-01 8.34887087e-01 -1.00080061e+00
-3.70968819e-01 -8.44644487e-01 -3.03437829e-01 4.73061204e-01
2.23746896e-01 -1.01320636e+00 -3.05539340e-01 -8.81873816e-02]
|
[7.912431716918945, -2.944645643234253]
|
e0da45ec-d582-4a3d-84e3-011de35954bd
|
neural-block-slot-representations
|
2211.01177
| null |
https://arxiv.org/abs/2211.01177v3
|
https://arxiv.org/pdf/2211.01177v3.pdf
|
Neural Systematic Binder
|
The key to high-level cognition is believed to be the ability to systematically manipulate and compose knowledge pieces. While token-like structured knowledge representations are naturally provided in text, it is elusive how to obtain them for unstructured modalities such as scene images. In this paper, we propose a neural mechanism called Neural Systematic Binder or SysBinder for constructing a novel structured representation called Block-Slot Representation. In Block-Slot Representation, object-centric representations known as slots are constructed by composing a set of independent factor representations called blocks, to facilitate systematic generalization. SysBinder obtains this structure in an unsupervised way by alternatingly applying two different binding principles: spatial binding for spatial modularity across the full scene and factor binding for factor modularity within an object. SysBinder is a simple, deterministic, and general-purpose layer that can be applied as a drop-in module in any arbitrary neural network and on any modality. In experiments, we find that SysBinder provides significantly better factor disentanglement within the slots than the conventional object-centric methods, including, for the first time, in visually complex scene images such as CLEVR-Tex. Furthermore, we demonstrate factor-level systematicity in controlled scene generation by decoding unseen factor combinations.
|
['Sungjin Ahn', 'Yeongbin Kim', 'Gautam Singh']
|
2022-11-02
| null | null | null | null |
['scene-generation', 'systematic-generalization']
|
['computer-vision', 'reasoning']
|
[ 2.86947012e-01 1.11060023e-01 -7.85715878e-02 -2.71592766e-01
-3.71372551e-01 -7.37574220e-01 7.35789537e-01 -8.57209191e-02
-1.89829126e-01 4.62481111e-01 4.14214015e-01 -3.07208449e-01
-4.30369318e-01 -7.55313277e-01 -9.83244479e-01 -7.06044316e-01
2.50964314e-01 3.24465185e-01 2.93226331e-01 -2.44823053e-01
2.98486590e-01 5.28253317e-01 -1.66191351e+00 6.47094905e-01
7.21979320e-01 5.09242415e-01 6.55610681e-01 3.98551315e-01
-2.98756629e-01 7.60727227e-01 -5.29419839e-01 -2.67861158e-01
3.34144980e-01 -5.68462968e-01 -7.68181324e-01 2.46735558e-01
3.30521405e-01 -2.77714729e-01 -3.04725796e-01 8.98624897e-01
1.42801464e-01 3.70431036e-01 9.25695121e-01 -9.86684799e-01
-1.02104163e+00 1.02293122e+00 -5.13533294e-01 2.27495998e-01
3.82512629e-01 5.95119074e-02 1.12093341e+00 -1.03621888e+00
6.20179653e-01 1.57605910e+00 8.44976902e-02 5.51503479e-01
-1.57412708e+00 -4.79022294e-01 3.99883896e-01 6.29964173e-02
-1.29548120e+00 -5.61320543e-01 5.97682059e-01 -5.48371494e-01
1.31020272e+00 2.85071224e-01 7.61012971e-01 8.38224769e-01
9.89153087e-02 9.65107024e-01 1.06574821e+00 -5.85908175e-01
4.24861550e-01 -2.86894798e-01 2.81361610e-01 9.35137928e-01
4.52679396e-01 1.33323092e-02 -1.05057359e+00 6.63305074e-02
1.36473775e+00 4.77911532e-03 -1.75257117e-01 -5.74937284e-01
-1.57761550e+00 4.89898115e-01 6.85867131e-01 3.66476536e-01
-1.96970105e-01 3.71471047e-01 1.17307261e-01 1.59705937e-01
-1.34610206e-01 8.15917730e-01 -3.68199110e-01 3.01844180e-01
-6.48514211e-01 4.37811822e-01 4.27319884e-01 1.01876879e+00
8.74425054e-01 2.59183317e-01 -2.55073905e-01 7.08980381e-01
2.85847962e-01 3.79653871e-01 6.38709128e-01 -7.92882621e-01
3.00060749e-01 7.48689234e-01 -9.87967998e-02 -9.52066779e-01
-5.85683048e-01 -4.96796280e-01 -8.81374836e-01 1.11132339e-01
3.48807842e-01 2.26204708e-01 -1.23960173e+00 2.19742203e+00
2.60535210e-01 -2.82932758e-01 1.60422567e-02 9.31404948e-01
8.91883731e-01 6.18150592e-01 2.04109088e-01 -3.13891396e-02
1.80465078e+00 -9.39239979e-01 -7.39404619e-01 -3.70308101e-01
3.75457346e-01 -3.00981253e-01 1.03261614e+00 6.84103549e-01
-1.19573188e+00 -6.04611099e-01 -1.10918760e+00 -3.68826151e-01
-6.27231777e-01 7.28384778e-02 1.29127252e+00 4.71973896e-01
-1.11972606e+00 1.95106447e-01 -6.51400983e-01 -1.92123979e-01
6.84851468e-01 2.73505718e-01 -8.70277584e-01 -9.05777514e-02
-9.14726555e-01 9.15271282e-01 6.78552568e-01 9.29950699e-02
-1.18912625e+00 -5.61806738e-01 -9.36514795e-01 3.73573154e-01
6.14064217e-01 -1.27005672e+00 1.07218373e+00 -9.44484413e-01
-1.28063464e+00 7.82837987e-01 -4.92949724e-01 -5.99380350e-03
-5.24548054e-01 -1.07185274e-01 -5.67853302e-02 1.38595432e-01
1.85354888e-01 8.06031227e-01 1.20712936e+00 -1.57447350e+00
-1.62565783e-01 -4.73960370e-01 5.98738253e-01 3.58649909e-01
2.17768885e-02 2.24341894e-03 -4.17230606e-01 -8.38433385e-01
6.55725479e-01 -4.84999418e-01 -2.54627198e-01 3.49742360e-02
-4.58764076e-01 -1.46245241e-01 -7.95086473e-03 -5.42496443e-01
1.01712263e+00 -2.33762574e+00 8.45090568e-01 1.58241719e-01
6.30716503e-01 -4.12000716e-02 -2.13229120e-01 3.84400368e-01
-3.60380143e-01 1.54939547e-01 -1.87412798e-01 -9.96452570e-02
1.74408123e-01 3.40918273e-01 -2.95862496e-01 1.81071162e-01
3.78356218e-01 1.30321229e+00 -9.50670183e-01 -2.24026799e-01
-3.98965776e-02 3.57435644e-02 -7.95155168e-01 5.54547971e-03
-5.54369509e-01 1.57058105e-01 -2.54171371e-01 6.29872680e-01
6.77214503e-01 -5.44671476e-01 3.32816631e-01 -2.86619037e-01
1.47358522e-01 3.65108371e-01 -1.23268557e+00 2.29103327e+00
-2.25282386e-01 5.04325032e-01 -7.56926164e-02 -1.08315814e+00
5.82543612e-01 1.47969499e-01 -2.49704719e-01 -7.67416120e-01
1.19860776e-01 -1.16634309e-01 2.23346040e-01 -1.05728678e-01
7.03100085e-01 -4.02899593e-01 -2.59943426e-01 5.63726902e-01
6.39433503e-01 -7.11705387e-02 3.28325868e-01 7.19236016e-01
1.08123243e+00 1.88723639e-01 6.24178171e-01 -4.04001713e-01
6.50379956e-02 -1.53481200e-01 2.05093011e-01 8.79749179e-01
1.93835467e-01 4.14631367e-01 5.31470418e-01 -3.60997051e-01
-7.23626673e-01 -1.70755351e+00 -5.29305711e-02 1.43382585e+00
1.65965036e-01 -8.04527342e-01 -5.15872896e-01 -2.45861247e-01
-2.31004551e-01 6.30373478e-01 -6.80686474e-01 -2.35683531e-01
-1.81694791e-01 -5.66523731e-01 4.49038744e-01 5.98389268e-01
4.70997334e-01 -9.68331099e-01 -5.05144835e-01 1.19050175e-01
5.80484346e-02 -8.21933091e-01 1.63513552e-02 6.35525823e-01
-7.26628780e-01 -9.43729222e-01 -3.61149073e-01 -7.18134940e-01
1.05373645e+00 7.04134762e-01 1.11945307e+00 7.68245161e-02
-3.07241857e-01 3.23802650e-01 -1.27666533e-01 -2.19758645e-01
1.07477717e-01 -1.39760986e-01 5.43442592e-02 -1.56246290e-01
1.52432606e-01 -7.97837198e-01 -4.03470784e-01 3.11193705e-01
-1.45625067e+00 6.18442595e-01 5.40344059e-01 9.57826674e-01
4.85437155e-01 1.38591766e-01 3.64662558e-01 -6.94190204e-01
6.53766334e-01 -3.62481117e-01 -4.80530858e-01 2.67714679e-01
1.05688848e-01 5.23609519e-01 4.56059963e-01 -4.93750155e-01
-1.11554360e+00 1.64060518e-01 3.18676233e-01 -1.34496123e-01
-1.93704382e-01 5.59059262e-01 -5.64361572e-01 2.46885940e-01
1.13985741e+00 5.34955263e-01 -3.08293849e-01 -3.76352102e-01
9.94055569e-01 1.05868421e-01 5.61382711e-01 -1.09030867e+00
1.02191806e+00 5.17308235e-01 7.75812333e-03 -7.74728298e-01
-9.17266488e-01 -2.06982210e-01 -7.25545526e-01 -1.25898849e-02
9.59856212e-01 -1.17413497e+00 -6.50057435e-01 4.43957299e-01
-1.21717656e+00 -2.39319384e-01 -3.66602957e-01 1.46156982e-01
-5.08001626e-01 4.21733297e-02 -2.34677672e-01 -3.86339247e-01
2.40345314e-01 -9.97747064e-01 1.07347977e+00 3.56141061e-01
-3.51014704e-01 -6.79275692e-01 -1.20386362e-01 2.73106515e-01
2.52450764e-01 -1.81554154e-01 1.17102587e+00 -4.95419711e-01
-1.05124366e+00 2.88232803e-01 -3.52098554e-01 2.16703564e-02
1.85302764e-01 -3.43583167e-01 -9.56702292e-01 4.46045101e-02
-1.12026863e-01 -5.42164087e-01 1.20970130e+00 7.28439763e-02
1.23206103e+00 -2.49675527e-01 -2.82265276e-01 6.20462596e-01
1.21251023e+00 1.00023396e-01 8.84905696e-01 1.02505878e-01
8.70999694e-01 3.57463986e-01 -1.33952811e-01 1.06488332e-01
4.77225453e-01 3.69158179e-01 1.65628359e-01 -1.66179985e-02
-1.79436907e-01 -4.09983575e-01 1.13898650e-01 7.17412651e-01
-1.91223249e-01 -2.30448037e-01 -7.65974998e-01 4.69849348e-01
-1.92859495e+00 -1.13215995e+00 2.38293409e-01 1.90062904e+00
9.90153015e-01 -8.06034915e-03 -2.22718507e-01 5.49495146e-02
3.93600464e-01 2.00694636e-01 -4.91198629e-01 -2.19742388e-01
-4.81306314e-01 2.38544300e-01 2.55327225e-02 3.55077893e-01
-7.74830639e-01 1.35823083e+00 6.58210325e+00 1.01473284e+00
-6.07686579e-01 1.01533934e-01 2.03895271e-01 -2.00294927e-01
-7.68845499e-01 2.21371561e-01 -5.64775348e-01 -1.52744472e-01
4.59918708e-01 1.10389732e-01 9.06895697e-01 6.33470953e-01
-2.84733653e-01 -4.29160208e-01 -1.29152763e+00 1.18790364e+00
1.81420878e-01 -1.56349015e+00 5.52134931e-01 -2.44569167e-01
6.40826941e-01 -4.83617246e-01 -5.61039969e-02 2.78481513e-01
4.74957138e-01 -1.38504446e+00 9.11488652e-01 6.65430367e-01
6.24132037e-01 -4.20195311e-01 -2.19705239e-01 3.22758138e-01
-1.12656939e+00 1.71767324e-02 -6.68657720e-01 -4.19216543e-01
1.09470084e-01 4.34661150e-01 -2.48568013e-01 4.61526722e-01
3.95220399e-01 5.23444116e-01 -6.53338850e-01 8.39042544e-01
-5.87813973e-01 1.86859906e-01 -1.94959968e-01 5.91770075e-02
-1.22540422e-01 2.62594044e-01 4.06686366e-01 9.36694562e-01
-9.60411727e-02 3.40528220e-01 -2.01680601e-01 1.18194473e+00
-9.54310875e-03 -1.70098275e-01 -6.94488406e-01 -1.77309647e-01
3.14499557e-01 1.24442422e+00 -9.74172950e-01 -5.00421166e-01
-2.57640064e-01 1.06934702e+00 7.96324849e-01 6.69228792e-01
-5.00400603e-01 -2.91172974e-02 6.13729954e-01 -7.07685277e-02
6.85997069e-01 -6.27479672e-01 -5.05760491e-01 -1.56789398e+00
-4.05754931e-02 -8.81140471e-01 3.61696780e-02 -1.30028379e+00
-1.27653515e+00 3.81463975e-01 3.54784995e-01 -8.42993140e-01
4.10268735e-03 -1.07216990e+00 -2.02778012e-01 9.19283926e-01
-9.03068900e-01 -1.13160002e+00 -4.52911705e-02 1.07791197e+00
4.72171336e-01 -3.45700592e-01 1.06257308e+00 -2.08807483e-01
-3.91574919e-01 4.23018813e-01 -2.36744180e-01 5.62898777e-02
2.67856956e-01 -1.28529572e+00 4.13418680e-01 9.97330785e-01
6.65651441e-01 1.49020731e+00 3.92026216e-01 -5.48172653e-01
-1.80652964e+00 -5.85797846e-01 5.82199633e-01 -5.83655119e-01
6.47398710e-01 -9.76037025e-01 -8.90294969e-01 6.50959492e-01
2.51325935e-01 -1.19873293e-01 8.53194416e-01 4.80645090e-01
-1.21316063e+00 1.37894213e-01 -8.23860943e-01 1.03548896e+00
1.52879632e+00 -7.52491176e-01 -1.21512604e+00 2.76883692e-01
9.95515764e-01 -4.77684945e-01 -4.16375041e-01 4.97795232e-02
6.27533674e-01 -1.06157470e+00 1.04664218e+00 -9.85724390e-01
5.64684093e-01 -8.16971719e-01 -5.01530528e-01 -1.40532243e+00
-9.98020470e-01 -4.91912305e-01 -3.23172957e-01 8.11895430e-01
6.21201575e-01 -5.70894837e-01 3.86111319e-01 4.95020121e-01
-1.39627531e-01 -3.60045403e-01 -9.63981867e-01 -6.14595175e-01
-6.92562982e-02 -7.13595152e-01 7.06820130e-01 8.55080724e-01
2.18682244e-01 9.22277629e-01 -8.63656104e-02 2.71012783e-01
4.73316789e-01 4.26877216e-02 7.76723921e-01 -9.39458549e-01
-7.04285085e-01 -4.01161522e-01 -5.00731051e-01 -1.46686268e+00
1.33798391e-01 -1.21692431e+00 -2.89923139e-02 -1.80835176e+00
5.24473727e-01 -1.75679520e-01 -2.96332896e-01 8.68804872e-01
-1.99311867e-01 -4.54030335e-02 3.90299797e-01 8.80990177e-02
-6.46422207e-01 5.32047749e-01 1.59023869e+00 -1.87130138e-01
2.85467897e-02 -6.77558422e-01 -1.44880819e+00 6.23492837e-01
5.97564936e-01 -2.08330959e-01 -6.79547489e-01 -6.47643447e-01
8.70235324e-01 -4.45515625e-02 6.96182311e-01 -8.05143654e-01
1.84058756e-01 -3.68056506e-01 7.62319267e-01 -3.74456823e-01
3.77633005e-01 -7.89599240e-01 2.10546404e-01 -3.08674779e-02
-1.60222054e-01 3.56684513e-02 4.69364405e-01 4.94503587e-01
-1.53466491e-02 -4.39586751e-02 3.19958061e-01 -4.93393630e-01
-7.99372792e-01 -9.91591066e-02 -5.79721451e-01 -1.35280639e-01
5.48464000e-01 -3.63646626e-01 -7.52198100e-01 2.22151112e-02
-6.93172634e-01 -2.49210969e-01 2.55546719e-01 4.09328014e-01
7.89671183e-01 -1.47344732e+00 -2.21646547e-01 4.90370095e-01
3.26931030e-01 2.41401136e-01 4.57132667e-01 3.58100802e-01
-1.97820708e-01 3.98073286e-01 -3.44140559e-01 -6.13985121e-01
-5.70514381e-01 8.74279439e-01 1.86559021e-01 -2.13029590e-02
-3.63534361e-01 1.02082849e+00 1.03665733e+00 -4.72214878e-01
-4.91783256e-03 -6.61671996e-01 9.40441936e-02 -1.91441774e-02
9.14750755e-01 -1.96538895e-01 -1.85410753e-01 -4.91342962e-01
-2.68857270e-01 4.36542362e-01 -4.57613766e-02 -3.66296083e-01
1.08179271e+00 1.33604873e-02 -5.50349891e-01 6.38049841e-01
5.89069307e-01 -4.44028452e-02 -1.00320613e+00 -2.49956831e-01
-3.01144391e-01 -5.43158352e-01 -1.94890529e-01 -9.10876870e-01
-7.61998236e-01 5.64515591e-01 3.27214263e-02 8.21359605e-02
1.28166306e+00 2.89554834e-01 5.91912121e-02 7.15682566e-01
4.61150676e-01 -5.74059069e-01 2.66259581e-01 6.08229339e-01
1.12769711e+00 -8.58493209e-01 -5.00147988e-04 -3.42973143e-01
-5.03741741e-01 8.80820155e-01 6.43092811e-01 -1.05800293e-01
5.07203937e-01 1.33929193e-01 -3.44847739e-01 -5.09733319e-01
-9.85919654e-01 -3.20006698e-01 5.97711205e-01 7.74668694e-01
2.25872427e-01 2.59890586e-01 -9.06299576e-02 1.00331664e+00
-3.58110130e-01 -2.16818109e-01 2.04121292e-01 9.14451778e-01
-6.99573159e-01 -8.70724559e-01 -5.02957880e-01 5.12503564e-01
-1.80433001e-02 -4.29813176e-01 -2.13900954e-01 6.14254594e-01
2.55351365e-01 6.52286232e-01 8.56389198e-03 -3.17751467e-01
6.75359517e-02 1.06411979e-01 9.65000093e-01 -9.47802484e-01
-3.99725437e-01 -4.25984152e-02 4.54130918e-02 -5.74612379e-01
-5.58446348e-01 -3.51878494e-01 -1.41797161e+00 3.50898840e-02
-1.51408985e-01 -2.04101831e-01 5.03887534e-01 1.18959475e+00
2.57941067e-01 8.12557995e-01 -4.80257496e-02 -9.60523903e-01
5.60181588e-02 -7.48175740e-01 -7.88248777e-01 3.01577538e-01
2.90956974e-01 -1.04111564e+00 -2.16907382e-01 3.18640172e-01]
|
[10.023052215576172, 1.2325763702392578]
|
c3bbfac3-11d5-4623-8e22-85c4bcf85d1b
|
salient-sign-detection-in-safe-autonomous
|
2301.05804
| null |
https://arxiv.org/abs/2301.05804v2
|
https://arxiv.org/pdf/2301.05804v2.pdf
|
Salient Sign Detection In Safe Autonomous Driving: AI Which Reasons Over Full Visual Context
|
Detecting road traffic signs and accurately determining how they can affect the driver's future actions is a critical task for safe autonomous driving systems. However, various traffic signs in a driving scene have an unequal impact on the driver's decisions, making detecting the salient traffic signs a more important task. Our research addresses this issue, constructing a traffic sign detection model which emphasizes performance on salient signs, or signs that influence the decisions of a driver. We define a traffic sign salience property and use it to construct the LAVA Salient Signs Dataset, the first traffic sign dataset that includes an annotated salience property. Next, we use a custom salience loss function, Salience-Sensitive Focal Loss, to train a Deformable DETR object detection model in order to emphasize stronger performance on salient signs. Results show that a model trained with Salience-Sensitive Focal Loss outperforms a model trained without, with regards to recall of both salient signs and all signs combined. Further, the performance margin on salient signs compared to all signs is largest for the model trained with Salience-Sensitive Focal Loss.
|
['Mohan Trivedi', 'Akshay Rangesh', 'Nachiket Deo', 'Akshay Gopalkrishnan', 'Ross Greer']
|
2023-01-14
| null | null | null | null |
['traffic-sign-detection']
|
['computer-vision']
|
[ 3.01219106e-01 -1.87096391e-02 -4.44087178e-01 -6.54904246e-01
-6.17222190e-01 -2.20242050e-02 8.53217244e-01 -3.04687694e-02
-6.64977968e-01 3.59107435e-01 4.75597590e-01 -4.09483641e-01
-3.46655816e-01 -3.79625231e-01 -5.71163833e-01 -5.56584060e-01
4.64576595e-02 -1.62985161e-01 9.00762916e-01 -2.85467327e-01
5.45673609e-01 5.38233519e-01 -2.16394281e+00 1.50284097e-01
6.49845839e-01 1.08306158e+00 3.38738151e-02 4.92037147e-01
1.68297231e-01 1.19341743e+00 -4.49055314e-01 -4.71368283e-01
4.58070666e-01 -1.77434728e-01 -2.11576015e-01 -3.98702443e-01
1.14844477e+00 -3.87275994e-01 -3.87929052e-01 1.01179063e+00
3.29172820e-01 1.67723134e-01 7.89145947e-01 -1.66174269e+00
-3.04993749e-01 2.01783553e-01 -5.24766982e-01 7.58399367e-01
-7.46665895e-02 5.60782135e-01 1.19733775e+00 -7.43178666e-01
5.18030345e-01 1.23351181e+00 4.79115635e-01 4.73447442e-01
-6.52293861e-01 -7.86890030e-01 4.25404698e-01 7.55311966e-01
-9.72087145e-01 -4.92161185e-01 9.73509848e-01 -5.98949194e-01
4.10262704e-01 4.39642817e-01 4.50171739e-01 7.47363150e-01
2.57600129e-01 1.09153688e+00 1.24830985e+00 -1.57437786e-01
-9.92446318e-02 3.62358391e-02 5.49804688e-01 5.37490726e-01
2.70921260e-01 7.45243907e-01 -5.60552061e-01 3.66601884e-01
1.21601775e-01 -6.35441393e-02 1.23080641e-01 -3.11623156e-01
-1.10817480e+00 6.60080254e-01 5.90112388e-01 3.67671065e-02
-4.28222299e-01 4.19797629e-01 1.78232715e-01 2.99707323e-01
1.47297427e-01 2.77976990e-01 5.03088385e-02 -4.02358562e-01
-7.86512375e-01 4.65931416e-01 2.14500859e-01 6.38506055e-01
7.29551315e-01 3.26183289e-02 -8.67951930e-01 9.04446721e-01
4.22166973e-01 7.36337304e-01 1.31301209e-01 -8.59770298e-01
3.97763729e-01 8.05732667e-01 -2.46849433e-01 -9.45640802e-01
-5.08240104e-01 -2.99268246e-01 -1.37724042e-01 6.22092128e-01
4.78771806e-01 5.20894444e-03 -1.26248527e+00 1.71948993e+00
2.58957058e-01 3.53212684e-01 -3.15912575e-01 1.15685296e+00
1.01115870e+00 1.96936980e-01 7.14019775e-01 3.78194630e-01
1.48345804e+00 -7.09375203e-01 -6.05147541e-01 -5.46726584e-01
5.97455263e-01 -6.90925360e-01 1.19994664e+00 1.19268661e-02
-8.30431879e-01 -7.16345906e-01 -1.17778635e+00 -4.26910907e-01
-5.88526845e-01 1.17193267e-01 6.87401772e-01 5.52376628e-01
-7.30207264e-01 5.70527837e-02 -1.50260448e-01 -1.05753087e-01
4.01038229e-01 6.80368915e-02 -4.26791087e-02 -3.89737524e-02
-1.39496303e+00 1.62220943e+00 7.89543912e-02 -1.33649549e-02
-8.05433095e-01 -9.85058963e-01 -1.04032755e+00 -5.20998053e-02
3.35120916e-01 -3.23748052e-01 1.18029857e+00 -5.83320320e-01
-8.26476276e-01 1.03337455e+00 -3.53587806e-01 -6.35719061e-01
8.58829856e-01 -1.41308010e-01 -6.93408608e-01 -9.58168507e-02
1.61762774e-01 1.12390184e+00 1.14213765e+00 -1.27708530e+00
-1.28048849e+00 1.25114113e-01 1.84761196e-01 1.25381770e-02
1.90589488e-01 3.04283977e-01 -2.27427959e-01 -4.85003561e-01
-2.27346987e-01 -1.01982224e+00 4.94028702e-02 2.54997700e-01
-1.50056586e-01 -4.62858230e-01 1.23485994e+00 -3.81190211e-01
1.32729852e+00 -2.37796140e+00 -5.59682906e-01 3.76585037e-01
3.64517212e-01 4.14765239e-01 -3.62016112e-01 -2.55728483e-01
-1.85188651e-02 -1.51892096e-01 -1.82649165e-01 -3.99102457e-02
2.92733103e-01 1.45766661e-01 -2.23144859e-01 1.99076161e-01
7.68448830e-01 1.08668506e+00 -8.84029806e-01 -5.37846982e-01
4.84039456e-01 3.51894438e-01 -3.15051973e-01 -2.53980458e-01
1.48528397e-01 -6.68733642e-02 -4.02242005e-01 5.79060316e-01
6.74075603e-01 3.71178001e-01 -7.94212103e-01 -1.04636386e-01
-4.13634509e-01 4.11225587e-01 -9.16164279e-01 4.28447574e-01
-3.08888346e-01 1.21899629e+00 -1.37878343e-01 -4.20545667e-01
1.07377827e+00 -1.43437669e-01 2.68635303e-01 -1.28476095e+00
1.80667564e-02 1.80852413e-01 4.90763396e-01 -6.17239594e-01
9.19181287e-01 -1.04435354e-01 -1.53019652e-01 1.97980344e-01
-6.46824002e-01 -1.53570414e-01 4.99086469e-01 1.47767246e-01
1.01070952e+00 -2.50569761e-01 -8.41742530e-02 -1.20574027e-01
5.85671127e-01 -7.60003866e-04 5.79330921e-01 7.72293687e-01
-9.09196436e-01 5.69127381e-01 6.54076338e-01 -3.75749052e-01
-6.32674396e-01 -9.44532812e-01 -3.38201374e-01 1.24630785e+00
4.81863946e-01 1.30747095e-01 -2.97518969e-01 -8.86453032e-01
6.88654363e-01 1.15279210e+00 -9.09550667e-01 -6.68407977e-01
-6.18405163e-01 -4.08188432e-01 3.99590880e-01 6.47293508e-01
6.35903299e-01 -1.33266807e+00 -9.26149070e-01 -4.86484736e-01
7.73067549e-02 -9.31189299e-01 -7.67284155e-01 -6.43230900e-02
-1.32298604e-01 -1.31992877e+00 -7.16512442e-01 -8.15977693e-01
6.84999824e-01 5.91714859e-01 8.75559568e-01 2.01670498e-01
-3.48165274e-01 3.71722311e-01 -1.94677413e-01 -1.04208386e+00
-4.40871388e-01 -1.80340827e-01 -3.14440191e-01 2.58097798e-01
7.71431148e-01 2.07614630e-01 -6.52784288e-01 6.28390908e-01
-7.30012119e-01 6.23838790e-02 8.44437420e-01 4.25267756e-01
2.32205018e-01 -3.32271695e-01 6.33090138e-01 -5.47670901e-01
7.37899721e-01 -2.43156761e-01 -4.27295476e-01 -3.68991010e-02
-3.77440155e-01 6.11577369e-02 -1.83534518e-01 -3.91216844e-01
-1.05579579e+00 -3.00161801e-02 -6.74952865e-02 -2.57628947e-01
-2.41449535e-01 2.35693589e-01 6.66262507e-02 -1.64731532e-01
5.35032451e-01 -3.97792459e-02 4.55406569e-02 -2.57585347e-02
2.94739276e-01 5.60546398e-01 4.93353456e-01 -6.40289634e-02
8.73982191e-01 3.10579747e-01 3.51422161e-01 -9.79336858e-01
-8.04945946e-01 -8.17883670e-01 -5.71939170e-01 -6.78444207e-01
6.42865956e-01 -6.57732546e-01 -6.52469456e-01 4.22204792e-01
-7.31018960e-01 -2.99021959e-01 -5.22868037e-01 7.13114202e-01
-3.49845052e-01 1.10892162e-01 1.82845499e-02 -7.80203521e-01
1.93787813e-01 -1.24575615e+00 1.05636716e+00 1.34411976e-01
-4.42454427e-01 -6.64264262e-01 -2.00206906e-01 5.41223526e-01
6.60320580e-01 1.29114181e-01 8.63106668e-01 -3.25156271e-01
-7.62971580e-01 -4.55670804e-01 -7.33904421e-01 3.66276383e-01
1.32101197e-02 1.14715971e-01 -1.15805471e+00 3.24730366e-01
-3.40499580e-01 3.55033316e-02 1.60066319e+00 6.12859786e-01
6.54183209e-01 1.17922977e-01 -2.33921915e-01 2.76987225e-01
7.15381742e-01 2.83798158e-01 8.61725092e-01 3.91999006e-01
6.21479273e-01 9.16419744e-01 8.81817162e-01 -1.41624257e-01
5.98342478e-01 6.11513555e-01 4.00489807e-01 -4.21734780e-01
-6.82922065e-01 -1.86154082e-01 5.05015135e-01 9.91942510e-02
2.06177264e-01 2.51989424e-01 -8.13032269e-01 8.59812021e-01
-1.72509277e+00 -1.17752826e+00 -5.10888040e-01 2.17903423e+00
5.53197443e-01 4.73441064e-01 2.52308458e-01 3.69014382e-01
7.09295988e-01 3.73820484e-01 -6.65495098e-01 -5.17720878e-01
-1.91368431e-01 -3.27572674e-01 6.96402311e-01 5.57911515e-01
-1.34158421e+00 8.85098279e-01 6.59650373e+00 7.14408398e-01
-1.34652352e+00 -1.63651884e-01 5.12392759e-01 -2.50426412e-01
-2.99578756e-01 -1.28010869e-01 -9.57839727e-01 6.84147358e-01
7.19810605e-01 -1.18941456e-01 -3.41145098e-01 8.30587089e-01
6.84531569e-01 -6.27972066e-01 -9.52341616e-01 6.78607404e-01
7.00473934e-02 -8.09135973e-01 1.38807356e-01 -6.34971708e-02
4.89960194e-01 -2.36489847e-02 4.94735718e-01 5.22916198e-01
2.71165639e-01 -8.68901789e-01 1.00787532e+00 6.08404100e-01
4.93865967e-01 -6.39127374e-01 6.80093586e-01 3.06493118e-02
-1.13574648e+00 -1.82711378e-01 -4.90589179e-02 -7.38691911e-02
2.77779460e-01 4.59069103e-01 -8.84340525e-01 -3.29179376e-01
3.07530731e-01 8.10494959e-01 -1.04805827e+00 1.59300220e+00
-5.82153082e-01 8.62038493e-01 -1.72704622e-01 -2.06643313e-01
4.01726961e-01 1.67687982e-01 7.60830760e-01 1.39044070e+00
-8.30227323e-03 -2.63863355e-01 -3.63353901e-02 8.56098473e-01
7.55353868e-02 -2.34837517e-01 -6.90453410e-01 5.00688553e-01
2.25671455e-01 1.13086665e+00 -3.55589002e-01 -3.49975288e-01
-5.92712998e-01 2.13285580e-01 -2.07086205e-01 2.34270334e-01
-1.08737326e+00 -4.95313883e-01 9.06366587e-01 4.41614896e-01
1.82973489e-01 1.09549537e-01 -5.60083747e-01 -3.87008131e-01
2.86740661e-01 -4.22517538e-01 3.48097026e-01 -8.27903330e-01
-9.35775340e-01 1.24781489e-01 2.53096074e-01 -1.49526322e+00
1.01259090e-02 -6.41666532e-01 -8.56157303e-01 8.27669621e-01
-2.46281958e+00 -1.20363700e+00 -4.70833331e-01 2.26189703e-01
4.43659842e-01 4.72492427e-02 -1.21469907e-01 4.96362358e-01
-4.91582096e-01 7.61603415e-01 -6.32325530e-01 1.15328841e-01
8.68666112e-01 -1.02557397e+00 3.94395679e-01 1.01785016e+00
-2.66316056e-01 3.31690997e-01 4.17948931e-01 -6.94145918e-01
-8.24829757e-01 -1.19776726e+00 1.23861575e+00 -6.88529670e-01
5.43271780e-01 1.89537838e-01 -7.56578863e-01 4.22107279e-01
-1.12100311e-01 -4.49560992e-02 2.07276911e-01 -6.49124980e-02
-3.38915288e-01 -3.35415423e-01 -1.19737756e+00 8.92058730e-01
1.09855342e+00 -5.14528275e-01 -9.31530297e-01 -2.00313076e-01
2.64806628e-01 -6.56105801e-02 -2.62860268e-01 6.59645259e-01
6.88000798e-01 -8.37291539e-01 1.03639686e+00 -3.24125350e-01
3.03372473e-01 -5.70649326e-01 2.73287803e-01 -1.09903789e+00
-5.12038410e-01 -2.79259890e-01 1.10073000e-01 8.80816996e-01
6.23495221e-01 -7.74657905e-01 5.77554166e-01 1.01974297e+00
-5.28709292e-01 -3.62774670e-01 -9.93338227e-01 -7.77591050e-01
-1.65021598e-01 -7.53435016e-01 6.21994734e-01 5.72387993e-01
-1.33416489e-01 6.19735084e-02 -1.63608775e-01 -5.63659593e-02
3.97892118e-01 -2.09630236e-01 8.24019372e-01 -1.35241652e+00
5.36974788e-01 -1.16178703e+00 -7.95126379e-01 -9.78828192e-01
8.53680074e-02 -9.31814313e-01 3.61649722e-01 -1.36322880e+00
1.75400898e-01 -5.47361195e-01 -6.70982182e-01 6.90576792e-01
-5.86681008e-01 2.24455163e-01 4.97954994e-01 -4.30068523e-02
-6.27160549e-01 4.76715952e-01 1.56564653e+00 -3.96236479e-01
-4.64375801e-02 8.84142518e-02 -7.56661355e-01 7.84625530e-01
5.97787976e-01 -2.03361526e-01 -3.55370373e-01 -1.10136837e-01
-1.20403938e-01 -5.67161024e-01 7.47389078e-01 -1.24974513e+00
3.36090118e-01 -3.94041151e-01 1.73101664e-01 -8.33656132e-01
3.88646275e-01 -6.30380332e-01 -5.59179127e-01 3.97502869e-01
-7.61393905e-01 -2.12524503e-01 4.42053556e-01 2.93794066e-01
-3.52095395e-01 -2.05118954e-01 8.45787466e-01 5.15155733e-01
-1.50957966e+00 1.14039198e-01 -4.79138374e-01 3.45670432e-01
1.23013687e+00 -7.26556361e-01 -3.87126386e-01 -3.76458704e-01
-2.93972909e-01 4.93539155e-01 8.01834688e-02 1.11937368e+00
8.94877791e-01 -1.31948900e+00 -8.90663028e-01 5.30053318e-01
5.68233669e-01 -4.25035626e-01 -3.00062504e-02 1.08974659e+00
-2.12487370e-01 5.86023033e-01 -1.70080796e-01 -5.44796348e-01
-1.61521661e+00 1.73843950e-01 2.78923631e-01 9.09893364e-02
-4.36262518e-01 6.68811977e-01 3.15441161e-01 -4.28536758e-02
2.80972987e-01 -9.44831610e-01 -5.05205929e-01 3.33713114e-01
6.90196037e-01 7.55620897e-01 5.26665375e-02 -1.13626206e+00
-3.64594221e-01 8.49170983e-01 -1.17592111e-01 1.35040447e-01
7.91491151e-01 8.41623992e-02 4.48569447e-01 2.15031028e-01
1.07456040e+00 2.51791719e-03 -1.69781375e+00 -1.62480786e-01
1.95941195e-01 -6.81432903e-01 3.96140754e-01 -9.20536995e-01
-1.14962018e+00 8.10084581e-01 6.99194074e-01 -7.53105432e-02
7.23295808e-01 -2.62482315e-02 7.42024958e-01 1.76584005e-01
6.06362112e-02 -1.15669382e+00 -1.22581922e-01 7.47882307e-01
1.02165687e+00 -1.73326838e+00 -1.41330197e-01 -3.27917099e-01
-1.03945196e+00 6.32390976e-01 5.93468666e-01 6.20810986e-02
7.70384431e-01 8.10276046e-02 4.81380492e-01 -4.00679499e-01
-6.15081072e-01 -9.62829351e-01 9.89726007e-01 5.33618152e-01
1.20105706e-01 1.67467311e-01 -4.41437960e-01 3.45224105e-02
-2.21016616e-01 4.26054299e-02 2.40750223e-01 7.98846483e-01
-7.18285143e-01 -6.13468766e-01 -2.41198733e-01 8.38167429e-01
-6.95561767e-02 -7.10779279e-02 -6.30367458e-01 8.73569846e-01
3.30837727e-01 1.08850908e+00 3.39572310e-01 -3.18366885e-01
9.19546545e-01 1.79766938e-01 -7.74082914e-02 -3.45159084e-01
-5.26145339e-01 -6.51470721e-01 3.06018114e-01 -4.96069819e-01
-2.97445148e-01 -9.78817701e-01 -1.35972977e+00 -1.54455572e-01
-1.54681310e-01 -3.16334665e-01 5.38010418e-01 9.32577252e-01
3.53793621e-01 6.71308756e-01 4.86524612e-01 -5.93101323e-01
-3.21647525e-01 -7.87033141e-01 -2.66590476e-01 9.53494251e-01
7.72020102e-01 -1.12845385e+00 -4.30530965e-01 -1.72397420e-01]
|
[7.950728893280029, -0.7848660349845886]
|
16eec3ef-6aa9-402d-b58e-01a02da31399
|
making-video-quality-assessment-models-robust
|
2304.13092
| null |
https://arxiv.org/abs/2304.13092v1
|
https://arxiv.org/pdf/2304.13092v1.pdf
|
Making Video Quality Assessment Models Robust to Bit Depth
|
We introduce a novel feature set, which we call HDRMAX features, that when included into Video Quality Assessment (VQA) algorithms designed for Standard Dynamic Range (SDR) videos, sensitizes them to distortions of High Dynamic Range (HDR) videos that are inadequately accounted for by these algorithms. While these features are not specific to HDR, and also augment the equality prediction performances of VQA models on SDR content, they are especially effective on HDR. HDRMAX features modify powerful priors drawn from Natural Video Statistics (NVS) models by enhancing their measurability where they visually impact the brightest and darkest local portions of videos, thereby capturing distortions that are often poorly accounted for by existing VQA models. As a demonstration of the efficacy of our approach, we show that, while current state-of-the-art VQA models perform poorly on 10-bit HDR databases, their performances are greatly improved by the inclusion of HDRMAX features when tested on HDR and 10-bit distorted videos.
|
['Alan C. Bovik', 'Sriram Sethuraman', 'Hai Wei', 'Yongjun Wu', 'Zaixi Shang', 'Joshua P. Ebenezer']
|
2023-04-25
| null | null | null | null |
['video-quality-assessment', 'video-quality-assessment']
|
['computer-vision', 'time-series']
|
[ 1.06635347e-01 -4.77324188e-01 -3.30894962e-02 -2.88062572e-01
-5.95949173e-01 -5.36620498e-01 6.17194533e-01 -1.98588073e-01
-1.39788285e-01 5.24015605e-01 6.41279399e-01 -3.76211922e-03
-2.21321642e-01 -5.89371800e-01 -6.85146332e-01 -5.34318805e-01
-4.67447519e-01 -7.36238156e-03 5.23462772e-01 -4.39544261e-01
1.23309560e-01 8.14278066e-01 -1.88345742e+00 5.05021513e-01
6.35639608e-01 1.11417711e+00 -6.99583665e-02 1.16678095e+00
6.03310764e-01 1.18824685e+00 -7.50315905e-01 -4.96783912e-01
4.80016083e-01 -2.99831629e-01 -4.76601690e-01 8.48372579e-02
8.31300378e-01 -9.92498577e-01 -1.14287567e+00 8.65572095e-01
4.34034795e-01 9.39306319e-02 5.95040739e-01 -1.20490563e+00
-8.69311929e-01 1.91864461e-01 -4.34749126e-01 7.66168237e-01
8.20268929e-01 5.24094582e-01 1.06823587e+00 -8.92772019e-01
8.28184664e-01 1.47852671e+00 6.17852092e-01 3.50523055e-01
-1.05590236e+00 -3.95724207e-01 -2.31972113e-01 7.19799876e-01
-1.29436338e+00 -7.34315515e-01 6.55052662e-01 -2.15984255e-01
1.09591293e+00 5.08400202e-01 5.89803874e-01 1.03335357e+00
3.89000118e-01 7.32028961e-01 9.30410922e-01 4.10159975e-02
2.31935635e-01 -2.80734539e-01 -3.53854835e-01 2.22508639e-01
-2.32440140e-02 5.74129641e-01 -6.54466391e-01 -1.73506828e-03
9.33191955e-01 -2.63390601e-01 -6.50348663e-01 -5.37026644e-01
-1.25483179e+00 6.00164473e-01 2.27710813e-01 7.34283179e-02
-3.60004544e-01 8.86939466e-02 4.02949810e-01 5.13425708e-01
3.34440142e-01 2.58312017e-01 -2.12354586e-01 -5.40163875e-01
-1.32314503e+00 2.34791920e-01 3.87593418e-01 1.02749968e+00
6.44028902e-01 3.48063648e-01 -4.62847441e-01 7.66387939e-01
2.84026176e-01 9.07515824e-01 1.34778425e-01 -1.49825561e+00
2.95298904e-01 2.85112429e-02 2.34733686e-01 -1.13547599e+00
-2.25429550e-01 -1.43904969e-01 -7.40219831e-01 3.71990502e-01
3.29909712e-01 4.04050201e-01 -1.02276933e+00 1.31062770e+00
-9.57809985e-02 8.17965269e-02 5.05696051e-02 1.29672515e+00
7.98287034e-01 9.24761415e-01 5.11553586e-02 -4.15869564e-01
9.46139753e-01 -5.00011027e-01 -7.68845260e-01 2.91187353e-02
1.44320324e-01 -7.94951618e-01 9.74616230e-01 8.54249060e-01
-1.47974432e+00 -8.57665718e-01 -1.23639059e+00 -2.07275480e-01
-2.13111322e-02 -3.00349027e-01 3.54660958e-01 8.10733557e-01
-1.46107757e+00 5.82305610e-01 -5.41894674e-01 -5.70257269e-02
3.76901090e-01 1.10309526e-01 -4.88942116e-01 -5.98577678e-01
-1.35075271e+00 9.82988954e-01 -1.35512576e-01 -2.38375599e-03
-1.38912845e+00 -9.61343825e-01 -8.93795311e-01 8.63561332e-02
2.52843261e-01 -2.69758075e-01 8.09340656e-01 -9.96053278e-01
-1.38122034e+00 6.22398138e-01 1.35807276e-01 -4.89336818e-01
7.26925254e-01 -2.84034580e-01 -8.57835412e-01 8.27761889e-01
-4.83687848e-01 5.87866545e-01 1.24231160e+00 -1.49589837e+00
-4.10339773e-01 -1.25634059e-01 1.63982570e-01 1.32209733e-01
6.15374409e-02 3.87726068e-01 -5.60830176e-01 -8.90514255e-01
-1.10072173e-01 -6.21800184e-01 1.02917545e-01 2.81660676e-01
-4.50220332e-02 4.09709513e-01 1.02333331e+00 -1.09620607e+00
1.29352403e+00 -2.29471779e+00 2.57629395e-01 3.09999645e-01
3.11954826e-01 5.45605719e-01 -3.45521390e-01 2.24626958e-01
-1.78577617e-01 1.43297777e-01 3.28360051e-02 2.63209373e-01
1.44453362e-01 2.57240504e-01 -2.90491521e-01 8.05432081e-01
3.66312981e-01 6.02159083e-01 -8.62253368e-01 -3.73567969e-01
6.19706690e-01 8.34411561e-01 -6.88197792e-01 4.73145992e-01
3.54202166e-02 2.48375729e-01 1.73220456e-01 5.53031683e-01
9.39933956e-01 6.09951280e-02 -9.92582068e-02 -6.24389410e-01
3.70497932e-03 -3.59818600e-02 -1.21893167e+00 1.24396062e+00
-1.56492084e-01 9.21534836e-01 -1.38720647e-01 -4.38856244e-01
7.79863179e-01 1.77718908e-01 6.32394135e-01 -1.05617356e+00
-1.57439291e-01 9.93419439e-03 -3.37093957e-02 -5.99053741e-01
7.43617117e-01 3.84409092e-02 2.61370957e-01 3.93052958e-02
2.57784784e-01 -1.80029586e-01 3.01628947e-01 4.35900539e-01
1.37528336e+00 8.43228251e-02 4.55992997e-01 5.88643923e-02
4.73464847e-01 -5.12351215e-01 4.92283136e-01 7.57458329e-01
-5.23310959e-01 1.19537711e+00 3.89177829e-01 -2.15317890e-01
-1.64765120e+00 -1.57721949e+00 -3.08277816e-01 1.00040078e+00
1.96372166e-01 -3.94733340e-01 -3.66398901e-01 -4.73208576e-01
-5.30332215e-02 5.79956830e-01 -5.19127369e-01 -2.38323659e-01
-4.83880281e-01 -5.97396374e-01 4.96194601e-01 5.24294019e-01
4.25977856e-01 -8.12565744e-01 -6.69517636e-01 7.33890757e-02
-1.10297278e-01 -1.25593436e+00 -2.21572697e-01 -1.43697292e-01
-5.41830301e-01 -1.04086637e+00 -1.12817025e+00 4.92183305e-02
3.53517383e-02 3.01635593e-01 1.56180310e+00 9.77165774e-02
-3.06640327e-01 8.69550407e-01 -8.61320734e-01 2.44488239e-01
-8.09409678e-01 -7.32988834e-01 7.23714009e-02 -3.21107209e-02
1.59262478e-01 -4.74619418e-01 -7.33869553e-01 5.96245050e-01
-1.20082271e+00 -3.07255089e-01 4.87234831e-01 6.27505898e-01
4.35905069e-01 8.10934156e-02 2.85735607e-01 -3.72593433e-01
1.77006796e-02 -5.54272711e-01 -3.49173427e-01 1.06620289e-01
-4.64335352e-01 -1.68940455e-01 6.53219104e-01 -3.95238101e-01
-1.04544854e+00 -5.59421897e-01 -4.53766823e-01 -8.40665460e-01
-2.76843071e-01 -1.43109396e-01 -3.79517376e-01 -3.91576201e-01
7.15280831e-01 2.14308083e-01 -1.84693709e-01 -3.02049994e-01
3.93701047e-01 5.87191701e-01 6.65232480e-01 -1.64881781e-01
9.90176618e-01 4.71080512e-01 -1.25233633e-02 -9.83145297e-01
-3.20473313e-01 -4.67606455e-01 -3.34924608e-01 -4.66442794e-01
7.23703384e-01 -1.12134516e+00 -3.54701221e-01 4.85675514e-01
-7.91000068e-01 -1.73797935e-01 -5.23224711e-01 4.18566197e-01
-9.09729600e-01 8.22365701e-01 -7.26769328e-01 -5.99702775e-01
1.67263955e-01 -1.14451194e+00 1.00808346e+00 -7.91045204e-02
5.76448664e-02 -8.30264926e-01 4.91794795e-02 2.23628581e-01
5.86710691e-01 -6.21666957e-04 9.22757804e-01 -3.80450070e-01
-6.33861363e-01 1.71500593e-01 -4.51992273e-01 7.64229000e-01
-1.21205918e-01 4.23035592e-01 -9.86208558e-01 -4.68218744e-01
-1.62878260e-01 -2.36031547e-01 9.07421291e-01 5.39231837e-01
1.15337694e+00 -2.74813801e-01 3.77330542e-01 8.86741698e-01
1.52078760e+00 3.27215567e-02 1.46237671e+00 2.94061899e-01
5.64947248e-01 1.51464000e-01 8.50790083e-01 7.69393206e-01
2.27673307e-01 1.01290357e+00 8.33816826e-01 -1.72040522e-01
-6.94938242e-01 7.07590356e-02 6.90015137e-01 5.87319851e-01
-4.00927722e-01 -4.80639279e-01 -6.36849999e-01 3.88845235e-01
-1.32400382e+00 -1.23573577e+00 -1.49313584e-01 2.13249898e+00
7.80417502e-01 -2.15467308e-02 3.06730956e-01 4.93840903e-01
4.81656611e-01 5.35839379e-01 -3.94095093e-01 -5.87375104e-01
-7.00802505e-01 1.86863661e-01 4.52610224e-01 1.94078371e-01
-1.02867460e+00 4.75713372e-01 7.38668251e+00 8.17493558e-01
-8.01783442e-01 -3.92498411e-02 5.29731929e-01 -1.59249261e-01
-4.28260833e-01 -3.17361325e-01 -3.03981692e-01 5.26050866e-01
1.20802188e+00 1.05761237e-01 7.11284578e-01 8.35058510e-01
4.51662600e-01 -7.77803361e-02 -1.14912057e+00 1.29339671e+00
2.55029798e-01 -1.06623530e+00 2.74589419e-01 -6.38769791e-02
7.58631170e-01 -1.98137313e-01 6.18053615e-01 2.37421975e-01
4.76297475e-02 -1.20422125e+00 8.09125364e-01 4.14752305e-01
1.03719676e+00 -7.72525966e-01 8.66772413e-01 -5.43328524e-01
-8.33401978e-01 -3.61174852e-01 -7.83521891e-01 3.24514478e-01
2.27497131e-01 6.10313714e-01 -3.64254951e-01 4.40725833e-01
9.64781582e-01 9.38059330e-01 -1.00636780e+00 1.25955677e+00
-2.60328334e-02 5.15584886e-01 8.61560404e-02 7.29808271e-01
-1.88753475e-02 2.81474769e-01 8.88217568e-01 1.40114021e+00
4.07844931e-01 8.29350129e-02 -3.73600274e-01 4.89801258e-01
8.97803232e-02 -1.30318999e-01 -5.26131630e-01 1.77085266e-01
1.63742453e-01 9.47332263e-01 -6.86696842e-02 -2.84809530e-01
-5.46486676e-01 9.33061182e-01 -2.45900467e-01 6.11271679e-01
-1.13119662e+00 -1.69080988e-01 8.72202694e-01 2.90760070e-01
7.00829208e-01 -4.34712693e-02 1.58097401e-01 -1.39817357e+00
-1.97904557e-01 -1.47722816e+00 2.41957560e-01 -1.29867697e+00
-1.38159585e+00 4.87598538e-01 3.84390689e-02 -1.60574329e+00
-3.70779872e-01 -8.01012874e-01 -1.09267294e-01 4.31260228e-01
-1.61851513e+00 -9.30225670e-01 -4.28339273e-01 8.52636158e-01
4.26253438e-01 -7.16506988e-02 4.41020370e-01 4.90028143e-01
5.02701066e-02 5.70811033e-01 2.71182388e-01 -1.06911950e-01
8.97041023e-01 -1.21580422e+00 2.34338254e-01 1.11487234e+00
1.15703151e-01 1.29117742e-01 1.21150887e+00 -4.15671349e-01
-1.61303258e+00 -1.03882277e+00 6.95292354e-02 -3.30619246e-01
4.68648642e-01 -2.66196113e-02 -1.08184731e+00 5.35023928e-01
6.12990670e-02 3.80339205e-01 5.41309237e-01 -2.57111311e-01
-7.22139895e-01 -2.12516353e-01 -1.37404895e+00 3.35622430e-01
1.01865673e+00 -6.61591172e-01 -6.17745459e-01 -2.86837071e-02
5.65313935e-01 -2.52630591e-01 -1.37051344e+00 5.03880441e-01
4.37008053e-01 -1.57310259e+00 1.41985679e+00 -4.48794156e-01
6.81670308e-01 -3.58644545e-01 -6.13184810e-01 -1.29652917e+00
-4.72958207e-01 -5.74115694e-01 -7.31883287e-01 1.04359567e+00
-8.64790678e-02 2.21102424e-02 4.63307709e-01 3.50057960e-01
-7.98393562e-02 -1.30980119e-01 -1.02192259e+00 -8.54183078e-01
-1.77963585e-01 -5.48137307e-01 4.97614264e-01 7.20624030e-01
-2.01176465e-01 -4.69840884e-01 -8.14138055e-01 1.73364803e-01
7.49998212e-01 -5.09048581e-01 5.92843711e-01 -7.29264498e-01
-4.86585170e-01 -1.01888329e-01 -1.12092066e+00 -1.00335240e+00
-2.59995401e-01 -1.75078705e-01 1.25175804e-01 -1.12571740e+00
3.31725657e-01 -2.96937197e-01 -3.83483887e-01 -2.78406329e-02
-1.24679074e-01 7.00826705e-01 5.18931210e-01 -1.06361462e-02
-8.79504859e-01 5.67601681e-01 1.18273807e+00 -9.90893617e-02
1.42818928e-01 -3.15171927e-01 -2.55877882e-01 4.95693833e-01
3.23805183e-01 -1.92905031e-02 -3.00919712e-01 -4.47673231e-01
3.55596900e-01 2.21578583e-01 6.16467476e-01 -1.37805092e+00
-2.23572239e-01 -8.02451149e-02 8.59903455e-01 -4.13492978e-01
4.36267048e-01 -8.85507941e-01 3.02544892e-01 2.69570816e-02
-2.08800584e-01 7.84404203e-02 8.67820531e-02 4.77466971e-01
-4.08730090e-01 1.45226747e-01 1.19449389e+00 1.50120676e-01
-1.08453083e+00 3.39153528e-01 -4.63362396e-01 8.78506005e-02
8.83700788e-01 -3.51382345e-01 -6.29984975e-01 -9.11046505e-01
-5.80000877e-01 -3.18018347e-01 1.00143385e+00 5.17692745e-01
1.08547604e+00 -1.27282846e+00 -8.52359772e-01 3.07698339e-01
7.06034005e-02 -5.50694644e-01 6.49318039e-01 6.19207382e-01
-7.44435370e-01 2.18640685e-01 -6.67942762e-01 -6.70993447e-01
-1.33014071e+00 9.85606611e-01 1.19319253e-01 -8.93046185e-02
-6.42663062e-01 5.33662975e-01 2.24635482e-01 4.12958592e-01
1.29390925e-01 -2.01535836e-01 -3.32972437e-01 -2.29507059e-01
9.07973886e-01 7.42789388e-01 9.45901349e-02 -1.02360821e+00
-3.18708569e-01 5.25470793e-01 4.67836708e-02 2.22454183e-02
1.26026475e+00 -5.24366379e-01 3.24285328e-01 9.65044647e-02
1.24160743e+00 2.29847863e-01 -1.72050202e+00 7.61976093e-02
-4.68971282e-01 -1.01468623e+00 2.50959188e-01 -8.73090565e-01
-1.17512918e+00 9.10779715e-01 9.81841028e-01 2.33659700e-01
1.61898649e+00 -2.34924644e-01 6.07651711e-01 -1.61712080e-01
4.23324615e-01 -9.03848708e-01 3.25605571e-01 3.24592769e-01
9.52643871e-01 -1.14193368e+00 1.87413499e-01 -3.57532412e-01
-8.35045636e-01 1.27997899e+00 2.24229977e-01 -1.29761130e-01
3.39395285e-01 2.40691006e-01 -8.90785009e-02 1.61980018e-01
-6.67979658e-01 -9.13099572e-02 4.27993089e-01 1.11860335e+00
8.35103914e-03 -4.15438443e-01 1.23606868e-01 1.38244748e-01
6.64613694e-02 -9.76973912e-04 9.73045945e-01 6.27171814e-01
-4.73715335e-01 -5.30986369e-01 -4.84795988e-01 2.33402058e-01
-6.53669715e-01 -1.60571069e-01 1.61993504e-01 8.60928953e-01
1.04835942e-01 1.06284022e+00 2.27602825e-01 -5.10993361e-01
2.50324488e-01 -5.27334273e-01 7.11452425e-01 -5.72404489e-02
-3.89760554e-01 -1.45270884e-01 1.36329845e-01 -1.10855329e+00
-3.82389396e-01 -4.43476856e-01 -6.91604197e-01 -7.11470604e-01
2.75491714e-01 -2.67923415e-01 4.75357741e-01 7.33280957e-01
1.56134993e-01 5.88951826e-01 7.76844561e-01 -9.81061816e-01
-6.25977278e-01 -6.69428408e-01 -8.49487603e-01 9.22046721e-01
6.63968205e-01 -6.45702541e-01 -7.07159340e-01 3.02404165e-01]
|
[11.59286880493164, -1.8814352750778198]
|
6174ba20-1658-49a7-bd10-eefcfd13d778
|
contextual-modeling-for-3d-dense-captioning
|
2210.03925
| null |
https://arxiv.org/abs/2210.03925v1
|
https://arxiv.org/pdf/2210.03925v1.pdf
|
Contextual Modeling for 3D Dense Captioning on Point Clouds
|
3D dense captioning, as an emerging vision-language task, aims to identify and locate each object from a set of point clouds and generate a distinctive natural language sentence for describing each located object. However, the existing methods mainly focus on mining inter-object relationship, while ignoring contextual information, especially the non-object details and background environment within the point clouds, thus leading to low-quality descriptions, such as inaccurate relative position information. In this paper, we make the first attempt to utilize the point clouds clustering features as the contextual information to supply the non-object details and background environment of the point clouds and incorporate them into the 3D dense captioning task. We propose two separate modules, namely the Global Context Modeling (GCM) and Local Context Modeling (LCM), in a coarse-to-fine manner to perform the contextual modeling of the point clouds. Specifically, the GCM module captures the inter-object relationship among all objects with global contextual information to obtain more complete scene information of the whole point clouds. The LCM module exploits the influence of the neighboring objects of the target object and local contextual information to enrich the object representations. With such global and local contextual modeling strategies, our proposed model can effectively characterize the object representations and contextual information and thereby generate comprehensive and detailed descriptions of the located objects. Extensive experiments on the ScanRefer and Nr3D datasets demonstrate that our proposed method sets a new record on the 3D dense captioning task, and verify the effectiveness of our raised contextual modeling of point clouds.
|
['Lin Ma', 'Jiebo Luo', 'Long Xu', 'Yufeng Zhong']
|
2022-10-08
| null | null | null | null |
['dense-captioning', '3d-dense-captioning']
|
['computer-vision', 'computer-vision']
|
[ 5.74608939e-03 -1.79455251e-01 -2.55853925e-02 -6.22145534e-01
-8.08947027e-01 -5.34423709e-01 6.24912679e-01 2.98479855e-01
1.35363936e-01 2.99562603e-01 3.57159734e-01 1.69259444e-01
-1.30541295e-01 -7.15467334e-01 -9.28593755e-01 -8.47885549e-01
3.20210040e-01 6.42085075e-01 3.59387904e-01 6.48424495e-03
2.69354045e-01 8.13524902e-01 -1.72402287e+00 2.62309104e-01
7.61849225e-01 8.96105647e-01 9.79742885e-01 2.46302307e-01
-7.63085365e-01 4.42520350e-01 -4.95291978e-01 3.90745029e-02
1.34098725e-02 6.60676137e-02 -3.88806373e-01 4.83968198e-01
4.63618338e-01 -3.77978653e-01 -2.85798818e-01 1.17337072e+00
2.97699988e-01 6.53787404e-02 4.51440126e-01 -1.20297694e+00
-8.45802724e-01 2.90671021e-01 -7.74856269e-01 -7.82212615e-03
2.68698573e-01 1.65544897e-01 8.42817843e-01 -1.12942934e+00
4.41345364e-01 1.52076471e+00 3.77076954e-01 4.60942209e-01
-8.76135230e-01 -7.16878772e-01 7.82470822e-01 1.86343312e-01
-1.64277458e+00 -1.13513447e-01 1.24541962e+00 -4.48285073e-01
5.36044300e-01 1.77036867e-01 6.42643929e-01 8.68755519e-01
-4.33976114e-01 7.61002481e-01 7.73279846e-01 -9.71496254e-02
1.57525852e-01 1.74798578e-01 2.60100037e-01 3.13861102e-01
2.33130306e-01 -1.02279417e-01 -1.45511851e-01 -2.38378182e-01
8.97346675e-01 4.30486739e-01 -2.78901100e-01 -4.98788774e-01
-1.42800987e+00 5.57167530e-01 6.50481761e-01 2.73337781e-01
-5.78157663e-01 2.81961083e-01 -4.40798849e-02 -5.81288576e-01
5.00414848e-01 -6.34323508e-02 -3.15667838e-01 2.43608624e-01
-6.18195474e-01 3.28906924e-01 2.61652231e-01 1.76751041e+00
1.32490158e+00 -3.70744109e-01 -2.36649737e-01 6.38571203e-01
5.70254982e-01 9.10634160e-01 -1.25094101e-01 -8.32676589e-01
8.10722649e-01 8.69915128e-01 3.88031036e-01 -1.39717650e+00
-1.39269590e-01 -4.34967041e-01 -7.28053749e-01 -2.43451014e-01
-2.41379842e-01 3.12241733e-01 -1.03458810e+00 1.72219574e+00
6.70191526e-01 7.45941043e-01 -3.32087912e-02 1.16962528e+00
1.22256196e+00 1.09165025e+00 2.67193824e-01 -5.81338964e-02
1.31124079e+00 -7.57875085e-01 -6.04475975e-01 -4.38215047e-01
4.08467144e-01 -7.29560673e-01 9.19643521e-01 -3.84566486e-01
-7.59581625e-01 -7.25109816e-01 -8.36301446e-01 -2.26675615e-01
-3.00257802e-01 1.50913581e-01 5.36864758e-01 -5.78624047e-02
-7.30114579e-01 -8.30428153e-02 -6.61864579e-01 -2.69097269e-01
4.87384886e-01 1.45307049e-01 -3.68957460e-01 -5.53115308e-01
-7.77474046e-01 6.78514421e-01 5.62640727e-01 3.39423627e-01
-1.01692557e+00 -8.48376930e-01 -9.19264257e-01 -6.20534122e-02
4.29689348e-01 -8.77493620e-01 8.85836899e-01 -5.24276555e-01
-6.44692361e-01 8.84856582e-01 -5.20937502e-01 1.07385144e-01
5.78045174e-02 -1.92849785e-02 -1.44306093e-01 7.99898580e-02
3.51235062e-01 8.40126038e-01 4.77129132e-01 -2.04758787e+00
-7.42217898e-01 -5.42058706e-01 1.56256258e-01 5.79799771e-01
4.39596802e-01 -1.50576130e-01 -1.01735592e+00 -5.10358691e-01
5.31491935e-01 -7.75972128e-01 -1.51878551e-01 7.71852583e-02
-3.50470871e-01 -3.15927982e-01 1.11746168e+00 -3.79807800e-01
6.98358953e-01 -2.32361388e+00 1.83005724e-02 5.88003453e-03
2.10158989e-01 6.44106120e-02 -2.36341715e-01 2.65192240e-01
2.96313204e-02 2.52019037e-02 -3.94965321e-01 -5.77621579e-01
2.65803393e-02 6.82781041e-01 -4.85570997e-01 3.00531298e-01
4.12928760e-01 9.65507507e-01 -1.12722111e+00 -7.19520748e-01
6.93926990e-01 6.41823411e-01 -3.23558360e-01 2.84981757e-01
-6.28199995e-01 5.96779764e-01 -1.06784356e+00 7.74162769e-01
1.13780499e+00 -2.12884068e-01 -5.78471959e-01 -4.43788677e-01
-3.68528366e-02 -2.64806356e-02 -1.31706357e+00 2.03618240e+00
-3.79670799e-01 1.78444207e-01 1.40929908e-01 -5.91002941e-01
1.19519317e+00 8.70514065e-02 5.61190009e-01 -3.73067796e-01
-8.25979039e-02 5.56919463e-02 -6.53771758e-01 -6.58250153e-01
4.87053812e-01 -9.76861790e-02 -8.06713477e-02 6.57311007e-02
-1.55316845e-01 -3.95654321e-01 -4.04116452e-01 3.04436594e-01
5.01870334e-01 1.03724733e-01 -1.16947889e-02 4.92361281e-03
5.70841491e-01 3.01800102e-01 5.18704832e-01 5.65999806e-01
-6.97111040e-02 1.08305645e+00 -7.84524679e-02 -3.91279429e-01
-1.18218470e+00 -1.01983082e+00 -1.17072910e-01 3.30748975e-01
1.00687420e+00 -3.21183056e-01 -3.69995236e-01 -5.27157247e-01
3.55160125e-02 8.75341773e-01 -5.11076689e-01 -1.10935763e-01
-4.94019985e-01 -4.30911124e-01 -9.37687904e-02 5.40859997e-01
5.64179361e-01 -9.97460365e-01 -2.15063050e-01 1.48488814e-02
-4.44324046e-01 -1.46169984e+00 -5.19784272e-01 -1.57061741e-01
-7.46229410e-01 -8.69749606e-01 -4.52721417e-01 -1.03290498e+00
8.74509454e-01 8.10225844e-01 1.02070165e+00 1.93623051e-01
8.73463601e-02 4.73922819e-01 -4.56131667e-01 -4.65093374e-01
-1.66536242e-01 -3.91192377e-01 -4.47614938e-02 1.21221177e-01
6.40959740e-01 -6.33942366e-01 -4.82113570e-01 4.10542220e-01
-7.52553105e-01 5.44716656e-01 8.05065572e-01 3.67200255e-01
1.30314457e+00 5.41471876e-02 2.42495805e-01 -5.97619534e-01
-3.17518339e-02 -6.83154583e-01 -4.29215997e-01 2.01052040e-01
2.70846183e-03 -1.15620635e-01 1.35704562e-01 -3.57721925e-01
-1.02179348e+00 3.99808198e-01 -3.55221257e-02 -1.00941563e+00
-5.66945910e-01 4.30792242e-01 -8.69809568e-01 5.35632297e-02
1.53737023e-01 5.97752988e-01 -3.66216898e-01 -7.21321046e-01
5.43113530e-01 5.82055926e-01 7.23244667e-01 -7.61201382e-01
1.12171125e+00 6.53450251e-01 6.78348914e-02 -6.91554606e-01
-1.03137398e+00 -9.23942566e-01 -7.76427507e-01 -5.94722293e-02
1.07306397e+00 -1.31604040e+00 -3.18553478e-01 9.30216014e-02
-1.57890069e+00 3.78260583e-01 -2.46349186e-01 3.70129406e-01
-5.44966340e-01 3.78715426e-01 -2.21255720e-02 -7.77783036e-01
-6.42065629e-02 -1.17430234e+00 1.76670718e+00 3.28215778e-01
3.30332667e-01 -6.62583888e-01 -1.71972379e-01 3.74992579e-01
-4.40914594e-02 4.87609863e-01 9.46043372e-01 -4.04099107e-01
-1.19819033e+00 -1.31987914e-01 -6.02247596e-01 2.16722876e-01
2.94187635e-01 -2.41640478e-01 -9.49817479e-01 1.49203271e-01
2.32644543e-01 2.37532571e-01 4.23573464e-01 3.56643617e-01
1.20509005e+00 -1.49885207e-01 -5.58094382e-01 6.33058965e-01
1.54181480e+00 1.50892198e-01 3.62347782e-01 1.09034732e-01
1.29376185e+00 6.18293881e-01 1.04731500e+00 3.34496915e-01
5.22782683e-01 7.20633805e-01 8.95371735e-01 -7.12687150e-02
-7.25802556e-02 -6.37521029e-01 -1.89512402e-01 7.36696362e-01
1.34995773e-01 -6.01413660e-02 -9.65625763e-01 7.26869643e-01
-2.06669855e+00 -8.81063402e-01 -2.96404928e-01 1.85175264e+00
4.25662547e-01 -2.16165051e-01 -3.44273806e-01 -4.22478348e-01
1.17816341e+00 3.01511258e-01 -6.25300407e-01 3.59034568e-01
-3.04689884e-01 -4.34820950e-01 2.99837232e-01 3.53837222e-01
-1.04144967e+00 9.91542935e-01 4.60724640e+00 8.12481523e-01
-5.91756046e-01 1.29685765e-02 3.23501289e-01 -7.48848915e-02
-5.80031157e-01 1.39749497e-01 -9.96665299e-01 4.97659832e-01
2.87150115e-01 -9.66000706e-02 -4.43350971e-02 1.02771473e+00
4.52619106e-01 1.88113555e-01 -1.20156181e+00 1.47593439e+00
9.86861363e-02 -1.36611581e+00 6.77436769e-01 2.41447687e-02
8.12915027e-01 8.89249146e-02 -1.69985667e-01 5.68083264e-02
-1.89255551e-01 -7.06345499e-01 9.26291049e-01 8.29755723e-01
4.78336394e-01 -7.71019101e-01 7.68003821e-01 6.58730030e-01
-1.40568423e+00 1.17274225e-01 -6.96062505e-01 2.44911343e-01
2.07031131e-01 6.51857078e-01 -6.46700442e-01 8.77850652e-01
7.55944967e-01 9.60811198e-01 -5.03397405e-01 1.17949402e+00
-6.12727180e-02 1.66039288e-01 -4.76346046e-01 -3.32903415e-02
5.54638863e-01 -3.36808652e-01 9.48838770e-01 9.15402651e-01
3.04902136e-01 6.10708117e-01 3.52779925e-01 1.39273584e+00
1.24463983e-01 3.44959460e-02 -5.18624306e-01 2.16810614e-01
7.58208096e-01 1.08360052e+00 -4.19828862e-01 -3.50323886e-01
-4.67666149e-01 4.70741034e-01 1.61709279e-01 3.26513231e-01
-8.35099638e-01 4.93309796e-02 8.53234828e-01 1.58744797e-01
4.25018936e-01 -5.49223900e-01 -3.75813484e-01 -1.16104507e+00
4.11463618e-01 -3.95175457e-01 5.68845533e-02 -1.44037867e+00
-1.32374549e+00 3.97834718e-01 4.12773222e-01 -1.58050871e+00
1.77067667e-01 -2.96273649e-01 -6.41170323e-01 1.24166429e+00
-1.67504907e+00 -1.57803619e+00 -9.12083566e-01 7.03889906e-01
7.13738441e-01 2.85047621e-01 3.76329750e-01 2.26841539e-01
-2.23494485e-01 -1.47118419e-01 -1.54832795e-01 -5.99133000e-02
3.18193972e-01 -9.38672006e-01 2.57637024e-01 6.97406769e-01
1.31601751e-01 6.50988638e-01 6.70827687e-01 -8.22846234e-01
-1.38524604e+00 -1.56266427e+00 7.97967494e-01 -6.84296668e-01
3.11409205e-01 -5.59988499e-01 -1.09740496e+00 5.96906602e-01
-5.04608691e-01 1.28534526e-01 1.85630769e-01 -1.47484735e-01
-1.59291700e-01 -7.80014023e-02 -9.56997693e-01 5.90793133e-01
1.25188971e+00 -5.70503175e-01 -9.71117079e-01 5.67412972e-01
1.46630478e+00 -6.32770360e-01 -4.91851717e-01 5.18770635e-01
-1.89983159e-01 -5.97437382e-01 1.30991340e+00 -2.10771143e-01
4.70205069e-01 -9.59940612e-01 -5.80799282e-01 -9.18901503e-01
-4.48655337e-01 5.50574213e-02 3.63662001e-03 1.65409660e+00
8.10661837e-02 -1.71240404e-01 8.09226453e-01 7.48476565e-01
-5.94493866e-01 -5.55216968e-01 -8.44252765e-01 -3.98771673e-01
-2.48248324e-01 -7.71012902e-01 1.20155799e+00 8.78130615e-01
-7.41403222e-01 2.08262488e-01 -1.85278699e-01 9.16699111e-01
6.53202951e-01 7.11866081e-01 7.92807043e-01 -1.14649606e+00
2.15837508e-01 -1.26966849e-01 -8.16402078e-01 -1.36492836e+00
2.48691559e-01 -8.65642250e-01 1.36227980e-01 -1.84978056e+00
3.87708992e-01 -8.19330156e-01 -2.01533541e-01 2.51236290e-01
-2.97037810e-01 7.60067115e-03 2.31266558e-01 5.23372531e-01
-6.58197165e-01 8.50765705e-01 1.40289712e+00 -4.78920043e-01
-1.64935440e-01 2.90228687e-02 -7.62743711e-01 6.27695799e-01
3.65684241e-01 -3.95063043e-01 -5.28694868e-01 -8.25023353e-01
-1.19909212e-01 -9.16902199e-02 9.30274189e-01 -9.88773465e-01
3.81953359e-01 -4.00124520e-01 3.70742321e-01 -1.43107486e+00
6.68925285e-01 -1.36974943e+00 2.83786267e-01 -2.74748176e-01
2.50063781e-02 -1.89507380e-01 2.61705816e-01 9.43758428e-01
-3.46373498e-01 -1.61017269e-01 4.59795237e-01 -3.87940198e-01
-1.04431248e+00 9.35891986e-01 4.22561347e-01 -1.89440355e-01
1.19110906e+00 -2.29960486e-01 -2.61186838e-01 -1.73844293e-01
-8.08617413e-01 6.25270069e-01 7.17349291e-01 6.82667375e-01
8.75337839e-01 -1.51259398e+00 -7.26645827e-01 2.62341857e-01
7.08054185e-01 1.04723656e+00 7.70559192e-01 3.81758362e-01
-4.53842908e-01 4.84900653e-01 1.14181668e-01 -1.25048697e+00
-1.18416655e+00 1.01178801e+00 2.80756623e-01 4.56071466e-01
-9.47382033e-01 7.12679446e-01 8.42409372e-01 -3.35417241e-01
1.85806200e-01 -5.98574758e-01 -2.51261294e-01 -2.42397070e-01
4.60492849e-01 -2.16115341e-01 -1.52812645e-01 -1.14958990e+00
-5.22231519e-01 1.20182502e+00 8.35132077e-02 2.99317837e-01
1.25617814e+00 -5.76354086e-01 -2.06398934e-01 5.45634031e-01
1.20822453e+00 -2.02756241e-01 -1.28183782e+00 -6.16604090e-01
-2.69703656e-01 -7.65357673e-01 2.02882320e-01 -5.47831833e-01
-8.28419685e-01 9.47884560e-01 3.65028501e-01 -1.76152363e-01
1.03122306e+00 5.69733799e-01 5.69104493e-01 1.86780468e-01
6.63581491e-01 -5.58082938e-01 -2.13276014e-01 3.04203898e-01
1.05488527e+00 -1.21507192e+00 -1.38959000e-02 -8.45236361e-01
-7.40566850e-01 7.48743117e-01 7.82267213e-01 -7.58237243e-02
6.31192207e-01 -1.52647704e-01 -6.31591529e-02 -3.99328560e-01
-5.70018351e-01 -3.00179154e-01 2.87236840e-01 8.39251220e-01
-4.45436031e-01 1.16956316e-01 4.09179300e-01 8.43530536e-01
-9.44761152e-04 -3.80932003e-01 2.68599182e-01 8.40882897e-01
-5.41298389e-01 -6.45264566e-01 -5.57129860e-01 6.00303560e-02
3.19874018e-01 -8.82855952e-02 -4.03289229e-01 7.29827166e-01
4.53478783e-01 8.99266541e-01 2.93255329e-01 -3.93015683e-01
5.52182794e-01 -8.58501345e-02 1.42588228e-01 -8.45324218e-01
-7.25894868e-02 6.65087551e-02 -3.01248401e-01 -4.44466412e-01
-6.73368454e-01 -6.39578104e-01 -1.44860554e+00 -1.50485402e-02
-3.75099242e-01 2.22159281e-01 8.46805453e-01 1.07925010e+00
5.91464102e-01 3.34581792e-01 6.64361238e-01 -1.24270022e+00
-2.12386549e-02 -7.66633868e-01 -6.12456024e-01 6.03491962e-01
5.60259283e-01 -7.95464516e-01 -4.72471535e-01 -4.37903591e-02]
|
[8.213172912597656, -3.1914725303649902]
|
c64d9fbf-e8a4-40d4-9ac8-d0c93b7be3c3
|
unsupervised-segmentation-via-semantic
|
2005.10513
| null |
https://arxiv.org/abs/2005.10513v1
|
https://arxiv.org/pdf/2005.10513v1.pdf
|
Unsupervised segmentation via semantic-apparent feature fusion
|
Foreground segmentation is an essential task in the field of image understanding. Under unsupervised conditions, different images and instances always have variable expressions, which make it difficult to achieve stable segmentation performance based on fixed rules or single type of feature. In order to solve this problem, the research proposes an unsupervised foreground segmentation method based on semantic-apparent feature fusion (SAFF). Here, we found that key regions of foreground object can be accurately responded via semantic features, while apparent features (represented by saliency and edge) provide richer detailed expression. To combine the advantages of the two type of features, an encoding method for unary region features and binary context features is established, which realizes a comprehensive description of the two types of expressions. Then, a method for adaptive parameter learning is put forward to calculate the most suitable feature weights and generate foreground confidence score map. Furthermore, segmentation network is used to learn foreground common features from different instances. By fusing semantic and apparent features, as well as cascading the modules of intra-image adaptive feature weight learning and inter-image common feature learning, the research achieves performance that significantly exceeds baselines on the PASCAL VOC 2012 dataset.
|
['Xi Li', 'Huimin Ma', 'Yidong Wang', 'Hongbing Ma']
|
2020-05-21
| null | null | null | null |
['foreground-segmentation']
|
['computer-vision']
|
[ 6.79550827e-01 -2.83419043e-01 -3.12414318e-01 -7.86401212e-01
-3.13252568e-01 -1.61645472e-01 3.68795007e-01 -2.03515887e-01
-3.49983662e-01 3.93663675e-01 1.41840741e-01 2.40881130e-01
-1.30831748e-01 -7.88720846e-01 -4.78096575e-01 -8.98137629e-01
4.88748103e-01 2.90740002e-02 8.29046905e-01 5.84038794e-02
5.76495707e-01 3.60726982e-01 -1.85276043e+00 4.51364815e-01
1.00678015e+00 1.19098318e+00 7.27806151e-01 3.20678324e-01
-8.86367917e-01 6.56049490e-01 -7.37323880e-01 -4.29250076e-02
-7.78579328e-04 -5.13247967e-01 -8.33981872e-01 4.61358011e-01
1.54250443e-01 -2.09198400e-01 1.23989955e-01 1.19698942e+00
3.05516034e-01 4.94193919e-02 6.14629269e-01 -1.37062919e+00
-7.40036666e-01 6.15819037e-01 -5.27333498e-01 3.80403310e-01
9.90478992e-02 1.79430977e-01 8.37404251e-01 -6.28970444e-01
3.71924102e-01 1.31199372e+00 3.07522088e-01 4.27283466e-01
-8.75649035e-01 -5.73367774e-01 5.87716222e-01 4.61364985e-01
-1.20122552e+00 -1.03086159e-01 9.45924401e-01 -3.75756562e-01
4.96550769e-01 2.93186486e-01 9.30464327e-01 7.56699800e-01
1.76227853e-01 1.19892693e+00 1.07583237e+00 -1.76027849e-01
-3.78224663e-02 3.31399411e-01 2.30059117e-01 4.53582853e-01
2.41764486e-01 -1.88576162e-01 -4.90207702e-01 3.82368237e-01
5.96458852e-01 2.89562255e-01 -2.57555783e-01 -1.48044571e-01
-1.21205723e+00 6.58468068e-01 6.45758033e-01 6.32722616e-01
-1.47696629e-01 -6.08208664e-02 2.11446539e-01 -1.01053126e-01
1.48802921e-01 2.54471034e-01 -6.30373776e-01 6.28209710e-02
-8.99485230e-01 -3.98376249e-02 3.43662560e-01 8.60446274e-01
1.26865017e+00 9.58742946e-02 -5.49706817e-01 7.32056737e-01
5.70753455e-01 5.73974669e-01 7.73842335e-01 -9.82466877e-01
-9.37169716e-02 1.05458236e+00 -3.33965331e-01 -1.36911559e+00
-2.44817808e-01 -2.13735238e-01 -6.13843560e-01 -6.28478080e-02
1.35699779e-01 -3.04348115e-02 -1.10106635e+00 1.56034994e+00
5.53394675e-01 3.52707297e-01 9.25314426e-02 1.04234958e+00
1.08929205e+00 8.22146952e-01 1.83017522e-01 -2.88390666e-01
1.29494917e+00 -1.05508721e+00 -1.01777792e+00 -2.48550177e-01
2.47078627e-01 -7.15823948e-01 9.30792809e-01 2.39289835e-01
-6.09954357e-01 -9.19769049e-01 -1.00961399e+00 -7.27875829e-02
-6.73158705e-01 1.74283069e-02 9.63110030e-01 4.87612724e-01
-7.99753189e-01 3.00316870e-01 -4.85026360e-01 -2.22159997e-01
5.70910096e-01 3.16905230e-01 9.19093378e-03 7.41408020e-03
-1.23242795e+00 6.82082057e-01 8.90433609e-01 3.98836315e-01
-7.22835600e-01 -5.41591227e-01 -7.33958423e-01 -1.26157746e-01
4.79605913e-01 -4.15345818e-01 8.68031979e-01 -1.71180689e+00
-1.45906079e+00 6.21769965e-01 -2.23785102e-01 -1.52980313e-01
1.33711249e-01 -1.92097083e-01 -5.79246283e-01 2.79405147e-01
2.72524089e-01 9.02090728e-01 1.05579615e+00 -1.35041714e+00
-1.11567879e+00 -2.13274568e-01 -2.33558834e-01 2.35023886e-01
-3.44533652e-01 5.09655625e-02 -4.90028918e-01 -4.67678219e-01
3.25957417e-01 -3.15921187e-01 -2.35287905e-01 -1.64020315e-01
-1.55341357e-01 -4.34334904e-01 1.24248290e+00 -5.76126814e-01
1.24499643e+00 -2.24331474e+00 1.40881494e-01 3.05154264e-01
7.16061518e-02 3.92596185e-01 9.39639881e-02 -3.51252884e-01
1.57756090e-01 6.04301319e-03 -5.29510558e-01 4.47887391e-01
-2.07181379e-01 4.38492209e-01 1.69021524e-02 9.62354802e-03
7.35131085e-01 9.79416490e-01 -9.89380419e-01 -1.01179850e+00
5.03057420e-01 2.34437570e-01 -2.80692339e-01 2.62799263e-01
-2.07166195e-01 3.63227725e-01 -8.52422953e-01 8.68746102e-01
6.56858385e-01 -1.74674779e-01 -1.09549671e-01 -4.96274322e-01
-1.75423518e-01 -3.63855034e-01 -1.35892844e+00 1.51105130e+00
-6.15895428e-02 3.61444205e-01 -7.55721182e-02 -1.21422839e+00
1.12338150e+00 2.46392898e-02 4.74618018e-01 -6.21881485e-01
3.58506829e-01 2.36896619e-01 -4.93044965e-02 -8.53443921e-01
3.44174564e-01 1.60385087e-01 -8.43985938e-03 -2.48457454e-02
2.67083913e-01 -2.69946724e-01 9.27905068e-02 1.00477636e-02
6.82316959e-01 3.90829563e-01 -1.34798037e-02 -3.42347234e-01
9.10340786e-01 6.82678586e-03 9.94445145e-01 4.70588446e-01
-5.17660916e-01 5.54106176e-01 2.80709416e-01 -4.52728868e-01
-4.65965420e-01 -1.05365705e+00 -2.81465977e-01 9.56520259e-01
8.65206420e-01 -1.58470750e-01 -9.20973659e-01 -6.63648427e-01
-3.84486653e-02 4.85087454e-01 -7.46711791e-01 -3.49565923e-01
-3.22916001e-01 -6.71204925e-01 1.79892778e-01 6.37829006e-01
1.05900228e+00 -1.31088352e+00 -7.28014767e-01 1.92398369e-01
-2.08429992e-01 -8.95914972e-01 -2.81300396e-01 2.44180411e-01
-7.16310799e-01 -9.03009057e-01 -5.11460185e-01 -9.95894074e-01
6.20217502e-01 6.89151704e-01 8.90018821e-01 3.58385742e-01
-4.44714665e-01 2.41072953e-01 -5.38048267e-01 -4.09759402e-01
2.61567235e-02 -1.02075092e-01 -3.39534193e-01 4.33583051e-01
5.20803392e-01 -2.43348256e-01 -5.11989534e-01 2.91934550e-01
-1.04264164e+00 2.15553597e-01 9.26047981e-01 8.18954885e-01
8.92940044e-01 8.71839672e-02 6.21446252e-01 -7.16072023e-01
2.07585007e-01 -3.63967776e-01 -4.07483637e-01 4.69355047e-01
-5.39971352e-01 -1.91379450e-02 2.54231960e-01 -4.05263394e-01
-1.45162773e+00 1.92242071e-01 2.40992513e-02 -2.83255547e-01
-3.86197120e-01 2.67621845e-01 -6.90525651e-01 -4.07646671e-02
2.43278340e-01 4.33464736e-01 -4.72561195e-02 -1.20787449e-01
5.02388179e-01 8.86570990e-01 5.37124336e-01 -5.80437958e-01
5.68986773e-01 1.85504571e-01 -4.03222948e-01 -7.69617260e-01
-1.01113784e+00 -4.60151911e-01 -7.74477482e-01 -4.99525040e-01
1.41572869e+00 -8.29224169e-01 -2.60127306e-01 7.28530824e-01
-9.15371060e-01 -1.25396565e-01 -2.76392490e-01 3.06240529e-01
-3.48658919e-01 3.40214223e-01 -4.60150629e-01 -6.22865617e-01
-3.05692869e-04 -1.47150970e+00 1.12449479e+00 9.61379766e-01
5.25774881e-02 -7.21501052e-01 -6.16080046e-01 4.07446176e-01
4.95564669e-01 2.41363063e-01 6.40305102e-01 -4.37112480e-01
-7.49324203e-01 1.65835589e-01 -5.93962193e-01 4.75321412e-01
4.53033537e-01 5.11753142e-01 -9.10470188e-01 2.10050732e-01
1.21492118e-01 -1.27265275e-01 9.74167764e-01 5.13690650e-01
1.53321409e+00 -1.09830178e-01 -4.49936181e-01 6.65070474e-01
1.25024855e+00 4.06626910e-01 6.49364173e-01 3.53682429e-01
7.90332615e-01 6.29492164e-01 9.54033494e-01 1.53165743e-01
3.10009152e-01 2.97728211e-01 3.15122575e-01 -1.61555365e-01
-1.17668666e-01 -6.29907101e-03 1.99801534e-01 7.63051510e-01
5.49816638e-02 1.83028623e-01 -4.83714044e-01 5.05259097e-01
-1.77930486e+00 -1.04688728e+00 -1.82704404e-01 1.69147170e+00
9.65547085e-01 2.54846841e-01 -1.45163789e-01 1.10148549e-01
1.12885439e+00 3.66381481e-02 -6.36954010e-01 -2.21773222e-01
-4.07798171e-01 -7.87622854e-02 3.15621793e-01 1.54151946e-01
-1.28978479e+00 1.18719351e+00 5.94121408e+00 1.09306014e+00
-1.23542845e+00 1.08694965e-02 6.70930326e-01 3.32214415e-01
-3.78430545e-01 8.57059807e-02 -7.97440231e-01 7.63742030e-01
5.49886405e-01 -1.40923085e-02 2.85398573e-01 8.83564055e-01
-1.37684658e-01 -3.21358383e-01 -6.04069352e-01 8.78758907e-01
9.51488018e-02 -1.25730288e+00 2.54092753e-01 -2.57017106e-01
8.83260608e-01 -3.16306859e-01 -1.07653104e-01 1.78599626e-01
3.03605963e-02 -8.78161371e-01 7.08845437e-01 7.63627470e-01
3.74336928e-01 -6.18487060e-01 8.09821188e-01 2.18309090e-01
-1.38845289e+00 -1.88911811e-01 -4.50712293e-01 1.08574763e-01
7.30640292e-02 6.38236523e-01 -2.98189461e-01 7.30729043e-01
8.68543863e-01 1.03511381e+00 -7.18263984e-01 9.97743130e-01
-3.72095764e-01 6.93798959e-01 -1.70998991e-01 -1.90663368e-01
2.92657822e-01 -2.35987186e-01 1.41711995e-01 1.32742774e+00
1.54265538e-01 7.39930123e-02 3.37513804e-01 1.06679082e+00
2.53408760e-01 1.87411189e-01 -2.73844868e-01 1.70693453e-02
4.72920030e-01 1.59813356e+00 -1.24740422e+00 -5.97332060e-01
-4.33639139e-01 1.01318300e+00 9.74670425e-02 3.60406309e-01
-9.57799673e-01 -4.82310236e-01 3.23881686e-01 -2.58773267e-01
4.07139122e-01 1.04394853e-01 -3.24131429e-01 -1.00584507e+00
-1.81896538e-01 -5.21274090e-01 2.84958005e-01 -7.04203665e-01
-1.16866875e+00 2.10772827e-01 1.47081658e-01 -1.00824416e+00
2.92223990e-01 -7.20967293e-01 -8.60257804e-01 7.64115691e-01
-1.58407283e+00 -1.09004402e+00 -5.48999131e-01 6.37822688e-01
6.49666309e-01 -9.51442719e-02 5.25014222e-01 1.83458835e-01
-7.96888411e-01 3.18683147e-01 -2.06942260e-01 1.46042421e-01
4.63888496e-01 -1.13464713e+00 -1.48274779e-01 9.51602280e-01
3.31600532e-02 4.75082517e-01 3.20487857e-01 -6.87250078e-01
-1.13600707e+00 -1.14394772e+00 4.97404575e-01 -1.79165989e-01
3.44400078e-01 1.49231581e-02 -1.06248116e+00 3.45951438e-01
1.58450872e-01 1.86145514e-01 6.27855599e-01 -1.95480764e-01
6.17423579e-02 -3.20643693e-01 -1.17014849e+00 5.14636576e-01
9.05361056e-01 -1.63719803e-01 -8.87318552e-01 1.33614600e-01
1.12078977e+00 -1.80258006e-01 -7.71803379e-01 7.57164478e-01
3.68591726e-01 -9.28031087e-01 7.89906681e-01 -2.49664605e-01
3.38518947e-01 -9.26358521e-01 -3.57584417e-01 -8.37389886e-01
-5.23534119e-01 -6.36484921e-02 1.14370417e-02 1.62368333e+00
2.87936807e-01 -5.35046756e-01 4.60178465e-01 3.43533695e-01
-3.31454426e-01 -7.62075663e-01 -6.90598786e-01 -3.27798009e-01
-3.55567485e-01 -3.37942332e-01 7.62286067e-01 8.42103362e-01
-4.42793220e-01 2.68245012e-01 1.79251190e-02 9.76059139e-02
2.77653366e-01 3.66509736e-01 5.26709914e-01 -1.24112117e+00
1.17434762e-01 -7.81318784e-01 -5.83732903e-01 -9.41532195e-01
2.59655118e-01 -9.96003807e-01 3.47746432e-01 -1.61167324e+00
4.54530597e-01 -4.22489583e-01 -7.00255036e-01 4.40337092e-01
-6.63662195e-01 1.79783806e-01 1.71725243e-01 -7.57924318e-02
-8.39233637e-01 5.60006976e-01 1.58708882e+00 -3.49772036e-01
-1.26613379e-02 -1.20974965e-01 -8.23083103e-01 8.23933542e-01
7.29559720e-01 -1.07375890e-01 -4.25006807e-01 -3.95189583e-01
-2.88696140e-01 -2.79883027e-01 2.87928045e-01 -1.17439842e+00
1.10418372e-01 -6.72325552e-01 9.13729548e-01 -7.28885472e-01
8.53310227e-02 -9.40209150e-01 -7.85950124e-02 3.14042032e-01
-2.63284713e-01 -3.18191677e-01 -3.26515269e-03 5.44808984e-01
-2.92318881e-01 -3.70965540e-01 5.96509039e-01 -2.14573696e-01
-1.45325315e+00 2.15257809e-01 -3.10843885e-01 1.61913812e-01
1.22424173e+00 -8.18333685e-01 -3.14848647e-02 2.02771828e-01
-4.64289904e-01 5.04181623e-01 2.37574652e-01 5.85424244e-01
9.12964284e-01 -1.42820370e+00 -4.07685518e-01 5.05463004e-01
2.60149222e-02 2.79153556e-01 5.04105449e-01 7.31723249e-01
-3.43104124e-01 -1.73155561e-01 -4.68649447e-01 -1.02352786e+00
-1.00976408e+00 5.05304575e-01 2.24619985e-01 1.79738760e-01
-3.34186673e-01 1.03499985e+00 2.91781574e-01 -1.83808938e-01
4.74969633e-02 -3.49234104e-01 -4.43736255e-01 2.62473911e-01
6.36700630e-01 1.21116363e-01 -2.47853026e-01 -9.21345711e-01
-3.14828992e-01 9.12657917e-01 -3.76591715e-03 2.75426358e-01
9.87041771e-01 -3.25113863e-01 -3.12376618e-01 5.49167395e-01
1.03599966e+00 -2.32706353e-01 -1.50924468e+00 -2.10677907e-01
5.14300764e-02 -5.96015930e-01 1.68443508e-02 -7.38480806e-01
-1.40579522e+00 9.56795454e-01 8.15194845e-01 1.32645503e-01
1.44755960e+00 -1.21934491e-03 7.72902250e-01 1.47221193e-01
2.73448497e-01 -1.37247145e+00 2.37753421e-01 3.79062951e-01
3.90351772e-01 -1.40110636e+00 -1.42238200e-01 -4.90389347e-01
-8.38363945e-01 1.15099502e+00 9.86410797e-01 -1.40786231e-01
5.43702781e-01 2.01603100e-01 2.35308066e-01 -3.74775492e-02
-3.37903798e-01 -4.54505920e-01 4.28272247e-01 6.98539019e-01
2.94158727e-01 7.69662261e-02 -4.78128403e-01 8.78251612e-01
2.88012866e-02 -1.72220275e-01 1.26801506e-01 9.65775311e-01
-9.17607784e-01 -8.13220382e-01 -2.81614721e-01 6.45075679e-01
-3.89346600e-01 1.38014153e-01 -2.83075780e-01 5.35896719e-01
7.58758128e-01 1.04273582e+00 1.41264170e-01 -6.64048910e-01
-1.24994349e-02 -7.50356987e-02 3.23371679e-01 -5.10546625e-01
-4.63049799e-01 6.95670694e-02 -5.30327380e-01 -6.09373868e-01
-9.99907196e-01 -5.39122045e-01 -1.71137464e+00 1.99270844e-01
-6.38880074e-01 1.15651920e-01 3.43991399e-01 1.20737946e+00
2.58200988e-02 8.99444818e-01 6.26986146e-01 -8.94172847e-01
-9.48014334e-02 -7.49407291e-01 -4.90714371e-01 5.32950699e-01
1.08856373e-02 -8.65114689e-01 -2.51004368e-01 1.77221641e-01]
|
[9.575494766235352, -0.3182496726512909]
|
43e273de-498d-4b47-b47f-2f881f16eb1a
|
mathematics-assisted-directed-evolution-and
|
2306.04658
| null |
https://arxiv.org/abs/2306.04658v1
|
https://arxiv.org/pdf/2306.04658v1.pdf
|
Mathematics-assisted directed evolution and protein engineering
|
Directed evolution is a molecular biology technique that is transforming protein engineering by creating proteins with desirable properties and functions. However, it is experimentally impossible to perform the deep mutational scanning of the entire protein library due to the enormous mutational space, which scales as $20^N$ , where N is the number of amino acids. This has led to the rapid growth of AI-assisted directed evolution (AIDE) or AI-assisted protein engineering (AIPE) as an emerging research field. Aided with advanced natural language processing (NLP) techniques, including long short-term memory, autoencoder, and transformer, sequence-based embeddings have been dominant approaches in AIDE and AIPE. Persistent Laplacians, an emerging technique in topological data analysis (TDA), have made structure-based embeddings a superb option in AIDE and AIPE. We argue that a class of persistent topological Laplacians (PTLs), including persistent Laplacians, persistent path Laplacians, persistent sheaf Laplacians, persistent hypergraph Laplacians, persistent hyperdigraph Laplacians, and evolutionary de Rham-Hodge theory, can effectively overcome the limitations of the current TDA and offer a new generation of more powerful TDA approaches. In the general framework of topological deep learning, mathematics-assisted directed evolution (MADE) has a great potential for future protein engineering.
|
['Guo-Wei Wei', 'Yuchi Qiu']
|
2023-06-06
| null | null | null | null |
['topological-data-analysis']
|
['graphs']
|
[ 1.42883971e-01 -7.04807788e-02 3.05915982e-01 1.83948845e-01
-1.19414367e-01 -4.49682802e-01 3.97722244e-01 2.35166788e-01
-2.40335956e-01 9.84984696e-01 -1.06491774e-01 -5.97619176e-01
-3.68881166e-01 -9.21184659e-01 -9.20767665e-01 -1.13905537e+00
-5.24035156e-01 5.49528658e-01 1.56093076e-01 -7.44781613e-01
3.64945173e-01 6.35845423e-01 -1.28063130e+00 6.32717237e-02
1.34623706e+00 6.24579251e-01 2.53972411e-01 6.27846062e-01
-5.17420530e-01 2.69968867e-01 -1.39541596e-01 -5.07955611e-01
1.27122357e-01 -6.27345204e-01 -6.59624040e-01 -6.32100582e-01
-3.62529382e-02 3.97673160e-01 -4.73258466e-01 1.03727996e+00
7.22296774e-01 7.62456805e-02 7.51823664e-01 -1.13281429e+00
-1.34711432e+00 3.04789424e-01 -6.00556254e-01 1.37081981e-01
2.47767761e-01 4.57595259e-01 9.37930703e-01 -1.20718455e+00
9.92631793e-01 1.23507583e+00 1.00278628e+00 3.60268354e-01
-1.38295555e+00 -2.77469605e-01 -2.82221675e-01 4.70461994e-01
-1.16501009e+00 1.64166495e-01 9.89530563e-01 -5.51407456e-01
1.20323884e+00 4.13481519e-03 1.03724575e+00 1.09163499e+00
8.17953646e-01 5.33526242e-01 1.01293945e+00 -6.54200196e-01
3.79996985e-01 -2.90773392e-01 -1.91142894e-02 1.19088078e+00
1.44387484e-01 4.13893700e-01 -6.59058630e-01 -6.22591913e-01
4.85340416e-01 1.15271673e-01 -3.44118297e-01 -8.01307797e-01
-1.31088388e+00 1.02683079e+00 4.63214338e-01 3.35330844e-01
-5.85591197e-01 -5.05480058e-02 4.10033971e-01 4.48937982e-01
2.22648516e-01 9.57266033e-01 -3.78102660e-01 -1.73411906e-01
-5.07231414e-01 2.58819401e-01 7.20717072e-01 4.95366782e-01
9.91059005e-01 3.20759676e-02 4.07960176e-01 5.44019699e-01
9.70973372e-02 2.48142436e-01 4.75583702e-01 -6.41788960e-01
-1.99372858e-01 1.09347975e+00 -2.52301246e-01 -1.29493117e+00
-3.76665324e-01 -3.85268144e-02 -1.03924274e+00 3.30881417e-01
1.83048472e-01 6.40431121e-02 -9.21480060e-01 1.96703172e+00
3.54068965e-01 1.79721322e-03 -6.44405931e-02 5.44092000e-01
2.27352887e-01 8.59254718e-01 -5.72213717e-02 -2.36651704e-01
9.51286256e-01 -5.78182459e-01 -4.59954083e-01 3.71509284e-01
7.85822034e-01 -4.48725790e-01 1.18602157e+00 2.72168040e-01
-8.85928512e-01 -3.11411619e-01 -1.23857331e+00 -5.63339936e-03
-1.03898871e+00 -4.65257674e-01 7.89619863e-01 4.24492925e-01
-1.32066762e+00 8.76043141e-01 -7.31225133e-01 -5.32746732e-01
3.33186865e-01 5.71918070e-01 -4.69139934e-01 1.52055591e-01
-1.60153615e+00 1.06366444e+00 4.63879228e-01 1.47411581e-02
-5.76921582e-01 -9.32989180e-01 -6.41113818e-01 -8.78565162e-02
2.05085114e-01 -7.09813893e-01 1.13242276e-01 -5.86554825e-01
-1.54468012e+00 7.67744064e-01 -1.77194364e-02 -5.39077997e-01
5.63064739e-02 1.97826296e-01 -3.23853195e-01 1.16237719e-02
-3.58776778e-01 6.96786582e-01 5.07124305e-01 -6.79599524e-01
-2.22919174e-02 -5.81480503e-01 -2.83731878e-01 4.53508757e-02
-5.55902958e-01 -3.56035233e-01 2.97827065e-01 -6.02418423e-01
1.18728526e-01 -1.15417862e+00 -3.33867043e-01 2.35732809e-01
-1.02719076e-01 -3.64274114e-01 1.07842147e+00 -6.06773496e-01
1.10719776e+00 -2.03717756e+00 9.98117685e-01 1.49931252e-01
5.50203800e-01 6.40709460e-01 -1.36266664e-01 8.33424449e-01
-2.28425458e-01 1.73026085e-01 -5.78039706e-01 5.44007778e-01
-3.32927182e-02 2.54751742e-02 -4.19839658e-02 3.67057979e-01
8.09330717e-02 1.39567983e+00 -8.95245433e-01 -2.39808962e-01
1.49042830e-01 7.64574111e-01 -4.89387125e-01 -6.32215664e-02
-4.90739644e-01 4.12515730e-01 -4.48112696e-01 5.68048716e-01
5.28331816e-01 -3.80038559e-01 2.45107636e-01 -4.15392965e-01
-3.72999370e-01 -2.39245623e-01 -4.29271370e-01 1.78861308e+00
-4.88503277e-02 6.49649858e-01 -1.75047666e-01 -1.20090079e+00
1.07833958e+00 8.23181197e-02 8.36082757e-01 -6.81067169e-01
-1.78374663e-01 3.45229447e-01 2.12323010e-01 -4.11807299e-01
7.18106702e-02 -1.09391347e-01 1.59476534e-01 4.22872216e-01
-6.75564781e-02 1.68756261e-01 1.04270585e-01 1.83235750e-01
1.39723504e+00 3.57120275e-01 1.85724437e-01 -5.80153108e-01
6.43035591e-01 4.61638510e-01 5.82846642e-01 1.39116138e-01
-4.36429232e-01 1.42992333e-01 4.16857958e-01 -8.62552524e-01
-1.67762589e+00 -1.00130439e+00 -1.18624307e-02 9.27988708e-01
4.75113392e-02 -2.81899899e-01 -6.97638631e-01 -2.81139880e-01
2.09027037e-01 2.91511059e-01 -5.48919022e-01 -7.92314947e-01
-1.02318752e+00 -1.10081303e+00 9.00494277e-01 2.02151522e-01
4.19073850e-01 -1.23175311e+00 -3.54414403e-01 3.39229017e-01
1.17005616e-01 -3.35343897e-01 -3.64058107e-01 3.11597556e-01
-7.83716023e-01 -1.12329900e+00 -9.78335440e-01 -1.15599430e+00
6.06287837e-01 -1.52297497e-01 8.18260550e-01 -1.20984659e-01
-8.48292828e-01 1.81281008e-02 -1.88979715e-01 -7.44404793e-02
-5.69190800e-01 -2.40279928e-01 5.94283700e-01 -3.21724027e-01
6.02834642e-01 -9.88444090e-01 -7.00123787e-01 1.29512772e-01
-8.67716789e-01 -6.36274833e-03 7.00912058e-01 1.22528279e+00
6.88928843e-01 2.13741168e-01 6.26601577e-01 -5.74454963e-01
8.25633168e-01 -3.53067815e-01 -5.21929383e-01 6.18683040e-01
-8.23169887e-01 4.57672983e-01 9.10410643e-01 -5.38164973e-01
-8.24217260e-01 -2.38229319e-01 -1.09133773e-01 -4.04541612e-01
1.35081574e-01 7.99772382e-01 -1.00121051e-01 -5.58186650e-01
7.62875199e-01 7.34820426e-01 3.73502403e-01 -3.15064311e-01
4.64896590e-01 4.99300539e-01 3.04103434e-01 -5.68883419e-01
6.10840261e-01 2.40048841e-01 3.40267330e-01 -1.13626409e+00
1.02013759e-01 1.29703999e-01 -8.18692744e-01 6.57947510e-02
1.00677180e+00 -2.36408591e-01 -9.86289561e-01 5.55542946e-01
-8.37473273e-01 -6.44810647e-02 -1.71079531e-01 1.96136639e-01
-6.13307536e-01 7.87990510e-01 -6.11695349e-01 -5.85612237e-01
-4.19903427e-01 -1.32362068e+00 5.08403599e-01 3.58625390e-02
-1.15793109e-01 -1.16870165e+00 5.07319272e-01 8.01446363e-02
6.85689926e-01 6.45634055e-01 1.87142408e+00 -4.95824873e-01
-4.87751871e-01 1.31573081e-02 8.80247653e-02 6.81313574e-02
1.59133315e-01 1.79107398e-01 -1.25617951e-01 -4.34196413e-01
-8.73962715e-02 -3.21248293e-01 7.64930129e-01 3.21828485e-01
8.49138737e-01 -1.10391667e-02 -5.31683922e-01 7.46677041e-01
1.34086847e+00 7.42001474e-01 5.52919924e-01 2.83555478e-01
7.11871088e-01 3.67683172e-01 2.19723493e-01 -1.48487994e-02
1.01700746e-01 5.40378571e-01 2.11299285e-01 3.89226689e-03
1.29689604e-01 -3.30183476e-01 3.69732767e-01 1.15870190e+00
-1.48587003e-01 -1.27148688e-01 -1.18762922e+00 2.20347211e-01
-1.77313137e+00 -8.72711420e-01 1.08591609e-01 2.02250433e+00
8.56514275e-01 -1.84064552e-01 -3.40098739e-02 1.14178240e-01
8.38890135e-01 -8.32139328e-03 -1.11928546e+00 -6.37232661e-01
-5.69372058e-01 3.36537927e-01 3.32781076e-01 2.43432045e-01
-7.70710170e-01 8.97007346e-01 6.33292484e+00 6.86788976e-01
-1.07185686e+00 -1.49389669e-01 3.72039676e-01 4.20626223e-01
-3.06606382e-01 7.60654882e-02 -3.21666062e-01 6.98253989e-01
1.11652052e+00 -3.81423861e-01 6.18184686e-01 7.48374760e-01
-1.43550068e-01 2.87398368e-01 -9.48967040e-01 9.63653564e-01
-1.31451935e-01 -1.75406659e+00 2.00756744e-01 3.32433164e-01
6.43563807e-01 1.11234672e-02 2.86367267e-01 2.20212653e-01
2.43088245e-01 -9.85214353e-01 -1.64748356e-01 5.49192011e-01
8.29293191e-01 -9.06661570e-01 4.29101139e-01 2.37248272e-01
-1.16682231e+00 -1.31518453e-01 -6.10512137e-01 3.23286027e-01
3.02774459e-01 5.80844760e-01 -3.92819941e-01 4.89488631e-01
4.62850571e-01 7.16119468e-01 -2.45303139e-01 7.20076084e-01
4.98833984e-01 1.27889186e-01 -1.68706000e-01 -3.84032577e-01
2.25210816e-01 -7.89894700e-01 8.35406244e-01 8.47680151e-01
2.67818928e-01 2.47921124e-01 -6.82053796e-04 1.20240569e+00
-5.17949834e-02 5.07540107e-02 -7.13808715e-01 -7.59174943e-01
3.47041458e-01 8.37179005e-01 -6.40383065e-01 -2.02385530e-01
-1.22231729e-01 1.11891878e+00 3.69604886e-01 3.32579851e-01
-8.97416472e-01 -6.86165035e-01 8.31633687e-01 -3.16348821e-02
1.60298616e-01 -5.55016935e-01 2.32891187e-01 -1.08332396e+00
-1.42741188e-01 -1.15407503e+00 9.33395475e-02 -5.45023084e-01
-1.38253438e+00 4.66831833e-01 -4.66017962e-01 -7.52181649e-01
2.84059644e-01 -9.82848883e-01 -4.89506245e-01 8.05785477e-01
-9.78104055e-01 -7.62730956e-01 1.81215122e-01 3.87070656e-01
3.97576094e-01 -5.68193316e-01 9.92987633e-01 3.94513398e-01
-5.64518631e-01 3.74050111e-01 6.18141532e-01 -3.33756775e-01
3.45623225e-01 -1.22637689e+00 6.85666025e-01 4.67249334e-01
-1.54283568e-01 8.90886605e-01 7.51264691e-01 -7.70451844e-01
-2.25407076e+00 -7.48737335e-01 5.72075784e-01 -2.41784647e-01
7.73856699e-01 -3.05705667e-01 -1.26508176e+00 2.74861366e-01
-5.78584075e-02 -3.25564563e-01 4.99646157e-01 -2.44679496e-01
-3.86736654e-02 1.70814201e-01 -1.13881969e+00 7.21219480e-01
1.40728128e+00 -7.72629619e-01 -5.82082272e-01 5.34544766e-01
1.01375604e+00 5.23656867e-02 -1.06275666e+00 6.25313759e-01
5.05631149e-01 -9.80159879e-01 1.15327621e+00 -9.05991852e-01
1.03064977e-01 -2.80128300e-01 -4.75634150e-02 -1.35097659e+00
-6.43961430e-01 -7.67625928e-01 -3.09127301e-01 6.29650056e-01
4.38155502e-01 -9.19781446e-01 9.32573974e-01 3.68670374e-01
-2.05252826e-01 -1.15056050e+00 -9.63700473e-01 -7.87403166e-01
4.59696531e-01 3.60206991e-01 6.43153369e-01 1.11360717e+00
3.85459721e-01 4.71877679e-02 -2.66018599e-01 -3.20731163e-01
5.54491818e-01 -3.21731605e-02 3.81896138e-01 -1.43380868e+00
-2.38720104e-01 -6.62489057e-01 -7.12325931e-01 -8.83553207e-01
1.76270649e-01 -1.02590132e+00 -3.06963235e-01 -1.27764964e+00
7.05405325e-02 -2.45698497e-01 -4.39638674e-01 1.66126147e-01
9.95789096e-02 -1.93528175e-01 -4.12320971e-01 3.01819295e-01
-2.56126940e-01 9.71051633e-01 1.16163409e+00 -1.25853464e-01
-4.10111487e-01 -6.00077510e-01 -2.56672770e-01 5.45885921e-01
6.89347327e-01 -1.58395976e-01 -3.86250049e-01 -9.05201659e-02
6.66651607e-01 -8.25856328e-02 1.60962105e-01 -8.18896890e-01
3.70477468e-01 -1.39624715e-01 2.44864076e-01 -5.01152635e-01
3.46147120e-01 -5.00876963e-01 4.22730178e-01 9.50330734e-01
-3.23757470e-01 5.51280856e-01 -9.85715017e-02 7.71835446e-01
-1.67643294e-01 3.77000898e-01 8.51639330e-01 -2.09047496e-01
-6.33239627e-01 5.25379300e-01 -3.99816871e-01 -1.78041905e-01
1.16935527e+00 -5.19528210e-01 -3.64346802e-01 2.33760297e-01
-8.00436139e-01 3.79527994e-02 7.79674828e-01 2.02891752e-01
9.35270309e-01 -1.20418048e+00 -3.60405505e-01 4.11257774e-01
-3.82710360e-02 -4.45380479e-01 1.61578745e-01 9.92481112e-01
-9.47717369e-01 6.21990681e-01 -5.49251497e-01 -4.61912602e-01
-1.13253736e+00 8.23420167e-01 3.32810581e-01 -4.40308303e-02
-8.55447352e-01 9.20354366e-01 1.12248778e-01 -7.22279251e-01
1.42168216e-02 2.08118483e-01 1.76542774e-01 -2.25000888e-01
1.05613761e-01 6.07873917e-01 -1.43069193e-01 -6.57387376e-01
-4.07079756e-01 8.07508647e-01 -4.39600088e-02 2.74479151e-01
1.48885763e+00 1.00709967e-01 -6.35378361e-01 2.03720152e-01
1.30635214e+00 -4.00348783e-01 -8.21972132e-01 -8.50380063e-02
2.87124127e-01 -7.97436759e-02 -1.10995561e-01 -7.48908639e-01
-6.58344626e-01 9.47132409e-01 6.69921339e-01 1.04139812e-01
9.09326732e-01 -3.03501934e-01 1.18903244e+00 7.44943440e-01
6.65189862e-01 -9.27567065e-01 1.81810349e-01 5.14823198e-01
6.91065967e-01 -9.17975307e-01 -2.68025070e-01 1.32394195e-01
-4.23020899e-01 1.04849982e+00 4.28947330e-01 2.43905112e-02
7.28556335e-01 1.05225362e-01 -2.82143354e-01 -5.25415540e-01
-7.50209749e-01 1.80091947e-01 -2.12944602e-03 6.85024679e-01
5.15889943e-01 -1.40821218e-01 -5.28636932e-01 2.28554313e-03
2.94334106e-02 -9.02738273e-02 9.96344686e-02 1.05627620e+00
-7.34002113e-01 -1.33536434e+00 -8.37245286e-02 2.71574676e-01
-6.47836104e-02 -1.25929683e-01 -7.77721226e-01 5.98525465e-01
1.58048674e-01 3.02873701e-01 -3.50749433e-01 -5.47504067e-01
3.06237079e-02 5.61900556e-01 4.56210554e-01 -1.65547982e-01
-2.52044111e-01 -5.24820030e-01 -4.54083443e-01 -3.40653539e-01
-1.54656157e-01 -2.68098116e-01 -1.50643933e+00 -7.46039510e-01
-5.17392099e-01 4.29397345e-01 7.27330983e-01 6.89979970e-01
1.02128541e+00 2.55023271e-01 4.54473346e-01 -6.49669290e-01
-3.95592749e-01 -5.34473419e-01 -4.52165365e-01 3.32962483e-01
7.47180730e-02 -8.19498003e-01 -3.46735835e-01 -1.90061197e-01]
|
[4.892721652984619, 5.639756679534912]
|
0299bd90-f2e3-48db-b955-830e0fd38c71
|
multi-source-pointer-network-for-product
|
1808.06885
| null |
http://arxiv.org/abs/1808.06885v3
|
http://arxiv.org/pdf/1808.06885v3.pdf
|
Multi-Source Pointer Network for Product Title Summarization
|
In this paper, we study the product title summarization problem in E-commerce
applications for display on mobile devices. Comparing with conventional
sentence summarization, product title summarization has some extra and
essential constraints. For example, factual errors or loss of the key
information are intolerable for E-commerce applications. Therefore, we abstract
two more constraints for product title summarization: (i) do not introduce
irrelevant information; (ii) retain the key information (e.g., brand name and
commodity name). To address these issues, we propose a novel multi-source
pointer network by adding a new knowledge encoder for pointer network. The
first constraint is handled by pointer mechanism. For the second constraint, we
restore the key information by copying words from the knowledge encoder with
the help of the soft gating mechanism. For evaluation, we build a large
collection of real-world product titles along with human-written short titles.
Experimental results demonstrate that our model significantly outperforms the
other baselines. Finally, online deployment of our proposed model has yielded a
significant business impact, as measured by the click-through rate.
|
['Peng Jiang', 'Hanxiao Sun', 'Xiaobo Wang', 'Fei Sun', 'Wenwu Ou', 'Changhua Pei']
|
2018-08-21
| null | null | null | null |
['abstractive-sentence-summarization']
|
['natural-language-processing']
|
[ 5.06245673e-01 3.11688632e-01 -5.83316147e-01 -2.41762877e-01
-7.27843761e-01 -4.42485809e-01 1.45367026e-01 4.75860864e-01
-4.46287185e-01 8.17811847e-01 7.17945576e-01 -2.06894130e-01
5.26777431e-02 -5.90602279e-01 -9.89422381e-01 -2.28588849e-01
2.32005939e-01 6.26979992e-02 2.71763533e-01 -3.60951036e-01
5.97317755e-01 -1.93501934e-01 -1.18911362e+00 4.20426041e-01
1.31000233e+00 9.47155833e-01 4.81261432e-01 2.30180264e-01
-2.73702711e-01 5.70558965e-01 -8.99467289e-01 -7.35984981e-01
1.41652614e-01 -3.66264820e-01 -8.37644577e-01 1.19940408e-01
5.89838684e-01 -5.97173393e-01 -3.30339700e-01 1.28632247e+00
6.14265263e-01 1.56537250e-01 1.90479890e-01 -9.32708859e-01
-1.05881047e+00 1.25082040e+00 -9.28304791e-01 3.04060727e-01
4.31313455e-01 4.15088795e-02 1.42291701e+00 -5.21503329e-01
7.88147628e-01 1.11330187e+00 5.23480475e-01 1.92341849e-01
-1.00733566e+00 -3.86714131e-01 5.04759252e-01 6.31059855e-02
-1.13250744e+00 -5.31088114e-01 7.42153049e-01 -2.14506183e-02
1.30514812e+00 3.18981975e-01 2.19425991e-01 1.11609614e+00
4.34840709e-01 1.10389340e+00 1.45365581e-01 -1.73432082e-01
1.41301125e-01 2.53490984e-01 5.27160466e-01 4.32161838e-01
7.08112061e-01 -5.57659090e-01 -6.65783763e-01 -1.47102430e-01
2.16459066e-01 -5.21987714e-02 -4.29058969e-01 8.38208105e-03
-1.05037773e+00 7.26860702e-01 3.63585174e-01 -8.56936816e-03
-6.30771339e-01 9.15884301e-02 7.50953972e-01 1.41040415e-01
3.97915214e-01 5.90479910e-01 -6.82958186e-01 -1.43140450e-01
-9.94659543e-01 3.58858854e-01 9.65793908e-01 1.54091179e+00
3.73568028e-01 -1.05753914e-01 -4.49187607e-01 7.64284551e-01
2.39005104e-01 4.11750853e-01 6.47203326e-01 -7.40515530e-01
1.00740540e+00 4.57581848e-01 1.68870360e-01 -1.01699150e+00
-2.93573052e-01 -8.22507322e-01 -9.62218285e-01 -7.75388181e-01
-2.12887853e-01 -2.66643822e-01 -8.83050263e-01 1.52672195e+00
-2.09294185e-01 -4.48720809e-03 6.65721446e-02 6.17899418e-01
1.34025967e+00 8.74269128e-01 -9.58122164e-02 -6.98512733e-01
1.47520697e+00 -1.30408299e+00 -1.22758889e+00 -3.25036675e-01
5.20370066e-01 -7.66160965e-01 1.08599436e+00 2.41832867e-01
-1.33824861e+00 -5.26075244e-01 -1.20693302e+00 -2.13674441e-01
-1.69157341e-01 2.06444502e-01 6.84286952e-01 4.44237858e-01
-6.43637955e-01 6.96374536e-01 -6.50234759e-01 -3.59516561e-01
3.55389416e-01 2.96502024e-01 4.55721430e-02 9.26778689e-02
-1.32085168e+00 3.91872793e-01 6.77457213e-01 9.84607935e-02
-9.64006260e-02 -7.90063024e-01 -1.12250042e+00 5.24417877e-01
6.68364525e-01 -1.01802909e+00 1.69164646e+00 -7.93087363e-01
-1.42407966e+00 3.03712040e-01 -3.68217528e-01 -4.84850913e-01
1.40666768e-01 -7.17165172e-01 -4.49523866e-01 3.34840454e-02
3.04390460e-01 4.28355068e-01 5.31798482e-01 -1.07185328e+00
-7.92971551e-01 -2.03660145e-01 2.45635454e-02 3.01211089e-01
-4.52492654e-01 -5.65334149e-02 -7.14587688e-01 -8.67642105e-01
1.38567835e-01 -5.83315194e-01 -4.98821139e-02 -8.34373116e-01
-9.67201352e-01 -2.48110458e-01 4.88913536e-01 -1.07349682e+00
1.94145799e+00 -2.31468225e+00 -2.24585850e-02 -1.03045225e-01
1.03409037e-01 3.38463306e-01 -1.68761283e-01 5.02427042e-01
2.23659158e-01 5.46836197e-01 -8.28257948e-02 -3.09392244e-01
1.16117597e-01 -5.07646836e-02 -6.21909440e-01 -2.17253253e-01
5.54890819e-02 1.28501451e+00 -7.89281368e-01 -5.24532318e-01
-3.79850388e-01 -6.87628016e-02 -6.81879461e-01 -2.05559492e-01
-3.38724554e-01 -1.64237738e-01 -5.47092140e-01 6.94618225e-01
7.89743125e-01 -3.90730530e-01 -1.17918253e-02 -2.68540442e-01
1.72614709e-01 8.58134568e-01 -9.13770020e-01 2.03349042e+00
-2.56924570e-01 1.20814644e-01 -2.26671454e-02 -4.49468315e-01
5.57593465e-01 1.68895751e-01 2.33966142e-01 -9.91345823e-01
6.90626353e-02 9.55625251e-02 -1.87604219e-01 -4.74496216e-01
1.26782477e+00 2.77082205e-01 -2.87330806e-01 3.02673340e-01
2.19184551e-02 1.79529324e-01 3.66736203e-01 6.34838283e-01
9.98309135e-01 3.68653536e-02 2.84233391e-01 1.00925686e-02
2.97572821e-01 -1.42859081e-02 1.04712224e+00 9.42553401e-01
-9.07681808e-02 5.01482666e-01 7.24693775e-01 -4.50396203e-02
-7.30717659e-01 -7.91208684e-01 2.08914116e-01 9.77020800e-01
4.39992636e-01 -7.93276787e-01 -7.96001732e-01 -7.70333588e-01
1.16808705e-01 9.80771840e-01 -2.56704509e-01 -4.90986019e-01
-7.43883789e-01 -5.69520831e-01 3.49637717e-01 6.47969723e-01
7.56191254e-01 -1.07887745e+00 -3.66678149e-01 4.22311515e-01
-4.99566615e-01 -1.10242152e+00 -1.03351939e+00 3.15378681e-02
-1.00862563e+00 -4.54879761e-01 -6.30764425e-01 -9.03876603e-01
2.99954951e-01 4.83244985e-01 1.01918578e+00 -1.38295263e-01
1.73461303e-01 -7.90716782e-02 -4.32272017e-01 -4.05636787e-01
-1.43473223e-01 7.82832980e-01 -8.44107866e-02 -3.42509240e-01
3.48044783e-01 -3.68562698e-01 -5.00020921e-01 -8.22626352e-02
-7.74690449e-01 -4.97372970e-02 9.52148974e-01 7.50714421e-01
5.48933923e-01 2.61810362e-01 1.11118209e+00 -1.16387093e+00
1.17571235e+00 -3.16101611e-01 -2.72764772e-01 4.08991903e-01
-6.27916276e-01 1.12164654e-01 4.77703273e-01 -3.59713703e-01
-1.34752750e+00 -1.08609013e-01 -2.26336479e-01 1.46715432e-01
2.28004023e-01 1.02651727e+00 -4.99526024e-01 6.97085798e-01
2.83110827e-01 2.94374764e-01 -3.29274982e-01 -7.56870449e-01
4.99806970e-01 9.34312165e-01 6.63466156e-01 -6.69654012e-02
3.34142476e-01 -5.91992736e-02 -5.90200901e-01 -7.46096194e-01
-8.49063218e-01 -6.15702331e-01 -3.51530999e-01 5.13381362e-01
4.75879490e-01 -9.37075257e-01 -7.78943598e-01 3.41014922e-01
-1.38409841e+00 2.77777255e-01 -3.93303603e-01 4.08159345e-01
-2.75945336e-01 6.93396568e-01 -9.00000572e-01 -4.03952628e-01
-8.75343502e-01 -8.80962670e-01 1.09015119e+00 4.87355143e-01
-4.50854689e-01 -4.46564883e-01 -2.96182096e-01 4.13517267e-01
3.80698442e-01 -2.38942042e-01 1.05049753e+00 -7.90080726e-01
-4.03204918e-01 -1.96715459e-01 -2.51527995e-01 3.44249636e-01
2.66572058e-01 -3.21674377e-01 -5.53578615e-01 -1.56480923e-01
5.62820360e-02 -7.42579028e-02 1.42535210e+00 5.27455986e-01
1.03337014e+00 -6.90448642e-01 -4.77336437e-01 4.47743237e-01
1.08758259e+00 3.58297884e-01 6.63964391e-01 2.90585190e-01
7.54773617e-01 3.67292136e-01 7.18811989e-01 5.39939582e-01
4.93053734e-01 5.97126961e-01 1.52861938e-01 1.31227255e-01
-8.47375393e-02 -6.76543117e-01 3.98694813e-01 1.16768920e+00
2.98553407e-01 -7.19947219e-01 -2.38920197e-01 3.81486028e-01
-2.12463570e+00 -9.11504388e-01 -8.47068280e-02 2.10503769e+00
9.99848723e-01 6.19930327e-01 -4.58746925e-02 -2.45610207e-01
6.61701024e-01 3.30505699e-01 -8.41548800e-01 -5.55279195e-01
-2.28208542e-01 -1.56653255e-01 7.08559573e-01 3.16406965e-01
-1.01145673e+00 1.04077423e+00 5.91914892e+00 7.48882174e-01
-8.59238803e-01 1.25379473e-01 3.69703263e-01 -3.16489488e-01
-5.59357226e-01 -2.59807825e-01 -1.21010113e+00 7.86986649e-01
7.15195239e-01 -5.02736688e-01 1.09073080e-01 7.32260823e-01
2.22810477e-01 -1.63288593e-01 -1.01290834e+00 9.10252750e-01
3.09803069e-01 -1.35282421e+00 2.57724255e-01 1.06588438e-01
6.53056502e-01 -9.01318938e-02 -7.63377547e-02 5.94623268e-01
-4.77683954e-02 -5.15182078e-01 8.01521599e-01 2.38120481e-01
5.13321340e-01 -9.30164814e-01 1.09498084e+00 2.92834371e-01
-8.49772573e-01 1.13608856e-02 -4.06256944e-01 9.11576822e-02
6.13199770e-01 6.49061382e-01 -5.03272593e-01 8.27221513e-01
4.55138892e-01 8.68001819e-01 -5.08915246e-01 1.07915759e+00
-2.04811305e-01 6.73334539e-01 -2.07822785e-01 -6.84095994e-02
1.28847837e-01 2.57916059e-02 5.08488119e-01 1.26073074e+00
2.66180545e-01 4.08018455e-02 -7.76898041e-02 8.75078022e-01
-6.38227165e-01 9.89298448e-02 -3.57369572e-01 -2.90809572e-01
4.38885689e-01 9.82084572e-01 -6.16688609e-01 -3.93533170e-01
-4.32748765e-01 1.37450433e+00 -6.78658187e-02 5.17009437e-01
-8.33324134e-01 -9.92207825e-01 3.15703005e-01 -8.61319453e-02
7.00137138e-01 4.61597964e-02 -3.52585405e-01 -1.25094628e+00
6.35274470e-01 -6.35679305e-01 1.58812881e-01 -6.16486251e-01
-9.35772359e-01 3.91391784e-01 -2.09075049e-01 -8.59063029e-01
-4.62038815e-02 2.62442194e-02 -4.21554625e-01 7.29909241e-01
-1.50435507e+00 -8.19135129e-01 8.81662145e-02 -1.32834449e-01
8.58862519e-01 -4.71326783e-02 3.83131266e-01 4.61785942e-01
-8.07735562e-01 8.12520862e-01 9.73758027e-02 -1.59577187e-02
7.26753414e-01 -1.09943938e+00 8.58346045e-01 9.05775666e-01
-1.25038818e-01 1.07703364e+00 6.88358426e-01 -1.09973311e+00
-1.34159732e+00 -1.03724158e+00 1.37483299e+00 -2.24775076e-01
3.62868398e-01 -3.82376134e-01 -1.01750529e+00 8.44102025e-01
5.91675222e-01 -6.35989606e-01 6.51579559e-01 3.24946344e-01
-7.87652880e-02 -1.37784988e-01 -8.97605360e-01 4.53714371e-01
1.07302022e+00 -9.83123258e-02 -9.81447637e-01 4.98936594e-01
1.64612126e+00 -3.90961170e-01 -6.38688624e-01 3.26271951e-01
3.72583658e-01 -4.45289671e-01 7.24203527e-01 -5.99678934e-01
6.64904952e-01 -1.68614075e-01 3.40912379e-02 -1.39719129e+00
-4.59410518e-01 -8.21727633e-01 -5.37826896e-01 1.69241810e+00
5.90775907e-01 -4.97939974e-01 7.00846374e-01 5.13529360e-01
-4.23369408e-01 -7.15321541e-01 -6.21503174e-01 -8.64701569e-01
-2.33210832e-01 1.03937484e-01 7.88467944e-01 4.95669395e-01
4.06942517e-01 1.08770585e+00 -4.24448490e-01 -4.07474972e-02
2.28987604e-01 2.30328932e-01 3.66423756e-01 -8.78522933e-01
-2.84979343e-01 -5.35875678e-01 1.75678022e-02 -1.70758271e+00
4.19890694e-02 -7.10349381e-01 4.55056205e-02 -1.93430209e+00
6.17230117e-01 -1.31893856e-02 -3.44690472e-01 3.49016279e-01
-3.28865260e-01 -3.74669224e-01 1.03190236e-01 1.91297904e-01
-1.05751884e+00 6.40851557e-01 1.11762726e+00 -3.08193386e-01
-5.56722224e-01 2.28403077e-01 -1.49300110e+00 3.20569068e-01
7.94843078e-01 -3.63447726e-01 -4.55582947e-01 -5.13598740e-01
5.92054188e-01 1.27119467e-01 -1.70829564e-01 -4.40558761e-01
4.35342818e-01 5.71325682e-02 -6.91085383e-02 -9.57298815e-01
-9.56424326e-02 -4.93038893e-01 -4.28667992e-01 3.70028675e-01
-5.39515436e-01 1.56341925e-01 2.11266667e-01 7.44914591e-01
-3.63134444e-01 -4.25618827e-01 4.90113571e-02 -7.62565881e-02
-7.36819327e-01 -1.47857189e-01 -1.44184381e-01 2.15302125e-01
6.09733582e-01 -1.37559012e-01 -5.66596627e-01 -3.57409894e-01
-2.73321897e-01 5.98360896e-01 2.40911752e-01 7.57896006e-01
5.22480130e-01 -1.08249140e+00 -6.49165690e-01 3.81482877e-02
2.34819651e-01 6.72187582e-02 2.38414451e-01 6.62212670e-01
-2.53030479e-01 8.94580185e-01 1.21896781e-01 -1.78158626e-01
-1.14041913e+00 8.47962379e-01 -1.59484535e-01 -3.54420602e-01
-5.76687515e-01 7.44124472e-01 1.54483110e-01 -1.73895881e-01
5.23793817e-01 -8.82721066e-01 -2.79923916e-01 2.32731000e-01
7.37136424e-01 2.90874481e-01 4.41856623e-01 -4.15168613e-01
-2.66367376e-01 8.08487982e-02 -7.99860477e-01 1.07082449e-01
1.28549659e+00 -5.17404735e-01 -3.17394771e-02 2.76307881e-01
1.06790316e+00 4.33922298e-02 -7.91082263e-01 -4.72143501e-01
1.57504171e-01 -1.90514088e-01 1.86605081e-01 -1.10566318e+00
-1.01224446e+00 4.78060514e-01 1.03343017e-01 1.22900642e-01
1.15757930e+00 -2.08014742e-01 1.42630088e+00 5.88716388e-01
2.50457227e-01 -1.36786687e+00 -1.44936830e-01 5.12412846e-01
7.63512909e-01 -1.17254663e+00 2.99196482e-01 -5.81645608e-01
-8.07461143e-01 5.68133891e-01 5.45990825e-01 3.74524444e-01
2.21842319e-01 -1.08142719e-01 -2.59976834e-01 -2.13161111e-01
-7.08886623e-01 9.75623205e-02 2.13690251e-01 1.53900981e-01
5.10630548e-01 -2.49344539e-02 -7.15371788e-01 1.15108931e+00
-2.16403693e-01 1.90324664e-01 7.09073186e-01 1.03991568e+00
-5.24365604e-01 -7.53445506e-01 2.23966807e-01 7.18477786e-01
-8.45896721e-01 -5.42490125e-01 -3.84593278e-01 4.91691440e-01
-1.63387194e-01 1.00834143e+00 3.48838903e-02 -2.57218212e-01
7.73416102e-01 -1.17035575e-01 1.95160717e-01 -8.85837555e-01
-8.36109817e-01 2.34928161e-01 2.45437384e-01 -4.22150880e-01
-8.62667039e-02 -6.75099730e-01 -1.18016410e+00 -3.06418389e-01
-7.11006582e-01 2.05118150e-01 6.13948762e-01 5.58749974e-01
8.89200330e-01 9.71253276e-01 3.90774339e-01 -4.40167576e-01
-9.06217456e-01 -1.22132325e+00 -7.52522051e-01 3.02114278e-01
2.78807729e-01 -3.16468418e-01 -1.74789995e-01 1.39634982e-01]
|
[12.393524169921875, 9.353321075439453]
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.