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
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
f5800eb1-c4d2-46b7-a126-1269a1fe54a7
|
msdc-exploiting-multi-state-power-consumption
|
2302.05565
| null |
https://arxiv.org/abs/2302.05565v1
|
https://arxiv.org/pdf/2302.05565v1.pdf
|
MSDC: Exploiting Multi-State Power Consumption in Non-intrusive Load Monitoring based on A Dual-CNN Model
|
Non-intrusive load monitoring (NILM) aims to decompose aggregated electrical usage signal into appliance-specific power consumption and it amounts to a classical example of blind source separation tasks. Leveraging recent progress on deep learning techniques, we design a new neural NILM model Multi-State Dual CNN (MSDC). Different from previous models, MSDC explicitly extracts information about the appliance's multiple states and state transitions, which in turn regulates the prediction of signals for appliances. More specifically, we employ a dual-CNN architecture: one CNN for outputting state distributions and the other for predicting the power of each state. A new technique is invented that utilizes conditional random fields (CRF) to capture state transitions. Experiments on two real-world datasets REDD and UK-DALE demonstrate that our model significantly outperform state-of-the-art models while having good generalization capacity, achieving 6%-10% MAE gain and 33%-51% SAE gain to unseen appliances.
|
['Liehuang Zhu', 'Bakh Khoussainov', 'Yiwei Liu', 'Yang Chen', 'Zijian Zhang', 'Jiamou Liu', 'Jialing He']
|
2023-02-11
| null | null | null | null |
['non-intrusive-load-monitoring', 'non-intrusive-load-monitoring', 'non-intrusive-load-monitoring']
|
['knowledge-base', 'miscellaneous', 'time-series']
|
[ 3.53452325e-01 1.27088368e-01 -8.23947072e-01 -4.98874784e-01
-8.25184464e-01 -3.37035656e-01 5.59190750e-01 -3.09382647e-01
2.02590451e-01 6.38244390e-01 3.19457710e-01 -5.17136574e-01
1.31407842e-01 -5.78960598e-01 -7.02774405e-01 -8.34819376e-01
2.94914003e-02 2.72252500e-01 -4.48586076e-01 1.66158482e-01
-4.29486722e-01 2.69877285e-01 -1.21424067e+00 1.52732626e-01
5.09845674e-01 1.43018305e+00 -1.57805115e-01 6.30464375e-01
6.07355610e-02 1.27104831e+00 -9.07694995e-01 -1.59365013e-01
-9.54673998e-03 -2.02599600e-01 -8.91750693e-01 -2.02980428e-03
-1.77871659e-02 -5.44157863e-01 -6.33133769e-01 1.06302536e+00
3.85145217e-01 -2.10623607e-01 4.86232638e-01 -1.55674350e+00
-8.22725952e-01 1.12354016e+00 -5.94251335e-01 2.69024342e-01
3.57067585e-01 3.11464667e-01 9.51300561e-01 1.80367261e-01
-3.15344900e-01 9.20497954e-01 6.73781216e-01 6.82808340e-01
-1.63862586e+00 -1.03705812e+00 2.65804887e-01 2.01699764e-01
-1.14322317e+00 -5.91780901e-01 1.01008809e+00 -1.36439204e-01
1.50442576e+00 4.26293671e-01 1.50114223e-01 1.76136482e+00
1.51871502e-01 1.21106994e+00 1.17230284e+00 -1.88829303e-01
3.23895544e-01 1.95001364e-02 4.37375158e-01 3.90455186e-01
6.81408346e-02 4.77297194e-02 -5.10579765e-01 -3.20858479e-01
2.94725657e-01 5.53749725e-02 -2.37695277e-01 1.62905484e-01
-8.05729926e-01 5.44439316e-01 1.89798981e-01 4.23581362e-01
-4.63020056e-01 3.81190211e-01 3.33131373e-01 -7.52555430e-02
4.90680933e-01 -7.31753111e-02 -6.90388024e-01 -2.19946355e-01
-1.06991613e+00 -2.08552510e-01 9.33470428e-01 9.63231444e-01
6.18696868e-01 4.45308357e-01 -5.22601545e-01 5.56081057e-01
4.22158986e-01 6.71879351e-01 6.39106393e-01 -7.00782657e-01
4.90022093e-01 6.09764874e-01 -5.20213321e-02 -2.53495723e-01
-6.52558863e-01 -4.29970026e-01 -1.30290616e+00 -1.52786970e-01
-1.37722984e-01 -2.81252980e-01 -1.19568133e+00 2.07949686e+00
-1.63058028e-01 7.45973945e-01 3.95036899e-02 3.47481936e-01
5.00972152e-01 6.66016400e-01 3.56271982e-01 -3.05868924e-01
1.25854397e+00 -7.87741423e-01 -9.83695626e-01 -2.91913301e-01
3.01079415e-02 -1.81184962e-01 5.52868545e-01 3.85896087e-01
-9.64767814e-01 -4.16043401e-01 -1.27597499e+00 2.11376205e-01
-4.67670411e-01 3.15521717e-01 7.60110199e-01 7.06459105e-01
-6.84710622e-01 8.04837406e-01 -1.19741750e+00 -6.29797578e-02
7.66123414e-01 7.02557862e-01 6.98580965e-02 3.64485711e-01
-1.11077678e+00 6.62703335e-01 3.15499336e-01 1.60201892e-01
-1.24370360e+00 -6.34522557e-01 -8.81753266e-01 2.64066875e-01
1.23906516e-01 -4.41533327e-01 1.62416053e+00 -7.29013085e-01
-1.96673656e+00 3.75924736e-01 -1.77165940e-01 -8.64877939e-01
8.41003880e-02 -1.66107699e-01 -1.19674158e+00 -1.52721450e-01
1.30956590e-01 1.68901831e-01 1.03360522e+00 -1.11256862e+00
-7.05425084e-01 -4.30135339e-01 -1.22447573e-01 -4.85667348e-01
-4.46311325e-01 -2.43252024e-01 -2.87683547e-01 -4.11343068e-01
-3.14879328e-01 -8.83795023e-01 -1.10519864e-02 -8.48985672e-01
-1.10550547e+00 -5.28498232e-01 1.05324268e+00 -7.63058305e-01
1.40847552e+00 -2.10594535e+00 -3.17314416e-01 1.22658417e-01
2.51209438e-01 3.92918617e-01 -8.72666612e-02 3.02850623e-02
-4.68815386e-01 -2.64445662e-01 -2.42270038e-01 -8.63086581e-01
5.90135574e-01 4.13929909e-01 -3.80935848e-01 6.23135388e-01
1.74846143e-01 1.31791234e+00 -7.74728835e-01 5.80623969e-02
6.49386466e-01 6.00987017e-01 1.41246896e-02 3.00181121e-01
-2.74931103e-01 3.08417141e-01 -1.81864008e-01 6.59808517e-01
7.73594618e-01 -6.25242352e-01 4.56394821e-01 -3.42122972e-01
3.37658435e-01 7.55785108e-01 -9.09696519e-01 1.65946198e+00
-8.20076764e-01 5.18059075e-01 1.08000532e-01 -1.06622219e+00
5.02509475e-01 6.06206000e-01 7.99005091e-01 -8.58122706e-01
4.50783938e-01 -3.59658122e-01 -4.33827907e-01 -1.55729994e-01
1.36979207e-01 5.57205677e-02 -4.64865029e-01 4.99822676e-01
5.06655514e-01 4.43171203e-01 -1.09095290e-01 -9.77830291e-02
1.30052459e+00 -6.83840886e-02 2.30660200e-01 -2.56177962e-01
2.49997824e-01 -8.02816749e-01 9.13726211e-01 6.10553563e-01
-2.77353793e-01 -8.20680186e-02 4.16295648e-01 -2.81337559e-01
-4.10169631e-01 -1.37594163e+00 -1.94890369e-02 9.60716188e-01
-1.43162251e-01 -3.08472127e-01 -7.12809145e-01 -7.92795777e-01
-1.95328481e-02 1.25180638e+00 -6.56747222e-01 -4.60609853e-01
-4.08744276e-01 -1.23043716e+00 7.16491401e-01 9.61780250e-01
6.04100466e-01 -1.08510590e+00 -5.17114580e-01 4.03000742e-01
-2.19196156e-01 -1.39820480e+00 -4.39030617e-01 1.15357256e+00
-4.96686697e-01 -9.35680687e-01 -3.65903080e-01 -5.69689214e-01
1.89375669e-01 -2.73482174e-01 1.50153399e+00 -6.93823993e-01
-2.25619644e-01 8.60161707e-02 7.67889321e-02 -6.59613073e-01
-6.18642092e-01 5.05332887e-01 2.81968266e-01 2.70276964e-01
1.04738176e+00 -1.14822745e+00 -4.77936566e-01 -5.50695807e-02
-6.35492861e-01 -2.21624240e-01 6.15610898e-01 3.08619648e-01
2.85636008e-01 3.39821547e-01 7.33275771e-01 -7.75171757e-01
5.78023314e-01 -6.96819365e-01 -7.01682150e-01 1.37785703e-01
-1.18348372e+00 1.33833840e-01 7.32384741e-01 -4.98632938e-01
-1.00242102e+00 3.13139409e-01 -1.68556958e-01 -6.06692731e-01
-6.32850170e-01 -2.66463757e-01 -3.69680732e-01 7.33729243e-01
1.65582106e-01 3.61129761e-01 -5.44969440e-01 -7.23430455e-01
3.71616274e-01 9.22710598e-01 1.04706287e+00 -2.67459214e-01
5.84971666e-01 4.77003366e-01 -1.43281385e-01 -4.48848218e-01
-9.20173466e-01 -4.29604739e-01 -3.86480093e-01 2.29424030e-01
1.16312802e+00 -1.13669145e+00 -1.41491103e+00 7.91238666e-01
-1.09514618e+00 -5.94072163e-01 -4.33523834e-01 1.07114911e-01
-4.61512148e-01 8.58888477e-02 -9.50733483e-01 -1.07129753e+00
-6.34739399e-01 -1.04211855e+00 1.38336599e+00 2.22278982e-01
-3.69729936e-01 -8.96039069e-01 4.08960953e-02 2.92213619e-01
3.97115022e-01 2.14471906e-01 1.15833950e+00 -5.40026605e-01
-2.52254307e-01 -2.91538149e-01 4.93310392e-02 6.70593917e-01
4.56389666e-01 -5.29885828e-01 -1.68128526e+00 -3.46589416e-01
1.00002311e-01 -1.27335846e-01 8.65745664e-01 5.14596105e-01
1.51939416e+00 -4.61891085e-01 -6.01679027e-01 5.64163744e-01
1.40953958e+00 3.13206643e-01 6.38243258e-01 -2.70542383e-01
6.28051341e-01 -2.34960437e-01 -4.24485743e-01 4.48017895e-01
6.22095346e-01 4.70267057e-01 5.78444362e-01 -2.32055217e-01
-1.07272610e-01 -4.85190094e-01 6.91460371e-01 7.81778216e-01
4.38926876e-01 -4.21839356e-01 -3.04288715e-01 6.16170406e-01
-1.83209324e+00 -8.35741639e-01 1.74551889e-01 1.73267078e+00
7.35446095e-01 3.09732795e-01 3.32155585e-01 5.01595974e-01
4.81365263e-01 4.58285779e-01 -1.12325525e+00 -3.06281000e-01
7.63785392e-02 4.20577973e-01 8.40058148e-01 1.25358656e-01
-1.20410872e+00 3.44109535e-01 6.94049788e+00 7.46572554e-01
-9.92160261e-01 3.77781808e-01 5.75533986e-01 -2.09951773e-01
-2.84583047e-02 -6.18593037e-01 -7.40873277e-01 8.75169396e-01
1.61512613e+00 3.72124910e-01 7.94496298e-01 8.06127071e-01
-3.35240550e-02 -1.10467881e-01 -1.41926682e+00 1.25864053e+00
-6.60689250e-02 -9.30708349e-01 -3.81606072e-01 1.59208775e-01
6.67442441e-01 4.53211665e-01 2.98566953e-03 4.56376553e-01
8.65004957e-01 -9.25414979e-01 5.69614232e-01 4.74061787e-01
7.49815524e-01 -8.22395504e-01 5.66176832e-01 4.26130325e-01
-1.31383085e+00 -4.03909504e-01 2.34626174e-01 2.68372837e-02
3.43021721e-01 8.66482735e-01 -5.81176519e-01 5.62338352e-01
8.97408187e-01 6.69933736e-01 -2.95832247e-01 1.69149876e-01
-2.69823343e-01 1.11567867e+00 -3.75241578e-01 2.03558221e-01
-4.18573655e-02 8.12824517e-02 1.95433483e-01 1.23255181e+00
-1.00901373e-01 -3.59206200e-01 9.37498137e-02 1.17155313e+00
-3.67731243e-01 -8.26236665e-01 -3.91353577e-01 -4.97634336e-02
4.02046382e-01 1.33364558e+00 -3.47684652e-01 -3.66602182e-01
-6.42809808e-01 1.46650481e+00 3.72922234e-02 4.36205626e-01
-1.07698238e+00 -2.67571509e-01 1.04439032e+00 -3.73421967e-01
4.22272414e-01 2.02966318e-01 -2.24082828e-01 -1.07652974e+00
-6.36855513e-02 -5.90086579e-01 2.40470245e-01 -5.73221147e-01
-1.58892941e+00 4.39007461e-01 -1.83205396e-01 -8.67608309e-01
-5.43896914e-01 -3.95903111e-01 -7.11554468e-01 9.24301207e-01
-1.76457322e+00 -1.22981131e+00 -6.75543696e-02 7.43000627e-01
5.33308804e-01 -1.62566692e-01 1.20440412e+00 5.19035816e-01
-1.12594903e+00 7.05610752e-01 2.14101255e-01 4.61479604e-01
1.08477131e-01 -1.57213879e+00 9.15675640e-01 7.12306142e-01
3.29573840e-01 2.09227845e-01 4.12182122e-01 -2.11798489e-01
-1.26112068e+00 -1.54177642e+00 4.88620788e-01 -7.33220756e-01
5.45273304e-01 -7.33708620e-01 -4.80532497e-01 1.01784515e+00
6.82294130e-01 -6.01178519e-02 8.95349205e-01 1.23978861e-01
-3.91803056e-01 -3.07685524e-01 -1.14823127e+00 1.45814702e-01
7.03506291e-01 -9.15378153e-01 -3.88664067e-01 1.82199597e-01
6.48041070e-01 -1.55622602e-01 -9.03006136e-01 3.74659300e-01
4.10303205e-01 -7.71249175e-01 7.33730376e-01 -5.03083825e-01
-1.09617189e-01 -2.01899841e-01 -4.77173835e-01 -1.37885690e+00
-6.35028958e-01 -9.87715423e-01 -1.20253932e+00 1.62889576e+00
3.86894256e-01 -6.46477222e-01 6.46923125e-01 5.62772214e-01
-3.28516923e-02 -6.55043423e-01 -8.89857113e-01 -8.58324051e-01
-1.15758829e-01 -9.38811421e-01 1.22564638e+00 6.49328470e-01
3.76070254e-02 8.00033092e-01 -5.79367816e-01 5.44111133e-01
6.19728148e-01 1.86782196e-01 2.34944448e-01 -1.12849402e+00
-6.13334537e-01 -4.06509668e-01 3.40293869e-02 -1.19578505e+00
8.27537775e-01 -8.57997179e-01 2.21234888e-01 -1.27453732e+00
2.17660576e-01 7.18915313e-02 -9.38670039e-01 9.39935148e-01
1.82354286e-01 4.11750019e-01 8.35301280e-02 -2.47269198e-01
-7.00062096e-01 3.99245888e-01 3.28074723e-01 -6.65505528e-01
-1.26856700e-01 4.18482244e-01 -9.42689955e-01 6.90504849e-01
9.34014082e-01 -3.42304021e-01 -4.79427457e-01 -3.03731471e-01
-2.71618664e-01 -3.11153680e-02 4.39488083e-01 -1.13702023e+00
-6.52359426e-02 1.80543125e-01 5.35285890e-01 -8.04325938e-01
5.17584145e-01 -9.78108227e-01 3.38007696e-03 3.15402925e-01
-2.81811506e-01 -2.19233721e-01 2.35440329e-01 7.59084940e-01
2.12999284e-01 2.41170511e-01 5.49857318e-01 7.96108842e-02
-5.61779618e-01 3.51805151e-01 -4.65965152e-01 -4.01512325e-01
7.22086132e-01 3.53853166e-01 -2.00133070e-01 -4.31950539e-01
-6.56438053e-01 2.53901690e-01 -3.61633271e-01 8.37619245e-01
-7.43604973e-02 -1.52614045e+00 -1.63541824e-01 4.86101151e-01
-1.65555859e-03 -1.58560649e-01 1.85738519e-01 5.87599933e-01
6.29345357e-01 7.88852632e-01 3.69835466e-01 -5.46498358e-01
-8.01013231e-01 7.63088942e-01 4.28702593e-01 -5.03240824e-01
-4.40081000e-01 4.39747661e-01 -9.55408290e-02 -1.10025510e-01
5.21550953e-01 -8.64580154e-01 1.52306765e-01 -2.85730660e-02
5.80232441e-01 6.45665228e-01 3.98282021e-01 -5.36802649e-01
-5.45865059e-01 1.28540128e-01 1.18794739e-01 3.70992720e-01
1.50675035e+00 -1.49411201e-01 4.85668033e-02 7.01764762e-01
1.51636565e+00 -5.39590478e-01 -1.31683755e+00 -2.02905759e-01
1.75449047e-02 3.77995074e-01 5.47679305e-01 -1.29557073e+00
-1.43329132e+00 8.17091346e-01 1.11099839e+00 6.33132577e-01
1.49813056e+00 1.34640411e-01 1.35349035e+00 1.20692343e-01
3.93162131e-01 -1.11808145e+00 -3.99394512e-01 1.34811029e-01
1.21858984e-01 -1.22217071e+00 -4.67606783e-01 2.10884824e-01
-2.30475888e-01 6.51054800e-01 4.53104883e-01 8.07001814e-02
9.83406723e-01 9.09205675e-01 -2.43534595e-02 -1.85306236e-01
-7.13548303e-01 -2.29552954e-01 1.33594662e-01 7.44154990e-01
5.14606312e-02 6.15767121e-01 5.58098316e-01 1.08680129e+00
-1.63592759e-03 2.91410148e-01 5.55138998e-02 6.66099966e-01
3.33643556e-01 -8.85742486e-01 -7.58170485e-02 8.12506557e-01
-7.23131776e-01 1.08082496e-01 5.85573502e-02 2.25240588e-01
2.08939120e-01 1.37753224e+00 2.98258185e-01 -6.17785454e-01
4.23855782e-01 4.23200697e-01 2.20854580e-01 -3.78794312e-01
-4.94340271e-01 -1.02946404e-02 -1.31685734e-01 -8.36446881e-01
-4.66537923e-01 -4.96178448e-01 -9.66384351e-01 -3.05184454e-01
-4.77534026e-01 -1.82985857e-01 5.92582762e-01 1.18378830e+00
4.56873536e-01 1.24223483e+00 9.23927248e-01 -8.86222482e-01
-7.13333666e-01 -1.24564850e+00 -9.51823413e-01 4.01071638e-01
9.65708792e-01 -4.12804604e-01 -1.71530023e-01 1.98894762e-03]
|
[16.072616577148438, 7.586677551269531]
|
83783b47-40ee-4f0c-a161-b12c0239b2d7
|
distributional-reinforcement-learning-for-3
|
2011.0184
| null |
https://arxiv.org/abs/2011.01840v1
|
https://arxiv.org/pdf/2011.01840v1.pdf
|
Distributional Reinforcement Learning for mmWave Communications with Intelligent Reflectors on a UAV
|
In this paper, a novel communication framework that uses an unmanned aerial vehicle (UAV)-carried intelligent reflector (IR) is proposed to enhance multi-user downlink transmissions over millimeter wave (mmWave) frequencies. In order to maximize the downlink sum-rate, the optimal precoding matrix (at the base station) and reflection coefficient (at the IR) are jointly derived. Next, to address the uncertainty of mmWave channels and maintain line-of-sight links in a real-time manner, a distributional reinforcement learning approach, based on quantile regression optimization, is proposed to learn the propagation environment of mmWave communications, and, then, optimize the location of the UAV-IR so as to maximize the long-term downlink communication capacity. Simulation results show that the proposed learning-based deployment of the UAV-IR yields a significant advantage, compared to a non-learning UAV-IR, a static IR, and a direct transmission schemes, in terms of the average data rate and the achievable line-of-sight probability of downlink mmWave communications.
|
['Mehdi Bennis', 'Walid Saad', 'Qianqian Zhang']
|
2020-11-03
| null | null | null | null |
['distributional-reinforcement-learning']
|
['methodology']
|
[-3.05417508e-01 3.76198798e-01 3.16429943e-01 4.57283296e-02
-7.98372626e-01 -4.77679908e-01 -1.05213761e-01 -2.51909733e-01
-7.66755268e-02 1.07404280e+00 -1.52200133e-01 -5.54238617e-01
-8.10550213e-01 -1.23142660e+00 -5.61655879e-01 -1.34399843e+00
-4.97593224e-01 -1.12434857e-01 -8.08343410e-01 -1.59173906e-01
-2.08306104e-01 5.93126357e-01 -8.43358099e-01 -5.77370226e-01
8.40817690e-01 1.12424505e+00 2.82181650e-01 6.24229014e-01
7.49021173e-01 1.92797244e-01 -5.66671193e-01 -3.13163996e-01
4.66743231e-01 -2.43978109e-02 1.88102033e-02 6.60450980e-02
5.86675704e-02 -5.46178401e-01 -6.81450963e-01 7.57018983e-01
8.85280967e-01 -5.34505993e-02 8.72552216e-01 -1.09922361e+00
-1.06723323e-01 5.07081747e-01 -5.70572019e-01 -2.49121860e-01
1.62861526e-01 -2.48199090e-01 8.39921176e-01 -3.67015123e-01
1.47379234e-01 6.50509834e-01 6.09303892e-01 -5.21466397e-02
-4.56542581e-01 -7.72017658e-01 8.98863971e-02 -3.36859494e-01
-1.52925396e+00 -3.29310708e-02 3.19808811e-01 -2.47540995e-01
3.65104437e-01 1.10050254e-01 8.38783205e-01 5.18249452e-01
1.10833097e+00 2.65955687e-01 4.84041423e-01 -6.36249483e-01
4.00332421e-01 -1.91109762e-01 -3.93303722e-01 8.95201802e-01
7.69294977e-01 3.92343491e-01 -1.56296059e-01 -3.79123926e-01
7.21831262e-01 -1.50797501e-01 -5.74110150e-01 -3.14111918e-01
-6.77329242e-01 6.58954859e-01 4.12254870e-01 2.16437072e-01
-9.91639197e-01 4.43397343e-01 -6.69097781e-01 5.29810190e-01
4.69923556e-01 6.07878804e-01 -4.34424907e-01 2.52923995e-01
-8.79368782e-01 2.25876689e-01 9.24271524e-01 1.28373587e+00
6.16302907e-01 2.74220377e-01 -3.99855524e-01 3.21079373e-01
1.04703391e+00 1.31832707e+00 -3.59823316e-01 -5.45690358e-01
6.68032050e-01 -6.55458122e-02 5.54819763e-01 -8.67748201e-01
-7.99491286e-01 -1.70125401e+00 -9.24321771e-01 -8.12737718e-02
-1.84339080e-02 -1.57874668e+00 -7.13751972e-01 1.52828360e+00
2.74037629e-01 2.85695523e-01 7.14358509e-01 7.00902820e-01
1.93368375e-01 8.74258637e-01 -3.83418143e-01 -7.72433281e-01
5.90676904e-01 -3.16956639e-01 -5.27832150e-01 -1.78358987e-01
7.72682428e-01 -4.51686949e-01 3.51769254e-02 5.03214657e-01
-8.88757527e-01 3.36596370e-02 -1.35494900e+00 1.19571614e+00
1.63557619e-01 4.08825696e-01 3.22645068e-01 1.16289735e+00
-8.43302727e-01 1.57552380e-02 -5.32961369e-01 -8.37843940e-02
1.94473237e-01 3.65106285e-01 3.86918753e-01 -4.31385756e-01
-1.28540838e+00 5.34582138e-01 -1.23244211e-01 2.36127034e-01
-7.95756340e-01 -9.40255821e-01 -6.77769840e-01 -6.09075576e-02
-1.02503136e-01 -1.17809331e+00 7.97528625e-01 -5.00778496e-01
-1.64710808e+00 -1.77009016e-01 2.29434907e-01 -5.58704734e-01
2.44849160e-01 -2.57163912e-01 -2.67395794e-01 2.60780573e-01
-3.32419187e-01 -2.08089784e-01 7.78166294e-01 -1.53074324e+00
-1.14078557e+00 -5.26349545e-01 2.46637568e-01 3.45111251e-01
-3.49305689e-01 -8.23958993e-01 1.90653563e-01 -6.45436883e-01
2.70921230e-01 -1.28797102e+00 -4.67321068e-01 -5.69774628e-01
-5.26719749e-01 3.06332648e-01 5.48985660e-01 -5.55100083e-01
1.24087965e+00 -1.94784153e+00 4.19452816e-01 7.34615862e-01
-3.89227480e-01 -1.05175801e-01 -6.79962486e-02 7.97684789e-01
4.89453614e-01 -4.27241951e-01 -1.26733392e-01 -6.04646988e-02
-2.89668083e-01 -2.66362846e-01 -2.72686064e-01 7.79183030e-01
-5.99655867e-01 3.12973946e-01 -7.39546299e-01 3.95464808e-01
8.63743424e-02 3.62418681e-01 -4.55829114e-01 4.35307890e-01
-7.27353916e-02 5.05513668e-01 -8.82771850e-01 5.30513406e-01
1.03150451e+00 4.03896600e-01 5.42594828e-02 -2.03615382e-01
-4.45845395e-01 -8.16320539e-01 -9.61881578e-01 1.48627901e+00
-1.10315633e+00 1.39994621e-01 3.45359623e-01 -7.44976521e-01
1.11825490e+00 3.88359606e-01 1.00784588e+00 -3.98788810e-01
2.15226457e-01 9.27007124e-02 -3.29507113e-01 -6.00998521e-01
2.14342982e-01 -7.23949820e-02 -5.59765846e-02 4.80568051e-01
7.55815282e-02 -3.12051177e-01 -3.63277048e-01 9.19494629e-02
1.08820534e+00 -1.28429517e-01 1.69939727e-01 -4.96316463e-01
4.34325188e-01 -4.85831261e-01 3.76983017e-01 6.32291794e-01
4.18065280e-01 8.05046633e-02 -2.04783589e-01 2.85585709e-02
-4.06265706e-01 -1.07052886e+00 -6.17470406e-02 4.67872083e-01
6.24999344e-01 9.03895348e-02 -6.49010599e-01 -4.50043101e-03
2.71536171e-01 1.21537805e+00 6.90591708e-03 -3.09490472e-01
2.70137280e-01 -1.10513210e+00 2.32532561e-01 -3.38503152e-01
5.60539544e-01 1.41279355e-01 -6.80242717e-01 3.90766650e-01
-2.19166294e-01 -9.80262339e-01 3.02424997e-01 7.55541772e-02
-6.57702208e-01 -7.82365441e-01 -9.84254360e-01 -4.46604609e-01
6.84230566e-01 5.68450272e-01 3.61629069e-01 -2.26809263e-01
-1.56830221e-01 1.26207781e+00 -5.52710772e-01 -8.16812754e-01
3.32174331e-01 9.02786106e-02 1.65508926e-01 3.05378348e-01
-3.72909904e-01 -7.08422482e-01 -7.00431585e-01 2.06210598e-01
-2.49934927e-01 -3.04143727e-01 9.24250841e-01 5.16364276e-01
3.59786868e-01 6.08945489e-01 9.23872530e-01 -4.13816094e-01
4.31322187e-01 -9.47391510e-01 -8.72213840e-01 5.23199975e-01
-5.27776718e-01 -1.30965471e-01 4.69040066e-01 2.94679046e-01
-1.15543807e+00 1.02094449e-01 -4.77490388e-02 1.57927543e-01
1.23006172e-01 9.21476305e-01 -3.72009844e-01 -7.07018375e-01
3.45238864e-01 1.91767216e-01 -3.42647165e-01 3.52335930e-01
2.87356853e-01 1.10153198e+00 -1.07612506e-01 -4.94591802e-01
1.19603610e+00 3.73030484e-01 5.02961695e-01 -1.57378983e+00
-9.63224173e-01 -4.33966905e-01 -5.46056867e-01 -6.28274322e-01
4.35200870e-01 -1.47507083e+00 -3.46833915e-01 3.74137670e-01
-9.82257485e-01 -2.55831540e-01 3.68460357e-01 1.07002878e+00
-8.59643281e-01 -1.43149961e-02 -3.54478089e-03 -1.50875020e+00
-4.50359881e-01 -5.93688250e-01 7.22429276e-01 2.91716814e-01
3.33403558e-01 -8.91923189e-01 9.85804014e-03 6.02317974e-03
5.05398512e-01 6.12736344e-01 7.85864413e-01 4.43762571e-01
-9.77766573e-01 -3.67410451e-01 1.08061962e-01 -1.52546301e-01
4.96629775e-02 -4.93886113e-01 -6.58183992e-01 -9.09412444e-01
-1.26144886e-01 1.05109945e-01 3.66254032e-01 1.19354689e+00
6.33476019e-01 -3.47446233e-01 -5.36214828e-01 1.05571568e+00
1.94023156e+00 3.05572003e-01 6.34071410e-01 2.71123022e-01
-2.45165899e-02 1.60478204e-01 1.11277735e+00 1.42039227e+00
3.88161033e-01 2.49456331e-01 1.10062408e+00 1.05164409e-01
6.41912818e-01 2.54085749e-01 2.08792835e-01 1.80821359e-01
-2.48136759e-01 -9.69499469e-01 -2.82742351e-01 8.32296535e-02
-1.67970085e+00 -6.17099822e-01 9.68546346e-02 2.56427503e+00
-1.47431895e-01 -4.66803879e-01 -1.55503169e-01 -2.57232517e-01
3.73123109e-01 -2.47052163e-02 -2.64000803e-01 2.45028049e-01
1.17704729e-02 -3.19672257e-01 1.30931497e+00 7.28591561e-01
-9.85741079e-01 2.84397393e-01 5.23055124e+00 3.44323725e-01
-1.07134199e+00 -2.28176922e-01 2.02725485e-01 2.75968388e-02
-6.87590122e-01 -2.75721312e-01 -7.08764374e-01 1.09747685e-01
7.93409705e-01 -2.88293064e-01 4.32357997e-01 3.33862245e-01
5.16789258e-01 -2.45646849e-01 -5.08782208e-01 1.03064418e+00
2.17278060e-02 -9.67282057e-01 3.55381332e-02 4.01372641e-01
7.21285105e-01 -2.01370090e-01 2.27712199e-01 4.41956371e-02
1.46217242e-01 -3.79244775e-01 5.62559664e-01 1.24044919e+00
4.90583062e-01 -1.46235693e+00 9.53087449e-01 5.44721723e-01
-9.57610369e-01 -6.76661074e-01 -5.38947344e-01 2.77630575e-02
2.91153997e-01 1.32698810e+00 -4.78291512e-01 1.45547140e+00
3.04059625e-01 3.82918715e-01 3.13046455e-01 1.55573916e+00
-2.71425754e-01 5.91958463e-01 -2.09594652e-01 -4.01049674e-01
3.40249419e-01 -4.55172628e-01 1.01102853e+00 9.43631768e-01
1.27271950e+00 5.34377635e-01 2.33752161e-01 1.64035648e-01
1.91022098e-01 1.24475151e-01 -7.72554338e-01 3.38767022e-01
6.46050572e-01 1.35178328e+00 2.66141002e-03 5.10336399e-01
-3.04769695e-01 6.15766644e-01 -3.39066923e-01 1.00550818e+00
-5.97532928e-01 -7.18220711e-01 9.07299161e-01 1.51948212e-03
2.54679739e-01 -8.49893391e-01 -1.32059827e-01 -5.10986388e-01
-2.06536159e-01 -1.05355017e-01 1.08791173e-01 -4.90687668e-01
-8.41357410e-01 2.57961869e-01 -3.41146350e-01 -1.57333875e+00
-2.21609712e-01 -3.48811656e-01 -4.02106494e-01 8.37267935e-01
-1.74739635e+00 -1.41446388e+00 -3.61954927e-01 3.63561839e-01
-9.21119526e-02 -6.03293777e-01 1.03210914e+00 1.04865097e-01
-4.21267986e-01 5.51588476e-01 7.71540463e-01 -2.89495528e-01
3.86316240e-01 -7.87986636e-01 -7.09074557e-01 7.54778504e-01
-6.92020357e-02 3.19250286e-01 9.49742317e-01 -5.93609273e-01
-2.09342074e+00 -1.58772850e+00 -4.79158871e-02 3.90046507e-01
3.10171366e-01 3.01641766e-02 3.48406792e-01 4.56887722e-01
3.45177650e-01 -5.15104771e-01 1.05354035e+00 2.22583339e-01
3.75257134e-01 -3.89049649e-01 -1.30080867e+00 5.87834477e-01
6.27280891e-01 3.01956177e-01 1.96954340e-01 6.49711728e-01
4.69333798e-01 -3.49103987e-01 -9.29682016e-01 4.29574579e-01
8.12690496e-01 -5.22409618e-01 9.72373486e-01 -8.57429877e-02
-1.71951465e-02 -2.72328150e-03 -5.81463277e-01 -2.12822390e+00
-6.28090382e-01 -8.58007073e-01 -1.13002136e-01 9.99897540e-01
8.11489820e-01 -6.76776409e-01 8.32964897e-01 -9.86415893e-02
-2.39288807e-01 -9.51939940e-01 -9.43541110e-01 -8.31410170e-01
-2.19326206e-02 -2.44755834e-01 2.44335547e-01 9.77514610e-02
-6.27209917e-02 1.94599032e-01 -5.65997124e-01 1.47550404e+00
1.13649476e+00 1.05185337e-01 7.55486488e-01 -1.23402011e+00
-3.44397455e-01 4.24804002e-01 -2.91081905e-01 -1.15742195e+00
8.68665650e-02 -6.50638878e-01 2.89269745e-01 -2.01810789e+00
-5.91963470e-01 -8.35161388e-01 -3.04209232e-01 -1.76269278e-01
3.43457907e-01 -2.66288042e-01 -2.38627017e-01 -2.80207336e-01
-2.01803699e-01 8.15921664e-01 1.20938635e+00 -3.18300426e-01
-3.79596174e-01 1.20909929e+00 -7.02285826e-01 4.83362108e-01
6.61472321e-01 -2.62269974e-01 -6.36970282e-01 -5.76412737e-01
2.68507987e-01 7.40698338e-01 -1.29624411e-01 -1.55838180e+00
3.41864049e-01 -2.92784601e-01 3.98991466e-01 -6.29619598e-01
3.57101649e-01 -1.14466250e+00 -1.05560683e-02 8.44464481e-01
2.17316955e-01 -5.65808058e-01 4.33520973e-03 1.21263230e+00
1.88381255e-01 -4.34058577e-01 8.80089879e-01 2.50857502e-01
-1.85385332e-01 7.05866516e-01 -9.81316984e-01 -4.55387920e-01
1.56146526e+00 2.22871616e-01 8.69009867e-02 -1.11520147e+00
-3.46924216e-01 4.57673222e-01 -2.64723361e-01 -1.13930367e-01
6.95766211e-01 -8.61869037e-01 -9.16031778e-01 5.05994558e-02
2.64562629e-02 -3.98648173e-01 3.73637646e-01 5.05684376e-01
-3.38187456e-01 5.54844141e-01 -1.01554997e-01 -2.97028184e-01
-9.16908562e-01 -2.41713300e-01 6.57602489e-01 2.83313449e-02
-2.93188363e-01 9.82993066e-01 -2.16811195e-01 -1.77333668e-01
2.76690602e-01 8.38827640e-02 -9.99292508e-02 -3.09437197e-02
4.20395195e-01 2.73468137e-01 2.11411625e-01 -1.63770840e-01
-3.67317870e-02 8.68109822e-01 4.52503294e-01 -2.75241524e-01
1.47240233e+00 -7.16343939e-01 -1.74143706e-02 -1.15080634e-02
4.19654310e-01 4.82239276e-01 -1.16641474e+00 -4.63363752e-02
-4.58957374e-01 -5.19424021e-01 6.02533698e-01 -7.72620738e-01
-1.07246161e+00 3.59660238e-01 7.13285506e-01 1.45672992e-01
1.03062761e+00 -5.72405577e-01 5.78064263e-01 9.80612814e-01
1.36203778e+00 -1.01365113e+00 -1.18856438e-01 4.92047608e-01
8.51418138e-01 -6.41915381e-01 2.59951025e-01 -4.81802702e-01
-1.89835262e-02 1.52326024e+00 3.49774867e-01 -1.16561972e-01
1.25457942e+00 6.23654246e-01 1.23063453e-01 7.30027929e-02
-3.32083285e-01 -2.39191309e-01 -1.35898679e-01 8.42556894e-01
3.89247417e-01 2.88366884e-01 -2.10038573e-01 5.49946368e-01
-2.83734381e-01 -3.09097797e-01 8.16313446e-01 8.51872981e-01
-1.04643977e+00 -7.75044084e-01 -3.67103726e-01 7.90872097e-01
-2.36638159e-01 1.06073700e-01 3.09304237e-01 4.18293953e-01
-4.10884954e-02 1.47729075e+00 -2.66339213e-01 -4.78124857e-01
4.13566738e-01 -6.23958051e-01 7.22001910e-01 -4.74284589e-01
3.36248219e-01 -1.66174814e-01 6.16020113e-02 -1.63930785e-02
-1.82017416e-01 -4.45561737e-01 -9.89340544e-01 2.05203667e-01
-6.41018867e-01 5.68780005e-01 9.68074203e-01 1.25565183e+00
4.35031563e-01 7.63716996e-01 1.39411163e+00 -5.40049076e-01
-6.13903761e-01 -6.50584102e-01 -1.26001489e+00 -8.61426473e-01
3.69036347e-01 -7.29682326e-01 -4.02297646e-01 -6.23297572e-01]
|
[6.141955852508545, 1.369336485862732]
|
c4d1d96d-bbb7-40ae-80de-e3ec0a2fa5a3
|
ustep-structuration-des-logs-en-flux-gr-a-ce
|
2304.12331
| null |
https://arxiv.org/abs/2304.12331v1
|
https://arxiv.org/pdf/2304.12331v1.pdf
|
USTEP: Structuration des logs en flux gr{â}ce {à} un arbre de recherche {é}volutif
|
Logs record valuable system information at runtime. They are widely used by data-driven approaches for development and monitoring purposes. Parsing log messages to structure their format is a classic preliminary step for log-mining tasks. As they appear upstream, parsing operations can become a processing time bottleneck for downstream applications. The quality of parsing also has a direct influence on their efficiency. Here, we propose USTEP, an online log parsing method based on an evolving tree structure. Evaluation results on a wide panel of datasets coming from different real-world systems demonstrate USTEP superiority in terms of both effectiveness and robustness when compared to other online methods.
|
['Mar Callau-Zori', 'Raja Chiky', 'Arthur Vervaet']
|
2023-04-24
| null | null | null | null |
['log-parsing']
|
['computer-code']
|
[-1.56262413e-01 -4.58144784e-01 -5.50116599e-01 -4.59895045e-01
-6.53521180e-01 -6.17922664e-01 3.90498757e-01 8.50083530e-01
-7.19385520e-02 3.86733651e-01 1.64650269e-02 -7.00723469e-01
1.71092391e-01 -6.99711859e-01 -3.93255472e-01 6.74526766e-02
-6.30889416e-01 4.48919892e-01 8.71001363e-01 -5.65549545e-02
3.64579260e-01 4.09307182e-01 -1.49226451e+00 3.97966981e-01
5.52958667e-01 1.00182700e+00 -4.07573208e-03 7.20732450e-01
-4.59172070e-01 1.08919275e+00 -7.51843095e-01 -1.60892218e-01
3.06588918e-01 -1.54670939e-01 -8.57824326e-01 1.06577642e-01
-3.47588718e-01 -4.91877496e-01 -1.09160572e-01 6.75929368e-01
2.99893953e-02 -3.50943625e-01 -7.84915239e-02 -1.23634768e+00
2.36661330e-01 1.08870769e+00 -8.37439358e-01 6.60605848e-01
5.68257451e-01 -2.30396420e-01 9.86561835e-01 -6.33198440e-01
7.87408173e-01 9.58156943e-01 3.08573008e-01 -2.88722426e-01
-1.05614364e+00 -5.86504281e-01 1.27816007e-01 2.65729964e-01
-1.02582431e+00 -6.79928482e-01 4.64026272e-01 -4.07771140e-01
1.28030169e+00 3.50793928e-01 3.12142074e-01 3.34153444e-01
5.16543925e-01 5.14623582e-01 8.79647791e-01 -5.85829377e-01
4.28972065e-01 1.63193733e-01 6.95517182e-01 4.35246497e-01
4.69235390e-01 -2.56017536e-01 -8.93346548e-01 -5.37344098e-01
2.50236183e-01 -3.46154496e-02 -2.03290861e-02 -6.89912215e-02
-5.14151990e-01 5.87927878e-01 -3.05683225e-01 4.98160303e-01
-2.28044450e-01 9.95549038e-02 6.04267836e-01 8.35926950e-01
3.94943297e-01 7.51489326e-02 -5.09961903e-01 -9.24956739e-01
-8.46526265e-01 -3.36324990e-01 1.37200069e+00 1.23326027e+00
6.45774305e-01 -2.24102944e-01 -2.26586815e-02 5.85874438e-01
4.40011799e-01 -3.40978126e-03 5.06734669e-01 -4.61128592e-01
6.24432147e-01 1.07352030e+00 -1.54687697e-02 -5.08796036e-01
-1.61531568e-01 -1.11710317e-01 -3.76351535e-01 2.23328383e-03
4.09442306e-01 3.01396608e-01 -4.84136254e-01 1.16727889e+00
4.00702000e-01 -2.06546053e-01 -5.49097240e-01 1.58443376e-01
1.65842012e-01 5.87997556e-01 -1.74105912e-01 -7.43350804e-01
1.48989904e+00 -5.61700284e-01 -7.35270917e-01 -3.05827349e-01
6.59457088e-01 -1.04341018e+00 1.08855474e+00 8.20014358e-01
-8.89419198e-01 -1.18110292e-01 -1.17914641e+00 3.64549041e-01
-3.49100977e-02 1.03044122e-01 6.98416352e-01 8.01388025e-01
-9.23895478e-01 8.84148002e-01 -1.47068381e+00 -4.11968976e-01
1.49198279e-01 2.91518927e-01 -1.73001081e-01 2.77152181e-01
-4.67969388e-01 3.51687431e-01 5.71534455e-01 -1.04767986e-01
-5.51173508e-01 -3.76848131e-01 -4.77495611e-01 2.84897745e-01
8.44713926e-01 9.06777233e-02 1.85764039e+00 2.22725607e-02
-1.41059422e+00 3.05396795e-01 -2.46638134e-01 -3.57161403e-01
4.68515605e-01 -4.64140058e-01 -6.10716224e-01 -1.10512279e-01
6.46178750e-03 -6.45353794e-01 7.25346029e-01 -7.03605235e-01
-8.90590370e-01 -2.71505386e-01 -7.60306492e-02 -3.84786457e-01
-5.59575498e-01 6.19408667e-01 -7.77962744e-01 5.88775985e-03
9.40806940e-02 -6.35543823e-01 -1.23259529e-01 -6.75036073e-01
-3.13930303e-01 -3.80272448e-01 1.00206470e+00 -6.95704997e-01
2.28700566e+00 -1.95898914e+00 -6.22973442e-01 3.37514699e-01
3.47810924e-01 6.38616085e-02 2.68737733e-01 1.10290265e+00
-6.41891733e-02 4.40669626e-01 2.18083143e-01 -1.57843620e-01
-1.14619188e-01 -1.57904141e-02 -2.38422841e-01 2.81737626e-01
1.62751347e-01 5.87056458e-01 -7.01586723e-01 -8.89175296e-01
4.35109995e-02 -2.23086774e-01 -1.91279143e-01 4.60403293e-01
-2.03413114e-01 -2.80487370e-02 -5.54665029e-01 8.04481208e-01
3.65665317e-01 -5.85387468e-01 5.25851011e-01 3.50125939e-01
-4.27556306e-01 5.33589542e-01 -1.04417765e+00 1.17563486e+00
-7.29708314e-01 5.68066061e-01 -8.51600170e-02 -6.17668867e-01
7.82723844e-01 2.50138193e-01 3.97258669e-01 -8.46503079e-01
3.13279539e-01 9.36693326e-02 -7.47723132e-02 -3.96001667e-01
5.91113865e-01 3.55242789e-01 -3.48569989e-01 1.13928604e+00
-4.70513999e-01 1.78788468e-01 8.91474545e-01 4.43218738e-01
1.92298532e+00 -2.35956386e-01 8.53702903e-01 -4.65094410e-02
3.67986977e-01 1.80710018e-01 7.03591347e-01 4.27529514e-01
-1.62377313e-01 7.45270103e-02 1.05859351e+00 -2.28075653e-01
-7.85166025e-01 -5.91534615e-01 -8.42412189e-02 1.31593382e+00
-1.17137380e-01 -1.12439907e+00 -6.39727116e-01 -7.73506939e-01
-2.01347530e-01 6.70328677e-01 -3.50929081e-01 9.64328721e-02
-8.95366371e-01 -6.77083492e-01 7.82155991e-02 6.11451805e-01
4.19694036e-02 -1.07777512e+00 -1.04138303e+00 7.52735317e-01
-4.84976247e-02 -1.39739025e+00 -5.90152383e-01 3.75847459e-01
-1.09345031e+00 -1.44180131e+00 2.94636726e-01 -3.72071385e-01
4.36115652e-01 3.92453164e-01 1.45179927e+00 4.77202594e-01
-4.98655438e-01 1.56415120e-01 -8.49258304e-01 -5.56661367e-01
-9.97811913e-01 -1.20339245e-01 -2.10504517e-01 -1.21065877e-01
3.45119447e-01 -8.46912265e-01 -2.81984985e-01 4.88318205e-01
-9.83442307e-01 -3.14850539e-01 6.28994524e-01 4.83113497e-01
6.00211322e-01 6.14116371e-01 2.48192698e-01 -1.47735476e+00
1.03374624e+00 -5.03344893e-01 -1.13741398e+00 4.63087142e-01
-1.24916482e+00 2.45131359e-01 9.07990456e-01 -2.84082562e-01
-1.02232325e+00 -5.33997230e-02 1.76594287e-01 -5.83178774e-02
6.57358840e-02 6.87730730e-01 -1.27674982e-01 4.01543498e-01
4.76894826e-01 4.82343230e-03 -1.74955532e-01 -7.29189157e-01
-1.57268971e-01 8.24920475e-01 2.39825353e-01 -5.04055262e-01
6.13294840e-01 3.79720360e-01 -3.89306486e-01 -7.91210532e-01
-3.99852186e-01 -9.34683621e-01 -5.18950760e-01 -2.90163811e-02
5.67322150e-02 -2.62992114e-01 -8.58525693e-01 4.16707486e-01
-1.10131490e+00 -2.14366645e-01 -9.96797532e-02 1.31644934e-01
-2.22482860e-01 6.06466651e-01 -8.21360886e-01 -8.52723658e-01
-5.34980059e-01 -9.42305326e-01 8.82643938e-01 1.65767670e-02
-3.67486000e-01 -7.10421383e-01 2.46533439e-01 -3.81001122e-02
4.31345552e-01 2.37198219e-01 1.04904997e+00 -8.74443233e-01
-8.88657808e-01 -7.08255827e-01 -2.06993416e-01 5.85661009e-02
2.22053528e-01 3.88624787e-01 -7.83691406e-01 -4.52088028e-01
8.07799622e-02 3.43320966e-02 1.97309583e-01 -2.69204646e-01
9.21072662e-01 -3.28666151e-01 -4.07587171e-01 2.32011884e-01
1.58711374e+00 5.74706674e-01 3.59144390e-01 3.98982435e-01
2.18070075e-01 3.85474175e-01 1.20567787e+00 1.07887089e+00
-1.23705022e-01 4.19397980e-01 2.61947781e-01 4.70925927e-01
2.02937648e-01 -4.40588057e-01 6.54192686e-01 1.40057993e+00
3.69423091e-01 -2.95657933e-01 -1.16543591e+00 5.70339680e-01
-1.71618021e+00 -6.23459041e-01 -4.44093853e-01 2.49977255e+00
7.44352818e-01 8.25832725e-01 2.41981938e-01 5.76527536e-01
5.59411824e-01 -2.08334461e-01 -4.88346606e-01 -5.17495990e-01
5.56925535e-01 3.21123213e-01 4.67721969e-01 2.63418518e-02
-5.44735193e-01 6.33864522e-01 6.28574562e+00 6.70350850e-01
-8.88516307e-01 2.01929122e-01 1.77062124e-01 2.07448125e-01
1.11714974e-02 6.34971917e-01 -8.77896369e-01 5.51028848e-01
1.59179592e+00 -9.00934637e-01 1.21569395e-01 1.08094442e+00
5.13563037e-01 -6.92734003e-01 -1.64906538e+00 8.85396421e-01
-5.57337582e-01 -1.28047109e+00 -3.17302406e-01 4.23687696e-01
9.19102803e-02 6.96029365e-02 -6.41503811e-01 1.45187214e-01
5.30508995e-01 -6.46734774e-01 7.60124445e-01 -1.45288259e-01
8.01112235e-01 -5.61244249e-01 8.46336186e-01 6.10669136e-01
-1.79402936e+00 7.10623013e-03 -3.93737182e-02 -1.77213997e-01
6.24908388e-01 9.42934990e-01 -1.32875323e+00 4.57689732e-01
1.00563002e+00 3.42434376e-01 -7.19042540e-01 8.74136269e-01
-2.20364735e-01 1.19556189e+00 -7.39728689e-01 -2.35295936e-01
-2.52092123e-01 -1.48911737e-02 2.28224352e-01 1.23549378e+00
2.62841493e-01 -9.02975276e-02 1.64689571e-01 5.80590069e-01
2.12600883e-02 3.74389052e-01 -4.52794492e-01 -5.80443025e-01
1.02567494e+00 1.13870883e+00 -1.25981736e+00 -3.67462218e-01
-5.34804642e-01 3.90123695e-01 -4.95547382e-03 -1.87592000e-01
-7.84182191e-01 -6.18121624e-01 3.79108578e-01 6.62731469e-01
3.08319569e-01 -3.96592617e-01 -1.11758120e-01 -7.45205581e-01
6.03365421e-01 -8.41988206e-01 5.26652515e-01 1.03858724e-01
-8.74371052e-01 8.21473122e-01 6.09265007e-02 -1.56374180e+00
-5.31138301e-01 -2.61366695e-01 -8.22529554e-01 3.41404885e-01
-1.26326668e+00 -4.35673505e-01 -4.94457334e-01 3.77017975e-01
6.43882394e-01 1.25582069e-01 5.47336221e-01 3.38887662e-01
-7.85426140e-01 7.75518477e-01 7.11103305e-02 -1.07749924e-01
6.24680698e-01 -1.20104933e+00 8.29947293e-01 1.17669392e+00
2.68919438e-01 8.52731287e-01 8.07760119e-01 -7.72061229e-01
-1.86861682e+00 -8.12959790e-01 7.98441052e-01 -5.36158621e-01
1.10941350e+00 -4.65351194e-01 -1.08317947e+00 6.46348596e-01
1.21333309e-01 -6.38888627e-02 6.97246253e-01 2.38616899e-01
-2.77561694e-01 -1.90767452e-01 -6.58073366e-01 -5.09402603e-02
9.22850847e-01 -4.14925069e-01 -2.44738474e-01 1.94335893e-01
6.38204038e-01 -3.31067055e-01 -1.12953031e+00 -6.15423173e-02
4.00809765e-01 -1.34491658e+00 1.80798844e-01 -3.85749578e-01
1.29538238e-01 -1.12654023e-01 1.83087111e-01 -7.04676092e-01
1.43174261e-01 -1.49531627e+00 -5.90232372e-01 1.51625025e+00
3.81762892e-01 -7.25189090e-01 9.23702359e-01 4.60705727e-01
4.98631485e-02 -5.91385007e-01 -5.95352948e-01 -1.21869743e+00
-6.91024780e-01 -9.27037656e-01 8.36715043e-01 4.98407990e-01
5.16831279e-01 6.74369037e-01 9.10072867e-03 -3.73191610e-02
4.42277312e-01 4.99362499e-01 1.14167762e+00 -1.38391137e+00
-6.93898559e-01 -1.94267511e-01 -2.70119190e-01 -1.08326447e+00
-4.29147869e-01 -5.74058414e-01 9.60653126e-02 -1.21040428e+00
2.59196430e-01 -5.54948032e-01 -1.75528944e-01 4.77467686e-01
1.71245128e-01 -6.06305242e-01 9.58966836e-02 5.85591555e-01
-7.79999375e-01 3.57774310e-02 5.16974032e-01 3.02757055e-01
-7.31813371e-01 5.23074627e-01 -4.99202996e-01 6.61460459e-01
8.05510581e-01 -9.19001281e-01 -5.88969231e-01 -1.55800790e-01
4.50377285e-01 5.91601491e-01 -5.05120158e-01 -8.44652772e-01
3.38656038e-01 -2.73046464e-01 -4.23346281e-01 -7.98234940e-01
-1.39105886e-01 -8.56388509e-01 3.73857796e-01 5.96342087e-01
3.37303728e-01 6.72668219e-01 1.89784378e-01 7.70563185e-01
-6.29906893e-01 -3.96377265e-01 4.64127332e-01 1.32382885e-01
-8.82880151e-01 3.76137137e-01 -3.01264852e-01 2.46952236e-01
1.19407475e+00 -4.44427222e-01 -3.32160801e-01 -2.00629666e-01
-3.62854004e-01 3.00711364e-01 6.85976446e-01 3.87592435e-01
3.34687412e-01 -7.99323142e-01 -4.04864043e-01 2.42475823e-01
4.93160963e-01 -1.37366638e-01 -4.80692118e-01 9.66072977e-01
-7.90807784e-01 5.37523270e-01 1.09225504e-01 -6.66021645e-01
-1.58266234e+00 4.29091573e-01 -3.42565924e-01 -9.23519671e-01
-7.82449186e-01 5.10663986e-01 -2.39831463e-01 3.34526539e-01
2.92594492e-01 -5.23395956e-01 1.60216652e-02 1.80395842e-01
7.28813708e-01 5.87365925e-01 6.90306067e-01 1.93445191e-01
-4.29420352e-01 4.98781502e-02 -5.80986679e-01 5.66663891e-02
1.51868868e+00 -1.91449806e-01 -3.30081165e-01 7.02088237e-01
9.43635345e-01 3.74371916e-01 -9.28059340e-01 -3.47389340e-01
1.02171350e+00 -7.06501305e-01 -8.57388601e-02 -3.98435861e-01
-8.60607803e-01 7.27445483e-01 3.26515853e-01 8.14176679e-01
1.24729812e+00 1.91097111e-01 9.20516372e-01 2.91635692e-01
1.05540621e+00 -1.25564826e+00 1.23569973e-01 1.87093183e-01
5.31227827e-01 -1.24070585e+00 9.68578830e-02 -8.06959152e-01
-1.72030196e-01 1.25725687e+00 6.17453158e-01 4.15746301e-01
8.19067538e-01 8.38099182e-01 -1.37053907e-01 -1.06339030e-01
-1.24481106e+00 1.48121208e-01 -4.06486690e-01 2.05675676e-01
6.86002135e-01 -3.70663367e-02 -7.47758150e-01 6.54326975e-01
-1.24080203e-01 7.44786337e-02 9.69338059e-01 1.87083733e+00
-6.99991107e-01 -1.81974900e+00 -2.85938770e-01 6.21863902e-01
-5.73691428e-01 1.82616726e-01 -5.00308692e-01 1.02480853e+00
-5.86766541e-01 1.17748165e+00 -2.26382688e-01 -6.36714041e-01
5.27337492e-01 -1.25361174e-01 1.82278425e-01 -1.03373730e+00
-6.37638152e-01 1.03072159e-01 4.49085861e-01 -9.98481035e-01
2.73925722e-01 -8.25742602e-01 -1.36153281e+00 -6.71969771e-01
-6.80592656e-01 5.08304358e-01 7.06299782e-01 5.40832520e-01
4.67317045e-01 5.47518492e-01 9.91009831e-01 -6.57790676e-02
-9.49518204e-01 -8.70753050e-01 -5.86584270e-01 2.90250331e-01
3.54139833e-04 -5.01196504e-01 -1.83782503e-01 3.53686243e-01]
|
[7.976958751678467, 6.853243827819824]
|
a733936b-c089-4672-85a3-655e6c173ebb
|
cheap-and-quick-efficient-vision-language
|
2305.15023
| null |
https://arxiv.org/abs/2305.15023v2
|
https://arxiv.org/pdf/2305.15023v2.pdf
|
Cheap and Quick: Efficient Vision-Language Instruction Tuning for Large Language Models
|
Recently, growing interest has been aroused in extending the multimodal capability of large language models (LLMs), e.g., vision-language (VL) learning, which is regarded as the next milestone of artificial general intelligence. However, existing solutions are prohibitively expensive, which not only need to optimize excessive parameters, but also require another large-scale pre-training before VL instruction tuning. In this paper, we propose a novel and affordable solution for the effective VL adaption of LLMs, called Mixture-of-Modality Adaptation (MMA). Instead of using large neural networks to connect the image encoder and LLM, MMA adopts lightweight modules, i.e., adapters, to bridge the gap between LLMs and VL tasks, which also enables the joint optimization of the image and language models. Meanwhile, MMA is also equipped with a routing algorithm to help LLMs achieve an automatic shift between single- and multi-modal instructions without compromising their ability of natural language understanding. To validate MMA, we apply it to a recent LLM called LLaMA and term this formed large vision-language instructed model as LaVIN. To validate MMA and LaVIN, we conduct extensive experiments under two setups, namely multimodal science question answering and multimodal dialogue. The experimental results not only demonstrate the competitive performance and the superior training efficiency of LaVIN than existing multimodal LLMs, but also confirm its great potential as a general-purpose chatbot. More importantly, the actual expenditure of LaVIN is extremely cheap, e.g., only 1.4 training hours with 3.8M trainable parameters, greatly confirming the effectiveness of MMA. Our project is released at https://luogen1996.github.io/lavin.
|
['Rongrong Ji', 'Xiaoshuai Sun', 'Shengxin Chen', 'Tianhe Ren', 'Yiyi Zhou', 'Gen Luo']
|
2023-05-24
| null | null | null | null |
['chatbot', 'science-question-answering', 'chatbot']
|
['methodology', 'miscellaneous', 'natural-language-processing']
|
[-1.95287526e-01 2.61646807e-02 -7.56647959e-02 -2.76607007e-01
-8.38388383e-01 -5.83187282e-01 4.88547742e-01 -4.42060351e-01
-8.43496382e-01 5.35391688e-01 -1.49875969e-01 -5.37207425e-01
4.03388143e-01 -5.57944238e-01 -1.03378999e+00 -5.49521804e-01
6.27656460e-01 4.03851062e-01 3.49463783e-02 -2.32048735e-01
-5.48877148e-03 1.34837613e-01 -1.16360128e+00 2.26736754e-01
1.14984298e+00 8.30117643e-01 7.98259556e-01 7.07511306e-01
-5.67854702e-01 7.33765304e-01 -4.40137088e-01 -7.25854814e-01
-1.80006668e-01 -4.99154299e-01 -8.89134586e-01 7.94740617e-02
2.73180842e-01 -3.89093101e-01 -4.27169979e-01 8.92480433e-01
5.32765627e-01 1.00648746e-01 2.02278748e-01 -1.35070252e+00
-7.07840323e-01 5.69401920e-01 -4.85556692e-01 -3.42607349e-01
3.27691436e-01 6.55694067e-01 8.60135257e-01 -9.78680491e-01
4.20368135e-01 1.41777873e+00 4.11152810e-01 9.15634274e-01
-8.78298223e-01 -7.36278951e-01 1.16203189e-01 2.88709700e-01
-1.29430938e+00 -6.66823626e-01 6.05499506e-01 -1.69127434e-01
8.28005373e-01 9.82235819e-02 4.18909848e-01 1.15672565e+00
-1.15533680e-01 1.26453066e+00 8.13762426e-01 -3.88156265e-01
-6.20181598e-02 3.96881521e-01 -1.98603734e-01 1.11973763e+00
-1.88794702e-01 -3.24335665e-01 -3.76334786e-01 1.79435492e-01
8.25773478e-01 -4.97838445e-02 -3.54080230e-01 -2.45451853e-01
-1.35171235e+00 8.96034777e-01 4.10667986e-01 2.27585971e-01
-1.25616238e-01 2.53378421e-01 4.50986832e-01 3.30514014e-01
-5.17654456e-02 3.20632547e-01 -3.48695993e-01 -3.18619668e-01
-3.20747435e-01 -2.49991968e-01 5.86846113e-01 1.04308224e+00
9.84149516e-01 -7.32023790e-02 -5.54869100e-02 1.14747274e+00
4.87660468e-01 6.48726523e-01 6.03903949e-01 -1.14403522e+00
7.84704983e-01 8.86280060e-01 -1.43698603e-01 -6.62839174e-01
-3.43358964e-01 -1.52733084e-02 -9.93081868e-01 -1.98291197e-01
3.77631664e-01 -2.59330779e-01 -6.97642803e-01 1.96203864e+00
3.32721680e-01 -1.23502210e-01 3.33213925e-01 8.82751465e-01
1.16070771e+00 8.52224827e-01 1.41679689e-01 -6.44512624e-02
1.26309466e+00 -1.45133710e+00 -5.68469882e-01 -3.44146699e-01
7.55317628e-01 -7.51685619e-01 1.77209198e+00 4.46888320e-02
-1.22518587e+00 -7.89300919e-01 -6.97563469e-01 -2.24249378e-01
-2.14586154e-01 3.82356852e-01 7.87155509e-01 3.20665121e-01
-9.63851869e-01 -1.24836884e-01 -6.15083218e-01 -4.43581730e-01
2.92361915e-01 4.00776178e-01 -4.88101095e-01 -1.84888810e-01
-1.21609545e+00 7.47716129e-01 4.46047604e-01 2.86826193e-01
-9.98176873e-01 -3.07716250e-01 -1.01494896e+00 -1.28752470e-03
4.99521226e-01 -6.97114110e-01 1.35802329e+00 -1.03660178e+00
-1.84577513e+00 8.44715297e-01 -2.11151704e-01 -9.53640342e-02
4.08485949e-01 -6.94913194e-02 -2.53295898e-01 2.52920538e-01
-4.00493145e-02 1.11748147e+00 6.60018265e-01 -1.38916433e+00
-6.01986945e-01 -7.91641846e-02 5.02048433e-01 3.63719523e-01
-6.33778930e-01 -2.09496990e-01 -9.86783445e-01 -1.27015039e-01
-2.32240245e-01 -9.18693662e-01 -1.35780603e-01 -1.32976219e-01
-3.29484195e-01 -4.10139769e-01 6.07730746e-01 -4.87735122e-01
1.06855965e+00 -2.15013647e+00 2.18223527e-01 -1.98188722e-01
1.94202751e-01 4.29957837e-01 -6.68986022e-01 3.91368508e-01
4.29149032e-01 -1.70634466e-03 -3.57503474e-01 -6.19725049e-01
1.29679009e-01 4.32687849e-01 -6.75792769e-02 1.70741349e-01
-2.14031842e-02 1.43385875e+00 -7.93257594e-01 -6.64905965e-01
2.60254025e-01 1.93094641e-01 -4.90419239e-01 6.46136522e-01
-4.40261215e-01 6.84099793e-01 -4.31147665e-01 6.63717985e-01
5.53463995e-01 -5.22547603e-01 1.43380295e-02 -3.41108412e-01
-1.75520793e-01 -5.92330582e-02 -6.82512164e-01 2.15642548e+00
-8.93535495e-01 4.57089603e-01 3.56445074e-01 -9.93644416e-01
7.88322568e-01 2.70286947e-01 2.08049536e-01 -9.13999259e-01
1.44293085e-01 2.26404518e-01 -1.19064920e-01 -9.52159882e-01
3.36174816e-01 9.60176736e-02 -2.66598493e-01 3.11183810e-01
2.80626982e-01 -7.97405094e-02 1.97377548e-01 3.29697251e-01
6.53730810e-01 2.15473980e-01 5.31250760e-02 1.59314901e-01
9.36118245e-01 -1.17005512e-01 3.03571194e-01 6.68728471e-01
-2.15389311e-01 2.16687158e-01 1.77589059e-01 -4.13654819e-02
-6.99256420e-01 -8.67772639e-01 2.30697900e-01 1.27391136e+00
4.37455148e-01 -2.21758887e-01 -9.18990970e-01 -9.26698685e-01
-2.93210059e-01 6.33222520e-01 -1.88459605e-01 -2.46640876e-01
-5.58062434e-01 -6.07804656e-01 7.92758048e-01 2.74060875e-01
1.06835353e+00 -1.34432602e+00 -3.43489289e-01 -7.67231807e-02
-7.07835555e-01 -1.37961948e+00 -7.25315988e-01 -2.89588094e-01
-5.60756385e-01 -8.57914209e-01 -7.16508508e-01 -1.05717695e+00
6.31501079e-01 4.83428657e-01 1.01655948e+00 2.40363866e-01
6.87654540e-02 7.37848938e-01 -3.10850292e-01 -1.80192009e-01
-5.06316721e-01 3.39757949e-01 -1.15329050e-01 1.28212407e-01
2.73840338e-01 -3.42505306e-01 -5.62584341e-01 4.08431709e-01
-9.49143589e-01 3.85727644e-01 1.08191621e+00 9.39317942e-01
4.51030582e-01 -5.15599906e-01 7.60028958e-01 -4.82710063e-01
6.98413551e-01 -3.64525676e-01 -5.65930843e-01 5.86774111e-01
-3.51632535e-01 -5.27564436e-03 7.38278925e-01 -6.40351772e-01
-1.15693617e+00 5.85824996e-02 -5.13224602e-01 -4.01416421e-01
-3.51877928e-01 6.53139830e-01 -4.90235418e-01 -2.76262939e-01
1.01927131e-01 4.95556623e-01 2.26228401e-01 -3.14457059e-01
6.39442325e-01 9.72702801e-01 6.26867950e-01 -5.51233351e-01
6.80674255e-01 2.33650014e-01 -3.88414115e-01 -8.37955713e-01
-7.06817210e-01 -3.62603515e-01 -3.51323277e-01 -2.80991346e-01
9.52429712e-01 -9.37524974e-01 -1.27720416e+00 5.49169064e-01
-1.29201663e+00 -6.78317487e-01 8.18625912e-02 5.66539347e-01
-5.59875250e-01 5.80793738e-01 -7.48177528e-01 -5.73312640e-01
-4.55129325e-01 -1.36377990e+00 8.95761609e-01 6.00138247e-01
1.29794970e-01 -1.10744858e+00 -1.15214668e-01 1.04880309e+00
3.63904178e-01 -4.53298092e-01 8.84038866e-01 -3.79980594e-01
-6.86072409e-01 -2.60084067e-02 -4.13162351e-01 3.53938133e-01
-4.19828445e-02 -3.26168090e-01 -9.86613452e-01 -3.84913713e-01
-2.47628782e-02 -9.59776461e-01 6.63211048e-01 9.61681604e-02
1.12818861e+00 -2.59906769e-01 -1.34007052e-01 7.83978403e-01
1.20531166e+00 2.36013070e-01 6.27554119e-01 4.46589112e-01
9.37397540e-01 5.07473052e-01 6.35982454e-01 1.16575532e-01
9.15227354e-01 5.47525942e-01 5.92046797e-01 -4.25639331e-01
-6.66096807e-02 -4.17860210e-01 7.03660131e-01 1.39868832e+00
6.60927743e-02 -2.90424705e-01 -7.72768319e-01 3.91220987e-01
-1.89948416e+00 -6.18829966e-01 1.19199336e-01 1.98880434e+00
9.77302134e-01 -2.98781157e-01 -1.28288835e-01 -6.73947096e-01
4.83957767e-01 1.16503410e-01 -7.05698371e-01 -3.82369876e-01
-1.28175884e-01 -2.72488922e-01 1.27915725e-01 6.45204127e-01
-8.89864981e-01 1.32650161e+00 4.88757277e+00 1.09473789e+00
-1.03916562e+00 2.39009768e-01 5.29446006e-01 1.01875849e-01
-3.77300322e-01 -1.34792566e-01 -7.52971530e-01 4.49999690e-01
8.50436628e-01 1.83361933e-01 6.43503249e-01 6.18004262e-01
1.55105218e-01 -1.51977390e-02 -8.87302041e-01 1.23411036e+00
2.12238953e-01 -1.14300299e+00 3.02548945e-01 -1.59982190e-01
5.06020606e-01 1.37927890e-01 -5.59991710e-02 7.29464769e-01
1.20606959e-01 -9.43980992e-01 4.57003593e-01 4.64420587e-01
9.11772430e-01 -7.06516683e-01 8.26425910e-01 7.54989743e-01
-1.14513218e+00 -2.59942692e-02 -4.42078233e-01 1.92797109e-01
2.56333143e-01 1.28628844e-02 -4.63295817e-01 4.99503523e-01
5.66246331e-01 3.47204089e-01 -4.34952915e-01 6.60714507e-01
-3.72875541e-01 4.79225188e-01 -1.16254717e-01 -8.81442353e-02
5.13445199e-01 -4.42290723e-01 2.60367036e-01 1.18224895e+00
1.64128408e-01 -4.00360152e-02 3.26454401e-01 8.38197827e-01
-3.65515083e-01 2.21064255e-01 -4.91556466e-01 -1.83632419e-01
3.33139539e-01 1.53506243e+00 -1.22429721e-01 -2.82971978e-01
-8.18192184e-01 1.16848993e+00 5.35689473e-01 4.97666299e-01
-1.17465794e+00 -3.58289808e-01 4.37761545e-01 -5.26509643e-01
7.50053525e-02 -3.40025961e-01 1.06386550e-01 -1.46958947e+00
-4.36524414e-02 -1.14863861e+00 1.65736258e-01 -8.27860773e-01
-1.15978408e+00 6.49037063e-01 -3.62202048e-01 -1.05389524e+00
-5.13027124e-02 -5.24836540e-01 -5.29433012e-01 8.16099048e-01
-1.60053957e+00 -1.57934105e+00 -5.47430634e-01 8.54465425e-01
7.92409360e-01 -2.85961926e-01 7.33614147e-01 3.98349732e-01
-8.34992051e-01 9.73367751e-01 -6.37534857e-02 1.81585953e-01
8.84442806e-01 -7.97138035e-01 9.45850164e-02 5.27814865e-01
1.90194651e-01 5.33665359e-01 2.78213620e-01 -2.10342854e-01
-1.80840528e+00 -9.44295585e-01 7.81005025e-01 -3.91666770e-01
7.19209194e-01 -5.12992263e-01 -9.27935779e-01 6.66533947e-01
5.35493374e-01 -4.38626319e-01 6.04125917e-01 -1.37059197e-01
-3.19252238e-02 -1.66580051e-01 -7.29384243e-01 8.99858952e-01
8.85565042e-01 -6.75342977e-01 -3.54486048e-01 4.71240073e-01
9.63894367e-01 -2.99778253e-01 -8.89246583e-01 4.27499026e-01
4.15452093e-01 -8.46437156e-01 9.70660090e-01 -3.89539510e-01
4.78900760e-01 -1.97803304e-01 -2.01124340e-01 -1.05930448e+00
-4.46634106e-02 -5.60221672e-01 -1.27526894e-01 1.48489428e+00
4.66045499e-01 -7.86228001e-01 5.46401739e-01 4.41601545e-01
-1.91661969e-01 -9.48788166e-01 -7.19506860e-01 -5.48682868e-01
3.55336852e-02 -4.35392320e-01 4.51843888e-01 9.16534483e-01
-1.16690330e-01 8.05495083e-01 -4.24509525e-01 1.23083949e-01
4.05960917e-01 3.02742481e-01 1.13212502e+00 -6.35424078e-01
-4.59446520e-01 -5.40265143e-01 2.34751388e-01 -1.73750782e+00
3.81762028e-01 -8.96678686e-01 1.93753079e-01 -1.53740048e+00
4.27519292e-01 -3.76288056e-01 -2.02162623e-01 6.24346673e-01
-2.96338588e-01 1.38090372e-01 2.39071012e-01 3.59369129e-01
-1.00785267e+00 9.87767577e-01 1.59066045e+00 -3.10960263e-01
-1.94825053e-01 -8.18365291e-02 -5.93450785e-01 8.36355746e-01
9.10613537e-01 6.09226804e-03 -5.64536512e-01 -8.36395919e-01
6.99131284e-03 1.07619248e-01 3.92186195e-01 -6.87975228e-01
2.79191107e-01 -3.21455359e-01 1.03342012e-02 -3.44065219e-01
4.22536373e-01 -6.96242511e-01 -3.49402487e-01 3.48749042e-01
-3.49555522e-01 3.16299982e-02 2.78540164e-01 2.09434167e-01
-4.34340119e-01 -3.89135689e-01 6.17515266e-01 -2.71177590e-01
-1.10062647e+00 4.17235494e-01 -1.91515788e-01 -5.41289672e-02
8.68207514e-01 9.28819329e-02 -5.08236825e-01 -5.72674215e-01
-3.84306699e-01 7.21372843e-01 4.68490630e-01 4.86421198e-01
7.32225835e-01 -1.16444027e+00 -5.30887544e-01 1.65766418e-01
2.17771128e-01 6.31306991e-02 5.79122305e-01 1.06146288e+00
-4.58374858e-01 5.42150676e-01 5.52345887e-02 -6.58787131e-01
-1.19687676e+00 6.29515111e-01 3.62013638e-01 -2.30605051e-01
-3.74358743e-01 8.80742490e-01 5.50235271e-01 -9.54927802e-01
2.83987075e-01 1.22309558e-01 -2.24515259e-01 -1.49752721e-01
4.04790580e-01 1.09931128e-02 -4.83176410e-01 -5.32198548e-01
-1.71063483e-01 5.57871878e-01 3.42898294e-02 -1.10358521e-01
8.71750116e-01 -5.68667352e-01 -3.32721353e-01 5.35182655e-01
1.12199998e+00 -4.22543995e-02 -1.29177248e+00 -2.35089555e-01
-3.35371792e-01 -4.78234887e-02 -2.94217288e-01 -7.70386577e-01
-1.09713411e+00 1.20416582e+00 4.08635616e-01 -1.03723675e-01
1.15454149e+00 2.19101876e-01 1.20228398e+00 8.01327050e-01
3.35193038e-01 -1.00856245e+00 2.95140356e-01 7.30857491e-01
8.58906627e-01 -1.67011189e+00 -4.64022547e-01 1.95042174e-02
-9.18346465e-01 9.02699649e-01 1.08912635e+00 4.09613073e-01
-1.23672867e-02 -2.75923666e-02 4.86792207e-01 8.39526951e-03
-7.13855743e-01 -2.51673758e-01 1.12650618e-01 3.48624110e-01
3.83299530e-01 -7.19595477e-02 -1.45442531e-01 6.14380836e-01
1.12275686e-02 -1.97379142e-02 1.76446795e-01 5.94494462e-01
-3.30740511e-01 -1.03678882e+00 -2.03878105e-01 5.25702126e-02
-4.05870676e-02 -1.65557697e-01 -2.08586916e-01 9.39560711e-01
1.31251857e-01 1.08009171e+00 -2.30438322e-01 -5.48103213e-01
3.00227046e-01 1.96545348e-02 3.57042551e-01 -3.03289771e-01
-4.78701085e-01 -1.17119774e-02 -5.06674834e-02 -6.54982746e-01
-4.76196826e-01 -1.89701974e-01 -1.45677602e+00 -2.94844180e-01
-3.54142785e-01 2.16749251e-01 6.91834450e-01 1.12702620e+00
3.52957994e-01 4.29049313e-01 5.59988916e-01 -6.90213025e-01
-4.92391378e-01 -9.65882719e-01 3.10328677e-02 3.48797768e-01
1.51676625e-01 -3.01181883e-01 -1.66840330e-01 -6.91676605e-03]
|
[10.774001121520996, 1.4639867544174194]
|
a1fbc783-9cd2-4979-a508-f3ab23552366
|
investigation-of-network-architecture-for
|
2212.10724
| null |
https://arxiv.org/abs/2212.10724v1
|
https://arxiv.org/pdf/2212.10724v1.pdf
|
Investigation of Network Architecture for Multimodal Head-and-Neck Tumor Segmentation
|
Inspired by the recent success of Transformers for Natural Language Processing and vision Transformer for Computer Vision, many researchers in the medical imaging community have flocked to Transformer-based networks for various main stream medical tasks such as classification, segmentation, and estimation. In this study, we analyze, two recently published Transformer-based network architectures for the task of multimodal head-and-tumor segmentation and compare their performance to the de facto standard 3D segmentation network - the nnU-Net. Our results showed that modeling long-range dependencies may be helpful in cases where large structures are present and/or large field of view is needed. However, for small structures such as head-and-neck tumor, the convolution-based U-Net architecture seemed to perform well, especially when training dataset is small and computational resource is limited.
|
['Quanzheng Li', 'Kuang Gong', 'Se-In Jang', 'Junyu Chen', 'Ye Li']
|
2022-12-21
| null | null | null | null |
['tumor-segmentation']
|
['computer-vision']
|
[ 5.61539046e-02 3.01110744e-01 -1.66804329e-01 -2.91571647e-01
-5.90160668e-01 -1.22013621e-01 2.93985158e-01 3.02871346e-01
-5.21093845e-01 5.35236120e-01 2.25043610e-01 -7.11497545e-01
3.64077874e-02 -6.72832131e-01 -4.33572203e-01 -6.82538688e-01
-8.27497244e-02 7.53365517e-01 3.75526965e-01 -1.16306096e-01
-1.88400567e-01 6.70836449e-01 -1.12039483e+00 1.38552815e-01
5.43545842e-01 1.27651155e+00 4.82909560e-01 5.61437845e-01
-3.30309778e-01 9.08951640e-01 -8.39914754e-02 -1.14936784e-01
-8.55581462e-02 -1.34266734e-01 -9.64073300e-01 1.22568822e-02
3.60654891e-01 -4.11766052e-01 -3.14820915e-01 8.91314209e-01
6.49154961e-01 -5.51038869e-02 7.21616805e-01 -7.57519186e-01
-2.45959178e-01 5.28426349e-01 -3.76150578e-01 4.62878734e-01
1.09299302e-01 -8.96195620e-02 6.67456686e-01 -6.32189393e-01
6.39978707e-01 1.01049137e+00 8.52927506e-01 7.07955062e-01
-8.53942752e-01 -3.99572343e-01 -7.03862235e-02 2.42449105e-01
-9.91038322e-01 -1.15361661e-01 5.65869927e-01 -3.52119029e-01
1.08093405e+00 4.33798432e-02 5.93618929e-01 9.10808861e-01
4.22113895e-01 1.12634909e+00 9.70054507e-01 -3.09986055e-01
-2.17541847e-02 1.32506609e-01 3.80517304e-01 8.66365075e-01
-1.10416241e-01 -2.20652428e-02 -1.37196139e-01 2.35999040e-02
8.36065710e-01 -1.98424887e-02 -3.68671268e-01 -1.81502596e-01
-1.05664611e+00 9.62677002e-01 7.80318856e-01 1.02155733e+00
-4.08992112e-01 7.95956329e-02 6.02790773e-01 1.56308621e-01
6.68159008e-01 7.93272257e-02 -2.89731175e-01 1.72323197e-01
-1.18981135e+00 -3.01496923e-01 4.96933997e-01 5.75273097e-01
2.82836497e-01 -1.52460728e-02 -3.00602257e-01 8.08190584e-01
3.40897173e-01 9.08900723e-02 8.63761604e-01 -5.75583518e-01
1.06784336e-01 5.31323433e-01 -4.82436687e-01 -2.08621517e-01
-9.06013608e-01 -4.94600177e-01 -1.08336902e+00 1.50381550e-01
5.88050425e-01 6.89006746e-02 -1.55846548e+00 1.46640527e+00
1.76347494e-01 -1.03692986e-01 -9.29620713e-02 8.82334888e-01
1.27824914e+00 9.22406837e-02 1.40873324e-02 -1.81580320e-01
1.38924897e+00 -8.82960856e-01 -6.18912637e-01 -2.93904185e-01
6.88244641e-01 -7.26330698e-01 7.69947827e-01 2.82746404e-02
-1.02776432e+00 -2.78543681e-01 -7.38556027e-01 -1.45453900e-01
-4.39119995e-01 -4.39135320e-02 5.88585734e-01 7.37375855e-01
-1.42422712e+00 6.08785331e-01 -1.02227902e+00 -8.08844328e-01
7.34557986e-01 6.19970739e-01 -3.31752837e-01 -2.81656325e-01
-9.58521008e-01 1.21992433e+00 3.05191815e-01 2.45803148e-01
-8.81470859e-01 -5.11406660e-01 -8.75023067e-01 2.17846967e-02
3.16006839e-02 -8.65328789e-01 1.36668777e+00 -6.44232512e-01
-1.17196894e+00 1.05108297e+00 -7.11871684e-02 -6.97354317e-01
6.60015404e-01 2.42855579e-01 1.03170902e-01 1.69445083e-01
-7.16566518e-02 8.36465418e-01 5.83938122e-01 -8.07293653e-01
-3.61287385e-01 -7.44977772e-01 -7.50931278e-02 7.91280344e-02
-9.10574496e-02 -2.33450136e-03 -2.92227656e-01 -3.78749847e-01
2.79701412e-01 -7.01397896e-01 -4.23164010e-01 6.10443838e-02
-3.66920054e-01 -3.24166507e-01 1.07646859e+00 -7.21746147e-01
7.31687725e-01 -1.89234352e+00 -8.27378482e-02 2.86086529e-01
4.52589482e-01 2.74044394e-01 5.55627272e-02 -1.62098005e-01
-3.30446571e-01 -4.99149673e-02 -1.95604786e-01 -3.66762221e-01
-3.75540078e-01 4.57231283e-01 4.09990102e-01 5.05913436e-01
-2.71569639e-01 1.05097950e+00 -6.19760811e-01 -8.03365886e-01
6.01771951e-01 5.79430938e-01 -1.49673805e-01 1.48486719e-01
-1.46466285e-01 6.57775700e-01 -5.65436006e-01 9.37351644e-01
6.10245228e-01 -3.30729395e-01 -5.00578769e-02 -4.31425720e-01
-6.23388449e-04 8.90618786e-02 -5.35980105e-01 1.75741804e+00
-6.31184638e-01 8.60049725e-01 4.99361247e-01 -1.18277824e+00
4.08031136e-01 6.49269760e-01 8.77899885e-01 -9.66646612e-01
6.58411205e-01 1.79020748e-01 1.54073998e-01 -5.68501234e-01
4.82523255e-02 -5.38977027e-01 4.04992431e-01 1.57058969e-01
1.63032487e-01 -1.91516876e-01 2.07476243e-01 6.61301911e-02
1.07272005e+00 -3.39744836e-02 9.92272720e-02 -2.36876369e-01
5.61469853e-01 -1.03797249e-01 1.11328043e-01 4.89387661e-01
-3.09045345e-01 7.99081087e-01 4.74346608e-01 -4.06811446e-01
-5.82198739e-01 -9.33362544e-01 -4.67773288e-01 7.74862528e-01
-1.95278049e-01 -1.72332630e-01 -8.01698089e-01 -6.67853653e-01
-3.53127390e-01 5.19395053e-01 -6.62725091e-01 2.31664479e-01
-4.77783948e-01 -8.28731477e-01 4.43385065e-01 6.88832819e-01
3.77895713e-01 -1.19202077e+00 -8.10443640e-01 4.48222548e-01
-2.23100916e-01 -1.30421519e+00 -3.06497723e-01 6.05855942e-01
-1.07628286e+00 -1.14535356e+00 -1.29297626e+00 -8.98302019e-01
5.45758545e-01 -1.64833322e-01 1.07792342e+00 -1.19782738e-01
-4.33203369e-01 6.30165994e-01 -1.16128437e-01 -4.12515491e-01
-2.74980634e-01 1.24236703e-01 -2.85165757e-01 -4.39095646e-01
1.08534940e-01 -4.57542270e-01 -4.71693069e-01 2.57214129e-01
-1.01336443e+00 -7.23822638e-02 5.91356575e-01 1.02239490e+00
5.12318373e-01 -1.20772235e-01 2.60361969e-01 -7.17518926e-01
6.82368100e-01 -1.70624301e-01 -2.35681996e-01 3.20550233e-01
-5.17387152e-01 -2.80243922e-02 2.23397821e-01 -2.55627394e-01
-7.75061667e-01 -5.47803231e-02 -6.89823747e-01 -6.62322700e-01
-2.57690847e-01 5.66145420e-01 2.08523721e-01 -3.65632355e-01
4.82800215e-01 1.32448375e-01 3.02848548e-01 -2.25658640e-01
1.74300950e-02 4.91667688e-01 3.58025163e-01 -2.20896482e-01
1.92920640e-01 4.20882821e-01 1.98666289e-01 -9.15815175e-01
-6.97171032e-01 -5.61680138e-01 -6.61429465e-01 -3.05708289e-01
1.26730776e+00 -5.41437149e-01 -6.29809260e-01 5.78271747e-01
-1.07623720e+00 -4.43608195e-01 -3.61217827e-01 5.03114998e-01
-5.67253828e-01 2.94357479e-01 -8.22679102e-01 -5.26166916e-01
-7.57366180e-01 -1.61767483e+00 9.98504281e-01 3.77833068e-01
5.35101667e-02 -1.30829632e+00 -2.47747809e-01 3.53327215e-01
7.75493205e-01 2.13847935e-01 1.08378172e+00 -5.04170895e-01
-4.07612622e-01 -2.12484360e-01 -3.67958933e-01 3.46557111e-01
1.62748426e-01 -4.02583748e-01 -1.05271780e+00 2.41139047e-02
1.77344047e-02 -3.56293648e-01 8.27144921e-01 1.03177726e+00
1.35461259e+00 3.28921050e-01 -4.55361515e-01 5.23015380e-01
1.16091955e+00 5.37364334e-02 4.38390255e-01 -2.21135244e-02
6.97823465e-01 5.26456356e-01 1.04190484e-01 -8.17938000e-02
3.83626908e-01 3.49382490e-01 7.23528743e-01 -4.96103227e-01
-2.15426415e-01 1.00114867e-01 -6.92121014e-02 6.02708995e-01
1.39160445e-02 -2.00928971e-01 -9.45433557e-01 5.94922841e-01
-1.65270293e+00 -4.08019364e-01 -6.26078919e-02 1.94348562e+00
2.75501907e-01 3.79813194e-01 8.84019118e-03 -7.71351755e-02
3.75387996e-01 -2.12686462e-03 -4.17739123e-01 -2.26569071e-01
1.01647168e-01 3.66161704e-01 5.90690136e-01 3.14771771e-01
-8.69908631e-01 7.58705199e-01 6.68176508e+00 9.47277188e-01
-1.55120885e+00 4.48979110e-01 8.87691617e-01 9.80300009e-02
-2.05221549e-01 -5.65803707e-01 -3.65208715e-01 1.24891281e-01
8.77248943e-01 4.70228642e-01 -9.04403329e-02 4.15582299e-01
1.76755473e-01 -5.30882716e-01 -1.02791524e+00 1.08351564e+00
-2.23298315e-02 -1.41487575e+00 -2.51304209e-01 1.68316767e-01
7.59270936e-02 6.07573450e-01 -8.30993056e-02 1.80410758e-01
2.41051205e-02 -1.32565129e+00 2.11358681e-01 1.12392187e-01
8.32332671e-01 -4.00818616e-01 1.13164616e+00 4.32775736e-01
-1.21963024e+00 2.46411845e-01 -3.96849141e-02 4.84032959e-01
2.39025623e-01 6.04329646e-01 -1.08602095e+00 3.50625426e-01
8.98848772e-01 4.82590646e-01 -3.56902122e-01 1.10242581e+00
3.22660469e-02 5.85436165e-01 -7.66316116e-01 9.13344771e-02
6.47885501e-01 -9.07038003e-02 3.59336793e-01 1.12449694e+00
2.30720744e-01 -2.19734455e-03 1.68845698e-01 5.85359871e-01
-9.50426310e-02 3.63506973e-02 -6.01195812e-01 1.21462993e-01
-4.12334263e-01 1.51245737e+00 -1.21926808e+00 -2.27362603e-01
-3.71362835e-01 7.76607454e-01 7.05225989e-02 2.56820560e-01
-6.39773071e-01 2.32051075e-01 1.38056144e-01 3.22174072e-01
3.61806571e-01 -1.53828010e-01 -3.87308985e-01 -9.16297376e-01
-2.70640939e-01 -3.11458439e-01 6.00163400e-01 -7.82216609e-01
-9.49130595e-01 8.68992567e-01 -5.23072444e-02 -9.24194872e-01
-1.97550505e-01 -9.78112817e-01 -7.91544914e-01 5.79310656e-01
-1.63438761e+00 -1.57629335e+00 -8.48318934e-02 8.40777099e-01
4.93145585e-01 4.65716273e-02 8.05342138e-01 4.14167166e-01
-3.13182652e-01 4.26397294e-01 -2.85476986e-02 2.37943858e-01
4.38935786e-01 -1.23075497e+00 -2.69238561e-01 4.66011316e-01
-5.52408770e-02 1.47551209e-01 5.08165479e-01 -2.62185812e-01
-1.05572855e+00 -7.51343012e-01 8.11638594e-01 -3.63212614e-03
6.38333321e-01 -1.94574282e-01 -6.47657216e-01 6.18264437e-01
4.86759186e-01 3.56660932e-01 5.44071257e-01 -7.92205855e-02
6.07563965e-02 -2.80648861e-02 -1.60267222e+00 1.59496069e-01
6.67911291e-01 -4.60457861e-01 -3.29243183e-01 5.28569221e-01
2.24641278e-01 -7.75158048e-01 -1.00014412e+00 4.22250897e-01
3.95049870e-01 -9.36908126e-01 8.75118971e-01 -2.08888412e-01
1.28133863e-01 1.96735561e-01 -1.01349659e-01 -1.26567376e+00
-9.16871130e-02 -1.73138902e-01 2.45101005e-01 7.21630514e-01
3.81641656e-01 -5.43914378e-01 1.00817871e+00 5.77437222e-01
-3.76775593e-01 -9.55229402e-01 -1.37730825e+00 -2.44368836e-01
1.41404971e-01 -6.49015665e-01 -4.83972356e-02 5.17239213e-01
-2.71271080e-01 5.55957496e-01 1.10871948e-01 -2.64731139e-01
5.94772518e-01 -9.17219743e-02 3.38392854e-02 -1.23085725e+00
1.14211343e-01 -6.40500247e-01 -5.41378975e-01 -7.86424577e-01
1.78090200e-01 -9.04395044e-01 -1.18265010e-01 -1.99035883e+00
2.43147393e-03 -3.24657619e-01 -2.54512280e-01 6.19363189e-01
3.26681584e-01 2.51323998e-01 1.14717297e-01 -1.46652937e-01
-5.58685325e-02 2.69377798e-01 1.61719966e+00 -4.82562661e-01
-5.34763522e-02 4.81066227e-01 -4.04848844e-01 7.41024971e-01
4.93118554e-01 -2.13866308e-01 -3.59061182e-01 -3.04247200e-01
-3.81528795e-01 4.72943604e-01 2.25357175e-01 -9.37798381e-01
4.58004326e-01 3.99216264e-01 4.77985710e-01 -8.81539047e-01
4.21249032e-01 -1.19013393e+00 -2.73911357e-01 6.37466133e-01
7.46524185e-02 -7.89256245e-02 3.16811532e-01 -1.37388175e-02
-5.34019351e-01 -3.34071249e-01 9.87775385e-01 -6.06070995e-01
-7.91079342e-01 5.88656604e-01 -7.20416367e-01 -2.74332762e-01
9.13337231e-01 -4.42871869e-01 9.11401138e-02 -3.94288123e-01
-1.18720639e+00 2.78569400e-01 -1.12065189e-01 1.62030354e-01
7.35051036e-01 -9.02311206e-01 -5.47106028e-01 7.19729662e-02
-8.39820057e-02 3.59168917e-01 3.00505757e-01 1.37662804e+00
-6.34108782e-01 6.42166972e-01 -1.80581048e-01 -1.00204122e+00
-1.42385733e+00 2.52231002e-01 8.73093009e-01 -7.43172884e-01
-7.58495331e-01 9.69823658e-01 3.17892402e-01 -5.60891569e-01
3.90769631e-01 -9.93974328e-01 -3.90028208e-01 3.55627805e-01
1.11613780e-01 1.44495564e-02 5.60240567e-01 -6.83694541e-01
-5.63215137e-01 5.70997238e-01 -1.84064955e-01 8.80155414e-02
1.38866174e+00 -4.11052331e-02 -2.26582646e-01 2.20515698e-01
1.33084285e+00 -6.28232896e-01 -6.74448192e-01 -3.77955019e-01
-1.69938609e-01 2.97057778e-01 6.42260611e-01 -6.94972754e-01
-1.56811011e+00 1.08875990e+00 1.11275530e+00 7.87227154e-02
1.30685914e+00 3.17017794e-01 7.90091336e-01 1.78607762e-01
2.90478110e-01 -8.33106518e-01 -2.04494193e-01 7.17554033e-01
5.01327336e-01 -1.45184147e+00 -5.84163144e-02 -2.34257936e-01
-3.67297262e-01 1.29991472e+00 5.61707437e-01 1.85236007e-01
1.08538115e+00 4.92784262e-01 1.43155664e-01 -4.57942545e-01
-2.06798553e-01 -5.61593115e-01 5.05619764e-01 6.32083833e-01
7.48171031e-01 1.67380601e-01 4.55163307e-02 4.41899002e-02
-1.10110298e-01 2.12633759e-01 3.21247339e-01 8.36338997e-01
-3.55651855e-01 -1.03501105e+00 -4.00012463e-01 8.12643409e-01
-7.28924811e-01 -1.66924849e-01 -1.22685261e-01 8.22771370e-01
2.07651347e-01 5.96344173e-01 2.48888940e-01 6.72961697e-02
1.41831174e-01 4.96795680e-03 8.63195419e-01 -4.05123323e-01
-9.58864033e-01 4.81479555e-01 1.23002566e-01 -4.60054427e-01
-7.42616117e-01 -5.11157095e-01 -1.35261118e+00 6.61021248e-02
-3.39017808e-01 -1.29611939e-01 7.81816244e-01 1.26843560e+00
-3.32845241e-01 7.40718901e-01 1.13347113e-01 -7.64887214e-01
-6.43806830e-02 -1.22687078e+00 -6.17588520e-01 -5.52814454e-02
5.72441816e-01 -5.07652760e-01 7.95467719e-02 -2.16860339e-01]
|
[14.539372444152832, -2.473653793334961]
|
7f823ee4-e19c-4fdf-86f5-ff5c014f431d
|
hyperthumbnail-real-time-6k-image-rescaling
|
2304.01064
| null |
https://arxiv.org/abs/2304.01064v2
|
https://arxiv.org/pdf/2304.01064v2.pdf
|
Real-time 6K Image Rescaling with Rate-distortion Optimization
|
Contemporary image rescaling aims at embedding a high-resolution (HR) image into a low-resolution (LR) thumbnail image that contains embedded information for HR image reconstruction. Unlike traditional image super-resolution, this enables high-fidelity HR image restoration faithful to the original one, given the embedded information in the LR thumbnail. However, state-of-the-art image rescaling methods do not optimize the LR image file size for efficient sharing and fall short of real-time performance for ultra-high-resolution (e.g., 6K) image reconstruction. To address these two challenges, we propose a novel framework (HyperThumbnail) for real-time 6K rate-distortion-aware image rescaling. Our framework first embeds an HR image into a JPEG LR thumbnail by an encoder with our proposed quantization prediction module, which minimizes the file size of the embedding LR JPEG thumbnail while maximizing HR reconstruction quality. Then, an efficient frequency-aware decoder reconstructs a high-fidelity HR image from the LR one in real time. Extensive experiments demonstrate that our framework outperforms previous image rescaling baselines in rate-distortion performance and can perform 6K image reconstruction in real time.
|
['Qifeng Chen', 'Ying-Cong Chen', 'Ka Leong Cheng', 'Xin Yang', 'Chenyang Qi']
|
2023-04-03
|
real-time-6k-image-rescaling-with-rate
|
http://openaccess.thecvf.com//content/CVPR2023/html/Qi_Real-Time_6K_Image_Rescaling_With_Rate-Distortion_Optimization_CVPR_2023_paper.html
|
http://openaccess.thecvf.com//content/CVPR2023/papers/Qi_Real-Time_6K_Image_Rescaling_With_Rate-Distortion_Optimization_CVPR_2023_paper.pdf
|
cvpr-2023-1
|
['image-super-resolution']
|
['computer-vision']
|
[ 1.02156460e+00 7.02178627e-02 -3.53880674e-01 -1.04227560e-02
-1.19645607e+00 -2.62099206e-01 2.58113354e-01 -2.91578501e-01
-2.74554700e-01 4.69896525e-01 6.73639655e-01 -1.88677236e-01
-3.12013309e-02 -9.07156944e-01 -1.10740721e+00 -5.94909072e-01
2.02140033e-01 -3.78316253e-01 5.70456460e-02 -1.76948443e-01
4.58418190e-01 2.66168773e-01 -1.63796830e+00 4.57146049e-01
7.48010635e-01 1.02087820e+00 7.80956328e-01 1.16241109e+00
5.44855773e-01 1.00041234e+00 -2.27582753e-01 -2.56001651e-02
5.91058731e-01 -5.27587771e-01 -5.55478752e-01 1.38886841e-02
8.53840828e-01 -1.18235433e+00 -9.70997453e-01 9.20103788e-01
5.67807317e-01 4.63118441e-02 2.37305701e-01 -5.71314394e-01
-1.25760317e+00 4.23316658e-01 -6.69280052e-01 2.16114998e-01
6.29879296e-01 1.81329831e-01 6.34472728e-01 -9.27886069e-01
7.27611065e-01 1.26043820e+00 5.07873476e-01 3.11341614e-01
-1.47415972e+00 -6.16632104e-01 -5.55595160e-01 1.40980542e-01
-1.53953183e+00 -8.65787327e-01 4.46320146e-01 3.17018121e-01
8.68610144e-01 4.97579247e-01 4.09142286e-01 7.28462100e-01
4.08987314e-01 6.54095933e-02 1.30934381e+00 -4.94945109e-01
3.01315598e-02 -9.73664224e-02 -4.93239492e-01 2.77857929e-01
-1.08476998e-02 3.34447056e-01 -9.37393844e-01 1.10686012e-01
1.62553298e+00 4.40827645e-02 -6.82100594e-01 1.78960547e-01
-1.59397936e+00 3.01758230e-01 4.94287580e-01 1.18387595e-01
-2.69616485e-01 4.85603362e-01 3.36121693e-02 4.73074049e-01
2.48874247e-01 2.32024714e-01 2.08985247e-02 -9.80691891e-03
-1.13533568e+00 5.52351438e-02 4.92828935e-02 9.80132043e-01
8.13404560e-01 -7.92382378e-03 -3.03312540e-01 9.90360081e-01
-5.12276217e-03 6.60399020e-01 5.53490400e-01 -1.57431126e+00
5.73897660e-01 -1.54213086e-01 1.65364221e-01 -9.84940112e-01
2.91578621e-01 -1.44074008e-01 -1.23718822e+00 1.75681069e-01
-2.37647042e-01 6.09606504e-01 -7.44519830e-01 1.34104335e+00
1.16489060e-01 2.46929452e-01 2.64208674e-01 1.01698935e+00
6.80869877e-01 1.10808599e+00 -4.99092191e-01 -6.61282718e-01
1.44445169e+00 -8.62798452e-01 -1.04790270e+00 -3.99338752e-02
-1.67250812e-01 -9.58686471e-01 1.29017627e+00 2.78284371e-01
-1.48349452e+00 -1.02321053e+00 -1.33061361e+00 -7.95004427e-01
2.61017054e-01 9.34578013e-04 -6.32713214e-02 5.17174900e-01
-1.37565267e+00 6.79889798e-01 -2.07045630e-01 -1.18724778e-01
1.30311847e-01 1.23398356e-01 -5.65813661e-01 -5.09261012e-01
-1.25161338e+00 6.83018684e-01 3.38739127e-01 -3.61779809e-01
-6.91475928e-01 -1.15849280e+00 -8.50836754e-01 -6.93515390e-02
2.18127221e-01 -6.16100788e-01 1.07304394e+00 -6.28661811e-01
-1.71535802e+00 7.55209625e-01 -1.40760005e-01 -5.95806122e-01
4.25134659e-01 -1.58276379e-01 -6.18784368e-01 6.42928302e-01
-5.72322309e-02 6.07022047e-01 1.51532209e+00 -1.43346536e+00
-6.22600198e-01 -1.65154308e-01 -1.15966856e-01 4.54205811e-01
-3.23244303e-01 -2.31578827e-01 -7.38106549e-01 -1.12053835e+00
2.50453919e-01 -5.53600073e-01 -8.45164582e-02 2.66402125e-01
-2.90481448e-01 7.18093693e-01 9.48671877e-01 -8.88944805e-01
1.51476383e+00 -2.49617505e+00 7.06583168e-03 -2.08753780e-01
4.77586538e-01 8.67746994e-02 -1.64444491e-01 3.56888235e-01
-8.60960931e-02 2.38463487e-02 -3.36578012e-01 -4.03640270e-01
-2.96556771e-01 1.92942977e-01 -7.08960116e-01 6.30927920e-01
-2.45866716e-01 9.10876334e-01 -7.94399798e-01 -5.58366299e-01
6.66078091e-01 1.10936093e+00 -4.87823993e-01 5.23874938e-01
2.79067069e-01 4.49306339e-01 2.39207849e-01 4.24166679e-01
1.06370795e+00 -3.03646445e-01 2.18593106e-01 -9.50049043e-01
-2.72722483e-01 -2.41182577e-02 -9.56167698e-01 1.77072251e+00
-7.33480632e-01 6.12271249e-01 -7.71832988e-02 -2.06298456e-01
8.65732968e-01 9.28851217e-02 4.00076002e-01 -1.26688075e+00
-5.54810584e-01 2.41629422e-01 -8.50735307e-01 -9.70701724e-02
1.20557976e+00 3.93278757e-03 -1.32589089e-02 4.22878772e-01
-3.25402349e-01 -2.49459863e-01 -2.74114400e-01 3.89618367e-01
9.90103185e-01 -3.40899378e-02 4.27146167e-01 8.65737721e-02
3.68761718e-01 -7.41961956e-01 1.26293644e-01 7.92955160e-01
-2.57180613e-02 1.23726833e+00 -8.41346458e-02 -2.93232292e-01
-1.95390093e+00 -1.27539206e+00 -3.89668494e-01 8.22423458e-01
5.26556492e-01 -6.82782054e-01 -8.00643265e-01 1.56743556e-01
-3.51578295e-01 5.07115245e-01 -3.96463841e-01 -2.05597624e-01
-6.90701008e-01 -3.40823174e-01 4.43242162e-01 -2.69505009e-02
1.00435698e+00 -7.05327630e-01 -7.73064256e-01 1.81652129e-01
-6.22919679e-01 -1.59841728e+00 -9.90775406e-01 -3.30020249e-01
-8.28803062e-01 -5.98898053e-01 -1.06458604e+00 -5.99615455e-01
5.60937762e-01 8.35013688e-01 1.05884421e+00 1.46791488e-02
-3.92742991e-01 4.52896237e-01 -4.61313158e-01 5.69222808e-01
-7.73830593e-01 -3.65903258e-01 3.47232744e-02 1.45336345e-01
-5.37852824e-01 -7.62573123e-01 -1.06157768e+00 1.26029700e-01
-1.42452145e+00 5.62339723e-01 7.06589758e-01 6.50174201e-01
1.26267743e+00 4.43144321e-01 2.09792137e-01 -4.70741123e-01
3.02147478e-01 -4.17967103e-02 -4.11688834e-01 1.04977362e-01
-9.24825728e-01 1.32071853e-01 1.01771617e+00 -3.76361310e-01
-1.02543020e+00 -8.29345584e-02 8.83528311e-03 -5.69289386e-01
2.98262686e-01 -2.15794727e-01 1.59720276e-02 -4.63495523e-01
4.97573853e-01 9.23911870e-01 2.14713085e-02 -4.17606503e-01
7.98561633e-01 8.12425077e-01 1.41935241e+00 -1.37951583e-01
1.05188084e+00 6.27063036e-01 4.30046655e-02 -8.49003136e-01
-4.46648657e-01 -4.79978882e-02 -4.56240952e-01 -2.38920689e-01
7.86983669e-01 -1.33614206e+00 -5.13738334e-01 4.02422935e-01
-6.66935802e-01 -4.05641705e-01 -6.35180116e-01 1.60312414e-01
-1.07671225e+00 7.65819490e-01 -8.96201670e-01 -4.51261193e-01
-4.79678512e-01 -1.15817022e+00 1.45182872e+00 1.41672075e-01
4.10548776e-01 -3.44157726e-01 -8.61427188e-02 4.10325110e-01
8.00944686e-01 9.88173112e-02 6.17599547e-01 9.00433004e-01
-9.92205262e-01 5.25713004e-02 -7.70884633e-01 3.37379903e-01
2.10175723e-01 -5.53499758e-01 -9.45368171e-01 -5.09267211e-01
-1.19764626e-01 -3.48665833e-01 8.63141775e-01 2.50992239e-01
1.46230853e+00 -7.13234842e-01 8.27466324e-02 1.16402233e+00
1.93541539e+00 -3.10301751e-01 1.53974438e+00 1.74954697e-01
7.81654418e-01 2.70553827e-02 8.25203300e-01 7.02561259e-01
4.97013420e-01 1.14757240e+00 3.14873278e-01 -2.32079223e-01
-8.16985607e-01 -6.42012417e-01 6.56595230e-01 7.66339958e-01
-7.97277167e-02 -6.18267693e-02 -1.62811518e-01 1.33937269e-01
-1.43979454e+00 -1.12511742e+00 2.19433963e-01 2.60869980e+00
1.26590264e+00 -3.03621799e-01 -1.98071435e-01 2.30932176e-01
6.23926878e-01 6.46810949e-01 -7.57989943e-01 -1.10512741e-01
-4.59400952e-01 6.56016618e-02 9.87575352e-01 6.95622146e-01
-7.75811791e-01 8.08222473e-01 6.28935003e+00 1.31714964e+00
-8.70510638e-01 2.31645837e-01 8.60272408e-01 -2.43257910e-01
-5.70882261e-01 -7.12569654e-02 -5.77970564e-01 5.28906107e-01
1.40403402e+00 -3.16056699e-01 1.08278489e+00 2.64710903e-01
3.77080292e-01 -1.76325917e-01 -7.37731278e-01 1.56111240e+00
1.90964341e-01 -1.63934779e+00 2.98030674e-01 4.27517891e-01
6.65569663e-01 -4.38763916e-01 5.52012026e-01 -2.13544667e-01
-1.72601283e-01 -1.14055443e+00 7.25413024e-01 6.37811542e-01
2.06344056e+00 -8.67564917e-01 7.80979469e-02 -1.17128775e-01
-1.59064901e+00 -1.11318655e-01 -6.28363132e-01 5.20380914e-01
2.58189380e-01 7.66851783e-01 -2.42681801e-01 5.52795410e-01
1.01127958e+00 7.64816701e-01 -4.27985817e-01 2.00880766e-01
2.30571538e-01 1.30667821e-01 -2.81587485e-02 1.04365253e+00
-2.85249263e-01 -1.01068862e-01 4.92273927e-01 1.00955737e+00
5.71223915e-01 5.10501087e-01 -1.04015417e-01 7.31687248e-01
-2.27831453e-01 -6.89130202e-02 -6.20278239e-01 1.26700550e-01
7.70366669e-01 9.25086141e-01 -4.47506547e-01 -3.73048514e-01
-1.85325533e-01 1.85479355e+00 -2.63122320e-02 3.35414588e-01
-9.15761948e-01 -3.42855513e-01 6.81500435e-01 3.71399879e-01
2.38268703e-01 -1.83415845e-01 -7.61317536e-02 -1.14293087e+00
1.46867275e-01 -1.09203148e+00 -9.96585712e-02 -1.19883192e+00
-7.88198948e-01 5.38476110e-01 -3.51153791e-01 -1.53245568e+00
-4.61938940e-02 4.35444675e-02 2.79884964e-01 7.59378016e-01
-1.75137138e+00 -9.91236091e-01 -5.91014504e-01 7.50383258e-01
6.68569267e-01 1.76707238e-01 7.75587380e-01 4.15379375e-01
-3.81655432e-02 7.86840558e-01 4.93828744e-01 -3.77122849e-01
1.02763164e+00 -8.91187608e-01 5.68461120e-01 8.72258782e-01
-1.71238318e-01 3.41816038e-01 8.25886786e-01 -6.02966011e-01
-2.12219930e+00 -1.28519201e+00 5.19458890e-01 -8.54501873e-02
6.83406815e-02 -2.10072860e-01 -9.37506318e-01 4.09756899e-01
2.46153325e-01 4.76560742e-01 2.51877308e-01 -1.02663589e+00
-6.74227655e-01 -3.72204781e-01 -1.46148622e+00 7.36228764e-01
1.15116429e+00 -1.10248697e+00 6.69389591e-02 2.70114273e-01
1.50108719e+00 -5.60290396e-01 -1.42722392e+00 2.98439115e-01
7.66783476e-01 -1.28055918e+00 1.62965345e+00 5.13660431e-01
9.54215765e-01 -5.71014404e-01 -7.82409549e-01 -8.58551621e-01
-4.80133772e-01 -9.13782716e-01 -4.53491092e-01 8.64967525e-01
-2.05874786e-01 -3.51667553e-01 2.89399207e-01 1.85856029e-01
2.00388387e-01 -4.37381148e-01 -8.76114011e-01 -7.03490794e-01
-5.26568711e-01 -2.68611670e-01 6.00619793e-01 6.78253293e-01
-2.59717077e-01 -1.25033543e-01 -1.12615287e+00 3.87807161e-01
1.29616845e+00 7.01066777e-02 6.91933870e-01 -3.36707562e-01
-4.10050303e-01 1.05340451e-01 -3.49799424e-01 -1.25845838e+00
-3.70833278e-01 -5.53897679e-01 7.80153945e-02 -1.33470833e+00
4.21063125e-01 -1.61203697e-01 -6.40038326e-02 1.03624552e-01
-1.04287058e-01 1.07301271e+00 4.29725587e-01 6.49329066e-01
-5.35094917e-01 5.29853225e-01 1.39207792e+00 1.34712830e-01
-7.67060369e-02 -7.72033989e-01 -7.16355264e-01 7.64685124e-03
4.25067186e-01 -1.40776321e-01 -4.78841841e-01 -2.83741802e-01
1.78439990e-01 6.23677492e-01 5.08472800e-01 -1.09581661e+00
1.19426042e-01 -4.92341407e-02 5.03372014e-01 -4.89930302e-01
5.07268906e-01 -7.68960357e-01 5.80722153e-01 2.55517900e-01
-5.30210674e-01 -4.34500426e-02 -1.15823910e-01 8.04345906e-01
-1.19524620e-01 3.49238604e-01 1.24016690e+00 -8.08947384e-02
-6.35996342e-01 5.05291522e-01 -3.19415964e-02 -2.76170343e-01
7.85069525e-01 -5.07161021e-01 -5.25264084e-01 -6.59001946e-01
-2.10733622e-01 -4.56647843e-01 1.03441679e+00 3.43676567e-01
1.30218112e+00 -1.46008551e+00 -8.81619096e-01 3.71566236e-01
8.62599313e-02 -2.45190412e-01 7.73745358e-01 3.78338218e-01
-8.06670368e-01 1.91221327e-01 -3.44174594e-01 -4.55639899e-01
-1.29142821e+00 7.10199833e-01 1.52637765e-01 -2.28739858e-01
-1.35975659e+00 4.27446783e-01 2.06104338e-01 9.32999998e-02
-5.25558293e-02 1.36989327e-02 1.79414004e-02 -5.13337791e-01
1.24641788e+00 3.13331604e-01 -1.28095657e-01 -8.34211588e-01
7.18847737e-02 1.02811158e+00 -1.12309977e-01 -4.74015892e-01
1.21531725e+00 -9.82654989e-01 -1.56545788e-01 2.04041809e-01
1.42461455e+00 1.17359094e-01 -1.56169879e+00 -4.38820571e-01
-7.27812409e-01 -1.23937976e+00 5.07492721e-01 -4.76733893e-01
-1.04224992e+00 3.56874079e-01 9.51282501e-01 -4.54798006e-02
1.80073917e+00 -1.74390689e-01 1.26953900e+00 -1.72060147e-01
5.81684768e-01 -9.75629330e-01 3.07575434e-01 -6.21959567e-02
1.18240881e+00 -1.09701598e+00 3.80055577e-01 -3.85188878e-01
-3.51182312e-01 1.24895489e+00 8.82968009e-02 -8.02106187e-02
2.46385157e-01 4.09027547e-01 -2.03369692e-01 3.29435527e-01
-7.38349438e-01 2.79492587e-01 4.65746038e-02 6.34476602e-01
2.13581249e-01 5.33912797e-03 2.82055326e-02 -9.15054604e-02
-3.65979463e-01 2.43089363e-01 9.42184508e-01 4.75313336e-01
-4.81096506e-01 -7.43750691e-01 -7.19525576e-01 2.28284925e-01
-3.68092448e-01 -5.35568178e-01 3.81717265e-01 1.62070423e-01
-9.51357037e-02 9.72490311e-01 1.08040221e-01 -5.84877670e-01
1.49927080e-01 -7.24024653e-01 6.79256260e-01 -2.10140999e-02
-1.52116105e-01 7.45639205e-02 -4.24082160e-01 -1.20960057e+00
-2.50845104e-01 -1.40181288e-01 -1.00486541e+00 -7.49482632e-01
1.82140380e-01 -3.94046873e-01 6.39357209e-01 3.29761267e-01
6.16069198e-01 4.27183121e-01 1.07420444e+00 -1.10662115e+00
-3.80292833e-01 -4.15678293e-01 -8.07705343e-01 3.41525912e-01
9.30608034e-01 2.48960435e-01 -3.92639875e-01 4.28193301e-01]
|
[10.962095260620117, -2.1103968620300293]
|
76fa3385-8ab6-409c-a583-ae68ffcf002e
|
event-causality-identification-via-generation
| null | null |
https://aclanthology.org/2022.starsem-1.28
|
https://aclanthology.org/2022.starsem-1.28.pdf
|
Event Causality Identification via Generation of Important Context Words
|
An important problem of Information Extraction involves Event Causality Identification (ECI) that seeks to identify causal relation between pairs of event mentions. Prior models for ECI have mainly solved the problem using the classification framework that does not explore prediction/generation of important context words from input sentences for causal recognition. In this work, we consider the words along the dependency path between the two event mentions in the dependency tree as the important context words for ECI. We introduce dependency path generation as a complementary task for ECI, which can be solved jointly with causal label prediction to improve the performance. To facilitate the multi-task learning, we cast ECI into a generation problem that aims to generate both causal relation and dependency path words from input sentence. In addition, we propose to use the REINFORCE algorithm to train our generative model where novel reward functions are designed to capture both causal prediction accuracy and generation quality. The experiments on two benchmark datasets demonstrate state-of-the-art performance of the proposed model for ECI.
|
['Thien Nguyen', 'Minh Nguyen', 'Hieu Man']
| null | null | null | null |
sem-naacl-2022-7
|
['event-causality-identification']
|
['natural-language-processing']
|
[ 4.98204499e-01 3.55423659e-01 -3.41357142e-01 -3.82585347e-01
-7.85295963e-01 -2.24195108e-01 1.00139380e+00 4.13436353e-01
-1.20849647e-01 1.10881948e+00 8.04939926e-01 -3.20757538e-01
-3.74885440e-01 -8.79506290e-01 -6.04802132e-01 -5.81071019e-01
-2.91195154e-01 3.75212729e-01 1.21606246e-01 1.86462849e-01
4.05057073e-01 -1.57043692e-02 -1.06146348e+00 5.49648046e-01
9.02998507e-01 6.52808666e-01 3.25316787e-01 4.75756705e-01
-9.35894474e-02 1.25902390e+00 -6.58714592e-01 -4.33232695e-01
-4.64436024e-01 -8.16530287e-01 -1.08652222e+00 -5.22292435e-01
-3.67167085e-01 -7.87621289e-02 -1.25478297e-01 6.68958783e-01
3.72220308e-01 6.85942322e-02 9.79824424e-01 -1.63637948e+00
-6.01206601e-01 1.35322082e+00 -6.00193381e-01 5.33159733e-01
4.12311018e-01 -2.26014063e-01 1.52979052e+00 -7.59809971e-01
6.32145226e-01 1.52636707e+00 1.52851447e-01 3.69441599e-01
-9.30334330e-01 -9.16785896e-01 5.18630147e-01 8.21424305e-01
-1.04864252e+00 -4.62042019e-02 8.94528687e-01 -3.60708952e-01
1.12005210e+00 -3.20412708e-03 4.63767618e-01 1.53981197e+00
5.82892120e-01 1.02858281e+00 8.15955579e-01 -4.06153768e-01
1.14180490e-01 -2.59280533e-01 3.87925714e-01 3.68556440e-01
-6.73635229e-02 2.71432281e-01 -8.90164256e-01 -2.42690891e-01
5.33313096e-01 -3.06151450e-01 -1.55894220e-01 5.51530302e-01
-1.31254816e+00 1.10133696e+00 3.74603897e-01 2.39330336e-01
-5.36171198e-01 5.16871929e-01 3.00928533e-01 -9.05892551e-02
4.54922318e-01 2.61872411e-01 -4.90120828e-01 -6.75312951e-02
-3.68721545e-01 4.68616456e-01 6.08284771e-01 8.97103488e-01
3.73485655e-01 -2.06759959e-01 -9.40836608e-01 6.47736192e-01
6.26288593e-01 9.15484056e-02 3.21665108e-01 -3.17537427e-01
6.98056161e-01 6.37132823e-01 -6.85342550e-02 -9.93342996e-01
-4.78391826e-01 -4.64496613e-01 -7.69319892e-01 -3.56850892e-01
1.41194627e-01 -4.39497292e-01 -5.57255030e-01 1.98944187e+00
4.76776838e-01 9.16128039e-01 1.88146159e-01 8.07389498e-01
9.88775790e-01 9.20979023e-01 6.97051108e-01 -5.30941367e-01
1.36773837e+00 -7.10637391e-01 -1.07157207e+00 -3.51600468e-01
5.20627797e-01 -6.69885397e-01 8.02653313e-01 2.30337292e-01
-6.79101110e-01 -3.50116700e-01 -9.50875998e-01 1.33681163e-01
-1.32546257e-02 1.28125235e-01 7.56130874e-01 -4.18329090e-02
-2.79348552e-01 4.77757424e-01 -5.41907132e-01 -3.11808139e-02
2.17475891e-01 1.13717511e-01 6.45336732e-02 2.33108029e-01
-2.00131488e+00 8.07880700e-01 8.31676304e-01 1.29100680e-01
-1.28842509e+00 -7.75951743e-01 -6.57261491e-01 2.90516078e-01
6.15014255e-01 -5.55657923e-01 1.13268173e+00 -4.91291523e-01
-1.05494082e+00 2.49319717e-01 -3.77248079e-01 -4.45192605e-01
4.79017198e-02 -5.22703171e-01 -5.14090896e-01 -1.63309097e-01
3.28552932e-01 3.31746340e-01 7.16243684e-01 -1.31784749e+00
-1.11266029e+00 -1.67519957e-01 -7.79386833e-02 2.01718792e-01
-2.19937846e-01 2.78138965e-01 -1.60189942e-01 -8.28819513e-01
-2.88510501e-01 -6.07822359e-01 -9.62371156e-02 -9.65645432e-01
-7.94995010e-01 -9.64643240e-01 7.69742310e-01 -5.87351859e-01
1.68028378e+00 -1.68865752e+00 2.06034824e-01 -7.76370615e-02
-2.62842011e-02 -1.19144984e-01 -1.43729419e-01 4.10314232e-01
-5.03385425e-01 3.57482314e-01 -2.07240790e-01 -1.55571431e-01
-2.08769590e-01 2.29128793e-01 -6.95701420e-01 -5.29488735e-02
7.83201635e-01 8.56017172e-01 -1.34951591e+00 -7.52164662e-01
-2.81112432e-01 1.88157663e-01 -3.01127076e-01 6.43837035e-01
-5.63310385e-01 6.13566101e-01 -7.29743779e-01 1.52459964e-01
2.12237626e-01 -3.17333966e-01 3.40494335e-01 -1.31895214e-01
-1.43720686e-01 8.72489750e-01 -1.11039042e+00 1.40289271e+00
-4.61685449e-01 8.74983892e-03 -7.14095771e-01 -1.17832530e+00
9.04199183e-01 7.60823250e-01 3.29174757e-01 -4.96092707e-01
7.45686376e-03 -8.90946761e-02 2.19147488e-01 -5.22002697e-01
1.54462054e-01 -2.42876172e-01 -2.77750313e-01 6.76676035e-01
1.68028116e-01 2.85291404e-01 3.29652905e-01 3.92179072e-01
1.06506252e+00 3.05092931e-01 4.75300729e-01 -7.80479684e-02
5.02546847e-01 -1.09488696e-01 1.04508483e+00 6.07342362e-01
2.64312565e-01 3.47228557e-01 1.04451442e+00 -1.31081715e-01
-4.80258048e-01 -9.23216939e-01 4.49589007e-02 9.43921208e-01
6.97251409e-02 -5.28972030e-01 -3.86358887e-01 -1.01771832e+00
-3.95833671e-01 1.39750385e+00 -8.80538106e-01 -2.20020890e-01
-8.59463274e-01 -1.25775683e+00 2.76994139e-01 8.14143896e-01
3.76684666e-01 -1.42975378e+00 -4.98846173e-01 5.97085118e-01
-6.57495558e-01 -8.97969663e-01 -3.08933288e-01 4.21556622e-01
-5.20793855e-01 -1.11458325e+00 -9.04349536e-02 -5.96794963e-01
2.24617273e-01 -2.15187520e-01 1.21983325e+00 -2.80821063e-02
-1.36609197e-01 -8.88844281e-02 -6.98648334e-01 -6.25662982e-01
-3.61010492e-01 1.01583898e-01 -2.83673853e-01 1.11457787e-01
2.09882349e-01 -6.63639605e-01 -4.93597418e-01 2.02186868e-01
-6.50645256e-01 3.37459534e-01 4.67271268e-01 9.80570912e-01
6.33570790e-01 2.69306719e-01 1.04206514e+00 -1.09803712e+00
7.34900534e-01 -9.81793642e-01 -1.52400449e-01 5.35531461e-01
-7.81509876e-01 3.71772677e-01 6.38334095e-01 -3.56214792e-01
-1.74920022e+00 -1.74234182e-01 -3.43304910e-02 2.00801432e-01
-1.20404474e-01 9.73791182e-01 -4.67255086e-01 9.88266468e-01
4.29313570e-01 -7.41237849e-02 -8.98824930e-01 -1.99636087e-01
4.40570414e-01 2.58839637e-01 3.20304871e-01 -7.72732973e-01
4.49147165e-01 -1.03937484e-01 1.26774341e-01 -2.16040999e-01
-1.30385292e+00 -2.65525669e-01 -5.22901416e-01 -2.21898019e-01
1.14359963e+00 -7.59430707e-01 -7.46516645e-01 7.38030532e-04
-1.74048328e+00 -2.73849934e-01 1.04219057e-01 4.70879227e-01
-4.81401712e-01 -4.18343171e-02 -3.63179445e-01 -1.00455546e+00
-4.81666833e-01 -8.74510884e-01 1.07493198e+00 2.75941074e-01
-5.13033330e-01 -9.95806813e-01 2.73281127e-01 -6.14549518e-02
-2.37034291e-01 4.18941468e-01 1.46134675e+00 -7.42124677e-01
-5.64466894e-01 2.69661695e-01 -2.99679637e-01 -3.38327736e-01
1.59485251e-01 6.07716441e-02 -7.44983733e-01 3.78487200e-01
-1.18274666e-01 -2.89799601e-01 9.57632244e-01 2.76227534e-01
1.03733134e+00 -6.19932711e-01 -6.69936836e-01 2.37661712e-02
1.17464304e+00 6.22806787e-01 6.10842288e-01 3.27675194e-02
8.56734931e-01 7.95680106e-01 1.12858367e+00 5.94519675e-01
4.98918205e-01 6.26029491e-01 3.97855788e-01 2.17647031e-01
-1.65487275e-01 -7.35277891e-01 2.15912282e-01 5.03558993e-01
-2.27342442e-01 -6.73594356e-01 -9.90351677e-01 6.74765289e-01
-2.25466442e+00 -1.11953688e+00 -7.75591850e-01 1.74260962e+00
1.21787512e+00 1.91640809e-01 -2.04321101e-01 2.02645317e-01
6.29351914e-01 2.18204316e-02 -2.46147111e-01 -1.14267252e-01
1.57293126e-01 2.12495998e-01 -1.23011805e-02 6.10617757e-01
-1.14717448e+00 1.05244458e+00 5.42752934e+00 8.89518917e-01
-6.61318779e-01 3.02050352e-01 8.91091824e-01 8.73449147e-02
-5.60101628e-01 3.46672148e-01 -1.07870090e+00 4.58813727e-01
9.85282063e-01 -3.51710141e-01 -2.28471700e-02 4.68279034e-01
5.54791927e-01 -1.39431059e-01 -1.42544270e+00 4.36656833e-01
-3.63962173e-01 -1.20633650e+00 2.10359380e-01 -1.83825538e-01
8.13056946e-01 -5.90242863e-01 -2.47511297e-01 2.07430109e-01
5.83889067e-01 -1.05123138e+00 7.38068283e-01 6.41946912e-01
4.23569858e-01 -8.38101864e-01 7.04918265e-01 3.16157728e-01
-1.41954005e+00 -2.10693851e-01 -1.22320570e-01 -1.22308671e-01
4.17757809e-01 7.83723712e-01 -1.03775656e+00 8.71392965e-01
5.03836691e-01 9.04304147e-01 -2.97277391e-01 6.81198597e-01
-1.18939495e+00 1.06725419e+00 1.86139420e-01 -2.71931112e-01
7.55010620e-02 6.60634562e-02 4.02970642e-01 1.38298035e+00
2.57502735e-01 4.80789304e-01 7.23062232e-02 1.04539192e+00
-5.76525293e-02 9.36487019e-02 -3.90926898e-01 6.66797012e-02
6.10920012e-01 1.07023454e+00 -6.42731309e-01 -3.13890129e-01
-1.70089990e-01 7.24597633e-01 5.69987595e-01 2.08053470e-01
-1.18264878e+00 -4.95104827e-02 2.93526620e-01 -2.88562357e-01
5.56981526e-02 4.25076298e-02 -6.59089804e-01 -8.21407378e-01
-3.18661332e-01 -4.65947568e-01 7.60424733e-01 -6.14756346e-01
-1.33963621e+00 4.31761265e-01 4.33029920e-01 -8.16305399e-01
-5.25166154e-01 -1.11382969e-01 -1.28977108e+00 1.00708902e+00
-1.61664093e+00 -1.26699138e+00 -2.37338501e-03 5.18819451e-01
6.91344440e-01 1.21046200e-01 7.19151080e-01 1.38639495e-01
-7.98356533e-01 3.09718996e-01 -5.82252562e-01 1.09661492e-02
6.65635288e-01 -1.44256234e+00 3.48085523e-01 1.09196246e+00
9.60672572e-02 4.16320235e-01 7.30601430e-01 -1.16795218e+00
-7.83211052e-01 -1.41390157e+00 1.51916051e+00 -2.94311106e-01
7.64059126e-01 -2.15711325e-01 -7.45809376e-01 6.89885557e-01
3.88248980e-01 -5.32572448e-01 7.06313610e-01 4.68232393e-01
-8.78465697e-02 1.04377739e-01 -5.33676028e-01 5.17875493e-01
1.21088243e+00 -3.29549871e-02 -7.45192528e-01 4.62643057e-01
1.07454205e+00 -6.15341328e-02 -6.69452012e-01 3.27552497e-01
1.62193567e-01 -3.69132459e-01 9.11563218e-01 -7.45017111e-01
1.29410362e+00 -1.95120901e-01 2.31765971e-01 -1.42224169e+00
-4.71120596e-01 -4.00080919e-01 -3.45264971e-01 1.84003150e+00
8.09292495e-01 -1.91024244e-01 2.69063175e-01 5.26752412e-01
-3.82861793e-02 -7.08963752e-01 -7.40794241e-01 -3.61508757e-01
-8.65440369e-02 -5.81604660e-01 6.11507058e-01 8.08559120e-01
2.28293985e-01 1.14718664e+00 -5.23121834e-01 3.67126375e-01
6.33572996e-01 4.16693002e-01 2.29322225e-01 -1.27459478e+00
-3.72393936e-01 -2.39722773e-01 4.16964948e-01 -6.67283595e-01
4.83537853e-01 -9.71710980e-01 3.29013377e-01 -1.69794774e+00
4.84977722e-01 -4.91591543e-01 -4.04797465e-01 7.89159715e-01
-9.70690310e-01 -4.58903372e-01 -4.31605289e-03 5.44271655e-02
-2.83872724e-01 9.24515128e-01 1.16802216e+00 -1.50449798e-01
-2.45706350e-01 2.16962900e-02 -7.55593896e-01 6.54704392e-01
7.01373398e-01 -9.14598227e-01 -9.23927665e-01 -4.12724167e-01
3.80950183e-01 6.20060921e-01 4.80659813e-01 -5.81477880e-01
1.63430601e-01 -6.26674354e-01 1.80033669e-01 -7.31926858e-01
-8.75631049e-02 -3.76572490e-01 1.12201519e-01 2.65468746e-01
-8.24955463e-01 -5.12919314e-02 -1.63025543e-01 7.89755702e-01
-2.16800645e-01 -4.54155982e-01 3.83037299e-01 1.72214210e-02
-5.98402977e-01 1.68205112e-01 -2.53097147e-01 1.93606466e-01
1.01952851e+00 6.48961663e-01 -3.38857502e-01 -1.07971586e-01
-6.82873428e-01 4.24221605e-01 -7.71829367e-01 5.84284186e-01
8.27168763e-01 -1.46787155e+00 -1.13467336e+00 -3.69163841e-01
-6.93057664e-03 -1.15315085e-02 1.12025909e-01 5.46610117e-01
2.54655927e-01 5.25889516e-01 2.28715748e-01 -5.51899076e-02
-1.22097874e+00 6.35355771e-01 -2.24131748e-01 -9.96981740e-01
-3.24007958e-01 1.06526804e+00 4.92917687e-01 1.53574556e-01
1.96554706e-01 -2.69405663e-01 -7.85551488e-01 3.45129758e-01
5.30977428e-01 2.21887395e-01 -2.19681904e-01 -1.29229128e-01
-4.52224910e-01 -1.19614229e-01 -2.44232342e-02 -3.34802657e-01
1.45941818e+00 2.86193788e-02 -2.91711062e-01 4.39801306e-01
9.70356464e-01 -4.97665644e-01 -1.09008014e+00 -1.99151769e-01
3.11468273e-01 -1.95434868e-01 1.33700609e-01 -1.20982349e+00
-8.03488255e-01 9.00970936e-01 2.26161301e-01 1.24643873e-02
1.07839119e+00 2.90245682e-01 7.22271860e-01 -1.84712052e-01
1.44575611e-01 -8.27135563e-01 3.53934556e-01 5.42285860e-01
1.32209253e+00 -9.68549907e-01 -2.01326177e-01 -7.51599550e-01
-7.36872196e-01 9.88042772e-01 7.66751766e-01 5.21351025e-02
6.71127558e-01 2.45645285e-01 -4.44489568e-01 -2.55082995e-01
-1.19716048e+00 -2.07910493e-01 4.32917267e-01 3.42781156e-01
8.88813198e-01 2.11103633e-01 -1.04830408e+00 1.09307933e+00
-5.45102768e-02 -7.03683272e-02 3.62362653e-01 4.93985265e-01
-1.06231391e-01 -1.30890226e+00 -2.02643886e-01 3.54899317e-01
-3.70755225e-01 -4.84594494e-01 -3.18431020e-01 3.61353844e-01
3.04584324e-01 1.35146737e+00 -1.22997738e-01 -3.31679553e-01
1.56538814e-01 2.45223776e-01 1.73229486e-01 -6.44802928e-01
-6.04405880e-01 1.04075216e-01 3.60538244e-01 -4.27150488e-01
-4.91715163e-01 -8.60739589e-01 -1.66492236e+00 2.35218868e-01
-2.32863292e-01 3.25750470e-01 3.06014299e-01 1.29798877e+00
6.50082231e-02 1.13604343e+00 7.29214072e-01 -2.48056546e-01
-2.04748526e-01 -1.15024877e+00 -2.86064595e-01 3.46521944e-01
3.12086614e-03 -8.81029010e-01 6.16599843e-02 3.42324138e-01]
|
[9.120917320251465, 9.075896263122559]
|
a0392c17-4291-4033-a847-cfa958d94d4f
|
timestamping-documents-and-beliefs
|
2106.14622
| null |
https://arxiv.org/abs/2106.14622v1
|
https://arxiv.org/pdf/2106.14622v1.pdf
|
Timestamping Documents and Beliefs
|
Most of the textual information available to us are temporally variable. In a world where information is dynamic, time-stamping them is a very important task. Documents are a good source of information and are used for many tasks like, sentiment analysis, classification of reviews etc. The knowledge of creation date of documents facilitates several tasks like summarization, event extraction, temporally focused information extraction etc. Unfortunately, for most of the documents on the web, the time-stamp meta-data is either erroneous or missing. Thus document dating is a challenging problem which requires inference over the temporal structure of the document alongside the contextual information of the document. Prior document dating systems have largely relied on handcrafted features while ignoring such document-internal structures. In this paper we propose NeuralDater, a Graph Convolutional Network (GCN) based document dating approach which jointly exploits syntactic and temporal graph structures of document in a principled way. We also pointed out some limitations of NeuralDater and tried to utilize both context and temporal information in documents in a more flexible and intuitive manner proposing AD3: Attentive Deep Document Dater, an attention-based document dating system. To the best of our knowledge these are the first application of deep learning methods for the task. Through extensive experiments on real-world datasets, we find that our models significantly outperforms state-of-the-art baselines by a significant margin.
|
['Swayambhu Nath Ray']
|
2021-06-09
| null | null | null | null |
['document-dating']
|
['natural-language-processing']
|
[-1.72916338e-01 -4.85977903e-02 -5.20404339e-01 -5.70334792e-01
-3.35497051e-01 -6.61824048e-01 1.30572987e+00 7.06647635e-01
-4.15351361e-01 5.79691350e-01 4.18398976e-01 -3.43326539e-01
-2.52971381e-01 -9.89219904e-01 -8.09032261e-01 -3.15834165e-01
-1.73118398e-01 6.76666021e-01 3.20726871e-01 -3.77341986e-01
5.82376719e-01 4.26481515e-01 -1.28414059e+00 4.07325000e-01
3.07593405e-01 9.36074495e-01 1.56067461e-01 8.52270901e-01
-8.72867584e-01 8.23605239e-01 -7.64067769e-01 -5.92186511e-01
-1.32527828e-01 -1.42905176e-01 -1.01433802e+00 -3.93369049e-02
3.43431205e-01 -3.41347545e-01 -5.60202062e-01 7.81667948e-01
7.36765563e-02 2.84977645e-01 5.71055233e-01 -9.77102280e-01
-8.18486452e-01 1.23206711e+00 -9.79372501e-01 6.43859744e-01
2.69892514e-01 -4.99653876e-01 1.15794027e+00 -5.16191244e-01
1.04169786e+00 1.19152617e+00 6.72613859e-01 3.35518867e-01
-5.20766199e-01 -3.02561820e-01 6.22451007e-01 2.82152593e-01
-8.22465599e-01 -1.47319570e-01 1.06719482e+00 -1.70132622e-01
1.13772273e+00 7.63991550e-02 5.44249713e-01 1.47340512e+00
6.53584123e-01 7.53561080e-01 4.86579120e-01 -4.60265338e-01
9.70538333e-02 -2.22487152e-01 5.19916177e-01 6.37541652e-01
2.52172381e-01 -3.84497344e-01 -9.16046917e-01 1.23597465e-01
2.83760518e-01 1.49971038e-01 -6.28669783e-02 1.33955002e-01
-9.96614814e-01 8.81179988e-01 4.66367394e-01 7.41283774e-01
-2.67572701e-01 5.49732387e-01 7.64609873e-01 1.23378657e-01
9.29159224e-01 7.08737597e-02 -5.79504490e-01 -2.03343615e-01
-1.15530872e+00 1.92689240e-01 7.43442059e-01 8.63060236e-01
3.52007389e-01 -3.27840783e-02 -8.44929516e-02 6.98372424e-01
2.85498828e-01 1.00569017e-02 8.32495391e-01 -3.90157610e-01
6.82659864e-01 6.57467127e-01 -5.64046111e-03 -1.13495028e+00
-6.38249576e-01 -2.38551751e-01 -7.16146767e-01 -4.21966940e-01
3.26638579e-01 -2.70317756e-02 -1.42585444e+00 1.51924193e+00
4.59675193e-01 1.08477119e-02 -3.40049565e-01 7.18073785e-01
1.17385697e+00 1.10708356e+00 -7.50178769e-02 -3.27579290e-01
1.36751199e+00 -6.67774618e-01 -1.12655365e+00 -2.20347255e-01
3.89927119e-01 -6.91291213e-01 7.46210933e-01 4.39036816e-01
-7.65921950e-01 -1.77822649e-01 -1.11654675e+00 -5.53729773e-01
-9.33838427e-01 -1.43580779e-01 1.18604803e+00 3.72694403e-01
-9.39955950e-01 1.02777839e+00 -8.82611454e-01 -7.29191542e-01
3.35381955e-01 3.03538591e-01 -4.54696983e-01 3.52648050e-02
-1.25557852e+00 5.92865646e-01 5.91353595e-01 3.52026641e-01
-5.36426783e-01 -3.98420721e-01 -8.56907248e-01 -1.11643020e-02
5.50640464e-01 -5.51572084e-01 1.63741219e+00 -6.03760064e-01
-1.21793985e+00 7.43905544e-01 -2.09176928e-01 -8.06288540e-01
5.72267711e-01 -1.96740180e-01 -6.24103248e-01 -3.62629779e-02
1.68538645e-01 1.58411622e-01 9.37139332e-01 -7.64775395e-01
-6.74521267e-01 -6.48591399e-01 4.61141542e-02 -2.03913391e-01
-4.44832951e-01 1.39013126e-01 -8.80332232e-01 -7.95637906e-01
-3.44003923e-02 -7.43430376e-01 7.92011898e-03 -3.83887321e-01
-5.77740371e-01 -5.35084724e-01 1.16534150e+00 -8.02919090e-01
1.52893841e+00 -1.72549772e+00 -2.34374050e-02 -1.10664338e-01
2.41782025e-01 -7.65029788e-02 2.20095113e-01 9.25967932e-01
1.54872611e-01 3.16270858e-01 -1.82782292e-01 -6.38594449e-01
-7.90480245e-03 2.31276006e-01 -6.44826889e-01 4.78424877e-01
-1.00560986e-01 8.41215909e-01 -1.03610861e+00 -5.46720862e-01
-5.66771068e-02 4.87039804e-01 1.15422018e-01 -2.41488904e-01
-7.10360825e-01 4.57900092e-02 -5.78863084e-01 6.01911008e-01
2.01927811e-01 -2.21364856e-01 1.58776984e-01 -5.09660132e-02
-2.52394855e-01 4.13379133e-01 -8.86731923e-01 2.05706429e+00
-5.83064377e-01 1.11647129e+00 -3.67849171e-01 -8.37997258e-01
8.39762807e-01 3.19369346e-01 5.25328636e-01 -7.78731167e-01
3.23564917e-01 8.29378963e-02 -2.69910365e-01 -2.26727039e-01
1.48158216e+00 2.82726884e-01 -3.21766615e-01 5.84477425e-01
2.14437380e-01 -6.11227117e-02 6.43633723e-01 8.28877687e-01
1.14372849e+00 6.50678873e-02 1.91745460e-01 -1.69337913e-02
2.20220387e-01 -5.27157225e-02 3.67550492e-01 6.80688143e-01
1.73187768e-03 7.18997478e-01 7.66431272e-01 -6.35166705e-01
-9.95812654e-01 -4.78866637e-01 -6.02756925e-02 9.20371711e-01
-1.41968966e-01 -8.00450325e-01 -4.10540432e-01 -8.93052280e-01
1.13403894e-01 7.82007396e-01 -1.16116965e+00 2.24299310e-03
-6.18075907e-01 -5.73766649e-01 1.76231816e-01 7.39561141e-01
2.34203547e-01 -1.05210042e+00 -2.58387357e-01 5.36524415e-01
-7.03813322e-03 -1.06181979e+00 -5.37660956e-01 1.65957734e-01
-9.19953942e-01 -1.02180243e+00 -5.34860015e-01 -3.61847997e-01
3.29857230e-01 3.00041020e-01 1.04685128e+00 -7.90377706e-02
-3.14159989e-01 5.21812141e-01 -4.46250767e-01 -9.28909004e-01
-6.54267669e-02 4.00959879e-01 -1.43192202e-01 1.01609744e-01
2.94580102e-01 -5.09264052e-01 -4.57472235e-01 -3.97754669e-01
-9.91977572e-01 -2.63738394e-01 7.92090446e-02 5.98317564e-01
3.97201538e-01 3.74369562e-01 5.17009676e-01 -1.50223100e+00
9.77581441e-01 -6.73395574e-01 -7.39858508e-01 1.51659369e-01
-7.28124797e-01 2.52621353e-01 6.06870174e-01 -2.81192571e-01
-1.19043636e+00 -2.76967168e-01 2.59723216e-01 -9.59001184e-02
4.02308494e-01 9.12998855e-01 1.55210942e-01 6.75899744e-01
4.27184939e-01 -5.80282025e-02 -6.46381140e-01 -5.84643543e-01
5.21587074e-01 5.76854765e-01 7.35366881e-01 -5.19955456e-01
4.34572041e-01 7.60978043e-01 6.70501366e-02 -7.68093765e-01
-1.06729329e+00 -7.17175484e-01 -6.75078154e-01 -3.42191637e-01
6.76706910e-01 -5.03389776e-01 -4.34369624e-01 5.30062079e-01
-1.51474202e+00 -5.85091747e-02 -1.14690885e-01 1.56861678e-01
-2.38687415e-02 3.55610579e-01 -6.59354150e-01 -6.76037550e-01
-6.95193768e-01 -5.49075842e-01 1.20705867e+00 2.92781949e-01
-1.96773678e-01 -1.25480247e+00 2.20747426e-01 9.07704607e-02
2.98225820e-01 5.34314334e-01 7.27848649e-01 -8.27663958e-01
-3.85012060e-01 -4.99011070e-01 -1.45775735e-01 -3.70325208e-01
3.29768926e-01 5.83917260e-01 -8.92356813e-01 2.77848542e-02
-2.78288484e-01 7.54826665e-02 1.43160093e+00 5.89816988e-01
1.21264517e+00 -4.18408871e-01 -4.77231503e-01 2.14614168e-01
1.38827789e+00 1.63569093e-01 3.48866820e-01 4.86192167e-01
9.77131367e-01 5.45504510e-01 5.52158535e-01 8.76296878e-01
5.72636485e-01 3.30180734e-01 6.88036084e-01 4.72889632e-01
-3.71376835e-02 -1.47347912e-01 1.19915605e-01 8.69165361e-01
-9.48677361e-02 -9.13402855e-01 -9.67510819e-01 1.03287292e+00
-2.09607768e+00 -1.07184494e+00 -4.62474227e-01 1.85612345e+00
8.12130094e-01 5.44326603e-01 -1.33917749e-01 2.71965802e-01
8.52918863e-01 7.75594711e-01 -3.58952761e-01 -8.76897514e-01
9.26758572e-02 1.27509147e-01 5.31870127e-01 2.14974612e-01
-1.31059420e+00 8.59503686e-01 5.28442955e+00 4.20866519e-01
-1.20860100e+00 -1.40165612e-01 2.46104345e-01 -1.26997948e-01
-2.69019574e-01 4.00945581e-02 -9.16516185e-01 6.08465672e-01
1.27857113e+00 -4.25719827e-01 2.53935188e-01 6.85342133e-01
1.74596682e-01 -1.94048911e-01 -1.32142353e+00 8.10469925e-01
-1.94751378e-02 -1.65120077e+00 -1.56606622e-02 -3.63784879e-02
7.18769133e-01 9.05550346e-02 -5.39563745e-02 1.39388964e-01
4.77405369e-01 -6.57667398e-01 9.92625415e-01 5.07927060e-01
3.44741911e-01 -9.11367714e-01 7.60343075e-01 1.22106457e-02
-1.33791482e+00 1.57213420e-01 -2.18890145e-01 2.14453205e-01
2.62146473e-01 1.12045121e+00 -9.53148901e-01 8.90543997e-01
8.87693226e-01 1.23229849e+00 -5.77232063e-01 9.07945454e-01
-3.53211105e-01 5.76881289e-01 -2.76727438e-01 -2.27950230e-01
5.98397493e-01 6.18565306e-02 4.63567555e-01 1.30744696e+00
3.16851974e-01 -1.68829128e-01 -2.31348649e-01 3.82492214e-01
-6.56122148e-01 1.35178333e-02 -8.26633334e-01 -5.64587355e-01
3.65477681e-01 1.23191726e+00 -1.17544901e+00 -2.03581795e-01
-3.84032249e-01 8.25706124e-01 3.17464381e-01 1.25974774e-01
-7.11615145e-01 -5.86097360e-01 1.31346062e-01 -6.48663044e-02
5.48598945e-01 -5.56160986e-01 -3.82892378e-02 -1.12336385e+00
1.90150291e-01 -2.70457476e-01 8.77868652e-01 -7.85538316e-01
-1.10737944e+00 5.16159654e-01 -1.27783433e-01 -9.01387513e-01
-5.38076162e-01 -4.06724066e-01 -7.04740763e-01 3.75998378e-01
-1.54266441e+00 -1.21378851e+00 -1.75484583e-01 5.86815357e-01
9.02185678e-01 3.22117768e-02 3.49819034e-01 1.31572559e-01
-6.36811376e-01 1.29604697e-01 3.48050863e-01 2.98036307e-01
9.20429289e-01 -1.64911699e+00 7.09311068e-01 1.11725998e+00
5.53210139e-01 6.97102010e-01 8.56759489e-01 -8.45179141e-01
-1.81616831e+00 -1.22336173e+00 1.17606068e+00 -3.91945362e-01
1.10985136e+00 -3.43649983e-01 -8.88091385e-01 9.64870274e-01
5.79318643e-01 -1.22831866e-01 2.31933326e-01 3.47337663e-01
-4.15730894e-01 -2.86753416e-01 -6.63494945e-01 5.02012610e-01
7.80383766e-01 -5.43160558e-01 -6.61019146e-01 6.79428875e-01
9.37092423e-01 -6.16507530e-01 -6.74895108e-01 -7.01753870e-02
3.55297744e-01 -6.05215788e-01 5.51752031e-01 -5.09817302e-01
7.28452981e-01 -2.06475332e-03 2.16218323e-01 -1.18002546e+00
-2.12730300e-02 -7.40630448e-01 -8.24321032e-01 1.87550914e+00
2.29324877e-01 -3.88704449e-01 6.26207829e-01 5.31815708e-01
-3.26328367e-01 -3.51237565e-01 -7.89979160e-01 -7.51809299e-01
-1.86298117e-01 -6.62300289e-01 7.30362415e-01 1.05329061e+00
-1.15356855e-01 3.52618188e-01 -4.29139555e-01 1.39645785e-02
3.53392661e-01 2.83930719e-01 4.28830296e-01 -1.52171934e+00
9.50363874e-02 -5.14944077e-01 -1.79493919e-01 -3.52883458e-01
2.55024612e-01 -7.51831472e-01 -1.37328878e-01 -2.23518252e+00
-6.53930679e-02 1.95384137e-02 -2.27634847e-01 4.67027009e-01
2.24622935e-01 -2.87680358e-01 -4.71908711e-02 6.61610812e-02
-6.70750320e-01 3.17833036e-01 9.57234144e-01 -5.89556336e-01
-2.05523238e-01 -1.15935812e-02 -5.68841457e-01 5.57358801e-01
7.91289926e-01 -6.77326441e-01 -5.11056781e-01 -6.51098788e-01
5.16356409e-01 5.65165021e-02 -1.53583854e-01 -4.03372556e-01
6.35549188e-01 -2.15199813e-01 3.65328789e-01 -1.16254318e+00
2.78217614e-01 -7.69419730e-01 -4.98825423e-02 1.17668502e-01
-9.09702554e-02 2.66962349e-01 2.78057665e-01 1.03325438e+00
-2.59332716e-01 -3.09490919e-01 2.04158783e-01 -4.46070880e-01
-9.94995773e-01 5.52740514e-01 -2.37732530e-01 -8.44714195e-02
6.92942262e-01 4.33212295e-02 -6.87760115e-01 -3.46037835e-01
-2.97774464e-01 2.41114765e-01 1.30548820e-01 8.44777763e-01
3.72273743e-01 -1.05645573e+00 -5.06036580e-01 -5.49587131e-01
8.95619690e-02 1.44706383e-01 5.25151715e-02 4.22110438e-01
-5.08790135e-01 7.13657677e-01 1.44943699e-01 -4.35957044e-01
-1.08634388e+00 7.87805080e-01 -1.47283450e-01 -5.39826214e-01
-8.25935006e-01 6.98992729e-01 -2.74860412e-01 1.40735328e-01
3.49173903e-01 -7.67656624e-01 -5.74477732e-01 7.93995082e-01
5.62648892e-01 2.10612178e-01 6.46475792e-01 -2.72210836e-01
-4.71893847e-01 4.11957711e-01 -7.47019947e-01 -1.78592369e-01
1.80208850e+00 -9.65306163e-02 -3.20177764e-01 6.52161181e-01
1.06466269e+00 4.11065854e-02 -9.52093482e-01 -4.44994658e-01
4.61336821e-01 -2.80974150e-01 3.50478262e-01 -5.37799358e-01
-1.25766563e+00 7.50324368e-01 -1.28482297e-01 6.07235909e-01
9.50062513e-01 4.88549396e-02 8.50565732e-01 4.29924518e-01
2.86370814e-01 -1.49691713e+00 9.87376422e-02 7.11713910e-01
9.61937964e-01 -1.28157902e+00 5.36275625e-01 1.63678825e-01
-4.25488830e-01 1.48419428e+00 3.88766587e-01 1.67588770e-01
5.88697195e-01 9.29770842e-02 -3.29616636e-01 -4.24259543e-01
-1.01963651e+00 -1.47727579e-01 4.14403468e-01 2.35898763e-01
5.13489187e-01 -3.16696376e-01 -6.06171370e-01 7.23621130e-01
-2.37028733e-01 -1.69547409e-01 9.47227597e-01 1.26051712e+00
-2.76424259e-01 -1.14984214e+00 -2.02511922e-01 4.56612527e-01
-8.59470665e-01 1.21810683e-03 -7.22398520e-01 9.99624431e-01
-4.83052313e-01 8.43338490e-01 -3.26245911e-02 -4.90519553e-02
8.38272497e-02 5.55199161e-02 4.11531895e-01 -5.76969564e-01
-7.49271870e-01 -1.37183353e-01 3.51073772e-01 -3.24128240e-01
-4.04379427e-01 -7.90786982e-01 -1.52243531e+00 -4.59565848e-01
-2.82477915e-01 -8.38804021e-02 1.27104867e+00 9.31419313e-01
2.16381684e-01 9.32318091e-01 5.82978487e-01 -7.79693663e-01
-3.73053551e-02 -8.58565032e-01 -6.57105505e-01 3.55182104e-02
4.85697955e-01 -4.96151745e-01 -7.17567354e-02 2.01820150e-01]
|
[12.21107292175293, 9.347225189208984]
|
512cafff-63e3-4317-96e8-65a63fd2c8a0
|
a-large-dataset-to-train-convolutional
|
1512.02134
| null |
http://arxiv.org/abs/1512.02134v1
|
http://arxiv.org/pdf/1512.02134v1.pdf
|
A Large Dataset to Train Convolutional Networks for Disparity, Optical Flow, and Scene Flow Estimation
|
Recent work has shown that optical flow estimation can be formulated as a
supervised learning task and can be successfully solved with convolutional
networks. Training of the so-called FlowNet was enabled by a large
synthetically generated dataset. The present paper extends the concept of
optical flow estimation via convolutional networks to disparity and scene flow
estimation. To this end, we propose three synthetic stereo video datasets with
sufficient realism, variation, and size to successfully train large networks.
Our datasets are the first large-scale datasets to enable training and
evaluating scene flow methods. Besides the datasets, we present a convolutional
network for real-time disparity estimation that provides state-of-the-art
results. By combining a flow and disparity estimation network and training it
jointly, we demonstrate the first scene flow estimation with a convolutional
network.
|
['Philip Häusser', 'Thomas Brox', 'Nikolaus Mayer', 'Alexey Dosovitskiy', 'Philipp Fischer', 'Eddy Ilg', 'Daniel Cremers']
|
2015-12-07
|
a-large-dataset-to-train-convolutional-1
|
http://openaccess.thecvf.com/content_cvpr_2016/html/Mayer_A_Large_Dataset_CVPR_2016_paper.html
|
http://openaccess.thecvf.com/content_cvpr_2016/papers/Mayer_A_Large_Dataset_CVPR_2016_paper.pdf
|
cvpr-2016-6
|
['scene-flow-estimation']
|
['computer-vision']
|
[ 0.25646833 -0.27890474 0.03063365 -0.37212718 -0.15884508 -0.4949053
0.5940998 -0.6686021 -0.3995022 1.0092087 0.08679529 -0.21955419
0.3988814 -0.8026942 -0.91988677 -0.18070114 -0.22527999 0.08991061
0.38235542 -0.10608427 0.4046022 0.56988037 -1.8412501 0.32052934
0.8542095 1.0403056 0.17556891 1.1857479 -0.04242019 1.5626091
-0.49911946 -0.14250818 0.6312013 -0.3285461 -1.0148561 0.11276418
1.4699032 -1.1241654 -0.6891566 0.5686537 0.5202333 0.19087982
0.31303367 -1.4244874 -0.13301638 0.08799656 -0.18745454 0.38234144
0.4817354 0.6431721 0.7013879 -0.74382085 0.9247487 1.2964176
0.55387396 0.68368655 -1.234341 -0.727819 -0.19914983 0.1406728
-0.8999491 -0.65451974 0.7873115 -0.77982134 1.1654298 -0.20553541
0.81797826 1.1625694 -0.08743335 0.8681845 0.8884521 -0.18041448
0.13002439 -0.30698702 -0.41029647 0.873891 0.19287851 0.62987417
-0.5192687 0.49168292 1.1741298 -0.48409525 -0.4823524 -0.43245837
-1.4964749 0.68809664 0.6850174 -0.27138513 -0.03043252 0.7759932
0.5784926 0.19446664 0.39328453 0.4149578 -0.42304397 -0.4551549
-1.1174178 0.32523632 0.93741757 0.9197968 1.113595 0.42839402
-0.04761979 0.4929797 0.0628143 0.6174431 0.24357435 -1.8474269
0.5864387 0.20070387 0.32574943 -0.91714466 -0.34328687 -0.0455735
-0.8417245 0.7566004 0.8013376 -0.42042983 -0.6337312 1.6731184
0.0396448 0.94972605 0.046556 0.96125484 0.97871685 0.4904878
-0.13436204 0.08982955 0.6545155 -1.2995548 -0.4600517 -0.30724746
0.708285 -0.8046089 0.7281112 0.3109633 -1.2506889 -0.96551263
-1.0803112 -0.28051397 -0.06363624 -0.1084742 1.061871 0.5696128
-1.4444612 0.69389606 -0.6498776 -0.2268291 0.7839511 0.2335058
-0.47298878 -0.20355752 -1.0929393 0.68743765 0.3917292 0.20357557
-1.1472957 -1.076555 -1.1961135 -0.16109002 -0.06263261 -1.2552892
1.0983112 -1.0293272 -1.7205127 0.8493566 -0.10608099 -0.5483941
0.9099504 -0.23686056 -0.11798532 0.56122804 0.14816017 1.3303295
0.7466701 -1.1236293 -0.7918561 0.15860902 0.47164908 -0.20726383
0.10738055 -0.28672236 -0.16941726 -0.36187485 -0.40406474 -0.749169
-0.17401426 0.4947734 -0.3821518 0.41090575 0.6993919 -0.2628524
0.780222 -1.7771664 -0.14404616 -0.4350143 0.34526384 0.49028024
-0.46600822 0.0234929 -0.03621808 -0.14006883 -0.21807094 -0.48169023
-0.2901968 0.27512676 -0.36183757 0.47823012 0.38835153 1.2148843
-1.2417942 -0.56222486 0.93932915 0.67686474 -1.0455256 0.44430697
-0.12470857 0.86597985 -0.0201088 0.3316828 0.95940536 -0.26484218
-0.1685533 -0.42612135 -0.18847752 0.16920075 -1.203032 2.0303636
-0.72738034 1.4845088 -0.17020944 -0.8130764 0.8089809 0.0186366
0.8486682 -0.7751287 0.18336672 0.25527525 -0.17046233 -0.67364764
0.6534838 0.12541133 0.39786902 0.40997314 0.430725 -0.6575505
0.6142665 0.13454305 0.9623131 0.5346068 -0.13909952 -0.2221446
0.8841948 -0.04148395 0.39336255 0.7225028 -0.58827287 0.81460536
0.3942097 -1.0150105 -1.2198533 -1.1572783 -0.24709329 0.622222
0.3547984 -0.11362743 -0.5625619 -0.5786849 0.15629932 0.07455165
-0.7235495 0.21504897 -0.9020249 -0.20910153 0.69345796 0.604926
1.0305957 -1.2667124 -0.82352436 0.25307217 -0.27337557 -1.8585148
-0.20483921 -0.29359126 -0.81518835 -1.5917641 -0.6748198 -0.850913
0.38911653 0.20403741 1.5536159 0.04140569 -0.40001777 0.28171486
0.0520751 -0.02117662 -0.41986257 0.0119634 -0.29216963 -0.079617
-0.03947066 -0.721482 -1.1654522 0.27949205 -0.91161877 0.3278601
-0.01977951 0.7176891 0.02585111 -0.6101266 0.19523239 -0.66436374
0.16660187 -0.03871775 -0.8844246 -0.19199519 -0.0285415 0.06830131
0.65145576 -0.15864906 -1.1966718 0.21138307 -0.2290667 -0.46630034
-0.31458527 -0.12655428 0.47281364 -0.53131163 0.8059295 -0.04205986
0.04689528 0.07463573 0.52676237 0.26484072 0.85289204 -0.43951657
0.6702624 1.0453179 0.44368958 -0.6320331 -0.866802 -0.3434216
-1.0154204 -0.54734814 0.83550084 -1.1707826 -1.1570536 0.81196547
-1.3851384 -0.80258375 -0.44377294 0.81448555 -1.0657057 0.32795596
-0.8221449 -0.39499658 -0.0650939 -1.20163 1.1248698 0.09099568
0.02155584 -1.4525243 0.4064486 0.3207786 0.7547724 0.5924571
-0.01562734 0.37533352 -1.1142354 0.3275425 -0.7309944 0.5064021
0.15106869 0.3818938 -1.41485 -0.25352335 -0.34465057 -0.7423023
1.1541741 0.76484174 1.139269 0.22329563 0.03024208 1.4619178
1.5729752 -0.17024408 1.0143116 0.39640138 1.0054185 0.52217466
0.16847718 0.41510767 0.45389 0.51561755 0.59307003 -0.26015618
-0.6414721 -0.12738611 0.11533999 0.4403991 -0.2421155 -0.20015071
-0.8007872 0.51119554 -1.6073527 -1.1845555 -0.2041768 1.8733892
0.5677062 0.11658753 -0.131965 0.01095541 0.46783614 0.488592
-0.3714898 -0.48017028 -0.19830526 0.3646487 0.43456647 0.8618266
-1.3547751 1.29749 7.275622 0.25783876 -1.4699913 -0.34005463
0.5309091 -0.1489161 -0.15739444 -0.14435834 -0.5367597 0.3833274
0.68188435 0.01490057 0.3992108 0.5371601 0.46207544 -0.23980233
-1.1762018 1.1712053 -0.13609318 -1.8426433 0.04677225 -0.02111802
1.1296 0.36257356 -0.11184592 0.11785606 0.5033588 -1.1310589
0.36868963 0.35699838 1.1152269 -0.36550623 0.58594203 -0.06309138
-1.2169133 0.01814543 -0.39258778 -0.4413855 0.63056266 0.5813747
-0.47508192 0.41501972 0.532289 1.6162603 -0.6147782 1.2803317
-0.28099594 0.33710706 -0.12425748 0.43505782 0.3652073 -0.16995393
0.25847286 1.3465116 0.10986266 -0.4271695 0.05552413 1.0909574
-0.18600431 -0.22692211 -0.8159985 0.41698882 0.05043234 1.1171598
-0.33409128 -0.5349978 -0.5003206 0.8572998 0.46441525 0.5632618
-0.65371644 -0.309476 1.201499 -0.1169036 0.27949443 -0.3446023
-0.24482588 -1.6083874 -0.16940293 -0.19476184 -0.02271004 -1.0568793
-1.0876691 0.7341011 -0.212918 -1.4837303 -0.7539053 -0.8491588
-0.67330515 0.8213835 -2.3131187 -0.9551833 -1.0473723 0.80645764
0.39770606 -0.08785588 0.55279917 0.4817732 -0.34628597 0.17916256
-0.21369793 0.5423139 0.82216257 -1.1901028 0.7618764 0.89912003
-0.09494261 0.04124823 0.3766057 -0.1743948 -1.0921293 -1.0894186
0.7379255 -0.522365 0.66587764 -0.36168906 -0.5731327 0.6955536
0.32448208 0.90494955 0.07492531 -0.5289552 -0.28319988 -0.10826074
-1.0958064 0.4021535 1.4179628 -0.6717831 -0.1260524 0.13235697
0.6470552 -0.6434728 -0.6801379 0.66849566 0.7449529 -1.6454248
1.3595222 -0.4732025 1.14052 -0.19564432 0.08536255 -1.3781352
0.05478434 -0.67633843 -0.21148705 0.7220618 0.01123446 -0.5086766
1.1339217 0.4628553 -0.06983355 -0.19823071 -0.72351485 -0.6159773
0.10462732 -0.71114606 0.42016563 0.9732879 -0.46056703 0.03972919
-0.47653323 -0.27778673 0.80896896 0.15324338 1.2343489 -1.0683675
-0.02367713 -0.4756794 -0.9209796 -1.6579373 0.6516879 -0.61638826
0.10093404 -1.4211 -0.2944196 -0.305545 0.13977087 -0.09773697
-0.16002262 0.6817058 0.24837542 -0.03392576 -0.44706002 0.33413714
1.8980467 -0.13996269 -0.12634827 -0.18683368 -0.04424841 0.55448204
0.75905967 0.08922333 -0.6108169 -0.7078016 0.08614098 0.18738393
0.57411844 -1.2625198 0.31676435 -0.08294859 0.54066586 -0.409585
0.3072373 -0.5416362 -0.23765598 0.5386843 -0.42320946 -0.1555854
0.4724492 0.14848804 -0.4559484 0.22519436 0.78249305 -0.24767005
-1.3505483 0.5862153 -0.14585736 0.6361474 0.8418299 -0.39577344
-0.65924144 -0.46845743 -0.41194516 0.14889468 0.39318365 0.40938932
0.73457175 -1.2351992 -0.9183282 0.551359 0.05063522 0.24028774
0.15461315 0.39266804 -1.3805797 0.49155873 -0.8896707 -0.95276725
-0.7301578 0.45824635 0.7464134 0.04541622 -0.49888003 0.774231
0.3342001 -0.50165296 0.08458385 -0.5188353 -0.14551264 -0.30318987
0.5943997 0.400286 -0.22719829 -0.6298076 -0.2661771 0.8747555
0.6366677 -0.04611596 1.17114 -0.2734367 0.03166693 -0.03403232
1.6344013 -0.35207212 -2.0877655 0.0202812 -0.5102364 -1.0796088
0.13603146 -0.38836253 -1.6085018 1.2267232 0.4148308 -0.08319713
0.94123214 -0.45365417 0.8710001 0.33327454 0.40196767 -0.74751294
0.32197022 0.74106973 0.5442212 -1.836265 -0.22874069 -0.6914857
-0.39360315 1.3970177 0.9745339 -0.48936117 0.74485505 0.4875258
0.3443193 0.0526354 -0.74391663 -0.45272002 0.31456286 0.8759553
0.49905482 -0.34329093 0.32665822 -0.70883524 -0.3182063 0.7059112
0.84015566 0.6697496 -0.12477772 -0.94315726 0.18832237 0.04887542
-0.23150317 -0.12541491 0.04058727 0.8660289 -0.05920108 0.8999892
0.48585635 -0.3139618 0.3659175 -0.43521196 0.858961 -0.21677044
-0.48320583 -0.6035029 0.18427934 -1.1821846 -0.9834301 -0.04984287
-0.8846376 -0.7208999 0.2947894 -0.50110257 0.4156461 0.83043146
0.21797833 0.6002921 0.76073176 -1.3615901 0.22040644 -0.71265304
-0.3453776 0.86865693 0.81752646 -0.57468164 -0.6190813 0.3867204 ]
|
[8.747321128845215, -1.916761875152588]
|
a8c8ef71-1354-4460-96ff-ad7527ef31df
|
morphte-injecting-morphology-in-tensorized
|
2210.15379
| null |
https://arxiv.org/abs/2210.15379v1
|
https://arxiv.org/pdf/2210.15379v1.pdf
|
MorphTE: Injecting Morphology in Tensorized Embeddings
|
In the era of deep learning, word embeddings are essential when dealing with text tasks. However, storing and accessing these embeddings requires a large amount of space. This is not conducive to the deployment of these models on resource-limited devices. Combining the powerful compression capability of tensor products, we propose a word embedding compression method with morphological augmentation, Morphologically-enhanced Tensorized Embeddings (MorphTE). A word consists of one or more morphemes, the smallest units that bear meaning or have a grammatical function. MorphTE represents a word embedding as an entangled form of its morpheme vectors via the tensor product, which injects prior semantic and grammatical knowledge into the learning of embeddings. Furthermore, the dimensionality of the morpheme vector and the number of morphemes are much smaller than those of words, which greatly reduces the parameters of the word embeddings. We conduct experiments on tasks such as machine translation and question answering. Experimental results on four translation datasets of different languages show that MorphTE can compress word embedding parameters by about 20 times without performance loss and significantly outperforms related embedding compression methods.
|
['Benyou Wang', 'Xiuqing Lu', 'Sunzhu Li', 'Peng Zhang', 'Guobing Gan']
|
2022-10-27
| null | null | null | null |
['learning-word-embeddings']
|
['methodology']
|
[-1.82532027e-01 -1.83549136e-01 -2.99574226e-01 -4.38624993e-02
-1.68646395e-01 -4.50911701e-01 3.31259102e-01 4.17388856e-01
-8.12550902e-01 1.07320160e-01 3.62179518e-01 -7.96225548e-01
2.02972203e-01 -1.07407331e+00 -5.00094771e-01 -5.52023172e-01
-2.80035585e-01 3.42349380e-01 -3.54638398e-02 -3.08000058e-01
1.39655575e-01 2.04253763e-01 -1.26358104e+00 1.57923952e-01
8.81367326e-01 1.10535777e+00 2.96509027e-01 5.18218338e-01
-6.61456704e-01 7.05336258e-02 -4.14069265e-01 -8.21892917e-01
2.65682489e-01 7.85823688e-02 -7.17353225e-01 -8.21966380e-02
2.79844701e-01 -6.28776908e-01 -9.58560586e-01 1.13949084e+00
2.31787547e-01 1.06575549e-01 3.91379207e-01 -9.48989511e-01
-1.47760665e+00 5.01988888e-01 -8.31467733e-02 3.90373468e-01
-1.00392073e-01 -5.13802357e-02 1.48177767e+00 -1.20851433e+00
1.40592411e-01 1.28935254e+00 4.13426548e-01 2.50466108e-01
-9.49130416e-01 -4.02068675e-01 4.59107682e-02 5.10720253e-01
-1.39744711e+00 -1.50035143e-01 6.06300294e-01 -9.12629068e-02
1.52605212e+00 2.96196371e-01 7.48620212e-01 6.09373569e-01
2.28468716e-01 5.57693183e-01 4.52220291e-01 -4.06028777e-01
2.72420168e-01 1.92492664e-01 1.28171831e-01 8.85159850e-01
5.15228927e-01 -4.17729616e-01 -3.78550619e-01 -2.89819181e-01
5.18944800e-01 3.83316606e-01 -1.18694089e-01 -1.55984223e-01
-1.28546846e+00 1.04284585e+00 4.76503223e-01 2.59996086e-01
-3.27777028e-01 3.70911092e-01 6.59683228e-01 3.65948409e-01
5.24482906e-01 5.33376515e-01 -5.08274078e-01 -3.82224262e-01
-2.79819608e-01 -1.06596492e-01 5.98013937e-01 9.77809906e-01
1.00802374e+00 -7.02266991e-02 1.78246945e-01 9.06398416e-01
2.25855649e-01 5.46922326e-01 1.01867867e+00 -5.96080184e-01
7.79938281e-01 9.75315988e-01 -2.26438075e-01 -1.27625477e+00
1.38510913e-01 -1.07151911e-01 -7.28976130e-01 -7.59724557e-01
1.49538508e-02 1.36244353e-02 -6.98383033e-01 1.58354306e+00
4.10096258e-01 -9.83040221e-03 6.39331667e-03 6.78913951e-01
4.58936900e-01 1.06734693e+00 -1.03531830e-01 1.29369810e-01
1.94420743e+00 -1.05654657e+00 -8.74706030e-01 -5.15833616e-01
1.07629299e+00 -6.57245278e-01 1.34409773e+00 -3.02059706e-02
-8.90175521e-01 -3.35468650e-01 -1.19576645e+00 -6.06227160e-01
-7.45003879e-01 -1.42457351e-01 1.13636875e+00 5.85389018e-01
-6.02392256e-01 5.42910159e-01 -9.30443764e-01 4.59513515e-02
4.31262672e-01 3.68882388e-01 -6.01193547e-01 -4.76554066e-01
-1.31096852e+00 9.05297458e-01 4.07236934e-01 1.91115618e-01
-2.39808813e-01 -6.83186710e-01 -1.14554000e+00 3.79605770e-01
-2.00161695e-01 -4.03412223e-01 7.28705466e-01 -2.33563945e-01
-1.05590618e+00 3.94692898e-01 -3.48286420e-01 -2.70529449e-01
-3.04759890e-01 -1.50248110e-01 -4.17127162e-01 3.29852402e-01
-2.05196187e-01 5.37911057e-01 9.12576377e-01 -3.24990034e-01
-3.05467099e-01 -5.40030420e-01 3.15294623e-01 1.48687989e-01
-1.44133675e+00 -2.54813492e-01 -4.65241313e-01 -6.31619513e-01
2.87498057e-01 -7.99751878e-01 -9.99085754e-02 1.92116141e-01
-1.55931674e-02 -6.72857761e-01 9.13201034e-01 -7.56808162e-01
1.53624725e+00 -2.53608847e+00 3.11294317e-01 -2.10692734e-01
4.26656008e-01 4.53233242e-01 -4.29512650e-01 6.45485818e-01
-1.51875140e-02 4.40211028e-01 -9.25670043e-02 -3.13232034e-01
3.51641417e-01 8.12322855e-01 -6.09270394e-01 2.08348498e-01
4.35996681e-01 1.17898452e+00 -8.53004277e-01 -5.05569756e-01
5.15438505e-02 4.28624481e-01 -8.92147124e-01 1.76174253e-01
-1.48984790e-01 -6.20912611e-01 -5.21707118e-01 4.42650378e-01
7.23984003e-01 -1.62601903e-01 2.31302038e-01 -4.28025126e-01
1.64824471e-01 8.85131299e-01 -7.42735982e-01 1.75867569e+00
-7.70089269e-01 5.79491496e-01 -4.18028653e-01 -1.05178070e+00
7.92782784e-01 2.06801102e-01 1.68642014e-01 -7.13813901e-01
5.68767637e-02 3.58016372e-01 1.61148787e-01 -7.38889694e-01
8.21218312e-01 -2.51547903e-01 -1.63874760e-01 7.88033366e-01
2.46709257e-01 -1.29693691e-02 3.84182036e-01 3.24365586e-01
1.15324271e+00 -4.82216507e-01 -1.62025526e-01 -5.47552183e-02
2.39539042e-01 -3.17893863e-01 5.08580029e-01 -7.98095241e-02
-3.94149534e-02 1.89143289e-02 5.30731559e-01 -6.29717171e-01
-1.40385211e+00 -1.01805520e+00 -2.24519849e-01 1.05552077e+00
-1.99642181e-02 -9.97575879e-01 -6.05971098e-01 -4.10072207e-01
2.05707073e-01 4.96122062e-01 -5.45636833e-01 -5.33921957e-01
-8.28366339e-01 -9.12957907e-01 4.59325910e-01 6.90195799e-01
2.28233427e-01 -6.20608151e-01 -3.11550796e-01 2.56913483e-01
-1.98396221e-01 -1.26463532e+00 -9.61961091e-01 1.32622987e-01
-1.16690385e+00 -6.15045071e-01 -2.46837750e-01 -1.15856290e+00
9.92071629e-01 4.20772046e-01 6.83465898e-01 2.89239228e-01
-4.17878538e-01 -6.77954257e-02 -6.30019486e-01 -1.36697024e-01
8.59070104e-03 1.05264708e-01 2.82019466e-01 3.26497406e-02
9.65538323e-01 -9.07772243e-01 -6.23967350e-01 -1.00239785e-02
-1.35796702e+00 -1.57841489e-01 6.51806414e-01 1.05117381e+00
4.00870115e-01 1.05827563e-02 1.54330045e-01 -3.19166154e-01
8.48515511e-01 -3.60240191e-01 -3.55725348e-01 1.58077121e-01
-5.70190489e-01 4.76653665e-01 8.47131550e-01 -8.76469970e-01
-2.60416985e-01 -5.71032465e-01 -3.94224897e-02 -4.60881382e-01
5.87235212e-01 6.06011868e-01 -1.13853209e-01 2.34811500e-01
2.64506429e-01 4.49556768e-01 1.26036331e-01 -7.45836675e-01
6.29356027e-01 9.70981717e-01 7.74929672e-02 -5.33298135e-01
8.43951821e-01 3.32719147e-01 -1.53579518e-01 -9.55098510e-01
-3.71545702e-01 -1.98725447e-01 -5.69229424e-01 6.10316515e-01
7.25607455e-01 -6.77230179e-01 -4.41367090e-01 -5.62990643e-03
-1.39897478e+00 2.45763257e-01 -4.48015451e-01 7.06677020e-01
2.88580135e-02 7.87236631e-01 -9.14344847e-01 -3.25022191e-01
-4.92287457e-01 -1.13296437e+00 6.85931146e-01 -1.30231485e-01
-7.34560341e-02 -1.01984906e+00 -9.57609937e-02 3.04291010e-01
4.96466219e-01 -4.53355819e-01 1.82331395e+00 -5.36948442e-01
-5.89339375e-01 -6.17072999e-01 -2.13082135e-01 7.68701792e-01
6.75416142e-02 -3.67531776e-01 -4.29796427e-01 -3.41432989e-01
3.83926416e-03 -2.61565328e-01 6.78617358e-01 -3.82771224e-01
1.41934192e+00 -6.24154449e-01 -3.47296111e-02 7.17873454e-01
1.29824674e+00 -5.18320203e-02 5.43692112e-01 1.59965158e-01
8.71454060e-01 4.05123860e-01 2.11403981e-01 3.43018979e-01
5.25477469e-01 4.39329118e-01 3.96218091e-01 3.77400756e-01
-8.48574191e-02 -6.01916909e-01 4.66378838e-01 2.05727887e+00
2.01265842e-01 -1.87102556e-02 -6.31766319e-01 8.49566579e-01
-1.29651666e+00 -3.71153861e-01 6.27648607e-02 2.11153865e+00
8.84477794e-01 -5.23640066e-02 -5.24999917e-01 3.62941474e-01
4.35633749e-01 3.34140748e-01 -5.42367280e-01 -1.02962112e+00
2.15785861e-01 6.03199005e-01 5.38779318e-01 3.75495464e-01
-6.33635044e-01 9.00256336e-01 5.86868572e+00 8.26432049e-01
-8.51545513e-01 4.41927403e-01 1.06186353e-01 -5.93092330e-02
-7.30159044e-01 6.01063296e-02 -4.96184736e-01 6.16347015e-01
1.11044836e+00 -3.81188959e-01 6.82494283e-01 6.78164601e-01
-3.13616991e-01 3.95962059e-01 -1.20415890e+00 1.01403570e+00
5.68950437e-02 -1.24659717e+00 5.18784404e-01 3.84023190e-01
3.02705765e-01 4.73992340e-02 3.64597768e-01 4.46417391e-01
-2.14569017e-01 -1.02537799e+00 4.03047770e-01 -2.77590007e-01
9.32882547e-01 -7.97201812e-01 6.96091473e-01 2.19392270e-01
-1.22124791e+00 -1.43716216e-01 -1.28346980e+00 -2.26990163e-01
1.06993876e-01 5.73799372e-01 -5.96284568e-01 2.39582077e-01
4.50266689e-01 5.38757324e-01 -4.00487661e-01 4.23233539e-01
-2.60866672e-01 5.17730474e-01 -4.10708874e-01 -3.05475742e-01
4.30547237e-01 -4.83991623e-01 1.62389219e-01 1.02722943e+00
4.49600458e-01 2.62511104e-01 -2.11889207e-01 6.54078722e-01
-5.31804681e-01 3.27359527e-01 -4.44904387e-01 -8.82960558e-01
7.52675831e-01 1.24208033e+00 -1.20612957e-01 -4.58084106e-01
-5.71491420e-01 1.15219975e+00 5.60010910e-01 1.08265065e-01
-6.63147390e-01 -8.91971290e-01 1.28089106e+00 -1.20451413e-01
6.00979149e-01 -7.86342502e-01 -1.59286574e-01 -1.33452833e+00
5.76590180e-01 -7.55081892e-01 -6.49784654e-02 -3.29468399e-01
-1.20779741e+00 6.60542011e-01 -3.42815131e-01 -1.16733873e+00
6.35626465e-02 -1.08263695e+00 -3.40110391e-01 8.60872090e-01
-1.46821761e+00 -8.46680701e-01 8.51645321e-02 2.34002754e-01
4.14468497e-01 -1.04541503e-01 1.28959548e+00 6.10387146e-01
-7.92873263e-01 8.28743517e-01 2.21546084e-01 4.24137026e-01
2.45293647e-01 -1.00227237e+00 8.17810595e-01 5.72078347e-01
3.60636950e-01 1.11512971e+00 3.22127104e-01 -2.15623006e-01
-2.33160067e+00 -1.03427625e+00 1.48512280e+00 -2.87039220e-01
1.04266357e+00 -8.11864734e-01 -1.13361251e+00 6.01681352e-01
1.07414752e-01 3.90943527e-01 1.14224458e+00 3.15527879e-02
-8.28356326e-01 -1.49511874e-01 -7.57243931e-01 7.52458215e-01
9.96459961e-01 -9.85382617e-01 -8.77900004e-01 5.88435054e-01
1.34875333e+00 -2.21512634e-02 -1.13208735e+00 -7.01723918e-02
4.57339197e-01 -3.38094473e-01 9.67076480e-01 -9.91257012e-01
5.49965620e-01 -5.69326282e-02 -6.18849099e-01 -1.24287093e+00
-2.40183771e-01 -2.98851609e-01 -6.64973319e-01 8.09889257e-01
4.10229653e-01 -8.41436625e-01 6.00836337e-01 6.52083516e-01
-1.05181172e-01 -1.23450506e+00 -1.09996688e+00 -1.00822294e+00
4.03596967e-01 -4.82566684e-01 1.05432963e+00 9.53540504e-01
4.13544655e-01 4.28737313e-01 -6.76992238e-02 -8.13086629e-02
3.44306082e-01 2.80172593e-04 3.64438295e-01 -7.66843438e-01
-2.85593152e-01 -1.79061502e-01 -7.85359442e-01 -1.48441696e+00
2.35474542e-01 -1.27836585e+00 -3.42874110e-01 -1.40090334e+00
1.32213235e-01 -4.99796718e-01 -2.66280293e-01 4.58750933e-01
-1.97786927e-01 2.02721044e-01 -2.77811531e-02 1.81309327e-01
-1.76404536e-01 1.03819072e+00 1.11970651e+00 -3.64255875e-01
3.46481532e-01 -8.78049254e-01 -6.02679372e-01 5.62068105e-01
5.72319448e-01 -4.31388199e-01 -3.04328322e-01 -1.31143606e+00
4.53411907e-01 -4.81352121e-01 -2.02350333e-01 -3.32508355e-01
1.07898436e-01 1.32359311e-01 1.98106486e-02 -3.09055984e-01
6.90645695e-01 -7.53488421e-01 -4.71337318e-01 6.15728855e-01
-1.06857635e-01 7.48645246e-01 1.03655383e-01 4.98893708e-01
-3.94806087e-01 -3.19829553e-01 2.56995738e-01 7.84821734e-02
-5.03436744e-01 6.53405845e-01 -2.49648884e-01 6.39649928e-02
7.61000693e-01 -1.23282261e-01 -2.05569610e-01 8.71218517e-02
-1.35434449e-01 4.38946784e-02 3.68427426e-01 6.86933339e-01
9.62709904e-01 -1.79244006e+00 -3.98831218e-01 4.90294218e-01
2.47918114e-01 -5.03621809e-02 7.28856251e-02 5.30324280e-01
-7.14012861e-01 6.94655061e-01 4.68901806e-02 -8.15469176e-02
-9.61230695e-01 8.61032605e-01 -8.64849761e-02 -1.18597418e-01
-6.05325162e-01 9.28989112e-01 1.18039839e-01 -4.55165178e-01
8.01335350e-02 -6.06039286e-01 1.11455478e-01 -3.43590491e-02
7.97748387e-01 2.69047707e-01 1.47231773e-01 -3.91356707e-01
-2.82024235e-01 3.94273758e-01 -5.24015427e-01 1.55896321e-02
1.29623008e+00 5.68876565e-02 -7.99075127e-01 3.15603554e-01
1.93034422e+00 -2.62644112e-01 -6.45257652e-01 -5.08598626e-01
-1.70282215e-01 -8.03772509e-01 3.85585248e-01 8.56450349e-02
-9.15895104e-01 1.45114362e+00 3.83895487e-01 7.58724064e-02
8.05408835e-01 -1.40551310e-02 1.73313999e+00 6.84393466e-01
3.47348809e-01 -1.09617496e+00 2.78294832e-01 6.46828413e-01
5.01463473e-01 -8.23574662e-01 -1.60583749e-01 -2.96701521e-01
-1.68315858e-01 1.08122778e+00 3.43349397e-01 -9.80983153e-02
6.65480316e-01 2.96747051e-02 -3.16380650e-01 3.45534272e-02
-9.65722740e-01 1.07201569e-01 2.94368029e-01 3.08143377e-01
2.45161623e-01 2.56534308e-01 -5.96029758e-01 4.28004265e-01
-3.31173241e-01 -5.80656886e-01 1.95453167e-01 9.28348660e-01
-5.43822885e-01 -1.49057662e+00 1.05556445e-02 5.77550411e-01
-3.34858000e-01 -5.87549388e-01 -6.60595968e-02 2.17482150e-01
2.42711872e-01 7.31038749e-01 4.71975118e-01 -6.81257844e-01
1.51637763e-01 2.48536214e-01 5.25532007e-01 -7.63920069e-01
-5.99423088e-02 -5.31645596e-01 -1.79141313e-01 -4.31843549e-01
3.55026811e-01 -2.85464406e-01 -1.27262461e+00 -6.68540597e-01
-4.68635172e-01 4.02100652e-01 1.14889479e+00 8.07251096e-01
6.03877902e-01 2.28742465e-01 6.54814482e-01 -3.23738337e-01
-9.54229832e-01 -1.06336844e+00 -5.37727118e-01 4.28187579e-01
1.30676657e-01 -4.49175447e-01 -4.41811144e-01 -1.43601149e-01]
|
[10.628108978271484, 8.697815895080566]
|
92e04f11-cfdb-4ba2-9afb-4eb01bc15aec
|
multi-centroid-task-descriptor-for-dynamic
| null | null |
http://openaccess.thecvf.com//content/CVPR2023/html/Cai_Multi-Centroid_Task_Descriptor_for_Dynamic_Class_Incremental_Inference_CVPR_2023_paper.html
|
http://openaccess.thecvf.com//content/CVPR2023/papers/Cai_Multi-Centroid_Task_Descriptor_for_Dynamic_Class_Incremental_Inference_CVPR_2023_paper.pdf
|
Multi-Centroid Task Descriptor for Dynamic Class Incremental Inference
|
Incremental learning could be roughly divided into two categories, i.e., class- and task-incremental learning. The main difference is whether the task ID is given during evaluation. In this paper, we show this task information is indeed a strong prior knowledge, which will bring significant improvement over class-incremental learning baseline, e.g., DER. Based on this observation, we propose a gate network to predict the task ID for class incremental inference. This is challenging as there is no explicit semantic relationship between categories in the concept of task. Therefore, we propose a multi-centroid task descriptor by assuming the data within a task can form multiple clusters. The cluster centers are optimized by pulling relevant sample-centroid pairs while pushing others away, which ensures that there is at least one centroid close to a given sample. To select relevant pairs, we use class prototypes as proxies and solve a bipartite matching problem, making the task descriptor representative yet not degenerate to uni-modal. As a result, our dynamic inference network is trained independently of baseline and provides a flexible, efficient solution to distinguish between tasks. Extensive experiments show our approach achieves state-of-the-art results, e.g., we achieve 72.41% average accuracy on CIFAR100-B0S50, outperforming DER by 3.40%.
|
['Yuan Xie', 'Chengjie Wang', 'Guannan Jiang', 'Yanyun Qu', 'Xin Tan', 'Zhizhong Zhang', 'Tenghao Cai']
|
2023-01-01
| null | null | null |
cvpr-2023-1
|
['class-incremental-learning', 'incremental-learning']
|
['computer-vision', 'methodology']
|
[ 8.33086371e-02 -1.01990774e-01 -4.84541148e-01 -5.63409269e-01
-9.34666991e-01 -7.57438779e-01 6.31240547e-01 -3.24044563e-02
-6.14687800e-01 4.55162138e-01 6.05434701e-02 -9.80290677e-03
-1.99786931e-01 -2.41692692e-01 -8.13295603e-01 -8.43531072e-01
2.77281404e-01 6.81642175e-01 2.82199383e-01 1.89882591e-01
1.93278000e-01 1.85695738e-01 -1.37474489e+00 4.92966950e-01
8.56111646e-01 1.22744763e+00 4.88957614e-01 8.61897096e-02
7.34068155e-02 3.92635822e-01 -6.12090349e-01 -4.70596641e-01
2.36052141e-01 -1.69922605e-01 -8.69777024e-01 -1.82739943e-02
6.95734918e-01 -3.34196985e-01 -2.80356079e-01 1.10299087e+00
5.60733914e-01 2.37648770e-01 8.61748874e-01 -1.59654617e+00
-6.85724139e-01 7.29998946e-01 -6.41869843e-01 3.25953029e-02
-1.09847598e-01 -9.78268906e-02 1.36207330e+00 -1.30252206e+00
5.11294365e-01 1.24503851e+00 4.53772753e-01 6.19689703e-01
-1.44115317e+00 -1.10162902e+00 7.58336425e-01 4.50176418e-01
-1.66869462e+00 -5.81263661e-01 8.23927820e-01 -4.71874833e-01
5.73733211e-01 -2.36319583e-02 2.20162347e-01 1.23313248e+00
-1.18662760e-01 1.16642272e+00 8.95319462e-01 -2.61126131e-01
2.18662456e-01 1.72790065e-01 4.83045161e-01 4.04917032e-01
2.12443978e-01 -2.48670384e-01 -4.77463633e-01 -1.22392729e-01
3.56768638e-01 3.76689315e-01 -3.04019928e-01 -6.40885949e-01
-1.52297747e+00 6.88999474e-01 7.67848790e-01 2.78683454e-01
-8.31910744e-02 1.30996138e-01 5.37962854e-01 2.35619143e-01
2.91872233e-01 2.34976590e-01 -4.38241512e-01 1.97371706e-01
-8.11665297e-01 2.39267513e-01 4.14019525e-01 1.17825806e+00
8.83836150e-01 -3.91293287e-01 -6.78861737e-01 1.09052801e+00
1.72816962e-01 4.87782300e-01 4.55990613e-01 -9.16650891e-01
7.50809550e-01 5.54001153e-01 4.28794476e-04 -9.30947602e-01
-2.92099535e-01 -6.64269686e-01 -9.22931552e-01 -3.29356134e-01
5.48219800e-01 -9.16197374e-02 -8.58166218e-01 2.05245781e+00
1.37323111e-01 2.31650591e-01 -1.41783640e-01 8.78603339e-01
5.71152925e-01 5.19487023e-01 1.32196113e-01 -1.13448657e-01
1.59643984e+00 -1.17436075e+00 -5.69882274e-01 -4.41679329e-01
4.97336686e-01 -5.46464860e-01 9.60631847e-01 3.28029931e-01
-5.51783025e-01 -8.63066614e-01 -8.14391851e-01 5.41793741e-02
-3.08379292e-01 4.53530163e-01 5.52757263e-01 1.36610329e-01
-7.62590289e-01 2.53173769e-01 -6.12227798e-01 -6.97513446e-02
5.15632689e-01 2.73954570e-01 -2.39205971e-01 -3.40462834e-01
-1.23978651e+00 4.82758254e-01 6.30791783e-01 -4.95212711e-02
-1.05038738e+00 -6.62207186e-01 -6.23844504e-01 2.10256323e-01
5.37338078e-01 -6.08925998e-01 1.41445100e+00 -8.24265480e-01
-9.83271480e-01 9.29789960e-01 -5.84198058e-01 -2.93893397e-01
4.10160631e-01 -5.94088882e-02 -2.12159097e-01 4.43360284e-02
4.11185116e-01 9.18255091e-01 8.17304194e-01 -1.25532877e+00
-9.89649773e-01 -4.05672222e-01 1.29919037e-01 3.29854518e-01
-6.54733956e-01 -7.38019496e-03 -8.50960672e-01 -5.82239985e-01
2.06278890e-01 -1.19824624e+00 1.01715751e-01 7.21474513e-02
-5.47172487e-01 -9.42036808e-01 6.66918814e-01 -2.82321870e-01
1.27408111e+00 -2.31704974e+00 3.68738137e-02 1.30036131e-01
4.03038740e-01 6.94314912e-02 -7.74121955e-02 7.80277923e-02
-1.20529257e-01 -6.99023455e-02 -7.15687051e-02 -5.42610466e-01
2.51446694e-01 -5.50965890e-02 -4.14071232e-01 2.51547664e-01
-8.72991756e-02 1.09504294e+00 -9.46617126e-01 -5.24144053e-01
-2.38981739e-01 2.23947823e-01 -4.09696281e-01 1.80821106e-01
-3.13011780e-02 1.99545249e-01 -5.27910113e-01 6.41439676e-01
7.26329923e-01 -3.68652672e-01 1.62278444e-01 -5.08762181e-01
1.65773898e-01 2.59317905e-01 -1.15039885e+00 1.94138014e+00
-4.23120767e-01 6.57337248e-01 -4.84073386e-02 -1.35705876e+00
8.26174080e-01 1.55128434e-01 3.09546769e-01 -6.63767993e-01
-8.30484405e-02 1.52894348e-01 1.15764663e-01 -9.50631648e-02
2.82286584e-01 1.52320743e-01 -2.65461862e-01 3.71603757e-01
1.15320541e-01 3.46685469e-01 2.45067045e-01 2.70733237e-01
7.09819555e-01 -6.61280900e-02 -1.25762224e-02 -3.68179321e-01
3.97053033e-01 -1.22872531e-01 8.71065617e-01 9.32556093e-01
-4.48394537e-01 5.55777013e-01 5.71886718e-01 -2.03793570e-01
-6.61894977e-01 -1.14265823e+00 -4.02441591e-01 1.42789757e+00
4.01199907e-01 -2.91730523e-01 -6.52653456e-01 -8.71042013e-01
1.15475722e-01 5.48689842e-01 -7.29211271e-01 -5.01579523e-01
-6.65604591e-01 -6.18385255e-01 2.17151031e-01 7.22252786e-01
4.26275879e-01 -7.93337226e-01 -1.43368334e-01 1.56682581e-01
-4.78932649e-01 -1.21155870e+00 -9.61268425e-01 3.47209692e-01
-7.27987528e-01 -1.11096013e+00 -9.23391640e-01 -1.12107241e+00
8.65824997e-01 5.95630348e-01 1.09632397e+00 -3.18895757e-01
-1.96975060e-02 1.84380725e-01 -1.80785656e-01 -3.23514968e-01
1.89103916e-01 3.35279793e-01 5.27113583e-03 2.52668917e-01
5.38212776e-01 -2.28547201e-01 -9.34005797e-01 6.58428729e-01
-4.65010285e-01 -1.10168077e-01 7.53446460e-01 1.04557371e+00
8.55794430e-01 -2.16001309e-02 6.93991780e-01 -9.25511897e-01
4.88742650e-01 -5.65782905e-01 -2.57216543e-01 5.78153670e-01
-5.83639383e-01 8.46537650e-02 8.63208771e-01 -8.45055580e-01
-1.04928911e+00 2.70478517e-01 3.93500745e-01 -6.59774184e-01
5.06824441e-02 3.59500468e-01 -2.93937773e-01 2.37857401e-01
4.96121496e-01 2.41943523e-01 -2.08422452e-01 -4.86284435e-01
3.72755736e-01 9.86199498e-01 4.67051029e-01 -8.16453516e-01
7.27398157e-01 4.93351072e-01 -3.52672577e-01 -2.93069571e-01
-1.23836434e+00 -8.27591836e-01 -6.50006354e-01 -7.01163150e-03
6.60317183e-01 -1.20915461e+00 -6.87525094e-01 5.84658504e-01
-1.14480603e+00 -4.17587280e-01 4.74250838e-02 5.26148796e-01
-2.60843396e-01 2.43197978e-02 -4.75838602e-01 -4.43461120e-01
-2.28802860e-01 -1.27256870e+00 1.27556050e+00 6.10337667e-02
3.35316285e-02 -8.26972961e-01 -2.71296382e-01 2.91766763e-01
2.69621670e-01 -1.41354948e-01 7.83116102e-01 -8.76653850e-01
-5.80540955e-01 -1.14040583e-01 -5.79393387e-01 2.45006546e-01
9.27355513e-02 -3.20268661e-01 -1.09899080e+00 -6.27014756e-01
-2.34584317e-01 -5.01613140e-01 1.11412597e+00 2.31015965e-01
1.52880144e+00 -3.56362015e-02 -8.24859023e-01 5.99679053e-01
1.13079011e+00 2.77429283e-01 2.70477325e-01 8.68143812e-02
8.95184577e-01 4.89556283e-01 8.57354641e-01 1.64791375e-01
6.92032933e-01 9.05379772e-01 1.93463728e-01 2.03998417e-01
-2.25741014e-01 -2.62312770e-01 2.93827623e-01 6.83822870e-01
2.73275286e-01 -7.35149384e-02 -9.41735864e-01 5.91746986e-01
-2.01815271e+00 -7.79552341e-01 2.23666191e-01 2.26626015e+00
1.01224887e+00 2.05681875e-01 -3.13940346e-02 -1.97816178e-01
1.07180572e+00 1.62510667e-02 -8.96394074e-01 1.16213173e-01
1.46280006e-01 -7.07197860e-02 3.81842524e-01 1.85038164e-01
-1.37301517e+00 9.11886930e-01 5.08154678e+00 1.28417170e+00
-1.04687500e+00 1.90353557e-01 7.13437915e-01 -1.00526474e-01
-7.14836419e-02 -7.74985179e-02 -1.23267853e+00 6.99625790e-01
5.52803218e-01 -3.20160389e-01 3.18541139e-01 9.83631670e-01
-2.57384300e-01 2.78009862e-01 -1.40540385e+00 1.12148607e+00
2.59484440e-01 -8.16061139e-01 8.48042369e-02 -9.22909677e-02
7.19907880e-01 3.63386683e-02 1.07090846e-01 8.32714140e-01
2.98889369e-01 -6.08763874e-01 8.11881721e-01 4.26847011e-01
8.99176478e-01 -6.55221522e-01 4.55053449e-01 5.22174478e-01
-1.47376347e+00 -3.12910050e-01 -6.50577188e-01 3.93360823e-01
-1.63110137e-01 6.33483768e-01 -5.19047260e-01 3.98321509e-01
8.16733658e-01 8.01170230e-01 -6.00502431e-01 1.19374120e+00
-2.03705430e-01 4.88378823e-01 -1.94201306e-01 1.00689523e-01
1.80167779e-01 9.47553385e-03 2.41743684e-01 1.20728946e+00
2.03335360e-01 -2.79600531e-01 5.76501131e-01 7.89193511e-01
-4.91713852e-01 -1.48151368e-02 -4.42292392e-01 4.36062276e-01
1.01443040e+00 1.23531163e+00 -6.43791199e-01 -4.35548931e-01
-4.20273334e-01 1.09806967e+00 7.22123981e-01 6.03579581e-01
-9.61186647e-01 -6.96853101e-01 6.88065171e-01 -1.37716413e-01
4.07071769e-01 -1.53046576e-02 6.13668449e-02 -1.22876000e+00
2.14822188e-01 -6.21071696e-01 6.14530683e-01 -5.43475091e-01
-1.54228806e+00 5.46457708e-01 1.67940676e-01 -1.31891751e+00
-5.37408143e-02 -6.97159410e-01 -5.53154230e-01 7.20299482e-01
-1.61312449e+00 -1.21708000e+00 -5.03395975e-01 6.26066506e-01
5.04160345e-01 -1.81986496e-01 5.58093250e-01 5.41363001e-01
-6.27293587e-01 1.16602540e+00 2.65567303e-01 4.13789451e-01
1.25635242e+00 -1.29371750e+00 2.28496596e-01 6.40626609e-01
2.06059068e-01 8.39468002e-01 5.81594044e-03 -3.07996660e-01
-1.07333398e+00 -1.38384104e+00 8.82098258e-01 -4.79956686e-01
6.44939125e-01 -9.62210178e-01 -9.31261837e-01 7.29158640e-01
-5.44744506e-02 2.92622983e-01 4.76733357e-01 3.91145706e-01
-7.98608482e-01 -5.73089123e-01 -6.60547435e-01 5.31131268e-01
1.38372552e+00 -7.55649507e-01 -6.35611713e-01 5.46041369e-01
8.66222262e-01 -2.21043691e-01 -6.02772176e-01 2.94752121e-01
4.00443435e-01 -5.58405936e-01 1.07240152e+00 -4.35743988e-01
-8.11182521e-03 -3.12030405e-01 -2.30042115e-02 -1.41894197e+00
-5.19954383e-01 -2.22428188e-01 -1.65440813e-01 1.47109365e+00
5.37696540e-01 -6.90499902e-01 6.44617021e-01 6.29539430e-01
-3.09861690e-01 -8.09905589e-01 -8.11486363e-01 -1.14298296e+00
3.44008476e-01 -2.61353970e-01 5.05663991e-01 1.14741445e+00
-1.56145036e-01 6.42638505e-01 -1.42623559e-01 -1.11155488e-01
8.68197978e-01 5.16582191e-01 5.62613785e-01 -1.44475520e+00
-1.41176805e-01 -5.72315574e-01 -3.89077105e-02 -1.62471414e+00
6.55981600e-01 -1.29073882e+00 1.16454296e-01 -1.37217021e+00
5.84668398e-01 -9.30557847e-01 -8.25552702e-01 8.16083848e-01
-5.10287821e-01 1.21670634e-01 1.33676320e-01 7.37985313e-01
-1.10388279e+00 4.99693692e-01 1.22436929e+00 -4.02630270e-01
-1.11730233e-01 6.08784482e-02 -9.98077631e-01 4.58047867e-01
8.65560889e-01 -6.24090493e-01 -5.95260203e-01 -6.11383557e-01
4.90311459e-02 -3.27463657e-01 2.85913110e-01 -7.99201608e-01
6.10347569e-01 6.22213781e-02 2.66819984e-01 -5.52500963e-01
1.50508732e-01 -8.94308448e-01 -3.34126234e-01 2.58817405e-01
-5.88571787e-01 -3.09513301e-01 -7.54271671e-02 8.19600642e-01
-2.74709940e-01 -2.83677340e-01 6.17841065e-01 1.04022726e-01
-9.16481912e-01 6.14556551e-01 2.37266794e-01 3.93060446e-01
1.03430533e+00 -5.76129816e-02 -5.59231281e-01 -1.60565376e-02
-5.83712459e-01 6.66918278e-01 1.53758630e-01 6.20956004e-01
2.71272957e-01 -1.60846961e+00 -7.77252913e-01 -8.64083841e-02
6.03609204e-01 2.51154751e-01 2.83580035e-01 7.68926501e-01
3.25318009e-01 6.04333758e-01 6.45844042e-02 -9.05002058e-01
-1.06930053e+00 7.25728333e-01 2.01723173e-01 -2.23862290e-01
-3.51466119e-01 9.02435064e-01 8.02090764e-01 -5.69991529e-01
6.25638604e-01 -9.80548114e-02 -2.78980225e-01 5.44885457e-01
5.56488156e-01 8.27436373e-02 7.26772323e-02 -4.82805163e-01
-5.75524569e-01 6.45087242e-01 -5.34922302e-01 2.67056108e-01
9.48615193e-01 -5.71520180e-02 5.22015728e-02 5.41299462e-01
1.63075161e+00 -3.82916093e-01 -1.44734871e+00 -8.21048141e-01
4.97548357e-02 -3.72539610e-01 -7.04621151e-02 -7.72106290e-01
-1.14923990e+00 9.26037073e-01 6.05928659e-01 -1.26232281e-01
9.58293915e-01 3.07512194e-01 7.10595369e-01 7.54319310e-01
4.51753020e-01 -1.32765257e+00 2.73681581e-01 5.24318457e-01
7.99203455e-01 -1.45869291e+00 -1.48887187e-01 -2.41790935e-01
-7.55762756e-01 7.55515873e-01 8.58627915e-01 2.67748255e-02
6.29074931e-01 -1.72199771e-01 -1.66054830e-01 -8.61333907e-02
-8.32216084e-01 -2.90141106e-01 5.33895373e-01 4.55118537e-01
3.49827975e-01 2.27044150e-01 -1.55322731e-01 7.78423905e-01
9.16098952e-02 -2.50991583e-01 -1.63099587e-01 5.63556075e-01
-4.78376478e-01 -9.56233501e-01 -1.49770156e-01 5.94632387e-01
-3.24787229e-01 -2.21415579e-01 -2.43343458e-01 6.96952283e-01
2.15076253e-01 8.13703299e-01 3.32530886e-01 -3.10498834e-01
5.20881772e-01 1.49892658e-01 2.96920627e-01 -6.31807148e-01
-2.60356009e-01 -9.24674422e-02 -1.92469209e-01 -4.06835407e-01
-1.82918832e-01 -6.60381913e-01 -1.12769258e+00 5.95516488e-02
-4.77988899e-01 3.83962572e-01 4.68407393e-01 6.64497137e-01
5.83207726e-01 4.55016941e-01 8.83462310e-01 -6.19000316e-01
-8.18816662e-01 -9.52313185e-01 -3.61865848e-01 6.64729059e-01
1.16182998e-01 -1.03650355e+00 -4.05269057e-01 8.26200172e-02]
|
[9.632356643676758, 2.9282469749450684]
|
72974bb9-2fcd-4c33-91f5-196a75faf18e
|
completer-incomplete-multi-view-clustering
| null | null |
http://pengxi.me/wp-content/uploads/2021/03/2021CVPR-completer.pdf
|
http://pengxi.me/wp-content/uploads/2021/03/2021CVPR-completer.pdf
|
COMPLETER: Incomplete Multi-view Clustering via Contrastive Prediction
|
In this paper, we study two challenging problems in incomplete multi-view clustering analysis, namely, i) how to learn an informative and consistent representation among different views without the help of labels and ii) how to recover the missing views from data. To this end, we propose a novel objective that incorporates representation learning and data recovery into a unified framework from the view of information theory. To be specific, the informative and consistent representation is learned by maximizing the mutual information across different views through contrastive learning, and the missing views are recovered by minimizing the conditional entropy of different views through dual prediction. To the best of our knowledge, this could be the first work to provide a theoretical framework that unifies the consistent representation learning and cross-view data recovery. Extensive experimental results show the proposed method remarkably outperforms 10 competitive multi-view clustering methods on four challenging datasets. The code is available at https://pengxi.me.
|
['Xi Peng', 'Jiancheng Lv', 'Boyun Li', 'Zitao Liu', 'Yuanbiao Gou', 'Yijie Lin']
|
2021-03-22
| null |
http://openaccess.thecvf.com//content/CVPR2021/html/Lin_COMPLETER_Incomplete_Multi-View_Clustering_via_Contrastive_Prediction_CVPR_2021_paper.html
|
http://openaccess.thecvf.com//content/CVPR2021/papers/Lin_COMPLETER_Incomplete_Multi-View_Clustering_via_Contrastive_Prediction_CVPR_2021_paper.pdf
|
cvpr-2021-1
|
['incomplete-multi-view-clustering', 'multi-view-learning']
|
['computer-vision', 'computer-vision']
|
[ 1.03308529e-01 -4.54239286e-02 -5.30157149e-01 -4.52575266e-01
-1.10065639e+00 -5.52628875e-01 2.93388188e-01 -3.06247383e-01
3.11362445e-01 3.86408180e-01 3.54587525e-01 3.16237926e-01
-4.16776448e-01 -3.37390274e-01 -5.36054611e-01 -9.58447576e-01
2.55504578e-01 6.16248012e-01 -3.98080856e-01 2.46587381e-01
1.54749528e-01 -6.44490495e-02 -1.45391488e+00 4.93259102e-01
8.11619461e-01 7.53709018e-01 2.66861260e-01 6.57582209e-02
2.52041399e-01 9.12501395e-01 8.00527185e-02 -1.99324161e-01
2.99172103e-01 -6.15964174e-01 -9.31888163e-01 6.88346863e-01
1.91350788e-01 -2.53807932e-01 -3.01989913e-01 1.01804483e+00
4.11838651e-01 1.30871400e-01 6.46395266e-01 -1.38877475e+00
-9.16916370e-01 5.55387437e-01 -1.07966292e+00 -1.32341623e-01
4.19003338e-01 -4.93875206e-01 1.21575403e+00 -1.09309530e+00
6.83657646e-01 1.01073587e+00 1.71099380e-01 4.59399432e-01
-1.37644994e+00 -5.94694436e-01 4.06375319e-01 3.57538819e-01
-1.38534868e+00 -6.72438979e-01 1.12011492e+00 -5.76716661e-01
9.76930559e-02 9.53846872e-02 2.55830258e-01 1.10052884e+00
-4.08156931e-01 1.08791006e+00 1.32397830e+00 -2.50661314e-01
8.78393129e-02 3.51160616e-01 1.15181893e-01 6.37007713e-01
3.37483376e-01 -1.63270488e-01 -6.11601055e-01 -1.92251369e-01
4.04593945e-01 5.03499150e-01 -5.88435829e-01 -1.07387412e+00
-1.15583336e+00 9.51787233e-01 1.87806353e-01 2.66510099e-01
-2.61861056e-01 -2.73462534e-01 2.64124721e-01 2.24335611e-01
5.01325488e-01 -1.20364696e-01 -3.26285928e-01 3.55575562e-01
-7.91714430e-01 -2.94219375e-01 6.88424468e-01 1.15754819e+00
8.57587457e-01 -1.22610211e-01 3.80780518e-01 9.14387345e-01
5.37806809e-01 6.05492711e-01 2.70427763e-01 -1.34167445e+00
6.61922753e-01 5.75126767e-01 -5.01709692e-02 -9.57028627e-01
-9.65365320e-02 -2.69947201e-01 -1.18103039e+00 -8.10882002e-02
2.67316133e-01 -1.11720473e-01 -3.10576051e-01 1.84835446e+00
3.42496544e-01 2.62813479e-01 2.07309172e-01 8.65930915e-01
9.32422161e-01 5.11247218e-01 -5.36102116e-01 -6.48984313e-01
1.10113990e+00 -8.55735779e-01 -8.72809887e-01 -8.47223550e-02
2.87953168e-01 -6.54785216e-01 4.46043462e-01 4.92803156e-01
-1.13039863e+00 -5.38605273e-01 -1.04653609e+00 7.88541883e-02
-2.40762588e-02 3.62293482e-01 4.57751334e-01 3.49822193e-01
-8.06605220e-01 2.23131746e-01 -7.80502021e-01 -2.40399703e-01
3.00541610e-01 2.07074776e-01 -7.75246084e-01 -5.61091602e-01
-6.87909901e-01 5.68693995e-01 3.26588273e-01 -9.57555994e-02
-7.15460837e-01 -5.26641548e-01 -9.08851445e-01 -1.65904254e-01
7.87894011e-01 -8.13461959e-01 8.35967302e-01 -9.11100805e-01
-9.98613060e-01 1.07421362e+00 -5.49910009e-01 -5.38493022e-02
1.65479884e-01 -2.68549085e-01 -2.20433682e-01 3.39895546e-01
3.61645609e-01 1.50127783e-01 8.21418285e-01 -2.07500887e+00
-3.18735421e-01 -8.33615959e-01 -1.63420379e-01 3.81832927e-01
-5.69983125e-02 -3.67795855e-01 -7.74568856e-01 -4.88013089e-01
5.98848045e-01 -1.14973617e+00 -4.93987873e-02 -2.16234729e-01
-4.92699534e-01 -1.17933499e-02 8.36884499e-01 -8.00341487e-01
9.51094627e-01 -2.17156410e+00 7.89047718e-01 2.34964162e-01
4.00780380e-01 -2.27324218e-01 6.90350458e-02 5.37744403e-01
-3.67736012e-01 -7.31846318e-02 -4.44887191e-01 -5.28442442e-01
-7.08275363e-02 2.34045133e-01 -3.32636088e-01 8.14779699e-01
-3.86695981e-01 6.58904552e-01 -7.72143841e-01 -4.76832032e-01
2.84423739e-01 4.73329425e-01 -3.70095104e-01 3.33930314e-01
4.23126787e-01 8.25204670e-01 -5.52888334e-01 6.29627705e-01
9.08828259e-01 -7.60751665e-01 6.56564415e-01 -3.65728587e-01
1.72470272e-01 -1.14665307e-01 -1.47788763e+00 2.03237772e+00
-3.26638609e-01 2.32172802e-01 3.77037525e-01 -1.37095404e+00
6.80100024e-01 5.40662229e-01 9.75149333e-01 -3.75196666e-01
-1.02532946e-01 -4.03383188e-02 -4.44282949e-01 -5.10370553e-01
2.25596532e-01 -3.24357420e-01 -1.81655809e-01 7.21815884e-01
1.13056876e-01 1.89024046e-01 3.41845788e-02 4.43784952e-01
4.46692616e-01 9.35076326e-02 5.92286885e-01 6.59300163e-02
6.45485997e-01 -4.71795261e-01 7.98208594e-01 5.07179558e-01
-6.01631962e-03 8.55942726e-01 2.51241207e-01 -1.66516304e-01
-8.31596136e-01 -1.17354882e+00 -1.39385806e-02 8.60469162e-01
2.24457562e-01 -4.94206220e-01 -4.71369475e-01 -7.92658687e-01
-1.60959035e-01 4.26848382e-01 -5.53267062e-01 1.54601216e-01
-3.32593292e-01 -5.56713045e-01 -1.97480589e-01 5.05762517e-01
3.43844622e-01 -5.12565613e-01 1.64830267e-01 -2.44081542e-01
-8.98247421e-01 -1.10762823e+00 -5.75366080e-01 1.45750955e-01
-8.63155961e-01 -1.37531233e+00 -5.81550896e-01 -8.29716682e-01
7.93586075e-01 1.01975572e+00 1.10824120e+00 -2.82767918e-02
4.18331511e-02 9.10636306e-01 -4.95042831e-01 6.99496493e-02
-4.32915986e-02 -1.61802903e-01 1.68672547e-01 2.84921676e-01
2.02624664e-01 -8.14080656e-01 -4.06135678e-01 5.07975459e-01
-8.48602951e-01 1.19218990e-01 5.23721874e-01 8.05485845e-01
1.08216333e+00 1.22272566e-01 5.06471694e-01 -1.21083331e+00
7.76105970e-02 -8.89167488e-01 -3.09309036e-01 5.63987970e-01
-6.59459770e-01 2.94339806e-02 5.98503172e-01 1.77745134e-01
-1.27252364e+00 2.95229584e-01 3.47147852e-01 -7.01558888e-01
-2.53149897e-01 4.03046638e-01 -6.25214696e-01 2.64730811e-01
2.99634803e-02 5.40143073e-01 -5.56738414e-02 -6.44021809e-01
7.47446477e-01 5.21113217e-01 3.99926931e-01 -6.19494677e-01
8.04078281e-01 9.83208537e-01 -6.93909377e-02 -5.98964155e-01
-1.26918924e+00 -8.84659231e-01 -1.06648648e+00 -9.79942381e-02
8.00989747e-01 -1.46811497e+00 -4.50769722e-01 1.01258241e-01
-7.43421435e-01 3.60075533e-01 -1.72027186e-01 5.90025842e-01
-8.88404012e-01 9.08304930e-01 -3.81518662e-01 -6.21826470e-01
-9.49253738e-02 -9.81082618e-01 9.55016851e-01 -3.25560942e-02
1.78679690e-01 -1.02277339e+00 2.54695535e-01 1.05558729e+00
-3.15798134e-01 2.35826001e-01 4.45700616e-01 -4.71244365e-01
-5.87625563e-01 1.07822254e-01 -1.14905968e-01 4.09844100e-01
3.78191024e-01 -3.27714622e-01 -1.08143246e+00 -6.63846254e-01
5.34411490e-01 -3.85503292e-01 1.10986423e+00 5.71696937e-01
1.21645975e+00 -3.39281678e-01 -3.73615742e-01 8.16090763e-01
1.57548034e+00 7.45849609e-02 3.77793789e-01 -9.89378542e-02
8.89357209e-01 7.92906404e-01 4.25972730e-01 6.90387130e-01
8.53098452e-01 5.56891203e-01 3.89160007e-01 2.33806655e-01
3.24266814e-02 -2.68000424e-01 1.84249952e-01 1.37803304e+00
-1.63943514e-01 -2.29114860e-01 -5.76149583e-01 5.25913715e-01
-2.18046761e+00 -1.47203100e+00 -1.72842711e-01 2.20497656e+00
2.34981090e-01 -4.93017226e-01 1.57015696e-01 3.07035651e-02
9.95301604e-01 4.01716679e-01 -5.44236600e-01 3.57304424e-01
-1.99953809e-01 -4.21157092e-01 4.86884415e-02 2.91465819e-01
-1.27323079e+00 2.75950998e-01 5.51183844e+00 4.59030837e-01
-3.85883212e-01 2.92659521e-01 5.74434936e-01 -1.88007310e-01
-4.50901151e-01 1.86660916e-01 -5.54892302e-01 4.37795430e-01
5.36406279e-01 -9.39303041e-02 6.67823076e-01 7.22917020e-01
1.08212270e-02 7.91868046e-02 -1.22332001e+00 1.36101592e+00
5.76845050e-01 -1.03201354e+00 -4.59473990e-02 3.52008671e-01
1.10866356e+00 -5.30822091e-02 1.79249331e-01 1.06279878e-02
3.54148269e-01 -6.75906420e-01 3.64451796e-01 7.84275770e-01
4.99013543e-01 -9.06640589e-01 5.21456361e-01 4.87795472e-01
-1.42634892e+00 -8.34558830e-02 -3.41271043e-01 2.12881416e-01
2.10885704e-01 5.75956285e-01 -1.77770615e-01 1.23873711e+00
5.56914806e-01 1.47697186e+00 -2.59568393e-01 7.55749166e-01
-1.75375670e-01 3.03436279e-01 2.39630073e-01 8.24226379e-01
-2.27188185e-01 -6.90698504e-01 4.19931978e-01 7.64424324e-01
3.54629368e-01 1.90874219e-01 5.26551127e-01 7.06160307e-01
-1.24293387e-01 -2.61817034e-02 -9.14964974e-01 2.97782451e-01
4.87464458e-01 1.35222006e+00 -4.72379982e-01 -2.46238783e-01
-7.72274077e-01 9.97020304e-01 4.79969829e-01 4.52094048e-01
-5.86757898e-01 3.52629155e-01 3.66287470e-01 -1.16777048e-01
5.80256462e-01 -2.47578714e-02 -3.40691984e-01 -1.76271331e+00
2.36265227e-01 -8.20594966e-01 8.69289637e-01 -6.97269440e-01
-1.83190513e+00 1.94031000e-01 1.15915775e-01 -1.67663157e+00
-2.84109145e-01 -1.53550267e-01 -3.44135523e-01 5.51990390e-01
-1.44016492e+00 -1.21225774e+00 -1.80405840e-01 8.71252239e-01
4.23639178e-01 -2.89101005e-01 7.07606196e-01 2.64241576e-01
-5.94214439e-01 2.40516454e-01 6.62687004e-01 1.24808602e-01
6.94706023e-01 -1.31852233e+00 -4.00989950e-01 6.85547113e-01
3.33100706e-01 5.21791220e-01 2.68685937e-01 -2.84618020e-01
-1.55225503e+00 -8.95265877e-01 4.86994892e-01 -5.06938756e-01
4.05755252e-01 -1.55821502e-01 -1.00344920e+00 8.88806760e-01
4.75205570e-01 9.66608748e-02 1.42981148e+00 2.49502584e-01
-6.02865577e-01 -8.07006434e-02 -9.57229912e-01 -2.89133172e-02
8.96784544e-01 -7.33724892e-01 -5.04169405e-01 3.79100144e-01
2.94500113e-01 -2.16236021e-02 -1.10695839e+00 2.64121264e-01
3.96641999e-01 -1.29025185e+00 1.25179029e+00 -4.07595694e-01
4.32476491e-01 -4.48222965e-01 -7.56376386e-01 -1.25751483e+00
-5.82566977e-01 -1.37309581e-01 -4.50349212e-01 1.58905125e+00
5.52620590e-02 -3.99404943e-01 6.23329878e-01 4.16267961e-01
1.39521375e-01 -6.37280822e-01 -8.35895956e-01 -5.76249063e-01
-3.39961201e-02 -1.05953686e-01 2.94559836e-01 1.28132820e+00
3.71663272e-02 5.10529757e-01 -1.01959634e+00 4.24847126e-01
1.18561256e+00 9.43587422e-01 5.92646718e-01 -1.32928240e+00
-5.68517089e-01 3.64206098e-02 -6.06538774e-03 -1.01120746e+00
4.53831047e-01 -1.21817946e+00 -9.34665948e-02 -1.60046196e+00
1.14768302e+00 -8.16301107e-02 -3.96385342e-01 1.76457286e-01
-2.34198228e-01 -2.05136128e-02 4.80851114e-01 6.29112720e-01
-1.16364348e+00 7.37863958e-01 1.11909544e+00 -1.69308230e-01
1.49526313e-01 1.52955770e-01 -1.34169221e+00 8.32335770e-01
7.88306057e-01 -5.79499245e-01 -6.39270067e-01 -3.41647774e-01
9.65962708e-02 4.87495750e-01 2.39842355e-01 -5.90417922e-01
2.33707920e-01 -2.44175136e-01 4.27898556e-01 -8.70512128e-01
4.75149512e-01 -9.86053288e-01 2.96550661e-01 5.44059463e-02
-2.13158220e-01 -6.58196509e-02 -3.87655020e-01 1.06492019e+00
-3.49178463e-01 -8.58297944e-02 7.36459136e-01 -4.83258337e-01
-6.80170834e-01 3.09875458e-01 -2.36000083e-02 3.49106491e-01
1.04931140e+00 1.28307547e-02 -3.00812364e-01 -6.48555756e-01
-9.67975020e-01 4.09684926e-01 6.12351060e-01 3.70401680e-01
7.42973804e-01 -1.67901647e+00 -7.51992881e-01 3.78119089e-02
2.60080755e-01 -8.30521062e-02 7.55107284e-01 8.58851016e-01
3.36736768e-01 4.04931098e-01 5.17556025e-03 -5.88285148e-01
-1.31597590e+00 1.01334643e+00 4.55305055e-02 -2.92161971e-01
-4.92621481e-01 4.23311830e-01 5.89912415e-01 -7.27860689e-01
3.11086662e-02 4.40539986e-01 -2.92494029e-01 2.95024395e-01
5.63361645e-01 6.68054998e-01 -3.02846462e-01 -1.01497149e+00
-4.07534182e-01 8.57788563e-01 -1.33390471e-01 9.27834213e-02
1.59866691e+00 -8.04245949e-01 -2.83470154e-01 7.35002816e-01
1.54066074e+00 -2.01605055e-02 -1.30364299e+00 -6.02886856e-01
-2.04819292e-01 -6.67716563e-01 -1.25324354e-01 -3.87861252e-01
-1.49936235e+00 7.12642193e-01 4.05304849e-01 1.31472036e-01
1.25163913e+00 5.05803108e-01 4.44214195e-01 2.93038696e-01
3.25639278e-01 -8.74061882e-01 2.75108933e-01 1.18824035e-01
8.25421989e-01 -1.61450160e+00 3.01257968e-01 -4.26660508e-01
-1.00308859e+00 8.41902554e-01 6.01738751e-01 -5.67670614e-02
8.22285414e-01 -1.03835255e-01 -1.87021703e-01 -4.89487112e-01
-8.24012458e-01 -2.69725800e-01 4.50935364e-01 6.61089897e-01
4.28075910e-01 2.74890125e-01 1.42504960e-01 5.63559353e-01
2.85972118e-01 -2.65914440e-01 4.80956584e-01 7.60049701e-01
-2.03792229e-01 -1.16232002e+00 -4.60668266e-01 3.88525337e-01
-3.08066785e-01 2.02503875e-01 -5.08431315e-01 4.41153020e-01
1.34565711e-01 1.32671642e+00 -2.94422179e-01 -2.96021312e-01
2.31096465e-02 5.17232791e-02 4.66019243e-01 -5.92085481e-01
9.23762396e-02 3.07348907e-01 -2.69248426e-01 -4.37995762e-01
-1.10749102e+00 -1.10107827e+00 -8.89744103e-01 -1.88487560e-01
-3.76420975e-01 1.68651462e-01 2.46111289e-01 8.49319160e-01
3.65816414e-01 2.81291068e-01 1.10558116e+00 -7.03826725e-01
-4.06521738e-01 -5.36850691e-01 -9.84683096e-01 7.10389495e-01
2.69421101e-01 -6.59104109e-01 -3.57372075e-01 2.19226688e-01]
|
[8.378715515136719, 4.541141033172607]
|
ce2135e9-64e2-46f1-a9f0-07ab50d7434b
|
attacking-perceptual-similarity-metrics-1
|
2305.0884
| null |
https://arxiv.org/abs/2305.08840v1
|
https://arxiv.org/pdf/2305.08840v1.pdf
|
Attacking Perceptual Similarity Metrics
|
Perceptual similarity metrics have progressively become more correlated with human judgments on perceptual similarity; however, despite recent advances, the addition of an imperceptible distortion can still compromise these metrics. In our study, we systematically examine the robustness of these metrics to imperceptible adversarial perturbations. Following the two-alternative forced-choice experimental design with two distorted images and one reference image, we perturb the distorted image closer to the reference via an adversarial attack until the metric flips its judgment. We first show that all metrics in our study are susceptible to perturbations generated via common adversarial attacks such as FGSM, PGD, and the One-pixel attack. Next, we attack the widely adopted LPIPS metric using spatial-transformation-based adversarial perturbations (stAdv) in a white-box setting to craft adversarial examples that can effectively transfer to other similarity metrics in a black-box setting. We also combine the spatial attack stAdv with PGD ($\ell_\infty$-bounded) attack to increase transferability and use these adversarial examples to benchmark the robustness of both traditional and recently developed metrics. Our benchmark provides a good starting point for discussion and further research on the robustness of metrics to imperceptible adversarial perturbations.
|
['Feng Liu', 'Abhijay Ghildyal']
|
2023-05-15
|
attacking-perceptual-similarity-metrics
|
https://openreview.net/forum?id=VUcI0pKic8l
|
https://openreview.net/pdf?id=VUcI0pKic8l
| null |
['experimental-design']
|
['methodology']
|
[ 7.00060368e-01 -2.67304033e-01 4.26035672e-01 -1.78721279e-01
-9.24602330e-01 -1.23465383e+00 7.40209818e-01 7.44182020e-02
-4.50300157e-01 4.78674054e-01 7.27177486e-02 -4.91664737e-01
-5.54427914e-02 -6.30869269e-01 -6.33573532e-01 -6.12212658e-01
-3.58900994e-01 -4.49068099e-01 2.49642834e-01 -3.25823963e-01
4.26218390e-01 6.45889997e-01 -1.21247363e+00 1.80560518e-02
7.43176162e-01 8.65775168e-01 -2.96182603e-01 1.02084970e+00
7.11193025e-01 7.52430916e-01 -1.10247517e+00 -7.87293851e-01
9.79348361e-01 -4.25337166e-01 -7.26322055e-01 -2.00725853e-01
7.87657738e-01 -3.32899719e-01 -4.57851738e-01 1.28316903e+00
7.78948724e-01 3.43639821e-01 6.90592885e-01 -1.40982938e+00
-9.90701973e-01 1.60421133e-01 -4.54633415e-01 2.88498789e-01
7.26131976e-01 6.96172178e-01 8.07023644e-01 -6.06133580e-01
5.31623542e-01 1.43610942e+00 6.85948253e-01 4.83334541e-01
-1.37252951e+00 -6.88824713e-01 3.75925153e-02 3.29021484e-01
-1.37017989e+00 -4.33483899e-01 7.41815150e-01 -3.97931546e-01
5.14671028e-01 6.90088153e-01 -1.00037754e-02 1.14985430e+00
2.83706754e-01 1.03745766e-01 1.61377323e+00 -4.05988902e-01
1.65994033e-01 6.62477687e-02 -4.23322499e-01 1.45300105e-01
1.17601618e-01 6.91112459e-01 3.85572724e-02 -2.70134151e-01
6.87419236e-01 -3.20207119e-01 -4.45235074e-01 -2.55515873e-01
-1.26887739e+00 6.48247421e-01 7.79602647e-01 5.30671775e-02
-7.87958056e-02 2.04403475e-01 1.10973142e-01 8.58991325e-01
9.92766246e-02 1.20909357e+00 -3.00215185e-02 6.74756318e-02
-4.84676659e-01 3.03164184e-01 7.99966455e-01 7.23279655e-01
6.26973271e-01 1.04366913e-01 -3.49210680e-01 6.25966430e-01
-1.02341443e-01 7.59808004e-01 1.82172418e-01 -1.29975712e+00
3.75895381e-01 7.66883418e-02 3.82465214e-01 -1.43201864e+00
-2.03435510e-01 -2.86397696e-01 -5.59656143e-01 8.26033950e-01
6.89173341e-01 -1.73317328e-01 -7.28452265e-01 1.80483449e+00
1.78707540e-01 -7.97267929e-02 8.28794837e-02 1.01766646e+00
3.09229732e-01 4.49242651e-01 5.54683544e-02 3.09459921e-02
1.01161003e+00 -7.22330451e-01 -3.62636417e-01 -1.80593714e-01
2.60699093e-01 -1.11735904e+00 1.46182120e+00 3.33363086e-01
-9.31157112e-01 -6.86717093e-01 -1.35061550e+00 1.21328138e-01
-4.99368340e-01 -5.68803251e-01 -4.70025986e-02 1.11776555e+00
-9.89742458e-01 6.76191807e-01 -3.84031028e-01 -3.23491961e-01
2.38055632e-01 2.82799512e-01 -4.52803850e-01 -2.29008973e-01
-1.38171184e+00 1.25222361e+00 -3.53778154e-02 -8.29327181e-02
-1.02214217e+00 -7.40494609e-01 -7.73168564e-01 -2.60376304e-01
2.54369408e-01 -5.32620549e-01 9.65442479e-01 -1.05889308e+00
-1.45841575e+00 8.45016181e-01 3.25791240e-01 -4.72424805e-01
8.09676766e-01 -1.93551388e-02 -7.87554026e-01 2.16155365e-01
-1.03505336e-01 6.74069703e-01 1.15818477e+00 -1.49994600e+00
-8.30628276e-02 -2.06412803e-02 5.31688750e-01 1.03141956e-01
-1.38223097e-01 3.25986952e-01 2.31486470e-01 -1.14985669e+00
-3.95247310e-01 -1.08150458e+00 -1.62663445e-01 2.24917501e-01
-5.75523555e-01 6.17074966e-01 6.60893679e-01 -4.63462353e-01
1.11460638e+00 -2.45129585e+00 -2.08587185e-01 4.50049639e-01
1.40777618e-01 5.09235144e-01 -6.84877217e-01 7.05903709e-01
-3.70551080e-01 3.24699879e-01 -4.61313039e-01 1.71440858e-02
2.49255553e-01 1.48076396e-02 -5.21279335e-01 6.81325734e-01
1.71040252e-01 9.07401800e-01 -9.56500828e-01 -1.91280231e-01
2.33909085e-01 4.53965336e-01 -4.97779459e-01 1.72897145e-01
2.26249844e-01 4.41615552e-01 7.75013044e-02 3.55614007e-01
8.29635382e-01 1.90438181e-01 -3.21746588e-01 -1.95601910e-01
1.21509165e-01 -2.01257750e-01 -1.23768950e+00 1.08307755e+00
-3.89068842e-01 8.07487190e-01 -1.25527173e-01 -6.17655039e-01
8.72826338e-01 -1.22741181e-02 -8.26920643e-02 -7.07468987e-01
6.10566661e-02 8.80089328e-02 4.07516956e-01 -1.34695649e-01
2.62587488e-01 -3.75758350e-01 -2.40003332e-01 4.30944681e-01
-4.04859364e-01 -4.44260567e-01 -6.81543872e-02 1.26237944e-01
1.36312330e+00 -2.63005078e-01 3.21662545e-01 -2.27297097e-01
6.95879042e-01 -2.95543373e-01 3.05567831e-01 1.06909120e+00
-5.81226945e-01 9.04291451e-01 4.91535217e-01 -2.61278391e-01
-1.10080409e+00 -1.62134528e+00 6.65949564e-03 8.91026914e-01
4.27109063e-01 -9.89744589e-02 -7.49484360e-01 -6.77155137e-01
2.22031519e-01 7.86137640e-01 -7.57675290e-01 -4.99914795e-01
-4.24451500e-01 -4.77956802e-01 9.82744277e-01 4.86881137e-01
7.30275631e-01 -9.35240567e-01 -4.84046578e-01 -2.31628701e-01
1.04380235e-01 -1.19914138e+00 -7.87902713e-01 -2.65226513e-01
-1.75924584e-01 -1.12596285e+00 -7.12722480e-01 -5.07483840e-01
7.21224189e-01 1.76447049e-01 8.94335806e-01 -1.14621662e-01
-2.58001775e-01 5.92975378e-01 -4.19552803e-01 -3.65006089e-01
-8.00103784e-01 -5.10323942e-01 2.22730517e-01 1.98173493e-01
-1.06107607e-01 -5.71731806e-01 -9.03752506e-01 8.46289456e-01
-1.22858810e+00 -5.67843735e-01 3.49730164e-01 6.15408540e-01
3.08197886e-01 1.90337673e-01 3.73041928e-01 -4.75523323e-01
9.60326254e-01 -2.42828801e-01 -4.64466661e-01 7.02215284e-02
-4.38250929e-01 -1.64707959e-01 1.12793672e+00 -6.97512090e-01
-6.63652241e-01 -4.53945875e-01 6.85132965e-02 -4.76443470e-01
-2.62702346e-01 1.76740795e-01 -3.05305034e-01 -8.03350091e-01
1.20216501e+00 -4.38048989e-02 -1.29496694e-01 -1.52623445e-01
6.31375611e-01 4.66409922e-01 9.42085922e-01 -4.46846753e-01
1.63184130e+00 4.25862402e-01 5.62345609e-02 -5.82520723e-01
-3.74789357e-01 1.69960633e-01 -4.40475941e-01 -6.43571392e-02
6.97467506e-01 -5.14951289e-01 -8.44610035e-01 6.93483531e-01
-9.27936852e-01 -5.34518182e-01 -4.83891219e-01 1.93504438e-01
-6.46242380e-01 6.39466524e-01 -4.05831665e-01 -4.93446350e-01
7.26333037e-02 -1.17835128e+00 6.03297830e-01 2.62865406e-02
-3.70783061e-01 -1.02888262e+00 7.22166747e-02 9.50149819e-02
6.98427379e-01 7.56945074e-01 6.86023533e-01 -8.04784536e-01
-3.96638960e-01 -4.81659859e-01 -1.91134527e-01 8.41169298e-01
3.19310874e-01 -1.77801140e-02 -8.77874553e-01 -5.81597328e-01
2.06255212e-01 -2.19531849e-01 4.44923520e-01 -4.71984930e-02
1.02153444e+00 -6.38379693e-01 2.00881839e-01 7.79349983e-01
1.29017675e+00 2.34120578e-01 9.07102704e-01 6.15012944e-01
5.80971897e-01 5.37986577e-01 4.19153064e-01 7.72987083e-02
-7.41328904e-03 7.10285544e-01 4.63350147e-01 -2.03269631e-01
-1.66304082e-01 -1.22299306e-01 5.55265963e-01 2.17853919e-01
4.23921049e-02 -4.40221518e-01 -7.80070007e-01 2.65317321e-01
-1.12448072e+00 -1.15043211e+00 2.65117079e-01 2.50781155e+00
7.93837368e-01 2.33413309e-01 1.77359462e-01 4.97051060e-01
8.82381082e-01 5.21298707e-01 -4.89028245e-01 -8.05740058e-01
-3.12111408e-01 2.86831886e-01 6.58524930e-01 6.57963514e-01
-1.24788284e+00 7.08029985e-01 7.17842674e+00 7.06384540e-01
-1.13229144e+00 -9.48848724e-02 5.87536752e-01 -3.42315361e-02
-2.58362174e-01 -1.17953099e-01 -3.90970148e-02 5.97418547e-01
8.20144296e-01 -4.55126107e-01 7.89355814e-01 4.40842897e-01
1.44231260e-01 1.04583204e-01 -1.16410959e+00 7.66741574e-01
1.63798988e-01 -8.25821042e-01 1.54067338e-01 -5.56354970e-02
8.67090940e-01 -4.76967752e-01 7.52236009e-01 -7.57994428e-02
5.58315158e-01 -1.36264610e+00 5.35382688e-01 2.64846087e-01
9.99337256e-01 -6.33823574e-01 5.90889394e-01 -1.58473760e-01
-9.37932730e-01 -1.83105618e-01 -1.59653902e-01 -2.82478184e-01
-3.86084057e-02 1.33497223e-01 -5.00873208e-01 4.15774077e-01
3.56346220e-01 1.96968287e-01 -8.53252530e-01 1.01959002e+00
-4.11633670e-01 3.86831403e-01 -9.84014273e-02 3.27941179e-01
2.91860610e-01 1.08577773e-01 8.52187574e-01 1.01454818e+00
1.77107990e-01 -6.13063807e-03 -1.58588350e-01 8.58287036e-01
6.57049790e-02 -5.83860390e-02 -8.01849246e-01 2.88620412e-01
7.07381248e-01 1.01279128e+00 -4.72780854e-01 3.10983732e-02
-2.55035907e-01 1.20211446e+00 -2.30217248e-01 6.50888324e-01
-9.67793226e-01 -9.72705424e-01 1.05149555e+00 2.31619980e-02
2.19557554e-01 -2.03873277e-01 -2.93976545e-01 -7.93348432e-01
5.26740029e-02 -1.43515086e+00 2.33789518e-01 -8.99307013e-01
-1.60972309e+00 6.17193997e-01 1.43321931e-01 -1.55165446e+00
-1.62285924e-01 -5.16776800e-01 -9.12535012e-01 1.11178935e+00
-1.17615891e+00 -8.63558233e-01 -4.45027590e-01 8.17137718e-01
-7.00859278e-02 -1.25907278e-02 6.90580428e-01 -4.56620753e-03
-3.82991761e-01 1.12984788e+00 7.11930078e-03 3.38228732e-01
1.07701683e+00 -1.32025194e+00 9.73159909e-01 1.32730222e+00
2.40687896e-02 6.66046023e-01 1.03503883e+00 -2.79477149e-01
-1.08100724e+00 -1.09520519e+00 2.51603097e-01 -7.31940925e-01
9.18068111e-01 -3.31061035e-01 -8.21599722e-01 5.30454934e-01
1.82401702e-01 3.13048840e-01 5.61142504e-01 -6.08223021e-01
-8.07982922e-01 -1.90348819e-01 -1.53586590e+00 1.02969933e+00
1.09446514e+00 -8.17513824e-01 -5.56710780e-01 -1.16252499e-02
8.52942526e-01 -2.52554595e-01 -1.05674756e+00 4.00643766e-01
5.90835214e-01 -1.09628630e+00 1.34119308e+00 -4.70342994e-01
3.21877867e-01 -5.39972663e-01 -5.01330554e-01 -1.62668908e+00
-3.54408026e-01 -1.27290964e+00 4.06649500e-01 1.07811475e+00
1.70098290e-01 -1.06504834e+00 1.44807428e-01 5.82647681e-01
6.54272214e-02 -3.21307063e-01 -9.21884239e-01 -1.29088843e+00
5.57428956e-01 -4.09223258e-01 7.72860646e-01 8.72805536e-01
-3.11501771e-02 -2.73606956e-01 -3.11457902e-01 3.24012697e-01
6.22220218e-01 -3.92481685e-01 1.00689793e+00 -5.79089463e-01
-3.34950507e-01 -5.83617806e-01 -1.00283062e+00 -7.21810520e-01
-1.03211239e-01 -5.17141283e-01 4.75292280e-02 -8.19410384e-01
-2.83486962e-01 -2.34070420e-01 -5.02751350e-01 1.60960570e-01
-5.18702984e-01 6.75620377e-01 5.67137301e-01 5.06095774e-02
-2.45367646e-01 2.59307563e-01 1.34489179e+00 -1.83796570e-01
1.05707534e-01 -1.32642910e-01 -9.18224514e-01 5.38724601e-01
6.89188600e-01 -3.74862134e-01 -5.05018771e-01 -2.17090726e-01
1.26175284e-01 -2.36486077e-01 7.29157090e-01 -1.26188028e+00
4.56979079e-03 -1.93551868e-01 2.71045446e-01 3.23448300e-01
1.75168946e-01 -7.24963486e-01 9.29189399e-02 4.61216182e-01
-5.39187431e-01 3.34897101e-01 3.21749151e-01 4.95359451e-01
-7.18043521e-02 6.91214055e-02 1.12518287e+00 2.29299068e-01
-5.18822134e-01 2.09373655e-03 -2.28302613e-01 2.89524585e-01
1.27496314e+00 -4.73625869e-01 -8.06277394e-01 -6.82942986e-01
-5.65019667e-01 -1.35213464e-01 1.00260067e+00 4.23677415e-01
6.14018202e-01 -1.33046043e+00 -7.86457300e-01 3.33626270e-01
1.43181816e-01 -7.31135130e-01 8.57604817e-02 7.11289823e-01
-6.14053369e-01 6.03493117e-03 -5.39846897e-01 -2.45920688e-01
-1.14851248e+00 1.10218823e+00 5.36087215e-01 4.80805114e-02
-1.22121364e-01 8.03208709e-01 3.50352168e-01 -2.11125910e-01
1.50221229e-01 -1.85588047e-01 1.80322543e-01 -3.64523083e-01
5.41202724e-01 6.65391862e-01 -1.46267489e-01 -6.45241618e-01
-5.20792603e-01 6.90235019e-01 2.73320049e-01 -2.89637834e-01
6.85013652e-01 -2.46546820e-01 1.66566074e-01 1.08411200e-01
1.45801568e+00 3.45865309e-01 -1.32829952e+00 5.50182164e-02
-4.48494554e-01 -8.64028335e-01 -4.61684972e-01 -9.95846927e-01
-8.37349355e-01 7.66604722e-01 6.97443724e-01 6.17119849e-01
1.37069380e+00 -4.45136875e-01 5.13726056e-01 1.96939558e-01
2.11685121e-01 -5.62065840e-01 2.32707605e-01 1.77265242e-01
1.21457982e+00 -1.09898770e+00 1.54553819e-02 -3.49961847e-01
-6.77708268e-01 7.10426986e-01 4.85167295e-01 -4.53007877e-01
4.89402801e-01 2.51658291e-01 4.84556466e-01 2.44271144e-01
-3.19140881e-01 -6.91927737e-03 5.07965565e-01 1.11257124e+00
-6.28680661e-02 -1.04809180e-01 -1.16067147e-02 7.25670978e-02
-6.11051142e-01 -4.81836349e-01 6.92177713e-01 7.17647076e-01
-6.91664964e-02 -1.06965172e+00 -8.39336574e-01 1.25152782e-01
-5.02872527e-01 -2.21916795e-01 -5.90767443e-01 9.12347913e-01
2.44543497e-02 1.28045142e+00 -2.47818306e-02 -7.86302924e-01
5.72508693e-01 -3.21920365e-01 5.23398817e-01 -2.19198450e-01
-7.47004807e-01 -5.26755512e-01 -1.16436385e-01 -7.30507672e-01
-2.68344611e-01 -5.84878385e-01 -5.22608042e-01 -6.21298969e-01
2.87021641e-02 -1.57172516e-01 3.02006274e-01 6.68154180e-01
2.39851579e-01 2.94340223e-01 1.10044312e+00 -8.67136121e-01
-9.24852431e-01 -8.17396879e-01 -2.68846869e-01 9.52240586e-01
4.98258173e-01 -5.73780358e-01 -9.39949036e-01 -2.63505057e-02]
|
[5.597940921783447, 7.848174571990967]
|
979e1d65-c43f-4f2c-a514-ce9a83a5fd67
|
lazier-than-lazy-greedy
|
1409.7938
| null |
http://arxiv.org/abs/1409.7938v3
|
http://arxiv.org/pdf/1409.7938v3.pdf
|
Lazier Than Lazy Greedy
|
Is it possible to maximize a monotone submodular function faster than the
widely used lazy greedy algorithm (also known as accelerated greedy), both in
theory and practice? In this paper, we develop the first linear-time algorithm
for maximizing a general monotone submodular function subject to a cardinality
constraint. We show that our randomized algorithm, STOCHASTIC-GREEDY, can
achieve a $(1-1/e-\varepsilon)$ approximation guarantee, in expectation, to the
optimum solution in time linear in the size of the data and independent of the
cardinality constraint. We empirically demonstrate the effectiveness of our
algorithm on submodular functions arising in data summarization, including
training large-scale kernel methods, exemplar-based clustering, and sensor
placement. We observe that STOCHASTIC-GREEDY practically achieves the same
utility value as lazy greedy but runs much faster. More surprisingly, we
observe that in many practical scenarios STOCHASTIC-GREEDY does not evaluate
the whole fraction of data points even once and still achieves
indistinguishable results compared to lazy greedy.
|
['Amin Karbasi', 'Baharan Mirzasoleiman', 'Ashwinkumar Badanidiyuru', 'Andreas Krause', 'Jan Vondrak']
|
2014-09-28
| null | null | null | null |
['data-summarization']
|
['miscellaneous']
|
[-1.19052399e-02 4.14428562e-01 -3.82530361e-01 -3.52915555e-01
-1.24653161e+00 -8.48822474e-01 -3.31890970e-01 4.00938660e-01
-2.95736700e-01 9.01445627e-01 1.85230419e-01 -9.23886746e-02
-6.43224776e-01 -8.32042634e-01 -1.01559865e+00 -8.22760820e-01
-5.56499422e-01 9.66566980e-01 -1.37940437e-01 1.98290229e-01
1.24797285e-01 3.57705891e-01 -1.25859535e+00 -1.47881702e-01
8.75389516e-01 1.08410335e+00 3.00929695e-01 6.15906060e-01
1.08007953e-01 4.83697176e-01 -4.73514408e-01 -3.36725526e-02
6.10789537e-01 -4.17423666e-01 -7.02532947e-01 5.00829577e-01
6.21357083e-01 -2.43699849e-01 -2.32460767e-01 1.01636100e+00
4.10735458e-01 -4.19647055e-04 4.14829791e-01 -1.38988709e+00
-2.19635874e-01 9.96902764e-01 -7.52149165e-01 -2.01523587e-01
4.25741941e-01 -2.25525513e-01 1.33140087e+00 -5.90390384e-01
7.82197416e-01 1.09890187e+00 7.94787645e-01 1.47038892e-01
-1.44832766e+00 -4.11516607e-01 1.57014132e-01 -3.01720768e-01
-1.43937802e+00 -2.68753439e-01 5.54556549e-01 9.62797627e-02
5.51974595e-01 6.31094456e-01 6.38013005e-01 6.03395030e-02
-1.76942453e-01 1.21035397e+00 7.41041303e-01 -1.44060209e-01
6.16438091e-01 8.56413916e-02 2.41503716e-01 7.40257978e-01
8.43592644e-01 -2.76948392e-01 -4.31590527e-01 -7.58025110e-01
3.77207488e-01 2.47741878e-01 -4.47176725e-01 -7.64159441e-01
-1.12609541e+00 8.61594439e-01 3.43942463e-01 4.66911960e-03
-5.58885515e-01 8.52246284e-01 1.71692967e-01 6.06070340e-01
3.70247394e-01 5.41089594e-01 -5.90813160e-01 -1.20230228e-01
-1.43921661e+00 6.05342627e-01 1.18201399e+00 1.44011390e+00
8.10708523e-01 -1.82692796e-01 -1.15694888e-02 5.43642879e-01
-1.24755546e-01 8.92326832e-01 -2.06833318e-01 -1.32901430e+00
6.04154885e-01 6.98958516e-01 2.87990987e-01 -8.44043791e-01
-7.11350977e-01 -2.50135422e-01 -6.70884013e-01 -1.00501835e-01
3.44349116e-01 -4.54097450e-01 -4.73180890e-01 1.71554101e+00
5.39690793e-01 -4.07397658e-01 -8.30784291e-02 7.25799441e-01
2.18102902e-01 7.43192971e-01 -5.51498890e-01 -9.54302371e-01
7.21954823e-01 -5.85088134e-01 -5.19070983e-01 2.66934365e-01
1.08006525e+00 -4.73912001e-01 8.12545359e-01 6.15916371e-01
-1.65522969e+00 3.05334836e-01 -8.59875083e-01 9.21784192e-02
6.16135746e-02 -2.55685240e-01 8.78818572e-01 7.67165482e-01
-1.16548514e+00 3.73459846e-01 -6.07290149e-01 -4.77062881e-01
5.02247989e-01 5.80426872e-01 -2.19846576e-01 -3.85957599e-01
-1.86004922e-01 1.86663687e-01 3.84404033e-01 -5.16776264e-01
-9.13289547e-01 -9.55105901e-01 -6.66591167e-01 1.74144283e-01
7.29476333e-01 -8.18201482e-01 1.02260590e+00 -3.60302091e-01
-8.17488313e-01 7.15451062e-01 -4.57651645e-01 -7.64939964e-01
4.49774474e-01 -5.99739254e-02 5.30064702e-01 1.88528851e-01
9.07784253e-02 6.07537448e-01 3.51414561e-01 -1.32381082e+00
-7.50263870e-01 -6.81684673e-01 2.48178363e-01 3.48114073e-01
-4.69048053e-01 -3.67712200e-01 -2.41260171e-01 -3.80658954e-01
4.58459675e-01 -9.45240498e-01 -6.63464189e-01 2.40688041e-01
-4.60751295e-01 -3.10824990e-01 5.78097045e-01 -1.54343084e-01
1.39351285e+00 -2.17267609e+00 3.28784823e-01 5.68495452e-01
4.88564938e-01 -4.19043839e-01 -1.16121275e-02 7.46448278e-01
4.36583310e-01 -1.08896546e-01 -6.60484612e-01 -3.07602227e-01
1.64185941e-01 3.95477355e-01 -1.10536493e-01 9.93570268e-01
-9.55539942e-01 8.91688824e-01 -9.53252137e-01 -3.85173827e-01
-1.37411803e-01 -4.81476188e-01 -7.41260111e-01 -2.04289272e-01
-5.49803019e-01 -4.51109439e-01 -3.43137443e-01 1.03051198e+00
1.07694638e+00 -4.42268372e-01 4.23203796e-01 2.31229573e-01
1.86908141e-01 -2.32323289e-01 -1.53052080e+00 1.74129784e+00
-3.40608716e-01 3.96409661e-01 7.44617820e-01 -1.19883561e+00
6.46360278e-01 4.08430398e-02 1.04445112e+00 -3.72645766e-01
7.92715047e-03 5.88302314e-01 -6.72545969e-01 -1.57427758e-01
5.55231392e-01 -1.61397859e-01 -4.30960953e-01 9.17301416e-01
-2.06556424e-01 -9.64103639e-02 2.53390372e-01 4.77480441e-01
1.41411924e+00 -7.17545390e-01 4.77123111e-01 -7.91494906e-01
-7.99173266e-02 2.61277974e-01 4.85203803e-01 1.00922096e+00
2.73974687e-01 5.52648723e-01 5.64507663e-01 -3.87962490e-01
-9.80448842e-01 -1.21958435e+00 -1.03282705e-01 1.13092601e+00
4.95339245e-01 -4.46074188e-01 -7.26155818e-01 -7.08993614e-01
7.97204673e-01 7.73942053e-01 -5.52513123e-01 2.85775453e-01
-2.61192352e-01 -1.00654376e+00 2.12102547e-01 3.63558143e-01
2.01362580e-01 -5.98384798e-01 -5.01210332e-01 3.52440387e-01
8.58811513e-02 -9.56514597e-01 -8.71378362e-01 4.09999341e-01
-1.38082218e+00 -1.05453193e+00 -6.33791268e-01 -5.92163920e-01
1.01728523e+00 8.24301660e-01 1.00351202e+00 8.96380246e-02
-2.20703647e-01 7.92842746e-01 -3.17406535e-01 -4.43432271e-01
1.11970671e-01 1.45662904e-01 -8.36479664e-03 -4.00654912e-01
3.47373933e-01 -6.89394295e-01 -5.67166269e-01 9.55893323e-02
-9.37402546e-01 -5.64065039e-01 4.30628300e-01 5.26730359e-01
8.83415520e-01 2.03114107e-01 5.17321885e-01 -1.06234217e+00
7.72818387e-01 -7.88227856e-01 -9.28526163e-01 3.60010266e-01
-7.83878744e-01 2.22777098e-01 5.94580650e-01 -1.48562759e-01
-1.80149660e-01 2.78562963e-01 2.83359230e-01 -1.82087466e-01
3.99998277e-01 5.25959015e-01 -4.02749702e-02 -2.39261135e-01
6.49836421e-01 4.47481722e-01 3.90677191e-02 -3.80640000e-01
4.74579662e-01 6.21767223e-01 3.93693238e-01 -7.20931411e-01
6.96031392e-01 1.10939431e+00 3.52766216e-01 -9.43360388e-01
-8.72599483e-01 -8.43111694e-01 -4.80422042e-02 1.57135263e-01
2.50836521e-01 -8.42956901e-01 -9.03869390e-01 7.83518329e-02
-9.11846399e-01 -3.48053485e-01 -8.82978737e-01 1.68168738e-01
-1.01581359e+00 2.43405491e-01 -3.39865834e-02 -1.03714609e+00
-6.01643741e-01 -5.26517808e-01 1.12135708e+00 -1.62185133e-01
2.11785603e-02 -9.82462466e-01 3.44597697e-02 1.19189478e-01
1.56143621e-01 5.40195942e-01 5.16388357e-01 -4.92917687e-01
-8.97804558e-01 -2.56516367e-01 -1.24337144e-01 -5.28942514e-03
-1.55956119e-01 -6.02376461e-01 -3.16719770e-01 -7.49327481e-01
-2.96113230e-02 -3.28555882e-01 8.58235657e-01 8.39558423e-01
1.26482308e+00 -8.46358180e-01 -5.12299180e-01 9.29532468e-01
1.69412720e+00 -2.16060549e-01 3.92646641e-01 -2.32179998e-03
3.62393439e-01 2.66567886e-01 7.48368740e-01 9.84574735e-01
5.10442913e-01 3.83014858e-01 5.63323855e-01 1.17649399e-01
3.49227339e-01 -1.04888000e-01 2.98381150e-01 3.34207356e-01
3.07398826e-01 -3.54170322e-01 -3.51452172e-01 1.14640939e+00
-2.17053986e+00 -9.41645503e-01 -1.64706334e-01 2.49080157e+00
6.53657794e-01 -2.01516166e-01 5.66085756e-01 1.84203774e-01
4.61614072e-01 5.33368811e-02 -6.80112898e-01 -4.94861066e-01
-2.89459199e-01 -3.94691620e-03 1.36228395e+00 5.86728215e-01
-7.25292802e-01 6.27888799e-01 7.33518219e+00 8.85293663e-01
-6.28004372e-01 8.55350867e-02 4.66405749e-01 -9.10190225e-01
-8.02005708e-01 -2.48003807e-02 -7.99232602e-01 4.04441327e-01
5.87742984e-01 -6.35544419e-01 8.39744568e-01 1.30106080e+00
7.09350407e-02 -4.47335720e-01 -1.28741181e+00 1.25711608e+00
8.99986401e-02 -1.59439373e+00 -2.08128422e-01 5.33133209e-01
1.31073749e+00 3.90937403e-02 -1.79433793e-01 -2.30664596e-01
5.48454940e-01 -8.63771856e-01 5.55676699e-01 5.05112223e-02
6.07167661e-01 -9.32370663e-01 5.33510983e-01 6.67573631e-01
-1.09688663e+00 -4.29786712e-01 -6.76171482e-01 1.24857366e-01
4.26801354e-01 1.10773242e+00 -7.07565010e-01 4.09022987e-01
6.60080075e-01 2.59923697e-01 1.30992115e-01 1.32374656e+00
3.41679692e-01 5.85266590e-01 -1.19609582e+00 -2.96806693e-01
3.38813037e-01 -1.15045518e-01 7.93364823e-01 1.09168029e+00
4.32861328e-01 2.15990067e-01 3.93508404e-01 5.92634976e-01
-4.99926835e-01 1.45683214e-01 -7.45564163e-01 1.76493358e-02
8.20748806e-01 1.06837606e+00 -5.18017530e-01 -2.00391591e-01
-1.76264137e-01 7.73717761e-01 2.40670666e-01 2.90813863e-01
-5.72682202e-01 -3.59529793e-01 4.63759094e-01 4.89374697e-01
5.73907018e-01 -4.42858130e-01 -7.58430779e-01 -6.44193828e-01
4.80584472e-01 -5.66364288e-01 8.09732139e-01 -1.19965538e-01
-1.11100531e+00 -1.44658357e-01 2.53418267e-01 -7.76180744e-01
-6.13749810e-02 -2.86560655e-01 -3.15647334e-01 2.30488166e-01
-1.11606467e+00 -6.74332976e-01 -3.01008999e-01 6.54963970e-01
2.15097860e-01 1.24685720e-01 4.43386644e-01 3.68537940e-02
7.21172094e-02 7.84531474e-01 6.09775901e-01 -5.41187882e-01
1.85100019e-01 -1.47456443e+00 -1.76153600e-01 5.54686904e-01
8.15569684e-02 4.34794456e-01 8.97349358e-01 -3.73383850e-01
-1.85081327e+00 -8.97292376e-01 7.52384126e-01 -1.60098791e-01
4.91633713e-01 -2.59667635e-01 -4.02569473e-01 8.06018770e-01
-1.84538826e-01 9.68032628e-02 7.33213902e-01 1.83797061e-01
6.19086251e-03 -4.94181156e-01 -1.72075319e+00 2.30586573e-01
1.50701392e+00 -2.59913541e-02 -1.92749858e-01 7.78652549e-01
8.14680874e-01 -3.17808837e-01 -8.60939443e-01 4.21575904e-01
4.95262176e-01 -8.37383687e-01 7.52252996e-01 -3.37053269e-01
-2.11362764e-01 -9.36939120e-02 -7.02213049e-01 -9.83876526e-01
-1.19386241e-01 -1.08326375e+00 -6.02239907e-01 6.40223384e-01
3.26416224e-01 -6.52189732e-01 1.30749083e+00 3.80196154e-01
1.23110950e-01 -1.18652022e+00 -1.18649805e+00 -1.22480166e+00
2.66837515e-03 -2.67914504e-01 5.90639949e-01 6.66321039e-01
2.40303829e-01 -1.46089658e-01 -4.04449075e-01 1.31806746e-01
1.16053450e+00 5.70536792e-01 1.12435579e+00 -1.09859860e+00
-4.86054391e-01 -2.12592199e-01 -3.62207621e-01 -1.71000278e+00
-2.83481628e-01 -9.83395934e-01 -8.32264796e-02 -1.67308271e+00
6.43924236e-01 -7.32246101e-01 -1.13100253e-01 2.99141318e-01
7.99265951e-02 -5.47867268e-02 1.37263879e-01 1.31685704e-01
-1.19500482e+00 3.86611581e-01 1.05734992e+00 -2.40788478e-02
-4.94191796e-01 1.51343778e-01 -1.17493021e+00 4.31581706e-01
5.27929723e-01 -6.62708759e-01 -4.76438522e-01 -4.41451311e-01
4.15647745e-01 3.40295672e-01 -1.83879286e-02 -8.17529082e-01
3.41126382e-01 -4.92677480e-01 -3.49110007e-01 -8.86408627e-01
4.43550274e-02 -1.01822889e+00 4.28396389e-02 5.53528965e-01
-2.24799141e-01 8.45759735e-02 -1.07012786e-01 6.95913196e-01
1.08944632e-01 -3.96223247e-01 7.31959820e-01 -1.92030817e-01
-1.70815751e-01 5.62874079e-01 4.61040996e-02 3.25213909e-01
1.53112125e+00 -4.61534172e-01 -2.43794888e-01 -7.32621908e-01
-4.81368959e-01 7.71870673e-01 7.79253542e-01 -3.24935943e-01
5.62623501e-01 -1.20275939e+00 -8.14723372e-01 -3.88128161e-01
8.68100971e-02 4.37478542e-01 1.14638709e-01 1.09729505e+00
-6.24985456e-01 5.46790659e-01 4.50510502e-01 -6.56700969e-01
-1.22465444e+00 7.88327932e-01 3.86456549e-02 -3.55089694e-01
-5.97585797e-01 8.80879879e-01 6.75023394e-03 -4.00337636e-01
5.57571232e-01 -2.22719744e-01 9.19615388e-01 -2.50529386e-02
4.39999133e-01 8.78954828e-01 -2.03675345e-01 1.31800994e-01
-3.74445438e-01 4.76054311e-01 1.13720454e-01 -7.87402391e-02
1.54518938e+00 -7.96498731e-02 -3.18741709e-01 2.18531236e-01
1.35085368e+00 2.41118744e-01 -9.87274826e-01 -2.03921199e-01
-1.00676589e-01 -5.39082289e-01 -2.03617021e-01 -3.55984151e-01
-1.00482619e+00 1.91280901e-01 1.30476281e-01 5.40615976e-01
1.18999982e+00 5.18643022e-01 1.20944715e+00 8.99138391e-01
1.05800271e+00 -1.33812582e+00 -2.36149877e-01 -3.02025769e-02
7.69909143e-01 -1.02543044e+00 5.19133389e-01 -4.31396067e-01
-2.22935408e-01 8.21513772e-01 4.22824509e-02 -5.05101919e-01
6.73966229e-01 4.88667578e-01 -4.41613674e-01 -4.17664587e-01
-8.36589575e-01 -1.36653751e-01 -1.43568128e-01 4.31885153e-01
-1.36054292e-01 2.47713253e-01 -6.75880015e-01 2.98873007e-01
-2.89902151e-01 6.26284676e-03 4.79675710e-01 1.08615744e+00
-1.16941881e+00 -7.72505462e-01 -5.35914898e-01 7.73084939e-01
-2.66831636e-01 1.19683184e-01 -4.37857717e-01 7.28067517e-01
-1.98893979e-01 8.99915457e-01 1.52388126e-01 -9.26032886e-02
3.84729616e-02 -3.61173868e-01 7.22848594e-01 -5.24585366e-01
-2.48109490e-01 -1.05362035e-01 -5.43348528e-02 -6.94517732e-01
-2.26801634e-01 -9.14558768e-01 -1.42207205e+00 -7.55469441e-01
-2.88436919e-01 3.69213492e-01 6.04247034e-01 6.65412188e-01
3.46082717e-01 -1.18674979e-01 9.54456687e-01 -2.77687579e-01
-1.08269560e+00 -5.00463903e-01 -9.44221973e-01 1.74705967e-01
2.35204339e-01 -2.54594356e-01 -5.31041801e-01 -4.27988052e-01]
|
[6.594685077667236, 4.924299716949463]
|
4770029a-654d-41b7-9ce2-6c460d5ed68a
|
analyzing-berts-knowledge-of-hypernymy-via
| null | null |
https://aclanthology.org/2021.blackboxnlp-1.20
|
https://aclanthology.org/2021.blackboxnlp-1.20.pdf
|
Analyzing BERT’s Knowledge of Hypernymy via Prompting
|
The high performance of large pretrained language models (LLMs) such as BERT on NLP tasks has prompted questions about BERT’s linguistic capabilities, and how they differ from humans’. In this paper, we approach this question by examining BERT’s knowledge of lexical semantic relations. We focus on hypernymy, the “is-a” relation that relates a word to a superordinate category. We use a prompting methodology to simply ask BERT what the hypernym of a given word is. We find that, in a setting where all hypernyms are guessable via prompting, BERT knows hypernyms with up to 57% accuracy. Moreover, BERT with prompting outperforms other unsupervised models for hypernym discovery even in an unconstrained scenario. However, BERT’s predictions and performance on a dataset containing uncommon hyponyms and hypernyms indicate that its knowledge of hypernymy is still limited.
|
['David Mareček', 'Michael Hanna']
| null | null | null | null |
emnlp-blackboxnlp-2021-11
|
['hypernym-discovery']
|
['natural-language-processing']
|
[ 9.41114686e-03 6.95130765e-01 -4.58568364e-01 -2.91922599e-01
5.74735142e-02 -7.10010767e-01 7.82019913e-01 5.76614559e-01
-6.82097733e-01 7.00183868e-01 3.66380543e-01 -5.65198600e-01
-3.94502968e-01 -1.15233529e+00 -3.19277853e-01 -1.69169486e-01
-3.55059020e-02 1.06258869e+00 2.08456770e-01 -6.95772529e-01
-1.49826586e-01 2.50312209e-01 -1.47536588e+00 2.22813264e-01
6.08282268e-01 5.49378395e-01 1.06182180e-01 2.33653307e-01
-4.23778623e-01 9.04007614e-01 -5.34486651e-01 -9.12862718e-01
2.29218259e-01 1.49445519e-01 -1.46894085e+00 -4.38565463e-01
4.52304035e-01 -1.63087472e-01 -2.96352148e-01 9.37030494e-01
1.05312094e-01 1.73651084e-01 5.93514442e-01 -1.35881007e+00
-5.59884071e-01 1.24441850e+00 3.59217599e-02 4.42544609e-01
5.78280151e-01 -1.01184975e-02 1.91761518e+00 -4.97709960e-01
8.35559249e-01 1.34982467e+00 4.96480852e-01 6.66903138e-01
-1.27372086e+00 -7.77908981e-01 -7.90068507e-02 4.83664781e-01
-1.50524354e+00 -1.69110626e-01 2.40554601e-01 -3.90334308e-01
1.44131768e+00 9.81767103e-02 4.71581310e-01 7.35028684e-01
-4.48625773e-01 5.09564936e-01 9.09346581e-01 -7.79271662e-01
4.95029725e-02 1.31242484e-01 3.78192246e-01 5.22199810e-01
6.48366749e-01 9.82405245e-02 -6.81164384e-01 -2.11332366e-01
5.77763796e-01 -3.46558094e-01 6.07942119e-02 -1.81375042e-01
-1.20489573e+00 9.80504811e-01 4.26701516e-01 5.63875556e-01
-1.99110761e-01 -6.98177470e-03 2.46627435e-01 3.43203753e-01
-4.14255587e-03 1.32793236e+00 -7.09176898e-01 7.68454224e-02
-5.81319511e-01 3.22604328e-01 1.39328313e+00 1.19283998e+00
8.18941474e-01 -3.06343079e-01 3.41198385e-01 8.38538468e-01
1.12103038e-01 2.74951130e-01 8.28593493e-01 -7.66483545e-01
4.68511134e-01 8.09627771e-01 1.97342023e-01 -1.04667664e+00
-7.56268263e-01 3.13991904e-02 -2.98103631e-01 -3.66904080e-01
4.64654684e-01 4.40235510e-02 -4.99871850e-01 2.01511359e+00
3.32068801e-01 5.79282977e-02 3.48311722e-01 7.19611406e-01
1.25292361e+00 3.13757926e-01 7.74471998e-01 -5.77318929e-02
1.50165284e+00 -2.89605647e-01 -4.98166949e-01 -6.74246132e-01
1.02745354e+00 -4.30376619e-01 1.17712414e+00 1.59923062e-01
-6.72995448e-01 -8.91140774e-02 -9.26529527e-01 -1.92889109e-01
-7.95116484e-01 -2.30256230e-01 1.36745799e+00 5.14160335e-01
-7.30198860e-01 6.28333211e-01 -3.10636938e-01 -1.03392148e+00
1.35188535e-01 5.02912700e-01 -5.27223766e-01 -6.72876164e-02
-1.78299570e+00 1.52369773e+00 1.05237567e+00 -3.06479812e-01
-6.82615697e-01 -5.40226758e-01 -1.11082447e+00 1.90621108e-01
7.73774683e-01 -4.91297871e-01 1.28456962e+00 -6.14369154e-01
-7.61444092e-01 1.42604184e+00 -9.33784917e-02 -8.75050783e-01
-1.06123962e-01 -8.15660805e-02 -8.14213634e-01 2.14674339e-01
4.07576144e-01 8.44566047e-01 4.45797116e-01 -9.82058108e-01
-6.53352141e-01 -9.66416448e-02 7.33007371e-01 1.90148279e-01
-4.40611333e-01 3.23095173e-01 3.08222353e-01 -2.55775303e-01
1.72390357e-01 -6.38401926e-01 6.95920037e-03 -3.56087774e-01
-3.78655046e-01 -6.88440919e-01 1.87454835e-01 -1.09020971e-01
1.33524334e+00 -1.76685798e+00 -3.02798867e-01 2.48469338e-01
6.48817360e-01 4.27967072e-01 -1.37881562e-01 7.72070885e-01
-4.31272179e-01 4.04262871e-01 -1.81270372e-02 4.57113177e-01
3.74436915e-01 8.39056611e-01 -5.12283146e-01 1.63935110e-01
6.99704140e-02 1.14357233e+00 -1.14349043e+00 -6.10012650e-01
2.57497221e-01 -2.67913222e-01 -5.30063510e-01 1.21141724e-01
-3.68325412e-01 -2.97271162e-01 -3.13989036e-02 5.40716708e-01
2.37988189e-01 -4.47408319e-01 6.29604578e-01 -9.18544605e-02
2.83390373e-01 8.80658388e-01 -1.01852429e+00 1.06401932e+00
-4.95120227e-01 5.07181585e-01 -2.39560351e-01 -1.01399243e+00
8.69490743e-01 4.35653836e-01 3.19545805e-01 -3.92442405e-01
1.29329398e-01 3.61260802e-01 6.81451976e-01 -6.69102550e-01
3.57494503e-01 -8.11716676e-01 -5.59805185e-02 5.63520968e-01
2.95821041e-01 -2.57948846e-01 5.17281353e-01 5.49476027e-01
1.18722093e+00 -2.67433912e-01 9.63746607e-01 -4.99056071e-01
4.46476132e-01 3.48268718e-01 4.71137583e-01 6.99339271e-01
-1.17574506e-01 -2.58372158e-01 4.81143832e-01 -5.16297936e-01
-8.74794185e-01 -1.12207150e+00 -3.98475558e-01 1.27994096e+00
2.29750291e-01 -1.12209284e+00 -8.86039361e-02 -5.93412995e-01
2.06005096e-01 1.22569942e+00 -3.25896233e-01 -2.07563356e-01
-4.75572914e-01 -4.88827974e-01 8.00077856e-01 3.57952327e-01
7.30595961e-02 -1.39649868e+00 -1.05749500e+00 2.45756760e-01
-3.42531413e-01 -1.38816214e+00 2.67292798e-01 4.63970602e-01
-4.48212802e-01 -1.35113299e+00 1.53052181e-01 -8.18203688e-01
3.85331988e-01 -2.41944611e-01 1.63735473e+00 4.70728934e-01
-3.60117972e-01 2.66561478e-01 -5.48158824e-01 -4.75646764e-01
-3.47840160e-01 1.82255954e-01 3.21606010e-01 -6.89791918e-01
1.26383400e+00 -7.29985833e-01 -5.52045554e-03 2.24046052e-01
-1.01203847e+00 -2.38985971e-01 2.89228529e-01 6.54709458e-01
-4.23148535e-02 1.98800191e-01 5.37836492e-01 -1.34326982e+00
6.30775273e-01 -6.59644544e-01 -4.10420030e-01 2.76963055e-01
-6.80215299e-01 1.73354581e-01 4.75536555e-01 -5.27580678e-01
-8.11107874e-01 -1.73837453e-01 4.45704199e-02 1.66131958e-01
-4.66722846e-01 5.74917793e-01 -2.90896207e-01 1.21608041e-01
9.61500108e-01 -2.53065795e-01 -3.90123397e-01 -3.82936925e-01
7.13552594e-01 4.13789302e-01 6.31108344e-01 -1.07612491e+00
1.01178133e+00 2.86019862e-01 2.00252548e-01 -8.79883468e-01
-1.40605283e+00 -9.16901231e-01 -8.78543794e-01 2.52894461e-01
6.85585022e-01 -8.58712792e-01 -1.23936880e+00 -3.38973939e-01
-1.19566739e+00 -4.32014376e-01 -3.85636240e-01 4.02073234e-01
-4.83948976e-01 3.43136817e-01 -5.19140780e-01 -6.39885604e-01
-1.27214447e-01 -4.02902484e-01 7.01868117e-01 -2.53619581e-01
-1.12864220e+00 -1.08856571e+00 -2.71614581e-01 2.57917106e-01
1.99855000e-01 -1.27691835e-01 1.54942226e+00 -1.62044537e+00
-1.45514280e-01 -1.08901478e-01 -2.27713883e-01 -6.86198846e-03
1.32539511e-01 -4.49664086e-01 -7.69926846e-01 2.25600645e-01
-3.79060060e-02 -6.62477314e-01 5.04972339e-01 -2.72113770e-01
6.91886008e-01 -8.89160812e-01 -2.81099290e-01 3.88892144e-01
1.40312243e+00 -1.87299162e-01 4.35034782e-01 4.63697314e-01
5.27482152e-01 1.07624292e+00 6.36645079e-01 3.02968740e-01
5.67778587e-01 3.26683819e-01 3.75821263e-01 3.64219189e-01
1.49196669e-01 -6.91958010e-01 -2.87261195e-02 3.82162213e-01
3.53153676e-01 -1.87021136e-01 -1.33985376e+00 7.56161809e-01
-1.64292240e+00 -9.34210122e-01 -3.21446061e-01 2.05067325e+00
1.24798214e+00 2.69438457e-02 2.89780628e-02 3.10694464e-02
5.73983729e-01 9.49813798e-02 -3.59349191e-01 -3.59846711e-01
-2.66320437e-01 4.75820243e-01 5.88108242e-01 6.53785706e-01
-9.94688392e-01 1.73515964e+00 6.61022377e+00 7.62398720e-01
-4.51940954e-01 4.94995639e-02 5.67896515e-02 2.25186080e-01
-5.08082569e-01 3.69864523e-01 -9.58921015e-01 7.05756992e-02
8.00768614e-01 -4.17098522e-01 6.04860008e-01 7.15146720e-01
-2.63432235e-01 -1.09505393e-01 -1.49338126e+00 9.12315309e-01
1.71889514e-01 -1.08429456e+00 2.90710747e-01 -8.18879530e-02
3.33637059e-01 -1.13383204e-01 -5.95569193e-01 4.47185755e-01
8.56692195e-01 -1.44887269e+00 3.39673400e-01 -1.72741771e-01
7.29204535e-01 -3.37531716e-01 6.09328866e-01 6.93935215e-01
-1.04410732e+00 -3.19056064e-01 -6.57285094e-01 -4.85950142e-01
2.02181235e-01 2.96161830e-01 -1.35373044e+00 6.74099028e-02
2.48206198e-01 5.42033255e-01 -6.87847555e-01 8.21736217e-01
-1.13672400e+00 6.06513977e-01 -6.46438181e-01 -1.44184440e-01
2.29478151e-01 1.48649246e-01 3.64595532e-01 1.13964343e+00
-1.32941768e-01 6.85956419e-01 3.07943046e-01 9.87047732e-01
-8.82703736e-02 2.81598896e-01 -6.65824533e-01 -3.26565832e-01
8.21346223e-01 1.35887694e+00 -5.05815625e-01 -6.00150406e-01
-1.08087592e-01 4.92919832e-01 5.49641609e-01 1.85532406e-01
-4.13027734e-01 -2.15723693e-01 5.33522725e-01 2.77837187e-01
-1.88261077e-01 -1.47001579e-01 -2.14916721e-01 -9.15029824e-01
-2.65266776e-01 -7.23762035e-01 1.01473224e+00 -9.33794320e-01
-1.49452734e+00 2.76784301e-01 4.41474229e-01 -6.30432904e-01
-5.17240345e-01 -1.08250308e+00 -4.02719826e-01 6.76246405e-01
-1.44293988e+00 -1.15654290e+00 4.53789085e-02 5.47008634e-01
1.64292336e-01 3.70875932e-02 1.07698834e+00 1.11633923e-03
8.23491216e-02 4.49930936e-01 -8.57300997e-01 3.37561727e-01
6.36954188e-01 -1.35213006e+00 3.32162648e-01 6.01679862e-01
5.77535212e-01 1.06200337e+00 8.96168709e-01 -8.18025291e-01
-1.07574058e+00 -6.53450608e-01 1.73242998e+00 -8.10346246e-01
1.33267558e+00 -2.00982153e-01 -9.32187438e-01 1.01798952e+00
-9.51186940e-02 -1.88510731e-01 9.12947774e-01 5.41242361e-01
-8.94522071e-01 2.29744032e-01 -1.07120955e+00 6.04541481e-01
1.47283185e+00 -6.65816128e-01 -1.53817880e+00 9.04278636e-01
9.92548704e-01 5.61071672e-02 -8.29850852e-01 3.83884668e-01
2.49034867e-01 -4.57639575e-01 9.76866066e-01 -1.39846122e+00
3.54941577e-01 -1.61201432e-01 -4.06018615e-01 -1.21262038e+00
-2.11783350e-01 -5.89116931e-01 1.25128310e-02 1.12632668e+00
5.80439925e-01 -7.04784989e-01 6.65070713e-01 1.11358225e+00
2.71998584e-01 -1.10289916e-01 -7.65943944e-01 -1.07756639e+00
1.42193794e-01 -6.98159456e-01 7.83610702e-01 1.22307909e+00
9.18471456e-01 8.16212595e-01 8.28767475e-03 3.60719138e-03
4.02119845e-01 1.59195855e-01 4.57429469e-01 -1.75682259e+00
-1.42918751e-01 -2.20063105e-01 -5.62095642e-01 -6.98749840e-01
9.38936770e-01 -1.32028031e+00 -4.24679853e-02 -1.39732778e+00
8.26854333e-02 -5.51511824e-01 1.82549939e-01 9.79477584e-01
8.36944655e-02 -8.21423456e-02 1.68264106e-01 1.57915056e-01
-5.80561519e-01 1.08652517e-01 8.17747891e-01 -9.10796039e-03
1.24117449e-01 -3.16989422e-01 -8.44948411e-01 1.19971693e+00
8.24514508e-01 -5.18422425e-01 -3.63226354e-01 -4.98068541e-01
9.03590918e-01 -4.16884065e-01 5.56215703e-01 -5.20710289e-01
4.34621155e-01 -3.59840721e-01 -1.95411727e-01 2.52147131e-02
3.43235821e-01 -1.14882195e+00 -2.12600499e-01 4.15314078e-01
-7.00450897e-01 5.86087182e-02 -1.21001368e-02 2.93527693e-01
-1.16494931e-01 -6.60999894e-01 4.81031209e-01 -4.93422002e-01
-1.17731750e+00 1.47264481e-01 -2.36257613e-01 5.80493927e-01
7.79309690e-01 7.49624968e-02 -4.29890811e-01 -3.54688853e-01
-7.23301530e-01 2.53922373e-01 3.47493380e-01 3.38749021e-01
4.56944376e-01 -1.00254357e+00 -4.25888807e-01 -9.99316946e-02
5.55321813e-01 -1.52826801e-01 -4.78438079e-01 4.99196231e-01
-3.84825051e-01 6.67851627e-01 1.18359588e-01 -4.98245731e-02
-1.09991753e+00 8.29367399e-01 1.40753672e-01 -2.64450498e-02
-6.07025445e-01 9.85886097e-01 1.39045179e-01 -8.05482328e-01
1.19368151e-01 -2.06221566e-01 -4.76498544e-01 2.05282703e-01
3.52772623e-01 5.56959100e-02 -1.96208224e-01 -9.34706151e-01
-5.54695845e-01 1.77368432e-01 1.19722061e-01 -1.52065575e-01
1.13125658e+00 9.91204008e-02 -5.61393797e-01 5.68752229e-01
8.62240255e-01 -7.04367086e-02 -1.04634263e-01 -5.42555809e-01
6.49538755e-01 -5.86171031e-01 -5.94063580e-01 -9.10499394e-01
-3.36856961e-01 6.69542313e-01 -1.47825897e-01 2.51447141e-01
5.52089155e-01 6.71934128e-01 7.11749971e-01 1.05541265e+00
5.48087120e-01 -9.85947073e-01 -3.80157888e-01 7.82698512e-01
6.18219197e-01 -1.17007697e+00 5.18271960e-02 -8.24096978e-01
-6.97263300e-01 9.36465561e-01 8.33249032e-01 8.44800845e-02
5.09466887e-01 -9.57057811e-04 2.62603555e-02 -7.49744356e-01
-9.12980080e-01 -7.44677365e-01 6.57824650e-02 8.16281736e-01
4.84984905e-01 4.55286562e-01 -5.14517367e-01 6.27384126e-01
-1.11395884e+00 -3.87671858e-01 5.19922435e-01 3.41084749e-01
-6.84310019e-01 -1.05240858e+00 1.73834749e-02 5.90554476e-01
-4.08450991e-01 -7.81080067e-01 -7.32698560e-01 9.93915498e-01
3.36960942e-01 1.17487884e+00 -6.60276487e-02 -2.22957730e-01
-4.87793870e-02 3.20094764e-01 3.75799149e-01 -1.41131055e+00
-5.19694149e-01 -6.70852363e-01 5.89969218e-01 -4.82442886e-01
-3.98385912e-01 -4.16052729e-01 -1.56384289e+00 -2.28157163e-01
-3.37027818e-01 3.88786227e-01 1.58546910e-01 1.40121043e+00
-2.57881045e-01 -1.61881328e-01 -1.07854895e-01 1.41483754e-01
-8.50734651e-01 -9.22972322e-01 -7.65712142e-01 8.51215422e-01
-1.63620085e-01 -6.87105060e-01 -5.15486062e-01 -1.77290514e-01]
|
[10.040850639343262, 8.770288467407227]
|
dc2c9655-7cb4-4500-850b-ee5444dddb61
|
learning-goal-conditioned-value-functions
| null | null |
https://openreview.net/forum?id=BkesGnCcFX
|
https://openreview.net/pdf?id=BkesGnCcFX
|
Learning Goal-Conditioned Value Functions with one-step Path rewards rather than Goal-Rewards
|
Multi-goal reinforcement learning (MGRL) addresses tasks where the desired goal state can change for every trial. State-of-the-art algorithms model these problems such that the reward formulation depends on the goals, to associate them with high reward. This dependence introduces additional goal reward resampling steps in algorithms like Hindsight Experience Replay (HER) that reuse trials in which the agent fails to reach the goal by recomputing rewards as if reached states were psuedo-desired goals. We propose a reformulation of goal-conditioned value functions for MGRL that yields a similar algorithm, while removing the dependence of reward functions on the goal. Our formulation thus obviates the requirement of reward-recomputation that is needed by HER and its extensions. We also extend a closely related algorithm, Floyd-Warshall Reinforcement Learning, from tabular domains to deep neural networks for use as a baseline. Our results are competetive with HER while substantially improving sampling efficiency in terms of reward computation.
|
['Jeffrey M. Siskind', 'Jason J. Corso', 'Shurjo Banerjee', 'Vikas Dhiman']
| null | null | null | null |
iclr-2019-5
|
['multi-goal-reinforcement-learning']
|
['methodology']
|
[-7.51862302e-02 2.36643910e-01 -4.80100334e-01 -1.26310736e-01
-9.61305976e-01 -6.20721817e-01 6.37744963e-01 3.78435493e-01
-1.06440866e+00 1.38172400e+00 2.03825846e-01 -2.37233892e-01
-3.74373972e-01 -8.98107708e-01 -6.35393322e-01 -8.84648860e-01
-4.17546064e-01 6.53318822e-01 2.07054690e-01 -5.35219252e-01
5.05691767e-01 4.14188415e-01 -1.36831141e+00 -6.17951229e-02
4.45745289e-01 7.65959024e-01 1.53067231e-01 8.39279652e-01
8.55892897e-02 1.26060152e+00 -5.02786934e-01 -3.55615765e-02
2.99796969e-01 -5.77067673e-01 -9.69605148e-01 9.40909889e-03
-1.42857090e-01 -7.34271646e-01 -2.70640433e-01 9.30308640e-01
3.05201441e-01 5.52644849e-01 4.25682813e-01 -1.53558528e+00
-3.99874032e-01 8.06300879e-01 -5.67203939e-01 -6.15400858e-02
2.46997669e-01 1.21165253e-01 1.20270050e+00 -5.18788062e-02
5.06088853e-01 1.26764405e+00 2.73359597e-01 6.94274366e-01
-1.18037117e+00 -2.98431307e-01 3.99596691e-01 3.56799424e-01
-4.70057279e-01 -2.51146227e-01 6.06900156e-01 1.31361276e-01
1.13773930e+00 3.15530524e-02 7.76989639e-01 9.30807412e-01
1.55777261e-01 9.64078546e-01 1.34513414e+00 -5.06547630e-01
9.53519225e-01 -4.62376654e-01 -1.38352856e-01 5.23016095e-01
2.31380731e-01 7.07063258e-01 -4.09621894e-01 -2.42545292e-01
8.97367299e-01 7.27326497e-02 1.14962801e-01 -8.21618319e-01
-1.08509099e+00 1.14842772e+00 3.62498850e-01 1.73463240e-01
-7.02989280e-01 7.91175485e-01 4.31620538e-01 8.50596189e-01
-6.45828396e-02 6.71206355e-01 -5.58268547e-01 -3.88005108e-01
-8.04761052e-01 7.31644869e-01 8.25155675e-01 7.23033190e-01
9.54686165e-01 2.01345846e-01 -3.31541300e-01 4.10092115e-01
1.04846302e-02 3.59554261e-01 6.29387379e-01 -1.57565248e+00
2.38054931e-01 1.38500094e-01 8.00396860e-01 -1.94942132e-01
-5.11080503e-01 -3.30268532e-01 -1.01059243e-01 8.13878357e-01
6.06996179e-01 -4.18715864e-01 -7.55527020e-01 1.89357102e+00
3.66677225e-01 6.72219023e-02 5.06540954e-01 7.80432999e-01
-3.43854316e-02 5.38987994e-01 8.37332830e-02 -4.04947162e-01
8.09483528e-01 -1.02227724e+00 -4.70787644e-01 -3.10318381e-01
7.30117321e-01 -1.80223897e-01 9.38398659e-01 5.20147026e-01
-1.27015007e+00 -1.67662099e-01 -9.89335299e-01 3.05312634e-01
1.62950642e-02 -4.35427904e-01 9.60467637e-01 5.42441428e-01
-1.18450594e+00 1.20539677e+00 -1.04851723e+00 -1.74903497e-01
3.59883457e-01 3.53735358e-01 -5.04300520e-02 9.27730501e-02
-1.01664865e+00 1.15961051e+00 5.71603298e-01 -4.50410068e-01
-1.48717594e+00 -1.22601047e-01 -8.62618446e-01 1.30987123e-01
9.45465267e-01 -2.46360287e-01 1.89936161e+00 -1.08136141e+00
-1.85514677e+00 5.25874674e-01 2.26043984e-01 -8.71645808e-01
4.67419833e-01 -2.06331834e-01 -5.79552874e-02 4.50229980e-02
1.84548780e-01 7.22356498e-01 8.48250985e-01 -9.69580293e-01
-8.30636024e-01 -2.97763914e-01 4.66410339e-01 5.91453731e-01
6.66720048e-02 -2.25268260e-01 7.73211569e-02 -1.90646157e-01
-2.16616929e-01 -8.55025947e-01 -7.64522314e-01 -2.10459977e-01
1.91344306e-01 -5.09190679e-01 2.29005411e-01 1.53242633e-01
8.03746343e-01 -1.71061361e+00 1.59514651e-01 1.38626039e-01
6.55128583e-02 -1.33938178e-01 -6.24262393e-01 6.90178275e-01
4.24773805e-02 -4.27283943e-01 1.60618164e-02 -1.37760743e-01
2.76109159e-01 5.53227663e-01 -2.80276924e-01 5.97527266e-01
-6.78542852e-02 9.63760316e-01 -1.40784812e+00 -2.19933242e-01
1.43668786e-01 -3.79236519e-01 -7.92079210e-01 3.93304646e-01
-5.97578347e-01 1.52436838e-01 -6.41739190e-01 3.22471857e-01
2.25239247e-01 -2.02157170e-01 5.40900290e-01 5.98347723e-01
-2.54669636e-01 5.84195733e-01 -1.21782827e+00 1.84231174e+00
-3.20005238e-01 -3.45706269e-02 1.45746619e-02 -1.02912509e+00
8.25357795e-01 1.52824163e-01 6.41462564e-01 -7.63482749e-01
2.68579841e-01 1.31190166e-01 -7.45132342e-02 -1.25194147e-01
7.73750067e-01 -3.21052969e-01 -3.04185480e-01 7.84834862e-01
9.35958475e-02 -3.57934862e-01 4.91766781e-01 1.51950151e-01
1.20942163e+00 9.29964185e-01 6.76418781e-01 -1.22140750e-01
6.51345029e-02 3.09504747e-01 4.85177904e-01 1.27152753e+00
-3.58552039e-01 -3.77862491e-02 7.98330009e-01 -4.24532741e-01
-1.03284287e+00 -1.07395864e+00 3.61762404e-01 1.54484880e+00
1.16717137e-01 -1.12923481e-01 -4.69895661e-01 -7.40924478e-01
9.60590243e-02 1.03692412e+00 -7.27243602e-01 -5.14171738e-03
-6.75171673e-01 -4.99331206e-01 2.51939327e-01 4.48857218e-01
3.00337791e-01 -1.63412154e+00 -1.48787713e+00 6.45508111e-01
7.38139451e-02 -4.19544041e-01 -2.38724887e-01 8.60305607e-01
-9.29324985e-01 -1.11379814e+00 -7.78408706e-01 -4.51846659e-01
3.32488358e-01 -7.24864006e-03 1.12172747e+00 -1.32115483e-01
1.98000148e-01 5.57318807e-01 -3.48588705e-01 -2.71125257e-01
-4.00668830e-01 -1.71366706e-01 4.10606451e-02 -6.67300701e-01
5.03384829e-01 -7.33486235e-01 -4.32976931e-01 -2.65615750e-02
-7.65317917e-01 -1.46148920e-01 4.59553033e-01 1.00929809e+00
4.46904242e-01 -2.13746667e-01 9.53155756e-01 -5.81456423e-01
9.09257889e-01 -4.81672764e-01 -1.08538210e+00 7.07230642e-02
-6.65101886e-01 6.90620720e-01 6.68195367e-01 -5.28043330e-01
-6.94072127e-01 -2.55588852e-02 7.48701915e-02 -2.91476786e-01
-5.95275015e-02 2.77483255e-01 2.93890893e-01 2.23821789e-01
9.31909561e-01 1.91746891e-01 2.64364511e-01 -1.77350223e-01
6.02366209e-01 1.09343410e-01 4.05259222e-01 -9.77603793e-01
3.75598341e-01 1.23390503e-01 2.07467481e-01 -9.68497545e-02
-8.13158572e-01 -2.69169182e-01 1.33015811e-01 -1.53315067e-01
4.22489494e-01 -7.58139193e-01 -1.24579406e+00 1.21263765e-01
-8.11306059e-01 -7.87100136e-01 -9.58057463e-01 5.37100255e-01
-1.42948055e+00 3.51341933e-01 -5.81412554e-01 -1.11058545e+00
-1.01115875e-01 -9.64946866e-01 6.19734824e-01 2.90488213e-01
-9.94718745e-02 -7.28966236e-01 4.42856520e-01 -3.24130446e-01
4.11048740e-01 1.26259625e-01 7.33184814e-01 -6.59053683e-01
-5.14507949e-01 3.13933045e-01 1.34758115e-01 -7.79270157e-02
-1.08421497e-01 -8.22869003e-01 -5.74870884e-01 -5.15892208e-01
-6.85616769e-03 -1.11711228e+00 7.68602788e-01 6.19171798e-01
6.81945741e-01 -3.79216194e-01 6.20452687e-02 1.26890242e-01
1.52452946e+00 4.49062318e-01 5.83959699e-01 9.89136577e-01
-1.25909507e-01 4.19684947e-01 1.02382731e+00 9.30256784e-01
3.50283682e-01 3.69338751e-01 1.01397538e+00 2.09973946e-01
4.17304844e-01 -2.76105016e-01 5.41793227e-01 -8.70016292e-02
-1.13163769e-01 -1.27170548e-01 -3.64492744e-01 7.22135186e-01
-2.35146451e+00 -1.20220423e+00 4.31187063e-01 2.28810072e+00
1.03652763e+00 3.12954247e-01 5.96510112e-01 2.66869478e-02
3.71923685e-01 2.72934526e-01 -1.05602169e+00 -6.49591446e-01
3.71945530e-01 5.68351686e-01 6.74023092e-01 7.97608793e-01
-8.21617186e-01 1.21993577e+00 6.63614702e+00 6.20490849e-01
-6.12507761e-01 4.61233146e-02 3.16570640e-01 -4.22540963e-01
-3.21951687e-01 1.57769352e-01 -6.99009657e-01 -1.03248619e-01
8.62621188e-01 -2.98071116e-01 1.12178719e+00 1.17773831e+00
1.18747935e-01 -4.10208553e-01 -1.32052577e+00 6.73622966e-01
-3.19729656e-01 -1.07317770e+00 -4.63697076e-01 -1.06877089e-01
5.12520730e-01 4.25161906e-02 -1.18200980e-01 8.93837929e-01
1.45529914e+00 -8.74785721e-01 7.71594405e-01 2.02192709e-01
5.18168747e-01 -1.07445085e+00 3.45925510e-01 5.05578279e-01
-7.26935387e-01 -4.04408783e-01 -5.48990667e-01 -6.00699186e-01
4.14805636e-02 -2.76344363e-03 -9.35576499e-01 3.77721965e-01
2.46726006e-01 4.29869771e-01 3.43028689e-03 7.74834514e-01
-5.17017007e-01 3.43905598e-01 -2.93573081e-01 -6.30199015e-01
9.26687241e-01 -2.44817361e-01 3.84056389e-01 6.63067639e-01
3.41896445e-01 -1.66144758e-01 5.29870033e-01 7.38391221e-01
2.22968534e-01 -3.29464413e-02 -4.52975094e-01 -4.45594887e-05
4.34533447e-01 1.06124365e+00 -5.53374171e-01 -2.42839903e-01
-1.29932076e-01 8.52316499e-01 7.30519831e-01 3.43717068e-01
-8.27872753e-01 -1.72644526e-01 6.68942869e-01 -1.37270480e-01
3.99736613e-01 -2.46204376e-01 9.72697586e-02 -1.01740110e+00
-3.03505421e-01 -8.53167236e-01 5.19453049e-01 -6.29091680e-01
-1.05004311e+00 3.38575453e-01 7.43856002e-03 -1.08772957e+00
-8.78907621e-01 -3.66270542e-01 -3.03359032e-01 7.65178740e-01
-1.99618089e+00 -6.75183594e-01 2.11470127e-01 8.62575710e-01
4.69115227e-01 -1.76699117e-01 9.12364781e-01 -4.44387227e-01
1.23008871e-02 2.82140642e-01 8.77315253e-02 -3.35792065e-01
5.34508288e-01 -1.55118084e+00 2.13302419e-01 6.09420717e-01
-1.13527015e-01 2.21244395e-01 8.02021146e-01 -5.95540166e-01
-1.15389359e+00 -8.00608277e-01 5.07733345e-01 5.09975888e-02
8.28703463e-01 1.99387580e-01 -2.17719793e-01 8.64628017e-01
3.62700969e-01 -2.24634826e-01 2.60729045e-01 4.01244283e-01
-9.76124033e-02 7.60381892e-02 -1.16659307e+00 9.87448215e-01
8.47276270e-01 -5.85375428e-02 -8.84098947e-01 1.14903569e-01
5.01644075e-01 -3.99873376e-01 -4.65605170e-01 -8.66843313e-02
4.49568480e-01 -1.06355691e+00 7.66580880e-01 -1.07961524e+00
4.86432225e-01 -1.18797526e-01 -1.83654860e-01 -1.59465063e+00
-5.40844560e-01 -8.98304999e-01 -3.48339945e-01 3.09818566e-01
4.47349139e-02 -5.25651693e-01 9.89386082e-01 5.28747737e-01
-7.86507353e-02 -7.64904737e-01 -1.09276783e+00 -9.31039989e-01
2.69578546e-01 -2.10642919e-01 6.71328247e-01 6.88281655e-01
3.71409655e-01 2.39595070e-01 -7.20351934e-01 -2.51735240e-01
7.61913717e-01 3.37492675e-01 7.30671942e-01 -9.07400429e-01
-7.55849540e-01 -3.66067708e-01 1.47796571e-01 -1.23897624e+00
8.04538950e-02 -8.56979668e-01 1.08335473e-01 -1.40085173e+00
2.20816559e-03 -5.54377496e-01 -5.99475503e-01 9.97108221e-01
-1.11784535e-02 -3.42795104e-01 4.91191536e-01 -8.81712213e-02
-1.06785035e+00 6.93083525e-01 1.45073104e+00 2.23849043e-01
-5.26794016e-01 -1.94281060e-02 -7.97028780e-01 5.72855115e-01
1.11173832e+00 -7.82246828e-01 -5.27566493e-01 -3.62270288e-02
3.42236131e-01 8.73064756e-01 1.76213846e-01 -9.98764396e-01
6.50527254e-02 -7.84891009e-01 2.86725551e-01 -4.67145234e-01
3.07568848e-01 -7.44946897e-01 -1.23152263e-01 7.82686591e-01
-8.62559617e-01 3.12897176e-01 4.72167181e-03 4.47044104e-01
2.85556763e-01 -8.53755176e-01 8.47737968e-01 -5.20107210e-01
-6.47190988e-01 2.73021221e-01 -7.19217598e-01 1.99086145e-01
1.16507173e+00 -7.21966028e-02 -2.25232497e-01 -8.15966487e-01
-9.06418443e-01 4.70836550e-01 4.06240821e-01 1.08603872e-01
5.20141661e-01 -1.23963773e+00 -5.88670433e-01 -1.78055912e-01
-8.63555893e-02 -3.19482863e-01 -1.11588486e-01 4.50676113e-01
-7.76848048e-02 2.21631110e-01 -5.84398508e-01 2.28938628e-02
-8.51319432e-01 7.50569284e-01 3.68879795e-01 -1.04252958e+00
-5.22620976e-01 4.30687815e-01 -2.35951230e-01 -3.22779745e-01
3.28186065e-01 -2.04247355e-01 -2.03064680e-01 -9.65724066e-02
4.84622926e-01 4.35739577e-01 -3.11297774e-01 3.09412032e-01
6.85280785e-02 -1.59374386e-01 -2.85088748e-01 -5.79350173e-01
1.60712957e+00 -2.44953018e-03 2.27171242e-01 4.08734530e-01
5.30762196e-01 -3.19335550e-01 -1.75523841e+00 -2.43572578e-01
1.09870387e-02 -3.10901552e-01 -1.28989518e-02 -9.45709288e-01
-5.62158585e-01 4.11249280e-01 5.07598400e-01 1.56417578e-01
9.69025552e-01 -3.34269345e-01 5.31184912e-01 9.18493748e-01
1.01909649e+00 -1.35273170e+00 3.93737048e-01 7.30034709e-01
6.50808811e-01 -9.82621670e-01 2.45965309e-02 5.85025549e-01
-8.72469604e-01 1.10637081e+00 5.54686844e-01 -6.63303196e-01
2.63611157e-03 2.73519844e-01 -2.57609934e-01 7.19178915e-02
-9.15328085e-01 -5.68115711e-01 -5.90935349e-01 7.23506689e-01
2.97761913e-02 1.96510017e-01 -5.02024651e-01 2.01602980e-01
-1.75769962e-02 4.05736417e-01 7.14049995e-01 1.36124539e+00
-8.77830744e-01 -1.51627278e+00 -3.64415020e-01 3.75811368e-01
-4.32282358e-01 -1.77358076e-01 2.12496728e-01 7.59155750e-01
-4.06446666e-01 9.27111030e-01 8.56537372e-02 8.61906707e-02
-3.46220024e-02 -1.00305825e-02 9.75775719e-01 -6.07627630e-01
-7.15286553e-01 2.34911367e-01 2.30037600e-01 -8.14558148e-01
-3.74192506e-01 -7.25019991e-01 -1.63694310e+00 -3.32104892e-01
-7.03407265e-03 4.12019700e-01 3.53308409e-01 8.82350564e-01
6.32686689e-02 5.06322920e-01 6.88044906e-01 -9.29175675e-01
-1.31596053e+00 -7.40391374e-01 -8.99893403e-01 2.57291615e-01
5.97869098e-01 -7.67875433e-01 -8.70034546e-02 -4.60825324e-01]
|
[4.031910419464111, 1.7652043104171753]
|
b6ad1d03-c13a-4263-b9c5-721b3107b9d9
|
distantly-supervised-aspect-clustering-and
| null | null |
https://aclanthology.org/2022.naacl-industry.12
|
https://aclanthology.org/2022.naacl-industry.12.pdf
|
Distantly Supervised Aspect Clustering And Naming For E-Commerce Reviews
|
Product aspect extraction from reviews is a critical task for e-commerce services to understand customer preferences and pain points. While aspect phrases extraction and sentiment analysis have received a lot of attention, clustering of aspect phrases and assigning human readable names to clusters in e-commerce reviews is an extremely important and challenging problem due to the scale of the reviews that makes human review infeasible. In this paper, we propose fully automated methods for clustering aspect words and generating human readable names for the clusters without any manually labeled data. We train transformer based sentence embeddings that are aware of unique e-commerce language characteristics (eg. incomplete sentences, spelling and grammar errors, vernacular etc.). We also train transformer based sequence to sequence models to generate human readable aspect names from clusters. Both the models are trained using heuristic based distant supervision. Additionally, the models are used to improve each other. Extensive empirical testing showed that the clustering model improves the Silhouette Score by 64% when compared to the state-of-the-art baseline and the aspect naming model achieves a high ROUGE-L score of 0.79.
|
['Anirban Majumdar', 'Deepak Gupta', 'Aniket Chakrabarti', 'Prateek Sircar']
| null | null | null | null |
naacl-acl-2022-7
|
['aspect-extraction']
|
['natural-language-processing']
|
[-1.78971179e-02 9.32568312e-02 -3.02654505e-01 -6.33722007e-01
-9.39817786e-01 -8.90503645e-01 3.80362660e-01 2.47747198e-01
-2.58943141e-01 3.91889125e-01 4.04193968e-01 -3.40692937e-01
1.83846936e-01 -7.76494622e-01 -3.91437411e-01 -5.97319186e-01
4.11532074e-01 9.36053097e-01 -1.41916284e-02 -3.94695312e-01
4.38822329e-01 6.54131826e-03 -1.18882847e+00 4.85346466e-01
1.03884423e+00 8.36596131e-01 8.54615048e-02 6.24327242e-01
-6.99039102e-01 4.37601626e-01 -5.30903041e-01 -1.00135148e+00
2.71527469e-01 -5.62069058e-01 -4.96111929e-01 5.46927214e-01
6.53098226e-02 1.79527953e-01 3.59138995e-01 1.22469532e+00
2.75240868e-01 -9.73137692e-02 7.41455078e-01 -1.10660601e+00
-1.20935893e+00 8.93226922e-01 -7.52408743e-01 -1.68209970e-01
2.63738483e-01 -2.36270726e-01 1.45547342e+00 -9.02024984e-01
6.40157819e-01 9.93257642e-01 5.87149799e-01 4.58285630e-01
-7.37089932e-01 -3.95275623e-01 2.66030580e-01 1.10313259e-01
-1.13431442e+00 -3.59630324e-02 6.89907014e-01 -2.64841110e-01
1.07032835e+00 5.89823984e-02 6.52707994e-01 6.22626483e-01
3.62436652e-01 7.76992738e-01 6.80740714e-01 -3.60103846e-01
3.39057088e-01 8.93204987e-01 4.28343236e-01 5.05484760e-01
3.98110002e-01 -6.27250612e-01 -2.44706273e-01 -9.72286053e-03
3.31067801e-01 1.99210584e-01 9.43725854e-02 -3.29994947e-01
-1.06929672e+00 1.18937325e+00 1.71576545e-01 2.57722795e-01
-5.87119460e-01 -2.68477470e-01 5.95456243e-01 1.68734819e-01
5.09777904e-01 7.11533487e-01 -9.41045821e-01 -8.37806016e-02
-5.18590271e-01 -1.37870714e-01 9.98800278e-01 1.69662786e+00
6.52629495e-01 -5.45112975e-02 1.66955471e-01 8.38556826e-01
3.11626017e-01 5.50626397e-01 6.71368659e-01 -3.45021725e-01
5.41530252e-01 1.01060641e+00 -5.02019450e-02 -9.03368831e-01
-5.71898781e-02 -4.47045594e-01 -6.23167396e-01 -2.53301978e-01
-7.53034279e-02 -1.74761012e-01 -9.91319001e-01 1.10201693e+00
3.11909854e-01 -4.52829123e-01 1.99913636e-01 7.32064605e-01
7.15387285e-01 9.07386124e-01 1.02613032e-01 -2.51537889e-01
1.83928239e+00 -1.39764869e+00 -9.98387933e-01 -2.62547016e-01
6.50753558e-01 -1.28903842e+00 1.08682227e+00 4.03639376e-01
-6.30605340e-01 -4.27877814e-01 -1.03696859e+00 -3.40781212e-02
-6.94036126e-01 3.34972560e-01 8.81594300e-01 7.74340093e-01
-6.40424848e-01 1.18960798e-01 -6.78701103e-01 -4.36039805e-01
2.90121317e-01 4.56701636e-01 -2.99581259e-01 -7.17177093e-02
-9.54941511e-01 7.51191974e-01 9.02362764e-02 1.03289150e-01
-4.06508058e-01 -4.65733558e-01 -1.06087983e+00 -8.73306021e-02
2.07632422e-01 -5.21252036e-01 1.30216014e+00 -1.37307751e+00
-1.37008071e+00 8.17253411e-01 -3.63408625e-01 -1.75182462e-01
-4.46180813e-02 -3.90151352e-01 -7.19503701e-01 3.94128561e-02
4.72960383e-01 5.19728124e-01 7.25589037e-01 -1.20995438e+00
-8.81979227e-01 -5.58418989e-01 -3.57487798e-02 3.90031248e-01
-3.71948659e-01 2.40201086e-01 -4.32728618e-01 -5.99540412e-01
1.22344173e-01 -9.83397305e-01 -4.94891435e-01 -4.18782026e-01
-5.25174499e-01 -3.96419227e-01 6.09836400e-01 -7.38175511e-01
1.19929802e+00 -1.97244489e+00 -7.49032944e-02 1.09252430e-01
5.93458898e-02 2.33875126e-01 -2.61226028e-01 4.31785762e-01
8.98805186e-02 3.85984391e-01 -2.37549871e-01 -1.68251306e-01
1.94295987e-01 1.33117996e-02 -2.14639574e-01 2.62037106e-02
3.70634139e-01 1.05916333e+00 -7.28578329e-01 -4.03164148e-01
-1.58126093e-02 5.01863778e-01 -4.59497333e-01 2.36665905e-01
-3.82015824e-01 -9.16713104e-03 -5.87585390e-01 7.31825173e-01
6.12719357e-01 -1.94683835e-01 1.65239006e-01 -2.38903597e-01
1.77963302e-01 4.82363820e-01 -1.10531282e+00 1.36294806e+00
-7.41430998e-01 1.87796101e-01 -2.85136491e-01 -6.41557872e-01
1.09533703e+00 4.82141584e-01 2.44856924e-02 -4.68334883e-01
3.60725909e-01 2.54211813e-01 -1.27118751e-01 -5.48609972e-01
9.36098218e-01 -4.68159825e-01 -4.42491025e-01 6.14194572e-01
3.16667818e-02 -3.33369881e-01 2.37800494e-01 1.42151967e-01
7.47360289e-01 -1.94046851e-02 5.68941653e-01 -1.46773919e-01
6.61495686e-01 3.74042451e-01 8.34651947e-01 1.10659949e-01
-1.21105850e-01 7.96222091e-01 5.34116447e-01 -4.77101058e-01
-1.30628252e+00 -9.45919633e-01 1.16895281e-01 9.48294878e-01
1.47928551e-01 -5.38109004e-01 -7.37041056e-01 -1.00195801e+00
-3.77938151e-01 1.01849043e+00 -6.16333544e-01 -7.96468779e-02
-5.51621199e-01 -6.55101359e-01 -7.81278238e-02 5.33735096e-01
3.36137861e-01 -1.22507453e+00 6.82881027e-02 4.26039159e-01
-8.77446011e-02 -1.17347538e+00 -9.22241926e-01 1.18387468e-01
-7.04831958e-01 -8.79358530e-01 -4.32491362e-01 -1.41278398e+00
8.57547581e-01 2.59221256e-01 1.39640903e+00 -1.74938112e-01
-3.51044536e-02 7.26414844e-02 -9.07706261e-01 -6.60359323e-01
-3.89009655e-01 4.52781916e-01 -5.01895063e-02 5.46206255e-03
1.40028310e+00 -2.33464643e-01 -5.25482655e-01 3.41002345e-01
-9.54309940e-01 -2.53980756e-01 7.61286557e-01 6.27985537e-01
7.76152730e-01 4.66611013e-02 6.88076854e-01 -1.54077983e+00
7.35957742e-01 -3.82246554e-01 -4.52546775e-01 3.11431676e-01
-8.70969951e-01 2.65918281e-02 8.98247242e-01 -3.70010078e-01
-8.97982240e-01 2.43897170e-01 -2.42344320e-01 2.25397050e-01
-7.15604499e-02 3.83219540e-01 -4.33124572e-01 6.04770601e-01
2.31275618e-01 2.58396000e-01 -1.88761488e-01 -3.07577938e-01
6.36561871e-01 1.01728880e+00 -5.67306625e-03 2.85686348e-02
8.57760906e-01 9.80374441e-02 -6.74472749e-01 -8.68025899e-01
-1.12852073e+00 -7.83439577e-01 -7.79978633e-01 3.28934729e-01
1.24651909e+00 -8.77932787e-01 -3.36903304e-01 8.08974802e-02
-1.26867485e+00 5.19252300e-01 -4.09669846e-01 4.33624059e-01
-2.89727151e-01 2.43757367e-01 -7.63165593e-01 -6.05995774e-01
-7.74185240e-01 -1.20841420e+00 1.17017901e+00 2.49025717e-01
-5.36938906e-01 -9.11455750e-01 -3.37476563e-03 5.43768406e-01
4.01280016e-01 -4.10242453e-02 1.07446027e+00 -1.12280703e+00
-6.59895122e-01 -5.01548588e-01 6.31854171e-03 4.42687660e-01
4.61078465e-01 -1.35435596e-01 -6.90789938e-01 1.72723100e-01
3.52055103e-01 7.65228122e-02 5.04904687e-01 2.00983346e-01
3.36477190e-01 -6.55429900e-01 -1.99591577e-01 1.91966429e-01
1.43281949e+00 6.13183260e-01 7.14610040e-01 4.17517245e-01
9.08727169e-01 6.83318675e-01 8.18786204e-01 2.56180614e-01
5.04673541e-01 3.76874536e-01 6.90758694e-03 -1.24383876e-02
1.83961928e-01 -3.92258674e-01 4.42232430e-01 1.71275938e+00
5.87833226e-01 -2.14240253e-01 -3.84876072e-01 8.71993423e-01
-1.53704202e+00 -6.92977548e-01 -3.73443455e-01 1.85947073e+00
9.82584953e-01 2.12570161e-01 -7.25579858e-02 2.45327111e-02
9.27113652e-01 -2.46609468e-02 -3.58434051e-01 -9.03934836e-01
1.99573841e-02 1.78542078e-01 4.27928865e-01 4.95031983e-01
-9.85377908e-01 1.25683761e+00 5.14204168e+00 7.24027634e-01
-6.71584070e-01 2.27564320e-01 6.35090411e-01 1.86649755e-01
-7.92841613e-01 1.07176371e-01 -1.09120214e+00 3.03509057e-01
7.73994207e-01 -1.73758596e-01 1.53103784e-01 1.15425289e+00
1.03758797e-01 2.16822088e-01 -9.62589264e-01 7.53208101e-01
5.69314301e-01 -9.44970131e-01 4.20632303e-01 -7.36923069e-02
1.07093668e+00 -3.43751758e-01 -5.75879067e-02 4.03112441e-01
4.89744455e-01 -9.04414535e-01 4.12017137e-01 1.23030245e-02
4.22896802e-01 -1.09047842e+00 1.29309893e+00 -2.10448746e-02
-1.36567628e+00 2.73931086e-01 -6.56138241e-01 1.68916583e-01
6.61073148e-01 9.34323192e-01 -7.34991252e-01 3.30191761e-01
3.84931177e-01 8.89220476e-01 -4.63326305e-01 7.23708749e-01
-5.47731400e-01 4.85565662e-01 2.18765922e-02 -6.85733020e-01
3.16795051e-01 -7.30250061e-01 2.27634415e-01 1.06912768e+00
3.62394631e-01 1.55681034e-03 2.16661044e-03 6.13157988e-01
-1.04784913e-01 5.32388389e-01 -7.02541113e-01 -4.22673434e-01
2.15046987e-01 1.57894325e+00 -9.30252552e-01 -4.35220718e-01
-4.94548500e-01 1.29243934e+00 -1.29961699e-01 1.99324191e-01
-6.71899199e-01 -8.97000492e-01 8.29264641e-01 5.71025573e-02
9.26166713e-01 -3.99058573e-02 -6.96608841e-01 -1.11369884e+00
2.37251237e-01 -1.00643051e+00 5.96127212e-02 -4.55764979e-01
-1.59316051e+00 1.13342369e+00 -7.97533751e-01 -1.56592250e+00
-1.06210914e-03 -5.28049469e-01 -5.98933816e-01 5.57891965e-01
-1.45908141e+00 -1.13762712e+00 1.07929341e-01 4.10581753e-02
9.51362371e-01 -4.83990461e-01 8.65743637e-01 2.97089994e-01
-2.93044567e-01 7.51427829e-01 2.68987566e-02 2.53328234e-01
6.03673995e-01 -1.30385911e+00 5.80372036e-01 6.39020264e-01
2.62713045e-01 1.05678558e+00 8.12624872e-01 -6.62506163e-01
-1.41622722e+00 -1.21551907e+00 1.66229486e+00 -6.37729704e-01
7.39055216e-01 -5.72326005e-01 -5.23147464e-01 7.19323695e-01
6.20809555e-01 -6.14022136e-01 1.14255631e+00 2.43917644e-01
-4.32616860e-01 -3.73545498e-01 -1.04343200e+00 5.64060450e-01
5.57319105e-01 -3.88717651e-01 -9.12039340e-01 5.31858683e-01
1.17095733e+00 1.10483721e-01 -7.89584756e-01 6.16359944e-03
2.91220665e-01 -7.25923002e-01 3.17035198e-01 -5.52982628e-01
7.03451395e-01 -6.32599473e-01 -1.60446048e-01 -1.28606415e+00
-2.40493074e-01 -3.50955158e-01 4.51127648e-01 1.73488402e+00
1.11299598e+00 -4.47590113e-01 7.95143425e-01 6.17346466e-01
-1.05379976e-01 -6.54987335e-01 -4.74649876e-01 -4.33981240e-01
-1.07240885e-01 -1.86340228e-01 6.34247184e-01 7.64318407e-01
2.77285159e-01 1.24600327e+00 -8.52333680e-02 -4.57900837e-02
3.15185696e-01 4.13422674e-01 5.32780290e-01 -9.79012668e-01
-1.84862867e-01 -1.68089271e-01 -5.25379777e-01 -1.12074971e+00
3.87835577e-02 -8.73845398e-01 2.11294398e-01 -1.81873095e+00
4.39335585e-01 -2.38693982e-01 -3.80498506e-02 1.84885502e-01
-2.59022087e-01 1.49123579e-01 8.12207833e-02 -1.42444402e-01
-9.43995833e-01 7.73633659e-01 1.13207209e+00 -3.41546088e-01
-1.92205831e-01 2.47959718e-01 -1.29409587e+00 5.24624169e-01
9.98662889e-01 -7.39948153e-01 -5.34363270e-01 -3.89040738e-01
6.07766569e-01 -3.97736460e-01 -6.45899713e-01 -5.14470518e-01
1.00087181e-01 1.14864655e-01 9.29784849e-02 -6.64407313e-01
-1.85489766e-02 -1.01705420e+00 -2.70833910e-01 6.57086000e-02
-4.01353687e-01 7.86198199e-01 -1.04341678e-01 5.65071523e-01
-5.33326447e-01 -4.61255252e-01 3.39574337e-01 -4.97077763e-01
-4.00101274e-01 2.35157967e-01 -5.81431985e-01 2.53128827e-01
1.07273030e+00 -1.86681375e-01 -7.90662542e-02 -4.61384654e-01
-2.83233702e-01 1.48368329e-01 4.90996361e-01 7.90924430e-01
6.03529632e-01 -1.37296832e+00 -5.62587440e-01 1.34095401e-01
3.88443559e-01 -1.64551921e-02 -1.10292129e-01 3.82204860e-01
-5.73153257e-01 5.09998202e-01 1.51785254e-01 -3.10003847e-01
-1.21792376e+00 7.94646740e-01 -1.62568957e-01 -3.37711394e-01
-2.45809317e-01 6.82539463e-01 9.83477831e-02 -9.57999885e-01
7.57263601e-02 -3.82447183e-01 -5.31847179e-01 1.22414976e-01
5.90540290e-01 -3.93957831e-03 2.27065280e-01 -7.32359886e-01
-2.93174833e-01 9.50406075e-01 -5.80104828e-01 -3.85987200e-02
1.47990596e+00 -2.50125021e-01 -4.23046410e-01 5.02283871e-01
1.48719609e+00 1.06719971e-01 -5.44891179e-01 -4.53692526e-02
3.41106445e-01 -7.93904141e-02 -3.81791443e-01 -7.27601588e-01
-1.14555967e+00 7.94598818e-01 3.33678603e-01 1.25026360e-01
8.79169881e-01 -2.05254685e-02 1.25629413e+00 3.40033054e-01
2.53142864e-01 -1.34975243e+00 1.25256469e-02 6.55449033e-01
4.74526614e-01 -1.30696881e+00 -1.10929534e-01 -6.75366938e-01
-1.30395782e+00 8.51287961e-01 6.02455556e-01 -2.14595914e-01
7.44771540e-01 1.71901613e-01 6.64722443e-01 -2.52608538e-01
-6.82796001e-01 -1.33580253e-01 1.54542997e-01 6.46345556e-01
7.08069146e-01 9.53614190e-02 -7.29215086e-01 9.14620697e-01
-4.98181522e-01 -2.02993602e-01 7.44423211e-01 8.16098928e-01
-5.43134928e-01 -1.33647037e+00 1.47695497e-01 6.37554765e-01
-6.85896754e-01 -5.33583224e-01 -4.31798786e-01 3.86045188e-01
-2.75479201e-02 1.19411254e+00 -6.82046860e-02 -4.16080654e-01
3.81510228e-01 1.17712736e-01 -1.29982665e-01 -1.04406667e+00
-8.51748228e-01 1.43477827e-01 3.01263958e-01 -5.67875765e-02
-3.27936053e-01 -5.30875862e-01 -1.36136293e+00 -1.32696137e-01
-5.26867151e-01 6.49526715e-01 8.46394300e-01 8.87585223e-01
4.67732698e-01 3.80715400e-01 8.32498074e-01 -2.05462486e-01
-5.00531375e-01 -1.04772723e+00 -7.60462821e-01 6.22661829e-01
7.71340961e-03 1.36474641e-02 -3.99218619e-01 2.86841542e-01]
|
[11.350112915039062, 6.6887006759643555]
|
fe0cc738-fed8-43cc-8a4f-8ebfbbe817a7
|
cognito-automated-feature-engineering-for
| null | null |
https://ieeexplore.ieee.org/abstract/document/7836821/
|
https://ieeexplore.ieee.org/abstract/document/7836821/
|
Cognito: Automated Feature Engineering for Supervised Learning
|
Feature engineering involves constructing novel features from given data with the goal of improving predictive learning performance. Feature engineering is predominantly a human-intensive and time consuming step that is central to the data science workflow. In this paper, we present a novel system called "Cognito", that performs automatic feature engineering on a given dataset for supervised learning. The system explores various feature construction choices in a hierarchical and non-exhaustive manner, while progressively maximizing the accuracy of the model through a greedy exploration strategy. Additionally, the system allows users to specify domain or data specific choices to prioritize the exploration. Cognito is capable of handling large datasets through sampling and built-in parallelism, and integrates well with a state-of-the-art model selection strategy. We present the design and operation of Cognito, along with experimental results on eight real datasets to demonstrate its efficacy.
|
['Horst Samulowitz', 'Deepak Turaga', 'Udayan Khurana', 'Srinivasan Parthasrathy']
|
2016-01-01
| null | null | null |
icdmw-2016-2016-1
|
['automated-feature-engineering']
|
['methodology']
|
[ 2.70260125e-01 -2.89323539e-01 -9.00160968e-02 -6.07423723e-01
-5.36461473e-01 -3.80650073e-01 4.84705448e-01 3.70361596e-01
-2.04899848e-01 6.25753224e-01 -3.12349677e-01 -2.58134842e-01
-5.43975770e-01 -8.85897756e-01 -3.17747653e-01 -4.82604861e-01
-2.41228968e-01 9.62227702e-01 8.75785276e-02 1.18768491e-01
6.04440212e-01 8.60588014e-01 -2.14003992e+00 3.46630782e-01
7.75893033e-01 9.83189046e-01 2.59002417e-01 4.75602925e-01
-3.52458000e-01 3.28952283e-01 -3.08310062e-01 1.05952755e-01
4.45154965e-01 -1.62521303e-01 -9.41454232e-01 9.10832956e-02
9.60616171e-02 3.20854574e-01 4.58625764e-01 5.63816488e-01
2.90750831e-01 1.87596250e-02 2.18375832e-01 -1.29130650e+00
1.88414022e-01 4.63628203e-01 -1.17834531e-01 -3.79091129e-02
3.12726110e-01 3.41650426e-01 1.00112152e+00 -1.21626937e+00
6.47577465e-01 8.39527130e-01 5.33377826e-01 1.91350862e-01
-1.47740126e+00 -6.61289036e-01 3.66667658e-02 2.46673673e-01
-1.74453819e+00 -5.09125710e-01 6.97418630e-01 -6.36178672e-01
1.25196755e+00 7.05445051e-01 1.00192988e+00 1.57726556e-01
1.44551575e-01 7.51910508e-01 8.90499949e-01 -6.22565150e-01
5.77798307e-01 2.97500700e-01 3.71875644e-01 8.07905674e-01
2.38821611e-01 1.75151527e-01 -9.43950653e-01 -6.01477563e-01
3.26244533e-01 7.86507651e-02 8.86816829e-02 -8.03176165e-01
-1.00870144e+00 8.11305821e-01 1.33766547e-01 4.72523570e-02
-5.37393808e-01 -2.85933822e-01 3.19013417e-01 3.67927819e-01
2.60569662e-01 1.08294249e+00 -9.14665878e-01 -3.27119738e-01
-1.10240746e+00 3.97202164e-01 9.22906756e-01 1.04822409e+00
1.17436230e+00 -3.55866283e-01 -3.64536829e-02 6.42194211e-01
1.48582563e-01 -1.34467587e-01 4.14719939e-01 -4.35754329e-01
-2.32580334e-01 1.29325950e+00 -9.13463458e-02 -5.23966551e-01
-7.52955198e-01 -5.82053065e-01 -2.95110792e-01 2.74069250e-01
1.78875849e-02 1.81264170e-02 -9.30182636e-01 1.25940943e+00
5.78530014e-01 -3.43528420e-01 -1.50063097e-01 5.67872345e-01
6.88921928e-01 2.21346721e-01 2.52393812e-01 -4.64134246e-01
1.23254204e+00 -6.57527506e-01 -2.23573372e-01 5.76614076e-03
7.12525189e-01 -5.53455889e-01 1.13857067e+00 8.80472898e-01
-8.42662036e-01 -4.75219637e-01 -8.32349837e-01 7.06857815e-02
-5.19487321e-01 1.88141093e-01 1.07393348e+00 5.44806600e-01
-5.84814370e-01 6.98770404e-01 -9.00713980e-01 -4.01134104e-01
4.70126301e-01 6.38785660e-01 -2.98392117e-01 4.37115245e-02
-7.70978630e-01 7.47057080e-01 7.78186679e-01 -2.72756070e-01
-6.73865974e-01 -1.09155977e+00 -6.72552288e-01 2.53359705e-01
7.26167798e-01 -8.00674200e-01 1.13166964e+00 -7.34618008e-01
-1.23328626e+00 4.63590413e-01 -2.66942114e-01 -2.94048190e-01
3.70032191e-01 -3.80033106e-01 -3.99186820e-01 -2.21066996e-01
-1.94194466e-01 5.57896316e-01 5.87118447e-01 -8.93823087e-01
-8.71004939e-01 -3.62713695e-01 -3.63538623e-01 1.43524706e-01
-2.96623051e-01 5.58714792e-02 -4.99879509e-01 -1.51468694e-01
1.98126882e-01 -9.27045882e-01 -5.69538951e-01 -2.10328978e-02
-3.65030080e-01 -4.13701832e-01 7.85708010e-01 -4.07217704e-02
1.70024097e+00 -1.89079964e+00 -1.58175737e-01 8.14936101e-01
1.67938530e-01 1.41686469e-01 1.96649581e-01 6.51299596e-01
1.56100770e-03 1.41662776e-01 -2.07699731e-01 1.90797642e-01
-2.46586412e-01 3.28198969e-02 1.53517038e-01 1.54107481e-01
4.02475446e-01 6.41748488e-01 -7.43543327e-01 -5.54675281e-01
2.30221987e-01 1.24419793e-01 -6.72841370e-01 3.38614374e-01
-4.47184533e-01 1.93634570e-01 -6.00939810e-01 9.15115535e-01
5.89870036e-01 -4.60013777e-01 4.01903242e-01 -1.00182667e-01
-5.26802838e-01 3.01946849e-01 -1.32952893e+00 1.42460275e+00
-2.73806572e-01 2.33542010e-01 -4.36613917e-01 -6.51759684e-01
1.27653623e+00 -2.22409859e-01 6.34719312e-01 -2.94150651e-01
-7.84015134e-02 2.59475887e-01 1.01126917e-01 -4.83847290e-01
5.46940625e-01 1.02317967e-01 3.42179760e-02 5.16532004e-01
3.92755903e-02 -7.71088600e-02 5.24230063e-01 -1.04991414e-01
1.22037554e+00 1.44232467e-01 1.00348616e+00 -5.69303274e-01
4.84924316e-01 4.87392753e-01 6.19732320e-01 7.31464326e-01
3.37833226e-01 2.82136768e-01 3.22704703e-01 -8.93219352e-01
-8.48446429e-01 -6.86792910e-01 -3.99362564e-01 1.23797810e+00
-3.90166968e-01 -7.99810171e-01 -3.47603977e-01 -7.93633103e-01
4.86813754e-01 8.95452797e-01 -7.03809977e-01 -6.53196424e-02
-2.40611345e-01 -6.08047485e-01 -5.49940467e-02 2.93968856e-01
2.42199361e-01 -1.26046836e+00 -1.30828726e+00 3.59126151e-01
4.62856144e-01 -3.00790012e-01 -1.30430236e-01 7.05296397e-01
-9.18158233e-01 -1.25896859e+00 2.92157501e-01 -3.92227173e-01
8.25416148e-01 1.11907296e-01 1.32660091e+00 2.07469493e-01
-8.44241619e-01 -8.54258388e-02 -5.51931679e-01 -6.16032898e-01
-1.12019733e-01 6.04707301e-01 -1.31355226e-01 -2.22848788e-01
8.38762522e-01 -4.07142848e-01 -3.42403620e-01 3.41672629e-01
-7.69541681e-01 3.93604606e-01 8.11680615e-01 1.09890532e+00
9.00059402e-01 1.14997156e-01 4.66738403e-01 -1.21575534e+00
5.67255259e-01 -4.99898970e-01 -8.20238650e-01 5.34299433e-01
-1.28001559e+00 2.77755350e-01 6.16159022e-01 -2.50968814e-01
-7.80437946e-01 8.47042739e-01 1.08345464e-01 -3.04821104e-01
-2.48619184e-01 9.54281867e-01 -3.87117684e-01 -2.22051308e-01
8.13430190e-01 4.31744009e-01 1.40552506e-01 -9.03743148e-01
2.08963230e-01 6.09004617e-01 2.64256950e-02 -5.93300104e-01
4.59229976e-01 2.53693853e-02 2.12988138e-01 -7.11112261e-01
-3.88612509e-01 -3.82347107e-01 -9.62464988e-01 -4.26196679e-02
-3.69767286e-02 -3.75074714e-01 -7.75049448e-01 9.28778872e-02
-4.63859022e-01 -5.20146936e-02 -4.53377515e-01 2.11438969e-01
-3.85140300e-01 -2.40312874e-01 2.15822577e-01 -6.23316526e-01
-6.39859080e-01 -1.13625824e+00 5.96870184e-01 3.39117855e-01
-5.70740342e-01 -6.18888557e-01 2.45789383e-02 -1.34718478e-01
4.36067045e-01 2.65292168e-01 9.89174724e-01 -1.15948308e+00
-5.74528694e-01 -2.83871144e-01 4.34993170e-02 -1.89972341e-01
2.73383975e-01 3.44460130e-01 -8.02420199e-01 -3.33330363e-01
-6.55777574e-01 -1.59418568e-01 5.57826102e-01 4.70133647e-02
1.47765791e+00 -9.83433500e-02 -6.98199868e-01 6.66834414e-01
1.32223630e+00 3.00896138e-01 2.33293563e-01 5.48025966e-01
2.69113034e-01 5.78755379e-01 1.18750966e+00 1.00147724e+00
1.03037573e-01 5.42402565e-01 1.93279162e-01 4.65279408e-02
3.78186315e-01 -9.43942517e-02 -2.83391774e-01 9.24064666e-02
1.38641417e-01 1.46553501e-01 -1.26730227e+00 4.80521262e-01
-1.75512552e+00 -6.02357030e-01 -8.12653750e-02 2.15549588e+00
9.46483612e-01 1.35625511e-01 2.29314744e-01 2.41542131e-01
2.93118149e-01 -5.93714654e-01 -8.66619587e-01 -5.09864569e-01
2.91634232e-01 3.98018152e-01 2.95363247e-01 1.06251821e-01
-9.36385512e-01 1.00419116e+00 6.79639673e+00 6.15629256e-01
-1.18223119e+00 -4.70534831e-01 3.73486042e-01 -4.77078646e-01
-2.06809223e-01 1.48270994e-01 -8.73871624e-01 1.08324230e-01
8.64271224e-01 -7.56921589e-01 3.70491952e-01 1.29145575e+00
8.09453130e-02 -3.60758007e-01 -1.36217570e+00 7.16692507e-01
-3.49281758e-01 -1.60544968e+00 1.11715250e-01 4.61929180e-02
2.11874530e-01 -9.69574004e-02 -2.63924628e-01 1.82438895e-01
2.63829798e-01 -1.00813150e+00 4.66153175e-01 6.38153613e-01
8.53694916e-01 -1.20365620e+00 3.97542268e-01 4.03854638e-01
-1.20235729e+00 -2.15313822e-01 -5.82185537e-02 -1.31166771e-01
-2.29636490e-01 7.28142440e-01 -1.36189067e+00 5.64024270e-01
8.13165307e-01 5.93586624e-01 -1.10706639e+00 1.27812397e+00
1.83143973e-01 5.00019312e-01 -4.90075171e-01 -2.52144039e-01
-1.58008039e-01 1.47217110e-01 3.52407485e-01 1.36829269e+00
2.97034472e-01 -9.23667476e-03 4.07466382e-01 6.96408451e-01
3.44679981e-01 3.58231306e-01 -3.78534585e-01 -2.26864964e-02
8.32868099e-01 1.46269894e+00 -5.02692103e-01 -3.08161348e-01
-2.03479230e-01 2.55259663e-01 4.96414155e-01 -1.95678294e-01
-2.29434729e-01 -5.85927248e-01 6.89973831e-01 2.35047087e-01
2.90802717e-01 -6.48378506e-02 -7.02238083e-01 -7.46810496e-01
-3.38789187e-02 -1.05285573e+00 6.73268497e-01 -2.99635023e-01
-1.13236797e+00 7.81084180e-01 1.68580398e-01 -1.38260782e+00
-3.31534028e-01 -3.80788118e-01 -3.96118969e-01 9.68626320e-01
-1.24893200e+00 -9.21774983e-01 -5.80673456e-01 4.48179781e-01
3.71088952e-01 -3.30620021e-01 1.05380058e+00 -7.13908002e-02
-5.77332795e-01 5.80861807e-01 -9.10351146e-03 -5.61582506e-01
5.09855330e-01 -1.04117036e+00 5.85509777e-01 4.72606450e-01
-1.44572929e-01 1.07428062e+00 7.90403008e-01 -8.19886565e-01
-1.61596823e+00 -1.01897490e+00 9.41928208e-01 -7.68383127e-03
4.21517611e-01 -2.99740940e-01 -1.13308764e+00 3.38104516e-01
-2.56155521e-01 7.77631104e-02 1.16059756e+00 5.37170947e-01
-3.19597237e-02 -1.87857836e-01 -1.47420597e+00 2.93095976e-01
9.94714260e-01 3.37611958e-02 -3.94427955e-01 5.98349310e-02
2.14274153e-01 -2.87046522e-01 -9.33580339e-01 5.38030326e-01
8.44439447e-01 -7.27502286e-01 5.93917489e-01 -9.03230429e-01
5.37984669e-02 -4.89890873e-01 5.28131984e-02 -1.21565747e+00
-6.49614692e-01 -7.46504962e-01 -1.62262619e-01 1.05451882e+00
7.30088413e-01 -4.36400652e-01 9.59791601e-01 1.09861028e+00
1.89865716e-02 -1.02841640e+00 -6.75793946e-01 -6.15958393e-01
-3.67010593e-01 -3.30153108e-01 1.17544079e+00 8.71671140e-01
3.01317483e-01 2.54247904e-01 -2.02462763e-01 -6.78338408e-02
5.27700067e-01 4.31598246e-01 1.10232902e+00 -1.74497497e+00
-1.61270753e-01 -3.37487459e-01 -3.59529197e-01 -4.67264116e-01
-4.00424331e-01 -8.80139410e-01 7.06640184e-02 -1.15500891e+00
2.52649784e-01 -7.56823659e-01 -3.94658476e-01 9.22727227e-01
-2.44010136e-01 -2.51006693e-01 -1.27270976e-02 2.99800038e-01
-5.35295844e-01 3.12904149e-01 6.86954677e-01 3.30901235e-01
-7.06274867e-01 9.34163257e-02 -7.64873326e-01 4.40893352e-01
8.31315696e-01 -6.46420538e-01 -3.55977744e-01 1.58158317e-01
1.77435189e-01 -2.49260351e-01 -5.32825999e-02 -9.10259962e-01
3.00992012e-01 -6.37610257e-01 7.17967987e-01 -9.39224184e-01
-3.88395451e-02 -1.11862099e+00 6.59308255e-01 6.16138697e-01
-4.83262330e-01 2.20319986e-01 2.82336235e-01 1.68107539e-01
-4.90797423e-02 -2.32821405e-01 8.62101734e-01 -1.75459698e-01
-9.45606053e-01 2.74286926e-01 -2.13336349e-01 -3.94878596e-01
1.33624697e+00 -1.21096261e-01 -4.82715331e-02 3.58159572e-01
-7.66298175e-01 5.16291440e-01 5.50953627e-01 3.28660190e-01
6.71084464e-01 -1.00171077e+00 -6.08671665e-01 8.53019595e-01
6.42391860e-01 5.30770700e-03 -1.58342093e-01 4.27431494e-01
-4.90087926e-01 3.49570423e-01 -2.97153771e-01 -6.94180191e-01
-1.66242552e+00 5.67829013e-01 1.90870777e-01 -4.46850300e-01
-5.56537449e-01 6.75833046e-01 -4.13515657e-01 -4.81449813e-01
-9.27182883e-02 1.25393514e-02 -2.42641345e-01 1.09259658e-01
6.65745199e-01 4.54348087e-01 4.84051168e-01 7.03737512e-03
-6.47233725e-01 1.73740298e-01 -3.92979801e-01 1.79824457e-01
1.50725150e+00 1.66624144e-01 -6.97786212e-02 2.26981923e-01
7.38795757e-01 -4.57015246e-01 -1.11757088e+00 -3.73879492e-01
5.73034406e-01 -8.17818522e-01 1.07353561e-01 -1.21965563e+00
-7.10124314e-01 3.69105965e-01 5.08583903e-01 2.11307064e-01
1.33325994e+00 -6.99416548e-02 1.51131600e-01 5.86320400e-01
4.19700831e-01 -1.20326757e+00 -3.05767357e-01 4.09371704e-01
8.78975511e-01 -1.12641788e+00 5.03437877e-01 -4.23926502e-01
-6.08985186e-01 1.35332048e+00 8.51524055e-01 1.57029420e-01
8.58365774e-01 4.06544626e-01 -1.31657049e-01 -2.83454150e-01
-1.35516143e+00 -3.03036124e-01 4.79693830e-01 3.16103429e-01
4.47856039e-01 1.09829932e-01 -5.29091954e-01 5.68276227e-01
-3.22269291e-01 3.78347278e-01 1.43595085e-01 1.43958747e+00
-6.89493775e-01 -1.39059329e+00 -2.32733175e-01 1.01383376e+00
-3.40409577e-02 -6.68100491e-02 -5.64102292e-01 8.81510854e-01
2.38382682e-01 7.11866498e-01 -2.34956332e-02 -6.55296028e-01
4.87712562e-01 4.44274157e-01 1.57285675e-01 -8.19726765e-01
-9.82154369e-01 1.23484746e-01 2.90700525e-01 -5.62067389e-01
1.35195777e-01 -1.09684670e+00 -1.27378416e+00 -1.43616512e-01
-3.81274849e-01 4.86973882e-01 8.39511752e-01 8.22673202e-01
8.83169353e-01 5.06073952e-01 8.56496632e-01 -6.13927245e-01
-3.99892479e-01 -9.00767148e-01 -4.01175439e-01 1.33847699e-01
5.86316809e-02 -7.91231394e-01 2.32058018e-02 -6.08372316e-02]
|
[8.30702018737793, 4.603633880615234]
|
37841b7c-2e3a-4c0b-919e-d5f49bc206a5
|
targeted-extraction-of-temporal-facts-from
|
2203.11054
| null |
https://arxiv.org/abs/2203.11054v1
|
https://arxiv.org/pdf/2203.11054v1.pdf
|
Targeted Extraction of Temporal Facts from Textual Resources for Improved Temporal Question Answering over Knowledge Bases
|
Knowledge Base Question Answering (KBQA) systems have the goal of answering complex natural language questions by reasoning over relevant facts retrieved from Knowledge Bases (KB). One of the major challenges faced by these systems is their inability to retrieve all relevant facts due to factors such as incomplete KB and entity/relation linking errors. In this paper, we address this particular challenge for systems handling a specific category of questions called temporal questions, where answer derivation involve reasoning over facts asserting point/intervals of time for various events. We propose a novel approach where a targeted temporal fact extraction technique is used to assist KBQA whenever it fails to retrieve temporal facts from the KB. We use $\lambda$-expressions of the questions to logically represent the component facts and the reasoning steps needed to derive the answer. This allows us to spot those facts that failed to get retrieved from the KB and generate textual queries to extract them from the textual resources in an open-domain question answering fashion. We evaluated our approach on a benchmark temporal question answering dataset considering Wikidata and Wikipedia respectively as the KB and textual resource. Experimental results show a significant $\sim$30\% relative improvement in answer accuracy, demonstrating the effectiveness of our approach.
|
['L Venkata Subramaniam', 'Hima Karanam', 'Shajith Ikbal', 'Dinesh Khandelwal', 'Sumit Neelam', 'Udit Sharma', 'Nithish Kannen']
|
2022-03-21
| null | null | null | null |
['knowledge-base-question-answering']
|
['natural-language-processing']
|
[-2.35687926e-01 4.87137347e-01 1.04337400e-02 -2.30000794e-01
-1.15034306e+00 -8.51814747e-01 5.63199043e-01 6.17688894e-01
-3.84054393e-01 1.41385448e+00 1.27706632e-01 -5.11592686e-01
-6.40259385e-01 -1.26133955e+00 -6.48190439e-01 -6.74134959e-03
-1.31658152e-01 6.85580790e-01 1.21739209e+00 -8.28283608e-01
2.07701102e-01 1.83196455e-01 -1.68511808e+00 6.08584106e-01
9.43003714e-01 1.21027637e+00 -1.30491838e-01 7.27029622e-01
-7.70362496e-01 1.67278588e+00 -7.21492231e-01 -6.47139907e-01
-1.42557681e-01 -3.38413119e-01 -1.60048413e+00 -5.64113319e-01
2.24442214e-01 -2.39173517e-01 -1.08568296e-01 6.30049407e-01
1.60117716e-01 2.80314535e-01 5.54838598e-01 -1.26616633e+00
-2.47134611e-01 4.65469301e-01 1.29556777e-02 9.69575465e-01
1.10381424e+00 -3.29450309e-01 1.11400545e+00 -4.22074378e-01
8.82396758e-01 1.10144949e+00 4.47103560e-01 2.78539240e-01
-5.43968976e-01 -1.14149153e-01 2.03511968e-01 9.94682908e-01
-1.29723656e+00 -2.85058171e-01 5.69128752e-01 -1.74832046e-01
1.60493767e+00 5.75519383e-01 3.69434863e-01 4.55851316e-01
2.12585956e-01 4.20712441e-01 7.22913384e-01 -5.76237559e-01
3.49566102e-01 3.62516165e-01 4.96098071e-01 7.50731051e-01
-3.46149169e-02 -3.07340264e-01 -6.96447909e-01 -4.20324862e-01
2.48217121e-01 -6.55830979e-01 -1.08532496e-01 2.98620939e-01
-8.94049108e-01 7.14070201e-01 1.19329747e-02 3.82551283e-01
-6.16015375e-01 6.32473826e-02 5.64011216e-01 4.60045516e-01
1.43470079e-01 3.04080218e-01 -8.70679617e-01 -1.05059326e-01
-4.40020204e-01 7.54772127e-01 1.27095401e+00 1.12284851e+00
5.55821180e-01 -4.65138823e-01 -4.13153797e-01 4.75432217e-01
4.52624857e-02 3.07141155e-01 3.85226488e-01 -7.86915123e-01
7.32067943e-01 1.04318118e+00 6.13559365e-01 -1.25850070e+00
-3.43267560e-01 1.28105149e-01 7.14326128e-02 -6.07517183e-01
4.60048378e-01 -1.03827946e-01 -6.00329936e-01 1.53696525e+00
8.08648467e-01 9.66328979e-02 5.16806781e-01 5.06700218e-01
1.10827732e+00 9.02035594e-01 1.89207286e-01 -5.69416463e-01
1.94355893e+00 -4.71191853e-01 -1.05478251e+00 -8.86586159e-02
5.53134084e-01 -5.70626318e-01 7.89119363e-01 2.22581029e-01
-9.66467261e-01 -9.79180261e-02 -8.85899901e-01 -3.38774711e-01
-8.51173401e-01 -2.41644979e-01 4.56350207e-01 1.00393109e-01
-6.70474946e-01 -2.81293206e-02 -4.87236083e-01 -4.51586515e-01
-9.21059176e-02 2.96414673e-01 -1.05189122e-01 3.76996212e-02
-1.86825836e+00 1.18258822e+00 8.32183599e-01 2.31038667e-02
-5.65035999e-01 -7.70526886e-01 -6.58697546e-01 -4.58982885e-02
1.08016944e+00 -7.70537257e-01 1.49335515e+00 -4.45615798e-02
-9.77681518e-01 6.06174529e-01 -3.80983144e-01 -9.12442029e-01
9.10491198e-02 -3.12223166e-01 -9.76606429e-01 6.10573351e-01
4.58323777e-01 1.22801311e-01 2.57265568e-01 -8.22439373e-01
-1.13997662e+00 -4.28292572e-01 6.94944978e-01 9.83898491e-02
4.77682464e-02 6.74342364e-02 -5.41488290e-01 -3.77776809e-02
2.63752434e-02 -5.34629941e-01 6.23975135e-02 -6.48330927e-01
-2.03352094e-01 -7.06102550e-01 9.57241118e-01 -1.09801877e+00
1.82678199e+00 -1.43610644e+00 -1.65657729e-01 9.42372829e-02
-2.25839496e-01 8.81515741e-02 4.00446653e-01 8.58908832e-01
1.86495855e-01 -7.00345554e-04 -1.12231344e-01 6.85050547e-01
-6.71306700e-02 6.29605830e-01 -9.84320045e-01 -2.88570791e-01
3.75840932e-01 9.55095887e-01 -9.20408010e-01 -1.05389655e+00
-1.63147807e-01 -1.02660328e-01 -3.00577521e-01 1.40897751e-01
-1.14558077e+00 8.31582919e-02 -8.79875660e-01 6.59096301e-01
1.04256243e-01 -7.02290377e-03 3.50065380e-02 -4.05842990e-01
6.79723322e-02 5.82795203e-01 -1.35435987e+00 1.38365209e+00
-3.97874355e-01 1.98249757e-01 -5.14395118e-01 -9.42684293e-01
7.06304312e-01 8.06044698e-01 1.64611310e-01 -1.00467706e+00
-1.14641406e-01 3.58498186e-01 -4.06984806e-01 -1.32600105e+00
5.83337069e-01 -4.06840891e-01 -3.40020239e-01 2.01820120e-01
1.21003911e-01 -4.56434160e-01 7.49235034e-01 5.18271446e-01
1.30162251e+00 2.40930930e-01 6.75071001e-01 -1.77496806e-01
1.09889650e+00 8.02734256e-01 4.12656903e-01 5.73130012e-01
-1.82616320e-02 -3.49087641e-02 7.53019989e-01 -6.63928032e-01
-6.61410034e-01 -8.20316136e-01 2.04761744e-01 8.40055048e-01
9.14541446e-03 -4.86396313e-01 -2.28654280e-01 -9.95852947e-01
-1.95056602e-01 1.05758953e+00 -6.54003322e-01 4.51763645e-02
-8.20904613e-01 -3.67564231e-01 6.99589729e-01 3.23067129e-01
4.71854299e-01 -1.18383873e+00 -1.00558257e+00 5.26349664e-01
-8.98715675e-01 -1.23178875e+00 1.96188539e-01 1.27498144e-02
-7.30555773e-01 -1.59477293e+00 -6.48922622e-02 -5.07965863e-01
2.04184622e-01 -2.74158716e-01 1.39025366e+00 5.25549017e-02
-2.41227865e-01 5.59499025e-01 -4.98014480e-01 -7.27225840e-01
-2.24956185e-01 -1.70892593e-03 -3.01584780e-01 -2.99117476e-01
5.63807905e-01 -4.76075351e-01 -4.86245662e-01 2.56985962e-01
-1.29715443e+00 -5.29209554e-01 3.16699184e-02 4.23743457e-01
5.40131867e-01 4.67090964e-01 1.17704046e+00 -8.83303761e-01
1.06406260e+00 -9.65167999e-01 -7.40766585e-01 9.06363428e-01
-3.82255554e-01 4.47994262e-01 6.08799279e-01 -1.52998507e-01
-1.45445549e+00 -3.97205532e-01 -5.98707460e-02 2.97703594e-01
4.56553176e-02 1.15260899e+00 1.14981689e-01 5.83661079e-01
9.14760888e-01 3.57380778e-01 -5.28398097e-01 -1.01105675e-01
3.49044710e-01 2.32822165e-01 6.35186613e-01 -1.06420970e+00
5.83900273e-01 4.59403723e-01 1.07471168e-01 -4.23509985e-01
-1.15701485e+00 -7.61266589e-01 -2.63153583e-01 -2.41590872e-01
7.79709637e-01 -6.55791700e-01 -9.13482606e-01 -3.36418957e-01
-1.18263495e+00 1.48259476e-01 -4.17138338e-01 1.17161408e-01
-6.56585336e-01 3.26705843e-01 -2.53795981e-01 -1.12009990e+00
-5.83687186e-01 -3.24059606e-01 8.82014811e-01 2.61963457e-01
-3.20768923e-01 -9.70189750e-01 3.15429568e-01 5.41193426e-01
1.82271555e-01 4.54089701e-01 1.39911699e+00 -8.79020691e-01
-5.97241640e-01 -2.96473652e-01 -2.11355053e-02 -1.39553428e-01
7.06520230e-02 -2.01058596e-01 -5.76355338e-01 4.67617869e-01
3.47456196e-03 -3.53635132e-01 3.02672714e-01 -1.67152032e-01
4.13930982e-01 -7.40529478e-01 -2.50553727e-01 -4.68790144e-01
1.63885772e+00 5.59350908e-01 7.68062592e-01 5.51911533e-01
-8.03070366e-02 8.16132367e-01 1.07164407e+00 3.47639501e-01
7.49463379e-01 6.62999392e-01 1.77853853e-01 7.99055874e-01
1.62678853e-01 -2.25484043e-01 -8.95633176e-02 4.53050017e-01
-1.27310112e-01 -1.74177021e-01 -1.18961167e+00 1.21403778e+00
-2.16596270e+00 -1.17452037e+00 -3.85407656e-01 1.85158730e+00
1.37693226e+00 1.53946951e-01 3.43263485e-02 2.54724324e-01
1.68053851e-01 -2.67639190e-01 -4.36626405e-01 -3.09043020e-01
-3.76942679e-02 4.07638907e-01 7.76587576e-02 6.45540476e-01
-7.05802560e-01 9.49547589e-01 5.16287899e+00 8.11439276e-01
-5.07527113e-01 -1.90929584e-02 2.64738463e-02 1.26414254e-01
-2.09147409e-01 2.46896610e-01 -8.56947482e-01 8.66784751e-02
1.18291163e+00 -6.53947353e-01 6.30390942e-02 4.89609838e-01
3.56893167e-02 -6.64744139e-01 -9.84429300e-01 4.69070584e-01
1.50378309e-02 -1.40395892e+00 2.28338480e-01 -5.23624659e-01
4.53223646e-01 -6.17755949e-01 -4.28928882e-01 5.74841321e-01
1.84198886e-01 -6.80320323e-01 7.69614220e-01 1.05400801e+00
2.26258457e-01 -7.84993231e-01 9.38154161e-01 5.79240978e-01
-1.38418400e+00 -2.47661561e-01 -1.63595229e-01 -4.71370667e-02
3.94308329e-01 5.14384270e-01 -1.34590673e+00 1.22578526e+00
9.43294108e-01 -1.13664106e-01 -4.86092120e-01 7.19591975e-01
-2.60513693e-01 4.27125573e-01 -6.46938860e-01 -2.65608132e-01
1.91176236e-01 1.61508128e-01 4.78795648e-01 9.99468923e-01
3.86749446e-01 8.81972909e-01 -4.35110688e-01 5.91072857e-01
2.06917718e-01 3.30428034e-01 -5.74907899e-01 1.19918995e-01
4.19144958e-01 8.22666883e-01 -5.41236103e-01 -6.21275008e-01
-3.05544615e-01 5.98575473e-01 2.98255235e-01 2.84497082e-01
-9.30681109e-01 -5.74251652e-01 -2.82560233e-02 1.12063348e-01
4.31701064e-01 -8.82175639e-02 3.92152727e-01 -9.70187545e-01
6.04236245e-01 -8.15920711e-01 1.27709436e+00 -1.05042839e+00
-1.07930481e+00 7.90937126e-01 5.91299236e-01 -7.55163789e-01
-6.81214690e-01 -2.42052644e-01 -2.34400019e-01 7.51033783e-01
-1.72589135e+00 -9.19367373e-01 -1.75875306e-01 1.01219118e+00
4.11221117e-01 3.55432302e-01 7.54857540e-01 2.70734608e-01
-6.21640757e-02 -1.74242571e-01 -7.49710083e-01 -1.92916870e-01
4.90527481e-01 -1.24587786e+00 -3.09354931e-01 8.37698936e-01
-1.45293046e-02 7.93479681e-01 9.95335162e-01 -8.06005776e-01
-1.46373558e+00 -9.37806666e-01 1.55840409e+00 -7.96967745e-01
8.59138906e-01 1.89119995e-01 -9.96645331e-01 8.23179722e-01
2.46161699e-01 -1.50030822e-01 6.11844063e-01 -9.95572358e-02
-4.14812982e-01 -2.66487032e-01 -1.24613929e+00 5.45968354e-01
6.97206616e-01 -6.19066834e-01 -1.46263123e+00 3.57748210e-01
8.89167905e-01 -3.68380576e-01 -9.71185744e-01 5.24448037e-01
2.40415975e-01 -6.60840631e-01 8.53220999e-01 -9.98008966e-01
2.98476517e-01 -7.94992805e-01 -2.97791392e-01 -6.42581522e-01
4.62590694e-01 -3.62819403e-01 -6.80225968e-01 1.15214992e+00
8.68694484e-01 -4.92668390e-01 3.86830211e-01 9.22625482e-01
1.89691946e-01 -6.50397360e-01 -1.20188200e+00 -4.81612056e-01
-2.55490482e-01 -5.18872619e-01 6.05624318e-01 7.83274293e-01
3.53000998e-01 5.21448314e-01 -1.22096986e-01 5.13307273e-01
1.45730739e-02 3.10490280e-01 3.97620469e-01 -1.05975413e+00
-1.93769671e-02 1.03651240e-01 -1.02277927e-01 -5.90520263e-01
-9.78235900e-02 -3.68417740e-01 -3.82488221e-02 -1.91832590e+00
-1.53595701e-01 -2.69264996e-01 8.45466107e-02 5.13114393e-01
-1.84294969e-01 -1.78848282e-01 -2.65869498e-01 -1.58041582e-01
-8.37324798e-01 2.24044994e-01 9.59850490e-01 -3.73183601e-02
-2.06788644e-01 -1.08178876e-01 -6.14825428e-01 4.39889699e-01
5.55339992e-01 -4.02970165e-01 -9.05686200e-01 -1.90926984e-01
1.12117124e+00 7.11135626e-01 4.45773154e-01 -8.34809840e-01
7.15224504e-01 -4.13926929e-01 -2.76438057e-01 -8.51177275e-01
3.37807417e-01 -1.05525911e+00 3.45613450e-01 2.09536344e-01
-1.57218516e-01 2.75866181e-01 5.93450129e-01 5.97616374e-01
-6.23477340e-01 -4.13559645e-01 1.34170949e-01 -3.04451227e-01
-8.83373082e-01 -1.09382801e-01 -1.65209725e-01 3.55183095e-01
1.30844545e+00 1.42474547e-01 -5.51114202e-01 -4.51440603e-01
-8.77066612e-01 5.26339531e-01 -5.66461623e-01 3.62106383e-01
6.08421683e-01 -1.05361533e+00 -5.40124714e-01 -6.17843449e-01
4.32051569e-01 6.92922994e-02 4.04393166e-01 8.37338626e-01
-7.01281786e-01 1.11305833e+00 1.89333618e-01 -4.53524105e-02
-1.05005002e+00 7.99707174e-01 5.06595612e-01 -6.69551790e-01
-3.30316514e-01 6.87505007e-01 -5.47131121e-01 -2.21680269e-01
8.37150663e-02 -6.44192696e-01 -6.66862845e-01 3.17079246e-01
6.06736243e-01 2.58635968e-01 4.34844613e-01 -3.02782863e-01
-6.23964787e-01 2.51429379e-01 1.15776330e-03 -4.01729107e-01
1.15432191e+00 -1.83332443e-01 -3.00151378e-01 4.71457571e-01
6.12059414e-01 2.37260699e-01 -3.65437061e-01 -4.88089383e-01
5.25706530e-01 6.99916808e-03 -4.38418478e-01 -1.19562805e+00
-4.44009364e-01 2.74655282e-01 2.00238861e-02 7.97962248e-01
1.11471188e+00 4.64366704e-01 8.47762704e-01 6.86815858e-01
5.54297030e-01 -1.06965375e+00 -1.83836684e-01 7.01890111e-01
1.25584161e+00 -9.02536213e-01 -1.45403370e-01 -4.63998765e-01
-4.27364290e-01 1.13373709e+00 6.83725059e-01 3.29858512e-01
6.08421683e-01 -2.50055380e-02 3.09577491e-03 -8.37264538e-01
-1.25984371e+00 -4.50640321e-01 2.29440466e-01 1.92347407e-01
1.42389640e-01 -3.90863687e-01 -7.90440321e-01 7.38964915e-01
-3.36293638e-01 3.15689892e-01 3.17354202e-01 1.42615843e+00
-5.27584851e-01 -9.31552529e-01 -4.06399399e-01 4.30081666e-01
-7.50448406e-01 -4.91541885e-02 -3.37085426e-01 8.75031233e-01
1.42502367e-01 1.28938508e+00 -2.90725738e-01 6.19077310e-02
7.94345438e-01 6.95596814e-01 4.28909421e-01 -4.47531402e-01
-6.38466895e-01 -6.40509427e-01 7.44827628e-01 -4.20595407e-01
-4.88562733e-01 -4.90038514e-01 -1.61328924e+00 1.32064909e-01
-2.83684909e-01 8.17216098e-01 3.64616692e-01 1.31770504e+00
2.92381853e-01 5.64362288e-01 8.67334753e-02 6.31335378e-01
-3.52088451e-01 -7.52104878e-01 -9.69846696e-02 3.72568846e-01
2.74906993e-01 -6.99248016e-01 -1.08492367e-01 4.04628038e-01]
|
[10.421568870544434, 7.961282253265381]
|
3ef69213-b2f8-4a69-ba74-a50e0ec8e002
|
the-transformative-potential-of-machine
|
2303.15832
| null |
https://arxiv.org/abs/2303.15832v2
|
https://arxiv.org/pdf/2303.15832v2.pdf
|
The transformative potential of machine learning for experiments in fluid mechanics
|
The field of machine learning has rapidly advanced the state of the art in many fields of science and engineering, including experimental fluid dynamics, which is one of the original big-data disciplines. This perspective will highlight several aspects of experimental fluid mechanics that stand to benefit from progress advances in machine learning, including: 1) augmenting the fidelity and quality of measurement techniques, 2) improving experimental design and surrogate digital-twin models and 3) enabling real-time estimation and control. In each case, we discuss recent success stories and ongoing challenges, along with caveats and limitations, and outline the potential for new avenues of ML-augmented and ML-enabled experimental fluid mechanics.
|
['Beverley J. McKeon', 'Steven L. Brunton', 'Ricardo Vinuesa']
|
2023-03-28
| null | null | null | null |
['experimental-design']
|
['methodology']
|
[-1.89267442e-01 -4.84660417e-01 -1.58336014e-02 1.48018315e-01
-3.75925660e-01 -6.06390573e-02 6.19168460e-01 2.38489762e-01
-4.71875131e-01 9.98201489e-01 8.41920525e-02 -3.65424186e-01
-3.05977434e-01 -6.53664410e-01 -7.87551343e-01 -7.61536181e-01
-6.23602331e-01 5.10977209e-01 -1.29979461e-01 -1.31898150e-01
2.57740021e-01 6.62728369e-01 -1.46497774e+00 -8.78751650e-02
3.21197778e-01 9.26414669e-01 -3.14668804e-01 1.02064991e+00
3.46235494e-04 5.97695470e-01 -6.23070076e-02 1.56905711e-01
2.21949235e-01 -1.69606492e-01 -3.52248311e-01 -4.36378956e-01
3.94576788e-01 -6.94935083e-01 -5.76247811e-01 2.35851854e-01
5.98938882e-01 2.71883279e-01 7.79261649e-01 -1.01619971e+00
-2.73505837e-01 4.57227230e-02 -3.11493874e-01 2.43741989e-01
5.93234710e-02 8.58712554e-01 6.92532897e-01 -9.37525928e-01
4.49311674e-01 1.23668683e+00 8.11484516e-01 5.40707111e-01
-1.27786767e+00 -8.28695893e-01 -2.20825255e-01 -2.13359445e-01
-8.20072472e-01 -6.07617617e-01 7.41029382e-01 -1.15276814e+00
9.71922934e-01 -9.62080732e-02 9.87843454e-01 7.49196887e-01
6.65321827e-01 5.76522291e-01 1.23095930e+00 -1.83884367e-01
4.56670731e-01 -1.45389214e-01 -1.43233716e-01 5.45638144e-01
7.32406139e-01 1.18029821e+00 -6.44930780e-01 -3.69128406e-01
1.15931809e+00 -1.78362235e-01 1.61198527e-02 -7.94808805e-01
-1.28866041e+00 8.49382877e-01 -2.10240036e-01 -8.99786595e-03
3.39217819e-02 5.51544726e-01 5.83188117e-01 3.85320842e-01
5.12419999e-01 1.01661587e+00 -9.78656113e-01 -4.66604352e-01
-8.19653273e-01 1.05764294e+00 9.30594862e-01 5.20968139e-01
5.91457844e-01 6.38929367e-01 4.18824881e-01 3.01114589e-01
6.03529453e-01 9.02060568e-01 6.74357563e-02 -1.08937168e+00
5.17570078e-01 2.06484348e-01 4.61297631e-01 -7.21188009e-01
-7.10097730e-01 4.20504808e-02 -7.73440778e-01 6.53683305e-01
4.88478273e-01 -6.83450580e-01 -4.76846874e-01 1.29161310e+00
5.05220652e-01 6.20566487e-01 -1.78249195e-01 8.50010037e-01
7.99437761e-01 4.10775244e-01 1.52340025e-01 -1.41193256e-01
5.37437916e-01 -4.42428708e-01 -5.83758652e-01 -3.90659086e-02
7.58952856e-01 -6.26289368e-01 8.01691949e-01 1.26212552e-01
-9.91730511e-01 -5.30860245e-01 -1.07870126e+00 1.98546156e-01
-3.90271902e-01 -5.65927625e-01 1.13513589e+00 5.17900288e-01
-5.29038191e-01 1.39208698e+00 -1.34303367e+00 1.06124848e-01
2.45656297e-01 4.96046066e-01 -4.07579929e-01 4.99097824e-01
-1.15059268e+00 8.77504170e-01 -1.35976747e-01 4.60288897e-02
-8.53470325e-01 -1.77065635e+00 -8.84965897e-01 -4.57183719e-01
-4.47371416e-02 -8.46950412e-01 1.20582819e+00 3.35062027e-01
-1.83430028e+00 5.58942795e-01 7.09931031e-02 -2.09706962e-01
7.36099124e-01 -8.36141169e-01 -1.23403057e-01 -7.05433637e-02
-2.84214765e-01 3.03427339e-01 4.11757678e-01 -1.20156348e+00
-5.44369102e-01 -2.67387390e-01 -6.82435393e-01 -4.36947107e-01
1.13082945e-01 -2.02696905e-01 4.98026341e-01 -4.37831193e-01
-1.49219543e-01 -8.95999253e-01 -6.52282476e-01 3.94449502e-01
-1.03344902e-01 1.83104634e-01 1.04713464e+00 -2.45162711e-01
1.06633925e+00 -1.43252051e+00 2.37807661e-01 1.06022939e-01
6.23008192e-01 3.82831812e-01 3.07365716e-01 6.93203390e-01
1.60291895e-01 1.77325383e-01 -4.41673808e-02 -2.98016310e-01
-3.49507388e-03 -1.30762637e-01 -4.55301344e-01 7.29142725e-01
2.39548519e-01 1.18583763e+00 -9.92590427e-01 -2.18718365e-01
1.02255356e+00 1.80062577e-01 -5.78918934e-01 3.97851855e-01
-2.70444155e-01 1.25627589e+00 -5.89117289e-01 5.82253277e-01
5.08364975e-01 -1.16175242e-01 -1.01555891e-01 -4.29666322e-03
-5.42802334e-01 8.80412757e-02 -1.44876003e+00 1.01173067e+00
-5.70592999e-01 6.34411871e-01 4.30065751e-01 -9.21326160e-01
9.25009072e-01 2.78823406e-01 1.26173544e+00 -5.10788083e-01
1.76164225e-01 5.17521322e-01 -4.13433798e-02 -9.00920212e-01
3.20777267e-01 -5.49340606e-01 6.87309131e-02 5.34004986e-01
-5.76610044e-02 -8.43506336e-01 -2.17567623e-01 -2.90681690e-01
7.61734784e-01 4.92494702e-01 2.03981742e-01 -6.68807745e-01
3.44910026e-01 -2.19800204e-01 2.55211830e-01 4.84802693e-01
-3.72865379e-01 9.15717781e-02 7.85564929e-02 -8.31544399e-01
-1.72002697e+00 -1.08085501e+00 -5.75835407e-01 7.07994461e-01
1.41532049e-01 -3.73643875e-01 -7.42095485e-02 1.31068796e-01
1.29158258e+00 -5.25169373e-02 -5.99046707e-01 -7.59745762e-03
-1.12970352e+00 -8.43724072e-01 4.17762190e-01 7.28028774e-01
-3.04375924e-02 -7.33240128e-01 -4.08090413e-01 3.60931396e-01
5.86640775e-01 -1.01849616e+00 1.61701828e-01 -1.09376293e-02
-1.20563173e+00 -1.18072748e+00 -1.88842416e-01 -3.46946061e-01
6.84190318e-02 3.43174301e-02 1.12678325e+00 3.33585620e-01
-5.64416826e-01 4.19690251e-01 4.34606783e-02 -4.46976691e-01
-5.05056262e-01 -2.97236204e-01 7.96341062e-01 -6.21438920e-01
1.58475265e-01 -6.85812891e-01 -6.11017466e-01 2.08436847e-01
-2.29062751e-01 -1.95895195e-01 3.43245506e-01 6.67072833e-01
4.25070435e-01 -3.52877438e-01 7.77802944e-01 -6.22917771e-01
5.00152826e-01 -4.50428963e-01 -1.00469995e+00 -5.68547249e-01
-6.62739456e-01 5.70483431e-02 1.03221071e+00 -4.05610323e-01
-6.41808271e-01 -2.54548341e-01 -2.72387326e-01 -4.41394091e-01
-8.01022649e-02 1.74171865e-01 2.02609122e-01 -6.57825351e-01
4.70678329e-01 -2.67449617e-01 5.10048270e-01 -5.82633436e-01
4.32281308e-02 4.74894017e-01 9.93825719e-02 -1.34339595e+00
8.85671258e-01 4.04228091e-01 8.15628290e-01 -1.14528728e+00
-3.26319695e-01 -2.88701445e-01 -9.99706268e-01 -3.32885474e-01
5.11017978e-01 -6.89545691e-01 -1.20950973e+00 6.29564404e-01
-3.95005852e-01 -8.35718513e-01 -3.99973780e-01 8.36325586e-01
-7.54935205e-01 1.41345695e-01 -9.59062934e-01 -9.87452090e-01
-2.63779342e-01 -1.38235629e+00 1.21898174e+00 2.16283962e-01
-4.30243164e-01 -1.48542023e+00 4.86169100e-01 1.17570274e-01
7.90974677e-01 7.28997767e-01 7.65556633e-01 -3.22477281e-01
-6.97288156e-01 -3.24257255e-01 2.61733949e-01 1.13002412e-01
7.20903184e-03 4.68551964e-01 -9.69597876e-01 -4.78857428e-01
-2.72105426e-01 -5.55952191e-01 7.68879652e-01 6.80890322e-01
8.66324902e-01 1.20877042e-01 -7.05145717e-01 6.07786179e-01
1.21485114e+00 -1.62293464e-01 4.23647985e-02 1.02609314e-01
7.26360798e-01 3.92655015e-01 4.12144095e-01 5.95673740e-01
5.03737517e-02 3.78698528e-01 -6.76222518e-02 -3.38903725e-01
-8.59840736e-02 -5.50547183e-01 7.01861084e-02 1.13578498e+00
-1.58018783e-01 3.50662082e-01 -7.95618296e-01 2.44701669e-01
-1.55399394e+00 -8.62576723e-01 -2.64169037e-01 2.26265264e+00
6.60563648e-01 1.44050688e-01 5.27077615e-02 -1.58017129e-02
2.33427659e-01 1.57975033e-01 -9.00727570e-01 -3.33340913e-01
9.10858363e-02 4.15910453e-01 5.05053341e-01 8.12270343e-01
-1.40184927e+00 7.89858699e-01 8.60576344e+00 2.83523768e-01
-1.22162807e+00 -3.95781398e-01 5.49677789e-01 -1.21720158e-01
-6.53426200e-02 -1.48911059e-01 -1.10947239e+00 4.58855897e-01
9.25748229e-01 -4.12853956e-01 3.38565111e-01 5.99286377e-01
7.13489056e-01 -1.07761644e-01 -1.38739443e+00 4.67636436e-01
-4.99629200e-01 -1.79100859e+00 -8.31154659e-02 2.35470042e-01
7.57310152e-01 1.55080631e-01 -7.26605356e-02 3.09808671e-01
2.09105313e-01 -1.05997026e+00 3.40893120e-01 7.53861606e-01
1.06886828e+00 -2.58227855e-01 6.89414978e-01 1.32890001e-01
-1.48286533e+00 3.72399837e-02 -3.46058458e-01 -6.54675663e-01
3.72288108e-01 7.48898208e-01 -3.60957086e-01 1.39053091e-01
4.93679941e-01 7.94030786e-01 1.78518742e-01 9.61590469e-01
2.45029956e-01 8.03932905e-01 -6.39662683e-01 -3.53095472e-01
-1.42779186e-01 -3.83486062e-01 7.56649554e-01 1.09136736e+00
-9.42886770e-02 6.05346933e-02 5.29992700e-01 1.00397730e+00
7.91222453e-02 -1.97862044e-01 -8.08754385e-01 -2.23970920e-01
5.72236955e-01 1.04993415e+00 -1.74072623e-01 -5.81458919e-02
-4.38024014e-01 -4.35809910e-01 -1.85449291e-02 6.45097271e-02
-5.37906766e-01 -9.98848230e-02 1.41931760e+00 5.73502004e-01
-2.06581578e-01 -8.33196402e-01 -5.29175818e-01 -1.00555563e+00
-3.38754594e-01 -4.28661436e-01 -1.01591036e-01 -1.83272824e-01
-1.57233787e+00 -4.83415216e-01 2.04745129e-01 -1.15863776e+00
-3.48675430e-01 -1.08061397e+00 -5.32269835e-01 5.94740033e-01
-1.45874596e+00 -1.04148233e+00 5.29010445e-02 -2.24388003e-01
2.85989314e-01 -2.95794308e-01 8.17379653e-01 4.47929651e-01
-4.62762266e-01 3.44387680e-01 6.97475016e-01 5.85137904e-02
7.35227048e-01 -1.07410300e+00 6.03969395e-01 2.35891566e-01
-5.51767886e-01 5.38757086e-01 1.09632587e+00 -1.05683041e+00
-2.20064211e+00 -9.70297635e-01 1.59141794e-01 -7.10377634e-01
1.21237791e+00 -3.43100011e-01 -9.15518463e-01 5.53049743e-01
-6.13507867e-01 2.53317535e-01 6.85665488e-01 1.47713199e-01
2.63438612e-01 -6.26235679e-02 -1.21279621e+00 5.93585968e-01
9.58766162e-01 -4.69415605e-01 -2.49803603e-01 3.10602456e-01
5.47805667e-01 -7.28435874e-01 -1.36325288e+00 7.58936763e-01
1.08322453e+00 -5.97567260e-01 1.11066270e+00 -1.09345937e+00
6.26360595e-01 -5.58315292e-02 6.77492470e-02 -1.15263879e+00
-2.08604738e-01 -9.39153671e-01 -7.45820642e-01 7.41972744e-01
2.26422906e-01 -5.95858514e-01 1.12597406e+00 9.01377738e-01
-1.88824087e-01 -1.28222227e+00 -9.58070397e-01 -7.59101331e-01
1.16944885e+00 -5.69646180e-01 5.61829865e-01 6.23849630e-01
3.41863334e-01 -7.89805576e-02 -5.18980503e-01 -1.51379377e-01
8.72167766e-01 1.81909844e-01 1.14557540e+00 -1.43394482e+00
-1.63964927e-01 -4.34702486e-01 -5.57356358e-01 -9.44427729e-01
2.29947910e-01 -4.71083701e-01 -1.00397907e-01 -9.85371113e-01
-4.78858836e-02 -5.81012845e-01 1.10568300e-01 -1.70405045e-01
-2.38771632e-01 1.11936443e-01 -7.09219894e-04 1.19903736e-01
-2.99689144e-01 6.20998859e-01 1.73878706e+00 1.21258430e-01
-3.63680393e-01 4.88279313e-02 -3.20139468e-01 6.37039542e-01
7.05559134e-01 5.50812781e-02 3.54081951e-02 -7.53876928e-04
1.51855335e-01 2.23175094e-01 3.44495863e-01 -1.18490088e+00
1.85252786e-01 -7.11074889e-01 7.94373155e-01 -2.67473489e-01
1.88218415e-01 -5.67592502e-01 -1.98356196e-01 7.25487113e-01
-2.49298319e-01 -1.48624144e-02 6.87483370e-01 5.44538140e-01
2.36225948e-01 4.79890049e-01 1.26777112e+00 -1.40835255e-01
-5.16613781e-01 5.68640530e-01 -5.37289798e-01 3.42181027e-01
1.13070560e+00 -1.17285870e-01 -4.31268066e-01 -2.89122146e-02
-4.78264630e-01 3.49494219e-01 4.14268106e-01 3.81964087e-01
4.40019995e-01 -1.09597421e+00 -7.56623328e-01 6.69303894e-01
-3.03624600e-01 -3.69220972e-03 2.62183577e-01 6.91950202e-01
-6.16387367e-01 7.19614387e-01 -1.14784755e-01 -5.92084765e-01
-7.03265607e-01 3.97622347e-01 6.43336296e-01 -1.69175237e-01
-9.06050146e-01 5.10915995e-01 -1.23389423e-01 -7.02350974e-01
-2.31663421e-01 -3.64255637e-01 3.58544588e-02 -6.77611530e-01
4.35871810e-01 8.84921670e-01 9.28355828e-02 -4.86201048e-01
-1.54821038e-01 1.05666924e+00 4.01241213e-01 2.07021683e-01
1.52935004e+00 7.27491975e-02 1.21877864e-01 7.99984872e-01
9.51367557e-01 7.39907101e-02 -1.62739444e+00 2.20055461e-01
-1.19206712e-01 -4.32952851e-01 1.59317389e-01 -4.97477710e-01
-9.18097079e-01 1.07554877e+00 3.85390013e-01 -2.50095695e-01
1.47594765e-01 -1.64741099e-01 1.16656995e+00 4.61716115e-01
5.56258917e-01 -1.06408262e+00 8.76680315e-02 5.56593955e-01
7.19975948e-01 -1.22316921e+00 5.17631054e-01 -9.79000926e-02
2.25924123e-02 1.30197299e+00 8.10491025e-01 -5.32686889e-01
1.46397686e+00 1.17895460e+00 -1.27634723e-02 -8.63171741e-02
-8.86830688e-01 2.97932357e-01 6.38010576e-02 5.70007205e-01
7.10857868e-01 1.22529097e-01 -3.09531894e-02 3.88717651e-01
-2.82750845e-01 3.66829365e-01 3.08939695e-01 1.15472639e+00
-8.17178845e-01 -1.03365993e+00 -4.57898706e-01 1.04546225e+00
-2.14440167e-01 2.70530850e-01 2.62581371e-02 8.68832111e-01
-2.43472066e-02 6.65078998e-01 1.10191204e-01 -5.08191884e-01
4.60631937e-01 4.60204203e-03 4.52043772e-01 -1.87587395e-01
-3.38270009e-01 2.36004777e-02 -5.79560995e-02 -5.27889252e-01
-1.19296618e-01 -6.83838129e-01 -1.36702502e+00 -9.52985644e-01
-5.01464486e-01 1.00046746e-01 5.64664841e-01 1.19749093e+00
3.42156589e-01 4.43312019e-01 4.52665389e-01 -1.65071595e+00
-5.30147552e-01 -8.10328066e-01 -7.50044227e-01 2.25574911e-01
7.29106963e-01 -1.33202136e+00 -7.51278102e-01 -4.16782387e-02]
|
[6.373892307281494, 3.568662166595459]
|
a322dded-2a29-4e6c-a2d1-44bc086298da
|
scapegoat-generation-for-privacy-protection
|
2303.0293
| null |
https://arxiv.org/abs/2303.02930v1
|
https://arxiv.org/pdf/2303.02930v1.pdf
|
Scapegoat Generation for Privacy Protection from Deepfake
|
To protect privacy and prevent malicious use of deepfake, current studies propose methods that interfere with the generation process, such as detection and destruction approaches. However, these methods suffer from sub-optimal generalization performance to unseen models and add undesirable noise to the original image. To address these problems, we propose a new problem formulation for deepfake prevention: generating a ``scapegoat image'' by modifying the style of the original input in a way that is recognizable as an avatar by the user, but impossible to reconstruct the real face. Even in the case of malicious deepfake, the privacy of the users is still protected. To achieve this, we introduce an optimization-based editing method that utilizes GAN inversion to discourage deepfake models from generating similar scapegoats. We validate the effectiveness of our proposed method through quantitative and user studies.
|
['Shigeo Morishima', 'Hirokatsu Kataoka', 'Hideki Tsunashima', 'Mariko Isogawa', 'Yoshihiro Fukuhara', 'Gido Kato']
|
2023-03-06
| null | null | null | null |
['face-swapping']
|
['computer-vision']
|
[ 3.72752100e-01 5.86858690e-01 2.61929572e-01 -1.23722911e-01
-1.80533469e-01 -9.29936528e-01 6.01941407e-01 -5.85846007e-01
-1.48875102e-01 5.95387697e-01 5.41306697e-02 1.65972468e-02
1.38791040e-01 -9.56118643e-01 -8.08171034e-01 -6.87000215e-01
5.56949258e-01 -1.59603700e-01 -1.54018953e-01 -8.54399875e-02
1.26960784e-01 5.56633592e-01 -1.33513927e+00 3.44808489e-01
9.06074405e-01 1.00642288e+00 -1.47224784e-01 3.87769133e-01
6.37025014e-02 5.48484862e-01 -8.76145244e-01 -1.05122685e+00
7.82210648e-01 -6.53765380e-01 -4.73227829e-01 1.70291856e-01
6.12526774e-01 -6.12331569e-01 -3.61547381e-01 1.32456052e+00
3.42215806e-01 -2.56511569e-01 4.55580384e-01 -1.49401248e+00
-9.01978731e-01 1.60707101e-01 -7.02942610e-01 -5.09782434e-01
2.76187420e-01 1.08200341e-01 3.24094504e-01 -6.00556016e-01
5.62820852e-01 1.27603734e+00 7.11733341e-01 1.05951846e+00
-1.34838963e+00 -1.11098754e+00 -1.58967689e-01 -2.01285809e-01
-1.52395988e+00 -6.16437435e-01 8.92642915e-01 -2.58999020e-01
5.33546507e-02 6.35375619e-01 7.80954182e-01 1.45813978e+00
9.43882614e-02 6.94255471e-01 1.22807670e+00 -3.14428270e-01
5.92697412e-02 5.54950595e-01 -3.41923177e-01 8.24873745e-01
3.01566154e-01 2.20413566e-01 -3.43394011e-01 -4.71466571e-01
8.94082665e-01 -2.18750387e-01 -3.31978410e-01 -4.09680843e-01
-6.71980441e-01 7.31180191e-01 2.15444297e-01 -1.99611679e-01
-3.41183037e-01 4.99050692e-02 1.39203951e-01 1.36711091e-01
2.97616720e-01 4.75528985e-01 -1.83135495e-02 1.59002662e-01
-8.61507058e-01 2.35673070e-01 5.97183466e-01 7.61327744e-01
7.17089534e-01 1.09481387e-01 -1.16723917e-01 7.66438305e-01
2.83840150e-01 4.93077964e-01 3.25147867e-01 -9.28913653e-01
2.88889915e-01 5.91445982e-01 1.22982770e-01 -1.36367559e+00
3.21400970e-01 -3.85315239e-01 -8.80751789e-01 6.72250628e-01
5.30570924e-01 -2.82053143e-01 -9.42637861e-01 1.77034879e+00
4.58475113e-01 9.77918208e-02 3.04162167e-02 8.22611392e-01
6.21446490e-01 6.98660672e-01 -1.66752815e-01 -1.19000547e-01
1.13275969e+00 -8.28594029e-01 -8.28256965e-01 -7.15498701e-02
3.03885788e-01 -8.96165550e-01 1.19316483e+00 6.40959144e-01
-9.74530458e-01 -3.45409393e-01 -1.05493927e+00 9.57441181e-02
-2.83499304e-02 1.63969353e-01 3.16331387e-01 1.55740249e+00
-8.33143294e-01 2.85238773e-01 -2.92094767e-01 -1.93568647e-01
6.94242895e-01 2.28521928e-01 -5.97379565e-01 1.89901114e-01
-9.41818476e-01 7.40405619e-01 1.46283299e-01 2.56126970e-01
-1.04732108e+00 -7.32708454e-01 -6.38040364e-01 -5.93241006e-02
5.18543720e-01 -6.29997492e-01 9.10478830e-01 -1.57982373e+00
-1.58374262e+00 9.65599179e-01 9.16832834e-02 -2.37900510e-01
1.11240327e+00 -2.82332212e-01 -2.89605260e-01 -6.04440160e-02
-1.87315062e-01 4.72335547e-01 1.43015182e+00 -1.74270439e+00
-2.14651167e-01 -5.50380766e-01 1.76046714e-02 1.84986424e-02
-5.27238965e-01 -1.10640854e-01 -3.43143016e-01 -9.89283919e-01
-4.13116887e-02 -1.13946092e+00 -4.28619497e-02 3.87069702e-01
-7.50910163e-01 4.24655080e-01 1.47327924e+00 -9.54392433e-01
9.92102683e-01 -2.36201859e+00 -1.96892366e-01 4.68155622e-01
3.44111741e-01 6.82417274e-01 -1.89524606e-01 1.88599274e-01
8.10965672e-02 4.62072313e-01 -1.23459212e-01 -4.20030594e-01
-2.33779028e-01 3.12298149e-01 -3.39957178e-01 3.65157962e-01
-2.37968042e-01 8.33454311e-01 -4.38197643e-01 -1.58559307e-01
-8.55640024e-02 5.35110116e-01 -6.35261834e-01 3.61860514e-01
3.06240059e-02 4.90889996e-01 -3.42739433e-01 4.39703822e-01
1.22815835e+00 2.79963315e-01 1.06872581e-01 -2.78332651e-01
7.23971128e-02 -2.31258452e-01 -9.46208775e-01 1.16992152e+00
-3.54921192e-01 4.07274544e-01 5.11232734e-01 -4.37399954e-01
9.39494729e-01 1.61807522e-01 1.29269332e-01 -5.00162661e-01
3.11044574e-01 2.91251671e-03 -1.38647720e-01 -5.24602473e-01
5.55743814e-01 -1.96640760e-01 4.39822972e-02 5.48104048e-01
-4.11022514e-01 -6.37845090e-03 -5.45908988e-01 1.93328351e-01
7.02750087e-01 2.19494264e-04 3.78526561e-02 -7.86993876e-02
4.06808853e-01 -4.52206075e-01 6.79713130e-01 7.84537554e-01
-1.78055167e-01 6.88100636e-01 6.40285671e-01 -4.17648733e-01
-1.12912011e+00 -1.09808207e+00 3.00030708e-01 5.79890788e-01
2.57191002e-01 -4.89433974e-01 -1.39307010e+00 -1.21193600e+00
1.81387346e-02 7.44236112e-01 -7.64998615e-01 -4.35318410e-01
-3.39395285e-01 -4.39759165e-01 1.14720380e+00 1.68256108e-02
9.64259148e-01 -7.82548606e-01 -4.35194552e-01 -1.66277602e-01
-3.36625516e-01 -1.03224158e+00 -7.20657945e-01 -6.64100111e-01
-4.79451597e-01 -1.06392038e+00 -7.87635326e-01 -5.43860197e-01
1.03358245e+00 2.13111877e-01 3.99185777e-01 3.28436613e-01
-1.89285025e-01 2.90548563e-01 -2.76978314e-01 -4.45790589e-01
-8.36735904e-01 -2.96787411e-01 -8.83723870e-02 5.96466720e-01
-1.95714787e-01 -4.52729553e-01 -6.73580408e-01 5.28619230e-01
-1.09279716e+00 1.71114832e-01 3.97285849e-01 7.06938744e-01
1.98386192e-01 3.20178062e-01 3.71285975e-01 -1.06680739e+00
8.95173967e-01 -4.25678752e-02 -3.74051601e-01 3.33623886e-01
-5.89571476e-01 -1.39633760e-01 6.86256170e-01 -7.11485207e-01
-1.26576662e+00 1.60320863e-01 -1.78967014e-01 -4.74112839e-01
-7.16609731e-02 -1.36614591e-01 -7.45359957e-01 -6.39144361e-01
5.76018035e-01 5.53238928e-01 4.43256646e-01 -3.95401180e-01
3.64610076e-01 7.46691644e-01 4.96342957e-01 -4.70330864e-01
9.97370899e-01 7.08196878e-01 -1.62399009e-01 -9.18632448e-01
-3.49276274e-01 2.58224189e-01 -2.14528769e-01 -4.89910364e-01
5.59394181e-01 -5.19124985e-01 -8.14441502e-01 9.96437013e-01
-1.20758498e+00 -2.55571455e-02 -2.25376766e-02 -1.77881457e-02
-3.74349743e-01 8.44774663e-01 -2.59967268e-01 -8.85934472e-01
-3.58362198e-01 -1.00539660e+00 8.34029317e-01 2.40992993e-01
-1.82647005e-01 -4.99653131e-01 -5.48574738e-02 7.55300522e-01
2.62986720e-01 4.39735025e-01 9.96142328e-01 -9.67605636e-02
-4.87348497e-01 -3.51786464e-01 -1.73412710e-01 7.36607373e-01
3.18185031e-01 3.78814824e-02 -9.89065409e-01 -2.32404992e-01
2.11868346e-01 -2.79581007e-02 5.50588369e-01 -1.85249791e-01
1.32972705e+00 -9.90191042e-01 -1.94414496e-01 7.95983851e-01
1.14700556e+00 3.43950033e-01 1.28246629e+00 6.59800023e-02
7.65117884e-01 7.94825196e-01 3.12580347e-01 4.60802466e-01
1.59426965e-02 5.26404262e-01 7.17928410e-01 -6.45945892e-02
-6.90067261e-02 -7.98023522e-01 3.21714103e-01 2.77613904e-02
-1.53131023e-01 -4.12127882e-01 -4.29203808e-01 3.27778429e-01
-1.66326988e+00 -9.19620097e-01 -1.94028735e-01 2.22723913e+00
4.55160469e-01 -1.10937074e-01 9.57691669e-02 6.41274452e-02
8.92759085e-01 1.42254233e-01 -2.60467440e-01 -4.87335920e-01
-7.71450698e-02 -1.29304384e-03 4.97018903e-01 2.81329274e-01
-8.70634258e-01 1.06874073e+00 6.14918613e+00 8.61759663e-01
-1.30353785e+00 1.79633021e-01 7.49678791e-01 3.66067402e-02
-5.59854031e-01 6.61133528e-02 -3.96770358e-01 5.92206717e-01
2.19769746e-01 -3.96054499e-02 5.85242212e-01 7.25438654e-01
4.13992822e-01 3.12193204e-02 -6.62300944e-01 1.12082696e+00
3.23323309e-01 -1.21182358e+00 2.20975742e-01 3.23834389e-01
4.61360991e-01 -9.42869842e-01 3.85025293e-01 3.64310294e-02
-6.91249967e-02 -1.08349621e+00 8.40474844e-01 4.41704184e-01
6.79470360e-01 -1.00801611e+00 3.37533176e-01 5.05375624e-01
-6.56568110e-01 -9.48922038e-02 -1.51873738e-01 2.05856577e-01
3.72762345e-02 3.70594382e-01 -8.79210293e-01 2.52105087e-01
4.67996836e-01 1.63355481e-03 -4.82240438e-01 6.48896456e-01
-3.88227552e-01 3.60775530e-01 -2.16458127e-01 1.41492710e-01
-3.09545528e-02 -2.27322727e-01 7.31856883e-01 8.33073795e-01
5.82884789e-01 1.12298541e-01 -2.17392206e-01 1.06756854e+00
-2.29995891e-01 1.51863575e-01 -9.08024788e-01 -2.05179349e-01
3.35086644e-01 1.20790625e+00 -3.14626575e-01 8.29384625e-02
-1.17918707e-01 1.52174115e+00 9.50976182e-03 3.00379157e-01
-9.37709212e-01 -2.16011643e-01 7.82819033e-01 5.57163358e-01
-1.24946646e-01 -1.71264216e-01 -3.83415341e-01 -1.06949604e+00
2.40971357e-01 -1.18979919e+00 1.54508457e-01 -8.32551241e-01
-1.14793837e+00 6.85276985e-01 -2.93059081e-01 -1.05228233e+00
6.99623451e-02 -2.00804278e-01 -6.37555659e-01 6.22897804e-01
-8.41012776e-01 -1.56020117e+00 -3.63812566e-01 7.92358398e-01
2.17884466e-01 -3.49569410e-01 6.79634631e-01 2.44977891e-01
-4.38296199e-01 1.14056826e+00 -1.49108902e-01 1.94600970e-01
4.94451523e-01 -5.45036852e-01 2.92206705e-01 9.42918003e-01
-4.50254828e-02 6.78374112e-01 5.57588100e-01 -9.16892707e-01
-1.27967989e+00 -1.12293935e+00 4.34058934e-01 -1.88691765e-01
4.66295592e-02 -6.27064764e-01 -7.54485846e-01 5.51797569e-01
1.74981385e-01 -2.50871658e-01 5.93856871e-01 -3.64648223e-01
-4.64681745e-01 -1.64024383e-01 -1.71500123e+00 8.21312606e-01
9.98734355e-01 -5.51593363e-01 -1.42744780e-01 -1.89744279e-01
2.50018835e-01 -9.94223356e-02 -4.05552089e-01 2.90490091e-01
9.95189905e-01 -1.14499855e+00 8.85178149e-01 -4.77869570e-01
1.94484398e-01 -2.18756527e-01 3.14809158e-02 -1.25746787e+00
2.42385864e-02 -1.04085660e+00 -6.76336959e-02 1.42729533e+00
1.40675500e-01 -5.65706074e-01 1.23001873e+00 8.11650991e-01
3.53009939e-01 -2.42449373e-01 -8.82459223e-01 -9.25252736e-01
-1.00099273e-01 -1.97113201e-01 6.84980214e-01 1.12369049e+00
-3.16206932e-01 -1.41749263e-01 -1.17319119e+00 3.62491548e-01
8.65399420e-01 -3.96545619e-01 1.24413586e+00 -8.92274201e-01
-9.02393684e-02 -3.99091020e-02 -4.22440439e-01 -9.34613764e-01
1.95652097e-01 -6.53243959e-01 -1.00117274e-01 -8.68519187e-01
1.64504722e-01 -4.05189902e-01 1.91294819e-01 4.90483046e-01
1.98082060e-01 5.09957135e-01 4.54168469e-01 1.04573987e-01
2.33377442e-01 5.94756126e-01 1.51116955e+00 -5.31326383e-02
-1.15315594e-01 3.03456429e-02 -9.83315587e-01 7.86319137e-01
7.89135456e-01 -6.90940142e-01 -7.32001305e-01 -3.49879563e-01
1.76572412e-01 1.04251347e-01 7.62003779e-01 -8.04554582e-01
-2.76941597e-01 -3.23915422e-01 2.70670027e-01 -1.69793472e-01
3.88103515e-01 -1.06923616e+00 6.07816756e-01 4.74534094e-01
-2.09172681e-01 -1.95465773e-01 9.91106927e-02 5.43965220e-01
1.20835043e-01 -3.13048810e-01 1.15609610e+00 -2.05314353e-01
-2.91427940e-01 1.85402200e-01 -5.41858613e-01 -4.01513964e-01
1.22162986e+00 -5.17496824e-01 -3.53100628e-01 -8.11303854e-01
-6.04512453e-01 -1.66578859e-01 7.33542740e-01 4.58859652e-01
7.91468382e-01 -1.38020587e+00 -6.00118339e-01 5.98984003e-01
-1.92683235e-01 -4.28920776e-01 4.26313937e-01 3.43921930e-01
-6.37941241e-01 -3.09956133e-01 -3.87842447e-01 -4.77434397e-02
-1.68498540e+00 5.34371495e-01 4.68619496e-01 1.97724521e-01
-5.21613598e-01 6.26582205e-01 6.32933319e-01 -4.40005392e-01
5.74631877e-02 2.58243710e-01 9.91272107e-02 -1.58687636e-01
4.87316459e-01 3.03880692e-01 -1.19917311e-01 -7.13806689e-01
-1.53019235e-01 1.46927580e-01 -1.25244871e-01 -6.94709569e-02
8.85087013e-01 -3.70482594e-01 -7.17318654e-02 -4.56423163e-01
1.07113695e+00 6.14497542e-01 -1.32585156e+00 1.88027114e-01
-5.07536113e-01 -1.19014668e+00 -1.96444675e-01 -8.47496331e-01
-1.26930940e+00 6.00763261e-01 7.31855154e-01 3.60738486e-01
1.15359974e+00 -3.48356575e-01 1.14147651e+00 7.86992311e-02
3.05229366e-01 -1.00663090e+00 3.12651455e-01 8.30763206e-02
1.10188746e+00 -8.52077365e-01 -3.22565973e-01 -7.74058402e-01
-7.13500738e-01 8.33843529e-01 8.48673224e-01 4.82431017e-02
5.29504418e-01 2.14095905e-01 3.55038196e-01 1.67902723e-01
-2.14759469e-01 4.11863178e-01 1.60263434e-01 8.68388772e-01
-2.39373147e-01 -4.90070656e-02 -1.94432259e-01 7.74050474e-01
-3.45781803e-01 -1.86468586e-02 7.68694997e-01 7.42328167e-01
-3.32923345e-02 -1.28469145e+00 -6.94984972e-01 1.11691214e-01
-5.27938783e-01 1.92073539e-01 -9.66381907e-01 7.30026484e-01
4.79500115e-01 9.45271492e-01 -2.19640300e-01 -6.68880761e-01
2.71046609e-01 -7.10307360e-02 5.71470976e-01 -2.49033108e-01
-5.98529637e-01 6.25194833e-02 1.47381485e-01 -3.33376467e-01
-3.83858606e-02 -2.45704398e-01 -6.32290184e-01 -6.69581056e-01
-2.03338712e-01 -2.72097602e-03 5.89113533e-01 7.81960666e-01
3.31978410e-01 -8.52983072e-02 8.63135397e-01 -3.52707714e-01
-3.63079727e-01 -5.56719899e-01 -7.21106827e-01 6.11470222e-01
2.06039026e-01 -3.34949791e-01 -2.34458819e-01 -4.81665507e-02]
|
[12.71765422821045, 0.816485583782196]
|
5cea2f44-a590-4dbb-b041-b25fe0c432c6
|
decision-making-under-miscalibration
|
2203.09852
| null |
https://arxiv.org/abs/2203.09852v1
|
https://arxiv.org/pdf/2203.09852v1.pdf
|
Decision-Making under Miscalibration
|
ML-based predictions are used to inform consequential decisions about individuals. How should we use predictions (e.g., risk of heart attack) to inform downstream binary classification decisions (e.g., undergoing a medical procedure)? When the risk estimates are perfectly calibrated, the answer is well understood: a classification problem's cost structure induces an optimal treatment threshold $j^{\star}$. In practice, however, some amount of miscalibration is unavoidable, raising a fundamental question: how should one use potentially miscalibrated predictions to inform binary decisions? We formalize a natural (distribution-free) solution concept: given anticipated miscalibration of $\alpha$, we propose using the threshold $j$ that minimizes the worst-case regret over all $\alpha$-miscalibrated predictors, where the regret is the difference in clinical utility between using the threshold in question and using the optimal threshold in hindsight. We provide closed form expressions for $j$ when miscalibration is measured using both expected and maximum calibration error, which reveal that it indeed differs from $j^{\star}$ (the optimal threshold under perfect calibration). We validate our theoretical findings on real data, demonstrating that there are natural cases in which making decisions using $j$ improves the clinical utility.
|
['Gal Yona', 'Guy N. Rothblum']
|
2022-03-18
| null | null | null | null |
['medical-procedure']
|
['medical']
|
[ 5.55221140e-01 8.51074576e-01 -6.44334316e-01 -6.29677117e-01
-8.67788196e-01 -4.21136111e-01 -2.14259654e-01 5.54578006e-01
-5.73344886e-01 1.16204977e+00 9.83187854e-02 -8.23838532e-01
-5.76431513e-01 -8.54352117e-01 -8.32594275e-01 -6.98871017e-01
-2.80266464e-01 4.52040404e-01 -5.48071623e-01 1.20968044e-01
9.21139121e-02 1.27187550e-01 -1.03565478e+00 9.32630803e-03
1.25637329e+00 1.14791441e+00 -3.81888330e-01 4.85383987e-01
3.55895668e-01 7.34494328e-01 -5.91086388e-01 -1.12613606e+00
4.83107954e-01 -7.15632141e-01 -6.30956411e-01 -2.01330930e-01
-8.94135796e-03 -3.31338614e-01 7.10549131e-02 1.22046471e+00
3.23592514e-01 -1.43231913e-01 7.45534599e-01 -9.91555512e-01
-2.14494333e-01 1.01086450e+00 -1.96521029e-01 1.15595132e-01
2.10605010e-01 2.28405058e-01 1.09113574e+00 1.09292552e-01
3.71788591e-01 1.02072608e+00 9.05481279e-01 5.13016403e-01
-1.39959478e+00 -8.06943595e-01 1.18300386e-01 -3.28509659e-01
-1.35947251e+00 -3.33249718e-01 2.79654741e-01 -7.13209927e-01
4.45463836e-01 6.45400822e-01 5.89887619e-01 7.79845357e-01
8.11221719e-01 3.01573545e-01 1.05571222e+00 -4.02789265e-01
4.68351275e-01 3.21294039e-01 2.72045076e-01 4.54697996e-01
8.09614062e-01 6.16080463e-01 -2.50913084e-01 -6.60077035e-01
5.45664430e-01 1.14671856e-01 -4.18694496e-01 -6.50042966e-02
-9.61110353e-01 8.40612292e-01 2.94250697e-01 -2.89173782e-01
-4.52139586e-01 1.82696238e-01 1.58765808e-01 3.52926701e-01
2.15743303e-01 7.20516205e-01 -5.56323647e-01 -1.21897034e-01
-7.45142817e-01 1.61306337e-01 7.01402962e-01 7.15943754e-01
3.37814122e-01 -4.38403279e-01 -3.71188700e-01 5.10394335e-01
-1.00866683e-01 5.33011913e-01 2.73058861e-01 -1.13015878e+00
4.16886449e-01 4.35022622e-01 5.97781956e-01 -7.47037232e-01
-4.74833310e-01 -8.42122436e-01 -1.02123976e+00 -1.49472177e-01
7.27843285e-01 -7.59194732e-01 -6.09524369e-01 2.06769657e+00
-2.46961825e-02 -8.23242143e-02 8.64474326e-02 7.01186180e-01
3.23373526e-02 2.99572498e-02 4.33843017e-01 -7.95745730e-01
1.12470460e+00 -3.58559797e-03 -4.92320269e-01 -4.34865564e-01
8.95491302e-01 -3.44182760e-01 5.88658214e-01 3.85272235e-01
-1.05443931e+00 1.95397779e-01 -9.53044116e-01 5.91796339e-01
2.27955058e-01 -8.76709595e-02 8.11604857e-01 9.83380497e-01
-5.59525430e-01 9.17851210e-01 -5.29356122e-01 -1.11747503e-01
4.64563519e-01 6.13121092e-01 -6.57708198e-02 3.59322950e-02
-1.32996714e+00 9.86445606e-01 1.91716805e-01 4.96208221e-02
-7.42649496e-01 -9.96314943e-01 -4.94242251e-01 1.50521040e-01
5.66726267e-01 -9.97658074e-01 1.23738885e+00 -1.15675378e+00
-1.03241456e+00 7.80449390e-01 -6.09725006e-02 -8.59228671e-01
9.54124749e-01 -2.69345120e-02 -2.06205681e-01 -3.01776797e-01
1.97012529e-01 2.56130755e-01 5.01037717e-01 -7.30647206e-01
-6.52004361e-01 -4.69082117e-01 5.76559715e-02 9.53012183e-02
9.58389714e-02 -2.92549402e-01 3.48770082e-01 -5.54135084e-01
1.34350285e-01 -9.50650871e-01 -7.97634542e-01 8.94309357e-02
-7.31433451e-01 3.16214621e-01 -3.72709185e-01 -4.79485720e-01
1.42376721e+00 -1.95554185e+00 -3.69204551e-01 4.21698570e-01
2.41872966e-01 -1.42902955e-01 4.89001632e-01 1.65300183e-02
-1.12966806e-01 4.75908995e-01 -5.24480760e-01 2.72041351e-01
-2.29107767e-01 2.22210269e-02 -2.44810954e-01 5.18377304e-01
-1.98101968e-01 6.80525124e-01 -8.79323006e-01 -2.67161965e-01
-4.66815233e-02 -5.67895509e-02 -7.21000314e-01 9.40665379e-02
-1.18528791e-02 2.05543265e-01 -5.28961599e-01 4.69109297e-01
5.75490892e-01 -5.76361537e-01 5.62416792e-01 1.26038596e-01
2.23222017e-01 1.12201117e-01 -9.39771116e-01 8.35109532e-01
-1.79196447e-01 1.25582561e-01 -6.38534129e-02 -1.21358454e+00
6.51628613e-01 1.88761711e-01 3.94260049e-01 -1.84082150e-01
4.24227685e-01 1.20447204e-01 3.72282863e-01 -4.21964340e-02
-1.17369480e-01 -9.64181364e-01 -4.25047278e-01 3.64411026e-01
-4.01673198e-01 1.76287219e-01 -6.08657420e-01 9.28489268e-02
1.36153901e+00 -7.13840127e-01 8.33561480e-01 -4.87080693e-01
3.49723361e-02 1.22376420e-01 1.13356686e+00 1.31390882e+00
-4.86462265e-01 4.07072335e-01 1.01199365e+00 -3.95093083e-01
-7.61939824e-01 -1.12695384e+00 -6.17967546e-01 4.90783125e-01
8.90338793e-02 -9.89238620e-02 -6.56792939e-01 -6.77504838e-01
4.29028630e-01 1.03320301e+00 -8.95694792e-01 -4.96714920e-01
-1.19767793e-01 -1.37646055e+00 3.82508487e-01 3.70499581e-01
2.81586915e-01 -3.97258997e-01 -8.23792875e-01 2.66469419e-01
-2.14044712e-02 -6.12907469e-01 -4.42869455e-01 2.52539307e-01
-9.59686697e-01 -1.11442268e+00 -6.08449697e-01 2.12254554e-01
8.31303656e-01 -2.51097560e-01 1.15184319e+00 1.25677167e-02
-1.08746774e-01 2.36120909e-01 5.80978729e-02 -6.00395262e-01
-4.18987006e-01 -3.65928143e-01 1.23989910e-01 -1.31233826e-01
1.98714063e-01 -4.77659523e-01 -1.12109327e+00 3.19750756e-01
-3.33769858e-01 -9.01628211e-02 7.41830647e-01 9.42061841e-01
5.59913397e-01 -8.09193403e-02 7.21666515e-01 -1.55667782e+00
6.51795208e-01 -6.14640534e-01 -7.69849718e-01 3.58039081e-01
-1.15512729e+00 1.68926552e-01 6.15752935e-01 -3.62030685e-01
-8.52545500e-01 3.96776460e-02 1.01394475e-01 -1.70601845e-01
3.43658067e-02 4.77338582e-01 -2.47234162e-02 2.50671566e-01
9.25730109e-01 -9.34627578e-02 -1.07160062e-01 -2.53619522e-01
1.89457044e-01 6.59563661e-01 3.39063227e-01 -5.67241549e-01
1.38467252e-01 2.46094301e-01 3.17966580e-01 -2.90332764e-01
-1.11714780e+00 1.91777572e-01 -9.64601561e-02 1.15087079e-02
6.25457227e-01 -8.07893276e-01 -1.37983656e+00 -1.74199104e-01
-6.55112565e-01 -1.99674129e-01 -5.50105870e-01 5.88150799e-01
-7.32307792e-01 1.31995872e-01 -8.96348581e-02 -1.04004443e+00
-3.41201961e-01 -1.03990054e+00 5.12053311e-01 9.73478705e-02
-5.24090469e-01 -8.78620565e-01 -2.17623353e-01 2.75071651e-01
2.73415387e-01 5.67161083e-01 1.24988818e+00 -7.13711500e-01
-3.94312948e-01 -4.34243262e-01 -2.82105189e-02 1.37418285e-01
4.41499501e-02 -3.41662288e-01 -7.01973379e-01 -2.01773703e-01
-4.64659370e-02 4.81066369e-02 5.13511121e-01 8.97182703e-01
1.32792068e+00 -8.89770746e-01 -6.94954872e-01 4.38288450e-01
1.14625835e+00 4.81822670e-01 3.33569258e-01 -4.90413569e-02
-1.06660664e-01 4.47898358e-01 6.15547359e-01 7.50467539e-01
2.75570273e-01 4.81528401e-01 2.60324210e-01 3.22227091e-01
6.15857661e-01 -4.24010038e-01 3.35011818e-02 -1.16434246e-02
1.96125358e-01 -1.69407427e-01 -9.38533127e-01 3.22369218e-01
-1.82904685e+00 -6.73380733e-01 2.81690568e-01 3.03565216e+00
9.56945062e-01 5.57337761e-01 1.11104049e-01 -1.10438339e-01
7.83083618e-01 -3.18095684e-01 -9.96667325e-01 -5.86086631e-01
3.22644673e-02 -2.74718422e-02 9.96613979e-01 5.85497677e-01
-8.11247945e-01 3.65816444e-01 7.29255772e+00 6.45907998e-01
-1.00727654e+00 1.19189128e-01 1.55240273e+00 -2.21734568e-01
-4.13611412e-01 2.07548991e-01 -6.06244266e-01 6.89592659e-01
1.22761023e+00 -6.96082652e-01 7.71803707e-02 1.00380468e+00
1.30119920e-01 -3.57621461e-01 -1.47138143e+00 8.10951948e-01
-4.03033227e-01 -1.56328332e+00 -2.49064028e-01 8.76181424e-02
5.15191734e-01 -4.61059511e-01 6.28888756e-02 1.33331716e-01
9.65436935e-01 -1.20264292e+00 4.48222905e-01 6.68541431e-01
1.14226413e+00 -8.32224250e-01 8.35458815e-01 6.39254630e-01
-2.98290461e-01 -3.44229549e-01 -2.38404915e-01 -3.11328351e-01
7.38197938e-03 1.23583055e+00 -1.21313274e+00 3.55915666e-01
3.89352620e-01 3.80251914e-01 7.28280470e-02 9.82311845e-01
-7.81827196e-02 6.99556887e-01 -3.45003396e-01 4.72674817e-02
-2.49045819e-01 -8.35008547e-02 4.22142982e-01 9.58023846e-01
3.16081583e-01 6.59816682e-01 -1.68566599e-01 7.19753563e-01
-1.39499873e-01 3.36821228e-02 -3.49356264e-01 2.26891320e-02
6.52398050e-01 5.88328660e-01 -3.56427670e-01 -3.94180834e-01
5.84017746e-02 6.18711531e-01 8.32549185e-02 7.24961534e-02
-5.67431927e-01 -1.26987651e-01 9.12078857e-01 3.85601252e-01
-3.04879069e-01 5.64997792e-01 -9.26909089e-01 -1.03530288e+00
-1.82724297e-01 -5.66027701e-01 9.98741090e-01 -5.27001798e-01
-1.37950611e+00 2.30927020e-01 3.23519390e-03 -1.24463642e+00
-3.13961625e-01 -3.36475402e-01 -3.93896401e-01 8.30043435e-01
-9.91480112e-01 -1.60579100e-01 2.87822217e-01 1.56183302e-01
-1.58868045e-01 3.21359754e-01 7.77556598e-01 8.82489327e-03
-6.92020059e-01 1.28765965e+00 1.37718305e-01 3.82351615e-02
4.99632567e-01 -1.03355825e+00 -3.15333635e-01 5.70233643e-01
-6.30257905e-01 6.18469357e-01 8.94582570e-01 -7.12911487e-01
-1.13796532e+00 -1.10389388e+00 6.94770455e-01 -4.74592388e-01
3.41118038e-01 1.75951459e-02 -5.54439545e-01 9.44202483e-01
-5.12018263e-01 -1.04411572e-01 9.58466709e-01 2.75513083e-01
-6.21764064e-02 -4.10727888e-01 -1.77456307e+00 7.18853951e-01
9.27098691e-01 -2.08509620e-02 -4.23243612e-01 5.10345697e-01
7.21572280e-01 -2.55345225e-01 -1.09802639e+00 7.10365295e-01
7.06062913e-01 -7.81327724e-01 9.17930722e-01 -9.87357795e-01
3.43808740e-01 3.10637921e-01 -2.21477255e-01 -1.24808371e+00
-1.55087322e-01 -8.33759844e-01 3.81009340e-01 6.09451532e-01
8.47095490e-01 -1.04313898e+00 9.31199789e-01 1.48044395e+00
3.70336890e-01 -9.70787227e-01 -1.28591716e+00 -5.35301268e-01
4.95066404e-01 -4.15349960e-01 4.70179439e-01 9.80317295e-01
5.11967003e-01 -9.47618857e-02 -4.81505692e-01 2.32610703e-01
8.61184180e-01 2.83821166e-01 2.45486349e-01 -1.20292962e+00
-5.10610223e-01 -2.99061596e-01 -4.30681705e-01 -8.59560788e-01
-1.35903716e-01 -7.57491052e-01 -1.10434242e-01 -8.76686454e-01
5.05200684e-01 -9.91558671e-01 -4.43345726e-01 5.00289500e-01
-4.81646806e-01 -4.54980105e-01 1.40010729e-01 7.67624602e-02
-2.47523114e-01 1.40434697e-01 9.69715416e-01 1.15185469e-01
-1.57666415e-01 5.56540012e-01 -1.44926131e+00 7.98563957e-01
6.30600333e-01 -7.06211746e-01 -2.23908231e-01 1.50612295e-01
5.10963380e-01 1.02372456e+00 1.88565418e-01 -5.86798906e-01
1.23032527e-02 -7.23856390e-01 4.60612059e-01 -5.71527220e-02
1.87750369e-01 -6.50993586e-01 5.30274272e-01 9.62772012e-01
-8.67026627e-01 -3.02386135e-01 -1.57545358e-01 7.30578244e-01
3.79670173e-01 -1.49441302e-01 9.64806974e-01 -1.94119960e-01
2.62034118e-01 1.28462046e-01 -3.51472050e-01 2.12686688e-01
1.17093325e+00 1.34559885e-01 -4.04488027e-01 -7.47995198e-01
-1.05307996e+00 2.40575299e-01 3.53562891e-01 -2.10804820e-01
3.21476221e-01 -8.26078653e-01 -6.02975786e-01 2.01591267e-03
-2.98510920e-02 -1.91343769e-01 2.21398145e-01 8.29753220e-01
-1.29611015e-01 4.34325278e-01 1.91700488e-01 -4.37661290e-01
-7.66051173e-01 4.90062118e-01 7.85199940e-01 -2.71121174e-01
-2.74788678e-01 8.81897926e-01 3.78817916e-01 -9.96668637e-02
1.10992655e-01 -3.91278833e-01 4.00537133e-01 -3.16606551e-01
3.77396107e-01 4.23244745e-01 -1.85273513e-02 -8.13859552e-02
-3.93041790e-01 1.53109118e-01 -1.36294141e-01 1.71382725e-02
9.44988072e-01 -1.71575025e-01 1.15482964e-01 1.44890159e-01
7.43200123e-01 -1.43760517e-01 -1.16646802e+00 -1.52731687e-01
-8.52680057e-02 -6.80430472e-01 -5.60811237e-02 -1.33991873e+00
-7.72519827e-01 6.77363813e-01 6.45387769e-01 -2.93444023e-02
1.15683889e+00 -1.16316840e-01 1.78847566e-01 2.50387788e-01
5.97788811e-01 -6.80157900e-01 -7.01242149e-01 -2.45949537e-01
6.12238169e-01 -1.17108834e+00 2.72053182e-01 -5.18914104e-01
-8.86651039e-01 5.55854917e-01 3.78231496e-01 -4.46432494e-02
1.03549278e+00 1.25558808e-01 -1.88628640e-02 1.47132590e-01
-9.70349491e-01 2.71730185e-01 -5.61632775e-02 2.04882041e-01
2.25653678e-01 9.45052743e-01 -7.92582929e-01 1.38147151e+00
-4.36621398e-01 1.76312417e-01 7.10332751e-01 7.39877939e-01
-4.42413330e-01 -9.75972056e-01 -3.84367168e-01 1.37684071e+00
-6.32361889e-01 -7.27322400e-02 -2.18294710e-01 3.76644373e-01
1.55953526e-01 9.25893188e-01 -3.91424969e-02 -3.50090146e-01
3.02519083e-01 9.55983028e-02 1.27803773e-01 -4.12630677e-01
-4.49153841e-01 -2.85090089e-01 1.58135951e-01 -5.71533978e-01
1.36711866e-01 -6.30628169e-01 -1.06100571e+00 -5.10264277e-01
-1.65941522e-01 3.88872236e-01 2.06151649e-01 1.04577172e+00
6.66836262e-01 1.89223498e-01 6.12569153e-01 -2.67823767e-02
-9.88158166e-01 -4.27873701e-01 -5.40342689e-01 2.61723489e-01
3.74212563e-01 -6.01705730e-01 -5.51920176e-01 -3.60106975e-01]
|
[8.174915313720703, 5.270421981811523]
|
896f2293-d803-4631-9e0f-b409857707dd
|
oscar-data-driven-operational-space-control
|
2110.00704
| null |
https://arxiv.org/abs/2110.00704v1
|
https://arxiv.org/pdf/2110.00704v1.pdf
|
OSCAR: Data-Driven Operational Space Control for Adaptive and Robust Robot Manipulation
|
Learning performant robot manipulation policies can be challenging due to high-dimensional continuous actions and complex physics-based dynamics. This can be alleviated through intelligent choice of action space. Operational Space Control (OSC) has been used as an effective task-space controller for manipulation. Nonetheless, its strength depends on the underlying modeling fidelity, and is prone to failure when there are modeling errors. In this work, we propose OSC for Adaptation and Robustness (OSCAR), a data-driven variant of OSC that compensates for modeling errors by inferring relevant dynamics parameters from online trajectories. OSCAR decomposes dynamics learning into task-agnostic and task-specific phases, decoupling the dynamics dependencies of the robot and the extrinsics due to its environment. This structure enables robust zero-shot performance under out-of-distribution and rapid adaptation to significant domain shifts through additional finetuning. We evaluate our method on a variety of simulated manipulation problems, and find substantial improvements over an array of controller baselines. For more results and information, please visit https://cremebrule.github.io/oscar-web/.
|
['Yuke Zhu', 'Anima Anandkumar', 'Viktor Makoviychuk', 'Josiah Wong']
|
2021-10-02
| null | null | null | null |
['robot-manipulation']
|
['robots']
|
[ 1.09156653e-01 -1.34211257e-01 -5.04549146e-01 1.37712270e-01
-7.72154570e-01 -8.08362007e-01 5.96804261e-01 -1.82405099e-01
-3.29781443e-01 7.55619228e-01 1.04365088e-01 -1.62176624e-01
-5.20169377e-01 -1.23580799e-01 -9.15339172e-01 -5.94525456e-01
-2.87763327e-01 6.92911088e-01 4.88172054e-01 -6.81666374e-01
2.91686356e-01 5.19936383e-01 -1.38315523e+00 -1.80360839e-01
9.01690781e-01 4.87769783e-01 5.13042212e-01 7.48004556e-01
4.35570866e-01 5.92208385e-01 -3.60439956e-01 3.59449148e-01
4.35386211e-01 -1.92809775e-01 -4.85110998e-01 -1.41512632e-01
1.17468849e-01 -2.40569875e-01 -6.22970223e-01 9.18414533e-01
4.91323918e-01 6.18102372e-01 7.85990596e-01 -1.40650725e+00
-1.80026069e-01 4.15051103e-01 -4.94978987e-02 3.12226103e-03
1.14937179e-01 1.07228172e+00 5.21768212e-01 -5.64022362e-01
6.43597722e-01 1.58134663e+00 5.20842314e-01 5.97929060e-01
-1.47382200e+00 -7.22028315e-01 4.54107165e-01 -2.53114589e-02
-1.09570122e+00 -5.64893186e-01 5.00161231e-01 -8.01165998e-01
9.72908676e-01 -2.42450133e-01 5.23506403e-01 1.59316099e+00
6.11898959e-01 7.18712449e-01 6.81445420e-01 -2.41326001e-02
4.81224626e-01 -3.11790973e-01 -1.82248339e-01 6.18468285e-01
1.58503577e-01 6.69639409e-01 -3.43478680e-01 -6.96804896e-02
1.01651347e+00 -5.46944104e-02 -2.41891906e-01 -1.00281477e+00
-1.28817177e+00 5.81014991e-01 4.76131588e-01 -1.63767874e-01
-2.73567975e-01 6.40527785e-01 4.45039034e-01 5.51710427e-01
-1.19306594e-01 1.05456173e+00 -8.24078083e-01 -5.55070162e-01
-2.40426704e-01 7.99662709e-01 8.00548255e-01 1.38181460e+00
3.20631027e-01 3.30467612e-01 -2.63974369e-01 6.26915038e-01
-8.78586918e-02 5.31390786e-01 4.16836202e-01 -1.50722456e+00
5.60225308e-01 3.31643194e-01 5.96500337e-01 -7.08341241e-01
-6.60592616e-01 -3.47334087e-01 -3.78024012e-01 5.61515391e-01
5.56463540e-01 -3.35958987e-01 -1.08552074e+00 1.91590345e+00
4.22547609e-01 -5.26386872e-02 -2.08625704e-01 1.00579464e+00
-1.50526494e-01 3.13998640e-01 -7.05296025e-02 1.14534967e-01
8.83167803e-01 -9.41773415e-01 -7.92373300e-01 -5.10087967e-01
5.41513443e-01 -4.83809769e-01 1.46356058e+00 4.76738483e-01
-1.02594674e+00 -4.59151000e-01 -1.05575359e+00 1.06978603e-01
-2.17460334e-01 5.58773205e-02 3.27938497e-01 -1.16639800e-01
-6.90429628e-01 9.36726213e-01 -1.47126341e+00 -3.92633975e-01
2.47808963e-01 5.86130798e-01 -1.49225846e-01 7.00168610e-02
-1.01320624e+00 1.28431416e+00 5.12013972e-01 -2.16812611e-01
-1.45209980e+00 -8.80247831e-01 -5.55062950e-01 -2.73796886e-01
1.08221674e+00 -7.60499418e-01 1.71715784e+00 -4.88508880e-01
-2.05947351e+00 2.94535831e-02 2.51029164e-01 -3.55058789e-01
6.89347327e-01 -6.74721777e-01 -2.83688828e-02 -4.32872772e-02
-5.78775853e-02 5.63572526e-01 1.22905397e+00 -1.13781202e+00
-4.49107409e-01 -1.69429958e-01 9.95193049e-02 2.21042782e-01
-6.25320077e-02 -2.93409824e-01 -5.88702738e-01 -7.01026320e-01
-1.24557294e-01 -1.66761756e+00 -3.68922651e-01 1.95324734e-01
-1.33635148e-01 2.66869478e-02 8.27324212e-01 -4.91724938e-01
1.05701578e+00 -2.12474275e+00 8.35729182e-01 -1.87945411e-01
-3.85189392e-02 2.30978727e-01 -3.12271506e-01 7.84313381e-01
1.38457209e-01 -1.82598561e-01 -2.72475332e-01 6.38499483e-02
2.52647310e-01 2.01194063e-01 -4.39322531e-01 4.40633178e-01
4.31538522e-01 7.85052717e-01 -1.13636053e+00 -4.67968322e-02
2.96174824e-01 2.99954534e-01 -7.71910429e-01 2.36799955e-01
-7.71366358e-01 9.69751894e-01 -7.10258245e-01 5.52891195e-01
5.29289283e-02 -2.48865802e-02 1.86838701e-01 -8.31402093e-03
-1.09532505e-01 1.90389782e-01 -1.15847194e+00 1.82857859e+00
-5.17496169e-01 2.86462575e-01 5.41564584e-01 -8.66759121e-01
7.32591927e-01 6.97119534e-02 6.34067357e-01 -4.60177660e-01
2.36696407e-01 3.13350707e-01 2.28418007e-01 -6.82306886e-01
4.74029869e-01 1.12715632e-01 -3.13028783e-01 -6.50901273e-02
1.47727251e-01 -7.94669151e-01 1.34178132e-01 -2.52128579e-02
1.32717586e+00 7.88448811e-01 2.67432004e-01 -3.11412364e-01
1.35504425e-01 5.09511888e-01 6.50210559e-01 7.66988695e-01
-4.84231740e-01 1.60125405e-01 4.59977865e-01 -3.14555019e-02
-1.06712699e+00 -1.03366566e+00 1.08877040e-01 1.11659646e+00
2.31241211e-01 -2.80039877e-01 -5.76496780e-01 -2.88910449e-01
5.64861059e-01 8.57453704e-01 -4.63722914e-01 -7.63237298e-01
-8.53265226e-01 -2.38899231e-01 2.04030335e-01 5.54680169e-01
5.17864302e-02 -9.34429407e-01 -7.23297358e-01 4.42299932e-01
-8.48014257e-04 -9.98672903e-01 -5.99821091e-01 3.65317523e-01
-9.04584408e-01 -1.16757131e+00 -4.14231151e-01 -4.74238753e-01
3.16957772e-01 1.63203478e-01 6.92945361e-01 -2.60580182e-01
-4.26968843e-01 6.75413966e-01 -1.93280011e-01 -4.05033916e-01
-5.52910745e-01 2.50854224e-01 6.52302086e-01 -5.11954427e-01
-2.37089604e-01 -5.55897057e-01 -4.36534226e-01 6.31060481e-01
-5.33627033e-01 -1.66786194e-01 5.49287379e-01 1.03131878e+00
5.53210974e-01 3.00813671e-02 4.68689740e-01 -5.02966344e-01
7.34997809e-01 -4.38701987e-01 -9.97746646e-01 -1.83415785e-01
-4.45181161e-01 3.22432309e-01 6.39052629e-01 -8.88485730e-01
-9.68317568e-01 3.27927500e-01 4.55564260e-01 -8.15124631e-01
5.85546670e-03 3.28569114e-01 1.24114193e-01 -7.70662576e-02
8.49618316e-01 -5.14958575e-02 4.31673139e-01 -4.20208186e-01
4.50923562e-01 3.53215307e-01 4.92445648e-01 -1.03573525e+00
9.76521313e-01 1.34728074e-01 4.92524058e-02 -6.32149220e-01
-6.95072770e-01 -3.58202845e-01 -7.01660216e-01 -2.89442301e-01
6.65470302e-01 -9.58462536e-01 -7.55862296e-01 4.59061235e-01
-7.92407572e-01 -1.34583795e+00 -2.15509802e-01 6.36147439e-01
-1.12920260e+00 -1.69496298e-01 -4.02959019e-01 -8.96062434e-01
2.45254606e-01 -1.36654294e+00 9.99728799e-01 -3.15822475e-03
-2.35603139e-01 -6.48509860e-01 1.14793018e-01 5.15248580e-03
5.64058423e-01 4.21512991e-01 7.49568641e-01 -2.67957956e-01
-6.47529125e-01 -3.32836837e-01 2.77306229e-01 1.55635729e-01
8.67016688e-02 -2.83405557e-02 -4.91035193e-01 -8.06878448e-01
-1.36601001e-01 -5.78510404e-01 6.05127752e-01 3.72833014e-01
1.02495825e+00 -2.61883587e-01 -5.68545699e-01 4.47295219e-01
1.16594887e+00 1.70770884e-01 9.73450616e-02 5.46482384e-01
7.08388686e-01 6.35571063e-01 1.14465630e+00 4.80085224e-01
2.35930517e-01 9.47452784e-01 5.83792925e-01 4.58481282e-01
-2.08690967e-02 -3.61831903e-01 7.68612802e-01 5.45143366e-01
1.15480542e-01 -1.96762189e-01 -1.05713809e+00 5.24889588e-01
-2.13698769e+00 -8.03674161e-01 6.79934770e-02 2.26151156e+00
7.93285370e-01 3.17196399e-01 4.21209902e-01 -4.24486160e-01
5.43714225e-01 -1.74577892e-01 -1.31372023e+00 5.86991794e-02
3.54619324e-01 -1.08102955e-01 7.60754228e-01 5.84337234e-01
-1.08613563e+00 1.21224654e+00 5.77848577e+00 7.55394101e-01
-1.00461483e+00 -1.25000834e-01 -2.04574957e-01 -4.61729497e-01
3.51835728e-01 -7.61249801e-03 -7.41431117e-01 5.35713136e-01
8.98022652e-01 -1.64053783e-01 8.68766308e-01 9.16399717e-01
5.81833839e-01 -7.29933828e-02 -1.20812464e+00 5.15506148e-01
-4.01997209e-01 -9.66365099e-01 -4.51270103e-01 -5.32552525e-02
7.44255304e-01 3.69142532e-01 2.57291377e-01 8.12861681e-01
7.69880533e-01 -6.58445358e-01 9.59232092e-01 4.95067120e-01
6.17781043e-01 -4.80588496e-01 1.26628548e-01 5.94363689e-01
-1.03059769e+00 -7.86948144e-01 -7.65147135e-02 -5.60186729e-02
2.09701598e-01 -2.37386562e-02 -6.59794688e-01 2.49935687e-01
5.53929806e-01 7.97768533e-01 -2.02825755e-01 9.38703716e-01
-2.09251076e-01 4.56584334e-01 -4.44749445e-01 -6.20750785e-02
1.96220458e-01 -1.11796133e-01 1.01967013e+00 8.18823576e-01
9.27564427e-02 6.82610087e-04 7.27877140e-01 8.34794819e-01
4.06237274e-01 -5.86047947e-01 -7.21445799e-01 -2.93820173e-01
7.00380921e-01 8.42244983e-01 -2.91849256e-01 -8.53873491e-02
1.57083571e-01 7.28456080e-01 4.34230119e-01 5.46090484e-01
-9.28038180e-01 -4.36046809e-01 1.11530006e+00 8.60144123e-02
3.60596180e-01 -1.02916479e+00 3.54648684e-03 -1.04469621e+00
-7.33147636e-02 -1.18349195e+00 4.02015410e-02 -4.88120139e-01
-9.28701997e-01 -7.92696700e-02 3.56377542e-01 -1.58542395e+00
-3.84017199e-01 -6.76416814e-01 -1.69370666e-01 4.77214724e-01
-1.35127854e+00 -6.66965544e-01 -1.68635920e-01 3.70795041e-01
7.27031052e-01 -1.33764088e-01 5.92469454e-01 3.39745544e-02
-7.01705456e-01 2.70126164e-01 3.68500769e-01 -3.93439382e-01
1.07603633e+00 -1.15433717e+00 3.56546819e-01 4.99552697e-01
-6.73032820e-01 6.49625957e-01 1.18283391e+00 -8.26327682e-01
-2.03411484e+00 -1.20751476e+00 6.87151849e-02 -8.54372621e-01
1.14218390e+00 -3.92916858e-01 -9.09688354e-01 7.63586998e-01
-2.65426725e-01 -1.62280470e-01 -1.50229707e-01 -1.41014934e-01
-9.09694880e-02 6.26880676e-02 -7.90677547e-01 1.03780437e+00
1.36713576e+00 -3.06646079e-01 -4.74833101e-01 5.09710848e-01
1.10513759e+00 -7.86077619e-01 -9.88970816e-01 4.85736877e-01
5.80501199e-01 -4.33559358e-01 9.31517065e-01 -7.30463862e-01
2.71858841e-01 -4.88319665e-01 -1.96106900e-02 -1.62534738e+00
-5.38568258e-01 -1.06716597e+00 -4.10074890e-01 6.30167067e-01
2.66851932e-01 -5.79157829e-01 3.64355236e-01 4.81611580e-01
-4.90298063e-01 -5.95105946e-01 -5.55183053e-01 -1.47858953e+00
2.24066794e-01 -3.33705842e-01 3.73330891e-01 7.40509987e-01
1.33058369e-01 2.62262255e-01 -4.06522393e-01 3.12564313e-01
4.61193621e-01 -2.30990753e-01 1.24540460e+00 -8.47635508e-01
-5.35566926e-01 -6.32207394e-01 -6.21110294e-03 -1.03983998e+00
2.83220738e-01 -5.60686231e-01 6.72482967e-01 -1.18632805e+00
-3.32125336e-01 -5.43554664e-01 -1.28802344e-01 4.35206085e-01
-2.19285131e-01 -5.58975041e-01 4.41247791e-01 3.51359844e-01
-5.59556127e-01 1.03107369e+00 1.51737881e+00 -3.52037959e-02
-6.56828582e-01 5.07888012e-02 -1.24713883e-01 5.35077214e-01
1.18984175e+00 -4.36585307e-01 -6.95817411e-01 -5.61532676e-01
-1.40758321e-01 3.55333716e-01 2.77458221e-01 -1.29975486e+00
1.42075881e-01 -6.84216619e-01 -1.74504444e-01 -2.22219080e-01
5.34624875e-01 -7.62387872e-01 1.54670030e-01 9.49884117e-01
-6.16723061e-01 1.60974413e-01 5.23481131e-01 1.13982356e+00
1.61532879e-01 2.09712952e-01 9.08766568e-01 -6.68505877e-02
-6.96051776e-01 3.35378945e-01 -5.50483882e-01 3.14244092e-01
1.05076814e+00 -3.53244394e-02 -3.26639771e-01 -1.98842406e-01
-6.51407063e-01 7.31020391e-01 5.93138993e-01 8.71593893e-01
2.13345021e-01 -1.10036969e+00 -3.77567619e-01 -5.94267771e-02
2.41648942e-01 3.10398024e-02 1.28584325e-01 1.07167637e+00
-1.90953180e-01 3.84721994e-01 -3.09254289e-01 -6.13793254e-01
-7.88508177e-01 5.73830962e-01 3.99134785e-01 -4.12378050e-02
-8.48459363e-01 6.58667922e-01 1.52473152e-01 -7.66934514e-01
4.81999725e-01 -4.82116431e-01 1.38468906e-01 -4.20493245e-01
8.12225193e-02 4.65668052e-01 -2.40401149e-01 -1.92673028e-01
-2.83071756e-01 4.99332339e-01 1.72090624e-02 -1.96247846e-01
1.10852695e+00 5.06485533e-03 4.51523155e-01 6.09845817e-01
6.99681282e-01 -6.62930965e-01 -2.19041777e+00 -3.88927534e-02
1.30757973e-01 -3.58886659e-01 1.11960746e-01 -8.76538277e-01
-5.52873731e-01 6.55324638e-01 4.76825625e-01 -2.71332026e-01
5.91311932e-01 -3.54304761e-01 6.26006126e-01 7.22536862e-01
7.08727598e-01 -1.55684507e+00 5.94301164e-01 1.04552698e+00
1.27988374e+00 -1.25236797e+00 -3.40247750e-02 -1.22840330e-01
-8.09429228e-01 9.30193961e-01 1.03304720e+00 -5.39615691e-01
6.33046031e-01 2.60075778e-01 -2.62637228e-01 -6.94623366e-02
-1.07702780e+00 -1.21416718e-01 7.42634237e-02 7.56015539e-01
-1.13301404e-01 1.20311789e-01 1.45964175e-01 2.35751450e-01
2.08613183e-02 -1.26257669e-02 3.45159620e-01 1.37319958e+00
-6.79781199e-01 -9.24679339e-01 -2.49292910e-01 2.28641495e-01
2.14677602e-01 4.73080486e-01 -2.60104805e-01 1.05170000e+00
-3.96564811e-01 6.81046426e-01 -2.14878887e-01 -3.64623249e-01
7.35862255e-01 -1.10325098e-01 6.25788748e-01 -7.70201027e-01
-3.73799860e-01 1.20616853e-01 1.73981577e-01 -1.16034937e+00
9.75444242e-02 -8.17608058e-01 -1.46934998e+00 -1.72841832e-01
-3.15646887e-01 -3.89455199e-01 6.25623524e-01 7.34950721e-01
7.53504992e-01 8.86059999e-01 4.57711220e-01 -1.32862771e+00
-1.46500814e+00 -8.76821339e-01 -1.75401345e-01 2.22887948e-01
8.00617576e-01 -1.43162835e+00 -4.28333104e-01 5.99866500e-03]
|
[4.48192834854126, 1.3637778759002686]
|
d583783c-5b29-45cf-9bc2-213195292b08
|
from-rewriting-to-remembering-common-ground-1
|
2204.0393
| null |
https://arxiv.org/abs/2204.03930v1
|
https://arxiv.org/pdf/2204.03930v1.pdf
|
From Rewriting to Remembering: Common Ground for Conversational QA Models
|
In conversational QA, models have to leverage information in previous turns to answer upcoming questions. Current approaches, such as Question Rewriting, struggle to extract relevant information as the conversation unwinds. We introduce the Common Ground (CG), an approach to accumulate conversational information as it emerges and select the relevant information at every turn. We show that CG offers a more efficient and human-like way to exploit conversational information compared to existing approaches, leading to improvements on Open Domain Conversational QA.
|
['Adrià De Gispert', 'Bill Byrne', 'Gianni Barlacchi', 'Xiaoyu Shen', 'Marco del Tredici']
|
2022-04-08
| null |
https://aclanthology.org/2022.nlp4convai-1.7
|
https://aclanthology.org/2022.nlp4convai-1.7.pdf
|
nlp4convai-acl-2022-5
|
['question-rewriting']
|
['natural-language-processing']
|
[ 3.87847155e-01 9.42674756e-01 1.95892411e-04 -4.53297883e-01
-1.52479613e+00 -1.04283583e+00 9.45136428e-01 3.49312156e-01
-6.21986389e-02 1.19464827e+00 9.50913548e-01 -6.62162185e-01
-8.56079385e-02 -7.37433732e-01 -3.19454908e-01 1.10044694e-02
1.02569714e-01 1.03004074e+00 4.54340249e-01 -1.20003784e+00
3.66509557e-01 -1.91050917e-01 -1.20066762e+00 8.65198255e-01
1.05712759e+00 2.87948757e-01 1.81072399e-01 1.00586271e+00
-7.67117441e-01 1.22808218e+00 -8.76221776e-01 -9.44443822e-01
-2.50561517e-02 -9.67251182e-01 -2.08861113e+00 -2.71051288e-01
3.16621810e-01 -1.57708853e-01 -1.33813217e-01 7.93555439e-01
2.52404332e-01 3.62964600e-01 8.68927091e-02 -1.03705466e+00
-1.86380342e-01 1.08329129e+00 3.22963268e-01 5.90787828e-01
1.28572083e+00 2.72132009e-01 1.43746829e+00 -6.00574613e-01
1.18933618e+00 1.52680731e+00 4.99838442e-01 9.27580595e-01
-7.86373496e-01 1.95113942e-01 2.00380445e-01 5.96863866e-01
-4.68952656e-01 -5.13815045e-01 6.90838337e-01 -1.37937218e-01
1.25379288e+00 6.55525208e-01 6.10680163e-01 1.03412926e+00
1.83464661e-01 8.69158149e-01 1.08196211e+00 -8.13089013e-01
-3.70083898e-02 -9.63951573e-02 7.77169347e-01 4.11028534e-01
-3.62708896e-01 -2.63161898e-01 -6.50121510e-01 -5.41857898e-01
-1.14794023e-01 -6.45978451e-01 -3.72183740e-01 2.53193021e-01
-1.04559565e+00 9.29180980e-01 5.16634770e-02 3.29671443e-01
-5.40944040e-01 -3.12893659e-01 3.25305641e-01 1.00905049e+00
2.88060874e-01 1.16316509e+00 -7.09683716e-01 -7.16967404e-01
-2.66849488e-01 7.65179157e-01 1.71783280e+00 7.03791440e-01
1.02539980e+00 -7.81963110e-01 -5.17727554e-01 8.42097998e-01
2.93035321e-02 1.57172650e-01 3.32586259e-01 -1.62258554e+00
8.71777833e-01 8.42340052e-01 3.94917876e-01 -7.38375068e-01
-1.60563260e-01 -1.28061818e-02 2.52810884e-02 -5.09643972e-01
4.68042254e-01 -4.36430812e-01 -4.96726394e-01 1.70407629e+00
4.70434785e-01 -3.26069862e-01 6.96615338e-01 4.77617562e-01
9.80212808e-01 6.92126334e-01 -4.29262128e-03 -4.10883188e-01
1.35838020e+00 -1.33466685e+00 -9.82424140e-01 -6.37198806e-01
7.55652726e-01 -9.84823346e-01 1.06156063e+00 1.60796627e-01
-1.35271788e+00 -1.48250544e-02 -6.76755548e-01 -2.80712217e-01
-1.31674200e-01 -6.49659157e-01 5.61655879e-01 4.24055874e-01
-1.14535117e+00 4.69988734e-01 -2.71742851e-01 -4.31439042e-01
-1.08314931e-01 5.96440732e-02 -8.66501257e-02 -2.24600345e-01
-1.73317683e+00 1.31616724e+00 1.42666087e-01 3.36659476e-02
-6.28173649e-01 -5.08900523e-01 -9.60326254e-01 -1.53867185e-01
7.85552919e-01 -1.04256427e+00 2.20229530e+00 -8.40107203e-01
-1.91254997e+00 8.45539153e-01 -6.26286089e-01 -6.92151368e-01
1.55662864e-01 -3.76736104e-01 -1.42624915e-01 3.29555631e-01
2.40729317e-01 4.34187710e-01 3.39962602e-01 -1.10708034e+00
-7.17204869e-01 -1.78524539e-01 9.00614560e-01 6.59124494e-01
3.13478857e-01 1.20618969e-01 -1.09770738e-01 1.48529103e-02
7.60550126e-02 -8.93800855e-01 -4.86875206e-01 -8.13327551e-01
-2.86931723e-01 -7.90444493e-01 5.61826587e-01 -8.30068111e-01
1.28157222e+00 -1.10431433e+00 4.28995252e-01 -1.37325704e-01
3.45791847e-01 4.01721120e-01 -2.68492073e-01 1.08774984e+00
3.26566517e-01 7.25781173e-03 -2.36134529e-01 -1.37282601e-02
-1.01838522e-02 5.89626610e-01 -4.92897600e-01 -4.21898723e-01
2.66025841e-01 1.26742375e+00 -1.33513117e+00 -4.20641512e-01
-4.85647582e-02 -1.84099004e-01 -5.92065036e-01 5.12256265e-01
-9.93613720e-01 4.76428866e-01 -6.14586055e-01 2.34328941e-01
2.55445480e-01 -3.52490813e-01 4.58763003e-01 3.44522446e-01
2.32689247e-01 1.46421409e+00 -5.25155783e-01 1.74021327e+00
-5.80056310e-01 7.45002806e-01 2.97920197e-01 -5.62061846e-01
7.72269666e-01 4.57334220e-01 -1.70535892e-01 -7.18911171e-01
-8.30155537e-02 1.77092701e-01 -2.20466219e-02 -8.80033672e-01
7.22951412e-01 -5.04026785e-02 -2.49915779e-01 7.44452477e-01
9.58294794e-02 -4.63296413e-01 4.42472041e-01 7.69641042e-01
1.32877636e+00 -3.97295170e-02 4.26033109e-01 -9.56901088e-02
8.74048173e-01 6.60959005e-01 2.75795072e-01 8.59651327e-01
-2.01971576e-01 7.48221800e-02 6.83836401e-01 -3.46044421e-01
-6.23758078e-01 -6.61494374e-01 5.39583862e-01 1.30518973e+00
1.68386653e-01 -7.89263427e-01 -9.95246708e-01 -9.86776829e-01
-5.75364649e-01 9.98253465e-01 -2.23184004e-01 -1.59053095e-02
-1.15502822e+00 -1.42597660e-01 4.02139843e-01 -8.83329064e-02
2.85464138e-01 -1.39749503e+00 -4.36504483e-01 5.61564445e-01
-1.30505848e+00 -1.01795864e+00 -2.35620275e-01 1.84504746e-03
-7.15855777e-01 -1.29180253e+00 -3.34473640e-01 -5.87576866e-01
1.59683496e-01 2.32550487e-01 1.81249118e+00 3.03927571e-01
3.15792620e-01 7.02280104e-01 -7.04945087e-01 -2.00029612e-01
-1.01863396e+00 5.21233439e-01 -7.24370182e-01 -4.72159266e-01
5.65647364e-01 -5.18742204e-01 -4.87658888e-01 4.63345461e-02
-3.91512483e-01 -5.82006387e-02 1.34513840e-01 8.85513246e-01
-2.14593232e-01 -8.08075249e-01 9.40443516e-01 -1.23579240e+00
1.53766978e+00 -6.11167014e-01 -3.52566093e-02 5.87675214e-01
-4.48037416e-01 3.50973576e-01 4.50805575e-01 1.91737786e-01
-1.44230139e+00 -5.24376392e-01 -5.73952496e-01 5.88960290e-01
-1.24497406e-01 7.27185369e-01 9.66390446e-02 1.14171870e-01
1.05808306e+00 -8.33662972e-02 8.59212037e-03 -3.69505554e-01
7.22135425e-01 6.07784867e-01 4.00408357e-01 -9.37050343e-01
2.92597741e-01 3.43782417e-02 -4.57258493e-01 -5.44628978e-01
-1.35361969e+00 -7.68692911e-01 -3.96236420e-01 -3.54493409e-01
5.08676291e-01 -5.18668771e-01 -8.08691561e-01 -2.74344794e-02
-1.77113914e+00 -3.82613689e-01 -4.02000725e-01 5.72162606e-02
-6.08509064e-01 6.58946157e-01 -7.15729654e-01 -8.77380610e-01
-6.55220509e-01 -7.63322473e-01 7.48344064e-01 4.72266674e-01
-9.14365053e-01 -1.07993841e+00 5.64801216e-01 9.85179067e-01
4.08062011e-01 9.38297994e-03 9.64293242e-01 -1.05670261e+00
-5.84429204e-01 3.33075523e-02 1.97275981e-01 -2.18213908e-02
-8.91182944e-03 -3.69825482e-01 -6.91503942e-01 1.51111111e-01
2.24202231e-01 -5.51475525e-01 5.15452325e-01 -2.70774126e-01
5.17432727e-02 -8.70798171e-01 -2.51042843e-01 -4.30806190e-01
7.78376460e-01 2.07786053e-01 7.65080273e-01 3.59669417e-01
1.21735401e-01 1.07622373e+00 7.43799329e-01 -1.43738240e-01
9.76617575e-01 4.87853736e-01 1.21957235e-01 4.39519793e-01
-2.76543945e-01 -3.22027802e-01 3.87036264e-01 1.18706727e+00
1.09284610e-01 -2.38764092e-01 -9.71574247e-01 8.47576499e-01
-2.01466942e+00 -1.06812596e+00 -3.54067177e-01 1.58563137e+00
1.33724487e+00 1.15436144e-01 -1.72202796e-01 -2.10153878e-01
5.54789722e-01 2.25151256e-01 -1.91813082e-01 -8.29082727e-01
-2.05785483e-01 4.19289201e-01 -3.29853147e-01 1.34484375e+00
-4.79037613e-01 1.17108822e+00 7.64979172e+00 3.39878500e-01
-4.61576879e-01 4.15801555e-01 3.20578486e-01 3.96419972e-01
-8.03800523e-01 6.37852848e-01 -7.46251702e-01 3.47995609e-02
1.21736133e+00 -3.61257672e-01 3.20698649e-01 5.88848948e-01
-9.02266279e-02 -4.48330104e-01 -1.03887987e+00 4.08920854e-01
1.98581740e-01 -1.59646451e+00 2.29421556e-01 -5.80994427e-01
7.01471567e-01 -4.35582176e-02 -7.12537229e-01 6.34721220e-01
9.30182755e-01 -7.34670162e-01 9.71420333e-02 5.25191128e-01
-1.83730587e-01 -5.30833662e-01 8.85957479e-01 7.76908100e-01
-7.68464506e-01 -1.98813677e-01 -6.13469593e-02 -4.86755282e-01
6.36607170e-01 1.60232842e-01 -1.33480442e+00 9.80419815e-01
2.39786521e-01 7.64132291e-02 -4.07981575e-01 6.81110144e-01
-6.07779801e-01 6.51482999e-01 -1.53510705e-01 -4.56878930e-01
5.36895454e-01 -2.44069308e-01 9.45207655e-01 1.03962123e+00
-9.58535001e-02 4.45822477e-01 2.97147006e-01 2.18433738e-01
-5.38779385e-02 6.44137412e-02 -5.44253111e-01 1.92279071e-01
4.81774718e-01 1.04090154e+00 -7.61903301e-02 -5.92725754e-01
-2.50526309e-01 9.84610856e-01 5.45793712e-01 2.64314115e-01
-1.32973149e-01 -3.34433794e-01 3.96203518e-01 -9.46822092e-02
-1.51738256e-01 -1.12250194e-01 1.79859728e-01 -9.99248147e-01
2.61636019e-01 -1.64670038e+00 9.72581923e-01 -8.92393768e-01
-1.21243489e+00 1.00113332e+00 4.58761863e-02 -6.22931182e-01
-1.08519876e+00 -1.53637230e-01 -8.35206747e-01 7.52308786e-01
-1.52573884e+00 -9.61551428e-01 -5.45858294e-02 4.40282404e-01
9.64276195e-01 2.40151852e-01 1.02112877e+00 -7.30065331e-02
1.20955363e-01 1.48532435e-01 -5.36081553e-01 -7.70051405e-02
7.01991260e-01 -1.25455475e+00 6.18437529e-01 6.89867198e-01
2.58522153e-01 8.90132427e-01 1.03509104e+00 -6.02226496e-01
-1.35796571e+00 -4.62813884e-01 1.74578798e+00 -1.10777617e+00
6.69127882e-01 -1.36709139e-02 -1.03472674e+00 8.82767022e-01
1.01010060e+00 -9.45260525e-01 6.99267983e-01 4.81702209e-01
-3.39428037e-01 6.72245249e-02 -8.71540427e-01 7.77315021e-01
9.28574383e-01 -7.74238110e-01 -1.79284012e+00 6.13402367e-01
1.18133199e+00 -6.76528156e-01 -5.66146851e-01 2.35733762e-01
3.02508697e-02 -8.77809405e-01 7.80664921e-01 -9.17470455e-01
2.82417387e-01 4.90673771e-03 4.97207828e-02 -1.50253475e+00
3.43927503e-01 -1.73439455e+00 -3.14685345e-01 9.96031940e-01
9.28018928e-01 -7.38099933e-01 6.89581335e-01 8.56254756e-01
-3.21362972e-01 -6.29087389e-01 -9.58615005e-01 -2.38891587e-01
1.50497645e-01 -1.40527666e-01 6.61334634e-01 8.21343660e-01
8.58796656e-01 1.13153505e+00 -2.36471668e-01 -1.40058458e-01
1.29408300e-01 4.08467978e-01 9.02696669e-01 -1.20512354e+00
-1.35684118e-01 -1.92325398e-01 2.41265967e-01 -1.56579423e+00
2.07776830e-01 -7.05903471e-01 3.13059628e-01 -1.97623146e+00
-1.70899510e-01 -1.20397173e-01 3.49860251e-01 1.97549820e-01
-5.83945334e-01 -2.01415792e-01 2.45602071e-01 -1.31680006e-02
-1.04730833e+00 3.62588018e-01 1.62139261e+00 -1.73273832e-01
-4.14338529e-01 1.75558597e-01 -1.03276455e+00 6.60026848e-01
7.57104039e-01 -4.15154576e-01 -4.74236220e-01 -3.11808109e-01
7.08162725e-01 6.94575369e-01 6.85500726e-02 -5.76290309e-01
6.68074608e-01 -1.63616344e-01 -6.14141047e-01 -6.98310256e-01
5.35435379e-01 -1.01324774e-01 -2.28339180e-01 2.70522326e-01
-8.24668527e-01 2.93119252e-01 1.52467517e-02 3.90525579e-01
-6.60380781e-01 -5.95560491e-01 1.55632228e-01 -5.82211018e-01
-6.36585236e-01 -1.84057549e-01 -7.27473080e-01 6.48894668e-01
5.88561416e-01 8.84442255e-02 -7.24877000e-01 -1.02975595e+00
-9.46218312e-01 7.00862527e-01 -7.96092972e-02 4.80130166e-01
3.15237761e-01 -8.30168307e-01 -8.66896749e-01 -5.06840587e-01
2.12300662e-02 3.64540843e-04 2.30838358e-01 4.90560204e-01
-2.91657001e-01 7.21680284e-01 9.16628912e-02 -3.74885440e-01
-1.41378176e+00 1.76035300e-01 3.82895440e-01 -8.99104416e-01
-5.58108985e-01 7.86827922e-01 -4.15896684e-01 -7.88520157e-01
-7.63538405e-02 -1.35531336e-01 -6.24879301e-01 2.73639530e-01
7.00116396e-01 1.51773453e-01 4.21244293e-01 -2.91660786e-01
-1.57961637e-01 2.45946288e-01 -2.95966446e-01 -5.36410332e-01
8.75010490e-01 -6.94144130e-01 -5.67604542e-01 3.69144052e-01
7.85365999e-01 1.22428499e-01 -6.38791502e-01 -5.65881491e-01
6.03697181e-01 -2.40120575e-01 -7.02100694e-01 -1.17441046e+00
9.67338234e-02 6.49078131e-01 -2.96148717e-01 8.55320275e-01
5.53613901e-01 3.78493398e-01 1.14757097e+00 1.09496272e+00
5.19544065e-01 -1.06179214e+00 2.71263927e-01 1.27699864e+00
1.15882969e+00 -1.07676816e+00 -3.43542814e-01 -6.36776149e-01
-7.95609593e-01 1.16579723e+00 6.93580806e-01 2.87908435e-01
1.90322861e-01 -3.66608389e-02 6.52173221e-01 -6.09844685e-01
-1.40113413e+00 -5.15461624e-01 1.64602287e-02 6.72654331e-01
2.06968322e-01 -1.58903211e-01 -5.82169712e-01 4.70052719e-01
-6.72956645e-01 -2.49418855e-01 7.01591492e-01 1.09401977e+00
-6.41991735e-01 -1.60544598e+00 -3.14079285e-01 2.02955365e-01
-4.44455922e-01 -2.75926262e-01 -1.16114736e+00 5.10063469e-01
-4.33490276e-01 1.73186123e+00 -2.17355385e-01 7.67412111e-02
4.14126605e-01 7.06072211e-01 4.92752612e-01 -8.99810135e-01
-1.17575908e+00 -4.88522768e-01 1.18891156e+00 -5.82218766e-01
-6.39694810e-01 -3.68437082e-01 -1.17802584e+00 -2.68979937e-01
-4.07857925e-01 9.92016912e-01 1.77318335e-01 1.41081357e+00
6.16617084e-01 1.85629204e-01 6.41292870e-01 -2.11968735e-01
-6.72040582e-01 -9.92743373e-01 5.50836980e-01 5.59420049e-01
4.60437179e-01 -2.24117920e-01 -3.28813970e-01 -5.64746670e-02]
|
[12.125896453857422, 8.006229400634766]
|
fd0cff62-6b93-4ef1-809a-efc7ade7b790
|
a-23-mw-data-centre-is-all-you-need
|
2203.17265
| null |
https://arxiv.org/abs/2203.17265v1
|
https://arxiv.org/pdf/2203.17265v1.pdf
|
A 23 MW data centre is all you need
|
The field of machine learning has achieved striking progress in recent years, witnessing breakthrough results on language modelling, protein folding and nitpickingly fine-grained dog breed classification. Some even succeeded at playing computer games and board games, a feat both of engineering and of setting their employers' expectations. The central contribution of this work is to carefully examine whether this progress, and technology more broadly, can be expected to continue indefinitely. Through a rigorous application of statistical theory and failure to extrapolate beyond the training data, we answer firmly in the negative and provide details: technology will peak at 3:07 am (BST) on 20th July, 2032. We then explore the implications of this finding, discovering that individuals awake at this ungodly hour with access to a sufficiently powerful computer possess an opportunity for myriad forms of long-term linguistic 'lock in'. All we need is a large (>> 1W) data centre to seize this pivotal moment. By setting our analogue alarm clocks, we propose a tractable algorithm to ensure that, for the future of humanity, the British spelling of colour becomes the default spelling across more than 80% of the global word processing software market.
|
['João F. Henriques', 'Dylan Campbell', 'Samuel Albanie']
|
2022-03-31
| null | null | null | null |
['board-games']
|
['playing-games']
|
[ 2.63384014e-01 1.64476603e-01 -1.88474849e-01 -4.47252244e-01
-5.80761969e-01 -4.90687758e-01 5.28771162e-01 2.57396430e-01
-7.78431892e-01 7.58791208e-01 -8.01123828e-02 -9.75969195e-01
-1.81577802e-01 -5.86575747e-01 -3.97833556e-01 -2.65694112e-01
-8.89043063e-02 3.77240360e-01 -1.46057829e-01 -6.37217283e-01
4.52797472e-01 4.21186000e-01 -1.52836573e+00 1.80199534e-01
5.45895278e-01 7.74661005e-01 -6.96446002e-03 7.64791012e-01
-2.50030547e-01 9.79255855e-01 -6.20714724e-01 -6.80467784e-01
1.12215064e-01 -3.02132159e-01 -8.77517641e-01 -4.06738818e-01
1.57867949e-02 1.04066886e-01 -1.73539072e-02 8.63364100e-01
4.60199594e-01 2.17839982e-02 1.93968832e-01 -1.02908909e+00
-7.07457840e-01 4.77736592e-01 -4.59098041e-01 2.33072370e-01
4.32997942e-01 2.84060955e-01 1.04391754e+00 -2.29532376e-01
1.48177266e-01 9.15613472e-01 7.14017630e-01 4.42746550e-01
-9.82475400e-01 -8.07054996e-01 1.37044698e-01 -2.70146038e-02
-1.39689159e+00 -7.44781494e-01 5.96338332e-01 -5.59475958e-01
1.34186137e+00 6.68528438e-01 6.98731899e-01 8.30159664e-01
3.76903385e-01 2.51444757e-01 1.25208318e+00 -8.72701049e-01
9.45658013e-02 1.04883075e-01 5.80803864e-02 9.08206046e-01
2.77760029e-01 8.10025632e-02 -8.72361481e-01 -3.99770528e-01
4.59356010e-01 -3.42596859e-01 2.80211754e-02 3.25271428e-01
-1.05980694e+00 7.85336435e-01 -2.67523557e-01 4.61579323e-01
-1.51826233e-01 1.99281037e-01 2.80460775e-01 3.45485628e-01
5.00988841e-01 5.52345514e-01 -7.77804315e-01 -7.18679309e-01
-9.17342722e-01 4.03318405e-01 8.58186722e-01 6.13757133e-01
3.15638453e-01 -2.37733141e-01 6.30203545e-01 7.81003714e-01
2.03466639e-01 3.62687796e-01 7.83707917e-01 -7.69572139e-01
1.84563443e-01 3.38734329e-01 1.45594940e-01 -1.13784456e+00
-5.26895940e-01 -2.01837018e-01 -7.62712836e-01 8.29618722e-02
8.65133584e-01 -2.54828036e-01 -5.42890012e-01 1.58968723e+00
-7.72869820e-03 1.49757192e-01 -3.59908521e-01 5.99816501e-01
1.56901985e-01 5.30362248e-01 4.50280279e-01 -2.72074103e-01
1.48155749e+00 -5.42159006e-02 -2.72751659e-01 -4.67234492e-01
7.26641655e-01 -8.63835156e-01 1.11198914e+00 9.60917830e-01
-1.06911206e+00 -5.85570991e-01 -1.16314363e+00 3.11598834e-02
-4.70713466e-01 -4.51961398e-01 1.11028469e+00 1.12351799e+00
-7.81849384e-01 6.54138982e-01 -9.60487723e-01 -3.94013107e-01
2.46525425e-02 3.73826325e-01 -2.14660257e-01 1.32822245e-01
-1.34839249e+00 9.92352724e-01 4.45540138e-02 1.66387439e-01
2.59496808e-01 -5.01809657e-01 -5.84285796e-01 -2.75786102e-01
1.96917653e-01 -5.39399385e-01 1.06124163e+00 -8.56138110e-01
-1.00165975e+00 1.42343509e+00 -2.23115563e-01 -4.79022473e-01
3.40248227e-01 -5.24026379e-02 -7.17188299e-01 -4.85576510e-01
-9.27939042e-02 2.17020065e-01 6.23039186e-01 -3.83956105e-01
-9.17372406e-01 -6.46534324e-01 -1.86828241e-01 -1.67454630e-01
-2.56845713e-01 5.70649803e-01 -1.20312482e-01 -4.34395164e-01
-2.57706251e-02 -8.77865255e-01 -3.14318329e-01 -4.08133894e-01
-9.38817635e-02 -4.92605299e-01 -1.00628763e-01 -4.98187989e-01
1.54774594e+00 -2.03322959e+00 -2.55252242e-01 2.73029178e-01
2.30267107e-01 1.40802965e-01 2.95141160e-01 3.38729411e-01
-3.04837197e-01 3.61153632e-01 4.20806855e-02 3.99941690e-02
4.65851545e-01 1.18890390e-01 -4.22278434e-01 6.66061997e-01
9.05864090e-02 8.47201407e-01 -7.58250833e-01 -2.16066957e-01
-1.85326710e-01 2.43715435e-01 -2.96557158e-01 -2.40986943e-01
1.17199473e-01 4.52165632e-03 -3.13859612e-01 5.52786708e-01
1.88932210e-01 -3.04491490e-01 5.28813660e-01 5.03755927e-01
-4.30933207e-01 2.26604223e-01 -9.29141223e-01 1.58426440e+00
-1.72323570e-01 5.02224207e-01 1.57775402e-01 -1.04413342e+00
7.75311172e-01 -4.47733933e-03 4.81028914e-01 -1.04077947e+00
2.06230283e-01 3.94167513e-01 3.03968132e-01 -4.90460157e-01
5.01301706e-01 -6.27053738e-01 -4.37814295e-01 4.92600203e-01
-6.35906339e-01 -4.58059870e-02 -1.84209988e-01 -1.29641280e-01
1.08313692e+00 1.31206155e-01 3.00094843e-01 -5.23595810e-01
2.78014153e-01 -1.80740131e-03 4.59783435e-01 6.71533525e-01
-3.89590353e-01 5.64486906e-02 3.80585372e-01 -8.21882546e-01
-9.71319616e-01 -7.52280772e-01 -1.49581820e-01 1.59970951e+00
-2.98322588e-01 -3.35295051e-01 -7.14461923e-01 2.05485597e-02
-3.41427326e-02 5.88812768e-01 -1.66558310e-01 -1.97930187e-01
-5.06616592e-01 -1.00995064e+00 6.61339521e-01 3.01242203e-01
2.26361364e-01 -1.05954576e+00 -9.89262104e-01 2.05964848e-01
6.33109137e-02 -7.79551744e-01 -2.50624716e-01 5.01028419e-01
-5.12607396e-01 -6.24800324e-01 -4.76022154e-01 -6.30639255e-01
1.05041385e-01 -1.51238188e-01 1.13646805e+00 4.86827701e-01
-5.59585750e-01 7.67327026e-02 -1.99917346e-01 -8.16256523e-01
-3.73072088e-01 7.55690932e-02 3.40213418e-01 -2.88176507e-01
1.16027248e+00 -7.28032053e-01 -3.60570133e-01 -2.61112511e-01
-6.30775213e-01 4.26138984e-03 4.02075410e-01 4.88937169e-01
-4.43229228e-02 2.55387306e-01 7.41698146e-01 -8.97911131e-01
6.98725045e-01 -4.06644017e-01 -5.60308158e-01 2.31154367e-01
-8.76920521e-01 1.76767949e-02 2.74164885e-01 -1.89202681e-01
-4.51511681e-01 -1.80549532e-01 -3.43155861e-01 3.91496748e-01
-1.97620854e-01 3.95090580e-01 5.08155450e-02 5.66579811e-02
6.89684451e-01 1.82068571e-02 1.07788006e-02 -4.05014217e-01
3.16144556e-01 1.18838775e+00 7.35729694e-01 -9.20990944e-01
7.07432926e-01 2.84160495e-01 -1.44783035e-01 -1.15995502e+00
-5.99314690e-01 -3.90394598e-01 -5.43169916e-01 -3.18268277e-02
7.18567610e-01 -7.39977121e-01 -1.30998576e+00 5.39473116e-01
-9.92111504e-01 -4.15068835e-01 8.85799080e-02 3.08212548e-01
-2.88343102e-01 3.97253722e-01 -2.68978506e-01 -1.14300823e+00
2.23028306e-02 -4.94100183e-01 7.35006332e-01 2.31787220e-01
-8.46195281e-01 -8.54088962e-01 -1.29075721e-01 6.45144522e-01
2.86184847e-01 2.43531242e-01 1.16614270e+00 -7.41932511e-01
-2.99894400e-02 -4.40676272e-01 -4.25539538e-02 2.99442261e-01
1.30434692e-01 -7.72592351e-02 -8.90004992e-01 -1.71889171e-01
1.91543698e-01 -2.91607141e-01 4.41010207e-01 1.62673056e-01
9.39506590e-01 -2.24407036e-02 -1.01701058e-01 4.34756964e-01
1.22147346e+00 4.34718370e-01 4.57186669e-01 3.69558871e-01
2.66436905e-01 6.27318978e-01 4.20071602e-01 5.23241341e-01
4.51472819e-01 3.74270499e-01 -1.45729855e-01 1.08893827e-01
3.29492807e-01 -3.05240542e-01 1.41866192e-01 9.61888134e-01
-3.21267098e-01 9.40236077e-02 -1.36618817e+00 3.87610584e-01
-1.46903825e+00 -1.08709538e+00 -1.91247221e-02 2.56700516e+00
9.57090557e-01 9.74117100e-01 3.36856246e-01 1.89106703e-01
4.31060135e-01 -1.12392731e-01 -4.53877836e-01 -1.04216719e+00
1.23574831e-01 5.95234692e-01 6.69911861e-01 6.04845464e-01
-7.63339579e-01 7.12920666e-01 6.93886137e+00 8.90311956e-01
-1.10127008e+00 -8.64026695e-02 9.86072481e-01 -7.01055899e-02
-3.02294403e-01 -1.57133922e-01 -7.66734600e-01 7.93142438e-01
1.41270554e+00 -1.96393371e-01 7.59950995e-01 5.23940504e-01
3.58853787e-01 -2.84114778e-01 -8.21870446e-01 1.11409807e+00
1.51880160e-01 -1.19271958e+00 -5.30741513e-01 4.23986256e-01
1.76659390e-01 1.18743673e-01 1.70431241e-01 3.20594698e-01
4.47833627e-01 -1.60494792e+00 7.08967626e-01 6.17581427e-01
8.39875221e-01 -9.28658247e-01 2.26587981e-01 8.18294764e-01
-6.28324449e-01 -1.58723503e-01 -2.94606537e-01 -8.43388081e-01
-3.62476707e-02 4.34426039e-01 -4.80202347e-01 2.69112229e-01
3.51767570e-01 -1.03232704e-01 -4.10727769e-01 6.43785059e-01
1.96725085e-01 6.37413442e-01 -4.58985597e-01 -2.17701003e-01
2.31233165e-01 -6.64456710e-02 4.77964990e-02 1.17332244e+00
9.05668959e-02 5.06099403e-01 -9.65075865e-02 3.55307043e-01
1.58265948e-01 -2.46515889e-02 -4.86631930e-01 -4.54029202e-01
4.66079414e-01 9.05556381e-01 -7.51902103e-01 -1.12027973e-01
-6.51777983e-01 8.15862000e-01 -4.84746322e-03 1.76129322e-02
-7.93674767e-01 -6.15218580e-01 7.66404510e-01 2.59798467e-01
-2.34780118e-01 -5.21354139e-01 -5.75339019e-01 -8.70678604e-01
-5.39662875e-02 -1.16342854e+00 9.50580314e-02 -4.78626847e-01
-1.36168480e+00 4.86361265e-01 -3.91734958e-01 -3.00634325e-01
-4.43414330e-01 -8.14296246e-01 -1.76061153e-01 1.15447247e+00
-1.08784461e+00 -9.30144787e-01 4.41717833e-01 1.80908903e-01
2.56144047e-01 -2.28322551e-01 9.59668577e-01 3.07938337e-01
-2.87567288e-01 7.13369727e-01 1.04461469e-01 -2.51424331e-02
6.35166168e-01 -1.03213692e+00 7.95218825e-01 4.22484010e-01
5.14555812e-01 9.81236637e-01 8.71553481e-01 -4.37352002e-01
-1.54311609e+00 -4.37170625e-01 1.52149618e+00 -9.20138299e-01
8.97788405e-01 -6.90223634e-01 -5.90664208e-01 6.05956733e-01
-1.27649143e-01 -4.45692062e-01 8.36263061e-01 5.53790331e-01
-3.16953003e-01 6.06848747e-02 -8.99658442e-01 4.33680177e-01
8.94105077e-01 -7.36822188e-01 -6.24645591e-01 3.45180064e-01
5.26018381e-01 -1.40625417e-01 -5.34705520e-01 1.62640750e-01
1.04552913e+00 -9.30014193e-01 8.30012977e-01 -1.16573846e+00
1.88653901e-01 1.04615264e-01 -2.86039263e-01 -7.68787742e-01
-3.06491047e-01 -1.41184735e+00 4.85191256e-01 9.61482644e-01
4.31228638e-01 -7.57070422e-01 9.81858552e-01 1.22969484e+00
1.20203540e-01 -9.33467627e-01 -9.20438528e-01 -7.92158306e-01
4.77159202e-01 -9.91675198e-01 5.86118102e-01 1.14684558e+00
6.06687844e-01 2.97351986e-01 -3.90599549e-01 -1.34716764e-01
4.23182249e-01 -1.15138277e-01 5.42888641e-01 -1.24064195e+00
-5.75431883e-01 -4.79402810e-01 -6.16015494e-01 -9.25177276e-01
-3.42862196e-02 -8.18089902e-01 -2.18564481e-01 -9.08037901e-01
1.23275697e-01 -3.75235200e-01 -5.43686450e-01 4.01504904e-01
-7.19709247e-02 4.53649580e-01 2.49759898e-01 -5.29329367e-02
-3.46540183e-01 -4.42271233e-01 6.43189609e-01 1.10369518e-01
-2.02094521e-02 2.90128160e-02 -1.32983255e+00 9.83503282e-01
7.61341453e-01 -1.18674897e-01 -3.35025400e-01 -2.53413826e-01
8.89642060e-01 -1.91856205e-01 3.66950750e-01 -8.26367915e-01
3.30102235e-01 -4.90825623e-01 3.91594261e-01 4.39668167e-03
2.43480399e-01 -5.91230214e-01 2.53529221e-01 4.70581442e-01
-3.67104620e-01 9.50771347e-02 2.41710052e-01 3.34340632e-01
2.61465758e-01 -1.51711687e-01 7.12625265e-01 -2.07472503e-01
-5.87054551e-01 -5.56644313e-02 -4.64813948e-01 2.06390098e-01
8.99557233e-01 -2.44163573e-01 -2.18069717e-01 -2.91140705e-01
-7.36645162e-01 1.28498077e-02 4.85049814e-01 3.90892744e-01
9.26323086e-02 -8.50008547e-01 -4.86533135e-01 4.00308132e-01
-1.76985815e-01 -6.07004285e-01 8.11891854e-02 6.64620042e-01
-5.39140999e-01 6.84390604e-01 -1.18410736e-02 1.06305316e-01
-1.15302479e+00 5.88412762e-01 4.22955677e-02 -1.24211408e-01
-5.41879058e-01 1.26205266e+00 -2.73011208e-01 -3.37378569e-02
1.93046853e-01 -2.39243180e-01 2.36408800e-01 6.41814172e-02
7.81467557e-01 1.75141767e-01 1.61782011e-01 -3.83156925e-01
-5.19567668e-01 2.83024848e-01 -9.89367962e-02 -1.98692054e-01
1.35777915e+00 -1.24449125e-02 -2.12412015e-01 7.30354249e-01
8.98654401e-01 2.89966047e-01 -7.41161466e-01 2.35340297e-01
2.76219428e-01 -3.13700467e-01 -1.13580979e-01 -9.23832417e-01
-1.98922515e-01 8.44594002e-01 5.34420907e-01 6.05961859e-01
8.82822216e-01 -9.25590843e-02 1.00941348e+00 3.46550554e-01
6.43591523e-01 -1.22512412e+00 -5.13467073e-01 2.46906891e-01
2.36882642e-01 -9.02465582e-01 2.35240519e-01 8.38108733e-02
-3.81036073e-01 8.76509368e-01 -8.74087587e-02 1.10992648e-01
5.97828209e-01 2.80837536e-01 -1.06437013e-01 -3.84869307e-01
-8.83051395e-01 7.56955072e-02 -2.37984270e-01 6.45810783e-01
8.16132784e-01 3.60354155e-01 -4.75076914e-01 7.68090665e-01
-5.53616643e-01 2.80237850e-02 2.60296136e-01 8.60678971e-01
-1.05035102e+00 -1.25698686e+00 -4.00025606e-01 4.20307070e-01
-8.00368130e-01 -3.72948378e-01 -3.30459177e-01 7.73132086e-01
3.91874671e-01 1.05356503e+00 1.38065308e-01 -2.79298633e-01
2.80175507e-01 5.74567258e-01 5.22002518e-01 -5.94357967e-01
-3.82473737e-01 -5.13453111e-02 2.16112882e-01 -3.99091661e-01
-1.24423712e-01 -7.38298833e-01 -1.23430598e+00 -1.01411283e+00
-3.82471606e-02 2.78041065e-01 8.94431591e-01 1.05866969e+00
1.63619846e-01 8.27838294e-03 2.96175748e-01 -2.66673416e-01
-8.07037175e-01 -6.22304022e-01 -7.84853756e-01 -1.27824873e-01
1.04767404e-01 -2.63087809e-01 -2.61596709e-01 2.00118423e-01]
|
[9.01220989227295, 6.490387916564941]
|
8c6ea4b1-e0a9-4c9e-9128-3f7d45bdd31e
|
releasing-inequlity-phenomena-in-l-infty
|
2305.09305
| null |
https://arxiv.org/abs/2305.09305v2
|
https://arxiv.org/pdf/2305.09305v2.pdf
|
Releasing Inequality Phenomena in $L_{\infty}$-Adversarial Training via Input Gradient Distillation
|
Since adversarial examples appeared and showed the catastrophic degradation they brought to DNN, many adversarial defense methods have been devised, among which adversarial training is considered the most effective. However, a recent work showed the inequality phenomena in $l_{\infty}$-adversarial training and revealed that the $l_{\infty}$-adversarially trained model is vulnerable when a few important pixels are perturbed by i.i.d. noise or occluded. In this paper, we propose a simple yet effective method called Input Gradient Distillation (IGD) to release the inequality phenomena in $l_{\infty}$-adversarial training. Experiments show that while preserving the model's adversarial robustness, compared to PGDAT, IGD decreases the $l_{\infty}$-adversarially trained model's error rate to inductive noise and inductive occlusion by up to 60\% and 16.53\%, and to noisy images in Imagenet-C by up to 21.11\%. Moreover, we formally explain why the equality of the model's saliency map can improve such robustness.
|
['Xiaohua Xie', 'Junhao Dong', 'Junxi Chen']
|
2023-05-16
| null | null | null | null |
['adversarial-defense']
|
['adversarial']
|
[ 3.00047606e-01 5.38490951e-01 2.30806202e-01 -3.35393697e-02
-5.02423108e-01 -6.60672605e-01 4.37583297e-01 -5.78066945e-01
-4.77999777e-01 1.04486167e+00 -1.85001001e-01 -5.07111669e-01
6.58225566e-02 -8.68893266e-01 -1.08455312e+00 -9.00343657e-01
-1.17692076e-01 -1.75689697e-01 3.52249473e-01 -4.54193741e-01
5.24582975e-02 5.45039117e-01 -1.31335580e+00 -3.08708325e-02
1.13338995e+00 1.10690427e+00 8.53148028e-02 5.36837399e-01
-8.57828036e-02 9.82158124e-01 -1.04427838e+00 -7.40662515e-01
7.60594010e-01 -5.35611212e-01 -5.19166291e-01 -5.93467236e-01
6.83333158e-01 -3.82149488e-01 -6.83716178e-01 1.79640639e+00
6.46299601e-01 1.17536761e-01 5.26515961e-01 -1.36923611e+00
-1.00558817e+00 5.86405754e-01 -4.44163233e-01 3.82004648e-01
-1.41857356e-01 2.58851200e-01 3.51522505e-01 -6.43243551e-01
5.95454276e-01 1.19609177e+00 5.05659819e-01 1.02997899e+00
-9.63181853e-01 -1.16621196e+00 3.88762951e-01 1.11292072e-01
-1.19885767e+00 -1.55640185e-01 1.05980313e+00 -2.26276860e-01
5.25286198e-01 4.13258791e-01 2.34252334e-01 1.11859775e+00
4.12113219e-01 6.17127657e-01 1.27818704e+00 -2.50968963e-01
1.25212640e-01 2.31981799e-01 -3.73936981e-01 6.98461771e-01
1.14180721e-01 5.31978011e-01 1.91332251e-02 2.56117553e-01
8.27662647e-01 -8.63433555e-02 -3.32894415e-01 7.66932815e-02
-5.98354697e-01 6.62909091e-01 1.07284021e+00 2.06852257e-01
1.36556625e-02 3.72299820e-01 3.23801786e-01 4.65745538e-01
5.40672541e-01 8.38373005e-01 -2.75334835e-01 2.52968729e-01
-7.15086102e-01 1.19825020e-01 2.63852507e-01 1.05471623e+00
6.86141372e-01 7.81642377e-01 -7.41648301e-02 6.19636416e-01
5.97058907e-02 9.73291755e-01 2.20332444e-01 -9.93581653e-01
4.77658987e-01 3.92664045e-01 -8.49343091e-02 -1.08602250e+00
-7.73810372e-02 -5.61327279e-01 -1.13661051e+00 8.07681680e-01
5.04605949e-01 -4.00408983e-01 -1.28471422e+00 2.07753110e+00
-7.63424039e-02 1.07616693e-01 5.43828905e-02 8.49187136e-01
6.60625756e-01 4.96320575e-01 2.59038180e-01 -1.03947492e-02
8.12869072e-01 -7.49672890e-01 -5.54554284e-01 -3.38478297e-01
5.68178622e-03 -9.32874799e-01 1.03915584e+00 8.18064511e-02
-1.14798164e+00 -6.41766429e-01 -1.21329749e+00 3.51486087e-01
-6.07419193e-01 -5.91811717e-01 4.32377785e-01 9.71910059e-01
-9.05828476e-01 7.26910710e-01 -4.23274308e-01 1.67068928e-01
7.89563537e-01 5.19714415e-01 -1.41785622e-01 -1.11311391e-01
-1.63576365e+00 1.05297267e+00 2.62020886e-01 3.35304923e-02
-1.28663337e+00 -6.87693655e-01 -6.88710749e-01 -1.88672528e-01
1.73343331e-01 -4.63774264e-01 7.40458250e-01 -1.26932001e+00
-1.26059914e+00 8.16661358e-01 3.10252666e-01 -8.45541179e-01
7.67569959e-01 -3.52132082e-01 -6.97303414e-01 2.40332052e-01
1.76475514e-02 8.33187580e-01 1.17255616e+00 -1.46994698e+00
-2.74126768e-01 -3.30840707e-01 4.51749742e-01 6.78472444e-02
-2.97629267e-01 1.10177465e-01 -5.80257550e-03 -1.07877791e+00
1.21138297e-01 -9.30662274e-01 -3.27813387e-01 -2.52406299e-02
-5.91596305e-01 3.85667622e-01 1.03762388e+00 -5.14015615e-01
1.08188021e+00 -2.11073494e+00 -1.47269979e-01 1.93259135e-01
3.61795217e-01 8.54344308e-01 -1.37202144e-02 -1.07215248e-01
-2.96796262e-01 5.42669654e-01 -4.18693036e-01 8.29668567e-02
-3.91762778e-02 3.24548662e-01 -6.13893509e-01 5.03114641e-01
2.12862790e-01 9.45490837e-01 -8.42965782e-01 -3.40769619e-01
3.56644452e-01 5.72184563e-01 -5.01244783e-01 8.87516588e-02
-1.26737282e-01 5.01701593e-01 -4.99408990e-01 7.23907828e-01
1.02936947e+00 4.76858228e-01 -4.86727983e-01 4.41470137e-03
1.73710868e-01 -2.59360820e-01 -7.37730503e-01 1.17472267e+00
-5.13843410e-02 7.92321622e-01 1.22855976e-01 -8.66674483e-01
9.96846974e-01 1.71795398e-01 1.13752246e-01 -1.00801826e+00
3.00465167e-01 2.47408181e-01 9.77054089e-02 -3.11295629e-01
3.41903031e-01 -5.13377428e-01 -1.34551689e-01 4.04550135e-03
-5.03267050e-02 -3.15869808e-01 -3.46272558e-01 3.05439025e-01
1.14719462e+00 -3.19695994e-02 -2.65895814e-01 -4.62437302e-01
5.54730654e-01 -2.15268701e-01 7.55962491e-01 9.47860539e-01
-6.13169789e-01 6.38861120e-01 4.29127395e-01 -4.89760846e-01
-1.09937811e+00 -1.34739316e+00 -1.66886330e-01 8.33696246e-01
4.48256284e-01 3.54587793e-01 -1.17124188e+00 -9.70822096e-01
-1.19442068e-01 8.35312843e-01 -7.13642836e-01 -6.60022020e-01
-7.53413856e-01 -5.98474562e-01 1.16809690e+00 4.72180605e-01
1.18523479e+00 -1.29642129e+00 -2.79630274e-01 6.59128593e-04
9.10621658e-02 -9.42242861e-01 -2.98140526e-01 3.50032091e-01
-7.51666605e-01 -8.36453974e-01 -9.54665422e-01 -6.94495380e-01
1.05268133e+00 -1.16905339e-01 1.05484116e+00 5.35415076e-02
-2.28698060e-01 -1.90807670e-01 -3.30783695e-01 -7.87587464e-01
-5.34733415e-01 -2.82218516e-01 1.67570814e-01 -4.00660843e-01
-6.06185906e-02 -7.44651854e-01 -6.53409719e-01 5.16543567e-01
-1.09813273e+00 -3.67260247e-01 4.32607174e-01 8.31114829e-01
6.20896339e-01 2.32057676e-01 6.82146013e-01 -9.27298546e-01
2.44435042e-01 -1.64255336e-01 -5.99961519e-01 -1.13605566e-01
-5.44202089e-01 6.78483099e-02 1.10174119e+00 -5.41309536e-01
-9.05282915e-01 -2.75792867e-01 -4.17406589e-01 -9.29646075e-01
-7.26086274e-02 -1.26159638e-01 -4.57604945e-01 -4.36795831e-01
9.97317791e-01 7.43780360e-02 -4.12784666e-01 -2.48926103e-01
5.03618181e-01 9.83627960e-02 9.27747130e-01 -3.33040535e-01
1.36712682e+00 4.35217202e-01 -1.48565974e-02 -4.37496036e-01
-6.75915837e-01 4.32238400e-01 -2.38862395e-01 -3.65466923e-01
8.06271613e-01 -6.80613220e-01 -4.54287618e-01 7.41750181e-01
-7.89933562e-01 -3.72328520e-01 -6.06861591e-01 2.91454107e-01
-3.85135233e-01 9.59033743e-02 -4.33806658e-01 -7.69238412e-01
-3.27036321e-01 -1.19606864e+00 4.23515975e-01 4.18693244e-01
2.42382988e-01 -8.32887292e-01 -4.62134749e-01 1.66801170e-01
7.49626398e-01 7.91223884e-01 9.00182247e-01 -4.78781492e-01
-5.15133381e-01 -2.69392759e-01 -3.26932549e-01 9.50683713e-01
-7.72897229e-02 -2.06625938e-01 -1.18590868e+00 -1.93102688e-01
4.54734713e-01 -1.57207474e-01 1.01371026e+00 3.77870381e-01
1.16253901e+00 -6.10415220e-01 -9.29854959e-02 8.72271001e-01
1.42620158e+00 6.11558318e-01 1.22327411e+00 4.02472079e-01
7.06325889e-01 2.13732421e-01 4.10379410e-01 -1.28809273e-01
-4.06977028e-01 2.99369186e-01 1.20014882e+00 -3.70163471e-01
-3.89909446e-01 -3.42039913e-01 5.01580477e-01 4.19300854e-01
2.91895699e-02 -2.90766567e-01 -5.52946687e-01 4.21565890e-01
-1.17137063e+00 -1.13343883e+00 4.20860425e-02 1.97502089e+00
8.46831620e-01 7.25213170e-01 -4.30549294e-01 2.00743645e-01
9.71566319e-01 4.03335065e-01 -8.02049458e-01 -7.10695148e-01
-5.72167397e-01 4.79432464e-01 8.17615986e-01 4.62127298e-01
-1.12462485e+00 1.22580743e+00 5.84662533e+00 1.09660578e+00
-1.18803275e+00 3.24564427e-02 8.04358423e-01 -1.99764460e-01
-4.62361425e-01 -2.00128302e-01 -4.26253825e-01 7.98909307e-01
6.50909662e-01 -2.17167377e-01 2.78987706e-01 1.00808525e+00
-1.78930253e-01 -9.27145854e-02 -6.51734173e-01 6.21721745e-01
4.25977521e-02 -1.10179126e+00 1.57496817e-02 -7.18642548e-02
1.03202844e+00 -7.39182159e-02 6.63750410e-01 4.57109839e-01
4.83573794e-01 -1.18161952e+00 7.15819776e-01 3.39184970e-01
1.04647422e+00 -1.05411470e+00 7.33607590e-01 2.09042698e-01
-8.11554849e-01 -2.85246074e-02 -4.44776744e-01 -6.29968271e-02
-1.83955953e-02 7.55374789e-01 -5.14728427e-01 3.46495956e-01
9.06521857e-01 1.27318352e-01 -6.12575591e-01 5.62086284e-01
-6.77693248e-01 6.29169345e-01 -1.10926062e-01 3.04853618e-01
3.62004697e-01 5.12756296e-02 8.38251591e-01 7.61737406e-01
1.52729914e-01 1.60274431e-01 -2.67461509e-01 9.49511945e-01
-4.99043077e-01 -3.72801185e-01 -9.19662654e-01 3.27927470e-01
3.50525618e-01 6.31412685e-01 -6.04450822e-01 -2.72165626e-01
1.96421385e-01 1.22054398e+00 -7.68021718e-02 5.31147718e-01
-1.46248436e+00 -7.61534750e-01 9.29433167e-01 -1.29614733e-02
1.50417238e-01 1.40774772e-01 -5.01684010e-01 -8.45509529e-01
1.47529580e-02 -8.24487448e-01 -2.77278060e-03 -7.81027198e-01
-1.05754602e+00 1.04751337e+00 -2.38867566e-01 -1.13817430e+00
1.28268998e-03 -4.97437626e-01 -6.14316702e-01 1.00690293e+00
-1.45802999e+00 -7.14729428e-01 -1.32821649e-01 9.23740089e-01
3.22699964e-01 -3.11213911e-01 8.02787662e-01 4.65851575e-01
-6.63095713e-01 1.12112379e+00 2.38391727e-01 3.82758230e-01
5.15537679e-01 -1.14038873e+00 4.52884614e-01 1.28197885e+00
-1.86293781e-01 4.32783067e-01 1.03977942e+00 -4.67956424e-01
-9.25248146e-01 -1.47802889e+00 6.01857603e-01 -2.87183851e-01
4.36598301e-01 -2.37468645e-01 -8.77104461e-01 6.63576961e-01
2.22230569e-01 2.68414199e-01 1.73787475e-01 -7.04596937e-01
-5.10179818e-01 -2.44852781e-01 -1.96134508e+00 9.70155358e-01
1.20826352e+00 -7.23679304e-01 -5.88205397e-01 8.48389566e-02
1.08964396e+00 -5.27818263e-01 -7.60068476e-01 6.40842557e-01
2.52693713e-01 -1.18147612e+00 1.15254962e+00 -4.36558336e-01
6.07804775e-01 -3.61404389e-01 -3.35376978e-01 -1.19642663e+00
-1.22110359e-01 -7.87025511e-01 7.97620267e-02 1.16039658e+00
4.22406912e-01 -9.18333530e-01 7.52898633e-01 4.75594163e-01
-5.59979200e-01 -6.67764843e-01 -1.25776684e+00 -1.01172602e+00
5.13563633e-01 -5.08926213e-01 3.76400560e-01 8.18893850e-01
-5.50643265e-01 -3.44862998e-01 -3.53033811e-01 2.74943084e-01
5.71454525e-01 -4.16247189e-01 4.17298824e-01 -6.61177456e-01
1.12719916e-01 -3.81033927e-01 -6.29808247e-01 -7.72614419e-01
7.01698884e-02 -6.06754601e-01 3.59542109e-02 -1.06589150e+00
-3.89845878e-01 -6.19415224e-01 -7.13508487e-01 5.07815301e-01
-3.57465029e-01 7.13153005e-01 3.07410270e-01 5.59707778e-03
-2.80722529e-01 4.40821379e-01 1.50794423e+00 -4.01382655e-01
1.51939392e-01 -1.32120088e-01 -8.22788239e-01 9.21654820e-01
9.82969522e-01 -7.50737906e-01 -3.50354820e-01 -5.29836476e-01
4.59500588e-02 -3.75737906e-01 5.23189604e-01 -1.22990251e+00
-1.29760737e-02 -8.68037120e-02 5.69543958e-01 -1.98982716e-01
2.92540252e-01 -7.58147418e-01 1.12162858e-01 5.99999607e-01
-1.94695964e-01 -1.20768145e-01 5.89878798e-01 3.08941960e-01
-2.51785100e-01 -1.02285802e-01 1.31437039e+00 -2.54673749e-01
-7.64189243e-01 1.50083825e-01 -1.42846078e-01 3.98063242e-01
1.25537288e+00 -3.15944582e-01 -5.76687992e-01 -2.36290485e-01
-6.73092484e-01 -4.21848483e-02 4.55795407e-01 3.46812159e-01
5.75771868e-01 -1.27359319e+00 -4.94511306e-01 4.01607037e-01
-3.43515694e-01 2.37233341e-02 4.35146123e-01 3.50001663e-01
-5.76818645e-01 1.03600681e-01 -5.49633324e-01 -2.30073005e-01
-9.89399016e-01 8.89334559e-01 5.33083975e-01 -2.09464043e-01
-3.34764004e-01 1.49143040e+00 2.14980990e-01 -3.31629515e-02
3.36318254e-01 -6.83824271e-02 1.07245661e-01 -2.37968192e-01
2.94532508e-01 4.68604863e-01 -1.25294551e-01 -6.28033638e-01
-3.50815296e-01 4.60356832e-01 -9.47664753e-02 3.74574400e-02
9.63176548e-01 4.18655053e-02 -1.19640101e-02 -1.49060920e-01
1.30336869e+00 -1.15317861e-02 -1.56604218e+00 -1.19779874e-02
-5.77793419e-01 -4.51215297e-01 -1.62333488e-01 -1.00456583e+00
-1.50536501e+00 9.74615753e-01 8.37915182e-01 2.79177457e-01
1.53099871e+00 -2.52160102e-01 8.98628294e-01 -4.22513345e-03
3.80205601e-01 -1.03440535e+00 -7.31397560e-03 6.24214768e-01
9.84067142e-01 -9.76397216e-01 -1.85223475e-01 -2.95693010e-01
-5.71841896e-01 5.36859989e-01 8.21189642e-01 -5.07759452e-01
6.86393201e-01 3.23903948e-01 3.62163305e-01 -2.30578450e-03
-1.52438045e-01 -6.87692240e-02 7.67424032e-02 8.23463857e-01
-1.57157615e-01 -2.95919627e-02 -7.47837424e-02 4.19260055e-01
-5.45420587e-01 -3.78498435e-01 2.43844002e-01 7.39545763e-01
-3.82937759e-01 -8.85095775e-01 -4.80245590e-01 2.61500448e-01
-8.81503582e-01 -3.00795346e-01 -3.33033025e-01 9.31586742e-01
7.88191497e-01 9.56043839e-01 -2.10519731e-01 -6.69546723e-01
4.11025822e-01 7.83358440e-02 3.09360832e-01 -1.98879540e-01
-9.37698960e-01 -2.94621944e-01 -3.77283126e-01 -5.94603360e-01
-1.89046100e-01 -4.86737229e-02 -1.30027938e+00 -4.84885871e-01
-3.02220374e-01 6.35742918e-02 5.73381305e-01 6.96246564e-01
7.56172091e-02 8.98436964e-01 8.96755457e-01 -6.59268677e-01
-4.53269809e-01 -7.88121045e-01 -5.36139011e-01 5.46199143e-01
3.11169744e-01 -5.04257262e-01 -8.30399573e-01 5.31457067e-02]
|
[5.564581394195557, 7.9494500160217285]
|
b2bbac02-ef25-4729-9468-b469eb725c25
|
the-metaverse-survey-trends-novel-pipeline
|
2304.0924
| null |
https://arxiv.org/abs/2304.09240v1
|
https://arxiv.org/pdf/2304.09240v1.pdf
|
The Metaverse: Survey, Trends, Novel Pipeline Ecosystem & Future Directions
|
The Metaverse offers a second world beyond reality, where boundaries are non-existent, and possibilities are endless through engagement and immersive experiences using the virtual reality (VR) technology. Many disciplines can benefit from the advancement of the Metaverse when accurately developed, including the fields of technology, gaming, education, art, and culture. Nevertheless, developing the Metaverse environment to its full potential is an ambiguous task that needs proper guidance and directions. Existing surveys on the Metaverse focus only on a specific aspect and discipline of the Metaverse and lack a holistic view of the entire process. To this end, a more holistic, multi-disciplinary, in-depth, and academic and industry-oriented review is required to provide a thorough study of the Metaverse development pipeline. To address these issues, we present in this survey a novel multi-layered pipeline ecosystem composed of (1) the Metaverse computing, networking, communications and hardware infrastructure, (2) environment digitization, and (3) user interactions. For every layer, we discuss the components that detail the steps of its development. Also, for each of these components, we examine the impact of a set of enabling technologies and empowering domains (e.g., Artificial Intelligence, Security & Privacy, Blockchain, Business, Ethics, and Social) on its advancement. In addition, we explain the importance of these technologies to support decentralization, interoperability, user experiences, interactions, and monetization. Our presented study highlights the existing challenges for each component, followed by research directions and potential solutions. To the best of our knowledge, this survey is the most comprehensive and allows users, scholars, and entrepreneurs to get an in-depth understanding of the Metaverse ecosystem to find their opportunities and potentials for contribution.
|
['Mohsen Guizani', 'Ernesto Damiani', 'Zbigniew Dziong', 'Chamseddine Talhi', 'Jamal Bentahar', 'Rabeb Mizouni', 'Omar Abdel Wahab', 'Hadi Otrok', 'Azzam Mourad', 'Mohamad Ajaj', 'Osama Wehbi', 'Mario Chahoud', 'Sarhad Arisdakessian', 'Mohamad Wazzeh', 'Mouhamad Arafeh', 'Ahmad Hammoud', 'Hani Sami']
|
2023-04-18
| null | null | null | null |
['business-ethics', 'culture']
|
['miscellaneous', 'speech']
|
[-3.57610822e-01 5.37124276e-02 -5.66605568e-01 3.35358858e-01
-1.57887731e-02 -1.13674474e+00 6.67489052e-01 -2.36908227e-01
-3.28987278e-02 5.98327875e-01 2.50261635e-01 -8.65826309e-01
7.01504424e-02 -8.75759661e-01 -4.49299872e-01 -2.75300413e-01
7.84042925e-02 -2.24625781e-01 -1.99262463e-02 -6.86094165e-01
5.17287552e-01 2.72542775e-01 -1.75225902e+00 -3.96517478e-02
9.95825350e-01 1.01694298e+00 -8.38229060e-02 4.03982341e-01
-2.36204758e-01 7.72366524e-01 -5.61068952e-01 -6.86961293e-01
7.71637261e-02 -1.27155796e-01 -6.03089929e-01 -5.79507291e-01
-2.44113535e-01 -5.33769965e-01 -1.53216958e-01 9.11840975e-01
3.51032138e-01 -4.87279594e-01 -1.34851396e-01 -1.48772752e+00
-1.01170588e+00 6.15856946e-01 -2.47270375e-01 -6.05597161e-02
6.89399123e-01 3.97381008e-01 7.27622747e-01 -7.23894835e-01
1.17197049e+00 5.20828724e-01 5.13703048e-01 3.87297899e-01
-7.64323235e-01 -9.77177083e-01 9.08146054e-03 -6.53370097e-03
-1.32792497e+00 -3.95869881e-01 4.90010679e-01 -5.67825794e-01
7.77735054e-01 9.88485038e-01 1.27580500e+00 1.12406158e+00
2.19342127e-01 3.50020558e-01 1.28918409e+00 -2.51816630e-01
1.45409644e-01 5.90966582e-01 9.57777947e-02 2.73368936e-02
5.86496413e-01 9.73233730e-02 -3.96924555e-01 -1.38496429e-01
9.24348354e-01 9.13757458e-03 -3.49237680e-01 -3.49408984e-01
-1.37768185e+00 3.14099371e-01 1.01828642e-01 6.83279514e-01
-4.87073123e-01 2.15936154e-01 4.73785281e-01 2.69826353e-01
2.08081856e-01 6.10662580e-01 -2.11063072e-01 -9.83020246e-01
-4.40321773e-01 -4.68784980e-02 1.22269809e+00 8.79198790e-01
2.71247685e-01 1.44281359e-02 2.45610431e-01 2.18279898e-01
5.88675439e-01 2.49736547e-01 -1.18612632e-01 -1.23265481e+00
4.57208194e-02 5.37784874e-01 3.89744073e-01 -1.09300601e+00
-5.40070757e-02 -5.60925603e-01 -5.24491906e-01 5.87383807e-02
-1.27221867e-01 -2.93357939e-01 -3.38548839e-01 1.32825577e+00
7.17080832e-01 2.96287805e-01 -5.90507593e-03 8.91851664e-01
1.44234788e+00 5.84295213e-01 3.43272276e-02 -2.07227796e-01
1.56500649e+00 -6.10821605e-01 -1.08577859e+00 2.36456677e-01
4.73592073e-01 -8.17033887e-01 9.90434706e-01 2.84496665e-01
-1.22136688e+00 2.55837798e-01 -1.20739746e+00 -8.33255611e-03
-6.82205200e-01 -6.28379881e-01 1.01776850e+00 1.32662594e+00
-1.06806350e+00 2.15855494e-01 -5.41446507e-01 -5.16636252e-01
2.04389080e-01 1.28846124e-01 -2.76359141e-01 1.17413521e-01
-1.26676774e+00 7.93381393e-01 1.16672158e-01 -4.83933017e-02
-2.01545119e-01 -5.76581001e-01 -4.17548865e-01 -6.16979599e-02
3.35981250e-01 -9.74742115e-01 9.82726395e-01 -6.99470460e-01
-1.87941456e+00 7.85367131e-01 4.80808675e-01 -2.01237991e-01
4.86305326e-01 -2.03872830e-01 -6.57777488e-01 4.66784053e-02
-1.43274412e-01 -8.99986029e-02 -1.42686889e-01 -1.43904829e+00
-6.67137802e-01 -2.06317499e-01 3.61009598e-01 3.29646431e-02
-5.14121652e-01 5.55953324e-01 -5.94594002e-01 -5.11684939e-02
-1.22589290e-01 -7.51631141e-01 -1.08531371e-01 -3.56014848e-01
-3.56498733e-02 3.37414324e-01 1.11935759e+00 -5.74968576e-01
1.58636546e+00 -2.34056234e+00 -3.06014031e-01 1.66048810e-01
7.41201043e-01 3.87517542e-01 3.20132285e-01 1.10909808e+00
4.48359519e-01 1.00318444e+00 6.21974289e-01 2.20071077e-01
2.12758854e-01 -1.46756917e-01 -2.11942986e-01 4.55451876e-01
-5.46373904e-01 1.02323294e+00 -9.88568127e-01 -1.63432248e-02
3.63960445e-01 9.03798699e-01 -5.14339626e-01 -2.70422101e-01
1.80361554e-01 6.77740097e-01 -5.79790771e-01 9.25192595e-01
8.19605589e-01 -3.95512611e-01 7.22595870e-01 1.95909709e-01
-1.02808630e+00 3.40518504e-01 -1.03321958e+00 1.28605497e+00
-6.05641782e-01 7.22909868e-01 5.71036220e-01 -1.84172064e-01
7.20310867e-01 7.82403588e-01 6.28063202e-01 -9.54146385e-01
3.10010374e-01 6.54293239e-01 -1.07332528e-01 -7.04539061e-01
8.34847033e-01 2.56623626e-01 -7.91066512e-03 4.80237663e-01
-4.69783455e-01 1.85959101e-01 -2.73063451e-01 2.59969592e-01
8.41971099e-01 2.01254785e-01 4.89755601e-01 1.18112542e-01
1.77111886e-02 -9.75370631e-02 5.21020353e-01 4.26235825e-01
-5.58892369e-01 1.22752234e-01 6.38145864e-01 -2.20921412e-01
-7.69294083e-01 -7.27179110e-01 -1.49930567e-01 6.60909295e-01
7.53935575e-01 -9.30059195e-01 -4.46771324e-01 -3.28900009e-01
-1.53431207e-01 7.30045140e-01 -2.37731189e-01 3.24623644e-01
-2.99723029e-01 -3.59901130e-01 3.22151959e-01 6.38784543e-02
8.36992919e-01 -7.20847547e-01 -1.01840568e+00 8.32801685e-02
-4.05027866e-01 -1.20571697e+00 2.97092259e-01 -7.06236720e-01
-6.23121619e-01 -9.37709391e-01 -4.84541468e-02 -3.37761551e-01
2.86630653e-02 4.90015835e-01 9.35256362e-01 4.04154211e-01
3.02153170e-01 5.62444627e-01 -5.10601819e-01 -3.36822093e-01
-2.21387923e-01 -1.88090876e-01 -1.13293849e-01 -3.70207042e-01
-3.32204835e-03 -7.93064177e-01 -9.67093229e-01 3.32381010e-01
-5.20274878e-01 2.31543541e-01 3.44024181e-01 4.36921537e-01
1.58912808e-01 -4.79403675e-01 6.18371665e-01 -7.34307408e-01
6.18257523e-01 -1.05284786e+00 -4.20507550e-01 6.19306043e-03
-5.85310519e-01 -1.16287887e+00 1.69530675e-01 -3.91970351e-02
-8.25361490e-01 -8.40064228e-01 -1.85025528e-01 7.07044154e-02
1.16297550e-01 7.91567326e-01 -1.81503758e-01 -3.64329398e-01
4.16737378e-01 -5.96432574e-02 1.96326569e-01 -2.70181775e-01
6.67523623e-01 1.21406591e+00 2.29381517e-01 -3.81861418e-01
4.77178454e-01 6.86947107e-01 -4.08533603e-01 -6.30529523e-01
1.97755948e-01 -2.90755123e-01 2.45003834e-01 -7.80574262e-01
6.70718968e-01 -7.26143837e-01 -1.29803705e+00 1.50410175e-01
-1.03839302e+00 -1.25513464e-01 -3.88208598e-01 5.74689150e-01
-4.42588888e-02 2.36856416e-01 -6.11746371e-01 -1.14901876e+00
-3.76116812e-01 -1.10685563e+00 4.29873616e-01 5.82151651e-01
-1.89363062e-01 -7.78993070e-01 2.77705528e-02 8.61517012e-01
8.50310624e-01 7.10042059e-01 2.34271154e-01 -4.55777586e-01
-1.12700522e+00 -3.71831506e-01 -2.06568077e-01 2.09257547e-02
-8.50789845e-02 6.94756329e-01 -8.87951851e-01 -1.51183948e-01
-4.82041501e-02 -3.22546298e-03 -4.09713797e-02 5.60099594e-02
6.73250258e-01 -5.13163984e-01 -4.44470137e-01 7.14578688e-01
1.48887849e+00 5.56033552e-01 1.14958763e+00 9.80445504e-01
4.17255133e-01 6.00952268e-01 8.84585738e-01 7.72614062e-01
9.35043097e-01 4.16932434e-01 1.00907505e+00 -1.53189853e-01
2.52230704e-01 -1.18478797e-01 2.26330042e-01 1.08153164e+00
-7.75040627e-01 -1.07154690e-01 -1.22930861e+00 3.96470219e-01
-1.69071651e+00 -8.35883081e-01 -5.46925008e-01 2.11329103e+00
8.88537988e-02 -1.72257587e-01 2.91888803e-01 -2.21975371e-01
7.35386729e-01 5.10878228e-02 -1.49516568e-01 -7.82562315e-01
1.77512988e-01 -2.01717541e-01 2.25237489e-01 -1.22038566e-01
-4.91860837e-01 9.37994480e-01 6.96372700e+00 3.32839072e-01
-1.53036296e+00 4.01628405e-01 4.27384526e-01 7.47791380e-02
-8.77365053e-01 4.74966019e-01 -2.17056964e-02 3.85941565e-01
9.24951315e-01 -3.58878791e-01 5.73893309e-01 7.26931870e-01
3.29034090e-01 1.01802379e-01 -3.92926693e-01 9.52759147e-01
-3.76694560e-01 -2.10692787e+00 -5.76078117e-01 7.37629890e-01
8.53674829e-01 2.98877805e-01 3.07263613e-01 1.54004946e-01
2.17801750e-01 -8.68685901e-01 1.02024603e+00 2.17816308e-01
1.04031265e+00 -7.73218036e-01 6.97063625e-01 1.00110359e-01
-9.75248396e-01 -3.10470641e-01 3.91085148e-01 -5.45516253e-01
2.88221985e-01 3.31710637e-01 -7.76790380e-02 1.01917541e+00
9.22881842e-01 4.06060398e-01 1.65079802e-01 1.23522651e+00
-7.54304975e-02 3.65374506e-01 -9.60079953e-02 -2.14161932e-01
-6.62759691e-02 -6.29501402e-01 7.20132530e-01 8.27275574e-01
4.39666331e-01 3.83886009e-01 -3.34314495e-01 6.81606352e-01
-1.43881232e-01 1.92608893e-01 -8.95905495e-01 -5.10409653e-01
1.14073992e+00 1.37109327e+00 -6.61573052e-01 -9.78231952e-02
-7.32915521e-01 3.55723143e-01 -3.58458161e-01 4.27905113e-01
-8.61417651e-01 -3.09360296e-01 1.00809932e+00 3.29336584e-01
-7.64515325e-02 -4.48247433e-01 -9.73409295e-01 -1.29890132e+00
-1.76544502e-01 -1.08722043e+00 1.43735781e-01 -5.03961325e-01
-5.47527552e-01 3.24097008e-01 -2.96500176e-01 -1.15996861e+00
-3.68775800e-02 -1.24471396e-01 -4.81400430e-01 3.80484939e-01
-1.34629190e+00 -1.36695910e+00 -3.16174537e-01 -7.64383450e-02
-3.43379021e-01 1.84111789e-01 7.37074792e-01 5.34056067e-01
-7.61877775e-01 1.72302455e-01 4.32994723e-01 -3.69325668e-01
4.27408904e-01 -5.69496334e-01 4.66568589e-01 7.35223234e-01
-5.00923157e-01 1.18684936e+00 6.18201792e-01 -6.30234420e-01
-2.13784933e+00 -3.25014681e-01 7.47267604e-01 -6.64554596e-01
8.27654183e-01 -5.44928551e-01 -2.25938037e-01 1.04024935e+00
5.21762669e-01 -2.16712862e-01 1.04465592e+00 4.03755516e-01
-2.13353708e-01 -9.74942092e-03 -1.39267337e+00 1.00251222e+00
1.25197816e+00 -5.54763138e-01 2.56117899e-02 -1.91954657e-01
7.36726284e-01 -4.88591075e-01 -1.20338428e+00 1.38400093e-01
1.14748812e+00 -1.30271256e+00 8.65853608e-01 -7.21351877e-02
3.45106542e-01 -2.96368420e-01 -1.35529056e-01 -5.50967097e-01
-9.90343392e-02 -1.35966361e+00 -3.23042005e-01 1.59949851e+00
-1.73641443e-02 -1.50080740e+00 4.89165783e-01 8.35252702e-01
-2.87304699e-01 -9.35822725e-01 -8.48989129e-01 -4.69955951e-01
-8.91425014e-02 -6.56466365e-01 1.30359411e+00 1.51921475e+00
7.45069563e-01 -1.90159492e-02 -2.86438972e-01 2.21352316e-02
2.53820777e-01 1.22486457e-01 1.19449401e+00 -1.04224122e+00
2.14763824e-02 -4.34682727e-01 -5.08011222e-01 -8.28627348e-01
-6.41143143e-01 -5.16866744e-01 -7.69943416e-01 -2.00341821e+00
1.38117954e-01 -4.80937690e-01 -7.22322147e-03 -1.76606551e-01
3.82517070e-01 3.90291423e-01 6.01460457e-01 4.87175077e-01
-7.08604515e-01 2.85034031e-01 1.52503848e+00 5.79362035e-01
-3.73961627e-01 -3.95506293e-01 -1.41604054e+00 5.68710208e-01
8.18918169e-01 1.30484790e-01 -3.03408355e-01 -2.50330269e-01
8.60179067e-01 3.22178364e-01 3.75774413e-01 -4.50300425e-01
3.34061295e-01 -5.30929327e-01 -3.20060700e-01 -3.44314396e-01
2.82111645e-01 -8.11310649e-01 8.56822550e-01 5.01050413e-01
7.78202474e-01 3.16626802e-02 -4.75345291e-02 2.02310104e-02
-1.09515227e-02 2.89820582e-01 1.28406927e-01 -5.90326637e-02
-5.99912167e-01 -1.16790399e-01 -6.04107261e-01 -2.39470243e-01
1.44201827e+00 -8.37607801e-01 -9.89634395e-01 -4.80470628e-01
-4.01546299e-01 2.69091755e-01 1.18504000e+00 2.78583527e-01
4.16468769e-01 -1.01407659e+00 -3.10311556e-01 5.98033704e-02
-1.33774623e-01 -4.20385242e-01 4.30237234e-01 1.02383757e+00
-7.65586853e-01 3.44809502e-01 -5.20290196e-01 -2.40590692e-01
-1.17418063e+00 3.53134006e-01 1.03634864e-01 -3.32278176e-03
-4.48846489e-01 3.35220069e-01 -8.19203723e-03 -3.54208559e-01
-4.53138351e-02 -1.39126834e-02 -4.68227565e-01 -5.39851300e-02
5.14808297e-01 7.46918142e-01 -4.07077253e-01 -7.04856336e-01
-5.13074338e-01 2.10075170e-01 3.51201236e-01 -3.87688696e-01
1.30341089e+00 -5.82093418e-01 -5.31872332e-01 3.74836773e-01
5.62862217e-01 5.03362358e-01 -7.32732236e-01 5.99139214e-01
-2.69852608e-01 -1.01240396e+00 1.22469268e-03 -8.19457173e-01
-9.46966112e-01 3.19088906e-01 4.07423407e-01 6.50135577e-01
9.15595829e-01 -1.44747794e-01 7.87810087e-01 -5.44233203e-01
6.78922892e-01 -1.13417101e+00 -1.28426149e-01 2.58632511e-01
7.83790946e-01 -5.47573924e-01 -3.08036804e-02 -1.05314803e+00
-4.98749882e-01 8.90541255e-01 5.01506567e-01 3.49675238e-01
7.50164449e-01 3.09832454e-01 1.32715732e-01 -3.55797917e-01
-6.60329580e-01 2.45200954e-02 -5.54469824e-01 8.81150961e-01
6.09221220e-01 3.64214659e-01 -1.03245330e+00 6.91124797e-01
-2.34891936e-01 4.13777918e-01 8.40522945e-01 1.19038606e+00
1.02176771e-01 -1.30047154e+00 -5.90898573e-01 1.40126020e-01
-8.16812277e-01 1.58843905e-01 -4.54965979e-01 1.09120727e+00
2.44418845e-01 1.14929533e+00 -1.98787779e-01 -7.54776657e-01
3.12956780e-01 -3.89367640e-01 5.76171465e-02 -1.99410632e-01
-1.15915930e+00 -6.29192144e-02 7.28874624e-01 -6.44154012e-01
-6.41162992e-02 -6.86175644e-01 -1.10752976e+00 -1.23924446e+00
-4.32709575e-01 -1.56560764e-02 1.41654158e+00 4.67986941e-01
1.02523816e+00 4.85684574e-01 4.40086842e-01 -4.55908448e-01
5.56307063e-02 -2.28772908e-01 -5.08455753e-01 -1.18466452e-01
5.65786771e-02 -3.76698166e-01 -1.73446491e-01 -6.75572872e-01]
|
[9.06573486328125, 6.541296482086182]
|
fdf58c3c-682c-41c3-97ab-9f4efed2647a
|
crossvqa-scalably-generating-benchmarks-for
| null | null |
https://aclanthology.org/2021.emnlp-main.164
|
https://aclanthology.org/2021.emnlp-main.164.pdf
|
CrossVQA: Scalably Generating Benchmarks for Systematically Testing VQA Generalization
|
One challenge in evaluating visual question answering (VQA) models in the cross-dataset adaptation setting is that the distribution shifts are multi-modal, making it difficult to identify if it is the shifts in visual or language features that play a key role. In this paper, we propose a semi-automatic framework for generating disentangled shifts by introducing a controllable visual question-answer generation (VQAG) module that is capable of generating highly-relevant and diverse question-answer pairs with the desired dataset style. We use it to create CrossVQA, a collection of test splits for assessing VQA generalization based on the VQA2, VizWiz, and Open Images datasets. We provide an analysis of our generated datasets and demonstrate its utility by using them to evaluate several state-of-the-art VQA systems. One important finding is that the visual shifts in cross-dataset VQA matter more than the language shifts. More broadly, we present a scalable framework for systematically evaluating the machine with little human intervention.
|
['Radu Soricut', 'Song-Chun Zhu', 'Piyush Sharma', 'Boqing Gong', 'Soravit Changpinyo', 'Arjun Akula']
| null | null | null | null |
emnlp-2021-11
|
['question-answer-generation']
|
['natural-language-processing']
|
[ 1.30848080e-01 6.62124436e-03 2.82367676e-01 -4.82817262e-01
-1.28174078e+00 -1.23132658e+00 6.43133163e-01 -1.85131654e-02
-1.13943927e-01 4.44493711e-01 2.43030056e-01 -4.68902946e-01
7.88011216e-03 -5.23439229e-01 -8.03462505e-01 -4.23238456e-01
3.89691204e-01 7.33657479e-01 1.76819474e-01 -5.03645241e-01
1.11337185e-01 1.35560155e-01 -1.63032448e+00 9.40404475e-01
9.95857179e-01 6.64463639e-01 -1.55631423e-01 9.78054941e-01
-1.92780316e-01 9.41399813e-01 -9.18869019e-01 -1.01097369e+00
2.68140703e-01 -8.16344619e-01 -1.13347363e+00 -2.18219236e-02
1.20878613e+00 -2.51582235e-01 -1.65143367e-02 7.65972793e-01
6.65323317e-01 -1.10901427e-02 9.37877119e-01 -1.60575211e+00
-1.26588833e+00 4.19390023e-01 -5.06183922e-01 4.43848133e-01
5.44876337e-01 6.84027195e-01 1.47988379e+00 -1.03075540e+00
9.03887808e-01 1.55706930e+00 3.88555706e-01 8.35911334e-01
-1.64356053e+00 -6.21283770e-01 -9.15322155e-02 3.97684842e-01
-1.22421467e+00 -3.47361952e-01 8.66895437e-01 -6.49962485e-01
6.34699643e-01 5.84054410e-01 2.17099071e-01 1.50825596e+00
-1.16082162e-01 9.91057813e-01 1.34655821e+00 -5.06787181e-01
1.88747942e-01 3.57082039e-01 2.02442065e-01 6.25139773e-01
-7.52807036e-02 -4.65211347e-02 -6.54306769e-01 -1.85652107e-01
2.04282746e-01 -5.31475604e-01 -6.07121468e-01 -9.13634837e-01
-1.24955583e+00 1.13515854e+00 6.64107323e-01 -5.55600338e-02
4.50197123e-02 -5.98580725e-02 4.04742390e-01 6.52756751e-01
2.31005117e-01 9.58965421e-01 -3.96730840e-01 1.48127481e-01
-5.73813736e-01 7.27752149e-01 5.26929021e-01 9.25692201e-01
7.03300297e-01 -2.00022966e-01 -8.71000946e-01 7.59010553e-01
1.38110101e-01 6.65977538e-01 3.06272417e-01 -1.01225317e+00
7.91496396e-01 8.58621597e-01 -1.68749709e-02 -7.02705204e-01
-1.45953670e-01 -7.37386495e-02 -4.85797375e-01 4.66676325e-01
7.33395100e-01 2.03202933e-01 -9.00858700e-01 2.06734753e+00
3.89610112e-01 -5.52958250e-01 2.67312199e-01 8.63606930e-01
1.40251422e+00 3.71778399e-01 2.89642125e-01 3.15633148e-01
1.62069476e+00 -9.22425807e-01 -4.77885932e-01 -2.83492327e-01
5.47119498e-01 -7.19791651e-01 1.93725944e+00 -1.36371600e-02
-9.71530497e-01 -7.52430201e-01 -1.01561153e+00 -5.09049833e-01
-4.64017034e-01 -1.14658982e-01 3.82054038e-02 5.19960403e-01
-9.26846683e-01 -2.75072102e-02 -1.14721172e-01 -2.83126086e-01
6.00662112e-01 -2.01114446e-01 -4.13042963e-01 -2.17907712e-01
-1.09695661e+00 1.02748871e+00 6.69205412e-02 -3.48821819e-01
-1.08689702e+00 -1.02745533e+00 -9.82525826e-01 -1.48401856e-02
2.30265886e-01 -1.08473122e+00 1.41105855e+00 -1.26899254e+00
-8.98709655e-01 1.40052640e+00 -1.13997206e-01 -7.82448053e-02
7.23940492e-01 1.03818670e-01 -2.39154607e-01 4.12302703e-01
2.57676452e-01 9.85836744e-01 9.45298910e-01 -1.64774442e+00
-2.25859344e-01 -6.39662981e-01 3.69726062e-01 2.11278826e-01
4.74025048e-02 -2.67297834e-01 -5.96892685e-02 -5.31544685e-01
-4.42496985e-01 -8.47892702e-01 1.66788191e-01 7.55772963e-02
-4.06019747e-01 -3.17697853e-01 6.70932770e-01 -7.06696272e-01
8.52387130e-01 -2.04773664e+00 4.29847032e-01 -7.26044131e-03
6.01873994e-01 1.44677430e-01 -5.61990261e-01 3.86112541e-01
-1.12451330e-01 1.89976037e-01 -3.44787806e-01 -2.56862879e-01
1.29093036e-01 1.06100462e-01 -8.06664705e-01 -3.60387862e-02
5.51956058e-01 1.47494531e+00 -9.25295830e-01 -6.38901412e-01
-2.50520468e-01 1.04620412e-01 -5.52600682e-01 5.00499964e-01
-5.47818184e-01 5.78806698e-01 -1.52589992e-01 5.14190495e-01
6.71236396e-01 -3.69074494e-01 -3.79560189e-03 -3.25200260e-01
2.95850277e-01 1.10697970e-01 -5.07682383e-01 1.58529472e+00
-2.71827012e-01 8.43076408e-01 -1.76123396e-01 -4.46301669e-01
8.31460118e-01 1.22543335e-01 -2.44582057e-01 -1.00244510e+00
-9.05590132e-02 -1.20043382e-01 -8.52326825e-02 -8.20059836e-01
5.33157647e-01 -2.95394212e-01 -2.43484557e-01 5.29532373e-01
4.87701148e-01 -4.58614945e-01 3.21554959e-01 6.14968479e-01
9.78165030e-01 7.98335448e-02 2.46109813e-01 -2.48884231e-01
3.71760637e-01 3.27729166e-01 1.27321765e-01 7.63422012e-01
-5.36707401e-01 9.63088751e-01 8.39536905e-01 -2.85872787e-01
-1.11490130e+00 -1.58428752e+00 6.70281202e-02 1.28944743e+00
1.77716574e-04 -2.68912375e-01 -7.02185392e-01 -1.08982849e+00
2.22857505e-01 9.16846693e-01 -1.13649821e+00 -3.27322364e-01
-2.32487112e-01 -1.96779341e-01 8.00588250e-01 5.12174010e-01
1.33378476e-01 -1.23669887e+00 -7.26727188e-01 -4.48858351e-01
-5.65939367e-01 -9.95011210e-01 -5.16032994e-01 -1.61518380e-01
-4.90319550e-01 -1.27204013e+00 -5.54462433e-01 -7.78878987e-01
4.53442365e-01 3.76881987e-01 2.03781724e+00 -8.77397805e-02
-3.22214901e-01 7.37871051e-01 -3.97234380e-01 -5.49429357e-01
-7.78720140e-01 1.07743330e-01 -6.07883751e-01 -6.62588328e-02
4.99587655e-01 -2.58133680e-01 -7.09345937e-01 2.32891545e-01
-1.11044490e+00 1.25519231e-01 2.86059290e-01 8.54041815e-01
4.84867990e-01 -8.70854557e-01 7.85166740e-01 -1.07465947e+00
9.68876183e-01 -5.03272772e-01 -3.93638760e-01 7.66964078e-01
-2.45636135e-01 2.24601492e-01 4.72784281e-01 -4.00584728e-01
-1.00993443e+00 -2.12836817e-01 1.40072584e-01 -5.62213123e-01
-1.96661904e-01 7.88161810e-03 -4.45632100e-01 2.12382823e-01
1.13225138e+00 -1.24121681e-02 -1.59530770e-02 -2.13637993e-01
1.12164640e+00 4.35193539e-01 6.91949129e-01 -6.34223759e-01
9.79860485e-01 4.80579615e-01 -3.49150062e-01 -4.27870244e-01
-9.93284523e-01 -2.98992664e-01 -5.48076570e-01 -1.87502399e-01
1.30812490e+00 -9.02894735e-01 -7.32713938e-01 5.56343654e-03
-1.23883533e+00 -3.92119467e-01 -5.15468895e-01 -3.08314651e-01
-5.19040585e-01 1.11828752e-01 -2.93238610e-01 -4.33256060e-01
-3.67487729e-01 -1.22763288e+00 1.15654421e+00 1.62204504e-01
-5.70684433e-01 -9.15409088e-01 5.92367649e-01 1.03935659e+00
2.74169594e-01 2.02989519e-01 1.33799398e+00 -7.58521795e-01
-5.99251330e-01 3.40630233e-01 -3.77494663e-01 2.81297624e-01
-9.80580375e-02 3.53661478e-02 -1.32343161e+00 -3.09693456e-01
-2.77561724e-01 -1.07143497e+00 8.10012102e-01 -1.02734961e-01
1.09883273e+00 -6.28821105e-02 1.42616346e-01 5.43304205e-01
1.14307880e+00 -2.63898462e-01 5.80477595e-01 1.06656857e-01
8.64472151e-01 8.75114560e-01 4.44918394e-01 1.87712200e-02
7.98983395e-01 6.30936861e-01 4.72509563e-01 -1.82537392e-01
-5.18076718e-01 -4.56310213e-01 2.03009710e-01 4.19544131e-01
3.89791071e-01 -3.07380259e-01 -1.05578387e+00 8.08729827e-01
-1.54984701e+00 -1.05528045e+00 -2.52117753e-01 1.87495887e+00
1.10336196e+00 -3.86497527e-01 1.34422988e-01 -1.89066499e-01
3.40936333e-01 1.82915211e-01 -5.19979715e-01 -3.95035595e-01
-3.09837371e-01 3.62973124e-01 -2.23722249e-01 4.96284753e-01
-8.24258566e-01 7.97647834e-01 6.33863020e+00 4.94335145e-01
-6.78254724e-01 7.51813650e-02 6.91957176e-01 -2.52329767e-01
-9.61125493e-01 -8.59575197e-02 -3.28194797e-01 4.96730842e-02
6.15659118e-01 -4.55477312e-02 2.55270839e-01 6.44099295e-01
-3.06255102e-01 1.64781556e-01 -1.40448809e+00 1.00872064e+00
4.61296767e-01 -1.22274923e+00 6.98172867e-01 -3.64923000e-01
7.51153648e-01 -2.07036376e-01 4.17871535e-01 6.15310133e-01
5.43336093e-01 -1.25302112e+00 8.69891107e-01 4.36052620e-01
8.49281490e-01 -6.26623750e-01 4.00339395e-01 -4.06178012e-02
-7.40732968e-01 3.58953141e-02 -2.24436954e-01 2.74761140e-01
-1.84373278e-02 1.34541065e-01 -8.64837945e-01 3.75520647e-01
9.78313327e-01 2.45615598e-02 -1.44878399e+00 5.29433846e-01
-4.44990218e-01 5.39883316e-01 4.71227258e-01 1.40135735e-02
-7.54097775e-02 5.17292647e-03 5.29394925e-01 9.35423791e-01
1.30656548e-02 -1.39921933e-01 -1.84576243e-01 1.22654366e+00
-4.64635879e-01 6.73970357e-02 -7.91778028e-01 1.07774809e-01
3.48629415e-01 1.25099635e+00 -1.63178742e-01 -3.35008413e-01
-5.01244664e-01 1.00429666e+00 8.61791790e-01 6.30183995e-01
-7.26932824e-01 -3.37948084e-01 7.22564936e-01 -9.15892795e-02
3.93375129e-01 2.68623888e-01 -3.56125087e-01 -1.37520564e+00
1.42015338e-01 -1.64366221e+00 8.19371521e-01 -1.34447753e+00
-1.80243683e+00 7.38979757e-01 -7.82698840e-02 -1.09054136e+00
-3.13588470e-01 -7.04425037e-01 -3.74329656e-01 9.44734275e-01
-1.25287664e+00 -1.39406872e+00 -7.24921405e-01 8.68856311e-01
5.41868806e-01 -2.19276518e-01 8.22510958e-01 -1.46112636e-01
-2.07219005e-01 8.77362072e-01 -2.56901175e-01 1.54995948e-01
1.18475044e+00 -1.58848894e+00 6.38220131e-01 8.28686059e-01
5.81824839e-01 4.39515412e-01 8.99099946e-01 -2.22232133e-01
-1.09190631e+00 -9.47727680e-01 7.48574376e-01 -1.53849816e+00
5.48994005e-01 -7.14220047e-01 -1.23208368e+00 7.93476641e-01
6.91476703e-01 1.53419659e-01 9.56273854e-01 1.86611284e-02
-1.22158074e+00 -9.12539512e-02 -1.02771926e+00 8.42980206e-01
1.00264132e+00 -8.83337796e-01 -1.02134275e+00 1.37727752e-01
9.28918362e-01 -3.46934050e-01 -5.21484852e-01 2.95216352e-01
3.11905116e-01 -1.20342219e+00 1.12125158e+00 -1.21374094e+00
8.37724626e-01 -3.56088400e-01 -2.48745799e-01 -1.58315289e+00
-3.77537191e-01 -2.30322063e-01 3.82559076e-02 1.47611678e+00
6.51979923e-01 -4.03728694e-01 3.87785643e-01 4.99115646e-01
2.29397759e-01 -3.46784800e-01 -8.82053673e-01 -3.97584200e-01
4.83384371e-01 -1.50885850e-01 7.03095496e-01 1.04760754e+00
-2.05879107e-01 8.86439085e-01 -7.00464323e-02 8.93187225e-02
5.29258311e-01 4.12993312e-01 1.24191093e+00 -9.05670822e-01
-4.96950090e-01 -4.04656887e-01 -3.04574907e-01 -7.36348033e-01
1.77238241e-01 -1.06002450e+00 -3.36720273e-02 -1.30202270e+00
5.90087295e-01 -9.59400367e-03 -2.19314367e-01 2.39905611e-01
-6.16179943e-01 3.05175453e-01 3.08237940e-01 4.56563523e-03
-8.70023847e-01 7.70666420e-01 1.46652639e+00 -4.73442495e-01
6.11727610e-02 -3.46959889e-01 -9.82335091e-01 2.81642139e-01
6.52765036e-01 -2.62092769e-01 -8.35898340e-01 -8.32916200e-01
4.52073365e-01 -1.48842722e-01 8.17520678e-01 -5.37893593e-01
-3.19119066e-01 -3.25530879e-02 4.94493574e-01 -4.53899771e-01
1.90700620e-01 -3.44394714e-01 -3.78235966e-01 6.03582412e-02
-7.23150551e-01 6.03498757e-01 4.01869893e-01 4.06798661e-01
-2.37013683e-01 -5.68580404e-02 7.26289213e-01 -3.64954844e-02
-5.85051298e-01 6.58331513e-02 -3.20585668e-02 9.85679269e-01
9.88956809e-01 2.84279913e-01 -1.03614140e+00 -6.80482626e-01
-3.64340901e-01 4.10509109e-01 5.83765447e-01 8.39962065e-01
6.42061532e-01 -1.51985526e+00 -1.02456510e+00 2.59206612e-02
1.08825660e+00 -2.69918054e-01 5.54638445e-01 3.05650473e-01
-4.94066089e-01 1.56373695e-01 -3.06715757e-01 -7.34401822e-01
-1.52295625e+00 8.09959829e-01 3.77544940e-01 -3.12615812e-01
-2.33116329e-01 1.05083609e+00 6.98653042e-01 -8.52886260e-01
-2.14031756e-01 -2.07405463e-01 -1.81852326e-01 8.61000940e-02
7.15332031e-01 1.81675792e-01 -3.03508416e-02 -7.24600434e-01
-3.63390416e-01 3.91808778e-01 -5.98468538e-03 -3.81709814e-01
9.33342874e-01 -6.46744370e-02 1.52173322e-02 5.53814292e-01
1.18377626e+00 2.21386373e-01 -1.15067601e+00 -1.73531249e-01
-2.32104033e-01 -4.69567120e-01 -6.65524602e-01 -1.00857830e+00
-7.66563952e-01 1.16817164e+00 7.14334548e-01 3.10313731e-01
8.74814510e-01 5.86584210e-01 2.92349845e-01 1.92050591e-01
-2.17933040e-02 -6.66282177e-01 4.69745517e-01 3.03621292e-01
1.43708718e+00 -1.54749298e+00 -3.54005456e-01 -1.45892382e-01
-1.14108133e+00 7.42939651e-01 1.08125579e+00 2.15721354e-01
2.02661864e-02 -4.32167977e-01 8.67689788e-01 -5.08194983e-01
-1.05331802e+00 -2.49531507e-01 7.77798355e-01 8.93390596e-01
4.54549044e-01 1.11545555e-01 1.73514947e-01 4.38012689e-01
-5.80111921e-01 -5.59724212e-01 4.27267224e-01 5.73835313e-01
2.39388039e-03 -8.86936784e-01 -4.06408638e-01 1.47866622e-01
-2.85535865e-02 -8.34094174e-03 -8.97711039e-01 9.55611825e-01
5.83929243e-03 1.10355031e+00 4.53688912e-02 -3.46947193e-01
6.77802801e-01 4.27484512e-01 8.28192353e-01 -4.23452258e-01
-4.76257771e-01 -8.91459882e-01 1.47191629e-01 -6.44025564e-01
-6.88166246e-02 -5.26523292e-01 -7.86763310e-01 -9.13218409e-02
1.70531601e-01 -6.73540458e-02 1.51641101e-01 8.29938650e-01
5.37455916e-01 5.07278979e-01 2.90695876e-01 -3.07857066e-01
-6.83838010e-01 -9.33600545e-01 -1.26042828e-01 1.43279827e+00
5.02791762e-01 -5.48151076e-01 -4.76356030e-01 4.01469320e-02]
|
[10.85538101196289, 1.8963358402252197]
|
a367195c-67dc-4be1-8065-21a5882193ab
|
artificial-intelligence-moral-agent-as-adam
|
2305.11519
| null |
https://arxiv.org/abs/2305.11519v1
|
https://arxiv.org/pdf/2305.11519v1.pdf
|
Artificial intelligence moral agent as Adam Smith's impartial spectator
|
Adam Smith developed a version of moral philosophy where better decisions are made by interrogating an impartial spectator within us. We discuss the possibility of using an external non-human-based substitute tool that would augment our internal mental processes and play the role of the impartial spectator. Such tool would have more knowledge about the world, be more impartial, and would provide a more encompassing perspective on moral assessment.
|
['Nikodem Tomczak']
|
2023-05-19
| null | null | null | null |
['philosophy']
|
['miscellaneous']
|
[-1.17224693e-01 7.54765332e-01 1.38278361e-02 -2.99830109e-01
2.47151315e-01 -7.69246042e-01 7.92209446e-01 6.47395104e-02
-6.98721707e-01 9.47526038e-01 7.58163691e-01 -3.25329393e-01
-1.29228488e-01 -7.41751611e-01 -1.39456421e-01 -6.36394083e-01
5.47393858e-01 3.93673927e-01 -3.11184227e-01 -6.15965128e-01
9.79171813e-01 2.64520377e-01 -5.27686834e-01 4.21549939e-02
8.54494035e-01 1.25391141e-01 -6.00207269e-01 3.36657256e-01
8.40258062e-01 1.50045598e+00 -7.14066923e-01 -1.02387631e+00
2.31919512e-01 -5.67620635e-01 -1.06465638e+00 -3.54030579e-01
-3.15512449e-01 -7.31479466e-01 -7.31082484e-02 1.12214994e+00
4.45136070e-01 2.08374500e-01 4.85819131e-01 -1.02610183e+00
-1.13762832e+00 6.70279443e-01 4.39272560e-02 1.58314988e-01
6.03140295e-01 7.18689203e-01 5.15634894e-01 1.11484915e-01
7.82253802e-01 1.69069302e+00 3.95767391e-01 9.87685800e-01
-1.22445750e+00 -8.70281637e-01 -5.54469287e-01 2.24907011e-01
-6.25289202e-01 -4.22917008e-01 8.47366691e-01 -4.44557637e-01
4.61454034e-01 2.35487163e-01 9.84087586e-01 1.26547229e+00
8.64009082e-01 1.77638233e-01 1.53674936e+00 -1.85954958e-01
4.61735636e-01 3.98140043e-01 2.41729543e-01 -9.61971730e-02
5.70551634e-01 3.45352501e-01 -4.40372378e-01 -7.57644296e-01
9.14168298e-01 -3.88850242e-01 -2.93763936e-01 5.85621316e-03
-1.14574802e+00 1.05141342e+00 3.20154697e-01 5.48927307e-01
-6.68323338e-01 3.80934894e-01 6.93826497e-01 3.88941675e-01
-3.02286576e-02 1.22625089e+00 1.16610378e-01 -6.01313710e-01
-2.29526371e-01 3.13161969e-01 9.55474854e-01 -3.50920409e-01
-1.44024968e-01 -2.17492148e-01 -1.48747504e-01 -1.36543527e-01
4.96536762e-01 3.86601955e-01 2.12667763e-01 -1.94476664e+00
-1.72963753e-01 4.25552189e-01 5.40243089e-01 -1.04031479e+00
-3.05777669e-01 -3.64782631e-01 -3.48736681e-02 8.85711312e-01
3.92262489e-01 1.82158574e-02 -2.64455825e-01 1.70296681e+00
2.79934704e-01 -6.20066881e-01 2.45836154e-01 1.07418334e+00
-2.21579429e-02 2.89626986e-01 4.27709311e-01 -1.01177864e-01
1.07909739e+00 -1.74767688e-01 -1.01954186e+00 -1.13892838e-01
7.20644891e-01 -6.00721240e-01 9.80877817e-01 5.51358759e-01
-1.10541749e+00 2.29267597e-01 -1.15410507e+00 1.35410652e-01
2.67028809e-03 -7.39646733e-01 9.50450838e-01 1.08810925e+00
-9.54722345e-01 8.97133112e-01 -4.93016422e-01 -5.81511915e-01
6.68495595e-01 -2.82759458e-01 -3.34637851e-01 3.13092768e-01
-1.30969572e+00 1.51517868e+00 4.72032756e-01 2.01656912e-02
-7.23085463e-01 -2.34104291e-01 -5.69324911e-01 -1.58157557e-01
2.44623289e-01 -9.02997255e-01 8.26407194e-01 -1.18391657e+00
-1.26217699e+00 1.47719204e+00 4.12649989e-01 -6.73588932e-01
9.59821343e-01 -1.71250060e-01 -5.18670022e-01 3.05166095e-01
4.12820339e-01 2.22850144e-01 3.33161831e-01 -1.27070439e+00
2.27352858e-01 -6.96412086e-01 3.41011316e-01 4.74360824e-01
1.01531595e-01 4.30851728e-01 1.18458414e+00 -3.67967099e-01
-5.55809848e-02 -6.45644605e-01 -2.22197548e-02 -1.88181147e-01
-1.61239043e-01 -2.79094249e-01 5.94161570e-01 -5.80614090e-01
9.25208986e-01 -2.26722836e+00 -6.02496386e-01 -3.17319445e-02
3.22591752e-01 3.12807947e-01 5.28730869e-01 9.55845833e-01
-9.75793600e-03 6.24021292e-01 1.13263719e-01 4.73084182e-01
3.40383321e-01 1.02951787e-01 -3.25780034e-01 8.96030962e-01
-2.55328536e-01 7.25914955e-01 -1.15213442e+00 -3.12495291e-01
8.05706978e-02 2.10453838e-01 -2.34394506e-01 -2.26754248e-01
7.20638812e-01 5.78169882e-01 -3.82067591e-01 4.34359372e-01
4.54927266e-01 -1.58940663e-03 2.44047552e-01 6.39065742e-01
3.73674966e-02 5.13634324e-01 -4.50469673e-01 8.09504688e-01
1.18695945e-01 7.97744811e-01 3.13520610e-01 -4.06994373e-01
9.59860563e-01 6.37946665e-01 -4.15655881e-01 -8.54240358e-01
6.52458727e-01 8.99244547e-02 4.34454739e-01 -4.56751138e-01
4.68097717e-01 -1.20842063e+00 -5.93515486e-02 1.16441596e+00
-4.72564399e-01 -3.61026645e-01 -5.20172298e-01 3.34915817e-01
9.67183650e-01 4.82960522e-01 6.03649378e-01 -7.59981513e-01
2.69628823e-01 1.21931046e-01 8.85832012e-01 4.18231934e-01
-1.21381748e+00 -2.75667101e-01 7.36389756e-01 -8.90553236e-01
-1.08487785e+00 -9.73669410e-01 -5.17498732e-01 7.85418093e-01
3.84068310e-01 3.83624673e-01 -8.90235543e-01 -4.08602268e-01
-1.75258219e-01 2.13810825e+00 -6.80744469e-01 -7.06241429e-01
-1.89370334e-01 -5.32437325e-01 7.19381452e-01 1.34549469e-01
4.43093300e-01 -1.21679962e+00 -1.57014418e+00 -5.70984744e-02
-2.18008265e-01 -4.75884616e-01 1.70603082e-01 -7.85737485e-02
-7.97984123e-01 -9.31101084e-01 -1.03641205e-01 2.80507565e-01
4.72347170e-01 -2.07261667e-01 7.60578156e-01 -4.77609634e-02
2.65217096e-01 1.01646967e-01 -2.44959816e-01 -8.08935106e-01
-1.14069450e+00 -1.22235370e+00 2.47788563e-01 -3.40199947e-01
6.12668335e-01 -6.66892946e-01 -5.02086341e-01 1.22409493e-01
-5.20734489e-01 -3.37589881e-03 -5.97025976e-02 2.98919410e-01
-4.63086694e-01 -1.99299812e-01 7.81837463e-01 -8.91628027e-01
9.79939878e-01 -3.70693207e-01 1.21083431e-01 -2.07060754e-01
-5.02208173e-01 -3.24212253e-01 5.30312002e-01 -1.34178266e-01
-1.40418315e+00 -1.04143786e+00 4.88528758e-01 2.42666706e-01
1.58734489e-02 -2.60930270e-01 3.90021354e-02 -1.58792019e-01
1.08553195e+00 -3.56869608e-01 7.08020031e-02 3.89234871e-02
-1.05830006e-01 4.17547017e-01 5.40239215e-01 -6.08838558e-01
5.32826245e-01 7.05558479e-01 -5.04414774e-02 -1.88662961e-01
-5.11637509e-01 1.71881095e-02 -4.47941810e-01 -6.39753699e-01
8.40214789e-01 -6.22130394e-01 -1.21016121e+00 1.62560359e-01
-1.11246371e+00 -3.63680154e-01 -2.72157937e-01 8.11145782e-01
-7.12213755e-01 4.34024900e-01 -3.79643828e-01 -1.19303310e+00
-3.04171890e-01 -4.19477910e-01 -1.51234092e-02 2.74826676e-01
-1.02371621e+00 -1.04361033e+00 2.41009369e-01 9.44318712e-01
3.21553737e-01 9.95469987e-01 6.97900116e-01 -1.08507836e+00
1.02431558e-01 -5.59568524e-01 2.09436283e-01 4.60198492e-01
1.34642534e-02 -1.17743656e-01 -7.95897424e-01 1.14146158e-01
1.09547567e+00 -6.13939524e-01 9.40851271e-02 -1.18731610e-01
2.12472051e-01 -7.00092137e-01 -1.48554519e-01 1.50388494e-01
1.10185063e+00 7.20491767e-01 1.14495862e+00 8.94994557e-01
9.13161132e-03 1.17209780e+00 1.00720453e+00 4.26601768e-01
3.32667083e-01 1.48624480e-01 4.60226059e-01 2.88179785e-01
5.34612894e-01 -1.85039327e-01 5.02542436e-01 5.87832592e-02
-5.56768179e-01 7.88562000e-02 -1.23087955e+00 5.23227692e-01
-1.65178263e+00 -1.31574106e+00 -2.76818067e-01 1.94077122e+00
6.06299639e-01 3.35532933e-01 1.52377244e-02 7.48197362e-02
8.92299592e-01 -9.57472716e-03 -5.10862708e-01 -1.35790896e+00
9.67513546e-02 -3.10606897e-01 4.02414322e-01 3.47738534e-01
-6.71897352e-01 8.29087794e-01 7.79787779e+00 2.17779681e-01
-2.95471281e-01 3.28716755e-01 9.78325784e-01 -1.93471685e-01
-6.48448408e-01 4.41887826e-01 -7.17798155e-03 4.11721826e-01
7.47205198e-01 -8.27439308e-01 2.08626762e-02 7.24601507e-01
5.12777090e-01 -6.64273977e-01 -1.13155818e+00 4.85374272e-01
-1.27532601e-01 -8.30643892e-01 -2.07658768e-01 4.20284152e-01
4.40818876e-01 -6.45832837e-01 -2.27657810e-01 -1.21640399e-01
8.41616869e-01 -1.23542452e+00 1.02568603e+00 5.89809954e-01
2.86042720e-01 -9.14038241e-01 1.29946816e+00 5.29662669e-01
3.00718278e-01 -2.68633589e-02 -1.95508704e-01 -1.05343103e+00
3.54605496e-01 -1.17701530e-01 -5.67236722e-01 -1.64966568e-01
3.80309939e-01 -9.26913098e-02 -4.86243255e-02 3.84391308e-01
-4.80660826e-01 4.71672893e-01 -2.10594088e-02 -1.89872891e-01
4.97682065e-01 -3.71439815e-01 1.06516099e+00 5.92523575e-01
-5.47987223e-01 4.73191768e-01 -8.16914797e-01 1.13379359e+00
4.03470308e-01 -5.00420965e-02 -9.04393971e-01 -1.43185511e-01
4.04381186e-01 1.00136673e+00 -8.81993234e-01 -2.10810468e-01
4.03344512e-01 6.96896076e-01 -3.02782923e-01 -3.17985266e-02
-7.49779284e-01 -1.45171866e-01 5.49431264e-01 2.31606781e-01
-7.27367043e-01 8.30217078e-02 -7.23650813e-01 -8.42716992e-01
-4.59407687e-01 -1.00626695e+00 1.59795582e-01 -8.44219148e-01
-1.06917179e+00 6.78723752e-02 -1.33732229e-01 -5.03998160e-01
-4.03222814e-02 -4.42242742e-01 -1.08391798e+00 8.30489397e-01
-3.32132190e-01 -9.54630613e-01 2.26028547e-01 1.49516642e-01
-2.88474411e-01 1.79309264e-01 6.00927413e-01 -5.79396546e-01
1.68862641e-02 1.87161013e-01 -3.02604318e-01 -6.16535693e-02
8.70251238e-01 -7.68769681e-01 7.47695863e-02 8.16102445e-01
-8.53192270e-01 9.52305734e-01 1.11925471e+00 -8.28484178e-01
-8.69594157e-01 -7.60803744e-02 7.23481774e-01 -1.05899906e+00
1.11785972e+00 -5.99703798e-03 -6.44904375e-01 7.44233727e-01
3.89788836e-01 -9.37988639e-01 1.08482790e+00 1.14151366e-01
-8.88558552e-02 3.60484570e-01 -1.85420418e+00 9.22959566e-01
9.22448397e-01 -7.14938700e-01 -1.37786353e+00 1.38776869e-01
8.07922423e-01 3.57107192e-01 -7.09658384e-01 -1.39632419e-01
6.33981764e-01 -1.10647380e+00 8.27778220e-01 -5.98296881e-01
5.00315130e-01 -1.16564944e-01 -4.75369254e-03 -7.53960013e-01
-5.16344965e-01 -9.35394049e-01 6.97586000e-01 9.69107389e-01
1.11674979e-01 -1.51422703e+00 2.72854120e-01 1.44169128e+00
6.97465017e-02 -4.78847414e-01 -1.07205558e+00 -5.13556123e-01
3.95969123e-01 -1.30089656e-01 4.76167172e-01 1.61393940e+00
1.07301223e+00 5.74688353e-02 -3.62252444e-01 -3.66832167e-01
9.18252409e-01 -4.38197255e-01 4.78856474e-01 -1.10824120e+00
-9.68165845e-02 -1.29565403e-01 -7.24497080e-01 2.85082757e-01
-1.28454477e-01 -7.85090864e-01 -3.34092438e-01 -1.57278681e+00
4.12367225e-01 -1.33993449e-02 -8.97159502e-02 3.99860084e-01
6.09608665e-02 4.45827544e-02 6.75105453e-01 4.48098958e-01
-3.35383296e-01 5.54818094e-01 1.61319768e+00 6.57054961e-01
-2.36753896e-01 -6.34595335e-01 -1.74897397e+00 1.36527824e+00
1.11158705e+00 -7.25566685e-01 -5.43408878e-02 1.86589703e-01
6.14868224e-01 1.78146109e-01 7.53279328e-01 -7.58585334e-01
7.68205002e-02 -8.37708652e-01 3.58672500e-01 -1.06205260e-02
2.10659325e-01 -4.68398869e-01 4.38583374e-01 1.06418121e+00
-2.21724838e-01 -4.06531334e-01 7.93702062e-03 1.13856681e-01
2.21929178e-01 -5.22178888e-01 1.00346947e+00 -4.10196990e-01
-1.27010867e-01 -8.02796423e-01 -5.10604620e-01 -1.88978195e-01
1.53871167e+00 -6.81040287e-01 -9.95033443e-01 -5.78904867e-01
-6.41170263e-01 2.27596506e-01 1.37160635e+00 -1.07001595e-01
1.85633242e-01 -1.05432117e+00 -8.12760115e-01 -7.25419760e-01
-3.91570538e-01 -9.18221116e-01 4.47096266e-02 6.81640983e-01
-1.00398445e+00 4.82820868e-01 -8.31458092e-01 3.42528671e-01
-7.59028256e-01 5.83449244e-01 4.25356358e-01 1.91901296e-01
-6.83421552e-01 9.56470072e-01 7.56933928e-01 -1.44116834e-01
-7.40107119e-01 8.78796935e-01 -4.20935541e-01 2.03075156e-01
6.53679013e-01 9.97384489e-01 -8.92710626e-01 -9.44491982e-01
-7.17883110e-01 1.25586286e-01 2.29828358e-01 -5.99945426e-01
1.24251401e+00 -3.15220147e-01 -7.15889037e-01 5.09870529e-01
4.35757637e-01 5.05908839e-02 -9.83100116e-01 6.49893403e-01
-3.04449528e-01 -1.12078178e+00 -1.02755569e-01 -1.26550877e+00
6.41328618e-02 6.41531944e-01 1.57759506e-02 3.51083010e-01
8.18888903e-01 -1.68670505e-01 2.41525218e-01 3.14901799e-01
8.92464042e-01 -1.64108074e+00 7.20146522e-02 -6.68835640e-02
1.11161339e+00 -7.56918311e-01 1.59747854e-01 -1.39874682e-01
-1.38360000e+00 9.81910169e-01 4.56860721e-01 -3.66391093e-01
3.82378437e-02 -1.06304623e-01 3.43171418e-01 -4.01668727e-01
-7.98170626e-01 6.41866386e-01 -3.90961140e-01 8.54699194e-01
3.44295025e-01 7.02593625e-01 -1.24719679e+00 6.78649843e-01
-5.32832682e-01 1.12587407e-01 1.25003707e+00 7.31183827e-01
-4.40593928e-01 -7.83911169e-01 -9.52829003e-01 -7.40581676e-02
-6.15570486e-01 5.61590642e-02 -9.09903586e-01 7.58060634e-01
1.97767511e-01 1.23355556e+00 -2.83794731e-01 1.06769232e-02
-4.89390232e-02 1.77169740e-01 2.88519979e-01 -2.28170812e-01
-8.29900563e-01 -1.09686397e-01 6.46299124e-01 -9.54634249e-01
-4.48660970e-01 -9.57354903e-01 -1.18580544e+00 -1.22658658e+00
5.02787866e-02 -4.09376845e-02 3.04857373e-01 8.99372995e-01
-8.08788987e-04 8.71817842e-02 3.31388414e-01 -2.91384190e-01
-6.09764516e-01 -8.49270821e-01 -6.73908949e-01 3.63237411e-01
-1.46127090e-01 -3.55478585e-01 -8.03290367e-01 -3.45867991e-01]
|
[9.035667419433594, 6.2300543785095215]
|
479a3a58-0c01-4cd3-bf78-6e08d4474eea
|
abstractive-text-summarization-using-the-brio
|
2305.13696
| null |
https://arxiv.org/abs/2305.13696v1
|
https://arxiv.org/pdf/2305.13696v1.pdf
|
Abstractive Text Summarization Using the BRIO Training Paradigm
|
Summary sentences produced by abstractive summarization models may be coherent and comprehensive, but they lack control and rely heavily on reference summaries. The BRIO training paradigm assumes a non-deterministic distribution to reduce the model's dependence on reference summaries, and improve model performance during inference. This paper presents a straightforward but effective technique to improve abstractive summaries by fine-tuning pre-trained language models, and training them with the BRIO paradigm. We build a text summarization dataset for Vietnamese, called VieSum. We perform experiments with abstractive summarization models trained with the BRIO paradigm on the CNNDM and the VieSum datasets. The results show that the models, trained on basic hardware, outperform all existing abstractive summarization models, especially for Vietnamese.
|
['Jugal Kalita', 'Khang Thua Pham', 'Thieu Gia Doan', 'Khang Nhut Lam']
|
2023-05-23
| null | null | null | null |
['abstractive-text-summarization', 'text-summarization']
|
['natural-language-processing', 'natural-language-processing']
|
[ 1.75608054e-01 2.67835021e-01 -4.23985600e-01 -4.45398539e-01
-1.04759824e+00 -4.73585755e-01 6.82860494e-01 4.59124863e-01
-4.05855745e-01 9.68738854e-01 1.01765001e+00 -2.71479458e-01
2.54611254e-01 -7.42649555e-01 -7.00573802e-01 -1.37055144e-01
1.25360668e-01 4.89146382e-01 6.31824955e-02 -2.14341298e-01
7.55986869e-01 2.89309472e-01 -1.13362193e+00 5.99048376e-01
1.33274198e+00 4.37121272e-01 1.30699262e-01 1.52489400e+00
-2.74993688e-01 1.10088599e+00 -1.67777181e+00 -3.92848998e-01
-2.04117566e-01 -6.04110241e-01 -7.82009184e-01 -3.17567348e-01
1.02590966e+00 -7.90870190e-01 -7.09874392e-01 8.02047431e-01
8.53276014e-01 2.86803365e-01 8.46842051e-01 -8.24839711e-01
-8.86239529e-01 1.31518734e+00 -3.29069704e-01 5.60427189e-01
2.73730814e-01 3.50452284e-03 1.10341048e+00 -5.51907539e-01
5.62983155e-01 1.22456563e+00 6.68316960e-01 7.51679778e-01
-9.25674319e-01 -4.37740803e-01 3.47419977e-02 -4.41015996e-02
-1.05602646e+00 -9.88149166e-01 4.56382453e-01 2.77874395e-02
1.71983254e+00 7.05146730e-01 5.13382852e-01 9.66564715e-01
7.26958513e-01 1.19552648e+00 1.33090556e-01 -2.63092041e-01
2.06778392e-01 -3.63735676e-01 7.06131458e-01 6.25208795e-01
7.91456044e-01 -5.47153473e-01 -7.78850913e-01 -3.14436108e-01
9.00716186e-02 -2.80298799e-01 -2.06003994e-01 7.10326433e-01
-9.25177693e-01 8.04194331e-01 1.16934299e-01 -2.38766987e-02
-4.32643920e-01 5.40850580e-01 1.05370915e+00 1.90257356e-01
6.55080020e-01 7.22393870e-01 -2.36268416e-01 -2.84994692e-01
-1.68734956e+00 5.74549496e-01 1.18537724e+00 1.56980467e+00
4.22339559e-01 5.66401303e-01 -9.78782773e-01 8.25430572e-01
-1.48450121e-01 7.11624026e-01 7.58323550e-01 -1.03931093e+00
9.06109035e-01 2.05524907e-01 -1.66250318e-01 -9.22588408e-01
-2.78178304e-01 -5.86745918e-01 -1.18300569e+00 -5.51701784e-01
-4.97083604e-01 -4.86600161e-01 -9.90050375e-01 1.21492648e+00
-4.86612737e-01 -2.98118800e-01 4.79874730e-01 1.75098702e-01
1.55314553e+00 1.26931489e+00 -1.75998300e-01 -4.25440937e-01
7.87694633e-01 -1.39050090e+00 -1.06106901e+00 -3.02361995e-01
6.98497891e-01 -5.91407657e-01 8.18944037e-01 3.87175471e-01
-1.64019728e+00 -6.08220577e-01 -1.41095161e+00 -6.58191383e-01
-1.51649751e-02 4.92851704e-01 4.64796007e-01 4.61374640e-01
-1.29613137e+00 9.31611657e-01 -1.01490283e+00 -5.49934149e-01
5.34648657e-01 2.20573723e-01 7.31969923e-02 2.96696782e-01
-8.72490287e-01 7.70614922e-01 8.86561215e-01 2.37327181e-02
-6.96292341e-01 -5.55527210e-01 -9.99579728e-01 7.15680957e-01
2.42008530e-02 -1.26914597e+00 1.84389889e+00 -2.83137113e-01
-1.77903342e+00 1.58253521e-01 -4.29856241e-01 -1.03649724e+00
1.86093122e-01 -5.06595016e-01 -1.67523533e-01 4.13365573e-01
1.16029382e-01 7.23228037e-01 5.16158223e-01 -8.98710966e-01
-7.38936126e-01 2.59603690e-02 -1.69300273e-01 2.34480754e-01
-4.79314804e-01 -2.08807051e-01 -4.68302071e-01 -8.06985676e-01
-5.92292368e-01 -4.73470867e-01 -1.90409690e-01 -7.52604425e-01
-1.13543665e+00 -3.32554221e-01 7.00765729e-01 -9.41981077e-01
1.96280324e+00 -1.50610149e+00 -1.34817794e-01 -2.66979009e-01
1.96455106e-01 7.11630642e-01 -1.44077525e-01 8.76489758e-01
6.07078671e-01 3.53294313e-01 -3.36128473e-01 -6.17232561e-01
1.58338860e-01 2.35353440e-01 -6.63406491e-01 -2.58227065e-02
2.12484896e-01 1.04918408e+00 -8.10704052e-01 -9.74015534e-01
-9.59947109e-02 -1.32016972e-01 -7.60846734e-01 3.11164528e-01
-5.25528908e-01 -2.61852741e-01 -5.60780585e-01 1.78944007e-01
4.87583190e-01 6.72037601e-02 1.00278519e-01 -6.77367076e-02
-2.69638687e-01 8.07444692e-01 -3.20789844e-01 1.93532562e+00
-3.30777287e-01 1.14028525e+00 -3.06267142e-01 -7.64323056e-01
1.07576406e+00 1.51017293e-01 -1.49291173e-01 -5.02789557e-01
2.43301809e-01 1.86512306e-01 -3.15671265e-01 -5.36221087e-01
1.58340251e+00 3.95371675e-01 -3.03915381e-01 5.59921026e-01
2.18750447e-01 -6.03277266e-01 8.37899685e-01 8.13169003e-01
1.25990725e+00 -2.66431212e-01 5.62183917e-01 -2.99937814e-01
2.49411508e-01 3.45325112e-01 4.45591301e-01 1.36051440e+00
3.02606791e-01 8.35000515e-01 6.85834706e-01 1.03233326e-02
-1.06950593e+00 -1.07431400e+00 1.94083780e-01 8.43614817e-01
-2.93606907e-01 -9.71136391e-01 -9.09616590e-01 -6.65119588e-01
-2.45135397e-01 1.48210323e+00 -2.54858583e-01 -4.06331658e-01
-9.17814970e-01 -5.70100486e-01 1.19360161e+00 8.59682143e-01
8.70400131e-01 -9.22804415e-01 -7.15738416e-01 3.04156125e-01
-2.50658333e-01 -7.42169201e-01 -9.18739974e-01 1.94793493e-01
-1.13132131e+00 -3.61377686e-01 -6.58565879e-01 -5.92725515e-01
1.52895555e-01 1.93303138e-01 1.28116477e+00 9.74504426e-02
1.64419159e-01 2.03017592e-01 -2.37419024e-01 -7.73022354e-01
-7.82971501e-01 9.64676023e-01 -2.72426963e-01 -7.97455013e-01
2.66326144e-02 -4.91697997e-01 -3.53123218e-01 -7.36900985e-01
-9.66369331e-01 1.45302847e-01 7.41514921e-01 7.59553432e-01
3.00427854e-01 -2.84562737e-01 1.05734599e+00 -1.16129065e+00
1.28301394e+00 -1.67769358e-01 -7.15436116e-02 4.14332926e-01
-5.34051836e-01 2.71348596e-01 8.82911325e-01 -1.55798271e-01
-1.23571622e+00 -5.00399947e-01 -3.75982612e-01 -2.16161199e-02
3.33666384e-01 9.14966822e-01 -2.63771061e-02 7.63933420e-01
8.39646637e-01 3.37530285e-01 -2.03482002e-01 -2.80231744e-01
4.27798569e-01 1.02704191e+00 1.02540982e+00 -3.63068342e-01
3.31897795e-01 1.36048257e-01 -4.41796780e-01 -1.29580665e+00
-9.47796047e-01 -2.73473948e-01 -3.26147318e-01 1.77919134e-01
6.46470547e-01 -8.05839181e-01 -2.69327313e-01 2.24754184e-01
-1.72269225e+00 -4.00458515e-01 -4.90301311e-01 2.93100089e-01
-4.64090317e-01 6.58838391e-01 -8.68968368e-01 -3.10701251e-01
-1.47713923e+00 -6.28911614e-01 1.08760226e+00 5.64761996e-01
-7.77936876e-01 -9.88774061e-01 2.23656297e-01 -8.35406408e-02
5.43542862e-01 8.45669582e-02 1.02573085e+00 -1.02588069e+00
-2.67998368e-01 -3.72582346e-01 -1.73525751e-01 6.43361032e-01
-9.95573476e-02 5.21137774e-01 -7.41397798e-01 -1.03015162e-01
-1.46803215e-01 -1.66300789e-01 1.46992242e+00 7.04532027e-01
1.22208703e+00 -8.94537628e-01 -2.09608734e-01 4.66217041e-01
9.37274396e-01 9.15656704e-03 6.87573612e-01 -3.30149606e-02
5.43517053e-01 1.09417088e-01 4.13613498e-01 5.86639285e-01
4.95665908e-01 5.94912209e-02 -7.94433728e-02 2.58381993e-01
-4.93035436e-01 -3.54405791e-01 4.98412877e-01 1.66909158e+00
2.04704046e-01 -1.11559796e+00 -5.23137450e-01 5.89399099e-01
-1.98565817e+00 -1.26884317e+00 -4.24518913e-01 1.64933801e+00
1.03435457e+00 3.55159491e-01 -2.90625930e-01 -2.33626783e-01
5.18595934e-01 6.80389881e-01 -5.18345535e-01 -1.30534220e+00
-2.12964386e-01 3.92041743e-01 5.73201895e-01 5.29965937e-01
-8.24121594e-01 9.75686729e-01 7.37468672e+00 8.83620322e-01
-1.05929399e+00 -1.45962358e-01 3.34915638e-01 -4.08356696e-01
-4.08364415e-01 -2.67233014e-01 -9.84964967e-01 1.61505610e-01
1.39670467e+00 -9.71584022e-01 -2.32496545e-01 8.50947499e-01
3.00027996e-01 -2.45177478e-01 -1.24815321e+00 7.02773631e-01
6.16637051e-01 -2.00259590e+00 6.77789688e-01 -3.09077710e-01
1.03557944e+00 1.17925830e-01 -2.58193374e-01 6.93148434e-01
4.16414678e-01 -8.32884729e-01 8.18353117e-01 7.12688446e-01
6.22915804e-01 -8.65538538e-01 8.81482303e-01 4.37197983e-01
-6.45217836e-01 2.20667392e-01 -7.98823297e-01 -6.28712252e-02
2.05341533e-01 4.92795676e-01 -9.59354997e-01 6.82831705e-01
2.13626981e-01 9.07128692e-01 -7.83420503e-01 9.55698311e-01
-2.53069580e-01 8.93082857e-01 2.87346658e-03 -6.04910493e-01
2.86679953e-01 2.19012052e-01 6.73589766e-01 1.89474761e+00
4.50253099e-01 4.24494110e-02 -5.27234487e-02 6.42261505e-01
-4.51557934e-01 -1.06347382e-01 -4.33848709e-01 -2.06728488e-01
4.67537463e-01 8.10389280e-01 -3.44511122e-01 -1.15486288e+00
-5.42148948e-02 9.58711088e-01 1.69013098e-01 4.39628810e-01
-8.64182413e-01 -1.04265940e+00 7.19259083e-02 -3.12646836e-01
2.90108681e-01 -3.25074226e-01 -7.35332191e-01 -1.37912643e+00
-1.00132182e-01 -7.81985879e-01 2.86226928e-01 -8.41914535e-01
-8.88891816e-01 7.44230211e-01 4.04125005e-01 -7.13687360e-01
-5.88872790e-01 -1.35022458e-02 -1.17406905e+00 7.24986792e-01
-1.31555557e+00 -8.76542687e-01 -3.05119455e-01 2.52288058e-02
1.21796942e+00 -2.20884651e-01 7.98084080e-01 -3.34539264e-02
-8.45941067e-01 6.84313178e-01 3.81032228e-01 -1.26178280e-01
7.10232079e-01 -1.29068995e+00 8.80692124e-01 1.00607193e+00
-2.62488395e-01 8.82610857e-01 1.06432176e+00 -8.97456229e-01
-1.31297398e+00 -1.43994379e+00 1.25302100e+00 -1.77374855e-01
2.74991781e-01 -8.79859254e-02 -8.51853907e-01 8.69668782e-01
1.00485599e+00 -9.05618668e-01 6.52785838e-01 8.20480287e-02
-4.16849367e-02 -4.59202528e-02 -7.09052503e-01 8.56163979e-01
9.17782485e-01 -2.96629131e-01 -1.20541894e+00 4.18001175e-01
1.17287910e+00 -4.52693760e-01 -5.90403080e-01 2.36452073e-01
5.15367687e-01 -6.59314454e-01 4.72272426e-01 -6.62191987e-01
1.00240934e+00 1.50076121e-01 1.53041119e-02 -1.56973517e+00
-1.87718749e-01 -7.24256337e-01 -5.79689443e-01 1.59085929e+00
4.57688451e-01 -3.62381160e-01 6.20187163e-01 3.09707016e-01
-1.01063001e+00 -4.45097119e-01 -5.26619017e-01 -7.72165895e-01
2.92179823e-01 6.53878180e-03 6.77344918e-01 3.02209020e-01
1.81597304e-02 1.01381302e+00 -1.27159938e-01 -1.77399442e-01
2.78192014e-01 7.89305642e-02 1.11022639e+00 -8.94162476e-01
-4.69155464e-04 -8.01989079e-01 9.67384428e-02 -1.49718153e+00
3.93121064e-01 -1.04783773e+00 2.88596511e-01 -2.29461241e+00
3.34529817e-01 2.92296618e-01 4.97433901e-01 2.43028536e-01
-2.92030513e-01 -2.04150721e-01 -1.12436330e-02 1.64826155e-01
-8.86858821e-01 9.20411944e-01 9.64538276e-01 -5.95667183e-01
-3.82056504e-01 -5.69225177e-02 -9.66398954e-01 5.15492380e-01
1.00123572e+00 -5.30778348e-01 -4.75507826e-01 -9.23963189e-01
1.98599733e-02 2.30702877e-01 -1.04292147e-01 -1.13434374e+00
7.29210675e-01 1.01608291e-01 2.87619174e-01 -1.46867418e+00
1.41100101e-02 2.04786092e-01 -3.94313782e-01 4.76381361e-01
-9.89805579e-01 3.87894243e-01 3.20358485e-01 4.19054478e-01
-3.20212185e-01 -6.22817874e-01 5.12694895e-01 -3.46639492e-02
-2.85483211e-01 4.61104624e-02 -7.96600878e-01 2.81966180e-01
3.68303031e-01 4.77937283e-03 -8.75740111e-01 -6.12607002e-01
-2.41241202e-01 5.02986908e-01 2.01356083e-01 3.40830646e-02
7.27501571e-01 -9.08050358e-01 -1.08281946e+00 -1.88483626e-01
-1.58728316e-01 2.26990089e-01 2.46139675e-01 3.44476938e-01
-1.08451104e+00 9.43883479e-01 -1.07230686e-01 -3.66753936e-01
-1.10425031e+00 3.76066566e-01 -6.04836941e-02 -5.64535499e-01
-7.85285711e-01 4.05355603e-01 -5.76895513e-02 -2.55529135e-01
2.93739617e-01 -8.35864961e-01 -1.63495883e-01 1.20568886e-01
8.22006166e-01 7.84622908e-01 3.19427997e-01 -5.89345247e-02
1.08900219e-01 -5.69702536e-02 -4.64178801e-01 -1.78470016e-01
1.17621505e+00 -1.34834483e-01 -2.01224923e-01 4.47574347e-01
1.28290009e+00 7.88785517e-02 -6.48885489e-01 1.72290485e-02
-3.02626640e-02 6.40536100e-02 3.12656462e-02 -6.10643983e-01
-6.06710434e-01 9.74900544e-01 -4.51929212e-01 2.67013311e-01
9.17787910e-01 -3.48874122e-01 1.24518013e+00 1.05870640e+00
-2.49136820e-01 -1.18992078e+00 9.28304866e-02 9.99014556e-01
1.16669786e+00 -6.00619435e-01 5.40633738e-01 -7.73137482e-03
-6.06513023e-01 1.46160340e+00 5.54627299e-01 -4.83615696e-01
1.04415275e-01 3.04318786e-01 -3.40061247e-01 -1.79448307e-01
-1.23014569e+00 2.28372321e-01 1.22214586e-01 4.39469159e-01
4.07377571e-01 -7.17456359e-03 -6.30484164e-01 6.84033692e-01
-8.17039073e-01 -1.46689508e-02 1.17496550e+00 9.59869981e-01
-7.72709429e-01 -6.27968967e-01 -7.20822737e-02 1.03999698e+00
-5.19017935e-01 -4.46374625e-01 -5.83598673e-01 8.31509888e-01
-7.31353700e-01 1.16208589e+00 3.19354683e-01 -2.06720814e-01
5.06665230e-01 -1.61080286e-01 3.79388839e-01 -1.02821386e+00
-1.03039551e+00 -9.06745791e-02 7.21448064e-01 -7.69763142e-02
-3.73030044e-02 -5.03515124e-01 -1.32513189e+00 -9.29577053e-01
-2.25978598e-01 4.37254518e-01 4.23542529e-01 6.79451168e-01
5.25942385e-01 9.91426826e-01 2.87127048e-01 -8.83346856e-01
-9.24689412e-01 -1.39822721e+00 -4.10719156e-01 -3.53979111e-01
4.82138306e-01 4.01525289e-01 -9.39640552e-02 1.37489513e-01]
|
[12.54194450378418, 9.506134033203125]
|
b456ca5c-353b-45a4-af70-bec2e7f2b549
|
semantic-discriminative-mixup-for
|
2206.06629
| null |
https://arxiv.org/abs/2206.06629v1
|
https://arxiv.org/pdf/2206.06629v1.pdf
|
Semantic-Discriminative Mixup for Generalizable Sensor-based Cross-domain Activity Recognition
|
It is expensive and time-consuming to collect sufficient labeled data to build human activity recognition (HAR) models. Training on existing data often makes the model biased towards the distribution of the training data, thus the model might perform terribly on test data with different distributions. Although existing efforts on transfer learning and domain adaptation try to solve the above problem, they still need access to unlabeled data on the target domain, which may not be possible in real scenarios. Few works pay attention to training a model that can generalize well to unseen target domains for HAR. In this paper, we propose a novel method called Semantic-Discriminative Mixup (SDMix) for generalizable cross-domain HAR. Firstly, we introduce semantic-aware Mixup that considers the activity semantic ranges to overcome the semantic inconsistency brought by domain differences. Secondly, we introduce the large margin loss to enhance the discrimination of Mixup to prevent misclassification brought by noisy virtual labels. Comprehensive generalization experiments on five public datasets demonstrate that our SDMix substantially outperforms the state-of-the-art approaches with 6% average accuracy improvement on cross-person, cross-dataset, and cross-position HAR.
|
['Xin Qin', 'Chunyu Hu', 'Sinno Jialin Pan', 'Yiqiang Chen', 'Jindong Wang', 'Wang Lu']
|
2022-06-14
| null | null | null | null |
['cross-domain-activity-recognition']
|
['computer-vision']
|
[ 1.01550473e-02 -2.25725651e-01 -4.01797116e-01 -5.79488218e-01
-9.92514312e-01 -4.19089019e-01 3.41952890e-01 -1.04991250e-01
-2.79217452e-01 1.08329308e+00 1.60872489e-01 1.39132112e-01
6.14938736e-02 -7.58273005e-01 -6.55260384e-01 -5.30271828e-01
3.91491085e-01 6.51165903e-01 3.49603206e-01 1.17365517e-01
-2.20039144e-01 7.73408040e-02 -1.24849439e+00 3.33868533e-01
1.23740685e+00 9.28044796e-01 6.10396899e-02 -7.44746998e-02
-9.09322128e-02 5.92384219e-01 -1.05805635e+00 -3.09042364e-01
1.84276953e-01 -7.04999566e-01 -5.96907377e-01 3.12931687e-01
3.49527121e-01 -1.28090128e-01 -2.31606781e-01 8.61496449e-01
7.32258677e-01 2.61705697e-01 8.66419494e-01 -1.50119531e+00
-1.00199711e+00 3.51266026e-01 -6.74010575e-01 1.72444180e-01
3.28036070e-01 5.94296260e-03 4.77654934e-01 -7.61080503e-01
4.62295324e-01 9.89489079e-01 8.21306109e-01 6.37727022e-01
-1.22593331e+00 -7.60837138e-01 2.16699556e-01 3.20774794e-01
-1.51227021e+00 -9.38835293e-02 8.65871370e-01 -3.32187384e-01
3.61080378e-01 1.10279724e-01 3.17727864e-01 1.73104095e+00
-3.69331419e-01 9.38605964e-01 1.33688104e+00 -2.29609147e-01
4.93619978e-01 3.03671509e-01 1.26395091e-01 2.89798230e-01
4.41112012e-01 -2.28484213e-01 -5.41184962e-01 -1.46090731e-01
5.71824431e-01 5.04660569e-02 -2.61606067e-01 -6.94004416e-01
-1.14042139e+00 5.53061545e-01 3.82642001e-01 2.31308833e-01
-3.21436465e-01 -3.71744692e-01 3.93784434e-01 2.49737017e-02
3.36506456e-01 2.62818038e-01 -6.10150337e-01 -1.19976103e-01
-8.69728208e-01 2.87525028e-01 5.37556231e-01 9.98848855e-01
5.11375129e-01 -1.42281055e-01 -3.73239577e-01 1.28312898e+00
1.08052470e-01 6.52339756e-01 7.06931353e-01 -6.13666415e-01
6.88966393e-01 6.28214836e-01 1.46909341e-01 -6.79463923e-01
-2.70094693e-01 -6.77227080e-01 -7.98291564e-01 -3.13359082e-01
7.57472634e-01 -2.40919054e-01 -1.00341845e+00 1.81583536e+00
3.57419312e-01 2.51869977e-01 3.18859592e-02 8.09294343e-01
6.29242778e-01 4.86786902e-01 4.84877199e-01 -1.16889402e-02
1.12122416e+00 -1.14662158e+00 -6.01356208e-01 -5.52739143e-01
6.45264924e-01 -3.67740780e-01 1.24263108e+00 3.58293980e-01
-6.48238957e-01 -7.72026777e-01 -1.12244415e+00 2.38622308e-01
-4.69266534e-01 2.54638255e-01 4.54215258e-01 7.62611866e-01
-2.84234703e-01 2.59550124e-01 -4.72704917e-01 -4.86947477e-01
6.31458223e-01 -1.10577634e-02 -4.75718081e-01 -4.27340895e-01
-1.38179457e+00 7.72023082e-01 5.17433107e-01 -9.37425345e-02
-9.49170887e-01 -6.76019669e-01 -8.12572420e-01 -8.35069865e-02
4.03995037e-01 -4.82281864e-01 1.25306857e+00 -1.10377276e+00
-1.03291273e+00 7.39539921e-01 -1.68526638e-02 -3.69337350e-01
8.42319846e-01 -1.36063144e-01 -8.36712122e-01 -2.25711673e-01
4.31694835e-01 4.19471592e-01 5.83648384e-01 -1.33903491e+00
-5.25241852e-01 -4.91482943e-01 -4.04735386e-01 1.89822644e-01
-3.28193754e-01 -5.08548737e-01 -3.61580163e-01 -8.09244812e-01
-7.73333013e-02 -9.25375760e-01 8.63612071e-02 -2.88555384e-01
-2.19641715e-01 -3.73498648e-01 7.64211297e-01 -8.33285987e-01
1.23030865e+00 -2.16416287e+00 -2.64361531e-01 1.46588504e-01
-2.55275339e-01 4.10072595e-01 4.16461863e-02 2.48098835e-01
7.89240077e-02 -2.94492930e-01 -2.73655951e-01 -4.74521816e-02
1.20232366e-01 4.07011867e-01 -1.16302736e-01 4.01698291e-01
-2.55583174e-04 7.69836307e-01 -8.43973815e-01 -5.78304350e-01
6.93143234e-02 3.20962727e-01 -3.35859030e-01 1.58258870e-01
-1.21009350e-01 6.90197527e-01 -5.60726047e-01 7.23848879e-01
8.13815475e-01 -4.03052956e-01 3.88907194e-01 2.88703712e-03
4.63360280e-01 1.00303657e-01 -1.23535836e+00 1.67584741e+00
-3.56465966e-01 1.83959603e-01 -5.98363578e-01 -1.31687355e+00
1.07350457e+00 9.29334089e-02 4.12232667e-01 -9.81947362e-01
-5.15187457e-02 3.32180172e-01 -3.10446113e-01 -4.71850604e-01
3.95490639e-02 -1.82720259e-01 -4.02097315e-01 -2.93476856e-03
1.69102494e-02 5.46267748e-01 -1.71436016e-02 -2.23320797e-01
8.53298068e-01 2.90132701e-01 4.02649879e-01 6.22736886e-02
5.32032371e-01 3.61485220e-02 1.02340889e+00 6.97057486e-01
-6.58726752e-01 6.83898866e-01 1.92296565e-01 -2.46987686e-01
-8.61402333e-01 -1.42700613e+00 -2.93829501e-01 1.32640779e+00
3.92930478e-01 -3.18919308e-02 -9.31837738e-01 -1.39353108e+00
4.12932597e-02 7.74000704e-01 -5.71283519e-01 -2.61796057e-01
-3.58595282e-01 -9.29974616e-01 6.64416015e-01 1.02745771e+00
1.23902595e+00 -8.65813315e-01 -4.29340154e-02 1.60923272e-01
-5.84905028e-01 -1.17531669e+00 -4.99179006e-01 2.28047613e-02
-5.84646702e-01 -1.07306898e+00 -1.00618136e+00 -8.48458171e-01
6.32231116e-01 1.82781875e-01 1.02826905e+00 -4.48179543e-01
-5.56316525e-02 1.59064531e-01 -4.08028036e-01 -4.09178823e-01
3.47849750e-03 1.38573304e-01 -3.01123243e-02 1.39756892e-02
1.12537122e+00 -3.87914062e-01 -6.15018129e-01 8.76565754e-01
-7.02727735e-01 -2.77028620e-01 6.58215165e-01 9.28471923e-01
6.48390472e-01 1.85463801e-01 1.13260829e+00 -7.86025941e-01
4.41188037e-01 -8.35157573e-01 -7.36420080e-02 5.34593165e-01
-6.98778033e-01 -3.66709828e-02 5.81149280e-01 -8.05114925e-01
-1.48337245e+00 3.00121102e-02 3.80351208e-02 -2.06391588e-01
-4.73278761e-01 2.43693158e-01 -6.88700378e-01 3.31842721e-01
9.50637996e-01 2.79186815e-01 -3.09585541e-01 -7.16081977e-01
1.40566766e-01 1.03260744e+00 6.44485772e-01 -8.82299006e-01
5.75679362e-01 5.26366353e-01 -3.35455567e-01 -5.50326288e-01
-1.28468633e+00 -6.55012965e-01 -6.17354631e-01 5.65687753e-03
8.74469995e-01 -1.15480256e+00 8.14013630e-02 5.14006495e-01
-6.74838722e-01 -3.76439750e-01 -2.40878895e-01 4.28892493e-01
-4.54158694e-01 3.65112782e-01 -1.56057134e-01 -5.75470746e-01
4.72590886e-02 -7.27794409e-01 9.71962571e-01 2.96964288e-01
-3.16844136e-01 -9.56625462e-01 2.46901080e-01 9.41056132e-01
1.14227921e-01 2.03536108e-01 5.34993589e-01 -1.03109396e+00
-1.52854979e-01 -2.03500643e-01 -2.20520556e-01 6.49320304e-01
1.87986791e-01 -6.01064026e-01 -1.08291996e+00 -1.06885150e-01
-4.29976642e-01 -4.71042871e-01 7.05353975e-01 2.10723653e-01
1.42096138e+00 2.46033445e-03 -4.90173995e-01 3.63654077e-01
1.06083632e+00 2.11100861e-01 8.76645803e-01 5.27215123e-01
6.12966180e-01 5.27602971e-01 1.04309070e+00 3.32144886e-01
5.05958021e-01 8.15435171e-01 -8.43470320e-02 -2.10633606e-01
-2.22979307e-01 -7.44461775e-01 3.26366603e-01 2.41301402e-01
6.82240631e-03 -3.02355319e-01 -7.91818142e-01 7.51857877e-01
-1.98778379e+00 -1.03517103e+00 5.29180979e-03 2.42546988e+00
6.88829124e-01 9.72666889e-02 4.45086241e-01 1.20990887e-01
8.63830626e-01 -3.38337645e-02 -5.98745525e-01 4.63565486e-03
-1.42702162e-01 8.69915783e-02 4.72313404e-01 -1.04776405e-01
-1.35499144e+00 5.85462689e-01 5.82694721e+00 1.07854879e+00
-7.72607446e-01 4.31068331e-01 5.74227393e-01 6.89645857e-02
1.15246866e-02 -3.38874787e-01 -8.22131097e-01 8.58620405e-01
8.54692996e-01 8.51126909e-02 1.65434368e-02 1.34283578e+00
-2.98090205e-02 -4.53287549e-02 -1.00212407e+00 1.29409409e+00
2.47304976e-01 -6.46698892e-01 -1.10784531e-01 1.13980718e-01
8.93371701e-01 -2.03311473e-01 -1.63838472e-02 8.78185093e-01
1.06350139e-01 -8.59887123e-01 4.94339347e-01 3.18168253e-01
6.63242698e-01 -6.88444734e-01 9.10711050e-01 4.87069905e-01
-1.03986382e+00 -1.37035385e-01 -4.14489448e-01 2.64994949e-01
2.54119840e-02 5.07660747e-01 -7.96264887e-01 5.06820917e-01
7.22690165e-01 5.92392623e-01 -7.64938414e-01 9.28157747e-01
-2.99413055e-01 5.97976089e-01 -1.15290664e-01 1.63895831e-01
-6.21689409e-02 -7.02450275e-02 9.06451792e-02 1.09838963e+00
4.69221115e-01 -1.87602922e-01 3.91831040e-01 6.00516319e-01
-2.18854487e-01 3.15990821e-02 -4.46933061e-01 -1.15311034e-02
5.28005183e-01 8.07092786e-01 -3.30759048e-01 -4.19102848e-01
-5.22739708e-01 1.31341970e+00 2.70245790e-01 4.82035488e-01
-1.24299908e+00 -2.09286109e-01 4.62547451e-01 4.47246641e-01
2.63527185e-01 -2.77500078e-02 -2.44008452e-01 -1.25890326e+00
3.85625839e-01 -7.97209620e-01 9.09732699e-01 -5.17423153e-01
-1.76155114e+00 2.28282273e-01 1.18747935e-01 -1.36955237e+00
-2.42489297e-02 -4.19042945e-01 -3.16104740e-01 6.97844684e-01
-1.37974918e+00 -1.17781246e+00 -6.30234897e-01 7.49571979e-01
5.98241508e-01 -2.34811127e-01 5.83100438e-01 7.52954006e-01
-6.06629491e-01 8.94511104e-01 1.59964174e-01 4.01137710e-01
1.20408452e+00 -1.14995027e+00 1.60601810e-02 6.52837455e-01
5.76525182e-02 4.95376468e-01 5.45409262e-01 -7.30736494e-01
-6.39357686e-01 -1.28720760e+00 8.45903575e-01 -6.38976634e-01
1.98636144e-01 -2.82208800e-01 -1.19773901e+00 8.97991240e-01
-1.55471668e-01 2.27341697e-01 1.07906103e+00 3.23044777e-01
-5.93801081e-01 -2.76295543e-01 -1.44473028e+00 2.13445246e-01
1.35468483e+00 -3.48353386e-01 -9.16707754e-01 3.15846771e-01
1.41620919e-01 -2.44318858e-01 -8.26852441e-01 4.95594412e-01
4.42643791e-01 -7.92080462e-01 1.02699125e+00 -8.13343942e-01
7.20900372e-02 -5.24718463e-01 -2.11977646e-01 -1.38250101e+00
-3.13593060e-01 1.97047777e-02 -1.10238992e-01 1.40942299e+00
3.29204440e-01 -5.90051532e-01 8.42872500e-01 5.49730420e-01
-3.14148441e-02 -3.10496002e-01 -7.48715878e-01 -1.29710472e+00
4.65766676e-02 -3.08300734e-01 7.29537427e-01 1.20237267e+00
-1.17719430e-03 3.38548720e-01 -4.93280143e-01 3.15529943e-01
7.23413587e-01 -8.21072236e-03 8.58690202e-01 -1.17977321e+00
-3.37552220e-01 3.07336040e-02 -3.77877414e-01 -1.11548734e+00
1.59472883e-01 -7.93101192e-01 2.03431025e-02 -1.57152069e+00
4.41085428e-01 -6.34401560e-01 -5.36080122e-01 5.19326866e-01
-2.10804939e-01 2.54760951e-01 -1.02091141e-01 2.36284867e-01
-9.76678371e-01 7.26875246e-01 1.08920014e+00 -1.88833043e-01
-2.30906248e-01 2.99747810e-02 -7.61043549e-01 6.31210327e-01
1.01171756e+00 -3.69798660e-01 -7.37941742e-01 -1.75551221e-01
-3.38964432e-01 -2.18566552e-01 5.31041980e-01 -1.41386974e+00
-1.72293082e-01 -3.24177474e-01 9.07918692e-01 -4.81254697e-01
1.75296590e-01 -7.03706920e-01 1.73151702e-01 5.41202649e-02
-2.63660878e-01 -4.34307843e-01 -6.30641505e-02 9.06711698e-01
-1.85125008e-01 -2.87237410e-02 9.09256399e-01 4.14188765e-02
-9.79007661e-01 3.23778600e-01 1.46266976e-02 4.69583720e-01
1.41655874e+00 -2.43348375e-01 -5.05424440e-01 -1.36571407e-01
-9.58306849e-01 2.41030127e-01 4.26360548e-01 6.60680056e-01
2.08652198e-01 -1.55606735e+00 -5.01626492e-01 1.50652915e-01
6.37765408e-01 -1.38208911e-01 5.96982777e-01 6.27752423e-01
-1.04116775e-01 3.42021763e-01 -2.49180511e-01 -4.48831856e-01
-1.13369942e+00 7.71642864e-01 4.37257618e-01 -2.20424965e-01
-3.38598430e-01 6.32727325e-01 3.16618681e-01 -6.86908603e-01
2.25698248e-01 6.99566081e-02 1.43790739e-02 -5.49981743e-02
5.92665732e-01 5.32224774e-01 6.94753975e-03 -4.69897807e-01
-5.10089099e-01 2.44211331e-01 3.26939784e-02 1.16998106e-01
9.03546035e-01 -7.46660605e-02 5.39685607e-01 3.45107436e-01
1.01235127e+00 -4.85560298e-02 -1.42821527e+00 -3.87207836e-01
1.36020482e-01 -5.73076427e-01 -3.97429228e-01 -1.13298357e+00
-7.63647974e-01 8.28454554e-01 8.13496351e-01 -2.12258808e-02
1.05578923e+00 1.64157718e-01 9.60076988e-01 1.96413681e-01
5.15701413e-01 -1.67094505e+00 1.89197242e-01 6.27788249e-03
5.79323173e-01 -1.46319771e+00 -2.16310382e-01 -4.02499467e-01
-1.17695594e+00 6.15766525e-01 9.91847277e-01 1.73183948e-01
3.97669166e-01 -3.16392034e-01 1.49546312e-02 1.80884227e-01
-2.03886122e-01 -2.40008473e-01 2.85114020e-01 1.08499682e+00
2.29723290e-01 3.13618541e-01 -3.62229615e-01 8.52282584e-01
2.32798532e-01 3.11429799e-01 5.88357784e-02 9.05463994e-01
-4.65945452e-01 -1.30717623e+00 -4.83987302e-01 2.92548776e-01
-3.00106585e-01 4.06255841e-01 -3.30327868e-01 9.15124059e-01
5.90747476e-01 8.51670563e-01 -1.52649716e-01 -2.70619631e-01
7.66148567e-01 3.84666353e-01 5.02569854e-01 -4.93129998e-01
-5.08124456e-02 -1.13003202e-01 3.39334160e-01 -3.45228702e-01
-4.55059141e-01 -7.83817708e-01 -1.21218073e+00 -1.08685307e-01
1.24825343e-01 1.79182470e-01 2.56620258e-01 8.65540445e-01
4.58834141e-01 4.69735712e-01 3.91047806e-01 -1.34741336e-01
-7.44345784e-01 -1.05868411e+00 -8.50193143e-01 9.44436908e-01
2.19791569e-02 -1.09746003e+00 -2.40642875e-01 7.38263279e-02]
|
[8.03231143951416, 1.06674063205719]
|
d0afcc62-0930-4b87-9903-71d9feddf1ec
|
scan2cap-context-aware-dense-captioning-in
|
2012.02206
| null |
https://arxiv.org/abs/2012.02206v1
|
https://arxiv.org/pdf/2012.02206v1.pdf
|
Scan2Cap: Context-aware Dense Captioning in RGB-D Scans
|
We introduce the task of dense captioning in 3D scans from commodity RGB-D sensors. As input, we assume a point cloud of a 3D scene; the expected output is the bounding boxes along with the descriptions for the underlying objects. To address the 3D object detection and description problems, we propose Scan2Cap, an end-to-end trained method, to detect objects in the input scene and describe them in natural language. We use an attention mechanism that generates descriptive tokens while referring to the related components in the local context. To reflect object relations (i.e. relative spatial relations) in the generated captions, we use a message passing graph module to facilitate learning object relation features. Our method can effectively localize and describe 3D objects in scenes from the ScanRefer dataset, outperforming 2D baseline methods by a significant margin (27.61% CiDEr@0.5IoUimprovement).
|
['Angel X. Chang', 'Matthias Nießner', 'Ali Gholami', 'Dave Zhenyu Chen']
|
2020-12-03
| null |
http://openaccess.thecvf.com//content/CVPR2021/html/Chen_Scan2Cap_Context-Aware_Dense_Captioning_in_RGB-D_Scans_CVPR_2021_paper.html
|
http://openaccess.thecvf.com//content/CVPR2021/papers/Chen_Scan2Cap_Context-Aware_Dense_Captioning_in_RGB-D_Scans_CVPR_2021_paper.pdf
|
cvpr-2021-1
|
['dense-captioning']
|
['computer-vision']
|
[ 1.43612355e-01 5.23153424e-01 -1.25121893e-02 -7.08972812e-01
-9.42995131e-01 -6.93960845e-01 6.39480114e-01 2.29784340e-01
3.65330428e-02 1.22941233e-01 4.73280877e-01 -1.89986937e-02
1.73608080e-01 -5.31432092e-01 -1.27804017e+00 -2.31950462e-01
-1.97494566e-01 9.01695192e-01 2.98784554e-01 1.68337613e-01
1.06902428e-01 9.95040655e-01 -1.33769190e+00 3.27562809e-01
1.61840007e-01 1.23218000e+00 5.60688376e-01 6.28665090e-01
-4.81227726e-01 7.57274985e-01 -4.50173140e-01 -1.61593527e-01
2.07012281e-01 2.00480539e-02 -9.52913642e-01 6.52641833e-01
8.09950709e-01 -4.55024987e-01 -6.21478975e-01 7.30530739e-01
2.99951553e-01 1.11988313e-01 7.18591452e-01 -1.23506379e+00
-8.87378454e-01 3.37208271e-01 -8.90661955e-01 3.99111100e-02
7.89168239e-01 2.41400376e-01 1.14737070e+00 -1.24218142e+00
7.24091411e-01 1.63702226e+00 2.47878328e-01 6.41943455e-01
-1.21563578e+00 -5.33898115e-01 3.60410661e-01 -1.86385751e-01
-1.48653626e+00 -1.97267950e-01 8.17683160e-01 -3.66902143e-01
1.34410274e+00 1.35099798e-01 4.59970862e-01 9.27387357e-01
-1.65980548e-01 8.67877185e-01 4.86686975e-01 -2.65039671e-02
1.58773825e-01 -1.63487270e-02 1.26931533e-01 6.20406806e-01
1.10814162e-01 -1.99255288e-01 -3.82861882e-01 -3.54435444e-02
1.00147045e+00 -2.68050097e-02 8.44674483e-02 -7.16356397e-01
-1.36879611e+00 5.21582961e-01 1.16599178e+00 -2.72315323e-01
-6.33447170e-01 6.47382855e-01 1.60267204e-02 -5.06795347e-01
5.15081584e-01 3.59775960e-01 -3.62749696e-01 3.42900723e-01
-2.61713654e-01 4.13841397e-01 4.21161503e-01 1.86369395e+00
7.91694760e-01 -5.10353506e-01 -3.64156246e-01 4.63688046e-01
6.38827622e-01 8.75031829e-01 -3.63050908e-01 -8.80900145e-01
9.26175654e-01 8.03421974e-01 2.39193872e-01 -8.25767696e-01
-3.33162516e-01 -1.65883943e-01 -4.72771376e-01 -1.86590970e-01
-2.64164120e-01 4.13564771e-01 -1.29059982e+00 1.46281922e+00
5.55383563e-01 2.98137039e-01 -1.42859891e-02 1.26736224e+00
1.23698914e+00 8.69870007e-01 3.82181346e-01 5.14373302e-01
1.45541275e+00 -7.95239151e-01 -3.86537910e-01 -6.07803047e-01
3.39814633e-01 -5.95215082e-01 1.01846898e+00 -4.19977337e-01
-1.14485610e+00 -6.26337051e-01 -6.24041498e-01 -4.97384846e-01
-2.63745159e-01 -5.35000488e-02 5.53655386e-01 -2.30452970e-01
-9.60158169e-01 5.44624142e-02 -9.01345372e-01 -4.83411044e-01
7.78959751e-01 3.29257071e-01 -5.03004909e-01 -4.12070692e-01
-4.55040514e-01 7.83513248e-01 5.07858872e-01 -1.15400143e-01
-1.27778208e+00 -7.04260051e-01 -1.25996542e+00 -3.98864895e-02
1.78156659e-01 -9.01859045e-01 1.41365170e+00 -1.75200313e-01
-7.19023824e-01 1.42944348e+00 -1.88001305e-01 -1.36628687e-01
1.97150901e-01 -3.77956361e-01 5.42340055e-02 3.78918856e-01
5.04917145e-01 1.34162092e+00 3.78770292e-01 -1.78271675e+00
-5.38820565e-01 -3.73961687e-01 1.86054900e-01 4.80236709e-01
5.69862902e-01 -7.65556917e-02 -1.11431754e+00 -1.89554319e-01
6.27020538e-01 -9.66601849e-01 -2.28585944e-01 4.72600371e-01
-9.24048543e-01 -4.14517403e-01 8.57245207e-01 -2.08459437e-01
1.97695926e-01 -2.38129735e+00 -7.65112340e-02 6.22544251e-02
2.25953966e-01 -2.36212194e-01 -3.33110034e-01 1.01972371e-01
-9.07143876e-02 5.36680371e-02 -4.08509880e-01 -8.01949620e-01
2.33586997e-01 5.58167517e-01 -5.64637542e-01 4.69486862e-01
9.91942704e-01 1.19252956e+00 -1.28778863e+00 -6.88938022e-01
5.38364053e-01 6.37195945e-01 -4.02733207e-01 6.55345082e-01
-7.78633416e-01 3.70144486e-01 -8.37410688e-01 8.44099879e-01
8.73770952e-01 -6.16372228e-01 -4.71707851e-01 -3.01492482e-01
8.33879933e-02 6.58175647e-01 -9.40703213e-01 2.22366500e+00
-5.15190244e-01 6.91013634e-01 -2.04805404e-01 -4.50652242e-01
1.05530202e+00 2.27754004e-02 4.20885921e-01 -4.75425035e-01
-1.48867771e-01 -1.09389856e-01 -6.52716339e-01 -6.99487746e-01
3.96325827e-01 1.52229965e-01 -3.98349583e-01 4.57913667e-01
1.68887302e-02 -8.89963150e-01 -2.58300334e-01 5.24048984e-01
1.17609835e+00 3.86087537e-01 5.59165403e-02 1.89161494e-01
2.14193128e-02 3.89256239e-01 -2.21079454e-01 7.55902946e-01
1.37295350e-01 1.02774549e+00 3.13667506e-01 -5.12237966e-01
-1.11461043e+00 -1.40610230e+00 -5.95233664e-02 6.33309007e-01
6.76133156e-01 -2.14423642e-01 -1.17473356e-01 -7.68475056e-01
2.19132230e-01 8.76024783e-01 -5.97983897e-01 -3.24510932e-02
-5.37478805e-01 -1.44992881e-02 2.42501631e-01 7.67219603e-01
3.92699480e-01 -1.01159406e+00 -7.61193156e-01 -3.65661643e-02
-1.44467533e-01 -1.68712187e+00 -4.29711729e-01 3.05084497e-01
-8.03468466e-01 -8.39816213e-01 -4.71397042e-01 -1.02807033e+00
1.14116848e+00 5.57072639e-01 1.59876394e+00 -2.71873593e-01
-2.28676364e-01 6.68936193e-01 -3.59027505e-01 -5.47869086e-01
-2.83253670e-01 -5.90658151e-02 -2.67459899e-01 -3.93480599e-01
5.21018565e-01 -3.08556139e-01 -5.12168348e-01 2.43232995e-01
-7.94030249e-01 4.88149345e-01 8.95819902e-01 2.03844666e-01
1.13545191e+00 -5.44711292e-01 -2.77154624e-01 -4.96413052e-01
-8.22258815e-02 -6.59753084e-01 -5.07164180e-01 6.50831535e-02
5.14409691e-02 3.27361465e-01 -6.28405018e-03 -3.69445443e-01
-7.19780982e-01 7.38308847e-01 6.88275993e-02 -8.58797014e-01
-6.56677008e-01 8.24380517e-02 -2.65112460e-01 2.06261039e-01
5.57608247e-01 2.21940503e-01 -4.88613427e-01 -5.71028590e-01
7.51970708e-01 5.40633559e-01 7.58398890e-01 -6.45732045e-01
9.99564707e-01 5.34671664e-01 1.99863359e-01 -4.64933693e-01
-1.29883718e+00 -6.11449897e-01 -7.88771212e-01 -2.74473373e-02
1.09381163e+00 -1.27976322e+00 -6.98583066e-01 -2.80277699e-01
-1.94527829e+00 -1.18039921e-01 -3.74668866e-01 4.43991572e-01
-7.54747033e-01 -3.03587884e-01 -3.33027989e-01 -6.83491707e-01
-1.24411732e-01 -9.30023611e-01 2.07376218e+00 2.74020154e-02
-6.31733462e-02 -4.01280731e-01 -2.39734188e-01 1.49754331e-01
-9.03013349e-02 6.78354025e-01 9.26179588e-01 -4.91774946e-01
-1.20257652e+00 -1.75671265e-01 -7.94633806e-01 -4.99630496e-02
1.55909285e-01 -3.22761118e-01 -1.04195571e+00 2.30994135e-01
-2.93815583e-01 -1.92873761e-01 4.13218498e-01 2.15051398e-01
1.46034825e+00 -2.11546198e-01 -7.34946489e-01 5.86906254e-01
1.35909891e+00 -6.61635250e-02 2.28496760e-01 -1.68048032e-02
9.52170968e-01 6.85221672e-01 6.76232696e-01 3.74342859e-01
5.37932158e-01 5.80708027e-01 1.05659139e+00 -3.14879954e-01
-4.22693938e-01 -7.55566478e-01 -1.01375803e-01 5.10641634e-02
4.83337998e-01 -4.94083703e-01 -1.20646584e+00 6.37917101e-01
-1.67044401e+00 -4.81954396e-01 -3.13340276e-01 1.62974262e+00
5.58437228e-01 2.84663618e-01 -2.29688853e-01 -5.02597809e-01
9.10654366e-01 1.07955471e-01 -8.10165405e-01 5.53438738e-02
3.88293937e-02 -2.93488503e-02 5.43946981e-01 3.83003265e-01
-1.10016894e+00 9.39067841e-01 5.82176161e+00 6.63259178e-02
-7.72031486e-01 -2.49740183e-01 5.25270522e-01 -3.65013331e-01
-3.30857426e-01 -1.68281823e-01 -8.57343733e-01 -5.89932650e-02
5.78869224e-01 8.53269398e-02 -7.74643719e-02 9.69413221e-01
1.24971524e-01 5.84242530e-02 -1.73675585e+00 1.30646706e+00
2.40974471e-01 -1.23622847e+00 3.42730433e-01 -2.42227986e-02
6.74164414e-01 3.95171195e-01 -7.03720152e-02 1.02491695e-02
3.13246995e-01 -9.47755039e-01 1.07369685e+00 4.46539104e-01
8.31831217e-01 -4.26181167e-01 5.49686909e-01 2.21526742e-01
-1.14157116e+00 2.89727420e-01 -4.60859507e-01 2.86388900e-02
4.31191087e-01 4.58799332e-01 -1.55264962e+00 3.40438187e-01
7.24155724e-01 7.72395194e-01 -4.15765673e-01 1.12708616e+00
-5.85678160e-01 3.17521483e-01 -6.59356952e-01 -2.54850209e-01
6.04742825e-01 2.58920759e-01 6.91025734e-01 1.16745710e+00
2.97611445e-01 5.35448790e-01 2.90811472e-02 1.41003680e+00
-4.03321296e-01 -2.52133727e-01 -8.52201521e-01 1.11729093e-01
5.35919845e-01 9.60296214e-01 -7.32132792e-01 -2.96796352e-01
-3.41224581e-01 1.08836520e+00 2.15574577e-01 2.56599009e-01
-9.72393215e-01 -2.07761213e-01 7.50010252e-01 3.35297644e-01
4.91846919e-01 -4.43546772e-01 -2.23654523e-01 -6.85464382e-01
4.38023001e-01 -3.69256944e-03 1.76147640e-01 -1.64811468e+00
-1.39049292e+00 6.21422231e-01 3.74677211e-01 -1.35876608e+00
-2.27009077e-02 -5.99263370e-01 -2.44408458e-01 9.01740849e-01
-1.62978244e+00 -1.43477511e+00 -8.63119602e-01 4.43013281e-01
6.49913669e-01 5.34542322e-01 7.37170339e-01 1.06162138e-01
-1.90967657e-02 -1.95576057e-01 -5.67544043e-01 2.64106572e-01
2.43307322e-01 -1.26609159e+00 1.10847485e+00 4.04433876e-01
4.69646394e-01 3.27701151e-01 6.19759560e-01 -6.68440223e-01
-1.47226369e+00 -1.51111341e+00 1.09593129e+00 -1.04242575e+00
3.65914226e-01 -9.31135893e-01 -6.82493925e-01 9.04914320e-01
-1.27258852e-01 5.00747979e-01 3.02630395e-01 -3.14246446e-01
-6.18724942e-01 1.33133844e-01 -1.08135426e+00 4.68170822e-01
1.64075112e+00 -6.01791859e-01 -7.96877921e-01 7.84002781e-01
1.40844226e+00 -1.07234204e+00 -5.20034969e-01 1.89775452e-01
1.28168017e-02 -3.86827648e-01 1.41101170e+00 -6.81099892e-01
6.69737220e-01 -5.83008885e-01 -4.91210729e-01 -9.02462602e-01
-2.86297053e-01 -1.02895617e-01 -2.19801769e-01 1.02814543e+00
4.42079157e-01 2.38867730e-01 8.51213157e-01 1.00514925e+00
-5.34300387e-01 -4.72376347e-01 -8.05594325e-01 -5.60622752e-01
-3.40271801e-01 -7.92478681e-01 9.03554618e-01 4.58848983e-01
-5.13258696e-01 5.30780315e-01 1.04149371e-01 8.33693266e-01
7.09404767e-01 4.61621970e-01 8.61694336e-01 -9.28119183e-01
3.91940437e-02 -1.49861559e-01 -7.45096385e-01 -1.68795228e+00
3.04282039e-01 -1.00993288e+00 5.22908211e-01 -2.08967924e+00
1.14566341e-01 -6.65242970e-01 -1.12783229e-02 5.76630473e-01
1.23461567e-01 3.42837572e-01 2.33406708e-01 1.80105105e-01
-9.62901473e-01 5.58062136e-01 1.33821070e+00 -5.61005473e-01
-1.21200927e-01 -6.11223914e-02 -6.43203318e-01 3.40461403e-01
3.82110834e-01 -6.88328981e-01 -2.52700150e-01 -1.21184742e+00
-5.12668304e-02 2.55914964e-02 8.67803991e-01 -9.48417723e-01
2.49905080e-01 -1.91643000e-01 6.05761468e-01 -1.26385725e+00
7.79441416e-01 -1.22495437e+00 9.86802876e-02 8.07563439e-02
-6.41211390e-01 1.50579214e-01 3.83302122e-01 6.50253713e-01
3.66416909e-02 1.73608631e-01 3.67357880e-01 -2.72668123e-01
-8.79594266e-01 7.61295915e-01 2.14226976e-01 -1.74654305e-01
1.15282547e+00 -3.73938354e-03 -2.35203370e-01 -2.71313041e-01
-6.96612835e-01 5.18895924e-01 4.27520156e-01 6.87022269e-01
1.02772868e+00 -1.45118320e+00 -8.64444852e-01 2.02394471e-01
8.74993205e-01 1.09737182e+00 5.48494644e-02 1.33909419e-01
-6.49461627e-01 5.22832990e-01 9.37829465e-02 -1.14166129e+00
-7.79219389e-01 6.02359772e-01 3.15565228e-01 5.07198691e-01
-9.33025301e-01 1.27310395e+00 4.52933818e-01 -3.40905160e-01
4.79126781e-01 -8.27107251e-01 3.06288093e-01 -5.40680170e-01
3.84634674e-01 -3.28853488e-01 -6.43595606e-02 -7.57299006e-01
-8.03422332e-01 7.90712237e-01 1.68153584e-01 -1.05037376e-01
1.56911445e+00 -2.02830359e-01 5.03531620e-02 3.30830961e-01
1.41454399e+00 -4.07520562e-01 -1.52784419e+00 -5.28537393e-01
-2.31075864e-02 -6.72601283e-01 -8.43202695e-02 -8.28255415e-01
-6.83350503e-01 8.03456604e-01 4.87274975e-01 8.68811011e-02
7.90618181e-01 1.17122471e+00 4.35210049e-01 5.48251688e-01
3.88838500e-01 -2.16301933e-01 2.46561408e-01 3.80245149e-01
1.23555827e+00 -1.37454081e+00 -8.96068364e-02 -6.85070455e-01
-6.71195328e-01 8.72062802e-01 7.72597194e-01 -2.39275411e-01
4.16621089e-01 1.59254938e-01 -4.43830341e-02 -7.34484315e-01
-7.43522763e-01 -4.06940609e-01 3.62503290e-01 9.00606692e-01
3.22447158e-02 -3.22499610e-02 7.16374815e-01 2.10583359e-01
-5.18345200e-02 -5.54797053e-01 1.81076795e-01 6.79511249e-01
-2.49868155e-01 -4.70393270e-01 -4.46961492e-01 1.94274008e-01
2.62576520e-01 4.82186601e-02 -6.82243526e-01 8.60165894e-01
9.97915789e-02 7.77466476e-01 7.35905051e-01 -2.92079121e-01
9.66447949e-01 -1.78117424e-01 4.03434873e-01 -1.05102587e+00
-6.06590062e-02 -1.80398002e-01 -4.90208007e-02 -7.63349235e-01
-6.00007713e-01 -5.82197964e-01 -1.73134768e+00 2.39256740e-01
-6.10673241e-02 -1.00974210e-01 1.02879393e+00 6.99682176e-01
6.83812201e-01 3.93118113e-01 7.13706672e-01 -1.10091734e+00
-1.78327486e-01 -8.15028250e-01 -2.59736627e-01 7.14730620e-01
5.52124321e-01 -5.87788343e-01 -1.90824807e-01 2.23505408e-01]
|
[8.239520072937012, -3.2197132110595703]
|
3c26ad95-f07c-43f5-8140-e67755db76ae
|
large-selective-kernel-network-for-remote
|
2303.0903
| null |
https://arxiv.org/abs/2303.09030v2
|
https://arxiv.org/pdf/2303.09030v2.pdf
|
Large Selective Kernel Network for Remote Sensing Object Detection
|
Recent research on remote sensing object detection has largely focused on improving the representation of oriented bounding boxes but has overlooked the unique prior knowledge presented in remote sensing scenarios. Such prior knowledge can be useful because tiny remote sensing objects may be mistakenly detected without referencing a sufficiently long-range context, and the long-range context required by different types of objects can vary. In this paper, we take these priors into account and propose the Large Selective Kernel Network (LSKNet). LSKNet can dynamically adjust its large spatial receptive field to better model the ranging context of various objects in remote sensing scenarios. To the best of our knowledge, this is the first time that large and selective kernel mechanisms have been explored in the field of remote sensing object detection. Without bells and whistles, LSKNet sets new state-of-the-art scores on standard benchmarks, i.e., HRSC2016 (98.46\% mAP), DOTA-v1.0 (81.85\% mAP) and FAIR1M-v1.0 (47.87\% mAP). Based on a similar technique, we rank 2nd place in 2022 the Greater Bay Area International Algorithm Competition. Code is available at https://github.com/zcablii/Large-Selective-Kernel-Network.
|
['Xiang Li', 'Jian Yang', 'Ming-Ming Cheng', 'Zhaohui Zheng', 'Qibin Hou', 'YuXuan Li']
|
2023-03-16
| null | null | null | null |
['object-detection-in-aerial-images']
|
['computer-vision']
|
[ 2.41391465e-01 -3.36963058e-01 -8.77873227e-02 -4.18259889e-01
-5.43625653e-01 -7.77588129e-01 6.15730286e-01 1.11218914e-01
-5.82847238e-01 5.08421004e-01 1.90134980e-02 -4.74518031e-01
-4.41946447e-01 -9.36493695e-01 -6.52919054e-01 -6.60245180e-01
-4.38389897e-01 1.78242996e-01 4.62028682e-01 -1.96339004e-02
7.57495761e-02 8.30120623e-01 -1.48279154e+00 1.15965120e-01
6.44498587e-01 8.63181233e-01 4.92793947e-01 7.19630241e-01
2.69997150e-01 3.49709332e-01 -3.49708349e-01 8.35471898e-02
7.72971153e-01 4.45464514e-02 -4.74576980e-01 -5.72425663e-01
1.00883043e+00 -3.99562329e-01 -3.35891694e-01 1.17653060e+00
6.77328467e-01 3.52100074e-01 7.92354763e-01 -7.98189044e-01
-5.70609093e-01 8.23924065e-01 -9.27015901e-01 8.24224651e-01
-4.26399112e-01 1.87344804e-01 1.03005385e+00 -1.07103193e+00
1.96235821e-01 1.09663999e+00 9.94340003e-01 3.16589743e-01
-1.16079533e+00 -9.61457729e-01 6.48694634e-01 7.80615509e-02
-2.02219605e+00 -3.19220185e-01 3.01893592e-01 -4.79888916e-01
1.00102592e+00 6.89905107e-01 4.74143267e-01 6.41636908e-01
-1.60579070e-01 5.72120726e-01 1.02901733e+00 -4.45467718e-02
1.81604639e-01 -2.07968708e-02 1.49432451e-01 1.75024286e-01
5.12027979e-01 3.46942246e-01 -3.99326198e-02 -1.51248127e-01
7.20992804e-01 3.15371841e-01 -3.85517091e-01 1.09416276e-01
-1.13546860e+00 7.80897081e-01 1.07138979e+00 6.07907400e-02
-2.70804405e-01 3.31098706e-01 5.08563519e-02 -1.16341956e-01
5.72808385e-01 3.16734642e-01 -4.91474628e-01 5.03205955e-01
-1.17934322e+00 2.07795501e-01 3.63398194e-01 6.68548822e-01
8.26070547e-01 6.21200539e-02 -1.38267040e-01 8.49535823e-01
4.91652519e-01 1.02012336e+00 -1.36673063e-01 -6.26023829e-01
4.49833304e-01 3.76231074e-01 3.21140081e-01 -9.96237099e-01
-6.91726267e-01 -6.61124051e-01 -9.79526520e-01 5.13920248e-01
2.89683938e-01 -1.05357185e-01 -1.31407785e+00 1.46656370e+00
3.08511198e-01 3.63677919e-01 -1.06507443e-01 1.05192912e+00
9.19096828e-01 9.27347302e-01 2.81992882e-01 3.93494308e-01
1.13486063e+00 -5.84750891e-01 -6.66599721e-02 -6.68502986e-01
6.93960786e-02 -4.56642270e-01 1.00174773e+00 1.38152555e-01
-4.45421129e-01 -5.49143374e-01 -1.10850894e+00 4.58647251e-01
-7.78590441e-01 3.87590319e-01 7.32137799e-01 6.49492502e-01
-1.05362213e+00 2.91265607e-01 -7.27118433e-01 -5.66872239e-01
6.65933132e-01 9.01971608e-02 -1.21858511e-02 -1.80452708e-02
-1.19035149e+00 7.97858179e-01 3.54605824e-01 5.36807835e-01
-9.70652103e-01 -1.03516555e+00 -6.05454028e-01 -7.33138546e-02
5.05602002e-01 -1.61644235e-01 9.75206316e-01 -6.36311591e-01
-7.31995821e-01 7.01919436e-01 3.49700660e-01 -5.66015542e-01
5.52258372e-01 -3.78345639e-01 -5.83082139e-01 8.78685713e-03
1.85634062e-01 9.56489265e-01 6.57057047e-01 -1.32758713e+00
-8.86940122e-01 -4.23722833e-01 7.83320442e-02 2.27776006e-01
4.74315360e-02 5.61278239e-02 -1.15163021e-01 -7.97109365e-01
2.69103020e-01 -8.86268795e-01 -3.88544381e-01 4.04459536e-01
-2.41758227e-01 1.37068070e-02 7.56280482e-01 -4.31273311e-01
1.19078755e+00 -2.10860682e+00 -4.06074941e-01 3.66939187e-01
3.45412120e-02 4.30377275e-01 -1.64558277e-01 2.14434743e-01
6.28037378e-02 5.07714152e-01 -4.36853260e-01 2.86628664e-01
-9.50610638e-02 1.20015182e-01 -6.33406878e-01 7.52674341e-01
5.29579222e-02 7.52516150e-01 -8.40770721e-01 -1.43297106e-01
4.61051434e-01 7.08022058e-01 -2.21110851e-01 -2.06276491e-01
-8.20645466e-02 1.21544376e-01 -4.47067589e-01 9.12267566e-01
1.05051136e+00 -3.43894273e-01 -1.52869225e-01 -2.47198030e-01
-4.42434460e-01 -6.63970262e-02 -1.44322026e+00 1.11072266e+00
-1.22491643e-01 7.15535402e-01 1.35720596e-01 -4.95506108e-01
8.71114075e-01 -1.30670756e-01 2.44997948e-01 -5.39069355e-01
-2.30423361e-01 1.15066789e-01 -4.44866903e-02 1.39148235e-02
6.48800015e-01 6.88596070e-03 8.48943740e-02 -1.61961429e-02
-6.77451789e-01 -4.63501811e-02 -1.13006972e-01 5.95869087e-02
7.65522242e-01 2.16257069e-02 5.24494708e-01 -4.57940221e-01
1.29530907e-01 1.16764382e-01 3.11835021e-01 1.55248523e+00
-3.61667991e-01 7.95923948e-01 -3.96828741e-01 -7.17734933e-01
-6.67073965e-01 -1.20785725e+00 -7.63461173e-01 1.50771308e+00
2.84468740e-01 1.29360229e-01 -1.84735999e-01 -5.53516388e-01
3.36493582e-01 5.42129636e-01 -8.80587578e-01 2.47045070e-01
-2.78490037e-01 -1.27836955e+00 1.01949966e+00 7.67489851e-01
5.50643563e-01 -1.01611924e+00 -1.03573120e+00 1.75794914e-01
1.58819407e-01 -9.60980952e-01 -1.94101587e-01 3.04620713e-01
-8.10409665e-01 -8.99742305e-01 -6.68053687e-01 -9.55635607e-02
2.67947912e-01 8.01868856e-01 1.08230150e+00 -5.71287125e-02
-4.74976361e-01 7.48656914e-02 -5.75824678e-01 -5.83826602e-01
1.13654181e-01 1.05930001e-01 -1.59956872e-01 -2.29439676e-01
3.32342833e-01 -5.32325268e-01 -9.41115558e-01 6.03702962e-01
-1.03455913e+00 -1.63525343e-01 8.13396454e-01 4.09465432e-01
7.06218600e-01 1.24431662e-01 4.16371018e-01 -6.33309901e-01
-2.12181777e-01 -7.53798425e-01 -9.48988974e-01 2.99027145e-01
-5.69514751e-01 -3.11239958e-01 2.14764848e-01 -4.83495712e-01
-9.63668168e-01 1.56274900e-01 3.16105448e-02 -1.73460588e-01
-3.72055233e-01 5.50815940e-01 -3.79257835e-02 -2.66256124e-01
1.12210464e+00 1.71629712e-01 -7.58318782e-01 -5.50162196e-01
3.95421028e-01 8.00093830e-01 5.22834480e-01 -3.44478756e-01
9.69547212e-01 1.09848022e+00 -1.80225909e-01 -9.20847654e-01
-9.73268151e-01 -9.12866294e-01 -4.20159638e-01 -1.37679964e-01
8.55487943e-01 -1.33469343e+00 -2.46307760e-01 5.39312005e-01
-6.96030557e-01 -7.53909647e-01 -1.24371648e-01 5.22531390e-01
6.12284765e-02 4.50838767e-02 -1.16288275e-01 -1.08237529e+00
-4.94630009e-01 -5.27022541e-01 1.00305593e+00 4.19281721e-01
2.21970901e-01 -6.10882342e-01 6.33747578e-02 1.48928817e-02
8.36568594e-01 2.73529887e-01 9.08679068e-02 -4.07322049e-01
-6.46386802e-01 -6.23963848e-02 -8.19566131e-01 2.14770779e-01
4.66813818e-02 -6.98019145e-03 -1.19768500e+00 -3.31111193e-01
-6.48482740e-01 -9.89057198e-02 1.54593837e+00 6.70702577e-01
1.17667806e+00 -1.58075273e-01 -4.60262477e-01 9.46825862e-01
1.66120327e+00 8.25379863e-02 7.81051040e-01 4.11374927e-01
7.86062360e-01 3.82385999e-01 5.65735877e-01 6.33804321e-01
4.15951192e-01 5.84059238e-01 8.96060944e-01 -2.45590478e-01
-8.28684419e-02 1.03564650e-01 1.62413895e-01 -4.04643118e-01
-3.55258733e-01 -4.19603616e-01 -1.31622779e+00 7.14936256e-01
-1.77488148e+00 -1.22249377e+00 -3.00690532e-01 2.08483195e+00
5.32568395e-01 -2.86968332e-03 -1.51477501e-01 -2.57464916e-01
7.90989697e-01 5.66508770e-01 -7.83734739e-01 3.77826810e-01
-3.58273357e-01 2.96128225e-02 1.32712734e+00 6.32994592e-01
-1.67206049e+00 9.01564538e-01 5.49512482e+00 9.51661229e-01
-1.19030845e+00 1.09522887e-01 5.13941944e-01 -1.09190531e-01
4.33696853e-03 -4.91834693e-02 -1.24488902e+00 2.46609971e-01
6.23792589e-01 4.11537617e-01 2.62207359e-01 8.51914465e-01
2.75558293e-01 -4.12744105e-01 -5.50947130e-01 8.11719835e-01
-1.22060105e-01 -1.38263822e+00 1.47811119e-02 7.36474153e-03
7.60403275e-01 1.04416537e+00 9.05324370e-02 2.68449515e-01
3.92108858e-01 -1.23286295e+00 8.99501920e-01 5.78894794e-01
9.09444153e-01 -4.37860847e-01 6.31335676e-01 1.70895398e-01
-1.69612730e+00 -1.36102676e-01 -7.25975633e-01 -1.63534936e-02
-1.46057382e-01 6.46462321e-01 -6.31013632e-01 4.44960833e-01
1.30106497e+00 6.71776831e-01 -7.76662350e-01 1.36442769e+00
-1.01397373e-01 8.23568881e-01 -8.11709344e-01 8.18561688e-02
5.63564956e-01 1.66537523e-01 6.95348382e-01 1.58568728e+00
2.98751771e-01 2.81607807e-01 3.88506174e-01 8.30006897e-01
7.35383555e-02 -7.89765567e-02 -5.66808343e-01 3.57956886e-01
6.08309507e-01 1.33703566e+00 -8.57234716e-01 -3.33736002e-01
-3.03978652e-01 4.82963562e-01 -1.16917469e-01 4.45943475e-01
-1.00029027e+00 -1.95335433e-01 7.09808230e-01 2.94183284e-01
4.92134273e-01 -7.76845962e-02 -2.53802329e-01 -1.00671494e+00
-1.11800432e-01 -4.09181505e-01 7.28237867e-01 -9.21765566e-01
-1.39493418e+00 5.09209692e-01 4.16197419e-01 -1.15976155e+00
5.81981838e-01 -6.86494470e-01 -5.32195628e-01 1.00220072e+00
-2.06083131e+00 -1.29085636e+00 -7.46286750e-01 2.78271317e-01
3.55037540e-01 8.70296806e-02 5.21757603e-01 2.98770338e-01
-2.62681812e-01 4.24500138e-01 2.82597631e-01 2.42637798e-01
6.51832819e-01 -1.17932701e+00 5.79971731e-01 1.01059210e+00
-5.43578062e-03 4.79200900e-01 5.16330719e-01 -7.31560469e-01
-9.46259856e-01 -1.72988367e+00 3.39495748e-01 -5.80664694e-01
6.90123320e-01 -1.26855612e-01 -1.07705903e+00 5.58376312e-01
-3.63692224e-01 5.55184901e-01 5.02735317e-01 -4.95987236e-02
-7.27131367e-01 -4.03709143e-01 -1.18191099e+00 3.90283883e-01
9.94555771e-01 -4.07859683e-01 -2.38457933e-01 2.82366753e-01
4.28267419e-01 -4.32159781e-01 -6.01889849e-01 7.91768789e-01
7.80581057e-01 -8.27014089e-01 1.35735321e+00 -1.45504460e-01
5.20404577e-02 -8.19931209e-01 -7.29756773e-01 -7.44142354e-01
-5.80842078e-01 -9.89769995e-02 1.46385834e-01 7.91275203e-01
5.86146295e-01 -6.88509703e-01 6.37143254e-01 1.40717492e-01
-3.21568608e-01 -3.41981322e-01 -9.95694578e-01 -1.08863533e+00
1.76350415e-01 -5.60056925e-01 6.05645955e-01 1.01920485e+00
-8.87754142e-01 -8.87543857e-02 -4.14482623e-01 1.06444252e+00
6.61268592e-01 1.76043198e-01 5.96405029e-01 -1.35791302e+00
-1.96469709e-01 -7.11552382e-01 -2.30667830e-01 -9.64361608e-01
-3.89089018e-01 -8.24216664e-01 3.48487854e-01 -1.55663621e+00
3.35474074e-01 -8.75104189e-01 -5.21071494e-01 8.50597143e-01
-3.81369442e-01 7.34429896e-01 1.33709565e-01 4.40265566e-01
-5.39217710e-01 2.33879909e-01 6.34584248e-01 -3.29880983e-01
-4.05084938e-01 2.98689622e-02 -7.59701550e-01 7.54909277e-01
1.32620335e+00 -6.85882926e-01 -1.36110023e-01 -6.23268306e-01
4.89342570e-01 -5.80505729e-01 9.98357117e-01 -1.18739390e+00
1.38051346e-01 -5.96594334e-01 6.18192077e-01 -9.70152736e-01
4.90595587e-02 -8.27931464e-01 3.55095297e-01 3.93042982e-01
-7.13122310e-03 -2.35344812e-01 5.44585288e-01 7.55484819e-01
2.27517277e-01 -3.78087536e-02 9.98347640e-01 -1.92056537e-01
-1.23034608e+00 4.61562127e-01 -3.09111506e-01 -7.94994161e-02
8.35523307e-01 -3.67962629e-01 -6.00104153e-01 -2.64236368e-02
-5.59998095e-01 2.61096448e-01 2.43429422e-01 2.88053960e-01
5.14506817e-01 -9.08276558e-01 -1.06995583e+00 -9.29018855e-02
5.36226273e-01 1.86841369e-01 4.55227852e-01 6.29181802e-01
-5.23784935e-01 7.65111670e-02 1.29461646e-01 -7.68078327e-01
-9.05797422e-01 1.17361084e-01 7.00087309e-01 6.31172815e-03
-7.86730230e-01 9.68410492e-01 4.16895360e-01 -6.58039510e-01
1.15077384e-01 -5.22628367e-01 -2.19919294e-01 1.80366129e-01
9.21307445e-01 4.61871356e-01 -9.70604736e-03 -6.70570076e-01
-9.06096697e-01 6.57417119e-01 1.62960440e-01 8.68539214e-02
1.37594688e+00 -7.83791617e-02 1.15368359e-01 1.97769761e-01
5.66910803e-01 6.24106787e-02 -1.49841750e+00 -3.12532783e-01
-2.03857481e-01 -5.41919529e-01 3.24233353e-01 -1.30030143e+00
-9.77978051e-01 7.33606219e-01 1.13106620e+00 5.62250428e-02
1.00934589e+00 2.19556212e-01 1.83963344e-01 6.83073342e-01
1.88180223e-01 -7.78804362e-01 -3.11495066e-01 7.90411174e-01
1.01099002e+00 -1.56451881e+00 3.51709276e-01 -2.09814116e-01
-3.52278203e-01 8.12037945e-01 6.29000962e-01 -9.40239355e-02
8.72762859e-01 1.86867774e-01 3.51389915e-01 -2.71282375e-01
-1.84475809e-01 -5.88915825e-01 5.13624072e-01 6.96840823e-01
8.89846310e-02 5.22756219e-01 1.06262922e-01 3.17385286e-01
1.51031658e-01 -3.74386221e-01 2.88356751e-01 7.88543463e-01
-9.27661240e-01 -2.99342424e-01 -6.98575437e-01 5.82715690e-01
-4.45391655e-01 -6.28019094e-01 -3.02037805e-01 8.84201288e-01
2.87416101e-01 7.87092090e-01 8.96097794e-02 -2.41341129e-01
2.90105134e-01 -3.97978783e-01 8.20404589e-02 -6.63551569e-01
-5.66452086e-01 4.46757525e-02 -1.26731932e-01 -3.91137332e-01
-5.30906498e-01 -5.52185595e-01 -1.21953440e+00 -1.97039619e-01
-3.05244565e-01 -1.22504018e-01 5.82231283e-01 5.07654011e-01
2.57911175e-01 2.12698817e-01 2.07409516e-01 -1.19919550e+00
-5.02550662e-01 -1.15897954e+00 -6.27431571e-01 -1.77425668e-01
4.63852137e-01 -5.26905119e-01 -5.64658701e-01 -2.82993346e-01]
|
[9.062548637390137, -0.8962389230728149]
|
e13fec07-3855-4309-a0b4-62414f2c3525
|
deep-rectangling-for-image-stitching-a
|
2203.03831
| null |
https://arxiv.org/abs/2203.03831v4
|
https://arxiv.org/pdf/2203.03831v4.pdf
|
Deep Rectangling for Image Stitching: A Learning Baseline
|
Stitched images provide a wide field-of-view (FoV) but suffer from unpleasant irregular boundaries. To deal with this problem, existing image rectangling methods devote to searching an initial mesh and optimizing a target mesh to form the mesh deformation in two stages. Then rectangular images can be generated by warping stitched images. However, these solutions only work for images with rich linear structures, leading to noticeable distortions for portraits and landscapes with non-linear objects. In this paper, we address these issues by proposing the first deep learning solution to image rectangling. Concretely, we predefine a rigid target mesh and only estimate an initial mesh to form the mesh deformation, contributing to a compact one-stage solution. The initial mesh is predicted using a fully convolutional network with a residual progressive regression strategy. To obtain results with high content fidelity, a comprehensive objective function is proposed to simultaneously encourage the boundary rectangular, mesh shape-preserving, and content perceptually natural. Besides, we build the first image stitching rectangling dataset with a large diversity in irregular boundaries and scenes. Experiments demonstrate our superiority over traditional methods both quantitatively and qualitatively.
|
['Yao Zhao', 'Shuaicheng Liu', 'Kang Liao', 'Chunyu Lin', 'Lang Nie']
|
2022-03-08
| null |
http://openaccess.thecvf.com//content/CVPR2022/html/Nie_Deep_Rectangling_for_Image_Stitching_A_Learning_Baseline_CVPR_2022_paper.html
|
http://openaccess.thecvf.com//content/CVPR2022/papers/Nie_Deep_Rectangling_for_Image_Stitching_A_Learning_Baseline_CVPR_2022_paper.pdf
|
cvpr-2022-1
|
['image-stitching']
|
['computer-vision']
|
[ 6.72406137e-01 7.86636919e-02 -6.11909553e-02 -1.11090019e-01
-3.95983011e-01 -3.64637643e-01 4.38355625e-01 -2.71648228e-01
4.09473106e-02 5.46713769e-01 4.24592942e-02 1.46145090e-01
-6.27319049e-03 -8.45368862e-01 -9.02159572e-01 -7.50854731e-01
3.84157658e-01 2.10463867e-01 1.60078958e-01 -2.82865256e-01
2.98552990e-01 5.05432427e-01 -1.44434071e+00 1.78150713e-01
1.30737257e+00 1.05743229e+00 2.92030334e-01 1.92544550e-01
-1.54840490e-02 3.53442132e-01 -3.07949543e-01 -4.28082466e-01
4.54012871e-01 -3.02723438e-01 -5.32714248e-01 7.04587400e-01
7.74889827e-01 -4.99194235e-01 -2.60005116e-01 1.18734884e+00
3.71036112e-01 6.28225282e-02 5.24542212e-01 -1.06107497e+00
-8.28179359e-01 3.93449932e-01 -1.09507775e+00 -4.56422061e-01
2.51708001e-01 2.55134404e-01 7.33323753e-01 -8.23587298e-01
5.35310328e-01 1.34697843e+00 8.03785741e-01 3.74789089e-01
-1.40517819e+00 -7.15400636e-01 1.29336945e-03 -2.33854428e-01
-1.21300793e+00 -2.87700027e-01 1.41564560e+00 -3.33170444e-01
1.67180404e-01 3.54891211e-01 6.13641322e-01 8.86300266e-01
3.06897461e-01 5.29815257e-01 1.09852338e+00 -1.16456598e-01
-8.62646326e-02 -8.85387659e-02 -5.03598213e-01 6.72072768e-01
6.22529350e-02 2.01107755e-01 -1.43831056e-02 2.44435728e-01
1.24930954e+00 2.34438375e-01 -5.44689775e-01 -5.90528190e-01
-1.38636672e+00 4.58703130e-01 7.01230049e-01 1.96062654e-01
-1.28838316e-01 6.97467402e-02 2.42082521e-01 1.17679596e-01
4.80683684e-01 5.92191160e-01 -1.12291455e-01 3.83819312e-01
-1.01848269e+00 1.68871626e-01 1.87873170e-01 9.32548046e-01
8.38912427e-01 2.32250050e-01 -3.50071080e-02 1.06054807e+00
1.62320033e-01 4.39941823e-01 1.52289972e-01 -1.04187202e+00
4.34174687e-01 7.20599711e-01 1.94352955e-01 -1.37956285e+00
-2.77772784e-01 -3.65990072e-01 -1.37492537e+00 3.11125487e-01
2.55754143e-01 3.26727740e-02 -8.55935872e-01 1.28308082e+00
4.77347255e-01 1.17303357e-01 -1.64386392e-01 9.68160093e-01
8.68572772e-01 9.49793458e-01 -3.02558899e-01 -3.14186633e-01
1.31231487e+00 -1.16262352e+00 -8.28939795e-01 -2.27153182e-01
2.97630012e-01 -1.03672695e+00 1.26144469e+00 5.81248522e-01
-1.30319726e+00 -7.21781373e-01 -1.00538707e+00 -3.44367981e-01
1.23339124e-01 2.76122719e-01 3.41996849e-01 2.66961455e-01
-9.27761912e-01 5.92926800e-01 -3.49248528e-01 1.20956361e-01
4.91670877e-01 1.44577742e-01 -1.76432714e-01 -1.94344178e-01
-9.54555273e-01 5.84281564e-01 3.90392244e-01 1.85666919e-01
-4.49208081e-01 -8.64547729e-01 -1.01556408e+00 -7.19532669e-02
4.30401295e-01 -6.71045363e-01 7.68073022e-01 -1.23854291e+00
-1.62514710e+00 8.15954328e-01 2.24352986e-01 4.18880470e-02
7.15433359e-01 4.46014479e-02 -3.46694291e-01 7.25462884e-02
7.75806606e-02 8.15779805e-01 1.23412263e+00 -1.80083287e+00
-2.98164219e-01 -1.35984376e-01 -1.56470358e-01 3.27391833e-01
-3.90013933e-01 -3.51222873e-01 -5.64242661e-01 -1.14091718e+00
6.49116114e-02 -7.41070628e-01 -2.25412279e-01 4.12502199e-01
-4.94901806e-01 1.21364288e-01 1.01613009e+00 -6.37288809e-01
1.25382066e+00 -2.16075253e+00 3.64736885e-01 -5.70454821e-02
4.04284686e-01 2.93167293e-01 -2.84084976e-01 2.54481256e-01
-9.34514254e-02 1.53946392e-02 -5.18527210e-01 -4.26731944e-01
-3.69865328e-01 1.25444442e-01 -4.10317510e-01 4.54736441e-01
2.72392035e-01 1.01573908e+00 -9.13044095e-01 -6.20795786e-01
4.93097514e-01 5.19202113e-01 -6.97787404e-01 1.71991304e-01
-3.20939720e-01 6.43434286e-01 -4.53455925e-01 7.44750738e-01
1.17545164e+00 -2.26773754e-01 -1.90818265e-01 -7.65962243e-01
-3.89917195e-01 -5.35960615e-01 -1.05159688e+00 1.68522990e+00
-5.47652125e-01 4.64624584e-01 3.25906992e-01 -9.55024123e-01
1.19945383e+00 -1.77434683e-01 7.06757426e-01 -6.50013089e-01
3.12545657e-01 4.22642648e-01 -4.59643781e-01 -5.52974761e-01
6.00869358e-01 -2.41104737e-01 3.51489037e-02 8.97199512e-02
-4.63758081e-01 -6.22444093e-01 -2.54795551e-01 -1.36821300e-01
3.72313738e-01 2.02397127e-02 -1.73801735e-01 -2.79327452e-01
6.45835400e-01 -1.91717386e-01 4.71127659e-01 2.13908833e-02
2.17923969e-01 1.20372188e+00 3.77193630e-01 -7.68665373e-01
-1.44557178e+00 -9.27782536e-01 -3.78271490e-01 2.71568269e-01
8.04554343e-01 -8.55210349e-02 -8.13122571e-01 -3.80370826e-01
-2.70010740e-01 5.14572620e-01 -5.76298237e-01 -2.02273980e-01
-8.32872331e-01 -2.74182528e-01 1.14912204e-01 2.06077904e-01
1.08890676e+00 -9.79244292e-01 -4.00209248e-01 9.47623625e-02
-3.95662099e-01 -1.10228586e+00 -1.17424786e+00 -3.40427428e-01
-8.21098208e-01 -9.27295566e-01 -1.09109783e+00 -1.12996030e+00
9.04744208e-01 6.10875547e-01 9.53700781e-01 2.48690203e-01
-3.27928424e-01 -1.10143468e-01 -1.36029333e-01 1.18467823e-01
-3.73246342e-01 -2.63286054e-01 -1.76729709e-01 2.60914743e-01
-6.54749811e-01 -6.22791946e-01 -9.38261509e-01 6.77135170e-01
-1.48447549e+00 6.70505643e-01 5.88025033e-01 1.02413094e+00
6.09305203e-01 3.54563087e-01 2.32708603e-01 -4.83920634e-01
3.61493081e-01 -1.43743947e-01 -7.08402812e-01 2.75425047e-01
-5.26854992e-01 -3.80580276e-02 8.54458213e-01 -8.04873586e-01
-1.09083009e+00 3.71359475e-02 -1.38556987e-01 -8.03804398e-01
1.16433717e-01 1.52257562e-01 -3.41374278e-01 -2.75278509e-01
3.33969980e-01 4.20990884e-01 2.47102767e-01 -3.46207649e-01
3.79209995e-01 3.86836320e-01 6.13145173e-01 -5.14726222e-01
1.03091335e+00 6.32319927e-01 -1.01273740e-02 -9.49942291e-01
-7.91091859e-01 -3.48724797e-02 -5.12099624e-01 -5.25913596e-01
7.76672363e-01 -6.98439837e-01 -7.14960694e-01 7.36136913e-01
-1.19195330e+00 -4.03827369e-01 -3.01004857e-01 1.64053470e-01
-5.95481932e-01 8.11516404e-01 -3.96600723e-01 -3.09669167e-01
-5.00476480e-01 -1.30891931e+00 1.31397927e+00 2.88983345e-01
1.69753581e-01 -8.44304621e-01 -3.21393348e-02 6.65677071e-01
4.53357816e-01 5.60928583e-01 7.63772845e-01 5.14391541e-01
-6.57344460e-01 2.28523418e-01 -5.72774053e-01 2.88170904e-01
2.53660321e-01 7.73544461e-02 -6.23850465e-01 -3.98683965e-01
8.48889258e-03 -3.74717236e-01 7.77449608e-01 4.94270146e-01
1.50225449e+00 -4.70746577e-01 -1.98070198e-01 1.19720769e+00
1.47762990e+00 1.15543596e-01 8.79628479e-01 2.23963603e-01
1.05932593e+00 6.36542380e-01 5.57800174e-01 4.01607364e-01
1.54144704e-01 8.71218801e-01 6.22525156e-01 -5.75787485e-01
-3.13504934e-01 -4.23307568e-01 1.53739586e-01 5.87277234e-01
2.61445642e-02 -4.26843226e-01 -5.94493508e-01 4.19532746e-01
-1.57247066e+00 -7.10429728e-01 -2.11668059e-01 2.07714033e+00
9.52948809e-01 -6.80052564e-02 -1.47984192e-01 -1.00454902e-02
8.95834863e-01 3.09541017e-01 -7.09112763e-01 -1.41731516e-01
-1.87348828e-01 -1.62972435e-01 4.37564373e-01 5.47021449e-01
-9.95899320e-01 1.06343830e+00 5.28389120e+00 1.36427248e+00
-1.38894641e+00 -3.94990683e-01 8.94607723e-01 1.36137396e-01
-7.07670867e-01 -5.57305850e-02 -4.71213937e-01 7.08951473e-01
6.39810264e-02 2.09867835e-01 5.47501087e-01 4.60018426e-01
3.21048051e-01 1.37468260e-02 -6.08909667e-01 1.36629212e+00
2.48666364e-03 -1.52218151e+00 2.89753795e-01 1.17311954e-01
1.14572978e+00 -3.96126151e-01 2.81243712e-01 -1.52115643e-01
-7.31462762e-02 -1.11977625e+00 8.42164814e-01 5.53622484e-01
1.32044208e+00 -8.13927829e-01 2.78946459e-01 1.78550079e-01
-1.30262852e+00 1.41277816e-02 -3.84900421e-01 2.88330197e-01
2.61704415e-01 6.81909978e-01 -1.21439114e-01 5.67474902e-01
5.34859300e-01 9.87155676e-01 -3.29738706e-01 9.68160033e-01
1.14112280e-01 -3.09036556e-03 -1.03919759e-01 3.59702438e-01
2.99679071e-01 -6.14983559e-01 5.72603285e-01 7.77849138e-01
5.42872846e-01 2.99614936e-01 3.30456197e-01 1.15829957e+00
-1.81279466e-01 1.21662095e-01 -6.93793714e-01 2.48240963e-01
3.51982832e-01 1.38413644e+00 -9.14760232e-01 -1.39993832e-01
-2.45241612e-01 1.16643023e+00 1.81443095e-01 2.09341437e-01
-9.38643754e-01 -2.90208101e-01 3.90779763e-01 4.96424347e-01
8.93020779e-02 -2.40777761e-01 -3.97403032e-01 -1.33806312e+00
7.60866776e-02 -9.63695168e-01 -1.70121074e-01 -9.51311707e-01
-1.23951733e+00 5.59420288e-01 -4.89837341e-02 -1.65409017e+00
4.40270513e-01 -2.65141606e-01 -7.09577918e-01 6.22675538e-01
-1.59403276e+00 -1.43623376e+00 -6.48229778e-01 5.15295684e-01
7.96593845e-01 1.75343633e-01 2.38691062e-01 2.92802155e-01
-5.61866105e-01 4.51198012e-01 1.69829845e-01 1.61458049e-02
8.93476546e-01 -8.96192312e-01 2.62531251e-01 6.35830224e-01
-4.70227242e-01 2.89968461e-01 5.87286472e-01 -7.73234189e-01
-1.38087404e+00 -1.32563269e+00 3.15827757e-01 1.14206083e-01
3.98378044e-01 -1.70551091e-01 -1.07534480e+00 1.69628128e-01
4.02021080e-01 7.82433227e-02 -3.95487025e-02 -7.09476650e-01
-5.26016988e-02 -1.59426644e-01 -1.17145240e+00 8.79055798e-01
1.09317625e+00 -5.10564484e-02 -1.45694017e-01 1.49099454e-01
8.12070131e-01 -6.12315118e-01 -9.75353897e-01 5.81380963e-01
5.89395165e-01 -1.08283019e+00 1.16697061e+00 1.26365498e-01
9.62509632e-01 -4.66636688e-01 2.91040361e-01 -1.37084961e+00
-3.34198892e-01 -9.80657399e-01 2.18484864e-01 1.28196800e+00
-5.42409979e-02 -5.17543256e-01 6.40435219e-01 4.80850786e-01
-3.83222044e-01 -1.00046682e+00 -6.00897789e-01 -6.12294197e-01
-6.58024219e-04 3.56250219e-02 6.45663440e-01 1.17211866e+00
-4.44685400e-01 1.82918817e-01 -7.07229316e-01 2.03618314e-02
8.85153890e-01 3.39831829e-01 8.01377475e-01 -1.06632447e+00
1.03801146e-01 -5.77783525e-01 -4.27843481e-02 -1.47217274e+00
-1.05814196e-01 -5.79614162e-01 4.81191464e-02 -1.31626534e+00
4.74431962e-02 -8.25258791e-01 4.54446256e-01 1.87390745e-01
-1.11506484e-01 4.28476572e-01 -5.11348201e-03 3.57839644e-01
-1.70207262e-01 9.55958784e-01 2.20137024e+00 -3.26354891e-01
-2.43404463e-01 -2.08612263e-01 -6.62796319e-01 7.73775280e-01
7.25817204e-01 -8.93308781e-03 -5.88774502e-01 -6.60429835e-01
2.07740054e-01 2.45484114e-01 4.04631913e-01 -8.35899532e-01
5.59796281e-02 -4.66936201e-01 5.53270400e-01 -8.05455446e-01
3.22084844e-01 -9.02815700e-01 3.29186171e-01 4.24582332e-01
-3.43863547e-01 -1.49453789e-01 1.53334096e-01 5.66323578e-01
-1.62992775e-01 -3.39117236e-02 1.39223695e+00 6.19742647e-02
-3.35216016e-01 5.54843485e-01 8.15643892e-02 4.78099324e-02
1.18552709e+00 -5.27411520e-01 -1.43727005e-01 -2.20321253e-01
-2.54593611e-01 3.78538102e-01 1.03838241e+00 4.46754366e-01
1.01468325e+00 -1.56002128e+00 -6.58458829e-01 5.70046782e-01
-1.38471514e-01 5.64627290e-01 5.88649809e-01 6.66790545e-01
-8.54313731e-01 1.42534757e-02 -3.49427491e-01 -6.25437915e-01
-1.01323378e+00 7.68577874e-01 3.55513722e-01 -1.10397771e-01
-8.77107978e-01 6.64013028e-01 6.08846486e-01 -3.72092277e-01
1.60871089e-01 -5.26635706e-01 -2.28167906e-01 -8.24239403e-02
3.33520174e-01 1.56154707e-01 -2.25122407e-01 -7.62989521e-01
1.26199141e-01 1.30351067e+00 -6.07094504e-02 2.42620811e-01
1.26882339e+00 -3.39741200e-01 -1.54990733e-01 -6.12966157e-02
1.34863353e+00 1.54878259e-01 -1.73941076e+00 -2.34191716e-01
-6.10891759e-01 -8.52194667e-01 2.68053412e-01 -2.92706072e-01
-1.42124498e+00 7.38673210e-01 4.23456550e-01 8.65474939e-02
1.23809099e+00 -2.05117434e-01 1.39702499e+00 -1.72784731e-01
1.04616791e-01 -8.17836344e-01 5.11592686e-01 1.08030148e-01
1.23395514e+00 -1.21283674e+00 1.34103879e-01 -6.20638311e-01
-6.08934939e-01 1.35410893e+00 7.01789737e-01 -2.10460708e-01
4.34813321e-01 2.64629692e-01 -2.23382980e-01 -2.13867635e-01
-2.63059258e-01 2.15801582e-01 5.40320337e-01 4.95393515e-01
1.62146106e-01 -1.19291447e-01 -1.87116131e-01 1.13370173e-01
-2.07974836e-02 -1.30876884e-01 4.58575726e-01 4.49846238e-01
-4.05960262e-01 -9.51740563e-01 -5.64142048e-01 2.98415244e-01
-1.62655637e-01 -3.69434431e-02 -1.31443247e-01 6.36553645e-01
2.54844397e-01 7.36414909e-01 5.07186130e-02 -4.12866145e-01
3.09730232e-01 -7.99099267e-01 5.07360995e-01 -3.17276329e-01
-3.73133481e-01 4.16311711e-01 -2.68454254e-01 -4.53769982e-01
-2.29572892e-01 -2.72994131e-01 -9.81600344e-01 -3.96855742e-01
-3.73446643e-01 -2.83693850e-01 1.80582002e-01 5.93789637e-01
2.28799269e-01 3.77557158e-01 1.01129901e+00 -1.29479551e+00
-2.16347232e-01 -5.26068568e-01 -5.04326522e-01 4.74136353e-01
4.04809892e-01 -6.17218077e-01 -3.40715766e-01 2.42049247e-01]
|
[9.377681732177734, -2.3122763633728027]
|
0f6b1771-d002-4621-81eb-7f820f31934d
|
the-impact-of-extraneous-variables-on-the
|
1904.01125
| null |
http://arxiv.org/abs/1904.01125v1
|
http://arxiv.org/pdf/1904.01125v1.pdf
|
The Impact of Extraneous Variables on the Performance of Recurrent Neural Network Models in Clinical Tasks
|
Electronic Medical Records (EMR) are a rich source of patient information,
including measurements reflecting physiologic signs and administered therapies.
Identifying which variables are useful in predicting clinical outcomes can be
challenging. Advanced algorithms such as deep neural networks were designed to
process high-dimensional inputs containing variables in their measured form,
thus bypass separate feature selection or engineering steps. We investigated
the effect of extraneous input variables on the predictive performance of
Recurrent Neural Networks (RNN) by including in the input vector extraneous
variables randomly drawn from theoretical and empirical distributions. RNN
models using different input vectors (EMR variables; EMR and extraneous
variables; extraneous variables only) were trained to predict three clinical
outcomes: in-ICU mortality, 72-hour ICU re-admission, and 30-day ICU-free days.
The measured degradations of the RNN's predictive performance with the addition
of extraneous variables to EMR variables were negligible.
|
['Eugene Laksana', 'David Ledbetter', 'Cameron Carlin', 'Randall Wetzel', 'Melissa Aczon', 'Long Ho']
|
2019-04-01
| null | null | null | null |
['icu-mortality']
|
['medical']
|
[ 2.30323046e-01 -1.27364293e-01 -2.26198304e-02 -2.88318247e-01
-3.67896527e-01 -2.21687004e-01 2.86541253e-01 4.86452758e-01
-6.08740449e-01 8.67250144e-01 4.99487907e-01 -6.50152266e-01
-5.77705503e-01 -7.63472617e-01 -3.87140870e-01 -5.80713630e-01
-3.51543844e-01 7.70476043e-01 -5.70693731e-01 3.91500331e-02
-1.06589729e-02 7.82223582e-01 -1.30252957e+00 7.38641545e-02
3.64299715e-01 9.68906999e-01 -1.52254581e-01 8.96408796e-01
-6.42816350e-02 9.74698663e-01 -7.77893364e-01 6.89278319e-02
3.00582021e-01 -5.69644690e-01 -3.10629606e-01 -3.57820511e-01
-3.27333897e-01 -3.22587159e-03 -5.10320723e-01 5.12637794e-01
7.79194951e-01 1.14337444e-01 7.53157139e-01 -8.82613122e-01
-6.53913260e-01 6.82273805e-01 1.26318574e-01 1.73410565e-01
-3.31794806e-02 3.33290607e-01 7.04885066e-01 -6.57378256e-01
3.36911738e-01 8.80226433e-01 8.08443546e-01 7.04657257e-01
-1.17125905e+00 -3.25193286e-01 -2.64086664e-01 -1.11515731e-01
-1.13556039e+00 -2.60811657e-01 6.10271454e-01 -7.16559291e-01
8.91759396e-01 4.00123388e-01 4.22041118e-01 1.32081056e+00
5.65474153e-01 2.31093213e-01 7.23458111e-01 -8.50746930e-02
4.01104182e-01 4.15076077e-01 5.38905442e-01 2.25479916e-01
3.85552377e-01 6.20185971e-01 -2.71310121e-01 -3.85086864e-01
8.91885698e-01 8.56958389e-01 -4.15625215e-01 1.22416563e-01
-1.31787801e+00 9.18532073e-01 1.78720340e-01 2.28608042e-01
-8.74296665e-01 1.24309972e-01 5.26582837e-01 6.71389341e-01
3.20520550e-02 9.99315798e-01 -1.06990325e+00 -4.43437397e-01
-6.42530978e-01 -4.41430569e-01 6.60674870e-01 7.23825634e-01
4.05675828e-01 3.88025641e-01 -4.06068087e-01 1.00613272e+00
-2.46796012e-01 5.41727245e-01 1.24939609e+00 -5.22928178e-01
2.54238665e-01 6.95633650e-01 1.13755085e-01 -8.44120324e-01
-1.05078554e+00 -6.29307628e-01 -1.47592342e+00 -1.91185102e-01
-2.55532586e-03 -6.69394851e-01 -1.13306880e+00 1.26980162e+00
-3.34383249e-01 -2.95139760e-01 2.99083978e-01 7.32623696e-01
8.44688237e-01 4.95940596e-01 5.56029677e-02 -4.31070924e-01
9.85698938e-01 -4.14939195e-01 -7.51267374e-01 5.25907613e-02
7.54191816e-01 -1.58681899e-01 7.94482470e-01 2.45221257e-01
-8.95969391e-01 -4.32235032e-01 -5.30850649e-01 3.56709838e-01
-3.60457361e-01 9.16651711e-02 3.19835037e-01 3.39589894e-01
-9.49463189e-01 1.09426701e+00 -7.94366062e-01 1.64926425e-01
3.15779507e-01 6.60318375e-01 -4.69973624e-01 1.87319905e-01
-1.45857668e+00 9.68891382e-01 2.18596429e-01 3.40044320e-01
-6.52850986e-01 -8.42794478e-01 -8.23709905e-01 3.04264545e-01
2.58766711e-02 -9.81460154e-01 9.18790996e-01 -8.81856263e-01
-1.44188058e+00 3.04990828e-01 -1.60011619e-01 -6.92117155e-01
3.74221176e-01 -3.34297508e-01 -4.46044773e-01 -5.48896156e-02
-5.22446930e-01 -2.32665800e-02 7.07872629e-01 -6.68373168e-01
5.74174896e-02 -4.22585636e-01 -6.91216528e-01 -9.78960693e-02
-2.55517721e-01 -2.20800921e-01 3.52239430e-01 -7.38976061e-01
9.28276554e-02 -8.85269344e-01 -6.67944849e-01 -4.35500860e-01
-3.41616660e-01 1.79225117e-01 2.82157809e-01 -5.27554333e-01
1.33559346e+00 -2.05478168e+00 1.29782394e-01 2.81974405e-01
6.12615466e-01 3.80029112e-01 -1.75341830e-01 4.75957274e-01
-4.70223665e-01 2.92590559e-01 3.63980755e-02 -5.51902205e-02
-2.43829593e-01 1.58975124e-01 -9.13272128e-02 3.32695752e-01
4.32080448e-01 1.09479213e+00 -7.49917150e-01 -4.77151163e-02
5.45255959e-01 6.82190478e-01 -3.53454798e-01 4.38218266e-01
7.37829059e-02 4.61772382e-01 -4.79753733e-01 3.87948155e-01
-7.99697936e-02 -5.62372148e-01 -1.44123361e-01 7.33747259e-02
2.75841296e-01 3.36360842e-01 -6.29857779e-01 7.10733294e-01
-5.33233881e-01 7.73364663e-01 -4.59810793e-01 -8.25468838e-01
1.39727819e+00 5.77196538e-01 8.01146984e-01 -6.97091639e-01
6.68405771e-01 -4.38270159e-02 3.49504948e-01 -5.16485035e-01
2.46044338e-01 -4.89930063e-01 8.81679729e-02 3.54554951e-01
-1.20056935e-01 4.77880359e-01 -3.83176029e-01 -5.91975488e-02
1.44759881e+00 -6.39815509e-01 4.25148249e-01 -1.08658120e-01
1.49750412e-01 -3.03408265e-01 6.16034448e-01 9.81090486e-01
-1.55580655e-01 9.33863163e-01 6.89972222e-01 -7.00188339e-01
-9.97742355e-01 -8.62434030e-01 -3.64527404e-01 6.07819200e-01
-5.10347605e-01 -2.63108045e-01 6.56688213e-02 -2.13994205e-01
2.33098879e-01 8.25253427e-01 -1.03588128e+00 -6.82107329e-01
-4.33325797e-01 -1.09795225e+00 3.63598555e-01 8.09951246e-01
-4.90370899e-01 -1.41555309e+00 -1.01738358e+00 6.28230274e-01
2.20686555e-01 -5.91554224e-01 -5.92620075e-02 8.95560145e-01
-1.12989080e+00 -1.03139818e+00 -8.49753082e-01 -1.53688908e-01
5.41829646e-01 -4.89402801e-01 1.18975532e+00 -8.38148519e-02
-5.84015191e-01 1.96214139e-01 -2.96826392e-01 -6.02484584e-01
-7.41931498e-01 7.32579008e-02 1.68534309e-01 -4.92988415e-02
5.11550248e-01 -3.90790790e-01 -7.81044602e-01 1.11711904e-01
-8.84367645e-01 -3.81450243e-02 6.52581275e-01 1.28355479e+00
4.89776254e-01 -4.33806628e-01 1.05986547e+00 -1.20575738e+00
8.47117364e-01 -7.40270436e-01 -1.75030887e-01 -4.12547588e-02
-8.48142922e-01 3.58784169e-01 1.19035780e+00 -6.25226438e-01
-2.51645923e-01 -1.16485164e-01 4.45314636e-03 -1.00029051e+00
-1.88948795e-01 5.99632204e-01 3.91418368e-01 7.40465522e-01
7.08718479e-01 3.72745574e-01 1.73432603e-01 -4.20346200e-01
-1.44296378e-01 6.41935825e-01 2.85487115e-01 -1.79312274e-01
1.69263944e-01 -1.10361859e-01 2.97396421e-01 -6.44980013e-01
-2.70122319e-01 -4.80695009e-01 -4.89340186e-01 1.05235621e-01
5.67187607e-01 -6.57815754e-01 -9.64435101e-01 2.88511455e-01
-7.49707699e-01 -2.66454667e-01 -6.33310914e-01 8.16411018e-01
-4.29128051e-01 -2.63631463e-01 -7.88697660e-01 -8.07162642e-01
-7.86469042e-01 -1.12003326e+00 5.65826833e-01 8.47065225e-02
-7.03412771e-01 -1.08436632e+00 2.01573223e-01 -2.96770573e-01
7.57774234e-01 4.80401516e-01 1.29955602e+00 -1.25933599e+00
2.01434866e-02 -5.30059099e-01 4.33761068e-02 5.05234361e-01
4.82413381e-01 4.76180837e-02 -6.65732682e-01 -1.54462978e-01
2.07540005e-01 4.02886793e-02 7.51029789e-01 7.35118330e-01
1.24311054e+00 -4.17510778e-01 -1.93101481e-01 6.64123476e-01
1.35189342e+00 5.79047620e-01 6.57081664e-01 1.08689085e-01
5.52376449e-01 2.52063274e-01 -1.60613451e-02 7.63182700e-01
-2.26316109e-01 -4.98533584e-02 1.95091143e-01 -2.42201984e-01
3.11377734e-01 3.52134965e-02 1.60110638e-01 1.10340655e+00
9.89088714e-02 -2.55069256e-01 -1.10824549e+00 5.26372671e-01
-1.41960168e+00 -7.49461055e-01 -1.77205607e-01 2.31749797e+00
7.28664398e-01 2.35529449e-02 -9.35767367e-02 1.90715164e-01
5.82658410e-01 -2.20228851e-01 -9.89931226e-01 -5.96233010e-01
-2.42688119e-01 3.46732378e-01 6.37057900e-01 1.01988837e-01
-6.88925683e-01 1.69162273e-01 6.70532608e+00 -1.34565815e-01
-1.44453013e+00 -3.97318006e-01 1.02948201e+00 -4.89624143e-01
-2.27472454e-01 -6.06237292e-01 -5.38252413e-01 5.89662313e-01
1.71510231e+00 -1.70681342e-01 2.50829756e-01 7.47894943e-01
5.45396924e-01 1.99253857e-01 -1.37451744e+00 1.03291690e+00
-1.47862807e-01 -1.37669301e+00 3.78835648e-02 -1.72656283e-01
5.47571301e-01 2.89758295e-01 1.38869017e-01 3.63423795e-01
3.63969922e-01 -1.52173924e+00 -1.29715562e-01 9.96376157e-01
1.12482417e+00 -6.31958306e-01 1.17922020e+00 9.04479921e-02
-3.75485361e-01 -3.14884990e-01 -2.21468315e-01 -8.38681236e-02
4.80744764e-02 7.98629582e-01 -1.08027852e+00 8.15368742e-02
3.89994621e-01 7.33140171e-01 -4.59401697e-01 7.49840796e-01
5.11226058e-01 8.57787311e-01 -2.33021975e-01 -2.02116936e-01
9.66881737e-02 -1.02141961e-01 4.32459235e-01 1.00725579e+00
1.83120415e-01 1.86176240e-01 -2.92628586e-01 7.91092813e-01
-1.66780978e-01 -4.25600223e-02 -6.92417502e-01 -2.98464596e-01
3.81567299e-01 8.24783266e-01 -4.71063733e-01 -4.06644821e-01
-1.36101559e-01 5.46315432e-01 8.68070200e-02 5.78568757e-01
-5.01988232e-01 -5.60505629e-01 8.05777907e-01 3.62758577e-01
8.30853507e-02 1.23101950e-01 -6.16728306e-01 -1.04421651e+00
-1.49052024e-01 -8.50595117e-01 3.07655960e-01 -6.39460564e-01
-1.34899771e+00 9.24049199e-01 -4.41147238e-01 -1.11149192e+00
-7.70874202e-01 -6.73134744e-01 -4.20937598e-01 1.33500969e+00
-1.29823148e+00 1.42152682e-02 -5.64672537e-02 5.62530339e-01
1.77463040e-01 -3.97535741e-01 1.15962839e+00 2.47912109e-01
-7.75023520e-01 6.19830906e-01 6.44545138e-01 5.31537652e-01
5.81139505e-01 -9.27070200e-01 1.37952566e-01 1.58570021e-01
-3.92848998e-01 1.00056136e+00 5.58845103e-01 -4.86098230e-01
-1.37454414e+00 -1.39392614e+00 7.46807814e-01 -6.11182392e-01
5.32577693e-01 4.68285717e-02 -9.69916165e-01 8.28083277e-01
-4.56732720e-01 -2.30155420e-02 9.98171151e-01 1.85736209e-01
-3.05232089e-02 1.38596743e-01 -9.78289247e-01 4.82587099e-01
5.51846981e-01 -4.74812090e-01 -7.45333970e-01 7.50297308e-02
7.66367376e-01 -1.41126245e-01 -1.24171877e+00 6.98372960e-01
3.93130571e-01 -5.77059925e-01 8.63933325e-01 -1.18127179e+00
6.57447398e-01 8.11464936e-02 7.02438504e-02 -1.40583205e+00
-3.48507375e-01 -6.12506986e-01 -1.72307312e-01 5.07449508e-01
4.17467743e-01 -1.13095689e+00 6.21028841e-01 1.09959495e+00
-9.56804752e-02 -1.17784655e+00 -7.54016936e-01 -4.86056477e-01
1.97705209e-01 -1.26952171e-01 5.15318394e-01 7.68230677e-01
-4.22942005e-02 5.16075909e-01 -4.93580520e-01 -1.66463286e-01
-1.06656522e-01 -2.53071003e-02 2.48944968e-01 -1.43393528e+00
-3.22369576e-01 -6.29920006e-01 -5.50428867e-01 -3.53835911e-01
-8.32752362e-02 -8.53413463e-01 -1.57541320e-01 -1.44045341e+00
3.71694453e-02 -5.22521675e-01 -8.86785150e-01 5.12356639e-01
-3.94725949e-01 -4.66089755e-01 2.44597405e-01 1.03068151e-01
-3.02116405e-02 6.95465565e-01 8.82565200e-01 1.67437166e-01
-5.73209822e-01 3.58681470e-01 -5.66235900e-01 4.37606484e-01
8.32300782e-01 -7.26673245e-01 -3.77853245e-01 -1.18836753e-01
8.31059292e-02 7.40830958e-01 2.02916235e-01 -7.90278614e-01
2.38182954e-02 -2.43994772e-01 8.64291668e-01 -5.11561692e-01
3.06195974e-01 -7.75484681e-01 4.69947606e-01 4.76950765e-01
-7.46550620e-01 5.15774965e-01 2.90991396e-01 4.32121545e-01
-2.34957993e-01 -1.10156953e-01 5.73474169e-01 -9.27354023e-02
9.75370258e-02 3.43143374e-01 -7.61048257e-01 1.22077540e-01
7.06033170e-01 -2.49844268e-01 -8.23712945e-02 -4.86241549e-01
-9.15646791e-01 7.17941523e-02 -6.11252747e-02 4.76345569e-01
8.93355787e-01 -9.29165065e-01 -7.67144322e-01 4.28460389e-01
-2.63052315e-01 -1.75437734e-01 2.83539832e-01 6.39343798e-01
-3.80078733e-01 5.77536404e-01 -1.41243219e-01 -3.29547852e-01
-1.02764666e+00 8.20061564e-01 5.43042541e-01 -2.25178614e-01
-7.38941193e-01 4.77183521e-01 -1.43092405e-02 -3.87767136e-01
2.65950561e-01 -4.26562190e-01 -1.39002681e-01 6.38872087e-02
5.11232138e-01 4.63956177e-01 2.48844042e-01 -1.47295535e-01
-1.43186256e-01 9.82068777e-02 -6.39447123e-02 3.42833132e-01
1.51545835e+00 2.06251070e-01 1.62328675e-01 1.02982163e+00
1.54559481e+00 -6.30804121e-01 -7.63891220e-01 -1.71428576e-01
-1.03635423e-01 4.29543257e-02 -1.16651244e-02 -1.03597641e+00
-1.15609872e+00 1.06869471e+00 6.48836553e-01 -6.95178881e-02
1.11426365e+00 -3.18174601e-01 6.86959445e-01 5.16663134e-01
8.45305324e-02 -7.72080123e-01 -1.74844176e-01 6.85998797e-01
7.08981991e-01 -1.15764153e+00 -1.58137739e-01 4.26545918e-01
-6.62832439e-01 1.27557814e+00 2.45675638e-01 -1.92808121e-01
1.05129039e+00 1.15708828e-01 4.19849426e-01 -2.30025053e-01
-1.26637185e+00 2.88644612e-01 1.76580399e-01 3.00006926e-01
5.45798004e-01 4.49398696e-01 -1.49625331e-01 9.50088263e-01
-2.70133346e-01 2.59612352e-01 7.65257061e-01 6.64429188e-01
4.99626324e-02 -5.36724925e-01 -2.04906821e-01 1.47207165e+00
-5.16118288e-01 -3.95610005e-01 -1.03543058e-01 5.51637292e-01
-5.32576501e-01 5.58329105e-01 1.60527125e-01 -4.86417025e-01
2.99770087e-01 4.66159880e-01 -2.46908039e-01 -5.53663015e-01
-1.05858171e+00 -3.67139429e-01 -1.77966580e-01 -3.64907682e-01
2.12025404e-01 -5.83363831e-01 -1.30042660e+00 -4.57637571e-02
-1.99891165e-01 1.37957394e-01 5.37964404e-01 5.85936069e-01
8.69557917e-01 1.14561200e+00 6.10060394e-01 -4.22532141e-01
-9.40538466e-01 -1.08161473e+00 -4.99078631e-01 5.91989636e-01
8.90565872e-01 -2.86658436e-01 -6.16409242e-01 -7.24300221e-02]
|
[7.978903770446777, 6.181511878967285]
|
89e4f490-d671-4cac-ad8e-90f170c03279
|
distance-based-hyperspherical-classification
|
2107.02067
| null |
https://arxiv.org/abs/2107.02067v3
|
https://arxiv.org/pdf/2107.02067v3.pdf
|
Distance-based Hyperspherical Classification for Multi-source Open-Set Domain Adaptation
|
Vision systems trained in closed-world scenarios fail when presented with new environmental conditions, new data distributions, and novel classes at deployment time. How to move towards open-world learning is a long-standing research question. The existing solutions mainly focus on specific aspects of the problem (single domain Open-Set, multi-domain Closed-Set), or propose complex strategies which combine several losses and manually tuned hyperparameters. In this work, we tackle multi-source Open-Set domain adaptation by introducing HyMOS: a straightforward model that exploits the power of contrastive learning and the properties of its hyperspherical feature space to correctly predict known labels on the target, while rejecting samples belonging to any unknown class. HyMOS includes style transfer among the instance transformations of contrastive learning to get domain invariance while avoiding the risk of negative-transfer. A self-paced threshold is defined on the basis of the observed data distribution and updates online during training, allowing to handle the known-unknown separation. We validate our method over three challenging datasets. The obtained results show that HyMOS outperforms several competitors, defining the new state-of-the-art. Our code is available at https://github.com/silvia1993/HyMOS.
|
['Tatiana Tommasi', 'Barbara Caputo', 'Francesco Cappio Borlino', 'Silvia Bucci']
|
2021-07-05
| null | null | null | null |
['universal-domain-adaptation']
|
['computer-vision']
|
[ 1.93705276e-01 -1.63860664e-01 -1.04070678e-01 -3.01847577e-01
-6.71716809e-01 -8.37111175e-01 7.61545777e-01 -1.04119234e-01
-5.93485653e-01 9.38034832e-01 -4.64589000e-01 2.21733555e-01
-2.37323925e-01 -6.68604493e-01 -6.89055681e-01 -1.00680745e+00
-3.39736007e-02 9.45185244e-01 4.22707140e-01 -1.30872428e-01
1.78125292e-01 6.18663073e-01 -1.74520743e+00 -4.69087530e-03
1.18295252e+00 9.77052867e-01 3.66503239e-01 4.67284024e-01
-1.44818559e-01 1.44319266e-01 -6.30677998e-01 -2.93670893e-01
6.83903098e-01 -1.28036380e-01 -4.63737816e-01 1.10428799e-02
6.34558201e-01 1.59128025e-01 1.33902490e-01 1.18924940e+00
5.88515222e-01 2.64046341e-01 9.27362263e-01 -1.34406352e+00
-6.92048490e-01 -1.10281669e-01 -5.06621003e-01 2.72314578e-01
-2.19566450e-01 4.69432205e-01 5.82745492e-01 -7.42585421e-01
7.19003141e-01 8.95390749e-01 6.92898512e-01 7.28558064e-01
-1.35957611e+00 -7.64521897e-01 3.45080465e-01 5.08050442e-01
-1.38237274e+00 -3.84020388e-01 6.98990166e-01 -5.84201634e-01
6.43996775e-01 5.53285033e-02 3.48673254e-01 1.44608128e+00
-3.52918021e-02 4.41967785e-01 1.29769242e+00 -2.81085223e-01
4.74825352e-01 5.94847441e-01 1.41158178e-01 3.41476887e-01
2.23990560e-01 3.51252764e-01 -2.87430882e-01 1.47643350e-02
4.08388078e-01 -1.87011793e-01 -3.75889778e-01 -1.03263330e+00
-1.09255946e+00 7.00146556e-01 5.18883109e-01 1.08767651e-01
-2.16002271e-01 -7.08361566e-01 4.67551827e-01 5.55463672e-01
6.97793603e-01 4.23599303e-01 -7.24870265e-01 1.03791572e-01
-5.76628923e-01 4.35957387e-02 8.17103267e-01 8.61164987e-01
9.29041803e-01 3.16881537e-02 -9.69706550e-02 9.57158923e-01
3.28953862e-02 9.24141586e-01 6.25151575e-01 -5.05578637e-01
3.88068527e-01 4.45346355e-01 7.30090588e-02 -5.08367479e-01
-4.20071870e-01 -7.29445457e-01 -6.60799205e-01 3.31463814e-01
5.67413986e-01 -3.23544502e-01 -9.59748268e-01 1.62004602e+00
6.55175507e-01 4.68617976e-01 3.86067331e-01 9.39364910e-01
7.36017644e-01 6.71039402e-01 -4.81196158e-02 -1.44492269e-01
8.60976696e-01 -1.08708012e+00 -3.07364553e-01 -4.94277060e-01
2.42161870e-01 -4.09359276e-01 1.26595867e+00 5.78173697e-01
-3.81309092e-01 -6.88357234e-01 -1.03694451e+00 3.80351096e-01
-8.14253211e-01 2.01663986e-01 1.65075764e-01 4.93627697e-01
-7.21545577e-01 3.41798723e-01 -3.84353220e-01 -5.88804841e-01
4.39786732e-01 2.36773610e-01 -4.31778580e-01 -7.86516666e-02
-1.10072577e+00 9.55547869e-01 6.47739649e-01 -2.31449142e-01
-1.02769113e+00 -7.17907965e-01 -5.78736544e-01 -1.21046551e-01
5.14605224e-01 -5.71442842e-01 1.02910542e+00 -1.40413010e+00
-1.53925562e+00 1.00845993e+00 3.58272970e-01 -5.42641521e-01
7.64245510e-01 -1.91455707e-01 -6.10535204e-01 1.13204807e-01
-3.06729861e-02 6.60557032e-01 1.32079518e+00 -1.31639242e+00
-6.26345873e-01 -5.39818048e-01 -1.82232425e-01 4.96758163e-01
-6.28192604e-01 -3.98021758e-01 -2.01893196e-01 -4.66072291e-01
-2.24450707e-01 -1.02762449e+00 3.43176126e-02 1.27260491e-01
3.68518196e-02 -1.16948910e-01 8.56731951e-01 -3.81245643e-01
5.60049713e-01 -2.31116104e+00 2.22478583e-01 -9.54737961e-02
-1.01459399e-01 7.20074594e-01 -4.23416555e-01 1.52549326e-01
-7.24939406e-02 -4.26308572e-01 -4.64156061e-01 -6.51221797e-02
9.08178464e-02 1.75305650e-01 -3.58508199e-01 5.22503078e-01
2.67857283e-01 4.56821173e-01 -8.23583186e-01 -2.62147069e-01
3.97227019e-01 4.18922722e-01 -1.08722642e-01 3.61114413e-01
-4.86639887e-01 6.95844114e-01 -1.13144346e-01 4.30566907e-01
1.11377907e+00 -7.82469958e-02 -6.89590424e-02 -3.09415031e-02
-2.81025693e-02 -3.12793195e-01 -1.36288583e+00 1.66915476e+00
-4.30693120e-01 4.00455564e-01 2.23361164e-01 -1.13504624e+00
1.29815459e+00 -1.28556803e-01 2.27302492e-01 -7.07738638e-01
2.18710095e-01 3.69437993e-01 -1.19766921e-01 -5.11489093e-01
6.17822073e-02 -2.30966061e-01 2.31072813e-01 -8.77821073e-02
4.84839529e-01 -3.15098435e-01 2.26014182e-01 -3.05289060e-01
6.76622808e-01 4.03678060e-01 3.04742306e-01 -3.51403624e-01
6.96937740e-01 1.12123959e-01 5.90313256e-01 6.58330798e-01
-5.72534978e-01 6.97376907e-01 1.30756691e-01 -4.80818123e-01
-6.82109177e-01 -1.32401264e+00 -3.70489955e-01 1.16190958e+00
3.38975251e-01 3.03295612e-01 -7.59627044e-01 -1.04382980e+00
9.34911668e-02 5.57614684e-01 -5.96489668e-01 -3.54690522e-01
-1.38864741e-01 -8.47159564e-01 2.72459775e-01 1.60717040e-01
6.48301184e-01 -1.08593667e+00 -7.30004370e-01 -1.31661192e-01
-1.49601147e-01 -9.95388269e-01 -8.71868953e-02 4.95921880e-01
-6.94152415e-01 -1.24533713e+00 -9.77995276e-01 -6.91163540e-01
5.34592330e-01 1.33395866e-01 1.01742041e+00 -6.59176111e-01
-3.23978662e-01 4.14473414e-01 -4.18407619e-01 -6.01247728e-01
-3.33094150e-01 2.38136560e-01 1.79986179e-01 4.57747400e-01
5.13034344e-01 -4.90178078e-01 -3.87918741e-01 4.74164069e-01
-8.00393403e-01 -3.01614255e-01 6.49091482e-01 8.81825328e-01
6.83182657e-01 -5.29481545e-02 7.31056154e-01 -8.78974259e-01
2.75166154e-01 -6.93738520e-01 -7.73913085e-01 4.16030496e-01
-5.98697722e-01 -5.37896249e-03 8.30768764e-01 -7.01341331e-01
-1.30470037e+00 2.54291415e-01 6.30532950e-02 -7.01098144e-01
-6.14695609e-01 -9.34664384e-02 -3.31489563e-01 -2.13237390e-01
1.24051344e+00 2.83068657e-01 8.56295452e-02 -3.62348855e-01
5.11461735e-01 6.98806047e-01 6.20672584e-01 -6.34407341e-01
1.12023830e+00 6.31348729e-01 -3.47035527e-01 -8.83400261e-01
-9.44692671e-01 -6.37333155e-01 -9.37488854e-01 -2.66351938e-01
5.63919008e-01 -1.06795800e+00 -1.55625194e-02 6.52511895e-01
-7.53251433e-01 -5.86384356e-01 -7.12777793e-01 4.08867627e-01
-5.56973100e-01 2.08740279e-01 1.26798049e-01 -5.35958111e-01
-3.28609526e-01 -7.59426057e-01 8.94466400e-01 4.92205769e-01
2.81234801e-01 -1.02037036e+00 3.41297358e-01 2.26709291e-01
3.51883590e-01 1.77353531e-01 3.69057447e-01 -9.83087301e-01
-3.38455379e-01 -5.75544462e-02 -2.48212084e-01 7.41490006e-01
9.95643586e-02 -1.55832991e-01 -1.31312537e+00 -5.79794347e-01
9.43881199e-02 -6.56107008e-01 1.03762758e+00 2.60307431e-01
9.32491302e-01 -3.17295524e-03 -3.00502062e-01 9.63095844e-01
1.37472451e+00 1.05278268e-01 4.32814568e-01 6.15200996e-01
4.44932878e-01 7.57264018e-01 9.68931317e-01 5.55824995e-01
2.44051099e-01 4.72266227e-01 6.10643804e-01 -1.07191078e-01
-1.08128235e-01 8.70719701e-02 4.21947807e-01 2.21119151e-01
2.16236934e-01 -2.26504311e-01 -9.80353355e-01 5.84721267e-01
-1.66745365e+00 -7.39646733e-01 7.62054920e-02 2.43010259e+00
6.60314083e-01 2.63975739e-01 9.97439846e-02 -1.84715658e-01
7.62283862e-01 1.47193924e-01 -1.16377580e+00 -8.88856053e-02
-3.98152441e-01 1.26658306e-01 4.92418826e-01 2.95560718e-01
-1.50889659e+00 8.88064861e-01 4.89024496e+00 8.39242399e-01
-1.37541425e+00 2.66924143e-01 2.94585973e-01 -1.32787794e-01
9.57344174e-02 -2.13095427e-01 -7.55518973e-01 4.14086550e-01
7.15345383e-01 -6.59909695e-02 3.66062343e-01 9.42677557e-01
-2.47920856e-01 9.00686532e-02 -9.04695213e-01 1.10917974e+00
4.32972640e-01 -8.77324700e-01 -1.58176720e-01 -2.09514707e-01
7.48982489e-01 5.44339001e-01 2.13203862e-01 5.84422648e-01
1.82807565e-01 -4.29672658e-01 5.03487408e-01 4.29329067e-01
7.89170921e-01 -4.79559958e-01 5.69946349e-01 5.76659143e-01
-9.40568507e-01 -4.25209373e-01 -3.70581269e-01 -2.35239509e-02
-4.02505666e-01 5.36461830e-01 -8.91831458e-01 7.29719281e-01
8.78156543e-01 9.03276742e-01 -8.88944149e-01 1.22401261e+00
-1.25871569e-01 3.89570653e-01 -3.57797354e-01 9.24399272e-02
-3.08139380e-02 -3.20363015e-01 8.21982443e-01 1.14692664e+00
1.41030163e-01 -4.13561046e-01 2.55949169e-01 7.24790990e-01
2.16775045e-01 1.58176735e-01 -8.01073074e-01 4.22020763e-01
2.71866351e-01 1.30925715e+00 -6.11705601e-01 -1.76349968e-01
-2.73124248e-01 1.03667951e+00 4.47281301e-01 3.95641387e-01
-9.19437885e-01 -3.11600417e-01 5.69024563e-01 -3.37030254e-02
4.65799123e-01 3.33375633e-02 -1.32659286e-01 -1.47600532e+00
8.14068615e-02 -8.56368005e-01 7.65777707e-01 -5.45953274e-01
-1.69105089e+00 6.96703255e-01 2.73059737e-02 -1.59611714e+00
5.76253943e-02 -7.07654595e-01 -4.29520726e-01 5.46863377e-01
-1.96308208e+00 -1.02106881e+00 -5.83825350e-01 7.89370179e-01
7.66192853e-01 -5.28738022e-01 7.95393705e-01 3.06533664e-01
-4.54835057e-01 4.69054520e-01 6.68074250e-01 -2.62236297e-01
1.31372261e+00 -1.29063666e+00 2.28202328e-01 7.27866113e-01
1.67627916e-01 4.86419983e-02 6.37104332e-01 -4.21241671e-01
-8.59564781e-01 -1.27180779e+00 2.89842755e-01 -4.69406247e-01
6.18160248e-01 -6.00433946e-01 -1.07492483e+00 4.31865990e-01
5.63794672e-02 4.38172698e-01 5.24733722e-01 5.87977767e-02
-5.60852408e-01 -5.57378769e-01 -1.37512445e+00 2.43698254e-01
9.38845396e-01 -1.58657894e-01 -5.35121799e-01 4.46438015e-01
5.10882854e-01 -3.27382982e-01 -5.35408676e-01 5.56532443e-01
1.83639228e-01 -9.31330740e-01 9.26902235e-01 -4.82082218e-01
-5.80770820e-02 -4.38971907e-01 -1.17875084e-01 -1.64266109e+00
-2.46453404e-01 -2.90023267e-01 -6.25841767e-02 1.21735263e+00
3.37200433e-01 -1.04485059e+00 5.84632635e-01 5.20207733e-02
-5.03906682e-02 -2.78334528e-01 -1.00798130e+00 -1.22856736e+00
2.41980702e-01 3.77374962e-02 2.98636287e-01 1.14326346e+00
-5.62552512e-01 4.07470167e-01 -1.83270514e-01 4.17614460e-01
7.80456245e-01 2.27868468e-01 9.22328651e-01 -1.70389163e+00
-3.12412143e-01 -1.97002515e-01 -2.81172931e-01 -6.36212826e-01
3.56726438e-01 -7.97203660e-01 8.26686993e-02 -1.14675105e+00
7.44534805e-02 -5.73319852e-01 -4.29235905e-01 4.05933529e-01
3.30617130e-02 2.15285957e-01 2.92209655e-01 1.97604641e-01
-6.63973689e-01 9.06802893e-01 1.05606163e+00 -3.86026353e-01
-3.47339451e-01 1.12158485e-01 -4.79241431e-01 6.83240116e-01
1.06537843e+00 -3.69603395e-01 -5.93982220e-01 -3.97851616e-01
-2.35459566e-01 -4.55273509e-01 4.73602593e-01 -1.47489083e+00
3.49492021e-02 -3.25781673e-01 3.17820877e-01 -2.74540633e-01
3.34963858e-01 -9.93043661e-01 -1.90919987e-03 3.48781079e-01
-3.43485586e-02 -5.23283303e-01 4.62682635e-01 7.14900196e-01
-1.01498365e-01 -3.36777866e-01 1.19398117e+00 -4.98583913e-03
-1.08043432e+00 3.51430237e-01 1.36033773e-01 3.77351642e-01
1.40194046e+00 -2.32832357e-01 -5.80570281e-01 2.16452837e-01
-8.00942481e-01 2.74860650e-01 5.62398255e-01 8.15283120e-01
3.54149222e-01 -8.38951290e-01 -8.24953139e-01 4.28526223e-01
6.25816584e-01 1.78832561e-01 3.92599612e-01 5.68124175e-01
-2.91211843e-01 -4.95068766e-02 -5.96451342e-01 -8.12049150e-01
-1.15104938e+00 7.05078602e-01 7.07164705e-01 -1.22120850e-01
-5.94142377e-01 8.08819890e-01 5.09506762e-01 -1.12363219e+00
3.28649968e-01 6.80977479e-02 -4.08505231e-01 2.43634209e-01
4.31292832e-01 2.78646290e-01 1.78906083e-01 -4.04121608e-01
-3.85288417e-01 7.25750268e-01 4.89330404e-02 2.57721126e-01
1.27379394e+00 -1.36306554e-01 2.07182080e-01 6.43072963e-01
9.12064433e-01 -4.61368203e-01 -1.78710997e+00 -4.88953292e-01
-2.07063079e-01 -4.22475874e-01 -2.65547365e-01 -1.24426937e+00
-8.76183867e-01 8.45828116e-01 1.33526969e+00 -5.63049503e-02
1.25219643e+00 -6.85641170e-02 3.79817039e-01 5.59736788e-01
3.28123868e-01 -1.39468443e+00 1.42162710e-01 6.94190621e-01
9.08887446e-01 -1.64298451e+00 -2.20982283e-01 -2.20338076e-01
-8.58623326e-01 9.32639003e-01 1.03822744e+00 -8.49156976e-02
6.66522264e-01 -2.63012946e-02 3.54639441e-01 7.13750571e-02
-5.26479065e-01 -3.98649693e-01 1.45863667e-01 1.18399584e+00
-2.76680619e-01 1.17534347e-01 7.56032020e-02 4.35010642e-01
6.07725866e-02 -4.31508645e-02 2.70372152e-01 6.41781032e-01
-5.06345093e-01 -9.28054035e-01 -5.58493793e-01 2.69864172e-01
1.81401685e-01 1.84360996e-01 -5.33245146e-01 5.82572222e-01
4.80194002e-01 6.42118216e-01 5.99592067e-02 -1.64853558e-01
5.01381099e-01 2.38544554e-01 4.49730933e-01 -6.53302014e-01
-2.96802223e-01 -2.23014131e-01 -1.59901872e-01 -1.45645082e-01
-4.99140263e-01 -8.37352574e-01 -9.50352430e-01 2.16216102e-01
-3.25967252e-01 8.29966292e-02 5.46490610e-01 6.69493258e-01
4.96694863e-01 4.12069649e-01 7.12101698e-01 -9.08777237e-01
-8.92905295e-01 -8.83906364e-01 -5.82101762e-01 5.07110119e-01
5.04936695e-01 -1.02188218e+00 -5.14472306e-01 2.55064636e-01]
|
[9.824564933776855, 2.1976022720336914]
|
c4e4ce07-29dd-4fec-aba2-4d5b7bd22dfa
|
zero-shot-object-counting
|
2303.02001
| null |
https://arxiv.org/abs/2303.02001v2
|
https://arxiv.org/pdf/2303.02001v2.pdf
|
Zero-shot Object Counting
|
Class-agnostic object counting aims to count object instances of an arbitrary class at test time. It is challenging but also enables many potential applications. Current methods require human-annotated exemplars as inputs which are often unavailable for novel categories, especially for autonomous systems. Thus, we propose zero-shot object counting (ZSC), a new setting where only the class name is available during test time. Such a counting system does not require human annotators in the loop and can operate automatically. Starting from a class name, we propose a method that can accurately identify the optimal patches which can then be used as counting exemplars. Specifically, we first construct a class prototype to select the patches that are likely to contain the objects of interest, namely class-relevant patches. Furthermore, we introduce a model that can quantitatively measure how suitable an arbitrary patch is as a counting exemplar. By applying this model to all the candidate patches, we can select the most suitable patches as exemplars for counting. Experimental results on a recent class-agnostic counting dataset, FSC-147, validate the effectiveness of our method. Code is available at https://github.com/cvlab-stonybrook/zero-shot-counting
|
['Dimitris Samaras', 'Viresh Ranjan', 'Vu Nguyen', 'Hieu Le', 'Jingyi Xu']
|
2023-03-03
| null |
http://openaccess.thecvf.com//content/CVPR2023/html/Xu_Zero-Shot_Object_Counting_CVPR_2023_paper.html
|
http://openaccess.thecvf.com//content/CVPR2023/papers/Xu_Zero-Shot_Object_Counting_CVPR_2023_paper.pdf
|
cvpr-2023-1
|
['object-counting']
|
['computer-vision']
|
[ 1.50416121e-02 -3.90795857e-01 -1.19722292e-01 -3.49558294e-01
-7.08647370e-01 -6.21009052e-01 4.89581287e-01 5.43445289e-01
-5.89944959e-01 6.40598893e-01 -4.11345452e-01 1.18415453e-01
8.86451174e-03 -1.15063918e+00 -6.90622568e-01 -4.37222779e-01
1.52814329e-01 9.67011571e-01 7.15112329e-01 3.04524541e-01
5.70276022e-01 4.53213245e-01 -1.97443974e+00 1.87853083e-01
9.53644156e-01 7.86807716e-01 3.34228277e-01 5.62653005e-01
-2.63090223e-01 3.59776944e-01 -9.13664401e-01 -2.69687414e-01
3.23671192e-01 -3.24863851e-01 -5.99877059e-01 5.72383516e-02
6.41061246e-01 -2.92949170e-01 3.12204093e-01 1.22978699e+00
3.37819278e-01 2.23251134e-01 6.91225350e-01 -1.12722695e+00
-3.74920785e-01 3.09808642e-01 -7.11773276e-01 4.21480834e-01
1.42466515e-01 1.39110506e-01 7.53653705e-01 -1.05043721e+00
3.84048223e-01 1.05819774e+00 4.70505208e-01 7.46475816e-01
-1.18670571e+00 -8.50666642e-01 2.06769064e-01 2.01185584e-01
-1.75965142e+00 -2.11899742e-01 5.59319854e-01 -4.06101644e-01
4.68790323e-01 3.80499929e-01 8.98322523e-01 4.45169628e-01
-4.24746752e-01 9.34884787e-01 9.25901592e-01 -5.43687582e-01
7.64356911e-01 1.70936033e-01 4.57669437e-01 7.83582747e-01
7.46087432e-01 -3.17472786e-01 -3.63801032e-01 -3.22387338e-01
6.91766262e-01 3.23329806e-01 9.19197053e-02 -3.86378318e-01
-1.09117258e+00 8.35023880e-01 4.99818981e-01 2.69711792e-01
-2.17800140e-01 2.68925905e-01 2.37500951e-01 -2.30455101e-01
4.57324952e-01 4.45446432e-01 -1.98397428e-01 1.04097717e-01
-1.00712752e+00 2.04746440e-01 6.15458190e-01 1.13442814e+00
7.62392819e-01 -4.00097996e-01 -5.64979076e-01 9.33506727e-01
-1.94036961e-01 5.32519281e-01 2.03054041e-01 -8.29421997e-01
2.07003325e-01 1.03935921e+00 3.27587485e-01 -6.48434281e-01
-1.64582714e-01 -3.82605016e-01 -4.72727418e-01 5.09762838e-02
4.14218187e-01 3.16946089e-01 -9.30789709e-01 1.46127903e+00
7.16024160e-01 4.09952015e-01 -3.83297861e-01 9.17283714e-01
5.58986545e-01 3.40048641e-01 1.73107788e-01 -1.72047049e-01
1.61879611e+00 -7.76592433e-01 -3.12487185e-02 -3.13275307e-01
4.56555188e-01 -4.42153275e-01 1.38301909e+00 2.03814000e-01
-7.83235848e-01 -3.69442075e-01 -1.03683031e+00 2.78348774e-01
-5.66351533e-01 3.71084392e-01 5.66603541e-01 5.62854588e-01
-4.69183862e-01 4.32375580e-01 -8.43205094e-01 -3.50803465e-01
6.34201407e-01 1.44262791e-01 -9.98151079e-02 -1.66254297e-01
-6.02726936e-01 6.31682754e-01 6.20108724e-01 -2.94039100e-01
-8.86586308e-01 -3.48123401e-01 -7.07366765e-01 2.17290699e-01
5.92144966e-01 -4.37061340e-01 1.29019201e+00 -7.22409904e-01
-8.41893554e-01 8.36439133e-01 -3.83183032e-01 -2.43293792e-01
3.87033492e-01 7.71351159e-02 -8.93137455e-02 3.31525743e-01
6.55963778e-01 6.34735227e-01 6.83880627e-01 -1.37883472e+00
-9.82713759e-01 -2.66634673e-01 1.61239833e-01 -1.65993735e-01
-3.79254580e-01 -7.36070946e-02 -6.16630495e-01 -5.09085536e-01
1.28258824e-01 -7.14002728e-01 -1.59784153e-01 2.04138368e-01
-3.58278215e-01 -5.22522151e-01 4.82552946e-01 -1.38227949e-02
1.20628250e+00 -2.02013540e+00 -3.74983281e-01 2.76029170e-01
3.52093160e-01 3.06720495e-01 -8.87040421e-02 5.48936091e-02
3.87369901e-01 1.93382036e-02 -3.32038403e-01 -1.84568450e-01
-1.25235915e-01 1.29524052e-01 1.94502175e-02 3.28408241e-01
4.67448652e-01 7.04261959e-01 -1.31342447e+00 -9.52328146e-01
4.45613086e-01 1.38381287e-01 -4.72086459e-01 5.60401827e-02
-3.07711005e-01 3.73754144e-01 -4.64387387e-01 1.02111506e+00
7.65156031e-01 -5.65552235e-01 -6.99841529e-02 5.81849962e-02
-3.68789136e-02 -3.11214656e-01 -1.25472677e+00 9.12433445e-01
-3.77042681e-01 1.69404864e-01 -5.46300530e-01 -8.71025801e-01
1.05237854e+00 -2.57979184e-01 1.99907720e-01 -5.52179694e-01
2.30313405e-01 4.33783293e-01 -6.10227622e-02 -1.70829669e-01
4.16227967e-01 -1.33286923e-01 -2.25381523e-01 3.89409900e-01
1.51768804e-01 -2.26831630e-01 9.06244159e-01 1.75735071e-01
1.22475541e+00 -2.57136524e-01 7.25463986e-01 -2.00632662e-01
3.99244338e-01 2.80377090e-01 7.74696052e-01 1.00073957e+00
-3.56574088e-01 6.79792225e-01 1.86777860e-01 -5.61502934e-01
-9.70159888e-01 -9.72103775e-01 -3.76469314e-01 1.10445583e+00
4.62234259e-01 -3.04592341e-01 -8.34319293e-01 -6.77278280e-01
7.47923255e-02 8.19926083e-01 -7.73728251e-01 -5.25604412e-02
-4.56776798e-01 -4.15744185e-01 2.44077325e-01 7.27265120e-01
2.76603341e-01 -1.17405736e+00 -9.75921631e-01 2.02084050e-01
-2.52515435e-01 -8.22171926e-01 -4.75863755e-01 2.05491573e-01
-7.22014725e-01 -1.47159541e+00 -7.22605169e-01 -5.98727942e-01
1.10193598e+00 6.70097351e-01 1.24642706e+00 5.57345927e-01
-5.77388942e-01 3.92270416e-01 -4.24926460e-01 -7.32663214e-01
-1.21176481e-01 -1.54195547e-01 -6.21014871e-02 -2.12486126e-02
8.68466496e-01 -2.33029529e-01 -8.59215677e-01 5.41764736e-01
-7.15399742e-01 -1.45101607e-01 4.21776742e-01 7.29803026e-01
1.14422739e+00 -3.86797334e-03 6.38856351e-01 -1.08805132e+00
3.63236934e-01 -5.21790922e-01 -9.38095689e-01 5.29491723e-01
-2.09727809e-01 -1.58863097e-01 7.89531648e-01 -8.41458857e-01
-5.47957182e-01 3.36374342e-01 4.88458008e-01 -5.35544932e-01
-8.89769346e-02 1.23637252e-01 -1.22975539e-02 1.22570032e-02
8.03778291e-01 1.13270126e-01 -5.11545718e-01 -2.74018079e-01
8.60364288e-02 4.68206912e-01 4.86816436e-01 -8.84160280e-01
7.86722720e-01 5.66146612e-01 -6.31489009e-02 -7.05310643e-01
-1.07522273e+00 -1.00389051e+00 -7.71326065e-01 -3.38328123e-01
5.70776582e-01 -6.71029031e-01 -8.26128125e-01 9.93477181e-02
-1.06610596e+00 -1.75063893e-01 -4.88414943e-01 3.26375604e-01
-4.23027188e-01 1.68484464e-01 -1.79846853e-01 -1.09111345e+00
-4.87152517e-01 -8.69975209e-01 1.39177287e+00 5.66601157e-01
-1.68893158e-01 -4.49680626e-01 -3.29734348e-02 1.13620847e-01
4.54919897e-02 1.82472318e-01 8.38140845e-01 -8.33644390e-01
-7.75210440e-01 -6.98586166e-01 -3.03245813e-01 1.64950490e-01
-7.72670433e-02 1.77170962e-01 -9.31257308e-01 -2.39459559e-01
-4.33059812e-01 -2.88738608e-01 8.88893425e-01 2.85525352e-01
1.63961828e+00 -4.23629656e-02 -6.03063583e-01 2.77773976e-01
1.45916212e+00 7.37823248e-02 3.38262975e-01 1.88659862e-01
3.80577147e-01 2.26859644e-01 1.00134683e+00 7.04759061e-01
2.48099938e-01 5.59178770e-01 3.54839385e-01 1.56036168e-01
1.76404357e-01 -3.95908117e-01 -3.41335624e-01 6.06426656e-01
-2.75331557e-01 -2.27547199e-01 -1.02008021e+00 8.18508863e-01
-1.83264208e+00 -9.00729895e-01 -4.82461564e-02 2.64074826e+00
7.82762349e-01 1.41321376e-01 2.21069604e-01 3.01066220e-01
1.10130990e+00 -4.52629387e-01 -7.00433016e-01 1.25696450e-01
2.86097229e-01 6.31542444e-01 4.14297462e-01 1.80059820e-01
-1.12604511e+00 8.35297167e-01 5.11308670e+00 9.21793222e-01
-5.18791854e-01 1.78027287e-01 4.61293519e-01 -2.44309813e-01
2.78260857e-01 1.01622835e-01 -1.04854894e+00 6.90840185e-01
5.20576298e-01 -3.58077466e-01 3.39038402e-01 1.14772630e+00
-1.07014328e-01 -4.26946759e-01 -1.10880840e+00 8.69124293e-01
6.61334097e-02 -1.03115451e+00 9.79796425e-03 -2.95184910e-01
5.45098364e-01 -3.45386297e-01 -4.08765018e-01 5.25150239e-01
4.96938527e-01 -6.24497056e-01 9.11083102e-01 5.40719390e-01
8.72839868e-01 -7.08915412e-01 7.02845395e-01 6.78912580e-01
-1.45187771e+00 -1.21251158e-01 -9.45661604e-01 1.12579547e-01
-1.34598598e-01 6.12423539e-01 -8.76800179e-01 -6.25697896e-02
7.31301427e-01 1.97457626e-01 -9.82944906e-01 1.67001617e+00
-3.08998853e-01 5.62060416e-01 -5.17710924e-01 -3.85339051e-01
-1.47467807e-01 1.31181002e-01 2.64411837e-01 1.15102017e+00
5.29521227e-01 3.55221152e-01 4.52797532e-01 8.99778783e-01
-8.19630250e-02 1.98024258e-01 -5.55118382e-01 2.21333712e-01
1.08784676e+00 1.28275549e+00 -1.44572270e+00 -6.42003417e-01
-3.02924633e-01 7.03154147e-01 6.85907304e-01 7.73631874e-03
-9.43588555e-01 -3.94482970e-01 1.67100176e-01 2.42405847e-01
2.21538737e-01 1.31466672e-01 -1.39202356e-01 -1.13702953e+00
1.81716949e-01 -5.47222674e-01 6.23468816e-01 -5.82255960e-01
-1.48216796e+00 2.36681864e-01 1.60853490e-01 -1.72305560e+00
1.64644048e-01 -5.57100594e-01 -7.76959717e-01 4.59343076e-01
-1.09382915e+00 -9.90491152e-01 -8.28135014e-01 4.04224187e-01
6.13389969e-01 5.08173704e-02 8.06447864e-01 2.29361504e-01
-4.17693138e-01 6.20149016e-01 -7.62680322e-02 2.36704290e-01
4.48711425e-01 -1.23904753e+00 3.45720232e-01 7.15820491e-01
3.96105468e-01 7.99324393e-01 4.22967285e-01 -6.31729484e-01
-9.87966895e-01 -1.39605010e+00 4.51422483e-01 -6.18307471e-01
3.45180303e-01 -4.38025862e-01 -1.06712234e+00 3.38617712e-01
-7.54105449e-01 5.76902986e-01 4.76673007e-01 2.29959153e-02
-4.77866530e-01 -5.42013049e-02 -1.40923059e+00 4.47912484e-01
1.04903638e+00 -1.72570959e-01 -3.01674366e-01 5.63993990e-01
5.01417339e-01 -8.37867409e-02 -4.82997745e-01 2.53547639e-01
5.03810048e-01 -8.03740561e-01 8.65701199e-01 -2.21227318e-01
2.73887187e-01 -6.16231859e-01 -7.10976496e-03 -9.84424293e-01
-3.09939831e-01 2.16006115e-01 -1.08168438e-01 1.11003041e+00
2.96207964e-01 -6.42321289e-01 7.40713418e-01 4.04415518e-01
4.10272889e-02 -5.88143587e-01 -1.05171359e+00 -1.10809660e+00
-6.23860434e-02 -1.47135541e-01 8.14980626e-01 7.95030892e-01
-1.18988968e-01 2.25206152e-01 1.55917972e-01 2.81052023e-01
8.07607174e-01 3.36819381e-01 1.05858910e+00 -1.63187253e+00
-1.87890321e-01 -4.28347886e-01 -7.15362072e-01 -6.80753291e-01
-8.71611238e-02 -8.95467818e-01 3.31052303e-01 -1.43592834e+00
7.46470630e-01 -7.17763066e-01 -2.61339724e-01 6.71596587e-01
-5.75132310e-01 3.69245052e-01 1.95282400e-01 3.82407844e-01
-1.03444386e+00 2.44006708e-01 1.03345096e+00 -2.78700918e-01
-1.85229957e-01 2.55920231e-01 -4.54225719e-01 7.13073254e-01
1.04194009e+00 -7.06828296e-01 -4.66055684e-02 -2.45715320e-01
-2.89707538e-02 -1.19094409e-01 5.64368546e-01 -1.33524287e+00
2.73306578e-01 -3.82726461e-01 4.91660535e-01 -7.49747455e-01
2.69606680e-01 -8.43756497e-01 2.41852030e-02 5.70684850e-01
-3.06367949e-02 -3.82187575e-01 1.60507590e-01 6.75101340e-01
-4.31662686e-02 -8.23634684e-01 1.13931930e+00 -3.81615490e-01
-6.31698132e-01 3.14962953e-01 7.28955045e-02 7.26211444e-02
1.17785633e+00 -4.13603187e-01 -5.55790186e-01 1.93003207e-01
-4.01575953e-01 3.49283636e-01 7.28501737e-01 5.25418334e-02
5.13482690e-01 -1.31860483e+00 -6.78181589e-01 9.05860290e-02
8.49698901e-01 3.10672909e-01 -1.30779848e-01 4.00739849e-01
-4.97289687e-01 1.80091277e-01 2.82457080e-02 -8.10883880e-01
-1.29283106e+00 8.00618768e-01 7.83002600e-02 4.78444472e-02
-6.16956890e-01 7.53417611e-01 3.70466620e-01 -4.58627284e-01
-3.57794750e-04 -2.66219556e-01 -1.87537730e-01 -2.32444406e-01
7.77835190e-01 5.31676471e-01 7.59624271e-03 -3.32075506e-01
-4.45178151e-01 5.66073298e-01 6.21115044e-02 2.35003963e-01
1.19391096e+00 3.14800948e-01 2.94490103e-02 6.66658700e-01
6.61964953e-01 -7.15450570e-02 -1.07927620e+00 -2.90172458e-01
2.44059414e-01 -9.06790793e-01 -4.57375556e-01 -5.48653841e-01
-9.30606544e-01 8.75044346e-01 5.70129693e-01 2.18163908e-01
9.21233773e-01 1.47180602e-01 3.21604788e-01 5.27502000e-01
9.25852358e-01 -1.05508876e+00 2.41769820e-01 3.66603941e-01
6.05314434e-01 -1.39520228e+00 5.74700721e-02 -4.71551955e-01
-2.45988011e-01 9.97246206e-01 1.02792966e+00 -1.89548433e-01
3.19901973e-01 2.14895263e-01 -3.67971361e-01 -4.38087106e-01
-6.00345969e-01 -5.85482895e-01 1.19720578e-01 5.36879301e-01
2.17371404e-01 5.48311293e-01 -4.12597835e-01 5.71978807e-01
-7.89605901e-02 4.24605645e-02 3.96398306e-01 1.10039282e+00
-8.74513328e-01 -5.73736429e-01 -5.65654814e-01 1.01853323e+00
-1.68973822e-02 6.91152662e-02 -4.43040580e-01 6.85231566e-01
2.92923093e-01 8.03343594e-01 2.70687312e-01 -8.38709921e-02
4.40771043e-01 5.35002304e-03 5.76627076e-01 -1.10553026e+00
-1.67702407e-01 -3.20799261e-01 -4.02404606e-01 -1.43238291e-01
-2.28751153e-01 -4.78859335e-01 -1.27602124e+00 -3.65290642e-01
-9.41484630e-01 1.14017062e-01 3.92647266e-01 6.26733780e-01
8.74439031e-02 2.23775059e-01 5.03469408e-01 -8.58447433e-01
-3.58858943e-01 -1.04143918e+00 -5.28336465e-01 3.57007444e-01
-1.39827758e-01 -1.16625476e+00 -3.48866999e-01 -7.14427903e-02]
|
[8.99374008178711, 0.5420737266540527]
|
0287449c-6356-47c0-9d53-deb771fe7449
|
safe-reinforcement-learning-using-data-driven
|
2211.11027
| null |
https://arxiv.org/abs/2211.11027v1
|
https://arxiv.org/pdf/2211.11027v1.pdf
|
Safe Reinforcement Learning using Data-Driven Predictive Control
|
Reinforcement learning (RL) algorithms can achieve state-of-the-art performance in decision-making and continuous control tasks. However, applying RL algorithms on safety-critical systems still needs to be well justified due to the exploration nature of many RL algorithms, especially when the model of the robot and the environment are unknown. To address this challenge, we propose a data-driven safety layer that acts as a filter for unsafe actions. The safety layer uses a data-driven predictive controller to enforce safety guarantees for RL policies during training and after deployment. The RL agent proposes an action that is verified by computing the data-driven reachability analysis. If there is an intersection between the reachable set of the robot using the proposed action, we call the data-driven predictive controller to find the closest safe action to the proposed unsafe action. The safety layer penalizes the RL agent if the proposed action is unsafe and replaces it with the closest safe one. In the simulation, we show that our method outperforms state-of-the-art safe RL methods on the robotics navigation problem for a Turtlebot 3 in Gazebo and a quadrotor in Unreal Engine 4 (UE4).
|
['Karl H. Johansson', 'Hazem M. Abbas', 'M. Watheq El-Kharashi', 'Amr Alanwar', 'Mahmoud Selim']
|
2022-11-20
| null | null | null | null |
['continuous-control']
|
['playing-games']
|
[-1.09294876e-01 4.64704692e-01 -2.89469719e-01 9.09464732e-02
-3.75290930e-01 -5.91843963e-01 6.27151370e-01 2.86151975e-01
-7.60555446e-01 1.06367850e+00 -3.32082868e-01 -7.64946699e-01
-6.64154530e-01 -9.27489877e-01 -1.04946172e+00 -8.73249531e-01
-4.67566967e-01 5.31210899e-01 5.66476107e-01 -6.46740735e-01
3.03716451e-01 5.01916766e-01 -1.49993312e+00 -4.50729609e-01
7.92060494e-01 9.82587814e-01 2.76349962e-01 5.40419161e-01
7.00583458e-01 7.92963684e-01 -1.23548917e-01 5.08145213e-01
5.32106400e-01 -3.19378406e-01 -9.18571830e-01 -2.44870037e-01
-2.99725443e-01 -3.28113139e-01 -1.75152589e-02 1.23638892e+00
1.08611211e-01 6.49713337e-01 5.32964528e-01 -1.73798025e+00
4.39412415e-01 4.83609110e-01 -5.87039962e-02 -3.07922363e-01
1.80078149e-01 5.42285621e-01 5.95486522e-01 -2.84187376e-01
6.65954649e-01 1.17912936e+00 1.96315467e-01 7.96526015e-01
-8.30269456e-01 -5.37483811e-01 3.84256423e-01 2.06378713e-01
-1.18156445e+00 3.11369300e-02 3.26374084e-01 -4.71032530e-01
8.74496639e-01 -2.78792083e-01 7.47116446e-01 7.22304523e-01
7.14323699e-01 2.47031748e-01 8.88898432e-01 -2.50309587e-01
8.08056891e-01 -2.87508965e-01 -3.77808511e-01 8.91911566e-01
4.52711254e-01 8.97687316e-01 5.43796569e-02 -1.07142895e-01
3.67415220e-01 -1.56097054e-01 4.62017069e-03 -7.81767428e-01
-1.08966768e+00 9.62989807e-01 5.68013370e-01 -5.65045737e-02
-7.21008956e-01 3.53313327e-01 4.37809527e-01 3.92079413e-01
-2.84154296e-01 8.73917878e-01 -4.24004912e-01 -1.78141952e-01
-3.24359000e-01 6.72214448e-01 7.80135751e-01 8.11494172e-01
5.59444606e-01 3.19627494e-01 1.05300799e-01 1.61357939e-01
3.27963054e-01 3.28254610e-01 -1.28833860e-01 -1.19396842e+00
3.12223911e-01 6.84368908e-01 6.51185334e-01 -5.15593648e-01
-5.01554906e-01 -4.66292761e-02 -4.20499951e-01 1.15669894e+00
2.73082107e-01 -4.89610463e-01 -4.59382117e-01 1.59005058e+00
6.38004005e-01 9.29893702e-02 3.72891933e-01 1.13468742e+00
-1.95362598e-01 5.88593423e-01 -4.43110503e-02 -3.52258295e-01
9.32228029e-01 -6.59032345e-01 -4.70550895e-01 -3.28400970e-01
7.05113173e-01 3.09648272e-02 6.27561688e-01 7.92443573e-01
-8.27518165e-01 -3.15950960e-01 -1.50441706e+00 5.97321749e-01
-2.06702173e-01 -1.20889038e-01 2.70468108e-02 -1.14569858e-01
-6.78723514e-01 7.98178136e-01 -9.53107119e-01 -4.14971024e-01
-1.28926247e-01 5.07500768e-01 -4.09845650e-01 1.04842052e-01
-1.37832260e+00 1.59758472e+00 8.20635259e-01 4.28572185e-02
-1.75196934e+00 -1.34717315e-01 -8.81822050e-01 -2.15184242e-01
1.28973675e+00 -1.74311802e-01 1.27567768e+00 -4.92683113e-01
-1.84732032e+00 5.23673445e-02 6.85973704e-01 -9.29185092e-01
5.93490481e-01 -2.50904173e-01 -8.23569670e-02 -3.70081030e-02
7.13005960e-02 5.05051076e-01 9.03831363e-01 -1.15623963e+00
-1.04170716e+00 -4.96549085e-02 5.22793710e-01 2.63549626e-01
1.97701320e-01 -6.29754841e-01 1.36191815e-01 -6.94940099e-03
-2.40232989e-01 -1.48612511e+00 -6.60899162e-01 6.57783169e-03
-2.86078483e-01 -3.29413652e-01 7.97929049e-01 -1.38965715e-02
1.06071925e+00 -2.00696635e+00 3.18878412e-01 3.39245617e-01
-2.94652462e-01 3.00276190e-01 -2.42407173e-02 5.87163925e-01
1.25876144e-01 -2.46615425e-01 -1.64539695e-01 1.52821958e-01
2.15180330e-02 4.76722747e-01 -5.02066791e-01 7.51616418e-01
5.63166067e-02 5.91617413e-02 -1.18131518e+00 -1.04913965e-01
4.61222053e-01 -6.87852725e-02 -6.54092193e-01 4.83461946e-01
-6.11102879e-01 6.58627450e-01 -9.01083469e-01 5.20730205e-02
2.77443260e-01 5.34305751e-01 1.42983213e-01 3.86244416e-01
-7.61926830e-01 2.87048995e-01 -1.27565348e+00 1.28022778e+00
-4.80261534e-01 1.90182015e-01 4.72334623e-01 -1.02280092e+00
9.64922607e-01 1.98194146e-01 3.78452629e-01 -4.56633121e-01
3.81648928e-01 2.66221225e-01 1.29444271e-01 -2.85412550e-01
3.81975651e-01 -2.47057885e-01 -4.23500687e-01 1.48511842e-01
-1.63062111e-01 -6.30417824e-01 9.51264128e-02 -1.85087398e-01
1.29159105e+00 5.65885305e-01 5.35788894e-01 -4.82869297e-01
9.76686597e-01 3.02420855e-01 7.28138685e-01 7.12170601e-01
-3.82192582e-01 -3.62002552e-01 6.42474115e-01 -3.22527647e-01
-6.89830542e-01 -6.17923319e-01 4.01570797e-01 9.18134093e-01
5.36357641e-01 -2.82642305e-01 -6.48959816e-01 -9.83412325e-01
1.37067840e-01 1.17073727e+00 -7.25492001e-01 -7.26175845e-01
-4.74984974e-01 2.57768333e-01 3.28366429e-01 2.00951174e-01
3.14505637e-01 -1.10301590e+00 -1.42468667e+00 3.31065506e-01
2.04114020e-01 -8.26174378e-01 -1.12705864e-01 6.98489249e-01
-5.82343400e-01 -1.55905485e+00 1.93026930e-01 -5.43780863e-01
7.37803519e-01 -1.21547922e-01 3.38567287e-01 9.20061916e-02
1.72542185e-01 2.19755128e-01 -2.66724378e-01 -3.36561024e-01
-8.30618620e-01 -3.40569437e-01 6.22426927e-01 -3.34388912e-01
-3.57262254e-01 1.12419873e-01 -3.70270610e-01 5.09216964e-01
-5.82324684e-01 -1.69171751e-01 2.86004692e-01 7.66034603e-01
7.65443683e-01 7.15085864e-01 4.28751081e-01 -2.01217905e-01
7.20436752e-01 -3.31387073e-01 -1.51679301e+00 2.83714682e-02
-6.45066798e-01 6.39279902e-01 1.08814001e+00 -3.81934375e-01
-5.99714458e-01 4.48270619e-01 5.01709469e-02 -3.71109754e-01
-1.12282269e-01 4.77623850e-01 -2.58492436e-02 -1.51306456e-02
5.39240539e-01 -6.93933219e-02 2.65631944e-01 -9.79239419e-02
2.37580135e-01 2.45836794e-01 4.03528571e-01 -6.43104613e-01
8.44008803e-01 2.14381948e-01 6.23420596e-01 -4.53929424e-01
-4.06031787e-01 -1.54340431e-01 -3.89511794e-01 -5.74963868e-01
8.70557785e-01 -5.11199176e-01 -1.39066064e+00 6.89808875e-02
-8.64094198e-01 -6.06504440e-01 -3.48868489e-01 3.72014612e-01
-1.07765698e+00 -8.11974108e-02 6.88771158e-02 -1.28205252e+00
-1.52994424e-01 -1.49292088e+00 6.55214190e-01 1.36819124e-01
-5.29699074e-03 -5.04766405e-01 1.49341196e-01 -3.65308285e-01
9.57958549e-02 4.71515805e-01 8.19195628e-01 -6.17578268e-01
-2.81107634e-01 -1.40094459e-01 5.67345977e-01 2.46230215e-01
-1.06859609e-01 -1.35566324e-01 -2.50938982e-01 -6.81659043e-01
7.66752586e-02 -4.55515265e-01 3.01875561e-01 1.88727066e-01
6.68386936e-01 -6.52707160e-01 -4.76853937e-01 7.22838044e-02
1.51227593e+00 7.92898357e-01 3.52481425e-01 7.95083582e-01
1.55041173e-01 7.23300040e-01 1.62905133e+00 5.50381124e-01
2.91866124e-01 6.89109743e-01 1.46475315e+00 5.52530587e-01
7.01759398e-01 -3.44480991e-01 7.71811545e-01 -1.26262069e-01
1.75173059e-01 -3.12651172e-02 -8.44038248e-01 3.57691258e-01
-2.21218348e+00 -7.70492256e-01 7.57047683e-02 2.62423253e+00
6.23026133e-01 6.13679767e-01 2.50428140e-01 2.81644493e-01
4.77982581e-01 -2.59877980e-01 -7.50703275e-01 -6.45720959e-01
5.65588713e-01 -3.65677625e-01 7.81113029e-01 7.34240234e-01
-1.15114319e+00 9.83371437e-01 4.64775038e+00 5.85432529e-01
-1.09654713e+00 -3.00867975e-01 -7.04216808e-02 8.95006359e-02
3.44386011e-01 2.43335158e-01 -7.09512949e-01 2.83399373e-01
9.22579587e-01 -1.59040168e-01 6.97714806e-01 1.21910608e+00
5.20703435e-01 -7.93433666e-01 -1.38313150e+00 2.99932420e-01
-4.41423833e-01 -9.03029799e-01 -5.01857638e-01 -1.97412632e-02
3.60777557e-01 -9.70086157e-02 -1.31000042e-01 5.56384981e-01
6.32211328e-01 -6.60504043e-01 1.09797502e+00 4.83178854e-01
3.66263509e-01 -1.29061520e+00 6.69451237e-01 9.40828919e-01
-1.04369581e+00 -6.45354033e-01 -6.59789219e-02 -2.52556443e-01
6.81412444e-02 -1.12087891e-01 -1.09620297e+00 5.72974265e-01
3.71656030e-01 4.06255037e-01 -6.77769780e-02 9.21680033e-01
-7.45258570e-01 1.11713268e-01 -3.57036412e-01 -5.56452751e-01
7.43887722e-01 -3.93220723e-01 9.16510940e-01 3.25677931e-01
2.55854994e-01 4.56802547e-03 6.03475094e-01 5.95703363e-01
6.96849585e-01 -4.76656675e-01 -1.00948834e+00 7.52697736e-02
3.02388102e-01 8.52706492e-01 -5.12120306e-01 -6.70920759e-02
2.26153120e-01 4.67181593e-01 2.25233883e-01 1.47743732e-01
-9.58660185e-01 -4.66932803e-01 8.47986281e-01 1.39640914e-02
1.74145922e-01 -4.09898996e-01 7.00988322e-02 -2.91142374e-01
-3.58405590e-01 -7.97327816e-01 3.37372005e-01 -7.48187661e-01
-7.18534708e-01 5.78469634e-01 7.74472505e-02 -1.75032496e+00
-5.99527180e-01 -5.46015143e-01 -4.64825898e-01 4.54433978e-01
-1.20365822e+00 -2.90279239e-01 2.06768624e-02 5.29860198e-01
3.22622269e-01 -1.71114355e-01 6.46584451e-01 -3.06534886e-01
-3.49505454e-01 -2.80873358e-01 -1.88750237e-01 -3.31945658e-01
5.49483061e-01 -9.22636449e-01 -1.42443538e-01 8.52826953e-01
-7.47372448e-01 2.58263171e-01 1.16696072e+00 -7.75369942e-01
-1.79215312e+00 -1.25222683e+00 2.73587137e-01 -1.17352970e-01
8.43192697e-01 -2.51473505e-02 -4.88534421e-01 5.43451965e-01
-1.12289242e-01 -8.03834498e-02 -2.89938867e-01 -4.92853254e-01
2.68762708e-01 -2.13030323e-01 -1.25698721e+00 1.04165387e+00
5.48365176e-01 -5.56173101e-02 -6.97032690e-01 8.56670365e-02
7.89774895e-01 -3.46371919e-01 -4.93248224e-01 7.07650900e-01
3.07935506e-01 -4.48809087e-01 5.39888978e-01 -6.25489295e-01
9.83935446e-02 -8.18772793e-01 1.16595395e-01 -1.69004154e+00
-7.27304146e-02 -7.67979920e-01 -7.67151117e-02 4.36531633e-01
1.31816387e-01 -5.43363035e-01 5.31516433e-01 2.55600184e-01
-1.97320744e-01 -8.78022373e-01 -1.50053930e+00 -1.18671823e+00
9.24090967e-02 -4.29202914e-01 3.96727353e-01 4.33054715e-01
5.12133539e-01 1.96986496e-01 -4.88511264e-01 4.87113744e-01
5.32414496e-01 -2.45733023e-01 6.86285257e-01 -9.31637347e-01
1.29014784e-02 -3.09262395e-01 -1.04362473e-01 -5.97864270e-01
6.03923202e-01 -4.80823040e-01 9.62857306e-01 -1.58816862e+00
-6.32423997e-01 -5.76086760e-01 -1.20253451e-01 6.96671665e-01
2.45744333e-01 -6.75426543e-01 2.88998067e-01 -2.59975046e-01
-8.08169425e-01 6.80690587e-01 1.28336906e+00 -1.81708679e-01
-5.44567764e-01 3.40418696e-01 -1.34658471e-01 8.21921885e-01
1.01530743e+00 -4.84676987e-01 -4.81926918e-01 6.90242574e-02
3.60560060e-01 7.49357462e-01 1.68553784e-01 -1.35519862e+00
4.88114566e-01 -7.42528141e-01 -5.63130558e-01 -6.04353607e-01
2.33419493e-01 -1.59011006e+00 9.75491703e-02 1.49579251e+00
-4.98888165e-01 6.47669062e-02 2.83811271e-01 6.90381527e-01
-2.01316793e-02 -3.68966401e-01 1.04793274e+00 2.29158387e-01
-7.80239344e-01 2.07643941e-01 -9.32060301e-01 -1.48554385e-01
1.53580809e+00 1.58357382e-01 -1.65726811e-01 -3.35046738e-01
-4.04428810e-01 1.02724266e+00 3.09486061e-01 5.47023296e-01
6.72720909e-01 -8.51584792e-01 -2.15927973e-01 2.10748166e-01
2.60815937e-02 3.72170620e-02 -2.00900555e-01 8.01674664e-01
-2.58913636e-01 3.72828722e-01 -4.64776665e-01 -2.52851814e-01
-1.09480548e+00 8.91687930e-01 5.94779372e-01 -1.35250106e-01
-4.40095872e-01 3.53365868e-01 -1.36438161e-01 -4.74967003e-01
4.51896846e-01 -5.82576990e-01 -4.44994152e-01 -1.77893668e-01
1.62525594e-01 7.05840468e-01 4.95239682e-02 -3.89176339e-01
-7.25609839e-01 3.42977285e-01 4.06370908e-01 -4.11388010e-01
1.17015290e+00 2.44513452e-01 1.18625470e-01 3.22576344e-01
3.43370616e-01 -5.00721753e-01 -1.70804584e+00 4.25334454e-01
2.22172230e-01 -5.87442704e-02 3.52700800e-02 -9.05396700e-01
-5.94302237e-01 6.37935400e-01 3.90723348e-01 1.05463803e-01
8.96312356e-01 -5.30026317e-01 1.67015493e-01 7.70186424e-01
1.12778234e+00 -1.47002447e+00 4.93830219e-02 1.20336103e+00
1.09616363e+00 -8.86873364e-01 1.77978817e-02 -8.51086080e-02
-8.73820901e-01 1.01312637e+00 1.10529828e+00 -5.13843834e-01
4.99350101e-01 3.53107452e-01 -3.25571030e-01 1.74466163e-01
-1.08142138e+00 -1.39548466e-01 -5.58852442e-02 5.01405656e-01
-5.80987275e-01 1.24938004e-01 -4.82818037e-01 3.67510319e-01
3.58682983e-02 4.72671166e-02 7.56736100e-01 1.27396297e+00
-1.00285828e+00 -8.22460771e-01 -4.45726007e-01 -7.57004470e-02
9.80409607e-02 4.82547671e-01 -1.78460926e-01 9.71086860e-01
2.09559023e-01 1.12177587e+00 -1.96338631e-02 -5.64725697e-01
7.13877857e-01 -1.74422443e-01 3.42559189e-01 -6.68337226e-01
-5.54107010e-01 -1.61500126e-01 2.77256876e-01 -8.90694022e-01
7.68741295e-02 -4.60883766e-01 -2.02195978e+00 3.73558961e-02
-3.32922310e-01 3.57989401e-01 6.47280097e-01 1.04794216e+00
2.11272776e-01 6.02864146e-01 8.66428316e-01 -7.43093431e-01
-1.08087647e+00 -5.00262201e-01 -3.25229883e-01 -2.06105903e-01
6.89711750e-01 -1.19680285e+00 -5.13652861e-01 -5.65797806e-01]
|
[4.737216949462891, 1.8641114234924316]
|
1f270e54-ecd7-4d20-b568-7916f30556ef
|
conditional-cross-design-synthesis-estimators
|
2109.13288
| null |
https://arxiv.org/abs/2109.13288v1
|
https://arxiv.org/pdf/2109.13288v1.pdf
|
Conditional Cross-Design Synthesis Estimators for Generalizability in Medicaid
|
While much of the causal inference literature has focused on addressing internal validity biases, both internal and external validity are necessary for unbiased estimates in a target population of interest. However, few generalizability approaches exist for estimating causal quantities in a target population when the target population is not well-represented by a randomized study but is reflected when additionally incorporating observational data. To generalize to a target population represented by a union of these data, we propose a class of novel conditional cross-design synthesis estimators that combine randomized and observational data, while addressing their respective biases. The estimators include outcome regression, propensity weighting, and double robust approaches. All use the covariate overlap between the randomized and observational data to remove potential unmeasured confounding bias. We apply these methods to estimate the causal effect of managed care plans on health care spending among Medicaid beneficiaries in New York City.
|
['Sherri Rose', 'Jacob Wallace', 'Tim Layton', 'Irina Degtiar']
|
2021-09-27
| null | null | null | null |
['design-synthesis']
|
['adversarial']
|
[ 3.19697708e-01 3.38031173e-01 -1.51036906e+00 -5.78278482e-01
-8.41523409e-01 -2.13651985e-01 3.19452852e-01 3.96073610e-01
-4.96844262e-01 1.08407915e+00 1.07117593e+00 -9.41466451e-01
-4.29549992e-01 -9.21768010e-01 -7.67806530e-01 -6.71228915e-02
-1.39142200e-01 3.01741302e-01 -5.29184580e-01 4.68407094e-01
1.85159579e-01 2.63303399e-01 -7.01155186e-01 1.83495395e-02
1.42854011e+00 2.00463966e-01 -3.73296171e-01 4.37945426e-02
4.01697814e-01 8.53302360e-01 5.80393337e-02 -1.34631112e-01
5.73108159e-02 -5.82210183e-01 -4.24396276e-01 -2.23119065e-01
5.66561997e-01 -6.76763594e-01 -1.34620667e-01 8.39964569e-01
6.08571112e-01 -3.17656845e-01 1.06020761e+00 -1.24140573e+00
-9.63998497e-01 1.09239662e+00 -6.18635297e-01 4.97689098e-02
5.62120318e-01 1.76256806e-01 8.29067051e-01 -4.25478339e-01
5.71526051e-01 1.43389440e+00 7.74174094e-01 1.95651218e-01
-1.69790077e+00 -1.26484621e+00 2.52873600e-01 -4.69407469e-01
-7.52996922e-01 -5.57726383e-01 1.67993739e-01 -8.44364703e-01
6.99743211e-01 7.25919306e-02 5.45840204e-01 1.06958115e+00
5.25598466e-01 -4.05824482e-02 1.29974282e+00 -4.74769443e-01
3.13803852e-01 -1.14177525e-01 5.55880666e-01 2.76434541e-01
1.00000036e+00 1.02886403e+00 1.47640407e-02 -1.14097023e+00
8.48517179e-01 4.20660615e-01 -2.46765405e-01 -3.86078686e-01
-1.51319098e+00 1.15137339e+00 3.20658296e-01 -8.31655785e-02
-8.96734297e-01 2.42631748e-01 2.81384408e-01 1.08957842e-01
2.50114441e-01 1.88957885e-01 -5.93291938e-01 2.98724324e-01
-9.11301196e-01 4.12447691e-01 5.83740532e-01 8.95321846e-01
3.46730411e-01 -1.07038170e-01 -5.17453790e-01 2.88467288e-01
3.84720534e-01 1.26545918e+00 2.59371340e-01 -1.05117965e+00
5.77181458e-01 7.62180865e-01 5.03890336e-01 -4.23952997e-01
-6.18678868e-01 1.53737769e-01 -9.45551753e-01 2.01738849e-01
3.58486861e-01 -6.87947035e-01 -1.01715815e+00 2.19490504e+00
1.67706951e-01 -2.06936985e-01 1.44712642e-01 4.48529094e-01
4.55061227e-01 -1.46114618e-01 8.32525551e-01 -7.55355597e-01
1.48657739e+00 -2.62585193e-01 -9.67331588e-01 -8.39691684e-02
7.49529243e-01 -4.47082013e-01 6.94561303e-01 -2.84729093e-01
-1.52655315e+00 -5.48452921e-02 -5.06513953e-01 1.20651595e-01
1.41505629e-01 -1.77672148e-01 5.41363299e-01 8.71982455e-01
-4.77227896e-01 3.01414460e-01 -6.29051685e-01 -5.04422963e-01
6.40018463e-01 3.38951528e-01 -2.29500294e-01 -1.57832190e-01
-1.10260952e+00 8.45707178e-01 1.01559393e-01 -5.02630770e-01
-6.47974491e-01 -1.45919394e+00 -8.81835580e-01 2.64325619e-01
1.35752857e-01 -1.50331390e+00 1.25233400e+00 -9.15994942e-01
-5.88972151e-01 5.21901667e-01 -4.77242738e-01 -1.69267759e-01
3.12460244e-01 1.02560081e-01 -4.37859535e-01 -1.42190382e-01
6.61687851e-01 6.44708872e-02 3.01727742e-01 -8.57706308e-01
-6.48457944e-01 -9.17644083e-01 -3.09521645e-01 1.97807625e-01
-5.23778461e-02 5.75008273e-01 4.82988715e-01 -9.09889877e-01
8.40503499e-02 -9.16883290e-01 -7.87867427e-01 -4.08774167e-01
-3.02083671e-01 7.53627717e-02 -1.16612418e-02 -4.17001992e-01
1.56786716e+00 -1.87916660e+00 -3.76545697e-01 1.87315851e-01
9.37198102e-02 -4.74754661e-01 1.58156097e-01 4.78400946e-01
-6.68244898e-01 4.78903860e-01 -3.45571250e-01 3.91535699e-01
-2.34263495e-01 -1.78318292e-01 -3.42261046e-01 7.19119310e-01
-3.70099470e-02 9.82931674e-01 -7.11842537e-01 -5.69592595e-01
2.58212000e-01 -1.03775136e-01 -9.93785799e-01 2.92117638e-03
4.46419001e-01 2.10721806e-01 -7.29990423e-01 4.21865165e-01
8.28732491e-01 -4.06481147e-01 6.49366915e-01 1.12854749e-01
-3.89390290e-01 6.11279249e-01 -1.02650511e+00 1.08981872e+00
-2.04402089e-01 -1.38842046e-01 8.70071650e-02 -8.45322013e-01
3.67520183e-01 6.45710707e-01 5.45497179e-01 -5.15332103e-01
-5.20568900e-02 2.21400633e-01 2.91396499e-01 -6.37620389e-01
-8.37281272e-02 -1.14039528e+00 -3.31629068e-01 6.11551344e-01
-4.76342142e-01 3.86930734e-01 -3.65672290e-01 7.68126175e-02
1.17826009e+00 -3.99851114e-01 1.06093550e+00 -7.02572465e-01
-1.15645513e-01 2.26864636e-01 1.05501103e+00 9.88145828e-01
-2.12737337e-01 6.04919016e-01 3.87089461e-01 -1.59936562e-01
-9.42009270e-01 -1.39254856e+00 -7.93557405e-01 4.34489965e-01
-3.45254391e-01 3.34760398e-01 -1.23987615e-01 -6.31659448e-01
6.38152540e-01 9.89679039e-01 -8.67772043e-01 -3.35620880e-01
-3.42180282e-01 -1.16375458e+00 3.91292512e-01 8.23878884e-01
1.61292329e-01 -7.22589374e-01 -6.46515608e-01 2.87318796e-01
-5.84223941e-02 -3.53778452e-01 -4.77738678e-01 -1.88383296e-01
-1.40310645e+00 -1.62121654e+00 -8.91693830e-01 -5.02890587e-01
5.72729707e-01 1.86809435e-01 1.17819703e+00 -2.76003420e-01
2.26027876e-01 4.21503991e-01 1.30530819e-01 -6.29427910e-01
-4.71856385e-01 -3.48655879e-01 2.44778514e-01 -6.33926034e-01
8.27623606e-01 -3.94880861e-01 -8.76947343e-01 1.09854594e-01
-6.17175400e-01 -2.67770261e-01 5.21000266e-01 7.79319286e-01
5.73670082e-02 -5.49884617e-01 1.10022187e+00 -1.15410149e+00
4.12161022e-01 -9.39275801e-01 -4.74387795e-01 3.13724458e-01
-1.15238822e+00 -2.03174025e-01 -6.41921088e-02 -6.11144006e-01
-1.44876242e+00 -3.56557637e-01 5.20802081e-01 3.24280113e-01
-2.41167843e-01 8.77606630e-01 -2.02034265e-01 5.51399529e-01
8.55581045e-01 -8.58898282e-01 2.68008947e-01 -2.54072994e-01
7.31473193e-02 8.28633666e-01 -9.55616459e-02 -5.55101395e-01
3.02690089e-01 6.01690173e-01 -1.49523035e-01 -2.56634876e-03
-4.47514862e-01 -4.09334362e-01 -2.87463851e-02 4.35352534e-01
8.63531172e-01 -1.34639907e+00 -1.10293853e+00 -2.23529845e-01
-6.00710630e-01 -3.97014916e-01 -4.89220530e-01 1.64579129e+00
-6.36829019e-01 1.45426476e-02 -2.23026931e-01 -9.03204381e-01
-2.15464711e-01 -1.15641606e+00 5.76981366e-01 -8.48366320e-02
-7.30410218e-01 -1.08270311e+00 5.21607697e-01 1.18600428e-01
2.43969887e-01 3.85642052e-01 1.44007099e+00 -3.73863369e-01
-1.86058074e-01 -2.78242201e-01 -4.68426138e-01 -3.95989597e-01
6.58675253e-01 2.06387881e-02 -3.37120414e-01 -3.64174128e-01
1.76210165e-01 -5.76370582e-02 6.08608723e-01 1.71031523e+00
7.97294855e-01 -5.50766170e-01 -8.58558416e-01 1.68621629e-01
1.45446503e+00 5.47816694e-01 5.19194126e-01 1.14228740e-01
1.91002980e-01 8.93548369e-01 4.00716484e-01 5.77813327e-01
6.46723568e-01 4.04385984e-01 2.36218479e-02 -4.37576383e-01
3.41857731e-01 -3.38809371e-01 -1.73888560e-02 -7.91379884e-02
-1.80608377e-01 8.73231515e-02 -9.78567660e-01 8.73315394e-01
-1.92698205e+00 -1.18962896e+00 -6.76174760e-01 2.73405457e+00
8.33622336e-01 -1.69930369e-01 3.78877074e-01 -4.82374758e-01
9.98206854e-01 -2.01810211e-01 -5.38655758e-01 -3.95653993e-01
4.83522192e-02 3.83659005e-02 1.05618286e+00 4.71403539e-01
-6.25049770e-01 7.65070617e-02 8.44371510e+00 4.99849506e-02
-3.86932015e-01 1.05972812e-01 7.96662092e-01 -1.34136185e-01
-7.39212275e-01 7.00277865e-01 -5.38568556e-01 2.50335246e-01
1.18078256e+00 -5.46020687e-01 -2.53511906e-01 5.39315939e-01
1.11469316e+00 -3.54904711e-01 -1.21047974e+00 1.91643775e-01
-4.94719386e-01 -1.10331857e+00 -1.89783633e-01 5.60400248e-01
1.19179761e+00 -2.83655077e-01 9.07291397e-02 2.72190362e-01
1.04581642e+00 -1.17136121e+00 4.26807702e-01 5.84553778e-01
1.14521170e+00 -3.01405549e-01 7.84021616e-01 4.59447280e-02
-4.78536487e-01 -3.92561048e-01 -1.81692570e-01 -5.53639054e-01
5.20946503e-01 1.08961320e+00 -4.55763042e-01 3.46300691e-01
5.96211791e-01 5.14005780e-01 2.00592786e-01 9.31267023e-01
8.01933706e-02 8.69362116e-01 1.43858582e-01 5.92368066e-01
-1.32656157e-01 -1.48437753e-01 1.70556843e-01 6.62672222e-01
4.72819954e-01 4.82806325e-01 -2.03035548e-01 9.40996170e-01
-1.95420399e-01 8.90832171e-02 -9.07950878e-01 1.97574526e-01
4.42937940e-01 4.81895447e-01 -1.57033741e-01 -6.18954182e-01
-9.88983810e-01 -1.03586242e-01 3.63306068e-02 6.05980039e-01
-6.63216531e-01 3.93866271e-01 6.07721388e-01 4.03679252e-01
-2.52606034e-01 4.24531609e-01 -8.03867161e-01 -1.17894495e+00
-3.51703078e-01 -1.07283413e+00 8.53520155e-01 -4.28392440e-01
-1.48308480e+00 -5.32327712e-01 6.26550734e-01 -8.56347919e-01
-7.03539625e-02 1.82554591e-02 -4.05131102e-01 1.27286804e+00
-1.20439231e+00 -7.81235635e-01 1.08757213e-01 3.75233740e-01
-5.85113093e-02 1.38293386e-01 8.21055591e-01 1.57320008e-01
-6.91806912e-01 3.07851791e-01 1.45827845e-01 -2.36815780e-01
1.17832673e+00 -1.10982680e+00 -6.71949610e-02 3.96971703e-01
-1.11837304e+00 1.11253834e+00 4.78616297e-01 -1.48357522e+00
-7.37166762e-01 -9.67531264e-01 1.22670674e+00 -2.78379798e-01
4.54738408e-01 2.83005178e-01 -6.00592732e-01 1.28223050e+00
1.37485638e-01 -4.25177038e-01 9.25978720e-01 8.04745436e-01
-3.84599417e-01 1.90744281e-01 -1.36993682e+00 7.63317227e-01
9.91210043e-01 -2.49627024e-01 -9.76890981e-01 1.07948937e-01
6.71556890e-01 9.41448212e-02 -1.39335871e+00 9.30660725e-01
8.61272931e-01 -7.91677773e-01 9.92057383e-01 -1.17501295e+00
6.66975796e-01 2.85780132e-01 -6.50611520e-02 -9.99607801e-01
-6.01758718e-01 -1.86156809e-01 4.68220592e-01 9.87745404e-01
6.77772641e-01 -9.26991463e-01 4.95668679e-01 1.35582340e+00
9.95710939e-02 -1.39210656e-01 -7.15860307e-01 -5.44379771e-01
7.01815426e-01 -1.34303598e-02 8.91828299e-01 1.35444641e+00
5.64008296e-01 1.36024430e-01 -1.94748327e-01 2.18909025e-01
8.45296741e-01 5.06239593e-01 5.28897047e-01 -1.22242463e+00
2.93472223e-02 -2.36831486e-01 2.67101437e-01 -2.19154924e-01
6.37064502e-02 -4.32770312e-01 -4.10628855e-01 -1.78806520e+00
1.16665435e+00 -6.99357331e-01 -1.16620116e-01 3.80247861e-01
-6.36868417e-01 -4.48546499e-01 -2.44756237e-01 1.74890965e-01
3.03167909e-01 5.09792507e-01 1.09118152e+00 5.47506697e-02
-4.96950865e-01 3.70052129e-01 -1.15993428e+00 6.85862541e-01
8.44529629e-01 -9.26270962e-01 -3.03454936e-01 4.36702417e-03
1.64836034e-01 6.71951771e-01 6.14756286e-01 -3.22708338e-01
-2.08220512e-01 -8.38414133e-01 2.80734718e-01 -3.05268288e-01
-5.63307166e-01 -1.09562707e+00 6.67168617e-01 9.45536673e-01
-8.51021111e-01 2.33951494e-01 1.34407625e-01 6.74567521e-01
8.20331424e-02 -2.63413638e-01 5.31561494e-01 -3.10043413e-02
4.31718469e-01 1.99107647e-01 -5.95319867e-01 2.57584065e-01
8.35917115e-01 -2.61956956e-02 -5.75967550e-01 -3.45136136e-01
-5.77123344e-01 1.88888341e-01 7.50916183e-01 3.80077725e-03
2.09014177e-01 -1.55377305e+00 -9.52963471e-01 -2.24616602e-01
3.60550106e-01 -6.05135560e-01 3.64633083e-01 1.21009958e+00
2.59780079e-01 6.47549093e-01 -4.90652919e-02 6.92894831e-02
-9.36049342e-01 1.13498688e+00 2.05197439e-01 -2.08698615e-01
-5.09340167e-01 -2.13541329e-01 8.50375414e-01 -3.27411264e-01
-1.06204093e-01 -5.40534139e-01 3.38242948e-02 -1.08796597e-01
5.53971946e-01 7.64575779e-01 -4.08495843e-01 -4.24290299e-01
-4.30353791e-01 1.91238314e-01 3.05401534e-01 -2.78012663e-01
1.14113259e+00 -3.90819848e-01 -1.81819126e-01 5.04003763e-01
1.12275302e+00 2.94150770e-01 -7.65414476e-01 -8.80621970e-02
-1.57229289e-01 -4.18879211e-01 -3.28371790e-03 -5.91756701e-01
-4.94109303e-01 2.05873385e-01 6.36411607e-01 -3.02584898e-02
1.11594868e+00 -1.57637045e-01 -1.34621114e-01 -5.10486424e-01
9.60060060e-02 -6.63205504e-01 -6.20289803e-01 -2.48526245e-01
7.10528195e-01 -1.51388049e+00 3.68347228e-01 -3.27571303e-01
-3.87258440e-01 4.67855871e-01 1.27489388e-01 -2.14340836e-01
9.36729908e-01 4.53862809e-02 -3.18773657e-01 -3.51877123e-01
-6.88030601e-01 7.96368462e-04 1.27810299e-01 5.93067229e-01
8.62872958e-01 4.78427649e-01 -1.17182386e+00 5.43386221e-01
2.77829945e-01 5.74818432e-01 6.07370734e-01 8.40180457e-01
2.66086422e-02 -8.96690428e-01 -7.54008293e-01 1.13126802e+00
-8.41480970e-01 -4.24961776e-01 4.68512215e-02 1.13026583e+00
-1.62335455e-01 1.15502954e+00 2.35793859e-01 5.19093513e-01
7.43455350e-01 -4.92702127e-02 8.00259858e-02 -6.55196846e-01
-6.67628884e-01 2.13580117e-01 3.39901447e-01 -3.67972612e-01
-1.21996689e+00 -1.07261050e+00 -9.56864953e-01 -4.84823734e-01
-5.74169576e-01 6.83599487e-02 -5.85183734e-03 7.94932902e-01
2.69711822e-01 3.87735903e-01 6.21752977e-01 -1.59276441e-01
-6.91164494e-01 -9.57779944e-01 -6.24176502e-01 3.56327564e-01
4.48212832e-01 -6.67504668e-01 -4.96329129e-01 -2.65418738e-01]
|
[7.974704742431641, 5.394338130950928]
|
e20d4d80-040c-4872-94b8-14cedb481b62
|
on-evaluating-adversarial-robustness-of-large
|
2305.16934
| null |
https://arxiv.org/abs/2305.16934v1
|
https://arxiv.org/pdf/2305.16934v1.pdf
|
On Evaluating Adversarial Robustness of Large Vision-Language Models
|
Large vision-language models (VLMs) such as GPT-4 have achieved unprecedented performance in response generation, especially with visual inputs, enabling more creative and adaptable interaction than large language models such as ChatGPT. Nonetheless, multimodal generation exacerbates safety concerns, since adversaries may successfully evade the entire system by subtly manipulating the most vulnerable modality (e.g., vision). To this end, we propose evaluating the robustness of open-source large VLMs in the most realistic and high-risk setting, where adversaries have only black-box system access and seek to deceive the model into returning the targeted responses. In particular, we first craft targeted adversarial examples against pretrained models such as CLIP and BLIP, and then transfer these adversarial examples to other VLMs such as MiniGPT-4, LLaVA, UniDiffuser, BLIP-2, and Img2Prompt. In addition, we observe that black-box queries on these VLMs can further improve the effectiveness of targeted evasion, resulting in a surprisingly high success rate for generating targeted responses. Our findings provide a quantitative understanding regarding the adversarial vulnerability of large VLMs and call for a more thorough examination of their potential security flaws before deployment in practice. Code is at https://github.com/yunqing-me/AttackVLM.
|
['Min Lin', 'Ngai-Man Cheung', 'Chongxuan Li', 'Xiao Yang', 'Chao Du', 'Tianyu Pang', 'Yunqing Zhao']
|
2023-05-26
| null | null | null | null |
['response-generation', 'multimodal-generation']
|
['natural-language-processing', 'natural-language-processing']
|
[ 7.38265812e-02 1.87199846e-01 1.29021183e-01 1.79578252e-02
-1.04101396e+00 -1.31887197e+00 8.62875700e-01 -5.82644224e-01
-4.48415637e-01 5.11546373e-01 1.00780167e-01 -6.99078858e-01
3.97525549e-01 -7.35543489e-01 -8.62049222e-01 -2.85548031e-01
1.11712843e-01 3.21412385e-01 -2.01697379e-01 -4.15138870e-01
-8.47779065e-02 3.68470788e-01 -8.62962067e-01 2.93456703e-01
6.97895527e-01 5.23078024e-01 -2.67255843e-01 1.27437055e+00
4.52644169e-01 1.00243592e+00 -9.34826553e-01 -1.05650020e+00
2.76635110e-01 -1.19153298e-01 -8.14299107e-01 -4.34087455e-01
6.37680769e-01 -8.37954462e-01 -8.11003447e-01 8.01687837e-01
8.14675748e-01 1.10846356e-01 4.37917024e-01 -1.71149337e+00
-1.07447493e+00 7.31833220e-01 -4.02821571e-01 -7.03842118e-02
7.54872024e-01 1.31613302e+00 7.58390844e-01 -5.61015844e-01
4.74835247e-01 1.67774987e+00 3.60590428e-01 1.13938606e+00
-1.22045183e+00 -9.06974852e-01 8.58231932e-02 -2.14722976e-01
-1.12550104e+00 -6.84934795e-01 2.80702084e-01 -2.01600656e-01
8.58599782e-01 7.53911972e-01 3.71837914e-02 2.12404013e+00
2.23187753e-03 8.85893703e-01 1.01567340e+00 -4.20589326e-03
-5.11368401e-02 2.90264130e-01 -6.15311079e-02 6.77122891e-01
-1.25437472e-02 4.76211697e-01 -2.88964808e-01 -7.40356326e-01
5.58584929e-01 -2.38085821e-01 -4.32233751e-01 2.85889089e-01
-1.02948153e+00 9.80029643e-01 5.12575805e-01 -1.63526505e-01
-2.79203922e-01 6.17683887e-01 2.69504815e-01 3.37443560e-01
3.13829295e-02 6.98581100e-01 -1.28738001e-01 -2.24079013e-01
-3.30234736e-01 4.80740160e-01 8.57701480e-01 8.32229018e-01
4.53924388e-01 2.42806420e-01 -8.05419743e-01 3.99788588e-01
1.16123185e-01 9.20079768e-01 3.09398361e-02 -1.25349367e+00
5.84384501e-01 1.54314891e-01 2.05950439e-01 -9.37417269e-01
-3.34170833e-02 -6.79405266e-03 -5.52250981e-01 3.75806659e-01
3.93648386e-01 -7.59511352e-01 -9.32813883e-01 2.07123160e+00
2.11040646e-01 9.54984035e-03 3.77061874e-01 9.73000050e-01
7.94579029e-01 7.83806920e-01 2.49710888e-01 2.85972208e-01
1.12348664e+00 -8.41295302e-01 -1.62922353e-01 -4.21081662e-01
4.53396112e-01 -6.98716223e-01 1.64596021e+00 1.82443500e-01
-1.14896655e+00 -2.59572208e-01 -4.21421915e-01 6.06055371e-02
-1.38629794e-01 -2.54115254e-01 6.26313686e-01 7.16740429e-01
-1.19627166e+00 5.16931228e-02 -6.20967209e-01 -3.53933960e-01
5.00099003e-01 2.61274844e-01 -3.59426558e-01 -5.78600056e-02
-1.37760365e+00 7.20886707e-01 -8.43101442e-02 -8.06457400e-02
-1.66004264e+00 -5.68770528e-01 -7.03872383e-01 -4.05260362e-02
5.45093656e-01 -9.61528659e-01 1.52352238e+00 -8.01496506e-01
-1.33377612e+00 7.12035716e-01 1.47758856e-01 -4.93456960e-01
8.33388805e-01 -1.61467299e-01 -1.42765641e-01 2.79111326e-01
-1.48834854e-01 9.66152310e-01 1.14925992e+00 -1.40313566e+00
-4.18140255e-02 -8.20875242e-02 8.00399601e-01 2.43886963e-01
-4.00235325e-01 3.34836721e-01 -3.71356636e-01 -5.64671040e-01
-9.97498870e-01 -1.14096177e+00 -3.03007901e-01 -2.20850021e-01
-1.06486511e+00 5.30386269e-02 8.03255558e-01 -6.15485311e-01
9.40053344e-01 -2.15383363e+00 1.02013141e-01 1.58774331e-01
5.41027009e-01 5.35166621e-01 -6.51213944e-01 6.87622428e-01
1.53961748e-01 6.98229969e-01 8.89428332e-02 -3.15162897e-01
1.68082759e-01 -6.34706095e-02 -7.84660101e-01 1.00196628e-02
3.08356006e-02 1.54637098e+00 -6.89144194e-01 -2.33498320e-01
3.34097706e-02 4.07415003e-01 -6.12532437e-01 4.18113858e-01
-5.22730768e-01 5.33778131e-01 -5.43583512e-01 6.71021163e-01
4.57059622e-01 -1.35811999e-01 -3.33290517e-01 2.03620777e-01
2.32945964e-01 -1.32897243e-01 -4.01740938e-01 1.10623348e+00
-4.86225456e-01 6.48425877e-01 2.47411430e-01 -8.91269594e-02
2.87674814e-01 2.20195621e-01 -1.37587920e-01 -5.30137718e-01
1.83800519e-01 -3.18507224e-01 -7.18544498e-02 -4.46227223e-01
4.82683957e-01 2.83243001e-01 -4.84011799e-01 6.75596833e-01
-3.00201535e-01 -1.91832885e-01 -2.96100795e-01 9.59946692e-01
1.29732990e+00 -4.84061956e-01 -3.14518422e-01 5.25047004e-01
1.57427803e-01 -1.98406689e-02 -1.21954903e-01 1.54179788e+00
-2.55329937e-01 4.57911193e-01 7.41677284e-01 9.01803598e-02
-8.23478580e-01 -1.20201290e+00 6.35029256e-01 1.27968991e+00
-4.51517925e-02 -3.77568811e-01 -1.00755477e+00 -9.42625880e-01
1.88311413e-01 1.07052457e+00 -4.72449303e-01 -6.80553317e-01
-2.75032282e-01 -4.05946940e-01 1.46030688e+00 2.22886816e-01
5.29633939e-01 -1.20746970e+00 -4.21604425e-01 -2.85441935e-01
-4.04402077e-01 -1.08640182e+00 -8.51459086e-01 -6.04440689e-01
-1.39780015e-01 -9.11850333e-01 -6.10497653e-01 -1.74688220e-01
5.05842566e-01 4.12764460e-01 1.01241171e+00 2.90533483e-01
-3.21046293e-01 9.72194731e-01 -2.65440911e-01 -3.51434320e-01
-8.46655667e-01 -3.22924228e-03 1.01167820e-01 -5.51942550e-02
-1.02235340e-01 -2.99190670e-01 -5.51819921e-01 3.20729405e-01
-9.41320837e-01 -2.01699883e-01 4.39828455e-01 6.14974201e-01
-2.60537826e-02 -4.24249411e-01 3.47825259e-01 -8.09197068e-01
1.29403794e+00 -5.16480625e-01 -6.16656005e-01 3.29062194e-01
-2.28002250e-01 -2.90203720e-01 8.30797315e-01 -9.66486752e-01
-9.42891359e-01 -4.14288461e-01 -2.12324142e-01 -7.86585748e-01
-1.97433636e-01 2.09238932e-01 -1.72961310e-01 -5.40784597e-01
1.07551706e+00 3.06722432e-01 -1.09497746e-02 2.24048505e-03
7.71543443e-01 6.22583151e-01 5.73638260e-01 -6.83697224e-01
1.35757053e+00 1.90719455e-01 -5.02346158e-01 -5.94253182e-01
-3.19395006e-01 3.39484334e-01 5.74378908e-01 -2.62277335e-01
8.38945270e-01 -7.55484402e-01 -1.48736012e+00 7.39558578e-01
-1.27662945e+00 -8.26009989e-01 1.81656331e-02 6.49433699e-04
-3.56332093e-01 3.75660539e-01 -9.11263943e-01 -9.62792575e-01
-6.01392865e-01 -1.20053256e+00 9.59620714e-01 3.43813807e-01
-4.02243853e-01 -8.23987305e-01 -1.13577656e-02 8.18880737e-01
7.33226478e-01 2.27357954e-01 8.00471961e-01 -7.06212044e-01
-9.15059686e-01 -4.44970936e-01 -1.64653197e-01 3.04811120e-01
-2.85068482e-01 1.66594058e-01 -1.06778681e+00 -5.11267900e-01
-2.51559347e-01 -1.02347720e+00 6.76053643e-01 4.00270894e-03
1.19097316e+00 -9.55356658e-01 -2.81109184e-01 6.71467602e-01
9.37780142e-01 3.19735110e-02 5.45944035e-01 5.29222824e-02
9.19811964e-01 3.66309971e-01 4.07480299e-01 3.39844137e-01
4.17375892e-01 3.92128319e-01 7.55753636e-01 -1.81856409e-01
2.42121503e-01 -5.53116620e-01 8.61676514e-01 -1.41737878e-01
1.79836765e-01 -8.73579621e-01 -8.46985102e-01 1.07676931e-01
-1.70145226e+00 -1.06599915e+00 2.03297272e-01 2.25879121e+00
8.81113231e-01 1.65525645e-01 1.39292151e-01 -6.68271303e-01
4.03611481e-01 3.23726743e-01 -7.10724115e-01 -6.01647556e-01
-9.99399200e-02 4.74473275e-03 5.44764459e-01 8.22130203e-01
-8.62268746e-01 1.42272615e+00 6.55386353e+00 8.46551955e-01
-1.05155134e+00 2.04839155e-01 8.83420169e-01 -5.85248232e-01
-5.79422176e-01 1.20325740e-02 -5.75818062e-01 5.30443728e-01
9.67631221e-01 -3.10454518e-01 1.01791728e+00 6.28965676e-01
3.19248319e-01 1.39618844e-01 -1.03009880e+00 6.90233409e-01
-3.40825394e-02 -1.25671351e+00 3.42274576e-01 2.62085438e-01
3.78031760e-01 5.38485758e-02 6.71537161e-01 6.38806641e-01
9.45057571e-01 -1.38515687e+00 5.13321340e-01 4.34023201e-01
1.01293898e+00 -7.90924549e-01 2.43327007e-01 5.72199404e-01
-4.01650518e-01 -1.24555014e-01 -1.56007111e-01 1.08193509e-01
2.14999437e-01 -2.14778587e-01 -8.65701675e-01 1.22626476e-01
4.70512688e-01 -2.71397382e-01 -5.98021686e-01 3.66317928e-01
-3.76578778e-01 8.33410621e-01 -2.84177005e-01 -4.67195883e-02
3.87811154e-01 2.06819639e-01 9.00147021e-01 9.49375570e-01
5.56778163e-04 1.51416972e-01 3.45775671e-02 1.19926000e+00
-3.06592166e-01 -3.71269792e-01 -1.05803859e+00 -5.58031261e-01
5.84486008e-01 1.38565850e+00 1.47475302e-01 -1.87290251e-01
-1.96375415e-01 1.23832476e+00 3.09529275e-01 8.45314324e-01
-1.32065713e+00 -1.60505101e-01 1.13151801e+00 -1.59121722e-01
-3.07151616e-01 -1.51597768e-01 6.10298999e-02 -1.13704860e+00
-1.02048218e-01 -1.59630501e+00 2.92714566e-01 -1.02270758e+00
-1.37108123e+00 7.41114318e-01 -5.09829149e-02 -6.12734497e-01
-5.57418942e-01 -2.79329121e-01 -9.39929128e-01 1.13997090e+00
-8.84807467e-01 -1.52560925e+00 -2.71498919e-01 1.14731836e+00
3.10907334e-01 -3.24663669e-01 8.50222766e-01 -1.26799745e-02
-7.20687807e-01 1.37070358e+00 -4.19223249e-01 2.80960947e-01
8.83326709e-01 -8.85939837e-01 1.01296771e+00 9.61684883e-01
4.42765504e-02 8.65070760e-01 7.48741925e-01 -7.18687057e-01
-1.79194832e+00 -1.20665336e+00 2.35200211e-01 -1.18146682e+00
6.96455300e-01 -7.06329048e-01 -6.96014702e-01 9.15533185e-01
3.62782449e-01 -3.28687459e-01 4.58700985e-01 -3.09535056e-01
-6.19598389e-01 2.75119513e-01 -1.24745548e+00 1.37916112e+00
9.99604046e-01 -8.87346983e-01 1.13196124e-03 5.93287349e-01
1.28086877e+00 -4.70825732e-01 -3.13709527e-01 -1.07713602e-02
4.09568131e-01 -9.00502980e-01 1.25820780e+00 -9.86584961e-01
4.94175524e-01 1.74404189e-01 -1.77433029e-01 -1.27920175e+00
-9.19181779e-02 -1.29403675e+00 -2.11934075e-01 1.28916824e+00
4.50882524e-01 -1.03526306e+00 6.82581604e-01 1.10996282e+00
4.29202169e-01 -4.58681464e-01 -4.99934047e-01 -5.14851391e-01
2.23381966e-01 -5.59861541e-01 3.52201253e-01 7.08422422e-01
-3.28243256e-01 2.15027183e-01 -9.22891438e-01 4.78563935e-01
7.34977007e-01 -4.01459873e-01 1.29199266e+00 -3.41103345e-01
-7.28658676e-01 -3.11530083e-01 9.12669450e-02 -8.25724721e-01
3.08298886e-01 -6.92721784e-01 -2.70683706e-01 -1.06551123e+00
1.23229347e-01 -3.38474512e-01 1.00531615e-01 8.52118492e-01
-3.67827594e-01 3.60213727e-01 6.47372305e-01 2.15727925e-01
-4.15183365e-01 3.41672808e-01 1.23585665e+00 -2.96334624e-01
-1.24688245e-01 1.16609752e-01 -1.26302958e+00 4.73867774e-01
8.97881091e-01 -3.80253404e-01 -6.13553405e-01 -6.30725682e-01
1.38000563e-01 3.82415116e-01 9.53118861e-01 -4.89311129e-01
2.54813373e-01 -4.56381381e-01 1.92227796e-01 1.15011908e-01
4.93104279e-01 -4.42634732e-01 1.61186270e-02 3.43516380e-01
-5.54536939e-01 4.03621644e-02 3.37326407e-01 4.17247534e-01
3.03567827e-01 5.50942086e-02 4.79171157e-01 -3.13370496e-01
-3.99544328e-01 5.17266989e-01 -5.45027316e-01 2.09168047e-01
1.07707405e+00 4.67586927e-02 -1.00844955e+00 -1.04807663e+00
-4.97800052e-01 6.25499725e-01 6.81564152e-01 6.43665731e-01
9.27935898e-01 -1.02233863e+00 -8.45963120e-01 1.60058796e-01
1.25048369e-01 -4.03562188e-01 4.83749539e-01 5.99192083e-01
-3.88560385e-01 1.15390331e-01 -2.97929905e-02 -2.31787503e-01
-1.41442871e+00 7.72310019e-01 5.49470961e-01 -3.21698748e-02
-1.65750325e-01 1.08897698e+00 6.61546707e-01 -4.30795044e-01
3.41278017e-01 3.43093008e-01 2.51232296e-01 -4.92036432e-01
4.55817729e-01 2.25624517e-01 -6.20842040e-01 -4.15658325e-01
-4.09111321e-01 -5.24358563e-02 -3.37616295e-01 -5.16080558e-01
5.64818919e-01 5.53212827e-03 -1.05924979e-02 -2.50422984e-01
9.13313687e-01 3.20024848e-01 -1.16754556e+00 4.15298939e-02
-7.81069160e-01 -5.91693401e-01 -4.00248617e-01 -1.18646669e+00
-9.54938173e-01 7.19283581e-01 2.68833250e-01 2.11730540e-01
9.67632055e-01 8.05946589e-02 9.80203450e-01 5.88830292e-01
4.92103159e-01 -5.06662130e-01 4.09154683e-01 3.91208917e-01
1.12849414e+00 -1.38247466e+00 -5.82829773e-01 2.52033807e-02
-1.09863436e+00 3.95487040e-01 1.10190451e+00 5.18128760e-02
-1.02774262e-01 2.93418199e-01 3.47680300e-01 -4.07433659e-02
-1.14151573e+00 2.12776691e-01 1.68522373e-02 8.07067096e-01
-1.02671281e-01 1.63297579e-01 3.34618896e-01 5.27961016e-01
-2.17171833e-01 -4.94453937e-01 7.24509537e-01 6.88057065e-01
5.00683673e-02 -8.92541111e-01 -5.78038633e-01 1.24105863e-01
-6.25926137e-01 -3.13041985e-01 -1.06187761e+00 7.04412460e-01
-4.03962761e-01 1.29759705e+00 -4.78403598e-01 -8.12653959e-01
3.54704529e-01 -1.71737373e-01 2.15605363e-01 -4.55123097e-01
-9.93565619e-01 -3.02521616e-01 2.78387189e-01 -9.30607319e-01
5.69351256e-01 -1.67805225e-01 -8.89339864e-01 -9.02319968e-01
1.01225056e-01 -7.89266229e-02 3.03761721e-01 5.16356230e-01
6.04940355e-01 1.23169824e-01 7.69487441e-01 -7.64174640e-01
-1.15910757e+00 -7.35507667e-01 1.15702786e-01 3.76931161e-01
4.86849934e-01 -2.56755063e-03 -5.44978082e-01 -3.43404174e-01]
|
[5.897946834564209, 7.947128772735596]
|
350febaf-ce3a-4f05-9f96-bdb4e26b583d
|
isbnet-a-3d-point-cloud-instance-segmentation
|
2303.00246
| null |
https://arxiv.org/abs/2303.00246v2
|
https://arxiv.org/pdf/2303.00246v2.pdf
|
ISBNet: a 3D Point Cloud Instance Segmentation Network with Instance-aware Sampling and Box-aware Dynamic Convolution
|
Existing 3D instance segmentation methods are predominated by the bottom-up design -- manually fine-tuned algorithm to group points into clusters followed by a refinement network. However, by relying on the quality of the clusters, these methods generate susceptible results when (1) nearby objects with the same semantic class are packed together, or (2) large objects with loosely connected regions. To address these limitations, we introduce ISBNet, a novel cluster-free method that represents instances as kernels and decodes instance masks via dynamic convolution. To efficiently generate high-recall and discriminative kernels, we propose a simple strategy named Instance-aware Farthest Point Sampling to sample candidates and leverage the local aggregation layer inspired by PointNet++ to encode candidate features. Moreover, we show that predicting and leveraging the 3D axis-aligned bounding boxes in the dynamic convolution further boosts performance. Our method set new state-of-the-art results on ScanNetV2 (55.9), S3DIS (60.8), and STPLS3D (49.2) in terms of AP and retains fast inference time (237ms per scene on ScanNetV2). The source code and trained models are available at https://github.com/VinAIResearch/ISBNet.
|
['Khoi Nguyen', 'Binh-Son Hua', 'Tuan Duc Ngo']
|
2023-03-01
| null |
http://openaccess.thecvf.com//content/CVPR2023/html/Ngo_ISBNet_A_3D_Point_Cloud_Instance_Segmentation_Network_With_Instance-Aware_CVPR_2023_paper.html
|
http://openaccess.thecvf.com//content/CVPR2023/papers/Ngo_ISBNet_A_3D_Point_Cloud_Instance_Segmentation_Network_With_Instance-Aware_CVPR_2023_paper.pdf
|
cvpr-2023-1
|
['3d-instance-segmentation-1']
|
['computer-vision']
|
[-8.52403715e-02 5.89954592e-02 -3.00795764e-01 -4.97528881e-01
-8.93698871e-01 -7.18639374e-01 6.01850033e-01 -1.47939958e-02
-3.96598816e-01 1.75650120e-01 -1.49096817e-01 -1.73664719e-01
9.80804395e-03 -8.94587159e-01 -1.05788314e+00 -4.83502775e-01
-9.11654308e-02 6.88372254e-01 1.02157438e+00 2.86047887e-02
2.97085881e-01 8.82932365e-01 -1.68104041e+00 4.32732493e-01
1.06011665e+00 1.27902675e+00 3.66416007e-01 4.65909213e-01
-4.81381983e-01 3.28435332e-01 -4.22010541e-01 -3.34197015e-01
6.64285302e-01 2.00923488e-01 -6.62858427e-01 -3.57485227e-02
8.00315559e-01 -2.44284347e-01 -2.30315387e-01 8.56080234e-01
3.96392822e-01 2.04441577e-01 6.96361780e-01 -1.02135551e+00
-2.95748353e-01 5.32235861e-01 -9.27941561e-01 7.19310790e-02
-1.29793838e-01 4.71178442e-01 9.82514322e-01 -1.18976498e+00
4.13959086e-01 1.18616951e+00 8.80048692e-01 4.00182724e-01
-1.31922185e+00 -7.60548770e-01 4.12629008e-01 -5.18031381e-02
-1.65298688e+00 -2.58600920e-01 5.60534358e-01 -3.41937006e-01
1.05285597e+00 3.14276129e-01 7.65641510e-01 7.63902962e-01
-3.56711417e-01 9.59418893e-01 6.78379595e-01 3.29596139e-02
2.23905727e-01 -7.22330660e-02 1.82556018e-01 6.79424822e-01
1.43661305e-01 1.19059399e-01 -3.45215231e-01 -2.50306368e-01
8.65362167e-01 2.30887067e-02 -1.01669215e-01 -5.84260762e-01
-1.32193673e+00 7.73309350e-01 9.12209749e-01 -5.21938317e-02
-4.47801292e-01 4.04447466e-01 2.56233573e-01 -1.91171393e-01
6.52833045e-01 4.86068964e-01 -6.32114053e-01 9.67641175e-02
-1.42509425e+00 3.66172910e-01 7.32631683e-01 1.34091389e+00
9.86350179e-01 -2.57454216e-01 -3.57137412e-01 8.76127303e-01
2.47591630e-01 5.96889734e-01 -7.64528364e-02 -9.09586191e-01
4.98147607e-01 8.99806738e-01 8.62363949e-02 -7.71292090e-01
-3.74096692e-01 -5.83705723e-01 -3.48958343e-01 2.59837151e-01
3.53295267e-01 1.36178493e-01 -1.48706353e+00 1.26419044e+00
6.14595115e-01 5.77338159e-01 -3.32615495e-01 1.17605782e+00
9.85058069e-01 6.82466149e-01 8.07068720e-02 6.86265647e-01
1.39167547e+00 -1.22123086e+00 9.21683088e-02 -2.59533256e-01
4.02564198e-01 -6.59715235e-01 1.13723576e+00 2.07021356e-01
-1.21171629e+00 -6.37057126e-01 -9.23642218e-01 -1.46851510e-01
-3.71193975e-01 3.11806232e-01 7.17696369e-01 3.75902176e-01
-1.03385341e+00 5.18503845e-01 -9.73170459e-01 -1.97199136e-01
1.10418904e+00 4.14884359e-01 1.21704288e-01 -1.38272077e-01
-6.47716582e-01 3.93946886e-01 4.47114766e-01 -9.52345207e-02
-7.50535309e-01 -1.19909370e+00 -6.23803318e-01 -5.39699160e-02
5.21771312e-01 -7.31244445e-01 1.09266913e+00 -7.44027734e-01
-1.14032602e+00 8.66971493e-01 -1.02891251e-01 -4.97313082e-01
6.22454047e-01 -3.58281255e-01 -4.30366062e-02 4.10195798e-01
3.90945375e-01 1.20183110e+00 7.44264305e-01 -1.37392342e+00
-1.14384079e+00 -3.15447569e-01 -3.23099084e-02 1.38813645e-01
2.12285131e-01 -3.26452672e-01 -1.06462038e+00 -5.89273095e-01
3.80351275e-01 -7.24740386e-01 -5.41966975e-01 2.58364767e-01
-8.01368594e-01 -3.34831774e-01 7.88053632e-01 -2.29003966e-01
1.01673114e+00 -2.26241446e+00 -2.22360194e-01 6.22086644e-01
3.25002551e-01 2.88025588e-01 -5.90854585e-02 4.37028781e-02
2.27488682e-01 1.59132510e-01 -3.86863559e-01 -4.19328719e-01
1.41730368e-01 1.16144426e-01 -3.06788385e-01 4.31961924e-01
4.36538339e-01 1.11938083e+00 -9.03193653e-01 -5.21408021e-01
6.14731729e-01 4.88823861e-01 -6.46221519e-01 -8.26136544e-02
-4.85974699e-01 1.95682883e-01 -6.82908475e-01 9.78133380e-01
1.04409480e+00 -4.73570377e-01 -4.92460370e-01 -6.33259237e-01
-3.30984116e-01 2.93938607e-01 -1.31824434e+00 1.93467307e+00
-1.73472986e-01 3.62917960e-01 -9.75801498e-02 -5.95360994e-01
7.42631733e-01 -2.07170367e-01 4.95936126e-01 -4.12164837e-01
-4.37526405e-02 3.22498202e-01 -4.13840145e-01 -1.79175615e-01
5.83842933e-01 4.90411699e-01 3.36135775e-02 9.19672176e-02
2.60168333e-02 -3.59159142e-01 1.39506295e-01 2.14535937e-01
1.11307836e+00 3.42917889e-01 -9.02556926e-02 -2.53296524e-01
2.11735796e-02 5.33373773e-01 3.38578045e-01 1.14543986e+00
8.52744933e-03 1.04570735e+00 3.08829755e-01 -4.56908137e-01
-1.02361095e+00 -1.28689873e+00 -3.73376906e-01 8.06055427e-01
4.73367900e-01 -1.61574349e-01 -6.26386940e-01 -8.25833559e-01
2.76551038e-01 6.79816604e-01 -5.27461886e-01 1.82821348e-01
-6.10448658e-01 -6.58524454e-01 7.64586926e-01 7.36773968e-01
5.42372286e-01 -8.41380298e-01 -5.53011537e-01 2.11480245e-01
5.41838482e-02 -1.10582626e+00 -4.91684318e-01 4.07708228e-01
-9.33405280e-01 -1.17604876e+00 -5.95392287e-01 -6.51983500e-01
7.60840595e-01 4.83499527e-01 1.32544875e+00 1.14696860e-01
-2.59014726e-01 1.40535340e-01 -4.65305805e-01 -1.90564275e-01
1.45115092e-01 4.46115941e-01 -3.98171782e-01 -2.03361079e-01
5.69981754e-01 -4.48028564e-01 -9.46202397e-01 4.75873351e-01
-6.59974515e-01 1.76022902e-01 7.42473602e-01 4.84865814e-01
1.15790141e+00 -3.23676676e-01 9.97109190e-02 -8.21311414e-01
-3.74577828e-02 -5.75849771e-01 -7.40384817e-01 5.86780682e-02
-2.82394767e-01 -4.77527194e-02 2.41388485e-01 -4.25847381e-01
-8.14590573e-01 4.11841840e-01 -2.12745637e-01 -7.61665940e-01
-4.78922427e-01 -1.61062460e-02 -4.53726761e-02 -2.50718385e-01
7.82176137e-01 1.53560832e-01 -2.85172760e-01 -5.09918094e-01
7.09593773e-01 4.33120281e-01 5.25682807e-01 -7.27917671e-01
9.57799911e-01 9.13902223e-01 -2.73118556e-01 -6.46887183e-01
-8.61673832e-01 -8.41887832e-01 -6.70613229e-01 -1.23277992e-01
1.01487768e+00 -1.09979129e+00 -4.12280500e-01 2.47434288e-01
-1.13492692e+00 -6.28070176e-01 -5.06103694e-01 4.30356205e-01
-4.28741366e-01 -1.06077671e-01 -5.25557339e-01 -5.40838599e-01
-3.56961638e-01 -1.20881271e+00 1.51255381e+00 1.95038840e-01
1.09284802e-03 -3.97822201e-01 -3.31358016e-01 1.22376919e-01
1.71252325e-01 3.28803629e-01 5.67084670e-01 -6.87376678e-01
-1.20337582e+00 -1.36026457e-01 -6.90552652e-01 2.09558681e-01
-2.19550774e-01 7.39222616e-02 -1.02659404e+00 1.28572822e-01
-5.43325186e-01 -2.81153154e-02 1.10649192e+00 6.21419609e-01
1.59489608e+00 -3.33265960e-02 -8.11445713e-01 1.10004604e+00
1.47451615e+00 -1.18196361e-01 5.36585391e-01 1.66247919e-01
1.02067852e+00 3.61440688e-01 6.28169179e-01 4.40927505e-01
4.21095669e-01 6.40307546e-01 5.43513000e-01 -3.95474613e-01
-3.69882524e-01 -2.04894930e-01 -1.94302171e-01 2.74004310e-01
-1.82159081e-01 -2.60842294e-01 -8.83379877e-01 7.82969356e-01
-1.83332837e+00 -8.49139214e-01 -6.44700289e-01 1.92844665e+00
6.82626009e-01 3.41683835e-01 2.04197824e-01 -3.48874122e-01
7.20495820e-01 1.31741345e-01 -6.80692911e-01 2.55649269e-01
-1.15784027e-01 5.33185482e-01 9.93599236e-01 3.41560245e-01
-1.29919231e+00 1.28333914e+00 5.21665716e+00 1.36806488e+00
-8.99467349e-01 4.70832363e-02 8.01427007e-01 -4.47022080e-01
-1.74014494e-01 -6.59637302e-02 -1.10205579e+00 5.34550369e-01
3.44445556e-01 6.00277185e-01 1.54198229e-01 1.03870249e+00
1.73392072e-01 -1.83370575e-01 -8.11364055e-01 8.84156168e-01
-2.68611282e-01 -1.65615177e+00 7.32602850e-02 -5.36848195e-02
7.54756629e-01 6.85759127e-01 -1.39276860e-02 1.71172023e-01
4.12010640e-01 -8.39927852e-01 1.08450294e+00 3.93738627e-01
7.28138983e-01 -8.46656263e-01 4.44340587e-01 1.28922209e-01
-1.46091998e+00 1.93484783e-01 -5.90692282e-01 4.96162176e-01
1.73692003e-01 9.06697333e-01 -8.51125181e-01 4.89878386e-01
1.16708398e+00 6.96293652e-01 -6.08926892e-01 1.41572762e+00
-2.52703637e-01 7.94051468e-01 -8.64517033e-01 5.70348538e-02
7.11269736e-01 -1.76091008e-02 5.90459704e-01 1.42154348e+00
3.00440013e-01 8.26786235e-02 2.95778006e-01 1.38683498e+00
7.56819695e-02 -1.28486350e-01 -1.55036747e-01 3.63351226e-01
7.26301134e-01 1.29237521e+00 -1.24984860e+00 -3.91018838e-01
-3.92600566e-01 9.27373648e-01 3.25866729e-01 4.04021591e-01
-1.10012126e+00 -3.86689156e-01 9.17849600e-01 4.17272419e-01
8.45234811e-01 -1.63675994e-01 -6.40488744e-01 -8.60135078e-01
2.68935338e-02 -2.29006290e-01 1.46426007e-01 -4.89336014e-01
-1.46983051e+00 5.75154960e-01 5.77638149e-02 -1.30687869e+00
3.61597806e-01 -4.98119831e-01 -5.32159925e-01 7.15632498e-01
-1.58003306e+00 -1.31088400e+00 -5.35569549e-01 4.91406381e-01
6.72068655e-01 2.06592605e-01 1.88738942e-01 3.98827910e-01
-2.39825845e-01 4.44926560e-01 -1.11480899e-01 1.55321255e-01
4.63292092e-01 -1.36248922e+00 9.48279738e-01 6.00457907e-01
1.14366002e-01 4.29360926e-01 1.41997904e-01 -7.90180087e-01
-9.43790734e-01 -1.56026649e+00 4.92194146e-01 -6.23708367e-01
4.47172850e-01 -6.49565041e-01 -8.83094490e-01 3.76885682e-01
-2.72258371e-01 3.84487629e-01 3.47824842e-01 -9.98758897e-02
-2.92353898e-01 -1.98917165e-02 -1.17814171e+00 6.31594658e-01
1.47353017e+00 -1.18054941e-01 -2.59878486e-01 2.34795198e-01
1.06726217e+00 -7.62775779e-01 -7.43389726e-01 4.69782561e-01
2.77843297e-01 -1.00041759e+00 1.36631644e+00 -1.34485051e-01
3.51254404e-01 -7.60248065e-01 -1.61412910e-01 -1.00715256e+00
-4.04753387e-01 -2.14595884e-01 -2.58969426e-01 8.50881815e-01
6.03714645e-01 -4.67119038e-01 1.10996604e+00 3.87135953e-01
-7.05136180e-01 -1.13635421e+00 -8.66636336e-01 -7.19334960e-01
-2.41450146e-01 -7.33639479e-01 1.06024456e+00 7.38252997e-01
-7.20935881e-01 -2.18192711e-01 2.40972504e-01 6.07290864e-01
6.94029391e-01 2.34860614e-01 9.33075666e-01 -1.18323684e+00
-7.61590526e-02 -6.30729258e-01 -3.67235243e-01 -1.68834603e+00
-2.16333523e-01 -9.57408309e-01 1.35056913e-01 -1.60187590e+00
-2.00471193e-01 -1.18285370e+00 -1.24314174e-01 5.30762553e-01
-2.32691213e-01 5.13519764e-01 6.45157397e-02 4.21554208e-01
-8.72867346e-01 2.76727647e-01 1.15196395e+00 -4.18690629e-02
-4.59779441e-01 2.00418979e-01 -5.54412842e-01 6.37318075e-01
8.45881402e-01 -4.60417330e-01 -2.98043430e-01 -4.68488336e-01
8.08914471e-03 -4.68411505e-01 7.90584266e-01 -1.24941325e+00
2.16543853e-01 -9.96750221e-02 6.42730534e-01 -1.33962595e+00
4.46459144e-01 -8.69673550e-01 8.44870508e-03 1.77382663e-01
-8.15093070e-02 -2.33096883e-01 2.63556182e-01 5.93121052e-01
7.04996362e-02 -1.13689788e-01 6.88063085e-01 -2.96450049e-01
-1.10597682e+00 8.39886487e-01 5.54628856e-02 7.48264641e-02
1.14663279e+00 -7.10561514e-01 -3.27453017e-01 2.79081285e-01
-5.31655729e-01 3.75440031e-01 6.36425197e-01 3.14965695e-01
6.75249577e-01 -1.13029945e+00 -6.24275208e-01 1.95717081e-01
1.97248310e-01 9.00558829e-01 2.85749227e-01 8.79958749e-01
-7.84683585e-01 1.56106710e-01 3.65142435e-01 -1.08511579e+00
-7.55759358e-01 2.19647616e-01 3.74633968e-01 1.85782820e-01
-9.41533506e-01 1.19213712e+00 3.20824295e-01 -5.30020595e-01
3.16841125e-01 -7.03382015e-01 2.51336396e-01 -9.51560885e-02
3.21626693e-01 3.69864732e-01 1.57439798e-01 -3.83052230e-01
-4.76247132e-01 4.81852353e-01 -1.06242232e-01 1.56900704e-01
1.42580211e+00 1.86226994e-01 1.31266832e-01 1.08672678e-01
1.03934729e+00 -4.82240804e-02 -1.71804047e+00 -1.62361026e-01
-1.03787445e-01 -6.48054183e-01 7.57764652e-02 -7.78173566e-01
-1.27987528e+00 4.68502074e-01 5.28325558e-01 2.36684196e-02
7.42605269e-01 5.50609291e-01 9.14689183e-01 1.61729708e-01
3.98716837e-01 -9.95954216e-01 -7.78148621e-02 4.27048981e-01
4.69723254e-01 -1.20262682e+00 -6.09142892e-02 -7.75729597e-01
-4.98087019e-01 9.79994476e-01 8.75428140e-01 -4.42972809e-01
7.93534696e-01 3.19318503e-01 -2.60589141e-02 -6.18474960e-01
-3.29777241e-01 -4.00917500e-01 5.10056317e-01 6.81636512e-01
8.01125541e-03 1.68271005e-01 1.58942565e-01 5.08206129e-01
-1.87450513e-01 -2.60668516e-01 -1.50141092e-02 7.64953375e-01
-5.76913118e-01 -6.69140637e-01 -4.25354809e-01 9.28964496e-01
-1.31429419e-01 -3.13398391e-01 -1.15458563e-01 8.09178293e-01
3.69221270e-01 5.67538977e-01 5.01075089e-01 -3.79179329e-01
5.10201514e-01 -2.88122356e-01 3.30901861e-01 -7.04500377e-01
-6.79677188e-01 9.96166244e-02 -9.78109539e-02 -9.75603580e-01
-3.57811838e-01 -5.36459446e-01 -1.48285604e+00 -2.34758481e-01
-4.57293957e-01 -1.74763560e-01 6.33778572e-01 5.96413732e-01
7.96567559e-01 5.14420390e-01 3.58994663e-01 -1.25379789e+00
-2.25371286e-01 -6.69678748e-01 -3.64465624e-01 1.72296256e-01
1.21111967e-01 -7.03711689e-01 -3.16030800e-01 -1.42466486e-01]
|
[8.00167179107666, -3.1968297958374023]
|
75cdcb07-b5a9-484d-91ec-e4e8b7d56432
|
autoformalization-with-large-language-models
|
2205.12615
| null |
https://arxiv.org/abs/2205.12615v1
|
https://arxiv.org/pdf/2205.12615v1.pdf
|
Autoformalization with Large Language Models
|
Autoformalization is the process of automatically translating from natural language mathematics to formal specifications and proofs. A successful autoformalization system could advance the fields of formal verification, program synthesis, and artificial intelligence. While the long-term goal of autoformalization seemed elusive for a long time, we show large language models provide new prospects towards this goal. We make the surprising observation that LLMs can correctly translate a significant portion ($25.3\%$) of mathematical competition problems perfectly to formal specifications in Isabelle/HOL. We demonstrate the usefulness of this process by improving a previously introduced neural theorem prover via training on these autoformalized theorems. Our methodology results in a new state-of-the-art result on the MiniF2F theorem proving benchmark, improving the proof rate from $29.6\%$ to $35.2\%$.
|
['Christian Szegedy', 'Mateja Jamnik', 'Charles Staats', 'Markus N. Rabe', 'Wenda Li', 'Albert Q. Jiang', 'Yuhuai Wu']
|
2022-05-25
| null | null | null | null |
['program-synthesis', 'automated-theorem-proving', 'automated-theorem-proving']
|
['computer-code', 'miscellaneous', 'reasoning']
|
[ 1.28274113e-01 6.46641910e-01 -1.66014984e-01 -3.52275640e-01
-8.75832558e-01 -7.11546600e-01 8.09665799e-01 1.33507550e-01
3.55265774e-02 9.34532225e-01 -2.83900172e-01 -1.25342059e+00
-1.05317466e-01 -9.82419312e-01 -1.35914004e+00 1.68311015e-01
-4.86192346e-01 4.07794178e-01 2.92205065e-02 -2.95252144e-01
2.08684560e-02 2.71454453e-01 -1.52970457e+00 3.99950445e-01
8.61791790e-01 7.46451735e-01 -4.07909781e-01 7.90612996e-01
-4.07850027e-01 1.06622839e+00 -4.25058037e-01 -4.21707839e-01
1.70426413e-01 -3.29335034e-01 -1.12228811e+00 -5.32459319e-01
6.81027591e-01 -2.32673317e-01 -2.99369171e-02 1.20178032e+00
-2.20410913e-01 -5.06821036e-01 1.89589858e-01 -1.94786811e+00
-5.14889479e-01 1.19290793e+00 -2.24943519e-01 -2.64577657e-01
2.83467054e-01 3.24481070e-01 1.30677724e+00 -3.59087169e-01
5.13445199e-01 1.37983739e+00 7.38686204e-01 7.52172410e-01
-1.39976776e+00 -8.70846927e-01 3.64346579e-02 9.96171534e-02
-1.42443776e+00 -5.88247836e-01 5.11316776e-01 -5.20952225e-01
1.50884867e+00 2.35257521e-01 4.11902279e-01 4.67257172e-01
4.45569485e-01 5.81319630e-01 9.89626110e-01 -6.06244385e-01
9.09994021e-02 2.09931150e-01 3.47744882e-01 1.27428091e+00
4.43065763e-01 1.38380408e-01 -8.39575455e-02 -2.88617641e-01
3.88886780e-01 -4.44134176e-01 -1.08684562e-01 -2.82655150e-01
-1.34742880e+00 8.12624633e-01 2.40787119e-02 2.62568474e-01
2.84316212e-01 8.52029502e-01 6.09952390e-01 6.08014405e-01
-1.24573953e-01 9.50714290e-01 -6.33759856e-01 -9.83315930e-02
-9.48198795e-01 6.47412956e-01 1.08505404e+00 8.90492380e-01
5.31134546e-01 3.25658947e-01 2.53726780e-01 -1.45043761e-01
2.17803568e-01 7.59663701e-01 -4.03508544e-02 -1.32920396e+00
4.07296598e-01 8.56853366e-01 2.05736101e-01 -7.21639216e-01
-2.04565525e-01 -4.14005637e-01 -6.29766941e-01 4.89293218e-01
5.16987979e-01 -1.95204660e-01 -4.04083341e-01 1.79267144e+00
-4.15044464e-02 -2.13854179e-01 2.33646259e-01 4.71031994e-01
4.82610881e-01 9.51293528e-01 -3.30158383e-01 -1.26739785e-01
1.05158317e+00 -7.47196853e-01 -2.79690474e-01 -7.28801787e-02
8.68017793e-01 -3.04353535e-01 9.44635808e-01 7.42969036e-01
-1.50184834e+00 -2.09249169e-01 -1.39931858e+00 1.95185229e-01
-2.57136524e-01 -1.40775621e-01 1.35876095e+00 3.87250334e-01
-1.28342783e+00 3.19679886e-01 -8.71773183e-01 1.71041489e-01
2.78737485e-01 6.68825626e-01 -3.52794379e-01 -1.77899450e-02
-1.02549577e+00 6.82349980e-01 3.69696796e-01 -6.03335574e-02
-1.23275995e+00 -1.01892292e+00 -1.19291973e+00 3.25629920e-01
6.50963247e-01 -7.25522339e-01 1.66713274e+00 -8.19340765e-01
-1.30335343e+00 6.15957975e-01 -1.26157418e-01 -1.08348143e+00
4.30330515e-01 1.88115373e-01 -4.10416007e-01 -2.51626104e-01
-1.62197906e-03 8.73567402e-01 4.04156387e-01 -9.01103497e-01
-6.88073993e-01 1.35401890e-01 8.60689044e-01 -7.50717521e-01
1.67323306e-01 7.43282586e-02 3.16466808e-01 -4.26906906e-02
-3.03848207e-01 -8.08772206e-01 1.71556957e-02 1.97045300e-02
-1.25164390e-01 -5.56365609e-01 3.58764470e-01 -3.21602583e-01
9.14895475e-01 -1.79830492e+00 1.00268185e-01 2.06559435e-01
6.71925128e-01 1.72131911e-01 -4.82065557e-03 4.05868053e-01
-1.91643998e-01 5.61098576e-01 -1.29501462e-01 2.65434533e-01
9.27917719e-01 -1.02547385e-01 -6.43902481e-01 3.42379183e-01
5.32367826e-01 1.35186040e+00 -9.05654907e-01 -4.50380296e-01
2.71343831e-02 -1.46569451e-02 -8.98145854e-01 -1.70014247e-01
-1.10909247e+00 -4.78328139e-01 -2.40279809e-01 3.95793229e-01
5.40467381e-01 -6.68298900e-01 4.18238133e-01 3.72743666e-01
-1.88444465e-01 7.27523446e-01 -8.75731885e-01 1.74856365e+00
-5.66359282e-01 8.97663951e-01 1.65036842e-01 -1.00643420e+00
4.96658415e-01 3.78179669e-01 2.07874849e-01 -7.10666776e-01
1.38487536e-02 2.48235554e-01 4.75962937e-01 -1.63236141e-01
1.87060326e-01 -4.58542645e-01 -3.12949330e-01 9.47502971e-01
-5.70001937e-02 -4.44148839e-01 5.54127872e-01 2.72354871e-01
1.18131745e+00 4.84720409e-01 2.57537574e-01 -3.36650997e-01
6.42612934e-01 4.75188255e-01 3.92300248e-01 7.34955549e-01
7.43401349e-02 -1.18358955e-01 9.21303630e-01 -8.81929874e-01
-1.06778097e+00 -9.90259051e-01 3.33298028e-01 5.89268446e-01
-4.36513335e-01 -9.39273894e-01 -8.42049599e-01 -8.49603891e-01
1.16684679e-02 9.90851045e-01 -3.01719844e-01 -3.72416556e-01
-6.46371543e-01 1.43563017e-01 1.28383613e+00 4.99937236e-01
4.06493545e-01 -1.09163499e+00 -6.97819591e-01 7.37260655e-02
-1.42992720e-01 -9.44188178e-01 4.24535498e-02 1.50899380e-01
-5.54748297e-01 -1.20740712e+00 1.01090059e-01 -6.73022747e-01
6.45751715e-01 -4.81430404e-02 1.38171327e+00 4.93929654e-01
-6.29269406e-02 -7.65628815e-02 4.55473661e-02 -5.09031773e-01
-1.18520105e+00 -2.77158106e-03 1.89700097e-01 -7.72159636e-01
1.86880901e-01 -5.38720548e-01 1.69641256e-01 -1.20043539e-01
-9.04616475e-01 3.68450463e-01 2.88215935e-01 6.58155918e-01
3.77487056e-02 4.46459651e-01 3.93715322e-01 -6.36278152e-01
5.02555311e-01 2.24630594e-01 -1.55066085e+00 4.12734598e-01
-9.05440271e-01 6.00907803e-01 1.21474314e+00 -4.46499251e-02
-3.74470472e-01 -2.34512582e-01 1.42489240e-01 -3.15001935e-01
-1.75549492e-01 8.08947444e-01 -6.73033819e-02 4.57839444e-02
7.38123596e-01 3.20172489e-01 1.17926911e-01 2.55350500e-01
4.19119298e-01 1.61474973e-01 7.08384037e-01 -1.18882740e+00
1.43362427e+00 6.72344565e-02 3.88107896e-01 -1.31288677e-01
-4.81607974e-01 4.09108847e-01 7.56237879e-02 3.43336523e-01
1.93650767e-01 -6.71949923e-01 -1.36871612e+00 -1.95206523e-01
-1.45801282e+00 -7.78122306e-01 -3.47843826e-01 1.19046800e-01
-5.62471092e-01 2.54944712e-01 -2.81605810e-01 -7.54958093e-01
-4.46310788e-01 -1.31267083e+00 8.65592897e-01 -2.15679467e-01
-6.78104162e-01 -7.90547669e-01 8.65067318e-02 9.30621549e-02
3.08940917e-01 2.46677220e-01 1.49474192e+00 -2.92673081e-01
-8.14801335e-01 -2.45792866e-01 -4.16890830e-01 6.70591533e-01
-3.51061337e-02 2.33708456e-01 -6.98082209e-01 -1.38570055e-01
-4.09402817e-01 -4.70923305e-01 3.06787997e-01 1.28971592e-01
8.98357928e-01 -6.88555777e-01 -6.85874447e-02 2.87272632e-01
1.29524124e+00 4.94981185e-02 5.49494267e-01 3.03124726e-01
7.32674003e-02 2.53610350e-02 5.50714768e-02 8.99877846e-02
4.95888799e-01 4.25553322e-01 4.36311096e-01 2.15787396e-01
-6.70013875e-02 -3.99107605e-01 4.79548007e-01 1.24761291e-01
1.67946965e-01 2.93688804e-01 -1.47531819e+00 4.98059481e-01
-1.60548115e+00 -1.02677774e+00 2.10184246e-01 1.83621478e+00
1.23524916e+00 6.28485441e-01 1.24393076e-01 4.32589442e-01
1.26127630e-01 -3.13725233e-01 -4.42063473e-02 -7.69232392e-01
2.37347528e-01 5.67491055e-01 1.22935630e-01 9.49764311e-01
-7.90834427e-01 9.26634192e-01 6.07878351e+00 5.42687178e-01
-1.32947922e+00 -3.91016245e-01 6.62291469e-03 5.96360825e-02
-6.36304796e-01 2.77622193e-01 -6.08813107e-01 2.82743964e-02
1.27813613e+00 -6.94595575e-01 1.04813528e+00 7.86864221e-01
3.13189104e-02 1.37685895e-01 -1.76840961e+00 5.19650877e-01
-6.72956780e-02 -1.95064342e+00 7.09266663e-02 -8.06653872e-02
5.88039339e-01 -4.71197739e-02 -1.23922199e-01 7.56329119e-01
6.61112607e-01 -1.43056035e+00 1.05021346e+00 2.25199103e-01
8.93646598e-01 -8.21755767e-01 5.93540668e-01 4.81220216e-01
-8.64940703e-01 -5.40814847e-02 1.05964519e-01 -6.63532674e-01
-2.57332832e-01 1.65636972e-01 -1.32057583e+00 5.84175587e-01
3.37850332e-01 4.61125284e-01 -6.23446941e-01 8.40263426e-01
-4.05789763e-01 6.79534376e-01 -4.32878375e-01 -2.55573362e-01
4.60608184e-01 2.75892764e-01 3.77498209e-01 1.10664403e+00
4.61085252e-02 -8.35422426e-02 -1.42966047e-01 1.51367664e+00
-1.64235786e-01 -6.89724326e-01 -6.83316588e-01 -7.45511353e-01
-1.07863873e-01 1.00031054e+00 -4.31992352e-01 -5.07185578e-01
-3.15885901e-01 1.69636816e-01 1.48519337e-01 2.50362307e-01
-1.24771261e+00 -6.43716693e-01 3.07446301e-01 1.80950791e-01
1.74430102e-01 -4.29329962e-01 -3.47001672e-01 -1.16083741e+00
4.18522269e-01 -1.39054644e+00 -3.27590331e-02 -1.03055871e+00
-6.07849836e-01 3.41384172e-01 2.99183339e-01 -7.63435960e-01
-6.39038444e-01 -6.98198378e-01 -3.72383386e-01 7.40742028e-01
-1.44985485e+00 -1.17281079e+00 1.55273333e-01 3.40581760e-02
1.03965439e-01 -2.92712659e-01 1.20169353e+00 2.56804645e-01
-1.08595327e-01 7.26299345e-01 -4.69302446e-01 1.59302488e-01
1.82884753e-01 -1.29081476e+00 6.08478367e-01 9.99252498e-01
2.21150994e-01 1.08097827e+00 9.29950476e-01 -1.84600011e-01
-2.12425590e+00 -1.06296527e+00 1.31744885e+00 -4.46696669e-01
1.12034917e+00 -5.66845775e-01 -3.72331738e-01 1.02215350e+00
3.34785044e-01 -3.07617933e-02 -2.38851197e-02 1.49230510e-01
-9.63095248e-01 -3.31709176e-01 -9.06241238e-01 8.19522798e-01
8.32301855e-01 -7.95238197e-01 -7.48132229e-01 2.89839476e-01
1.04830050e+00 -3.04415613e-01 -8.07180107e-01 3.21220458e-01
8.04730773e-01 -6.12702608e-01 7.22541034e-01 -9.74942327e-01
9.09309447e-01 -6.47960544e-01 -2.84742385e-01 -8.33212078e-01
-6.01265952e-02 -1.15078092e+00 -3.94176900e-01 9.55331028e-01
6.63964748e-01 -6.93150699e-01 6.73477650e-01 5.52151442e-01
-3.14446062e-01 -7.69966483e-01 -8.23760450e-01 -8.50669265e-01
5.62642753e-01 -7.90497541e-01 8.08117986e-01 8.06814790e-01
7.75682092e-01 5.05764365e-01 2.65764356e-01 4.98365276e-02
3.60477626e-01 5.57741404e-01 1.09405589e+00 -1.10299540e+00
-5.14608979e-01 -9.74827707e-01 -3.01477700e-01 -5.41567028e-01
8.37956786e-01 -1.33889723e+00 -9.36703607e-02 -1.52814353e+00
1.93136647e-01 -1.25831410e-01 -1.13669232e-01 1.07091796e+00
5.26453197e-01 9.36249197e-02 1.86295152e-01 -4.36782330e-01
-7.81535029e-01 1.21481903e-01 9.89893198e-01 -9.26784754e-01
3.59856367e-01 -1.52850732e-01 -9.97709215e-01 6.30413592e-01
6.77274525e-01 -2.74616450e-01 -3.72392714e-01 -5.82882822e-01
9.53894436e-01 3.38211775e-01 8.36154878e-01 -1.11230910e+00
3.83453742e-02 -5.62922180e-01 -4.27346647e-01 -2.86679119e-01
-1.71272486e-01 -7.53985822e-01 1.35022879e-01 7.04886258e-01
-5.09401679e-01 3.51726770e-01 6.84671521e-01 -1.02115072e-01
-2.40599349e-01 2.40371004e-02 4.32923466e-01 -1.97217450e-01
-5.19658148e-01 -7.99895152e-02 -3.17744017e-01 2.20806941e-01
7.51630127e-01 1.96011126e-01 -5.14256358e-01 -3.67592067e-01
-2.30067566e-01 3.45706791e-01 4.71535772e-01 1.98447555e-01
4.86743718e-01 -8.84441435e-01 -5.60853601e-01 1.82153255e-01
-4.92275432e-02 4.07400727e-02 -4.54138130e-01 8.65831971e-01
-6.00287318e-01 1.13962543e+00 -2.79725164e-01 -3.64038497e-01
-1.04135370e+00 5.50438881e-01 6.64313972e-01 -4.24475163e-01
-5.33845127e-01 4.74010497e-01 1.06121190e-01 -7.39625633e-01
3.12796205e-01 -1.12378514e+00 5.64673543e-01 -1.00613058e+00
5.52173972e-01 1.16933219e-01 8.77692178e-02 1.13911979e-01
-6.86871469e-01 1.17724106e-01 1.59071058e-01 -2.78938293e-01
1.42917192e+00 6.81191027e-01 -6.75259650e-01 1.92865193e-01
8.46499145e-01 3.94093432e-02 -6.14499450e-01 6.40992671e-02
-5.08307554e-02 3.30622435e-01 -1.34503767e-01 -1.11786366e+00
-4.23881412e-01 1.09610367e+00 -1.12414308e-01 4.02430385e-01
7.53036499e-01 -1.06257685e-01 6.81492209e-01 1.18460941e+00
6.86542571e-01 -4.42828834e-01 -3.83311450e-01 7.43314445e-01
7.17477500e-01 -9.81790602e-01 1.67881399e-01 4.26643901e-02
9.34123024e-02 1.16260481e+00 3.86986345e-01 -2.95617968e-01
3.25278901e-02 8.19340467e-01 -3.49493325e-01 -1.57708317e-01
-1.08280253e+00 4.42067176e-01 1.13835990e-01 3.77902865e-01
5.27753234e-01 1.00433886e-01 4.79630679e-02 7.22493231e-01
-5.89134455e-01 6.14500821e-01 6.72670066e-01 8.93796027e-01
-2.09593222e-01 -1.26790571e+00 -2.12143421e-01 2.01289970e-02
-5.80223918e-01 -1.65231004e-01 -4.86161739e-01 1.32354271e+00
-7.05896989e-02 7.58563161e-01 -2.81990230e-01 -2.26011813e-01
-6.40206859e-02 4.25076514e-01 1.14091396e+00 -5.66664636e-01
-4.97227877e-01 -3.57860386e-01 5.48378229e-01 -5.39282918e-01
-1.03529751e-01 -3.52337778e-01 -1.65074635e+00 -8.16752136e-01
3.34037654e-02 7.50848174e-01 5.69223702e-01 1.13789475e+00
1.35073826e-01 5.47879159e-01 9.51713398e-02 -4.21251446e-01
-9.30205286e-01 -4.85923678e-01 -6.06790818e-02 -1.82402447e-01
7.27587044e-01 -1.75538182e-01 -2.02388585e-01 2.94305325e-01]
|
[8.913480758666992, 7.07528829574585]
|
7517b96a-4498-45c9-8e88-069bfe182a7e
|
learning-selective-communication-for-multi
|
2109.05413
| null |
https://arxiv.org/abs/2109.05413v2
|
https://arxiv.org/pdf/2109.05413v2.pdf
|
Learning Selective Communication for Multi-Agent Path Finding
|
Learning communication via deep reinforcement learning (RL) or imitation learning (IL) has recently been shown to be an effective way to solve Multi-Agent Path Finding (MAPF). However, existing communication based MAPF solvers focus on broadcast communication, where an agent broadcasts its message to all other or predefined agents. It is not only impractical but also leads to redundant information that could even impair the multi-agent cooperation. A succinct communication scheme should learn which information is relevant and influential to each agent's decision making process. To address this problem, we consider a request-reply scenario and propose Decision Causal Communication (DCC), a simple yet efficient model to enable agents to select neighbors to conduct communication during both training and execution. Specifically, a neighbor is determined as relevant and influential only when the presence of this neighbor causes the decision adjustment on the central agent. This judgment is learned only based on agent's local observation and thus suitable for decentralized execution to handle large scale problems. Empirical evaluation in obstacle-rich environment indicates the high success rate with low communication overhead of our method.
|
['Jia Pan', 'Yudong Luo', 'Ziyuan Ma']
|
2021-09-12
| null | null | null | null |
['multi-agent-path-finding']
|
['playing-games']
|
[-6.48525655e-02 3.43140960e-01 -3.87067348e-01 -6.98426589e-02
-3.65962356e-01 -4.22424167e-01 7.34849453e-01 5.35935581e-01
-6.31474435e-01 1.32018793e+00 1.05390340e-01 -2.26361319e-01
-4.58624661e-01 -1.14865077e+00 -4.28863585e-01 -1.03607154e+00
-5.40608108e-01 9.13526356e-01 3.76173824e-01 -3.75494182e-01
3.07384908e-01 4.06851828e-01 -1.06227148e+00 -2.19367042e-01
8.51334989e-01 5.52653253e-01 5.12963235e-01 6.42717242e-01
5.94903007e-02 1.03602064e+00 -1.03307188e+00 3.01199317e-01
2.26617396e-01 -3.62699926e-01 -1.13669384e+00 -1.42138600e-01
-6.43617392e-01 -7.34452963e-01 -8.20438489e-02 6.12062216e-01
5.84310591e-01 2.07060561e-01 4.99786347e-01 -1.46901929e+00
8.89756903e-02 9.37102854e-01 -6.13018215e-01 1.99406058e-01
6.26499355e-01 1.75974622e-01 7.01524377e-01 1.18245035e-02
7.97442079e-01 1.16896451e+00 1.01069063e-01 3.85393739e-01
-8.87741089e-01 -6.38793528e-01 5.71311057e-01 4.49841589e-01
-9.70892131e-01 -9.33868065e-02 7.07132280e-01 -7.42184743e-02
7.78067589e-01 4.50327247e-03 9.19901907e-01 8.88654411e-01
5.04718244e-01 6.13213718e-01 1.10620844e+00 -2.61791497e-01
6.28609955e-01 -1.07304119e-01 -3.07232469e-01 5.69031715e-01
1.23374209e-01 3.31044137e-01 -6.16191864e-01 -3.03572327e-01
7.60314822e-01 -4.55805331e-01 -1.89361140e-01 -1.15001045e-01
-1.50692987e+00 1.05691266e+00 7.27348924e-01 3.38115031e-03
-9.20471132e-01 5.18597484e-01 2.14041904e-01 8.39816034e-01
-9.33514535e-03 4.88673806e-01 -3.54152828e-01 -1.53636888e-01
-3.08160912e-02 3.08204710e-01 9.87807810e-01 4.14040595e-01
1.04218984e+00 -6.99184164e-02 -7.28887767e-02 4.22044516e-01
4.22042638e-01 4.34521824e-01 -1.04320072e-01 -1.17605138e+00
3.63670915e-01 6.83327734e-01 2.93849975e-01 -1.26495528e+00
-7.89285123e-01 -3.27579230e-01 -9.07531202e-01 6.13297760e-01
8.28540996e-02 -7.70231366e-01 -2.89676696e-01 1.71218729e+00
8.81660521e-01 2.32477263e-01 4.04369593e-01 1.04471529e+00
6.38696134e-01 9.40528452e-01 2.56341528e-02 -5.04597604e-01
8.15624297e-01 -9.38860834e-01 -4.44943160e-01 -8.50160420e-02
8.24332893e-01 -5.23382366e-01 2.96642810e-01 2.58212268e-01
-7.93895245e-01 -1.76621694e-02 -1.00897098e+00 6.61208689e-01
-8.77959356e-02 -2.35816091e-01 8.63134205e-01 1.70998931e-01
-1.18404245e+00 3.24284315e-01 -6.26773834e-01 -4.93797123e-01
9.12785754e-02 8.78335476e-01 -2.64861166e-01 5.39174639e-02
-1.26624835e+00 7.90384889e-01 2.59051472e-01 -2.25815643e-02
-1.57972372e+00 -1.55834377e-01 -3.72019410e-01 4.21622954e-03
9.12705719e-01 -8.86752903e-01 1.21047485e+00 -9.17808473e-01
-2.04642224e+00 -7.42149726e-02 5.60691394e-02 -5.67921042e-01
4.52514291e-01 2.72308201e-01 9.14485008e-03 2.45626539e-01
3.34724903e-01 8.05431426e-01 7.07762301e-01 -1.39582992e+00
-1.11885905e+00 -5.45101911e-02 6.81568623e-01 7.45877266e-01
-8.87227729e-02 -1.86602786e-01 -1.52855262e-01 -1.33219570e-01
-1.18670322e-01 -9.99458194e-01 -6.65579796e-01 -2.93249995e-01
-3.99198830e-01 -5.92423260e-01 9.13008094e-01 -6.11417741e-02
8.30785036e-01 -1.55036747e+00 3.24020654e-01 5.14269769e-01
2.65943855e-01 -6.42589107e-02 -4.39971507e-01 9.39399302e-01
6.71012998e-01 -1.42793596e-01 1.59380466e-01 8.55281055e-02
-2.02117696e-01 4.29173142e-01 1.01049542e-01 5.10455728e-01
-2.03599006e-01 4.82575655e-01 -1.05085254e+00 -3.46544415e-01
1.93582401e-01 2.14562461e-01 -6.72660768e-01 2.40779623e-01
-2.48731092e-01 9.51878726e-01 -1.04098356e+00 2.23707065e-01
4.38298941e-01 -3.07049274e-01 4.06646520e-01 3.91941696e-01
-2.26269364e-01 4.51538354e-01 -1.45484650e+00 1.20148063e+00
-4.86906648e-01 4.63838607e-01 5.48894286e-01 -1.15713811e+00
7.70921886e-01 3.56913894e-01 7.82561779e-01 -7.56460786e-01
-2.01174133e-02 7.18109682e-02 2.80423164e-01 -3.78375232e-01
2.50481199e-02 4.44194794e-01 -3.50432061e-02 8.93673241e-01
-5.55831611e-01 9.48016793e-02 2.42472142e-01 4.52764153e-01
1.55727816e+00 -2.11343795e-01 2.80863941e-01 -1.73217222e-01
7.22790360e-01 3.56898844e-01 8.14427614e-01 9.46812987e-01
-2.51103103e-01 -4.58347231e-01 6.06013477e-01 -5.72151721e-01
-3.52000296e-01 -6.66585445e-01 7.34738648e-01 1.09729028e+00
7.88731813e-01 -3.56288016e-01 -4.40770417e-01 -7.02097237e-01
-1.03667244e-01 5.11608422e-01 -4.97294903e-01 -6.70674890e-02
-8.49431098e-01 -5.39168596e-01 -3.71823236e-02 -2.74957474e-02
9.39834177e-01 -1.20531750e+00 -8.52660239e-01 6.67897224e-01
-1.90027624e-01 -8.41792226e-01 -2.91413903e-01 1.53517425e-01
-3.88894171e-01 -1.31398785e+00 -2.75426686e-01 -8.13202381e-01
9.49133575e-01 5.60092330e-01 6.34093404e-01 4.12394375e-01
2.77464032e-01 4.72540289e-01 -5.16188979e-01 -2.26130411e-01
-7.05247283e-01 3.07985216e-01 1.65940732e-01 -1.66017741e-01
-1.35856852e-01 -4.40905333e-01 -8.78843963e-01 7.56209373e-01
-4.56364781e-01 1.35109395e-01 8.21570814e-01 6.31798923e-01
2.53801048e-01 4.88571197e-01 9.35759246e-01 -5.38992047e-01
9.78041232e-01 -6.03001654e-01 -7.65194058e-01 7.76330456e-02
-5.84413052e-01 -8.23709071e-02 8.16818774e-01 -3.06643158e-01
-9.49378252e-01 -2.64655910e-02 2.29881704e-01 3.88685465e-01
-1.70930654e-01 5.97231030e-01 2.40182787e-01 -4.15187240e-01
3.73295635e-01 5.47335260e-02 1.01354189e-01 1.72875360e-01
-8.15978944e-02 5.02752364e-01 -2.12059334e-01 -5.63387454e-01
6.80953264e-01 4.59921479e-01 2.92193115e-01 -6.02950156e-01
-2.98459046e-02 -1.86084151e-01 -2.66773641e-01 -5.19017875e-01
5.13378739e-01 -8.92005563e-01 -1.53046346e+00 2.01163307e-01
-1.31608760e+00 -6.20554745e-01 2.66485244e-01 5.37996769e-01
-2.91663080e-01 5.54585792e-02 -6.21694088e-01 -5.86761177e-01
-2.01493889e-01 -1.39904702e+00 4.96520996e-01 3.05595875e-01
-1.18499897e-01 -1.04183471e+00 1.06555924e-01 2.69357234e-01
6.86959147e-01 1.64220288e-01 6.94483042e-01 -3.67141247e-01
-1.03510642e+00 1.22137807e-01 -8.93245190e-02 -4.83294189e-01
1.22018278e-01 -7.08896220e-02 -2.40431488e-01 -5.92298627e-01
-4.64123487e-01 -2.99909979e-01 4.06268865e-01 4.67261970e-01
4.69493300e-01 -5.88973641e-01 -7.02154338e-01 -3.35651860e-02
1.14741492e+00 4.38858867e-01 2.03374952e-01 6.43270850e-01
2.28606552e-01 6.77443087e-01 6.54258549e-01 6.47787750e-01
1.01713872e+00 7.41905928e-01 8.24141562e-01 -4.82915342e-02
-1.23847544e-01 -1.64392576e-01 6.53588593e-01 6.25240326e-01
-1.62138820e-01 -4.57722902e-01 -8.20740700e-01 2.38162577e-01
-2.08513761e+00 -6.33601129e-01 -3.98063585e-02 1.95955646e+00
5.75170696e-01 1.04443565e-01 1.53428748e-01 -6.86257929e-02
5.79136491e-01 -7.64580444e-02 -5.88362396e-01 -2.53287405e-01
-7.81358220e-03 -5.40613711e-01 5.63096642e-01 8.09639871e-01
-6.58139408e-01 1.00629103e+00 5.69583082e+00 6.29249871e-01
-1.01280868e+00 8.71174708e-02 4.68633890e-01 1.78617090e-01
-2.24963099e-01 2.69391298e-01 -6.58224821e-01 3.00431222e-01
7.31247902e-01 -1.84657350e-01 7.58005917e-01 4.52746302e-01
8.25802386e-01 -7.83036411e-01 -8.49142432e-01 5.57505369e-01
-4.64000136e-01 -1.58990538e+00 9.72883299e-06 2.58074194e-01
9.06877160e-01 5.83775751e-02 -4.22518760e-01 1.89625293e-01
1.00872409e+00 -7.33394146e-01 1.83272600e-01 -3.62038426e-02
1.44033507e-01 -8.86364937e-01 8.01803708e-01 6.98726952e-01
-1.21547186e+00 -5.66452384e-01 -2.76979297e-01 -5.25398314e-01
1.76880196e-01 2.03732088e-01 -1.23656225e+00 5.62484860e-01
6.10249162e-01 5.99840462e-01 -7.99563248e-03 9.57283437e-01
-5.28791487e-01 4.76485431e-01 -4.91720885e-01 -5.44566751e-01
6.03113949e-01 -2.08932042e-01 8.39294493e-01 4.74172145e-01
5.41919097e-02 2.95634985e-01 7.89149463e-01 4.02973890e-01
1.10299893e-01 1.80083975e-01 -4.72021729e-01 3.54890913e-01
7.51544833e-01 1.02913082e+00 -8.18455100e-01 -1.64630562e-01
-2.63008088e-01 7.27859318e-01 5.52693009e-01 5.93218565e-01
-5.00100791e-01 -1.39781028e-01 6.10431612e-01 -9.53232124e-02
3.75278816e-02 -3.04946482e-01 6.33990988e-02 -5.16322076e-01
-3.28658015e-01 -7.94332206e-01 3.59965175e-01 -2.81775504e-01
-7.15805829e-01 4.49191481e-01 -1.75803304e-01 -9.85280514e-01
-3.84423763e-01 9.42533836e-02 -9.61964667e-01 2.37188727e-01
-1.85618007e+00 -8.08402658e-01 -2.54618138e-01 8.30987155e-01
4.81069952e-01 -4.49669063e-01 8.84051681e-01 4.42071967e-02
-6.72332883e-01 4.95231375e-02 -8.75949636e-02 -1.18196577e-01
3.96658152e-01 -9.00116503e-01 -3.32734674e-01 4.80263382e-01
-3.85825872e-01 1.71171412e-01 7.52007246e-01 -6.86164975e-01
-1.81862068e+00 -7.79541373e-01 4.51436818e-01 1.05729572e-01
4.66644317e-01 1.27571449e-01 -3.95257562e-01 3.52953196e-01
4.02669728e-01 -5.11877000e-01 3.67272884e-01 -1.31197795e-01
4.53707039e-01 -4.59589720e-01 -1.05565786e+00 7.91236103e-01
6.65741682e-01 2.45631337e-01 1.87768377e-02 5.33582032e-01
6.26419187e-01 -1.18306711e-01 -5.37793875e-01 2.51362741e-01
2.59950962e-02 -7.86454022e-01 5.75114727e-01 -1.57469943e-01
1.08867604e-02 -5.98704994e-01 1.49199843e-01 -1.78496027e+00
-1.97498292e-01 -1.14273357e+00 1.79926783e-01 7.85841644e-01
4.61243719e-01 -1.04577029e+00 9.22640502e-01 2.00043827e-01
2.03703806e-01 -6.05381608e-01 -1.18379378e+00 -4.16476369e-01
-1.22477010e-01 8.17226991e-02 6.65520728e-01 8.06417406e-01
2.26279497e-01 5.16896725e-01 -3.79036784e-01 6.60938203e-01
6.54939771e-01 1.70829341e-01 1.30572546e+00 -1.25901091e+00
-2.73280114e-01 -2.93713212e-01 6.03753403e-02 -1.29623497e+00
2.48744473e-01 -5.51741958e-01 4.19206508e-02 -2.03191805e+00
-6.27571996e-03 -1.03198087e+00 -5.26020071e-03 4.91006017e-01
3.29496771e-01 -4.48565751e-01 1.71411201e-01 4.54649270e-01
-8.51610363e-01 5.63014388e-01 1.68327022e+00 -2.18987882e-01
-4.49885219e-01 3.90488178e-01 -4.19094354e-01 5.10324657e-01
1.19056976e+00 -4.67832297e-01 -6.00551009e-01 -4.73417908e-01
3.96855235e-01 7.38179922e-01 2.23315746e-01 -8.42021048e-01
9.63482738e-01 -6.52776957e-01 -7.19519109e-02 -2.72791773e-01
2.57670105e-01 -8.65715027e-01 6.70206696e-02 1.10036922e+00
-5.04354119e-01 1.01761140e-01 -2.62337774e-01 8.98756087e-01
-1.04447559e-01 8.59561563e-02 4.60917532e-01 -1.55525133e-01
-9.05465543e-01 2.64752001e-01 -9.85684454e-01 -4.69049424e-01
1.49968350e+00 7.91397169e-02 -4.72583294e-01 -1.06805527e+00
-3.07043672e-01 9.96124804e-01 1.19070657e-01 1.22767754e-01
8.08164120e-01 -8.85841250e-01 -9.40471292e-01 -3.68835554e-02
-2.52042532e-01 2.68760864e-02 -7.66976550e-03 9.89340901e-01
-4.39984769e-01 3.98767501e-01 -1.69105172e-01 -3.76508623e-01
-1.09015977e+00 1.03669859e-01 3.41466725e-01 -2.40033507e-01
-5.94483972e-01 5.19785523e-01 -2.53583603e-02 -4.60669607e-01
4.45825845e-01 3.88105176e-02 -5.47919869e-01 -1.72567651e-01
3.76460969e-01 5.77890813e-01 -4.38606262e-01 -3.44734639e-01
-3.62072825e-01 4.31210130e-01 -1.28116250e-01 -1.17427401e-01
1.23076808e+00 -5.42226493e-01 -2.96715915e-01 -1.81618497e-01
7.12192178e-01 -3.18418264e-01 -1.23289561e+00 -1.91628739e-01
-2.63636589e-01 -3.91683757e-01 3.80152792e-01 -8.58121753e-01
-1.09211648e+00 1.77847043e-01 1.23319976e-01 3.17325830e-01
8.47336650e-01 -8.14536586e-02 6.25730634e-01 8.94777179e-01
9.37485576e-01 -1.31890070e+00 3.19559097e-01 5.21030188e-01
7.79349029e-01 -1.50588417e+00 7.70451650e-02 -5.47275543e-01
-6.33503497e-01 1.17139113e+00 9.88347709e-01 4.07750234e-02
6.84249043e-01 1.48340434e-01 -5.40160388e-03 -2.36275971e-01
-1.17726576e+00 -1.36204571e-01 -6.78473294e-01 7.80537963e-01
-3.60119522e-01 1.54632360e-01 -3.36052507e-01 -2.50187725e-01
9.77588370e-02 -1.17554508e-01 8.34159791e-01 9.48024452e-01
-7.94147313e-01 -1.16064298e+00 -3.04290831e-01 2.92787850e-01
1.07704699e-01 4.90478307e-01 -3.38704228e-01 7.09807396e-01
-1.41234651e-01 1.52219021e+00 -7.26995543e-02 -1.09680749e-01
-3.72842960e-02 -9.83799040e-01 -8.30061510e-02 -3.96522880e-01
-5.64937770e-01 2.95221768e-02 2.39420563e-01 -6.79492354e-01
-6.89294636e-01 -3.24355781e-01 -1.74058115e+00 -6.86414003e-01
-1.22447014e-01 4.69969183e-01 6.52261198e-01 9.39635336e-01
5.23112416e-01 5.12624204e-01 1.07509577e+00 -7.39079475e-01
-3.06464165e-01 -5.66496968e-01 -1.73650756e-01 -2.41997421e-01
5.30946612e-01 -8.40443015e-01 -3.67129803e-01 -6.74289346e-01]
|
[3.786374568939209, 1.9764885902404785]
|
66bd98dd-6058-4ee9-8d21-926568f22c0f
|
graph-based-multi-view-fusion-and-local
|
2207.04081
| null |
https://arxiv.org/abs/2207.04081v1
|
https://arxiv.org/pdf/2207.04081v1.pdf
|
Graph-based Multi-View Fusion and Local Adaptation: Mitigating Within-Household Confusability for Speaker Identification
|
Speaker identification (SID) in the household scenario (e.g., for smart speakers) is an important but challenging problem due to limited number of labeled (enrollment) utterances, confusable voices, and demographic imbalances. Conventional speaker recognition systems generalize from a large random sample of speakers, causing the recognition to underperform for households drawn from specific cohorts or otherwise exhibiting high confusability. In this work, we propose a graph-based semi-supervised learning approach to improve household-level SID accuracy and robustness with locally adapted graph normalization and multi-signal fusion with multi-view graphs. Unlike other work on household SID, fairness, and signal fusion, this work focuses on speaker label inference (scoring) and provides a simple solution to realize household-specific adaptation and multi-signal fusion without tuning the embeddings or training a fusion network. Experiments on the VoxCeleb dataset demonstrate that our approach consistently improves the performance across households with different customer cohorts and degrees of confusability.
|
['Andreas Stolcke', 'Venkatesh Ravichandran', 'Yixiong Meng', 'Long Chen']
|
2022-07-08
| null | null | null | null |
['speaker-identification']
|
['speech']
|
[ 1.57086715e-01 9.41658467e-02 -1.44498155e-01 -1.03409612e+00
-1.09805655e+00 -5.57957530e-01 1.83326378e-01 1.32880256e-01
2.95154989e-01 3.77039373e-01 6.60650551e-01 -1.31924655e-02
1.53372124e-01 -4.81854796e-01 -2.72454500e-01 -6.49081826e-01
2.43298598e-02 5.82659006e-01 -2.50371695e-01 -1.52373925e-01
-3.45416844e-01 3.19461286e-01 -1.39006114e+00 1.36058226e-01
1.01566446e+00 7.68095851e-01 -4.94370610e-01 8.65478218e-01
-4.85998355e-02 6.79295242e-01 -8.28493893e-01 -1.04965317e+00
1.11177698e-01 -6.93611085e-01 -4.32632059e-01 3.24749082e-01
8.75347733e-01 -2.19399333e-01 7.28908135e-03 1.13991463e+00
1.19536841e+00 3.42536345e-02 8.10008943e-01 -1.68542433e+00
-6.83363378e-01 1.17935979e+00 -7.60886133e-01 -9.95457992e-02
6.70061350e-01 -2.00852677e-01 9.61691022e-01 -7.27396548e-01
1.16978943e-01 1.58978105e+00 9.94906366e-01 6.75051451e-01
-1.59510016e+00 -1.05770230e+00 2.71867603e-01 6.80896640e-02
-1.84664690e+00 -9.57351685e-01 9.59311664e-01 -3.08266282e-01
5.14458537e-01 6.23723090e-01 2.20014319e-01 1.24544775e+00
-6.52024686e-01 5.80438375e-01 7.13849485e-01 -2.57943541e-01
3.93722922e-01 5.31199813e-01 1.77009672e-01 4.13104624e-01
2.72376299e-01 -5.01999497e-01 -7.35163808e-01 -3.95691663e-01
1.71729699e-01 -1.20197698e-01 -3.24665487e-01 -4.89598423e-01
-1.04647863e+00 1.03119016e+00 -2.42597498e-02 9.40478519e-02
-3.48408699e-01 -3.37150037e-01 6.13848150e-01 2.43654996e-01
4.79732692e-01 -6.75821379e-02 -1.42567426e-01 1.94304109e-01
-1.40998018e+00 -8.58771950e-02 1.20298839e+00 1.08524418e+00
5.74318111e-01 4.46524888e-01 -1.59614876e-01 1.08994687e+00
3.16289485e-01 8.22672904e-01 4.51680183e-01 -8.02712023e-01
6.15750790e-01 4.75016057e-01 -3.18918079e-01 -1.06723118e+00
-4.12337363e-01 -5.05568445e-01 -9.71405804e-01 -2.50243068e-01
3.28127205e-01 -4.73498255e-01 -6.45996094e-01 1.98901618e+00
4.46679324e-01 2.44262382e-01 1.35942191e-01 7.35272288e-01
1.06224191e+00 3.02039564e-01 1.23838149e-01 -2.54119873e-01
1.20784199e+00 -8.37116897e-01 -9.86976862e-01 -4.04641330e-02
4.80587900e-01 -6.37544870e-01 9.09341931e-01 1.97925746e-01
-7.12981880e-01 -2.91196108e-01 -1.03073430e+00 3.50093722e-01
-3.90656978e-01 -7.64352530e-02 4.71615076e-01 1.64656508e+00
-1.27350104e+00 6.33936748e-03 -2.53254086e-01 -8.04102600e-01
4.41247255e-01 6.94607735e-01 -4.35202539e-01 -9.14223194e-02
-1.12831342e+00 6.25547647e-01 -6.84119835e-02 -2.91300118e-01
-5.18119991e-01 -7.89269269e-01 -1.04434514e+00 7.99885392e-02
1.00551732e-01 -5.96032441e-01 9.61359322e-01 -1.04331744e+00
-1.61217034e+00 7.35997617e-01 -2.32283428e-01 -3.94545883e-01
5.11630952e-01 3.61630589e-01 -9.63630080e-01 -9.43303704e-02
8.59229341e-02 3.53772908e-01 9.86596942e-01 -1.28764677e+00
-3.49904954e-01 -6.78381741e-01 -3.75696212e-01 3.53074163e-01
-6.31888211e-01 2.63555348e-01 7.99500346e-02 -4.93391484e-01
9.57388431e-02 -5.58452785e-01 2.02123985e-01 -7.97389030e-01
-6.29036665e-01 -9.44992006e-02 8.00669789e-01 -1.12839735e+00
1.27334166e+00 -2.24609780e+00 4.69036773e-02 4.07690942e-01
1.33917168e-01 -1.56862624e-02 -2.92909034e-02 2.48014778e-01
-1.88491508e-01 2.28678077e-01 -9.78964046e-02 -7.54608154e-01
4.58554208e-01 -1.17744155e-01 1.12816185e-01 6.33782923e-01
-1.54547319e-01 6.30835354e-01 -6.08585835e-01 -5.53486347e-01
3.05749267e-01 5.88583112e-01 -6.94134176e-01 3.40786725e-01
5.15763044e-01 2.20122829e-01 1.06493846e-01 1.04252255e+00
9.73055303e-01 9.15159732e-02 4.51814502e-01 -4.36309129e-01
4.51682121e-01 3.80082391e-02 -1.64906383e+00 1.35706568e+00
-3.38099092e-01 1.39010683e-01 7.16716707e-01 -1.01747096e+00
9.47382867e-01 4.15762275e-01 3.90480965e-01 -1.31022155e-01
1.92830309e-01 -3.53451213e-03 -2.41869822e-01 -3.59040767e-01
3.32398236e-01 -3.00273091e-01 -4.28882569e-01 4.91009146e-01
4.55588758e-01 -5.70299439e-02 -1.15165129e-01 3.36502045e-01
7.41307259e-01 -6.41453683e-01 3.39364380e-01 -1.07443668e-01
6.48868382e-01 -6.94128454e-01 6.67689979e-01 4.59610760e-01
-5.40502548e-01 9.05895352e-01 2.57062256e-01 2.66361147e-01
-6.36704087e-01 -1.24687731e+00 1.61680534e-01 1.63223934e+00
-2.36490160e-01 -1.83003277e-01 -1.05631733e+00 -8.07660878e-01
1.58656001e-01 1.02288175e+00 -3.85041833e-01 -2.55313516e-01
-3.56093347e-01 -8.76601338e-01 8.27533245e-01 4.93663818e-01
2.58570403e-01 -3.76809716e-01 3.06882471e-01 2.26191804e-01
-3.91674459e-01 -1.21081638e+00 -1.10202777e+00 3.71532887e-02
-2.10302830e-01 -6.95089936e-01 -6.82556450e-01 -8.28575432e-01
5.57084799e-01 1.95703179e-01 9.21594322e-01 -6.04426920e-01
1.32539466e-01 7.54861176e-01 -2.17435062e-01 -4.17796910e-01
-6.41710281e-01 2.29291797e-01 4.52173322e-01 6.79290712e-01
6.61911964e-01 -6.60686314e-01 -3.23045850e-01 5.54594815e-01
-4.42688525e-01 -3.98397803e-01 4.32340726e-02 7.54057884e-01
-7.22157285e-02 -6.46040635e-03 1.26479769e+00 -7.29791224e-01
7.16443717e-01 -4.69620407e-01 -1.23805232e-01 4.87382084e-01
-5.65667450e-01 -3.73948365e-01 4.45837379e-01 -5.96934915e-01
-1.05486000e+00 6.64658919e-02 -5.90683892e-02 -1.89273566e-01
-4.03125107e-01 -2.14876942e-02 -9.10896659e-01 -1.55467659e-01
4.77977097e-01 1.38107622e-02 1.63530871e-01 -3.51261228e-01
6.88904107e-01 1.29031897e+00 5.24166048e-01 -2.09731072e-01
6.58297777e-01 1.52440995e-01 -6.50775075e-01 -7.51691878e-01
-2.96423644e-01 -6.11430407e-01 -7.52681375e-01 -3.01168084e-01
8.24711800e-01 -1.37939358e+00 -8.91195297e-01 5.02784193e-01
-6.55115366e-01 4.61865542e-03 -2.54172623e-01 4.69302863e-01
-2.72316247e-01 3.95339966e-01 -4.34062064e-01 -1.06254137e+00
-5.61961114e-01 -9.81366813e-01 1.19833767e+00 1.74360141e-01
-6.24727488e-01 -9.54325497e-01 -1.87845096e-01 8.81342173e-01
5.32808840e-01 1.95643127e-01 6.72686517e-01 -1.21554589e+00
6.87220395e-02 -3.73059750e-01 -7.16312900e-02 4.96321857e-01
5.64181864e-01 -1.55229911e-01 -1.29708159e+00 -4.91616189e-01
-1.32310465e-01 -2.28329018e-01 4.38844770e-01 3.94230068e-01
5.84358811e-01 -5.26979029e-01 -1.07226901e-01 5.26879013e-01
9.53899026e-01 -7.94552341e-02 -6.20981082e-02 -3.57196391e-01
9.99288142e-01 9.57552612e-01 -8.89277458e-03 5.25530994e-01
9.57127273e-01 6.22289240e-01 6.77129030e-02 -8.57742950e-02
-3.20804119e-01 -3.46882135e-01 7.52048373e-01 1.18340659e+00
4.31191295e-01 -3.59049529e-01 -6.96499169e-01 5.95632374e-01
-1.60053420e+00 -1.07309258e+00 1.69942185e-01 2.21868777e+00
6.76684737e-01 -4.53514785e-01 7.40225136e-01 2.92728603e-01
1.39018548e+00 2.05627248e-01 -5.99773467e-01 -5.87152183e-01
-4.06309366e-01 -1.43755734e-01 7.08962440e-01 5.66781461e-01
-1.08174384e+00 6.13697886e-01 6.56375170e+00 6.38022065e-01
-1.12909949e+00 4.26416218e-01 7.31425583e-01 -3.24728757e-01
-3.13045055e-01 -6.98494017e-01 -8.31074893e-01 4.29191351e-01
1.17474365e+00 -3.31055492e-01 6.82369173e-01 7.24079072e-01
4.41640019e-02 4.34251875e-01 -1.23169208e+00 1.31727195e+00
8.79076242e-01 -7.00840116e-01 -1.65197238e-01 -3.23031917e-02
7.32611895e-01 -1.09220907e-01 2.16476500e-01 3.81959021e-01
6.70251071e-01 -9.34295952e-01 7.84529030e-01 -1.11595556e-01
8.33147287e-01 -8.26355934e-01 7.53467917e-01 8.40566456e-02
-1.17302489e+00 -2.42544502e-01 1.16770931e-01 3.78311843e-01
3.02516162e-01 5.08704364e-01 -1.04934525e+00 6.15444541e-01
5.97826004e-01 2.69534171e-01 -6.58833206e-01 5.50203145e-01
1.68162763e-01 8.62833917e-01 -4.65498716e-01 5.60803786e-02
-5.44749856e-01 -5.04094642e-03 4.15532202e-01 1.47623098e+00
4.76635277e-01 -1.86524749e-01 1.78622410e-01 4.61228848e-01
-4.54888999e-01 4.31970000e-01 -6.51664138e-01 1.97187439e-01
7.09620774e-01 1.33802128e+00 -3.29332858e-01 -4.31351244e-01
-4.69663709e-01 9.83788073e-01 5.92435114e-02 5.81769943e-01
-7.83596396e-01 -1.45953134e-01 6.23458028e-01 7.19023049e-02
3.19727391e-01 1.78147748e-01 -5.39048076e-01 -1.35445154e+00
-1.79347873e-01 -1.07882965e+00 6.88413501e-01 -1.13900229e-01
-1.52310801e+00 3.00299495e-01 -1.38988510e-01 -8.02738547e-01
-4.95805681e-01 5.90398610e-02 -7.44515777e-01 7.25570679e-01
-1.20838809e+00 -1.46924138e+00 -2.08602980e-01 7.49604821e-01
3.40612113e-01 -4.26873088e-01 9.15252864e-01 7.35400319e-01
-8.07068348e-01 1.45764410e+00 2.57587600e-02 1.39907509e-01
1.04608798e+00 -1.31020844e+00 1.10928930e-01 8.32948148e-01
-1.07241146e-01 3.22497666e-01 7.94199765e-01 -3.47696990e-01
-1.24853373e+00 -1.26374435e+00 9.53751624e-01 -3.65102291e-01
4.15043831e-01 -9.96495724e-01 -6.90204799e-01 7.60139704e-01
4.65793043e-01 -2.06153750e-01 1.36075509e+00 3.18048269e-01
-4.44814265e-01 -6.24270558e-01 -1.81174767e+00 3.98418278e-01
1.05339158e+00 -7.76779354e-01 -2.80531794e-01 1.62084907e-01
5.91557145e-01 -1.66123994e-02 -1.09854198e+00 2.56930500e-01
4.50782478e-01 -9.18861032e-01 1.02991652e+00 -2.49898583e-01
-6.06957734e-01 -1.40314236e-01 -6.90307856e-01 -1.44898653e+00
-4.55526590e-01 -7.34669864e-01 -5.00934385e-02 2.35049438e+00
5.31108320e-01 -8.08536410e-01 7.25248694e-01 9.36614990e-01
1.74187943e-01 5.67819886e-02 -1.03010416e+00 -6.15157366e-01
-2.47775853e-01 -2.30441257e-01 1.32353199e+00 1.29140639e+00
2.41339833e-01 4.76407528e-01 -5.94278455e-01 5.39280236e-01
1.08429658e+00 -2.17597163e-03 9.28223252e-01 -1.13810885e+00
5.71442321e-02 -4.26804513e-01 -5.72535813e-01 -3.21996778e-01
4.26495910e-01 -1.09510779e+00 -7.27509111e-02 -1.34479380e+00
1.42824888e-01 -2.11654663e-01 -2.95052618e-01 2.10427210e-01
-5.30322716e-02 2.25149170e-01 1.16267115e-01 -2.87757039e-01
-6.75166249e-01 3.43362749e-01 4.46656972e-01 -5.71275413e-01
-3.29643101e-01 1.13281578e-01 -1.21528065e+00 4.87367421e-01
8.38061929e-01 -2.47739345e-01 -3.50398988e-01 -2.55845189e-02
-4.31197464e-01 -4.57389578e-02 8.94942880e-02 -8.31803083e-01
2.98292249e-01 1.85628235e-01 4.11986262e-01 -4.32261378e-01
3.27368528e-01 -7.75945485e-01 3.54440242e-01 2.52070259e-02
-3.26682359e-01 -2.11815149e-01 -1.72469959e-01 4.32342410e-01
-1.12934709e-01 1.49874598e-01 6.49983943e-01 2.61808634e-01
-3.49697769e-01 2.08981752e-01 -2.29326129e-01 2.63983738e-02
9.25060570e-01 -2.50368685e-01 -1.67780071e-01 -8.45255852e-01
-7.82951355e-01 3.70752573e-01 5.17973602e-01 5.22266090e-01
3.60056520e-01 -1.58871162e+00 -9.78711963e-01 3.99191856e-01
1.87459573e-01 -3.20104688e-01 4.05389607e-01 7.84562111e-01
1.29066929e-01 1.50112277e-02 1.83726296e-01 -5.16321063e-01
-1.59148681e+00 3.30541581e-01 3.17512244e-01 6.15066849e-02
3.28385197e-02 9.88015831e-01 1.52624533e-01 -9.57376957e-01
5.15602946e-01 -1.67234078e-01 -2.68685490e-01 6.70788705e-01
3.75827253e-01 6.39244676e-01 2.14349926e-01 -1.32039201e+00
-7.44066000e-01 4.34695423e-01 6.69080243e-02 -7.75712682e-03
1.13105392e+00 -5.66211283e-01 1.19263738e-01 3.96178514e-01
1.40776396e+00 2.13713333e-01 -7.94820011e-01 -2.28350937e-01
-1.57701790e-01 -2.75609940e-01 -8.81825835e-02 -7.11476266e-01
-1.07588947e+00 6.96561158e-01 8.38294744e-01 4.24332768e-01
9.58622038e-01 1.51352867e-01 7.08695590e-01 -1.95568457e-01
2.67880231e-01 -1.18962252e+00 -2.01791823e-01 -1.66070342e-01
7.19617188e-01 -1.37801266e+00 -1.45079076e-01 -4.94304389e-01
-1.05525911e+00 7.05340326e-01 4.41500723e-01 5.81052482e-01
6.86412156e-01 2.62893856e-01 3.40439081e-01 8.56726468e-02
-2.01095402e-01 6.10821173e-02 1.31435588e-01 1.22720754e+00
3.46241653e-01 7.38397181e-01 2.92675555e-01 9.67061579e-01
-5.22161901e-01 -4.98376489e-01 4.52677190e-01 3.95293027e-01
-1.69346496e-01 -9.57891762e-01 -6.13855183e-01 6.63363934e-01
-5.54573417e-01 -2.24259775e-03 -4.41083401e-01 1.32083341e-01
-9.84665751e-03 1.60075915e+00 -1.49168551e-01 -6.67258978e-01
4.38771993e-01 4.82239068e-01 2.52706856e-01 -5.37832022e-01
-7.22592890e-01 1.18086874e-01 3.64858776e-01 -2.37697259e-01
-6.14166439e-01 -1.16281581e+00 -9.86034274e-01 -6.61956489e-01
-6.04325891e-01 7.77670890e-02 7.15051830e-01 4.33827013e-01
4.14852768e-01 4.07623470e-01 1.03231537e+00 -7.17561364e-01
-7.30039179e-01 -9.48391855e-01 -1.03860855e+00 6.23517573e-01
5.15549660e-01 -3.58028769e-01 -6.93762243e-01 8.69488791e-02]
|
[14.300603866577148, 6.089914321899414]
|
ab63bc6f-6e9a-4b88-9191-74f608959fea
|
task-oriented-clustering-for-dialogues
| null | null |
https://aclanthology.org/2021.findings-emnlp.368
|
https://aclanthology.org/2021.findings-emnlp.368.pdf
|
Task-Oriented Clustering for Dialogues
|
A reliable clustering algorithm for task-oriented dialogues can help developer analysis and define dialogue tasks efficiently. It is challenging to directly apply prior normal text clustering algorithms for task-oriented dialogues, due to the inherent differences between them, such as coreference, omission and diversity expression. In this paper, we propose a Dialogue Task Clustering Network model for task-oriented clustering. The proposed model combines context-aware utterance representations and cross-dialogue utterance cluster representations for task-oriented dialogues clustering. An iterative end-to-end training strategy is utilized for dialogue clustering and representation learning jointly. Experiments on three public datasets show that our model significantly outperform strong baselines in all metrics.
|
['Xiaojie Wang', 'Caixia Yuan', 'Wei Wu', 'Huixing Jiang', 'Shuyu Lei', 'Hengtong Lu', 'Chenxu Lv']
| null | null | null | null |
findings-emnlp-2021-11
|
['text-clustering']
|
['natural-language-processing']
|
[-8.06384534e-02 4.01805282e-01 3.03288624e-02 -7.14508593e-01
-8.63800704e-01 -7.06235290e-01 6.84043884e-01 -1.24435678e-01
3.92040201e-02 3.35759491e-01 7.08050370e-01 -2.12145030e-01
-2.80657522e-02 2.37245690e-02 2.62966931e-01 -4.18308854e-01
2.91305155e-01 1.00272262e+00 -1.42385453e-01 -5.93919516e-01
4.51492667e-01 -2.88641721e-01 -1.29984522e+00 8.83547306e-01
1.16592324e+00 4.60770309e-01 3.92072767e-01 7.59268284e-01
-7.84013629e-01 1.36701286e+00 -9.62298751e-01 -3.24269921e-01
-4.37397748e-01 -7.08350301e-01 -1.65589702e+00 5.15107274e-01
-1.41018972e-01 2.05299575e-02 8.73044729e-02 8.22625160e-01
5.08311391e-01 5.06653428e-01 1.10547125e+00 -1.17867064e+00
-3.75593513e-01 1.04586458e+00 -4.79395062e-01 -2.45255187e-01
5.63661337e-01 -2.82976538e-01 1.20766628e+00 -5.84131122e-01
5.40362298e-01 1.59362650e+00 3.52833748e-01 1.06061780e+00
-1.10264802e+00 -3.25520158e-01 1.88262969e-01 -3.07785673e-03
-9.54470217e-01 -4.16862160e-01 1.11729956e+00 -8.02096367e-01
9.62738574e-01 3.89110327e-01 8.20900574e-02 1.19650495e+00
-4.03382003e-01 1.06661034e+00 8.57148051e-01 -5.75203240e-01
9.45304781e-02 2.66329288e-01 6.43434048e-01 4.53169435e-01
-8.71218026e-01 -8.59962642e-01 -3.49035561e-01 -3.82827878e-01
3.48708004e-01 -2.67760694e-01 -2.01043263e-01 -1.55770987e-01
-8.64699125e-01 1.14300740e+00 -2.01535746e-01 6.06831431e-01
5.43093495e-02 -3.10115784e-01 9.39894736e-01 4.89053547e-01
7.67830968e-01 7.50966072e-01 -5.07300913e-01 -6.90942287e-01
-5.37180007e-01 1.95397273e-01 1.23629665e+00 1.14470971e+00
5.50312817e-01 -1.93285003e-01 -4.70359981e-01 1.64752626e+00
3.13243359e-01 -2.10939243e-01 8.26966763e-01 -1.36904240e+00
6.45913124e-01 9.56227541e-01 -2.23478213e-01 -7.47273445e-01
-6.01994634e-01 3.85248274e-01 -9.09820914e-01 -3.59869242e-01
3.22408080e-01 -6.88710928e-01 -2.06218556e-01 1.45803201e+00
1.82240307e-01 -1.31708741e-01 4.79930013e-01 7.05550730e-01
1.33343875e+00 5.73510289e-01 1.36072397e-01 -4.53300416e-01
1.38223708e+00 -1.18505228e+00 -1.02840781e+00 2.61753201e-01
1.02117836e+00 -9.51721549e-01 1.25707746e+00 3.07659835e-01
-8.95847321e-01 -4.67233807e-01 -5.10023654e-01 -1.48328424e-01
-1.99245661e-01 1.51362106e-01 5.09170175e-01 8.20178747e-01
-9.66773152e-01 4.42233421e-02 -3.73447329e-01 -4.13086206e-01
-4.70522121e-02 3.59937668e-01 1.18486919e-01 3.65297109e-01
-1.01023650e+00 4.76947159e-01 4.02056873e-01 -2.66315669e-01
-6.33494258e-01 -6.19964182e-01 -8.60425651e-01 -4.81986254e-02
2.69207090e-01 -2.47720197e-01 1.68327880e+00 -8.76287043e-01
-2.05678988e+00 8.39379430e-01 -2.67397702e-01 -1.63347095e-01
1.46038815e-01 -2.14875460e-01 -7.38993064e-02 -4.40431237e-02
4.69859317e-02 5.41358650e-01 5.30155361e-01 -1.39061499e+00
-6.78424835e-01 -2.41979808e-01 1.59067288e-01 8.55199575e-01
-7.04623699e-01 3.96924973e-01 -6.54133201e-01 -2.72505820e-01
-8.23376104e-02 -7.80604541e-01 -3.63015711e-01 -1.05523026e+00
-8.47304940e-01 -1.12107253e+00 1.01519597e+00 -4.76601928e-01
1.29564214e+00 -1.95016372e+00 4.19843733e-01 -1.36863023e-01
5.89420259e-01 -9.64503065e-02 1.02913193e-01 5.41388214e-01
1.15996912e-01 1.91867352e-01 -1.17934383e-01 -6.22143209e-01
1.84596092e-01 6.01514280e-02 -3.11626513e-02 1.14954198e-02
-1.38897719e-02 5.45947194e-01 -9.92434442e-01 -7.48861670e-01
2.88795263e-01 1.39917940e-01 -4.79092330e-01 7.65865982e-01
-5.83687901e-01 7.12475717e-01 -5.97359836e-01 3.00492436e-01
1.61168113e-01 -1.02520153e-01 7.76278496e-01 5.67550361e-02
-1.75898131e-02 6.05470061e-01 -5.92909217e-01 2.05265784e+00
-7.44260371e-01 8.74485195e-01 3.44839543e-01 -1.11427879e+00
1.05288625e+00 7.91849613e-01 5.68617225e-01 -4.45466310e-01
8.01606774e-02 -3.76633018e-01 2.75276810e-01 -6.48453653e-01
8.14987838e-01 2.85181582e-01 -6.26284420e-01 1.02860832e+00
1.26057416e-01 -2.31818616e-01 7.39020258e-02 6.33015454e-01
8.70332718e-01 -1.94951102e-01 1.47812739e-01 -3.83333713e-01
7.56404698e-01 9.48390439e-02 4.04284358e-01 3.46534461e-01
-1.11458749e-01 5.57817757e-01 9.90115404e-01 -1.18923023e-01
-4.59485590e-01 -3.67169678e-01 4.97560427e-02 1.94293678e+00
-2.47245058e-01 -8.13091338e-01 -1.01613784e+00 -1.23454690e+00
-3.56980383e-01 6.40590966e-01 -5.37028074e-01 -1.99413300e-02
-6.01128578e-01 -5.55881500e-01 7.66961753e-01 4.63960320e-02
4.93635744e-01 -9.82774734e-01 -1.82370216e-01 1.74713269e-01
-6.59253061e-01 -9.41612720e-01 -4.96443927e-01 2.20301852e-01
-4.21399146e-01 -1.23630941e+00 -4.36472446e-01 -1.11485302e+00
5.17542243e-01 4.73398149e-01 1.42626369e+00 1.53266758e-01
7.32368156e-02 6.83014750e-01 -7.30508626e-01 -1.65937766e-01
-8.22930336e-01 3.94610286e-01 -7.29391351e-02 -1.11178458e-01
6.99393749e-01 -2.09462389e-01 -1.96649864e-01 5.73176324e-01
-3.80201757e-01 -4.03869040e-02 -4.79827486e-02 1.12493420e+00
-1.61540255e-01 8.39322656e-02 6.57166660e-01 -1.51590240e+00
1.56831825e+00 -6.49768770e-01 4.60971482e-02 4.60350364e-01
-3.16551924e-01 -6.64653704e-02 6.85453415e-01 -3.20824355e-01
-1.66983628e+00 1.51954845e-01 2.14081779e-02 -3.04436594e-01
-5.34621477e-01 3.45698059e-01 -3.76126021e-01 4.07633871e-01
6.70399368e-01 6.85074925e-03 -1.99360750e-03 -5.20190299e-01
7.20454514e-01 1.35815275e+00 3.75755101e-01 -1.18834281e+00
2.31211647e-01 -1.07913889e-01 -8.65890384e-01 -1.27987528e+00
-7.48373866e-01 -9.85214114e-01 -9.71631765e-01 -4.13332373e-01
1.22414315e+00 -8.32468092e-01 -1.00165606e+00 5.21188900e-02
-1.37368429e+00 -4.53219056e-01 1.91922352e-01 1.61528366e-03
-5.90294063e-01 6.92435384e-01 -6.05593085e-01 -1.11799097e+00
-3.59308839e-01 -1.28918207e+00 8.06893408e-01 2.31771618e-01
-7.89233267e-01 -1.42647064e+00 9.78647545e-02 8.88268113e-01
9.17797238e-02 1.06247000e-01 1.05391049e+00 -1.24073064e+00
3.74025106e-01 3.77524555e-01 -7.34558627e-02 2.50898331e-01
3.72747481e-01 6.74671456e-02 -1.14464486e+00 1.87475961e-02
2.34693229e-01 -1.00797820e+00 5.21129549e-01 2.34673157e-01
1.15024388e+00 -3.40516835e-01 -2.87389725e-01 1.03461459e-01
4.53621089e-01 3.98612440e-01 3.42944503e-01 -2.62862705e-02
9.51920807e-01 1.35511136e+00 8.93118739e-01 6.30756617e-01
8.17078054e-01 8.00367951e-01 5.95388338e-02 -5.19378595e-02
1.88852102e-01 2.33154491e-01 3.76179487e-01 1.17351305e+00
1.30899191e-01 -1.57250627e-03 -1.19374585e+00 5.88116407e-01
-2.20163965e+00 -8.68102908e-01 -3.41083407e-01 1.61741042e+00
1.25461006e+00 -1.01991743e-01 5.04307985e-01 -2.20935851e-01
8.51562619e-01 1.55476093e-01 -2.62338340e-01 -6.64560914e-01
2.29845479e-01 -3.41542184e-01 -3.42481196e-01 4.99886632e-01
-1.25865746e+00 1.29155529e+00 5.83451414e+00 7.47857869e-01
-5.18304408e-01 3.62510771e-01 8.19550037e-01 1.43086597e-01
-1.46251559e-01 -1.13180436e-01 -6.78555191e-01 3.78534198e-01
9.22479212e-01 -2.32055560e-01 1.47234067e-01 8.98334384e-01
1.66854784e-01 3.38165492e-01 -1.23616195e+00 8.76339197e-01
2.80658662e-01 -1.01748192e+00 -5.67762330e-02 5.38790505e-03
6.49400592e-01 -3.54822338e-01 -4.33358829e-04 7.58455575e-01
1.14047408e+00 -1.09801865e+00 -1.22515909e-01 5.34193125e-03
4.43170965e-01 -9.97984052e-01 5.84544063e-01 5.82376301e-01
-9.78896201e-01 1.17294312e-01 -4.42036986e-01 6.58340752e-02
-2.49748394e-01 1.00153238e-02 -1.43003762e+00 5.21514118e-01
4.61405873e-01 7.33931959e-01 -2.57356644e-01 2.05035850e-01
7.09109157e-02 7.95669973e-01 1.03422016e-01 -2.84813195e-01
4.26355839e-01 -4.26657587e-01 2.99791962e-01 1.61709154e+00
-4.05145139e-01 2.20586285e-01 5.73687136e-01 8.12046885e-01
-1.27761543e-01 3.02441061e-01 -6.02453411e-01 -1.12633985e-02
8.52225542e-01 1.58779085e+00 -5.07524729e-01 -1.60732105e-01
-4.81191158e-01 8.97490084e-01 5.36863804e-01 3.11572820e-01
-4.38000888e-01 -4.61965293e-01 8.07447493e-01 -4.99923587e-01
-1.50758579e-01 -1.58896685e-01 -2.85986423e-01 -9.76415277e-01
-4.66772079e-01 -1.15702319e+00 6.08820200e-01 5.87868243e-02
-1.44632185e+00 6.39893234e-01 -4.61394787e-02 -1.08902121e+00
-8.41360390e-01 -1.86091989e-01 -1.14139056e+00 5.60132205e-01
-9.57128823e-01 -1.21240532e+00 -1.58214658e-01 8.93291473e-01
1.41878843e+00 -9.00381804e-01 1.05683291e+00 -1.51171997e-01
-8.19061339e-01 7.62802005e-01 1.33232638e-01 6.00253701e-01
9.29870307e-01 -1.81269360e+00 1.54528096e-01 1.24965586e-01
1.31640062e-01 8.55018973e-01 5.67201853e-01 -5.14835536e-01
-1.07179761e+00 -1.00826979e+00 6.10001981e-01 -6.87255561e-01
6.27905428e-01 -7.29577184e-01 -9.45569456e-01 3.93426150e-01
9.35257375e-01 -1.00491846e+00 1.43370962e+00 9.55385983e-01
-2.60239035e-01 4.70626444e-01 -9.12232280e-01 5.11898160e-01
5.19675791e-01 -5.95185220e-01 -8.73007774e-01 9.33071792e-01
8.39441538e-01 -3.84967387e-01 -1.23235512e+00 -1.41546682e-01
4.03750055e-02 -8.56533647e-01 7.38681555e-01 -6.83013380e-01
6.13228619e-01 2.33992979e-01 -6.99688047e-02 -1.30146337e+00
-1.16561092e-01 -1.03107607e+00 4.02757525e-02 1.96666992e+00
4.44225132e-01 7.43961185e-02 7.20111012e-01 8.13356340e-01
-4.02943939e-01 -3.69322449e-01 -7.10587204e-01 -3.45105708e-01
4.68973875e-01 -2.34965473e-01 4.30995554e-01 1.49986720e+00
1.00175047e+00 1.15585041e+00 -4.06655520e-01 -1.61537990e-01
3.67039293e-01 1.27687827e-01 1.04763961e+00 -1.57246256e+00
-1.43654794e-01 -6.09717667e-01 1.46253094e-01 -1.36986816e+00
8.99113238e-01 -7.69320011e-01 3.70979607e-01 -1.38220322e+00
1.81330800e-01 -5.14795363e-01 1.76359549e-01 3.28158379e-01
-3.99289608e-01 -3.29533130e-01 -7.92179629e-02 4.90762293e-01
-1.36715388e+00 7.11420298e-01 7.91133463e-01 -2.36884966e-01
-6.57859862e-01 2.08501741e-01 -8.23264897e-01 8.26325357e-01
1.03080416e+00 -1.89885050e-01 -8.66295934e-01 -3.42792243e-01
-3.19754750e-01 2.90264785e-01 -4.97904986e-01 -5.95395505e-01
5.10254562e-01 -8.43096673e-02 -8.84446055e-02 -6.64227962e-01
4.24209505e-01 -4.39511031e-01 -7.21907318e-01 -8.01047385e-02
-9.62696731e-01 -2.42616355e-01 -1.43342674e-01 5.74498653e-01
-3.64235103e-01 -4.29902524e-01 5.00350773e-01 -1.95340097e-01
-5.64237893e-01 -3.53161961e-01 -8.90401542e-01 4.51007932e-01
1.01579809e+00 9.91888270e-02 -3.68368655e-01 -8.67128789e-01
-6.61547244e-01 7.51654744e-01 1.98157966e-01 6.77536130e-01
5.39348602e-01 -1.07028127e+00 -8.63587916e-01 -4.09758687e-01
8.44019055e-02 1.29297361e-01 2.50455827e-01 4.45437998e-01
-2.72358693e-02 5.84150553e-01 1.16884284e-01 -8.51002157e-01
-1.92017138e+00 4.06654514e-02 1.71348006e-01 -4.65690404e-01
-3.83222878e-01 8.46273005e-01 3.65831465e-01 -1.10839915e+00
6.98581755e-01 1.58004314e-01 -9.21828687e-01 3.34106654e-01
3.93812031e-01 2.73053199e-01 -2.97845542e-01 -7.49214649e-01
-6.19012527e-02 1.73732147e-01 -6.95028245e-01 -1.58130839e-01
1.01021075e+00 -5.32730520e-01 -3.42130065e-01 7.93872297e-01
1.03795350e+00 -3.99842590e-01 -8.83707225e-01 -2.39465028e-01
5.58370233e-01 -9.93139893e-02 -2.42267326e-01 -5.82972944e-01
-6.80883348e-01 1.00002277e+00 1.03523687e-01 7.78861165e-01
6.64812982e-01 2.45781034e-01 4.84126359e-01 8.54202330e-01
-1.45098835e-01 -1.55655730e+00 5.34687579e-01 9.91133869e-01
7.64736533e-01 -1.42940855e+00 -1.99242309e-01 -4.63965863e-01
-1.39380717e+00 1.05053782e+00 1.19994688e+00 3.98525834e-01
3.98287863e-01 9.47504491e-02 4.76594865e-01 -5.43039978e-01
-1.15106869e+00 -1.21545903e-01 2.35195637e-01 9.50477004e-01
1.28235793e+00 -4.63633798e-02 -1.31705016e-01 8.31312776e-01
-2.74238527e-01 -9.11927938e-01 6.25637650e-01 6.77796304e-01
-4.40288126e-01 -1.27008939e+00 -2.01704279e-01 4.11107302e-01
-4.57017899e-01 -6.01041391e-02 -1.33334661e+00 6.20466709e-01
-4.01653200e-01 1.79346418e+00 1.51332721e-01 -7.01265872e-01
5.76323606e-02 4.41291064e-01 -1.31721631e-01 -1.13200498e+00
-1.28015220e+00 3.18944067e-01 6.54108703e-01 -2.57578969e-01
-7.00246215e-01 -7.10859895e-01 -1.44973314e+00 -6.63565695e-02
-4.40362096e-01 8.87451828e-01 4.44550693e-01 8.00046682e-01
2.88364619e-01 5.96054316e-01 1.12733543e+00 -5.09998262e-01
-3.69696021e-01 -1.40413582e+00 -3.78237754e-01 6.05239630e-01
-1.79028928e-01 -4.23804373e-01 -1.40223339e-01 1.43818408e-01]
|
[12.663139343261719, 7.782310962677002]
|
2a03e575-ee87-4c83-a8a9-294e1ad75bf7
|
distilling-translations-with-visual-awareness
|
1906.07701
| null |
https://arxiv.org/abs/1906.07701v1
|
https://arxiv.org/pdf/1906.07701v1.pdf
|
Distilling Translations with Visual Awareness
|
Previous work on multimodal machine translation has shown that visual information is only needed in very specific cases, for example in the presence of ambiguous words where the textual context is not sufficient. As a consequence, models tend to learn to ignore this information. We propose a translate-and-refine approach to this problem where images are only used by a second stage decoder. This approach is trained jointly to generate a good first draft translation and to improve over this draft by (i) making better use of the target language textual context (both left and right-side contexts) and (ii) making use of visual context. This approach leads to the state of the art results. Additionally, we show that it has the ability to recover from erroneous or missing words in the source language.
|
['Lucia Specia', 'Julia Ive', 'Pranava Madhyastha']
|
2019-06-18
|
distilling-translations-with-visual-awareness-1
|
https://aclanthology.org/P19-1653
|
https://aclanthology.org/P19-1653.pdf
|
acl-2019-7
|
['multimodal-machine-translation']
|
['natural-language-processing']
|
[ 5.20722449e-01 1.75569326e-01 -5.47617499e-04 -3.11909735e-01
-8.83591592e-01 -8.84397447e-01 9.77536201e-01 3.00045878e-01
-4.62379158e-01 9.80900824e-01 1.80032447e-01 -7.08862245e-01
5.17529011e-01 -4.65705812e-01 -7.51863241e-01 -5.17790496e-01
5.70895910e-01 7.25543022e-01 1.30835995e-01 -3.36855233e-01
1.26806557e-01 3.15753162e-01 -1.22325265e+00 6.07376337e-01
7.95699537e-01 3.89512300e-01 5.21947086e-01 6.90130830e-01
-3.32285196e-01 4.81710136e-01 -4.53938246e-01 -5.73516071e-01
2.28202552e-01 -9.83082712e-01 -7.07514405e-01 3.50766093e-01
4.06574577e-01 -4.16356623e-01 1.26600802e-01 9.75704730e-01
4.12098646e-01 -2.65055180e-01 6.27822459e-01 -7.51183808e-01
-2.45843992e-01 4.77444559e-01 -6.74305797e-01 -2.13150486e-01
4.72327352e-01 1.26114905e-01 8.53964150e-01 -1.07890630e+00
9.61753011e-01 1.05396199e+00 1.42752990e-01 5.73410153e-01
-1.43438435e+00 -2.53507763e-01 2.39819974e-01 4.74467874e-02
-1.21018314e+00 -6.50360644e-01 6.13126159e-01 -2.50134110e-01
8.98573458e-01 1.96838602e-01 3.06342185e-01 1.12060666e+00
3.45135201e-03 7.21949816e-01 1.22186315e+00 -9.41791356e-01
8.40213671e-02 4.94167447e-01 -3.79386961e-01 6.35484576e-01
1.41560912e-01 -3.58107165e-02 -3.11003149e-01 3.30172814e-02
3.18801075e-01 -3.64993185e-01 -3.59686434e-01 -3.89979035e-01
-1.24790335e+00 7.04177201e-01 8.91083293e-03 6.27116859e-01
-3.53804022e-01 -1.64536498e-02 1.50741115e-01 5.23152292e-01
4.16084290e-01 4.36328322e-01 -1.34684220e-01 -7.89468959e-02
-1.34663403e+00 1.98653802e-01 8.09905827e-01 8.85546505e-01
8.38202477e-01 -2.06669316e-01 -7.45305419e-02 9.13700640e-01
2.68129647e-01 6.09202325e-01 1.46169111e-01 -6.08280718e-01
7.16195941e-01 3.78276616e-01 3.58398944e-01 -5.91559827e-01
-9.87289175e-02 -3.76664042e-01 -3.51117045e-01 4.45049226e-01
5.14720976e-01 -2.49366224e-01 -1.29854572e+00 1.77151513e+00
2.03501865e-01 -2.36466363e-01 1.87455177e-01 9.46166575e-01
3.94839436e-01 6.70446694e-01 -4.26861681e-02 -3.13821375e-01
1.18867922e+00 -7.54721940e-01 -7.30026424e-01 -6.24205768e-01
7.42229462e-01 -1.26247716e+00 9.01658475e-01 2.71426171e-01
-1.17511094e+00 -2.99493879e-01 -1.12017345e+00 -1.31118670e-01
-2.51555562e-01 4.17608112e-01 -1.49928937e-02 5.64341784e-01
-1.10824251e+00 5.06361246e-01 -5.49655616e-01 -4.15177077e-01
2.07723454e-01 3.83541465e-01 -5.69943309e-01 -5.51760197e-01
-9.39012289e-01 1.20232308e+00 2.65671611e-01 2.31206343e-01
-7.02683628e-01 -1.15435332e-01 -8.36484849e-01 -1.64750278e-01
5.24153471e-01 -6.59562528e-01 1.31031907e+00 -1.58761847e+00
-1.28565431e+00 8.73710513e-01 -5.72707534e-01 -1.27899483e-01
9.67994571e-01 -1.34730235e-01 1.19792074e-01 4.14339513e-01
-5.20222634e-02 8.30824435e-01 9.40335155e-01 -1.69906831e+00
-5.85973263e-01 -3.08655709e-01 1.81105047e-01 3.94827038e-01
-2.96229757e-02 2.42334176e-02 -7.96914160e-01 -6.29033029e-01
2.42092479e-02 -1.23633015e+00 -5.70399389e-02 6.21055216e-02
-3.18253875e-01 1.39088184e-01 8.78810346e-01 -1.19349110e+00
9.28499699e-01 -1.91675317e+00 5.30600846e-01 2.44132206e-01
3.79269570e-02 2.98892915e-01 -2.86077648e-01 7.27076471e-01
-5.70826866e-02 1.17273837e-01 -4.31089818e-01 -6.82947397e-01
-2.33413354e-01 3.46737027e-01 -2.99713761e-01 2.71899670e-01
3.20902944e-01 7.67997026e-01 -7.44312167e-01 -6.05947196e-01
2.18805522e-01 3.81907642e-01 -2.33764455e-01 1.49913043e-01
-3.62283170e-01 5.63527644e-01 -8.35360587e-02 4.02226448e-01
6.16872013e-01 -5.08791544e-02 5.44470668e-01 4.82058488e-02
-1.02268398e-01 3.23934704e-01 -9.79462385e-01 1.66152024e+00
-5.95913887e-01 9.54294920e-01 2.63396680e-01 -7.14492679e-01
5.97643673e-01 5.58187306e-01 1.15839802e-01 -7.75845706e-01
1.05937637e-01 4.66999620e-01 1.34969890e-01 -5.12494266e-01
3.59208137e-01 -4.71846491e-01 2.61704892e-01 5.82778096e-01
3.04110553e-02 -1.71444237e-01 4.11697268e-01 4.38958704e-01
7.03617454e-01 4.77927119e-01 2.35224232e-01 2.02085242e-01
4.98987019e-01 1.65883288e-01 2.68887520e-01 5.73440850e-01
1.72037601e-01 9.26077485e-01 8.28655899e-01 6.73543587e-02
-1.42815399e+00 -6.15395069e-01 4.30036098e-01 7.43350029e-01
-4.12035594e-03 -3.26749504e-01 -8.50565851e-01 -9.80610609e-01
-4.97476101e-01 1.21743309e+00 -6.26999021e-01 -3.76506038e-02
-8.76676083e-01 -2.94897914e-01 1.23287804e-01 3.23358953e-01
9.01121870e-02 -8.46172988e-01 -5.60539663e-01 1.52284384e-01
-4.77448702e-01 -1.19094419e+00 -4.49898392e-01 5.02377637e-02
-8.51002932e-01 -8.32203805e-01 -1.14284050e+00 -7.55220294e-01
9.57201540e-01 1.79859951e-01 9.78895605e-01 2.84007102e-01
2.01410875e-01 1.62782684e-01 -4.41614121e-01 -2.01782778e-01
-7.83835769e-01 -2.68658847e-02 -4.53827143e-01 5.86532578e-02
4.12022974e-03 -3.35968703e-01 -2.12260380e-01 1.79051027e-01
-1.07506609e+00 5.33019304e-01 8.26202989e-01 8.93025875e-01
2.61389941e-01 -5.20339668e-01 2.40479842e-01 -8.32664609e-01
5.21038473e-01 -1.17560528e-01 -5.51219523e-01 4.32229310e-01
-4.50488240e-01 2.25301504e-01 5.67459583e-01 -5.56490421e-01
-9.82685030e-01 2.57852584e-01 -1.57315284e-01 -2.72903651e-01
-2.43457153e-01 4.96521294e-01 -2.10535303e-01 6.73711076e-02
3.75839025e-01 2.52347201e-01 -9.23443064e-02 -5.05261064e-01
2.66722739e-01 6.81969106e-01 1.87110752e-01 -3.63654882e-01
6.98031843e-01 3.19615543e-01 -2.36661304e-02 -9.15740788e-01
-2.51594365e-01 -2.06907436e-01 -8.34458649e-01 -2.53883570e-01
6.95925951e-01 -7.63305485e-01 8.29341412e-02 8.20677876e-02
-1.38277364e+00 -3.22011530e-01 1.39219472e-02 4.65972006e-01
-4.79499370e-01 5.49910486e-01 -3.38968396e-01 -9.87406492e-01
7.47971162e-02 -1.34931123e+00 1.13289249e+00 1.17973527e-02
-2.69674540e-01 -7.55894363e-01 -3.31551768e-02 1.71359971e-01
1.91767797e-01 1.64903363e-03 1.15147424e+00 -6.82075620e-01
-4.71762687e-01 -1.39774069e-01 -2.48057708e-01 1.68476060e-01
-2.08452623e-02 -2.32460778e-02 -7.76581287e-01 -1.30414650e-01
-9.84097570e-02 -1.58403903e-01 9.19934273e-01 -3.07019241e-02
2.70627856e-01 -1.08521298e-01 -3.65143061e-01 4.95122522e-02
1.35546005e+00 1.21814124e-01 6.92932069e-01 8.25127065e-02
5.98439097e-01 9.97451663e-01 4.42738593e-01 -3.52592655e-02
3.85000527e-01 1.02111626e+00 3.83945286e-01 -2.55870551e-01
-3.56754899e-01 -2.84636915e-01 3.60173792e-01 5.29974163e-01
-2.54745875e-03 -4.84487921e-01 -9.55280542e-01 9.43508148e-01
-1.92597532e+00 -7.27065444e-01 -1.63503900e-01 2.33213878e+00
8.57731104e-01 -3.60437832e-03 2.55482011e-02 -4.56665456e-02
6.72878444e-01 -3.93855870e-02 -5.50746173e-02 -6.02876008e-01
-1.98984042e-01 3.03320419e-02 4.51648831e-01 7.68700302e-01
-7.14839220e-01 1.06603301e+00 6.09313536e+00 6.98710620e-01
-1.18608236e+00 1.23971999e-01 4.03608710e-01 5.73258258e-05
-6.80747747e-01 4.52268332e-01 -4.13050115e-01 3.65944684e-01
6.87143803e-01 3.67344201e-01 3.86889070e-01 2.16135651e-01
3.32522720e-01 -6.06898963e-01 -1.11135685e+00 8.14381659e-01
2.24607408e-01 -8.86087716e-01 1.33223206e-01 1.52939066e-01
5.89153767e-01 -3.97656769e-01 -2.18990102e-01 4.27065864e-02
5.00077568e-02 -1.03245008e+00 9.83661890e-01 4.79142964e-01
8.87073159e-01 -6.49492860e-01 9.69285071e-01 5.58655679e-01
-7.68536448e-01 2.20230892e-01 -1.17648490e-01 7.49384016e-02
4.44103450e-01 4.16791350e-01 -1.08977425e+00 7.37225533e-01
1.48035020e-01 3.49180758e-01 -5.50013125e-01 9.46086347e-01
-6.61990762e-01 5.46177447e-01 -3.41087699e-01 -1.37742773e-01
3.42328668e-01 -2.26873681e-01 6.53205454e-01 1.25855541e+00
4.06806737e-01 -2.53779143e-02 1.28520476e-02 6.07415974e-01
1.47590619e-02 1.95686892e-01 -7.77237594e-01 -1.68097317e-01
-1.16852567e-01 9.19956923e-01 -8.36648881e-01 -4.96265233e-01
-5.81809282e-01 1.51806295e+00 2.86556661e-01 6.22063518e-01
-4.92864609e-01 -2.02971116e-01 2.71116197e-01 1.29201531e-01
6.22668445e-01 -5.18923938e-01 -3.41509998e-01 -1.16431034e+00
1.41910389e-01 -9.94623780e-01 6.44409955e-02 -9.69061553e-01
-5.59616268e-01 5.33701360e-01 -5.91510348e-02 -1.04171979e+00
-5.08405030e-01 -6.34601057e-01 -2.48480320e-01 1.16342533e+00
-1.52325499e+00 -1.30668986e+00 2.75629401e-01 2.93154508e-01
7.28852689e-01 1.43030941e-01 6.37235105e-01 2.42534712e-01
-6.88733906e-02 3.23052585e-01 -1.06179327e-01 3.80935445e-02
1.01350331e+00 -1.25953043e+00 9.03121680e-02 1.09483647e+00
3.54082912e-01 5.12471855e-01 1.18593490e+00 -7.87128448e-01
-1.30723274e+00 -6.35569453e-01 1.63379145e+00 -3.28196257e-01
2.59113103e-01 -5.77538490e-01 -7.65256703e-01 5.93685210e-01
5.80721140e-01 -4.67779487e-01 4.41644102e-01 -2.04885781e-01
-2.35707656e-01 7.24067837e-02 -9.61564600e-01 8.89679074e-01
5.37326276e-01 -5.01252532e-01 -7.13132501e-01 2.36659735e-01
3.02931547e-01 -4.09918964e-01 7.11055938e-03 1.44802630e-01
6.27031505e-01 -8.25880885e-01 4.77909595e-01 -4.11708057e-01
6.88237906e-01 -4.89741862e-01 -1.21514797e-01 -1.33916223e+00
2.34522745e-01 -6.06240034e-01 1.32264853e-01 9.32982147e-01
9.52561021e-01 -2.82103598e-01 3.86995167e-01 4.72360432e-01
1.24690183e-01 -5.52748084e-01 -9.17048395e-01 -3.73646647e-01
-3.71045247e-02 -2.29573667e-01 1.28952429e-01 7.51454115e-01
-8.61699358e-02 6.76341057e-01 -7.63051391e-01 -1.27211388e-03
1.16028681e-01 2.09156826e-01 7.77383506e-01 -7.81591713e-01
-2.11203739e-01 -3.29722047e-01 -6.11540936e-02 -1.17436171e+00
-1.25191838e-01 -6.16465926e-01 3.18673074e-01 -1.83212757e+00
2.26718977e-01 -1.06555715e-01 1.37986287e-01 4.57018107e-01
-2.18178779e-01 5.07793665e-01 4.82050538e-01 1.53212026e-01
-2.97098428e-01 2.30565056e-01 1.18615925e+00 -1.20763324e-01
-4.87093143e-02 -1.17773190e-01 -6.08718991e-01 3.96025479e-01
6.53894603e-01 -5.58372021e-01 -2.36192942e-01 -6.34317040e-01
4.41545159e-01 1.17483571e-01 2.74831027e-01 -4.95337605e-01
-5.99126443e-02 -1.77530795e-01 4.68996614e-01 -4.82200563e-01
4.10659254e-01 -1.08559787e+00 5.30774519e-02 4.26334441e-01
-3.01593482e-01 2.19016880e-01 2.71734238e-01 2.80164719e-01
-1.63435489e-01 -5.75683773e-01 6.34395301e-01 -3.07456046e-01
-3.73439074e-01 -3.13677579e-01 -5.18077910e-01 -3.81189406e-01
9.03267860e-01 -2.53858745e-01 1.97028797e-02 -8.74661863e-01
-7.74135351e-01 8.51773396e-02 7.92207718e-01 3.26155871e-01
5.00608921e-01 -1.12350857e+00 -9.24049497e-01 3.10209896e-02
1.90585945e-02 -4.56240505e-01 -9.53184441e-02 1.09666491e+00
-5.03873289e-01 3.37512672e-01 9.86590758e-02 -5.77104807e-01
-1.63575959e+00 5.93329251e-01 9.23248157e-02 -2.97750652e-01
-5.25741756e-01 4.57470894e-01 1.92247301e-01 -1.51439518e-01
-1.42121792e-01 -9.26988646e-02 -2.55538076e-01 6.57824725e-02
5.84305584e-01 -1.79957867e-01 2.20177904e-01 -1.00788283e+00
-3.22373331e-01 6.71414673e-01 -4.11304645e-02 -9.24323320e-01
1.08057666e+00 -4.05658513e-01 -1.62644107e-02 4.48690891e-01
9.43078339e-01 5.42666256e-01 -1.05077302e+00 -8.95788223e-02
7.62017891e-02 -4.66043085e-01 -1.40324876e-01 -1.18399143e+00
-6.90448403e-01 1.20106041e+00 3.23930919e-01 -7.35274181e-02
1.17431021e+00 -3.10775265e-02 5.31410754e-01 1.05370097e-01
1.63446024e-01 -1.08930957e+00 -1.91619471e-01 5.32590389e-01
9.15336370e-01 -1.17503130e+00 7.27034509e-02 -3.71271878e-01
-8.20714295e-01 1.34431958e+00 1.08314425e-01 2.61297613e-01
6.59865141e-02 1.21759638e-01 3.38033408e-01 1.94772959e-01
-8.38285208e-01 -4.44097012e-01 4.42881763e-01 5.15748739e-01
4.67956334e-01 -1.14055276e-01 -5.64439535e-01 6.84925318e-02
1.46920353e-01 -1.88345805e-01 4.93063748e-01 1.11857200e+00
-3.47165555e-01 -1.55308867e+00 -4.98475015e-01 6.17349260e-02
-5.81762016e-01 -2.26387814e-01 -9.65379298e-01 7.99471080e-01
5.14784604e-02 1.14620972e+00 -2.87743211e-01 -1.11841386e-04
2.37754375e-01 4.02289748e-01 7.65018761e-01 -6.23474240e-01
-5.20486534e-01 6.66395247e-01 4.37483758e-01 -3.74636710e-01
-2.64196932e-01 -6.18960857e-01 -9.03956234e-01 1.02317473e-02
-3.66541415e-01 3.98302072e-04 1.00363135e+00 1.27538836e+00
3.30960661e-01 3.64779115e-01 1.85813501e-01 -7.86859214e-01
-1.03341922e-01 -1.01103973e+00 -8.84234533e-02 3.35142255e-01
6.15625203e-01 -4.16177124e-01 -1.28955185e-01 2.83074528e-01]
|
[11.445738792419434, 1.4403221607208252]
|
86b5f27f-272d-4fa5-8b12-9ad8945e6acf
|
acr-loss-adaptive-coordinate-based-regression
|
2203.15835
| null |
https://arxiv.org/abs/2203.15835v2
|
https://arxiv.org/pdf/2203.15835v2.pdf
|
ACR Loss: Adaptive Coordinate-based Regression Loss for Face Alignment
|
Although deep neural networks have achieved reasonable accuracy in solving face alignment, it is still a challenging task, specifically when we deal with facial images, under occlusion, or extreme head poses. Heatmap-based Regression (HBR) and Coordinate-based Regression (CBR) are among the two mainly used methods for face alignment. CBR methods require less computer memory, though their performance is less than HBR methods. In this paper, we propose an Adaptive Coordinate-based Regression (ACR) loss to improve the accuracy of CBR for face alignment. Inspired by the Active Shape Model (ASM), we generate Smooth-Face objects, a set of facial landmark points with less variations compared to the ground truth landmark points. We then introduce a method to estimate the level of difficulty in predicting each landmark point for the network by comparing the distribution of the ground truth landmark points and the corresponding Smooth-Face objects. Our proposed ACR Loss can adaptively modify its curvature and the influence of the loss based on the difficulty level of predicting each landmark point in a face. Accordingly, the ACR Loss guides the network toward challenging points than easier points, which improves the accuracy of the face alignment task. Our extensive evaluation shows the capabilities of the proposed ACR Loss in predicting facial landmark points in various facial images.
|
['Mohammad H. Mahoor', 'Ali Pourramezan Fard']
|
2022-03-29
| null | null | null | null |
['face-alignment', 'facial-landmark-detection']
|
['computer-vision', 'computer-vision']
|
[-2.38975823e-01 8.15789104e-02 -1.36258885e-01 -7.17952609e-01
-6.38560593e-01 8.74299109e-02 3.00263584e-01 -5.06369919e-02
-3.70547056e-01 2.17712864e-01 7.60429054e-02 3.62577379e-01
-2.95884252e-01 -5.57212889e-01 -5.61240852e-01 -8.05283368e-01
2.11269632e-01 5.46212733e-01 -1.35864556e-01 -3.65156472e-01
2.96054989e-01 9.49327230e-01 -1.40979183e+00 -2.81013817e-01
8.94059002e-01 1.07106197e+00 -9.45349857e-02 -8.91192853e-02
-4.51252870e-02 2.46254727e-01 -4.48203266e-01 -6.00726008e-01
4.20494676e-01 -1.59990966e-01 -3.61553729e-01 -3.33110965e-03
8.31713438e-01 -1.55575842e-01 -1.90071687e-02 1.13578093e+00
7.72738218e-01 2.68909603e-01 7.86035061e-01 -1.53670490e+00
-5.45309961e-01 2.23406777e-01 -1.03775322e+00 1.40557773e-02
2.95195699e-01 -1.27858937e-01 6.70395911e-01 -1.15261352e+00
2.61287421e-01 1.45436394e+00 7.81353176e-01 7.92312384e-01
-9.87702072e-01 -8.43133092e-01 3.04093331e-01 4.37651455e-01
-1.69191849e+00 -9.35910046e-01 1.21370494e+00 -3.25438321e-01
2.72637546e-01 3.89434658e-02 5.62886119e-01 5.37436545e-01
-2.31292360e-02 4.80755091e-01 5.55205226e-01 -3.03074479e-01
1.45286202e-01 -9.64711085e-02 -1.50678590e-01 8.59159231e-01
-2.25566067e-02 -2.24228963e-01 -4.46971774e-01 -2.06545383e-01
8.63035321e-01 8.52893814e-02 -3.56410027e-01 -4.65837240e-01
-6.16615593e-01 6.53563261e-01 7.75340617e-01 1.38099104e-01
-5.08905053e-01 -6.03509471e-02 1.94092728e-02 -1.66336000e-01
5.92916071e-01 1.60686582e-01 -1.30308315e-01 2.77221203e-01
-9.55707967e-01 2.12457970e-01 2.49962449e-01 7.91823328e-01
7.54417658e-01 1.81810766e-01 -1.64256170e-01 1.21382320e+00
6.13729537e-01 3.38087499e-01 3.90888423e-01 -8.68769288e-01
4.92604047e-01 8.73450518e-01 -7.15956613e-02 -1.80513883e+00
-5.20143390e-01 -1.70986488e-01 -9.37610924e-01 2.87629992e-01
4.19054389e-01 4.54080775e-02 -6.32910132e-01 1.81701434e+00
6.00231707e-01 2.89153606e-01 -3.90065998e-01 1.13039196e+00
8.28063846e-01 4.98038620e-01 8.55447538e-03 -3.40036869e-01
8.46288979e-01 -8.09255064e-01 -7.03627288e-01 -1.43622667e-01
3.10270488e-01 -7.38328934e-01 9.22393322e-01 3.41416821e-02
-1.12238121e+00 -5.60266614e-01 -7.79030263e-01 3.46710570e-02
8.38850513e-02 3.03131551e-01 1.18627980e-01 4.21322346e-01
-1.19432604e+00 4.61373419e-01 -6.15670979e-01 1.35403257e-02
6.34376049e-01 5.74550271e-01 -4.80837822e-01 1.39321625e-01
-7.96392977e-01 8.26854050e-01 -2.04674542e-01 6.52382731e-01
-4.28547025e-01 -7.13062942e-01 -8.86837423e-01 2.16665819e-01
2.22461328e-01 -2.92861760e-01 7.27233410e-01 -1.19520450e+00
-1.62694025e+00 8.03883731e-01 -2.60760665e-01 -3.08342166e-02
6.52113676e-01 -1.36594519e-01 -8.66621062e-02 2.98471260e-03
-1.84396386e-01 1.03182876e+00 1.15114355e+00 -1.06346512e+00
-2.05578074e-01 -8.60058606e-01 -2.22509190e-01 3.98982853e-01
-4.07692105e-01 8.81639272e-02 -5.04135430e-01 -4.31604981e-01
4.52555537e-01 -8.47835243e-01 -1.58433139e-01 5.14016867e-01
-1.49384022e-01 -5.96938789e-01 7.05455124e-01 -6.64403677e-01
9.21825826e-01 -2.25185251e+00 2.93207258e-01 4.08681810e-01
1.62086830e-01 2.53875405e-01 -4.32460517e-01 -2.90151864e-01
-2.31087431e-01 -4.18711044e-02 -2.91649234e-02 -4.18606281e-01
-1.30043447e-01 -1.90264836e-01 -1.51310578e-01 5.57329178e-01
2.90610909e-01 6.81305289e-01 -5.90490341e-01 -4.96178269e-01
-6.86763301e-02 7.25584567e-01 -7.22917318e-01 1.48080766e-01
2.79250294e-02 4.10605401e-01 -2.85942584e-01 6.66976392e-01
8.46374631e-01 1.53013289e-01 -6.85222968e-02 -5.19879937e-01
9.93052498e-02 -3.08704287e-01 -1.11049783e+00 1.24922884e+00
-4.11053121e-01 5.01339793e-01 -2.14882568e-02 -7.39804745e-01
1.53859150e+00 2.93119460e-01 6.47026777e-01 -7.11430311e-01
2.38195196e-01 4.04146798e-02 -3.45966667e-02 -3.06038707e-01
3.05735588e-01 2.43134543e-01 5.25430560e-01 7.35793114e-02
-2.30305970e-01 1.07895896e-01 -2.47692451e-01 -3.86661559e-01
4.29533571e-01 2.45695412e-02 2.90626019e-01 -1.70676738e-01
7.51338959e-01 -5.49855232e-01 7.95792997e-01 -9.37910005e-02
-5.09700418e-01 6.53048396e-01 4.95309114e-01 -7.24061847e-01
-9.76999462e-01 -5.79202473e-01 -2.14242622e-01 9.39446509e-01
2.72721410e-01 -1.93948686e-01 -9.58422244e-01 -6.96248233e-01
-7.14681596e-02 3.00852209e-01 -6.24080896e-01 -2.84262627e-01
-8.09815645e-01 -6.58479452e-01 3.01093429e-01 4.56994832e-01
6.82603180e-01 -1.00653672e+00 -7.41623864e-02 -1.46517172e-01
-1.73883289e-01 -9.22831774e-01 -6.73309743e-01 -6.90260947e-01
-6.37032509e-01 -1.00448942e+00 -7.84811199e-01 -8.26611698e-01
1.33060873e+00 7.37454742e-02 7.88394690e-01 3.50544542e-01
-2.01132894e-01 1.85065255e-01 -9.67278332e-02 -4.64809388e-01
-3.96114662e-02 -7.86848813e-02 1.66619822e-01 3.30235481e-01
2.48093575e-01 -2.94359624e-01 -7.41831422e-01 8.26756239e-01
-4.68016118e-01 -1.18545935e-01 1.96907058e-01 6.33770525e-01
6.13269866e-01 -8.36857036e-02 4.59501415e-01 -3.78870547e-01
5.97126424e-01 -1.96405813e-01 -7.37977266e-01 2.42876932e-01
-3.68361115e-01 -2.10341945e-01 5.31296372e-01 -5.49559951e-01
-6.68146849e-01 2.11076558e-01 -3.39883775e-01 -6.58583939e-01
5.61227389e-02 1.31018654e-01 -3.03960055e-01 -4.42441702e-01
5.79981685e-01 -1.50269806e-01 3.13077211e-01 -2.22969741e-01
8.72503147e-02 4.07497138e-01 3.14780772e-01 -5.13383389e-01
7.89600790e-01 3.05033565e-01 2.76506066e-01 -6.19885027e-01
-7.26944506e-01 -1.58259690e-01 -8.01355481e-01 -5.97131729e-01
6.48076236e-01 -5.31856716e-01 -9.61786270e-01 4.75038320e-01
-1.07569444e+00 -1.91340894e-01 1.27304658e-01 1.68936729e-01
-4.94386852e-01 2.87043959e-01 -6.12887144e-02 -7.46318579e-01
-5.80810249e-01 -1.42209983e+00 1.04488850e+00 4.78628188e-01
-1.40863463e-01 -8.73613894e-01 -1.61384016e-01 2.89182514e-01
3.24172556e-01 2.55476475e-01 8.71609688e-01 -4.49853450e-01
-2.90893883e-01 -4.59545791e-01 -8.98716375e-02 1.36179343e-01
1.95128530e-01 3.90420943e-01 -7.47911274e-01 -2.67184824e-01
2.44312510e-02 -8.26802999e-02 2.52119452e-01 5.60092747e-01
1.50616407e+00 -5.94763815e-01 -1.31473482e-01 7.16455340e-01
9.93474662e-01 3.21892947e-01 8.36443961e-01 1.56098545e-01
6.71094418e-01 7.74925590e-01 7.23247826e-01 3.44380677e-01
1.81629807e-01 1.17024541e+00 4.36229765e-01 -3.25688094e-01
1.69052780e-01 -6.29464239e-02 1.49922147e-01 5.98185003e-01
-3.99079412e-01 2.61021048e-01 -9.74906683e-01 1.23478167e-01
-1.77889359e+00 -8.64624798e-01 2.98709214e-01 2.37336469e+00
6.97234035e-01 -1.37093201e-01 2.97872365e-01 2.17569754e-01
9.75732088e-01 -3.48993316e-02 -5.49326718e-01 -2.15069816e-01
2.29048997e-01 -1.90451398e-01 -2.83496361e-02 4.79902387e-01
-8.69431794e-01 8.77414823e-01 5.54461384e+00 9.25964952e-01
-1.45491445e+00 -9.26551148e-02 9.04918849e-01 2.38987301e-02
1.01885080e-01 -4.00622308e-01 -1.10497236e+00 6.34859443e-01
2.60294527e-01 1.87954248e-03 4.82397825e-01 1.14183712e+00
3.63767415e-01 3.03393424e-01 -1.07753813e+00 1.20951474e+00
2.75870800e-01 -1.06346476e+00 2.26027876e-01 -1.11002393e-01
4.79971319e-01 -5.19006491e-01 3.57714355e-01 -2.87356088e-04
-2.54621446e-01 -9.85180855e-01 6.10195518e-01 7.40798414e-01
6.55068457e-01 -8.69758487e-01 6.58782959e-01 2.73219615e-01
-1.10217869e+00 -1.42051280e-01 -5.87520003e-01 3.82131398e-01
-1.69591427e-01 2.48760030e-01 -8.74726653e-01 5.41501269e-02
6.13400340e-01 4.39606369e-01 -6.59019172e-01 1.10594797e+00
-2.92941350e-02 8.67180452e-02 -2.76389033e-01 -4.81775627e-02
-8.30746815e-02 -4.66535687e-01 4.82284695e-01 5.39083779e-01
3.00430536e-01 3.81710939e-02 -3.61735635e-02 7.97490180e-01
-1.61471829e-01 5.72248816e-01 -3.79934311e-01 4.24031466e-01
6.86749876e-01 1.43552268e+00 -5.67608774e-01 1.26862302e-01
-7.00267255e-02 5.29809654e-01 6.47182047e-01 2.08596334e-01
-8.82109404e-01 -1.63026363e-01 5.51817656e-01 3.76085818e-01
-2.26101592e-01 -7.60082752e-02 -2.58835256e-01 -6.21020615e-01
2.76225042e-02 -7.37221003e-01 2.10783988e-01 -8.98698628e-01
-1.09075332e+00 7.16984391e-01 -1.33461788e-01 -1.23359036e+00
3.69623415e-02 -3.74314129e-01 -8.04965794e-01 7.33975053e-01
-1.12080657e+00 -8.76768470e-01 -6.92312002e-01 5.97640216e-01
4.90879864e-01 -4.12069470e-01 5.18950105e-01 3.35394353e-01
-9.77595210e-01 1.08989429e+00 -2.48127624e-01 4.47580248e-01
5.91500580e-01 -6.86087370e-01 1.94472998e-01 4.58143890e-01
-8.51262435e-02 7.51075685e-01 3.87876838e-01 -3.66865575e-01
-9.90967870e-01 -1.05250478e+00 5.45124531e-01 -6.52586818e-02
1.50855631e-01 -1.13354489e-01 -9.10377204e-01 5.36755860e-01
-3.63900572e-01 4.59349304e-02 3.79613727e-01 2.56342590e-02
-3.42309475e-01 -6.66799963e-01 -1.39969599e+00 9.46641386e-01
8.97468626e-01 -1.50512278e-01 -2.10533887e-01 2.91826725e-01
3.52257937e-01 -4.74658549e-01 -6.54457271e-01 8.56467128e-01
6.80203199e-01 -7.48858750e-01 1.09252846e+00 -3.97270888e-01
3.32083017e-01 -1.89541861e-01 1.70590982e-01 -1.24346280e+00
-4.90253747e-01 -3.97522509e-01 1.34535581e-01 1.19007492e+00
2.81311005e-01 -5.65959692e-01 8.83289158e-01 8.24982762e-01
4.85968813e-02 -1.20490587e+00 -1.06206775e+00 -2.89383352e-01
-1.93965539e-01 1.06749319e-01 8.54723394e-01 9.33850348e-01
-3.28548789e-01 -1.54376756e-02 -1.31642103e-01 1.04829028e-01
4.95940477e-01 -1.92055553e-01 8.54925454e-01 -1.27570045e+00
3.43124866e-01 -7.10972309e-01 -7.75103271e-01 -6.96077228e-01
4.34051782e-01 -7.12281823e-01 -7.59671405e-02 -1.12455702e+00
-1.87566187e-02 -7.72216320e-01 -5.32485656e-02 7.16788828e-01
-2.54267871e-01 4.67960447e-01 1.75726801e-01 3.64682078e-01
-3.31702918e-01 7.97805905e-01 1.27312112e+00 -1.41949639e-01
-4.12860096e-01 2.17677295e-01 -3.80372167e-01 1.03884518e+00
7.59647727e-01 -3.53948921e-01 -2.30748475e-01 -4.33182299e-01
2.72136241e-01 6.39310181e-02 2.44662777e-01 -8.91039550e-01
3.78396839e-01 -1.84301510e-01 7.44399726e-01 -4.68277097e-01
5.17043352e-01 -9.72473621e-01 4.57832962e-02 2.28478462e-01
-4.41193432e-01 2.50282288e-01 1.67488143e-01 2.15443492e-01
-1.94478538e-02 -1.36803061e-01 1.28744686e+00 2.03263938e-01
-2.91408122e-01 7.51098037e-01 2.61182308e-01 -1.45749480e-01
1.09765863e+00 -3.59656602e-01 -2.58073192e-02 -3.98590267e-01
-7.14764297e-01 6.70905784e-02 2.59886026e-01 5.56739688e-01
9.27382708e-01 -1.46717596e+00 -7.56467342e-01 5.55120587e-01
-1.41211033e-01 1.22876890e-01 2.16716126e-01 1.02024305e+00
-6.03184760e-01 8.16543996e-02 -4.21532065e-01 -7.59623110e-01
-1.57742071e+00 2.79887050e-01 7.76681006e-01 3.70362848e-01
-2.93418378e-01 7.86754310e-01 2.39851475e-01 -4.01416808e-01
4.98440951e-01 1.68980092e-01 -6.89128339e-01 1.04293138e-01
4.92528856e-01 5.69138944e-01 2.69509763e-01 -9.44594681e-01
-2.82227486e-01 1.18206406e+00 -7.70637244e-02 3.34923923e-01
1.19966137e+00 1.16511069e-01 -3.67174149e-01 -5.37369810e-02
1.11724675e+00 -1.28677320e-02 -1.21376479e+00 -1.21300414e-01
-2.31431633e-01 -6.90006375e-01 -3.69617753e-02 -3.84196460e-01
-1.49021733e+00 7.92849064e-01 8.28382134e-01 -3.40670556e-01
1.14888513e+00 -2.91469723e-01 3.84027690e-01 2.09273979e-01
2.97061354e-01 -9.15034056e-01 3.19520861e-01 1.93394169e-01
1.24478233e+00 -1.08399272e+00 -1.26071442e-02 -6.07849479e-01
-5.66851377e-01 1.20939815e+00 1.05852461e+00 -1.12973772e-01
7.36580789e-01 -2.97502782e-02 2.89909273e-01 -1.71119049e-01
-3.54062021e-01 1.91232607e-01 5.88299990e-01 3.89596045e-01
2.49260128e-01 -2.05810264e-01 -2.49407724e-01 2.77464330e-01
-3.81322324e-01 -2.91981012e-01 5.56042511e-03 3.75419140e-01
-3.23370755e-01 -5.91463447e-01 -5.94565272e-01 2.27294996e-01
-2.28663966e-01 2.95102894e-01 -3.71564835e-01 5.69141567e-01
9.56826732e-02 6.71296358e-01 4.35275882e-01 -2.91027308e-01
3.39559317e-01 -1.45004749e-01 6.14929318e-01 -2.91163594e-01
-1.88869014e-01 -3.61693278e-02 -5.62579334e-01 -4.51948285e-01
-2.36796081e-01 -4.02156264e-01 -1.17884552e+00 -5.03852725e-01
-5.28422058e-01 6.71388209e-02 7.94481337e-01 8.04594874e-01
4.03564572e-01 -3.48373614e-02 1.06776059e+00 -1.04038537e+00
-3.77535403e-01 -8.03556621e-01 -1.86483830e-01 4.86330569e-01
6.81762695e-02 -9.56714213e-01 -4.51098621e-01 -1.86912015e-01]
|
[13.466218948364258, 0.4248078167438507]
|
827787b5-495c-4f53-838e-365dfaae1479
|
learning-cross-lingual-word-embeddings-via
| null | null |
https://aclanthology.org/P15-2093
|
https://aclanthology.org/P15-2093.pdf
|
Learning Cross-lingual Word Embeddings via Matrix Co-factorization
| null |
['Yang Liu', 'Tianze Shi', 'Maosong Sun', 'Zhiyuan Liu']
|
2015-07-01
|
learning-cross-lingual-word-embeddings-via-1
|
https://aclanthology.org/P15-2093
|
https://aclanthology.org/P15-2093.pdf
|
ijcnlp-2015-7
|
['cross-lingual-document-classification']
|
['natural-language-processing']
|
[-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01]
|
[-7.241291046142578, 3.901024103164673]
|
4ab66947-3525-42df-849f-83a53ac18c9b
|
node-classification-on-graphs-with-few-shot
|
2007.02914
| null |
https://arxiv.org/abs/2007.02914v2
|
https://arxiv.org/pdf/2007.02914v2.pdf
|
Node Classification on Graphs with Few-Shot Novel Labels via Meta Transformed Network Embedding
|
We study the problem of node classification on graphs with few-shot novel labels, which has two distinctive properties: (1) There are novel labels to emerge in the graph; (2) The novel labels have only a few representative nodes for training a classifier. The study of this problem is instructive and corresponds to many applications such as recommendations for newly formed groups with only a few users in online social networks. To cope with this problem, we propose a novel Meta Transformed Network Embedding framework (MetaTNE), which consists of three modules: (1) A \emph{structural module} provides each node a latent representation according to the graph structure. (2) A \emph{meta-learning module} captures the relationships between the graph structure and the node labels as prior knowledge in a meta-learning manner. Additionally, we introduce an \emph{embedding transformation function} that remedies the deficiency of the straightforward use of meta-learning. Inherently, the meta-learned prior knowledge can be used to facilitate the learning of few-shot novel labels. (3) An \emph{optimization module} employs a simple yet effective scheduling strategy to train the above two modules with a balance between graph structure learning and meta-learning. Experiments on four real-world datasets show that MetaTNE brings a huge improvement over the state-of-the-art methods.
|
['Xiaohong Guan', 'Lin Lan', 'Xuefeng Du', 'Pinghui Wang', 'Kaikai Song', 'Jing Tao']
|
2020-07-06
| null |
http://proceedings.neurips.cc/paper/2020/hash/c055dcc749c2632fd4dd806301f05ba6-Abstract.html
|
http://proceedings.neurips.cc/paper/2020/file/c055dcc749c2632fd4dd806301f05ba6-Paper.pdf
|
neurips-2020-12
|
['graph-structure-learning']
|
['graphs']
|
[ 3.10823262e-01 5.57390511e-01 -5.43782115e-01 -1.92807391e-01
-1.35035500e-01 -9.14787054e-02 6.18021190e-01 3.80613416e-01
-6.35816306e-02 4.86981928e-01 -1.42364502e-01 -1.11998603e-01
-3.04810166e-01 -1.28531063e+00 -3.32405984e-01 -7.80937850e-01
-2.51358211e-01 2.23580271e-01 3.79902720e-01 -4.63668257e-01
1.16136502e-02 1.84395328e-01 -1.59248590e+00 -1.31410575e-02
9.85318601e-01 8.04625452e-01 2.90492117e-01 4.20970619e-01
-4.54492271e-01 7.99431980e-01 -3.49435329e-01 -5.21938741e-01
2.53158331e-01 -6.10619724e-01 -8.36309016e-01 3.24462026e-01
-3.87107558e-03 4.85896431e-02 -5.52780509e-01 1.19502437e+00
3.51200372e-01 3.86817276e-01 7.68440664e-01 -1.54985440e+00
-6.16818607e-01 8.28998089e-01 -6.29762828e-01 1.98420435e-01
9.57861245e-02 -3.47574092e-02 1.31463361e+00 -6.58449233e-01
7.35266685e-01 1.06482160e+00 6.80997312e-01 5.92976213e-01
-1.08656847e+00 -4.06855077e-01 4.31107223e-01 2.28696808e-01
-1.33895957e+00 -1.25925541e-01 1.06192172e+00 -3.92475694e-01
5.03152847e-01 1.21119842e-01 6.76494420e-01 9.37406003e-01
9.18699503e-02 4.50527281e-01 8.19659650e-01 -4.96917933e-01
3.28308374e-01 4.91481572e-01 3.30720335e-01 1.14640856e+00
2.63949893e-02 -2.47685969e-01 -2.01996058e-01 -2.35124677e-01
4.12151903e-01 3.54482651e-01 3.12993936e-02 -4.50216442e-01
-9.96029735e-01 9.56622601e-01 5.98453581e-01 6.46648049e-01
-1.61183491e-01 1.33140646e-02 5.14059007e-01 5.93671441e-01
8.80694509e-01 3.44317824e-01 -2.33889952e-01 3.52408469e-01
-4.48949993e-01 -3.66615176e-01 9.98565316e-01 1.11041391e+00
1.27764821e+00 1.58457026e-01 -1.37342498e-01 1.00472128e+00
2.37489790e-01 7.83439130e-02 4.97974902e-01 -5.97836018e-01
3.07407141e-01 9.20130968e-01 -3.64096254e-01 -1.29053402e+00
-5.11718392e-01 -7.57783115e-01 -1.03595018e+00 -1.18904918e-01
-1.35017708e-02 -1.32046074e-01 -6.91907227e-01 1.85560369e+00
5.39204121e-01 6.93021238e-01 -2.67408729e-01 1.72933683e-01
1.05557656e+00 7.26971924e-01 1.03065446e-01 -5.72970212e-01
1.14556444e+00 -1.33556664e+00 -6.72701597e-01 -1.44990206e-01
9.46278214e-01 -2.62139767e-01 8.98283720e-01 -1.15532890e-01
-7.05652773e-01 -6.74888194e-01 -1.08530474e+00 2.60744393e-01
-7.51058042e-01 -3.12894911e-01 7.05655634e-01 7.28636920e-01
-1.12983775e+00 8.50883842e-01 -2.75813878e-01 -6.88211322e-01
1.61937088e-01 3.40684712e-01 -2.81863123e-01 6.14191331e-02
-1.16468418e+00 6.62061691e-01 6.79276049e-01 -2.58410752e-01
-5.67591786e-01 -6.03889763e-01 -1.08877075e+00 3.50153655e-01
8.73508513e-01 -6.80630982e-01 8.11102688e-01 -1.06874394e+00
-1.47305763e+00 8.52686048e-01 -2.32466199e-02 7.37477094e-02
1.99416786e-01 7.36371696e-01 -7.51057446e-01 1.13181733e-01
1.53281450e-01 3.68680269e-01 1.07913959e+00 -1.28396094e+00
-8.56992185e-01 -2.46370211e-01 3.28257829e-01 3.21103185e-02
-1.05691135e+00 -4.17042106e-01 -4.42056894e-01 -8.60435545e-01
1.06470317e-01 -9.10281837e-01 -2.17675686e-01 -2.29095012e-01
-3.12351078e-01 -5.91625035e-01 8.36595416e-01 -2.21080855e-01
1.61708450e+00 -2.11228013e+00 3.54027241e-01 4.10214245e-01
8.25188577e-01 2.81114757e-01 -4.11857814e-01 8.02003920e-01
-1.18851840e-01 2.57289141e-01 -2.58562677e-02 -3.55629951e-01
1.04281008e-02 1.31770089e-01 4.10356261e-02 2.78869182e-01
-1.18966557e-01 9.72662449e-01 -1.39009547e+00 -6.59514606e-01
2.43621022e-01 6.73542917e-02 -1.91366971e-01 2.55095333e-01
-1.04460001e-01 1.87700823e-01 -4.52652246e-01 6.51978552e-01
4.98258173e-01 -6.96473360e-01 4.98844415e-01 -1.63650066e-01
1.56630471e-01 -2.27968082e-01 -1.23958910e+00 1.49819565e+00
-4.07219887e-01 5.34907058e-02 -5.39703034e-02 -1.23798156e+00
9.32550848e-01 1.75085232e-01 6.76038861e-01 -2.23525494e-01
3.71245831e-01 4.65817861e-02 -1.61385447e-01 -5.12699485e-01
2.57719994e-01 -5.10825589e-02 6.87361434e-02 7.45580018e-01
4.84410882e-01 3.19707692e-01 4.85659719e-01 4.76017445e-01
1.38431656e+00 -3.83263588e-01 5.70456147e-01 -2.26858959e-01
6.37025177e-01 -3.30555111e-01 7.51832962e-01 8.07083488e-01
-2.22957551e-01 7.52109569e-03 5.43593824e-01 -4.76619601e-01
-7.32318997e-01 -8.32761884e-01 2.47982383e-01 1.58940434e+00
2.03157857e-01 -8.58854294e-01 -4.81380403e-01 -1.05770922e+00
-6.89843521e-02 4.90801483e-01 -1.01878273e+00 -5.43275416e-01
-3.03059459e-01 -6.82071745e-01 -4.15256023e-02 3.61643434e-01
4.30977255e-01 -1.00289237e+00 -1.35100354e-02 2.17375398e-01
1.11517392e-01 -9.46813703e-01 -5.10149121e-01 8.33151713e-02
-1.04758263e+00 -1.05119550e+00 -4.24308062e-01 -1.12337565e+00
1.02806628e+00 7.51180530e-01 8.64520133e-01 5.44467926e-01
-1.69093370e-01 5.55893362e-01 -7.36789107e-01 -5.17106391e-02
-5.36916196e-01 5.03192544e-01 2.08107326e-02 3.02525073e-01
2.35690743e-01 -1.02071810e+00 -2.88993627e-01 2.57211924e-01
-8.48162591e-01 1.11464180e-01 6.04538083e-01 9.97947276e-01
2.49987394e-01 4.95317638e-01 8.69564116e-01 -1.63552725e+00
6.47138298e-01 -8.89837801e-01 -3.41633350e-01 5.34391701e-01
-1.02331281e+00 -2.12685168e-02 8.86041641e-01 -5.66509247e-01
-8.61807466e-01 -1.29260913e-01 1.26224026e-01 -2.08033562e-01
2.00017393e-01 7.14224756e-01 -1.22738548e-01 -2.81245828e-01
4.24190342e-01 9.49596912e-02 1.34518415e-01 -4.68694717e-01
6.05719507e-01 5.96583307e-01 1.48548797e-01 -4.01371777e-01
1.18640399e+00 3.82396907e-01 2.69675016e-01 -8.57005954e-01
-9.62801754e-01 -5.90616167e-01 -8.88991356e-01 -3.93648773e-01
4.67725992e-01 -5.55737853e-01 -6.97566152e-01 2.78630197e-01
-7.19803870e-01 -2.85212636e-01 -5.29528201e-01 2.51263380e-01
-3.85375023e-01 5.62766910e-01 -7.49362409e-01 -6.27079427e-01
-3.07897836e-01 -8.12808752e-01 5.67555189e-01 3.68277073e-01
1.63088143e-01 -1.55441356e+00 1.20092280e-01 1.14315391e-01
2.98660994e-01 3.68129551e-01 1.41588199e+00 -7.91615009e-01
-4.06118095e-01 -1.99398726e-01 -1.82591885e-01 1.53982088e-01
5.06104767e-01 -7.63660446e-02 -9.68868554e-01 -8.27875793e-01
-2.17620209e-01 -1.79061010e-01 6.97779059e-01 -3.77793461e-02
1.10732746e+00 -2.43093386e-01 -6.41759694e-01 4.94001865e-01
1.32758915e+00 -1.30790714e-02 1.21108785e-01 1.10778682e-01
8.91195774e-01 5.91713667e-01 4.41450417e-01 5.24800599e-01
5.77105224e-01 5.58048010e-01 3.39344978e-01 -2.93654706e-02
-1.70008585e-01 -4.67948467e-01 2.18956798e-01 1.48849118e+00
8.95350147e-03 -2.14441687e-01 -6.85997367e-01 3.47497106e-01
-2.16087794e+00 -8.17916751e-01 2.30931103e-01 2.06256032e+00
5.19805491e-01 1.58906832e-01 1.70133173e-01 1.51577666e-01
1.06761181e+00 5.21704853e-01 -6.11012220e-01 -1.47983491e-01
2.77323961e-01 -6.98130392e-03 1.87736154e-01 2.61203200e-01
-1.08591199e+00 9.21661139e-01 5.55367565e+00 1.02206731e+00
-8.83511841e-01 4.61957783e-01 2.70058960e-01 5.11871040e-01
-3.83735210e-01 2.38337919e-01 -7.35734224e-01 4.23689812e-01
8.08173299e-01 -4.57087427e-01 4.43152905e-01 9.57038522e-01
-2.70282388e-01 2.99845606e-01 -1.16914618e+00 8.32736731e-01
3.80974770e-01 -1.28517592e+00 1.15006611e-01 1.97282687e-01
9.51372206e-01 -1.67494297e-01 -1.58468157e-01 7.91646540e-01
2.48128891e-01 -5.21368265e-01 1.80626944e-01 5.24065197e-01
7.19878018e-01 -7.05897927e-01 6.27755344e-01 6.67894900e-01
-1.63868380e+00 -3.74452829e-01 -5.84607124e-01 5.42453192e-02
-5.67139834e-02 5.90978742e-01 -8.45245898e-01 9.47465420e-01
3.64187330e-01 1.12275743e+00 -8.36266994e-01 8.54216516e-01
-3.19302052e-01 3.66660506e-01 1.28503889e-01 -2.35301688e-01
1.93812147e-01 -3.21632981e-01 6.09523773e-01 8.52992833e-01
2.03094482e-01 7.14308023e-02 5.97773254e-01 5.44353843e-01
-4.08607483e-01 3.51263463e-01 -7.37921059e-01 -1.38575792e-01
6.03885889e-01 1.71165705e+00 -1.07750905e+00 -3.51745963e-01
-6.35907888e-01 8.54475081e-01 7.16592491e-01 2.65765935e-01
-4.68718439e-01 -5.22053480e-01 2.32244134e-02 1.21856056e-01
1.87311664e-01 4.41189557e-02 1.52360812e-01 -1.25252676e+00
-2.66598582e-01 -5.96129835e-01 6.67486310e-01 -3.48006219e-01
-1.57988620e+00 5.99309444e-01 5.80707341e-02 -1.23333728e+00
-7.09168389e-02 -4.64190483e-01 -8.85222554e-01 3.31744313e-01
-1.51259887e+00 -1.26812196e+00 -6.08156085e-01 4.23658669e-01
4.94482338e-01 -4.65791404e-01 9.75666404e-01 3.34087878e-01
-8.83264184e-01 7.39650548e-01 1.67394325e-01 1.38023630e-01
5.45135140e-01 -1.25188851e+00 2.73271203e-01 6.99360073e-01
2.52562940e-01 4.88475472e-01 3.60308647e-01 -6.99047923e-01
-1.35375774e+00 -1.26513696e+00 8.32714379e-01 -1.79709151e-01
6.99377418e-01 -4.53750134e-01 -7.95597613e-01 7.45529354e-01
8.90450627e-02 2.54939854e-01 9.19190884e-01 4.06618595e-01
-3.00826520e-01 -4.14055824e-01 -1.10147607e+00 5.38295865e-01
1.25370467e+00 -4.79734957e-01 -5.04789591e-01 6.27221942e-01
9.08219695e-01 1.88679799e-01 -1.03251386e+00 5.64788461e-01
2.34171778e-01 -7.07486212e-01 6.68252409e-01 -7.22080052e-01
7.33912662e-02 -9.97347385e-02 1.85025588e-01 -1.58156908e+00
-8.79090011e-01 -7.24506974e-01 -6.77174509e-01 1.50846422e+00
3.16846550e-01 -7.43501425e-01 8.37470293e-01 3.12292367e-01
-2.53533553e-02 -9.51275170e-01 -6.47847950e-01 -8.18065047e-01
-4.00089234e-01 8.96419361e-02 5.79603195e-01 1.38160956e+00
2.92748094e-01 6.32873535e-01 -4.83001590e-01 -1.30059615e-01
6.61073208e-01 1.41452178e-01 7.80465066e-01 -1.68607998e+00
-3.04621041e-01 -3.59892309e-01 -4.40709442e-01 -7.43036807e-01
5.44075370e-01 -1.31683278e+00 -2.99109519e-01 -1.59113741e+00
5.00519931e-01 -6.50659263e-01 -6.91206217e-01 5.87678850e-01
-3.06474209e-01 -1.04478598e-01 1.39690906e-01 1.48330659e-01
-9.26043749e-01 7.04185724e-01 1.09584033e+00 -3.06214064e-01
-2.67251134e-01 3.07501078e-01 -7.88883507e-01 6.86113000e-01
6.18191361e-01 -7.23035395e-01 -7.37581670e-01 7.26429522e-02
3.78970981e-01 -1.53879702e-01 -1.28967807e-01 -8.60677063e-01
4.56606388e-01 -7.54908621e-02 -6.84686080e-02 -2.93047011e-01
2.43834615e-01 -9.39914882e-01 -8.02971721e-02 5.19587815e-01
-1.38981655e-01 -8.24970156e-02 -3.96742284e-01 9.61666405e-01
1.01105928e-01 -6.02497458e-01 8.24128747e-01 -3.60928178e-01
-9.83213544e-01 6.37081504e-01 -1.21969834e-01 1.51776403e-01
1.23087883e+00 -1.56980500e-01 -3.64609390e-01 -3.19372147e-01
-7.93870866e-01 4.22263771e-01 2.61591822e-01 5.48572242e-01
4.06747490e-01 -1.45991445e+00 -4.35885578e-01 1.54637679e-01
3.81343752e-01 -2.72612274e-01 3.16674948e-01 7.55955756e-01
-5.56738488e-02 -1.31330848e-01 -9.05426517e-02 -3.86046231e-01
-9.79246259e-01 8.26336503e-01 4.71702665e-02 -4.48160350e-01
-6.94019496e-01 7.09914505e-01 -1.24048330e-01 -6.35190547e-01
2.90310115e-01 1.72378808e-01 -4.91232008e-01 4.98205543e-01
3.80023777e-01 7.07106650e-01 -2.21463159e-01 -6.38597667e-01
-7.99782351e-02 4.89569277e-01 -2.79880702e-01 4.59200680e-01
1.50589097e+00 -3.56982589e-01 -3.75570089e-01 8.31914723e-01
1.22278655e+00 -1.53243825e-01 -8.87130201e-01 -7.67079532e-01
1.14959277e-01 -3.83038729e-01 -8.58804956e-02 -2.62407184e-01
-1.03929245e+00 5.50800622e-01 2.86601692e-01 5.90974331e-01
9.19472933e-01 -1.92512814e-02 7.34941185e-01 6.19650483e-01
4.88941401e-01 -1.34336579e+00 4.37156677e-01 4.31491405e-01
2.48103485e-01 -1.23716068e+00 1.70657426e-01 -7.89384902e-01
-3.95834565e-01 1.16935039e+00 7.29832172e-01 3.29580940e-02
1.13165486e+00 -2.92290658e-01 -3.17117900e-01 -3.55693638e-01
-7.15366662e-01 -3.17110866e-01 3.67772579e-01 3.32970500e-01
-1.10171819e-02 -6.51732460e-02 -4.52392250e-01 6.06329620e-01
2.65706837e-01 -2.91135907e-01 4.08740401e-01 9.15492833e-01
-7.18304455e-01 -1.30461192e+00 3.74641538e-01 6.85504794e-01
1.69285551e-01 1.13528341e-01 -2.27490291e-01 6.88790560e-01
3.59872818e-01 9.91613507e-01 -3.37126017e-01 -7.53983557e-01
2.02603027e-01 2.18648612e-01 1.86891064e-01 -1.22570610e+00
-5.63372910e-01 -1.37463927e-01 -9.50412303e-02 -2.59711325e-01
-4.99554962e-01 -1.69892132e-01 -8.89425755e-01 -3.45840842e-01
-6.25436306e-01 3.62858206e-01 3.17093939e-01 9.88721550e-01
3.56743306e-01 7.37884641e-01 1.27695882e+00 -6.00306630e-01
-6.34084702e-01 -1.00655890e+00 -1.10851765e+00 4.36078161e-01
-4.77887169e-02 -7.77336121e-01 -6.30913317e-01 -2.28876233e-01]
|
[7.378446102142334, 6.178509712219238]
|
2cbb7a0d-0746-43b5-9b92-fd029a39b093
|
a-robust-multimodal-remote-sensing-image
|
2202.13347
| null |
https://arxiv.org/abs/2202.13347v1
|
https://arxiv.org/pdf/2202.13347v1.pdf
|
A Robust Multimodal Remote Sensing Image Registration Method and System Using Steerable Filters with First- and Second-order Gradients
|
Co-registration of multimodal remote sensing images is still an ongoing challenge because of nonlinear radiometric differences (NRD) and significant geometric distortions (e.g., scale and rotation changes) between these images. In this paper, a robust matching method based on the Steerable filters is proposed consisting of two critical steps. First, to address severe NRD, a novel structural descriptor named the Steerable Filters of first- and second-Order Channels (SFOC) is constructed, which combines the first- and second-order gradient information by using the steerable filters with a multi-scale strategy to depict more discriminative structure features of images. Then, a fast similarity measure is established called Fast Normalized Cross-Correlation (Fast-NCCSFOC), which employs the Fast Fourier Transform technique and the integral image to improve the matching efficiency. Furthermore, to achieve reliable registration performance, a coarse-to-fine multimodal registration system is designed consisting of two pivotal modules. The local coarse registration is first conducted by involving both detection of interest points (IPs) and local geometric correction, which effectively utilizes the prior georeferencing information of RS images to address global geometric distortions. In the fine registration stage, the proposed SFOC is used to resist significant NRD, and to detect control points between multimodal images by a template matching scheme. The performance of the proposed matching method has been evaluated with many different kinds of multimodal RS images. The results show its superior matching performance compared with the state-of-the-art methods. Moreover, the designed registration system also outperforms the popular commercial software in both registration accuracy and computational efficiency. Our system is available at https://github.com/yeyuanxin110.
|
['Guo Zhang', 'Qizhi Xu', 'Chao Yang', 'Tengfeng Tang', 'Bai Zhu', 'Yuanxin Ye']
|
2022-02-27
| null | null | null | null |
['template-matching']
|
['computer-vision']
|
[ 2.90376246e-01 -6.29509568e-01 3.78324836e-01 -2.46639714e-01
-6.66006386e-01 -2.38308132e-01 6.44823790e-01 -8.16955343e-02
-5.26548088e-01 2.92398512e-01 6.21765889e-02 1.41854763e-01
-4.57733780e-01 -9.71946776e-01 -1.71056598e-01 -9.43843722e-01
-2.79094595e-02 9.05904099e-02 3.07756811e-01 -5.08623540e-01
4.33485746e-01 7.99880028e-01 -1.52590573e+00 -2.67160535e-01
1.18601382e+00 9.17630851e-01 4.09059644e-01 2.66941249e-01
9.94528830e-02 -6.95869029e-02 -1.63711399e-01 3.47741432e-02
4.14462656e-01 -2.02074036e-01 -4.99303669e-01 -1.26717597e-01
3.17015737e-01 -2.60811567e-01 -1.42297462e-01 1.25226164e+00
7.64672458e-01 3.58465016e-01 4.86653894e-01 -6.75247908e-01
-4.26694065e-01 6.12930879e-02 -8.94821942e-01 9.80021134e-02
4.44174111e-01 -6.90335184e-02 6.21136308e-01 -1.12687397e+00
3.66791338e-01 1.21176171e+00 8.52776468e-01 -6.68026060e-02
-1.09609795e+00 -7.23336518e-01 -2.81291217e-01 2.58059442e-01
-1.83988440e+00 -3.54467064e-01 8.16187263e-01 -4.65162516e-01
3.76220971e-01 5.29887617e-01 4.74778742e-01 1.89380080e-01
1.56385586e-01 -1.16199300e-01 1.29786897e+00 -4.15281355e-01
-8.25690553e-02 -2.18021244e-01 2.39226930e-02 6.23993278e-01
2.02005014e-01 2.15386972e-01 -2.73135100e-02 -3.09437871e-01
8.61443877e-01 3.11125040e-01 -6.42903626e-01 -1.31211225e-02
-1.16341448e+00 6.04512393e-01 6.66195333e-01 6.46496832e-01
-4.45077956e-01 -3.54119390e-01 2.53814328e-02 -8.84955823e-02
2.74966508e-01 2.26650685e-02 -9.99404639e-02 4.30988878e-01
-8.30120921e-01 -1.89352185e-01 3.39248657e-01 5.47928393e-01
1.23181391e+00 -1.22391231e-01 4.83686626e-02 9.53532398e-01
6.71721280e-01 9.81901824e-01 5.02328694e-01 -4.44110513e-01
4.52719450e-01 5.78133225e-01 -8.80131647e-02 -1.58574367e+00
-5.80668986e-01 -2.39232555e-01 -1.18220210e+00 2.07441617e-02
9.59382355e-02 2.54524350e-01 -6.13939643e-01 1.26187050e+00
5.13905048e-01 3.00617337e-01 -8.43592361e-03 1.07806492e+00
9.60983992e-01 7.61169851e-01 -6.97153881e-02 -3.45908999e-01
1.52879846e+00 -2.21381307e-01 -7.71676362e-01 6.13653511e-02
3.75784300e-02 -1.26897705e+00 8.10426474e-01 -1.52293429e-01
-7.59003460e-01 -8.26839745e-01 -9.93131816e-01 1.71400174e-01
-2.76645720e-01 5.00402272e-01 2.66322643e-01 4.43432868e-01
-8.34706903e-01 4.28952664e-01 -6.98783517e-01 -6.09317720e-01
-2.44892631e-02 2.91129887e-01 -4.96955037e-01 -7.25614047e-03
-1.20595634e+00 8.89224172e-01 9.07786340e-02 7.57544518e-01
-2.54385769e-01 -4.78405267e-01 -9.51291561e-01 -1.20180249e-01
-7.08039328e-02 -2.97712684e-01 3.96332651e-01 -6.29476368e-01
-1.37891042e+00 8.51898730e-01 -2.33061075e-01 3.07249904e-01
3.73523444e-01 6.04705438e-02 -7.62285709e-01 2.62318760e-01
1.93515822e-01 -4.17294428e-02 6.92320049e-01 -1.20166302e+00
-4.04353321e-01 -6.50011420e-01 -4.93145645e-01 4.28900748e-01
-1.72882065e-01 2.56428242e-01 -3.25272530e-01 -5.65371394e-01
8.80784631e-01 -8.56489778e-01 -1.42855287e-01 -1.83030427e-01
-1.39263108e-01 1.55849189e-01 8.02040040e-01 -9.54586923e-01
1.22413492e+00 -2.33981228e+00 -1.78711012e-01 8.16507280e-01
-8.95276368e-02 3.62459838e-01 -1.64570034e-01 3.93526703e-01
-9.04095918e-02 -2.20191389e-01 -3.64007980e-01 1.32494569e-01
-4.09622401e-01 -2.17135280e-01 6.65536523e-02 9.81538594e-01
-1.25327006e-01 4.38354850e-01 -7.52656102e-01 -4.83429253e-01
4.53345209e-01 6.66929245e-01 8.04846436e-02 1.08802184e-01
6.64743423e-01 6.96867645e-01 -5.50078630e-01 8.26670587e-01
1.41236830e+00 1.29146799e-01 1.96647812e-02 -8.71574461e-01
-6.39410198e-01 -2.50087529e-01 -1.56395209e+00 1.32552958e+00
-2.94530928e-01 1.44973576e-01 2.98342079e-01 -8.12490284e-01
1.49619341e+00 2.61575669e-01 6.15925014e-01 -7.61968613e-01
1.55167133e-01 4.94859099e-01 -3.23470086e-01 -5.87507069e-01
4.73681837e-01 -4.47213836e-02 3.31447214e-01 -6.17532479e-03
-4.28460479e-01 -1.05604708e-01 -6.06639199e-02 -2.33780339e-01
5.15244603e-01 6.75855801e-02 5.09353220e-01 -4.46594328e-01
1.27365673e+00 -1.27545521e-01 5.73222518e-01 3.19389194e-01
-4.96728383e-02 5.68872094e-01 -4.03816253e-01 -4.57964659e-01
-7.85273254e-01 -9.20129538e-01 -5.24556577e-01 4.71898586e-01
7.17856526e-01 4.86983284e-02 -4.80197519e-01 -7.00108111e-02
2.88895350e-02 -2.86363289e-02 -3.27436328e-01 6.49126396e-02
-5.84305048e-01 -1.05926394e+00 3.73077661e-01 6.36669323e-02
1.03784490e+00 -6.45813286e-01 -3.24715495e-01 1.64464280e-01
-2.36106604e-01 -8.48960519e-01 -5.17206252e-01 -6.06595278e-01
-1.12987792e+00 -1.22494721e+00 -7.83834159e-01 -8.04784119e-01
8.48650694e-01 8.75045776e-01 5.33777893e-01 2.50550598e-01
-2.85398483e-01 1.17881194e-01 -4.09356654e-01 1.86598882e-01
-5.97257726e-02 -1.61258310e-01 -2.59231422e-02 4.90267187e-01
2.45685995e-01 -6.60355687e-01 -8.09478939e-01 9.61049676e-01
-9.91513491e-01 -2.37033799e-01 8.16024423e-01 7.76808560e-01
6.98644638e-01 8.74432698e-02 1.53004795e-01 -2.45367035e-01
4.35599148e-01 -7.23632202e-02 -9.16791916e-01 3.79351467e-01
-4.53558028e-01 -2.73495883e-01 3.99748296e-01 -2.08850384e-01
-1.18001521e+00 1.59162477e-01 -8.74551758e-02 8.83870721e-02
-2.34990045e-01 6.00435555e-01 -2.53233463e-01 -7.04012752e-01
5.67619741e-01 5.14563382e-01 3.16130459e-01 -6.34519160e-01
9.91653278e-02 9.54109967e-01 6.49916410e-01 -4.29435730e-01
1.23642683e+00 7.96506882e-01 1.97439209e-01 -1.10813761e+00
-2.84589887e-01 -7.44833589e-01 -7.04345644e-01 -3.90709311e-01
7.94337213e-01 -1.04847383e+00 -8.31107557e-01 8.49371910e-01
-1.09911191e+00 2.96761274e-01 2.83404350e-01 9.92292881e-01
-7.14364573e-02 7.20767736e-01 -5.14936328e-01 -5.83036423e-01
-5.82481980e-01 -1.23428237e+00 9.68190789e-01 6.12675488e-01
3.43717307e-01 -8.22138309e-01 2.05083698e-01 2.86144555e-01
7.02638924e-01 2.53837794e-01 3.98232967e-01 -2.16129735e-01
-5.42942584e-01 -1.45849138e-01 -3.98824453e-01 1.75537899e-01
3.90513897e-01 1.66046575e-01 -7.51684070e-01 -3.55877519e-01
1.14448898e-01 3.23477477e-01 4.42284465e-01 3.35545868e-01
5.15283942e-01 3.48068662e-02 -3.10420036e-01 8.51143122e-01
1.64647913e+00 2.62127578e-01 1.02528715e+00 6.37608826e-01
6.38859987e-01 5.31686127e-01 8.79661500e-01 4.15007770e-01
4.21169996e-01 7.37042844e-01 3.21055353e-01 -4.93021131e-01
1.50050148e-01 4.39736173e-02 2.44909182e-01 9.88905609e-01
-7.72616863e-01 5.01312196e-01 -7.38977790e-01 2.21646562e-01
-1.74309218e+00 -1.11935735e+00 -6.48441911e-01 2.47681284e+00
4.79976743e-01 -6.68707550e-01 -4.01673973e-01 9.18385945e-03
1.21943760e+00 2.25982100e-01 -1.14062298e-02 1.59851015e-01
-3.95471632e-01 1.69258237e-01 6.76059663e-01 6.29520655e-01
-1.21313190e+00 5.41953862e-01 5.01607800e+00 7.84427166e-01
-1.31158864e+00 6.03871085e-02 9.24413651e-02 7.67924011e-01
-1.61735937e-01 1.80123776e-01 -6.64044201e-01 4.18289095e-01
2.28346854e-01 2.19561875e-01 3.84637386e-01 4.67751324e-01
5.18653631e-01 -2.41547853e-01 -1.61450446e-01 1.03924835e+00
2.84649264e-02 -9.36887443e-01 -1.02222718e-01 5.75214475e-02
5.40765584e-01 -4.90874685e-02 -2.09410787e-01 -3.69454831e-01
-3.67746890e-01 -5.89493692e-01 4.13584799e-01 8.69477391e-01
8.87198150e-01 -5.65069437e-01 9.87323642e-01 -3.36937606e-02
-1.85652113e+00 1.75047651e-01 -5.56052387e-01 1.77021757e-01
1.29212156e-01 8.18285525e-01 -2.60160714e-01 1.11414850e+00
6.61332905e-01 6.63842797e-01 -5.10436356e-01 1.17209446e+00
-1.84102267e-01 5.75534999e-02 -3.88738662e-01 3.10418963e-01
-5.59021644e-02 -9.42467332e-01 5.75837076e-01 1.04131198e+00
7.73471415e-01 3.45697463e-01 1.21458665e-01 5.84364772e-01
4.24250722e-01 5.56893289e-01 -3.47542018e-01 5.55519044e-01
6.59225285e-01 1.61536098e+00 -5.55562675e-01 -6.81970865e-02
-4.46499974e-01 7.25814402e-01 -3.57125193e-01 2.96339482e-01
-6.50775313e-01 -6.20720804e-01 3.77188772e-01 6.91299662e-02
5.92191517e-02 -3.10892314e-01 -1.19545907e-01 -1.26371515e+00
9.22788754e-02 -7.08456635e-01 4.13664907e-01 -7.15211987e-01
-1.14492583e+00 6.08746469e-01 -1.31722882e-01 -1.61158037e+00
4.36101019e-01 -2.94925570e-01 -7.76036203e-01 1.33346331e+00
-1.73504329e+00 -1.32723856e+00 -9.77495372e-01 9.03988898e-01
-5.67230135e-02 -7.72253936e-03 7.59567201e-01 6.86863005e-01
-4.74691004e-01 3.34546417e-01 3.55660051e-01 7.95653909e-02
9.00778592e-01 -5.72802424e-01 -1.04065984e-01 9.44114268e-01
-4.30616438e-01 7.84577847e-01 3.92860711e-01 -7.59610236e-01
-1.38798118e+00 -7.93642938e-01 8.49031389e-01 2.65232861e-01
5.27408898e-01 9.87275317e-02 -9.64982212e-01 4.08033043e-01
-1.82655707e-01 2.08467573e-01 4.76867676e-01 -3.76406312e-01
-1.32233083e-01 -6.73622847e-01 -1.34078133e+00 2.90995598e-01
6.27741337e-01 -5.01162171e-01 -4.69590545e-01 1.25445768e-01
8.60266760e-02 -3.35140049e-01 -1.23511624e+00 7.53741920e-01
7.47768164e-01 -9.17641461e-01 1.14129615e+00 4.01778638e-01
-1.20481603e-01 -1.07861483e+00 -2.60007739e-01 -1.00655127e+00
-5.37380636e-01 -4.82497156e-01 7.02446938e-01 1.44822061e+00
3.91689241e-02 -1.04680049e+00 2.33220369e-01 2.70684361e-01
-8.31049606e-02 -1.36267602e-01 -8.98772836e-01 -5.92310667e-01
-7.40550756e-01 2.30708808e-01 7.17701793e-01 1.11325622e+00
-2.25507632e-01 -6.52306974e-02 -2.39433274e-01 7.49321401e-01
8.58575046e-01 2.31096163e-01 6.98092222e-01 -1.37102985e+00
1.24792449e-01 -2.11105704e-01 -6.84336662e-01 -8.08177710e-01
-2.83704102e-01 -6.37037873e-01 8.16593021e-02 -1.31680608e+00
1.14438899e-01 -6.91054702e-01 -2.13418677e-01 1.04835667e-01
-2.67506391e-01 4.71868843e-01 -2.25543585e-02 7.64380276e-01
-2.69413479e-02 5.30968547e-01 1.27582419e+00 1.95212662e-02
-3.69918197e-01 8.49951953e-02 -2.86396027e-01 5.87329865e-01
7.05006838e-01 -1.48333356e-01 6.26360625e-02 -2.60424852e-01
-1.86375789e-02 -7.58606801e-03 5.24504900e-01 -1.09021294e+00
3.91565263e-01 -2.08308205e-01 3.14877868e-01 -6.93076491e-01
1.59330413e-01 -9.85483110e-01 6.86342180e-01 6.28704011e-01
2.90209830e-01 1.62388161e-01 -7.39069954e-02 3.08650017e-01
-5.59706748e-01 -8.69538933e-02 1.03333819e+00 8.47707540e-02
-7.91925251e-01 3.29890281e-01 7.34544769e-02 -5.79346180e-01
9.24513459e-01 -3.17107767e-01 -5.02329886e-01 -1.61229093e-02
-4.89298820e-01 -6.26668185e-02 5.43031514e-01 -8.39907080e-02
7.63090968e-01 -1.47041392e+00 -8.55068505e-01 4.24028695e-01
2.53246188e-01 -1.30088657e-01 6.64784551e-01 1.36685181e+00
-9.70693290e-01 -1.07295081e-01 -3.61913532e-01 -6.37478292e-01
-1.46887755e+00 -1.30322827e-02 5.16806424e-01 -7.99636692e-02
-3.55639666e-01 3.64309251e-01 -7.77717158e-02 -5.68919957e-01
-2.75629371e-01 1.20191798e-02 -5.50698221e-01 2.54748881e-01
5.96635342e-01 6.70203209e-01 2.56661057e-01 -1.21063280e+00
-6.18146479e-01 1.47731364e+00 3.20598722e-01 5.78442514e-02
1.39311433e+00 -4.19639319e-01 -7.16163397e-01 -1.92739844e-01
1.15231526e+00 3.41134310e-01 -9.48284268e-01 -3.93498182e-01
-2.83777595e-01 -8.31125796e-01 1.12388311e-02 -3.79051834e-01
-1.10642421e+00 6.61497593e-01 1.05270684e+00 1.10091716e-01
1.39986587e+00 -3.37333649e-01 6.25012815e-01 7.97678307e-02
3.59887272e-01 -9.30760026e-01 -4.34790164e-01 4.66748476e-01
9.16955173e-01 -1.20014870e+00 3.78618538e-01 -6.23720646e-01
-2.18425244e-01 1.41729021e+00 2.27599621e-01 -7.82123953e-02
6.74565732e-01 -1.67211920e-01 3.83789629e-01 -2.09758744e-01
4.56185967e-01 -3.59775841e-01 4.87554580e-01 4.73387063e-01
3.94180566e-01 2.47570733e-03 -8.89153779e-01 1.28681868e-01
1.04372971e-01 -1.60523728e-01 1.81365684e-01 6.53125525e-01
-5.71188390e-01 -1.00782335e+00 -1.09280384e+00 8.20336491e-02
-1.66382313e-01 -1.08349353e-01 9.29194689e-02 6.83155596e-01
1.63000956e-01 1.00487053e+00 -4.69915904e-02 -5.94578922e-01
5.46283782e-01 -4.70737070e-01 1.45816728e-01 -3.92670110e-02
-5.32970369e-01 3.57665032e-01 -2.25336075e-01 -6.46715999e-01
-1.05450618e+00 -6.95307434e-01 -1.18733549e+00 -3.68071765e-01
-5.26573241e-01 2.48303056e-01 1.02809620e+00 7.38016486e-01
2.45612130e-01 -4.84595783e-02 1.12671089e+00 -1.12688577e+00
-4.10614759e-01 -9.01192605e-01 -7.99649119e-01 4.78406280e-01
1.82018384e-01 -5.74260354e-01 -5.71431279e-01 -5.59244072e-03]
|
[10.295793533325195, -1.83228600025177]
|
f531ce87-fd60-4835-81ee-9f50da0235ab
|
saliency-based-multiple-region-of-interest
|
2209.03656
| null |
https://arxiv.org/abs/2209.03656v1
|
https://arxiv.org/pdf/2209.03656v1.pdf
|
Saliency-based Multiple Region of Interest Detection from a Single 360° image
|
360{\deg} images are informative -- it contains omnidirectional visual information around the camera. However, the areas that cover a 360{\deg} image is much larger than the human's field of view, therefore important information in different view directions is easily overlooked. To tackle this issue, we propose a method for predicting the optimal set of Region of Interest (RoI) from a single 360{\deg} image using the visual saliency as a clue. To deal with the scarce, strongly biased training data of existing single 360{\deg} image saliency prediction dataset, we also propose a data augmentation method based on the spherical random data rotation. From the predicted saliency map and redundant candidate regions, we obtain the optimal set of RoIs considering both the saliency within a region and the Interaction-Over-Union (IoU) between regions. We conduct the subjective evaluation to show that the proposed method can select regions that properly summarize the input 360{\deg} image.
|
['Kiyoharu Aizawa', 'Satoshi Ikehata', 'Yuuki Sawabe']
|
2022-09-08
| null | null | null | null |
['saliency-prediction']
|
['computer-vision']
|
[ 2.13553995e-01 3.31590503e-01 -2.34500214e-01 -3.78500253e-01
-3.79960001e-01 -4.15001184e-01 1.91148013e-01 8.93843845e-02
-1.85616076e-01 4.92879212e-01 5.16652107e-01 -1.59226716e-01
-2.10116595e-01 -6.11714184e-01 -7.45398879e-01 -5.98714292e-01
4.04138654e-01 -1.94314599e-01 7.93753028e-01 -2.16237009e-01
5.92390358e-01 4.01129425e-01 -1.89614391e+00 4.41231094e-02
1.15962636e+00 1.10712647e+00 1.00383675e+00 1.37291476e-03
2.68800586e-01 7.35348821e-01 -6.46026671e-01 4.13923599e-02
2.28043824e-01 -1.48258731e-01 -5.36602139e-01 3.28817308e-01
3.02284330e-01 -2.47756734e-01 -1.18725337e-01 1.25462365e+00
3.92443240e-01 2.96188444e-01 4.65383470e-01 -1.04109657e+00
-3.91976923e-01 3.55570793e-01 -8.77211213e-01 5.95096171e-01
3.46352845e-01 -2.38172442e-01 7.94510067e-01 -1.04087055e+00
7.75495350e-01 8.13335121e-01 6.94217309e-02 2.06170917e-01
-5.56852877e-01 -4.24526900e-01 4.77930963e-01 2.48753577e-01
-1.45760727e+00 -2.79889315e-01 1.30647838e+00 -1.59727007e-01
2.17313260e-01 4.90343422e-01 7.34824955e-01 5.70047975e-01
-2.26132423e-02 1.04619706e+00 9.44466531e-01 -2.80724466e-01
1.50049597e-01 5.00798821e-01 -8.09793919e-02 4.96540815e-01
2.47039109e-01 -1.28261715e-01 -3.89038384e-01 2.21973464e-01
9.92878079e-01 1.85736760e-01 -5.89524806e-01 -7.60758281e-01
-1.50725830e+00 5.56587160e-01 7.61788785e-01 3.01505595e-01
-1.90507516e-01 -4.42144603e-01 4.76129092e-02 -3.45744759e-01
4.84478623e-01 5.38846254e-01 -3.75714868e-01 3.34185839e-01
-9.68391001e-01 9.50837061e-02 -1.17127690e-02 1.13708568e+00
9.86788034e-01 1.32244006e-01 -1.33119419e-01 7.82507479e-01
6.05202243e-02 4.58995879e-01 3.71185601e-01 -7.37427115e-01
5.55763662e-01 1.02148914e+00 4.21671808e-01 -1.38103664e+00
-4.94971812e-01 -7.12275505e-01 -7.70677447e-01 -3.42399217e-02
2.65331775e-01 -1.06335178e-01 -6.23362124e-01 1.61236537e+00
5.37209392e-01 -1.94799438e-01 -4.82049100e-02 1.40234435e+00
1.22443175e+00 5.55256963e-01 -3.76429677e-01 -3.97496760e-01
1.20051634e+00 -9.47430849e-01 -4.68758315e-01 -5.53324699e-01
2.49414071e-01 -8.24652135e-01 1.08434844e+00 1.99136093e-01
-9.40627515e-01 -6.68304741e-01 -9.46357429e-01 -3.05676810e-03
-1.23885497e-01 6.61323607e-01 5.66447377e-01 2.44905084e-01
-1.00055027e+00 -2.15005502e-01 -4.14998904e-02 -1.20655753e-01
2.59872794e-01 1.79875210e-01 -2.84922957e-01 -1.32568166e-01
-1.04353797e+00 7.46031582e-01 5.45357585e-01 -1.20986380e-01
-7.83308148e-01 -5.82587302e-01 -9.21128690e-01 4.80485260e-02
7.43207932e-01 -3.04389924e-01 6.82705343e-01 -1.14213443e+00
-6.46190107e-01 6.26730621e-01 -4.52089399e-01 -9.29574221e-02
1.85011476e-01 7.38067031e-02 -3.17294627e-01 3.00066888e-01
4.87186521e-01 8.10074389e-01 7.31272638e-01 -1.56438529e+00
-9.04207170e-01 -5.34714460e-01 1.57266259e-01 7.94107437e-01
-1.55290857e-01 -3.08118481e-02 -5.58013916e-01 -7.86414206e-01
6.39590800e-01 -7.18284905e-01 -2.94070303e-01 -3.50892723e-01
-6.54402852e-01 -3.26327831e-02 9.95281458e-01 -6.01297915e-01
1.34340107e+00 -2.03806996e+00 -1.47862703e-01 1.62806198e-01
1.51994973e-01 1.55497044e-01 2.86475599e-01 -1.45979881e-01
-1.42306164e-01 -7.98468292e-02 3.25991698e-02 1.08304448e-01
-7.56152570e-01 -1.62109628e-01 -1.63424149e-01 3.10507506e-01
-2.75562316e-01 6.38372123e-01 -1.02726495e+00 -7.27386177e-01
5.47157884e-01 7.56126493e-02 -4.22207624e-01 2.20667899e-01
-6.49351552e-02 7.71627128e-01 -7.29405463e-01 8.04897368e-01
9.67137754e-01 -2.68768489e-01 -2.67866611e-01 -4.70775813e-01
-4.11026686e-01 -7.14676604e-02 -1.18841553e+00 1.53704691e+00
-1.17232069e-01 5.44582129e-01 -2.35267520e-01 -8.55253756e-01
1.18085599e+00 -9.59834158e-02 7.17569649e-01 -7.05416501e-01
7.74280578e-02 4.19160798e-02 -3.06256384e-01 -4.06332135e-01
8.84051144e-01 2.57690012e-01 -1.67459503e-01 2.87109554e-01
-2.09081471e-01 2.75291270e-03 -8.64503086e-02 1.62041873e-01
5.91829181e-01 1.81551799e-02 5.18642187e-01 -3.96145076e-01
6.30796552e-01 7.80450180e-03 7.16278791e-01 4.81849700e-01
-1.47244751e-01 1.08007109e+00 3.06537181e-01 -5.69195449e-01
-9.62422788e-01 -8.96327496e-01 -2.43571736e-02 8.53716969e-01
1.10804033e+00 -8.15097690e-02 -8.99554551e-01 -6.15772605e-01
-4.45499718e-01 6.12606823e-01 -6.46567822e-01 -1.96227029e-01
-3.34983170e-01 -5.40395141e-01 -1.89351678e-01 3.18466723e-01
7.06613541e-01 -1.21066999e+00 -1.04741728e+00 -1.61411598e-01
-7.94179201e-01 -1.01592147e+00 -7.19835877e-01 -8.54497626e-02
-7.05112755e-01 -1.04543996e+00 -8.89406145e-01 -9.78181183e-01
1.10563469e+00 1.18481231e+00 1.12851691e+00 -1.81482866e-01
7.01211393e-02 9.54584830e-05 -3.96840781e-01 -4.88236725e-01
1.54494897e-01 -8.63145590e-02 -8.25675130e-02 2.29467507e-02
2.17102900e-01 -2.66637146e-01 -1.03747249e+00 7.93234348e-01
-8.35876942e-01 5.60613096e-01 6.38788223e-01 6.66443229e-01
8.19339335e-01 1.30948678e-01 6.22884154e-01 -5.40877402e-01
2.76296020e-01 -5.68657100e-01 -5.42629242e-01 3.08947682e-01
-3.40194553e-01 -1.78505257e-01 4.18285668e-01 -1.18975654e-01
-1.17729306e+00 9.16985944e-02 3.04450542e-01 -5.38252890e-01
-2.07501069e-01 2.57753700e-01 -5.68461955e-01 3.29916999e-02
5.60080826e-01 5.29386103e-01 -2.70029455e-01 -2.70863533e-01
2.30532244e-01 6.57605171e-01 5.57598114e-01 2.17481013e-02
4.04322416e-01 5.89474082e-01 -1.55738369e-01 -8.04268777e-01
-1.14540136e+00 -7.84615278e-01 -5.49835622e-01 -5.31410277e-01
6.80674314e-01 -1.06755483e+00 -6.07026041e-01 4.01758291e-02
-1.12720239e+00 3.76678556e-01 -2.04077825e-01 5.61187446e-01
-4.20033097e-01 4.51355875e-01 2.59505272e-01 -8.82792652e-01
-3.02129000e-01 -1.39916503e+00 1.24132836e+00 7.48600185e-01
2.30394062e-02 -5.11030257e-01 -6.02019250e-01 5.09806931e-01
-3.03080101e-02 1.99018851e-01 5.87625921e-01 -4.33802783e-01
-8.89860928e-01 7.43164122e-02 -5.38331449e-01 1.94932625e-01
3.30572933e-01 -2.48750389e-01 -7.87819207e-01 -1.23862587e-01
1.78509906e-01 1.27134115e-01 6.83048904e-01 8.63908172e-01
1.40971994e+00 -4.14865166e-01 -4.38254654e-01 5.14481306e-01
1.44198120e+00 4.06198233e-01 5.72709978e-01 3.65109712e-01
7.40868211e-01 6.43970191e-01 1.40254760e+00 6.55156255e-01
4.90243375e-01 7.45209634e-01 9.41451490e-01 -4.57507879e-01
4.64414991e-02 -3.72960359e-01 8.52077305e-02 2.89669275e-01
-1.49420738e-01 -2.50771284e-01 -5.86661875e-01 1.03781021e+00
-1.63779211e+00 -9.04299796e-01 -9.91274491e-02 2.39221859e+00
5.45328617e-01 -3.08303349e-02 1.21215284e-01 8.89260545e-02
1.01361573e+00 3.47982138e-01 -6.70421720e-01 7.56932702e-03
-3.41535866e-01 -4.97453123e-01 6.25418723e-01 2.78443545e-01
-1.11891222e+00 9.75386858e-01 5.60853577e+00 1.07852006e+00
-1.16692328e+00 -2.19553068e-01 9.40807998e-01 -3.98439579e-02
-7.52454758e-01 4.51147370e-02 -8.93385828e-01 7.38940656e-01
4.42003980e-02 -5.95361255e-02 2.15879291e-01 1.04072714e+00
4.57685500e-01 -8.20182920e-01 -4.80891198e-01 1.09829640e+00
4.43377286e-01 -1.25088692e+00 4.03027013e-02 1.12426132e-01
1.16252327e+00 -1.91751242e-01 2.93813229e-01 -3.90997231e-01
2.71603875e-02 -8.47439647e-01 8.85550082e-01 6.25592768e-01
7.96836197e-01 -9.33133364e-01 6.90625072e-01 6.42357349e-01
-1.06874526e+00 -1.84403405e-01 -5.69235325e-01 2.02133045e-01
-9.64419767e-02 1.03875577e+00 -1.09081709e+00 7.57709742e-01
9.47818518e-01 7.96129227e-01 -1.03015387e+00 1.05321741e+00
-1.14192300e-01 6.94366693e-02 -1.17423415e-01 -8.15739930e-02
4.81110904e-03 -4.38235193e-01 8.55210245e-01 5.43775380e-01
6.00323141e-01 5.77126332e-02 2.77888048e-02 8.03985000e-01
2.22893491e-01 2.64746368e-01 -9.11768496e-01 6.78710580e-01
5.92174053e-01 1.22014725e+00 -9.29150224e-01 -2.45247513e-01
-2.40116864e-01 7.01210201e-01 -5.03485613e-02 3.97500783e-01
-8.86934042e-01 -2.50564724e-01 3.03770512e-01 3.42744142e-01
3.50268483e-01 1.08780779e-01 -7.61425793e-01 -1.26037991e+00
3.82311732e-01 -5.09833574e-01 1.68271199e-01 -1.35944021e+00
-4.83994722e-01 6.09625757e-01 -3.59098390e-02 -1.82638216e+00
4.10709679e-02 5.59416935e-02 -3.80316794e-01 7.68688858e-01
-1.39068079e+00 -1.08893180e+00 -7.21863091e-01 5.49786270e-01
7.14549899e-01 -1.74790636e-01 2.47119278e-01 -1.42887965e-01
-1.73687994e-01 3.34153146e-01 -4.27404229e-05 -1.24831907e-01
5.54191887e-01 -9.19571638e-01 -1.43155038e-01 1.05086040e+00
-1.63212851e-01 4.89304215e-01 8.09224129e-01 -5.22345960e-01
-8.77555490e-01 -1.20955086e+00 7.57117212e-01 -2.83742309e-01
-5.29734790e-02 -7.27065802e-02 -5.89635432e-01 4.94685262e-01
3.41262482e-02 1.40984952e-01 2.15528741e-01 -3.42705458e-01
2.81274647e-01 -3.63021493e-01 -1.22019672e+00 8.98068070e-01
9.99003053e-01 -3.11408229e-02 -4.29401577e-01 1.56850275e-02
8.08914423e-01 -5.09796977e-01 -5.28435647e-01 6.81500793e-01
2.46426642e-01 -1.39405763e+00 1.19711459e+00 1.45803809e-01
5.98852515e-01 -7.11926460e-01 -2.99780052e-02 -1.25581646e+00
-4.49702553e-02 -3.02749217e-01 3.69105011e-01 1.09622931e+00
1.77036002e-01 -2.80140787e-01 9.14007723e-01 2.23506674e-01
-3.07419181e-01 -8.52321327e-01 -8.25124621e-01 -2.08335072e-01
-6.68066978e-01 -7.07807764e-02 5.89341998e-01 7.13146031e-01
9.52847824e-02 1.33321792e-01 -4.54771191e-01 2.21934870e-01
5.33489883e-01 4.59605157e-01 7.06659377e-01 -1.20912409e+00
4.02844518e-01 -7.41273984e-02 -4.44994986e-01 -1.31201816e+00
-3.13648701e-01 -4.72192854e-01 1.76464897e-02 -1.63048244e+00
3.87660146e-01 -5.83715260e-01 -2.39778966e-01 2.28028029e-01
-3.47358286e-01 9.95208323e-02 -5.01583889e-02 2.05044612e-01
-8.30238998e-01 5.73962629e-01 1.78831601e+00 -7.20911426e-03
-1.89994827e-01 2.86980182e-01 -1.23873591e+00 7.94047117e-01
7.36660302e-01 -1.49006516e-01 -8.10906410e-01 -1.19425558e-01
1.23741053e-01 4.05482233e-01 2.49307185e-01 -1.01398838e+00
1.44772619e-01 -4.95186508e-01 6.16639674e-01 -1.22317255e+00
3.43228281e-01 -7.61369884e-01 -1.35176629e-01 -2.98752934e-02
-2.10028410e-01 4.38727662e-02 -5.72362319e-02 5.40263355e-01
-3.88263226e-01 -1.04613028e-01 8.39648664e-01 -3.83204669e-01
-1.05191886e+00 1.67307988e-01 -1.75595954e-01 -2.76015494e-02
1.24346483e+00 -6.02771878e-01 -3.18406880e-01 -5.85473359e-01
-6.27860427e-01 2.12577820e-01 9.42527890e-01 5.43720067e-01
9.90760386e-01 -1.30105853e+00 -3.50940645e-01 4.43331927e-01
3.69132817e-01 5.07647812e-01 5.17624497e-01 7.58239031e-01
-2.64829218e-01 6.55625820e-01 -3.56697738e-01 -6.97093725e-01
-1.41692293e+00 9.03168797e-01 3.30438651e-02 -3.14089842e-02
-4.33623165e-01 6.77516639e-01 8.93804789e-01 -2.19486296e-01
-8.45660865e-02 -4.30848122e-01 -8.79495621e-01 -4.91260365e-02
3.53603870e-01 2.57072598e-01 -7.76714906e-02 -1.03834784e+00
-3.11050177e-01 6.34572685e-01 2.20022306e-01 1.52732870e-02
1.09239066e+00 -7.27538049e-01 -5.73709272e-02 2.22271308e-01
9.59717095e-01 2.85614431e-01 -1.37900996e+00 -2.70603627e-01
-3.91135991e-01 -9.12069976e-01 -7.58144408e-02 -6.43837571e-01
-1.18696821e+00 6.59721017e-01 6.04136884e-01 -3.99163589e-02
1.35366929e+00 2.18456551e-01 4.88156796e-01 -2.56208539e-01
5.69603264e-01 -1.21144736e+00 1.73376381e-01 2.32902378e-01
9.92689252e-01 -1.35652792e+00 1.63035363e-01 -8.80901098e-01
-1.13523269e+00 8.23955774e-01 9.30358410e-01 3.71311605e-02
4.80565101e-01 -3.64191175e-01 -5.46955764e-02 -2.14648813e-01
-2.72207081e-01 -4.12310332e-01 4.76003766e-01 7.14443862e-01
5.98703623e-02 -5.10316119e-02 -3.75889927e-01 6.94903016e-01
-3.15935105e-01 -1.15106173e-01 8.85493159e-01 5.59734285e-01
-7.15628028e-01 -4.05894607e-01 -5.61750770e-01 4.65424746e-01
-3.64414304e-01 4.15498987e-02 8.27796385e-03 4.09895688e-01
3.11385930e-01 1.19356692e+00 1.45472586e-01 -4.63819712e-01
1.41809419e-01 -6.79654360e-01 1.14640325e-01 -4.83103544e-01
-3.76176648e-02 2.24884525e-01 -2.10005984e-01 -2.95556515e-01
-4.73202646e-01 -5.67556977e-01 -1.21181405e+00 1.64650187e-01
-3.80355269e-01 1.53631464e-01 4.42900240e-01 8.49682212e-01
3.28448296e-01 4.33846861e-01 1.11204875e+00 -9.07948554e-01
2.47673318e-01 -9.21296418e-01 -7.26433516e-01 2.97040671e-01
2.78244108e-01 -7.93843925e-01 -3.97434801e-01 -5.79721332e-02]
|
[9.766566276550293, -0.4847685396671295]
|
cd3a7f86-b968-423f-967c-99f342e9bde3
|
joint-event-detection-and-description-in
|
1802.1025
| null |
http://arxiv.org/abs/1802.10250v3
|
http://arxiv.org/pdf/1802.10250v3.pdf
|
Joint Event Detection and Description in Continuous Video Streams
|
Dense video captioning is a fine-grained video understanding task that
involves two sub-problems: localizing distinct events in a long video stream,
and generating captions for the localized events. We propose the Joint Event
Detection and Description Network (JEDDi-Net), which solves the dense video
captioning task in an end-to-end fashion. Our model continuously encodes the
input video stream with three-dimensional convolutional layers, proposes
variable-length temporal events based on pooled features, and generates their
captions. Proposal features are extracted within each proposal segment through
3D Segment-of-Interest pooling from shared video feature encoding. In order to
explicitly model temporal relationships between visual events and their
captions in a single video, we also propose a two-level hierarchical captioning
module that keeps track of context. On the large-scale ActivityNet Captions
dataset, JEDDi-Net demonstrates improved results as measured by standard
metrics. We also present the first dense captioning results on the
TACoS-MultiLevel dataset.
|
['Vasili Ramanishka', 'Kate Saenko', 'Leonid Sigal', 'Huijuan Xu', 'Boyang Li']
|
2018-02-28
| null | null | null | null |
['dense-captioning', 'dense-video-captioning']
|
['computer-vision', 'computer-vision']
|
[ 2.33787894e-01 -2.51494590e-02 -2.82056034e-01 -4.92705941e-01
-1.15524912e+00 -4.70769405e-01 7.00290501e-01 9.98687744e-02
-2.23856717e-01 7.12400317e-01 1.04432392e+00 4.03575569e-01
4.45718378e-01 -3.83564949e-01 -1.18335021e+00 -2.26697430e-01
-4.75353509e-01 4.56393361e-01 4.30628389e-01 1.94629893e-01
-7.31080398e-02 1.27076551e-01 -1.46723545e+00 9.82954204e-01
2.11560383e-01 1.23523927e+00 2.88972735e-01 9.73087788e-01
6.09410405e-02 1.26747954e+00 -4.98703271e-01 -2.04616323e-01
-5.05137593e-02 -6.94575310e-01 -8.05032551e-01 3.16580921e-01
7.52707064e-01 -8.45452666e-01 -1.01523888e+00 5.20836949e-01
2.36996174e-01 6.28223792e-02 3.99475753e-01 -1.63121188e+00
-6.87402070e-01 5.80307961e-01 -4.06600505e-01 7.40793288e-01
7.87506104e-01 2.90026844e-01 1.11422074e+00 -8.91950071e-01
8.57068956e-01 1.16707325e+00 4.28890884e-01 5.65812886e-01
-9.78088021e-01 -5.85606039e-01 4.71157342e-01 5.35011232e-01
-1.24386740e+00 -3.35338563e-01 5.72630584e-01 -6.61855817e-01
1.09935582e+00 -7.20491037e-02 6.86070859e-01 1.66323853e+00
-6.79488387e-03 1.10163724e+00 2.21673772e-01 3.01013708e-01
1.28456160e-01 -5.23295701e-01 2.06155498e-02 6.22708321e-01
1.85095202e-02 -1.55839249e-01 -9.86767888e-01 -1.53621376e-01
9.50881600e-01 1.50284156e-01 -3.90402913e-01 -2.15728104e-01
-1.82677877e+00 7.47385025e-01 3.73672158e-01 8.46680030e-02
-8.22125077e-01 7.04215765e-01 7.63751507e-01 -1.91874400e-01
3.52616310e-01 2.25846589e-01 -3.61249506e-01 -4.07897919e-01
-1.20054817e+00 3.84540260e-01 4.89120185e-01 1.18432558e+00
6.18389606e-01 -2.76667118e-01 -1.06368828e+00 3.23652744e-01
2.73368746e-01 2.56781250e-01 3.51412386e-01 -1.00316048e+00
7.44300723e-01 2.46266246e-01 3.03837448e-01 -8.05042684e-01
-6.26062974e-02 -6.74021170e-02 -5.27057946e-01 -5.23179293e-01
-3.53310890e-02 -2.32925713e-01 -1.22717607e+00 2.02707839e+00
1.80888139e-02 1.19229746e+00 7.00258985e-02 1.28891575e+00
1.18842614e+00 1.36234391e+00 6.35752499e-01 -2.06877485e-01
1.55246699e+00 -1.29686856e+00 -7.64720321e-01 -4.22548205e-01
2.68925846e-01 -3.55937243e-01 5.31429172e-01 -1.89661264e-01
-1.22901344e+00 -6.33238435e-01 -7.81724572e-01 -3.47287923e-01
-7.47593399e-03 1.13740705e-01 4.99356747e-01 -3.74225646e-01
-1.09302104e+00 1.95865929e-01 -9.51433659e-01 -4.99425203e-01
5.67198515e-01 9.74251628e-02 -5.46593428e-01 -8.22888985e-02
-1.31398892e+00 4.14556950e-01 6.13717735e-01 6.48064464e-02
-1.60748267e+00 -7.11605012e-01 -1.42227125e+00 3.98541480e-01
1.75846815e-01 -9.05327618e-01 1.44056344e+00 -1.04739273e+00
-1.04010713e+00 7.78938413e-01 -5.70746839e-01 -7.77797937e-01
1.57137930e-01 -2.74492592e-01 -3.85710537e-01 8.03021967e-01
4.08717066e-01 1.26125848e+00 7.91316092e-01 -9.17754233e-01
-7.44900346e-01 2.13451818e-01 6.98818415e-02 3.01504344e-01
7.95186087e-02 3.35305572e-01 -1.07881141e+00 -9.82480884e-01
-3.85985672e-01 -7.76322365e-01 -1.72787488e-01 5.16464561e-02
-1.98746130e-01 -1.63216278e-01 9.34673011e-01 -8.43698800e-01
1.14162791e+00 -2.21749902e+00 2.17881933e-01 -3.29848468e-01
2.43311092e-01 4.85868417e-02 -5.40667534e-01 2.59494573e-01
-1.61096439e-01 -9.88711789e-03 -2.24330246e-01 -5.68291128e-01
-3.57550047e-02 1.18176073e-01 -6.13620877e-01 1.02252878e-01
8.05905223e-01 1.23600328e+00 -1.23599124e+00 -6.81484938e-01
1.59692526e-01 6.59229994e-01 -6.11138284e-01 7.87853360e-01
-4.88472193e-01 1.13590986e-01 -4.94560599e-01 5.83163798e-01
1.88645318e-01 -7.03668714e-01 -1.29211828e-01 -4.92209047e-01
8.84146020e-02 1.59704596e-01 -8.17141652e-01 2.21050334e+00
-1.58260703e-01 1.11308050e+00 -2.82677889e-01 -7.38879383e-01
5.21870792e-01 8.64187777e-01 6.85952663e-01 -4.70184118e-01
-7.24684820e-02 -1.85856432e-01 -7.38394439e-01 -6.39788389e-01
6.52218342e-01 2.27073342e-01 -6.03981674e-01 3.54717046e-01
5.72045565e-01 4.51172858e-01 4.68719035e-01 6.25979841e-01
1.40205228e+00 4.27195162e-01 1.06664851e-01 2.76358157e-01
2.51956433e-01 1.46904290e-01 5.69177449e-01 6.72553778e-01
-4.43846017e-01 1.36315393e+00 6.29772365e-01 -6.15893543e-01
-1.12903428e+00 -1.19813931e+00 3.71147662e-01 1.10017943e+00
3.82640958e-01 -6.63541973e-01 -6.00472927e-01 -6.94633543e-01
-2.31312484e-01 3.34582448e-01 -9.04219329e-01 -1.10033214e-01
-7.50069916e-01 -3.56476635e-01 3.80362630e-01 7.98408866e-01
5.04545391e-01 -1.39797997e+00 -6.39813721e-01 5.69720626e-01
-7.57583618e-01 -1.63478267e+00 -1.00747442e+00 2.62099714e-03
-4.75393534e-01 -1.01503837e+00 -1.03692174e+00 -1.09426272e+00
4.74986345e-01 1.95170134e-01 1.41589284e+00 -2.05575526e-01
-2.09903494e-01 4.35670108e-01 -5.16849637e-01 8.79289880e-02
-1.76182926e-01 -4.17168699e-02 -3.50197107e-01 2.66229719e-01
4.88148123e-01 -2.90452063e-01 -6.53230190e-01 3.73310685e-01
-8.65419686e-01 4.67829227e-01 4.83895719e-01 4.75731105e-01
8.75957549e-01 -8.04812789e-01 6.80367291e-01 -2.06686184e-01
2.29483142e-01 -1.01942182e+00 -3.03186029e-01 1.92392156e-01
4.58820522e-01 -3.27211656e-02 3.05578977e-01 -3.20256978e-01
-9.17639017e-01 4.03210729e-01 1.23168834e-01 -1.04741216e+00
-3.94928187e-01 3.53201270e-01 -1.81742981e-02 6.08693242e-01
2.66600668e-01 3.85189384e-01 -3.17187726e-01 -2.13351011e-01
3.56675059e-01 4.42655712e-01 1.10709429e+00 -3.82140517e-01
5.84688902e-01 4.33780193e-01 -4.16066200e-01 -4.91213739e-01
-1.07504475e+00 -5.39417446e-01 -4.32742208e-01 -3.08763057e-01
1.49420345e+00 -1.73374951e+00 -5.31960964e-01 2.66705096e-01
-1.65322793e+00 -3.17190528e-01 -3.07482868e-01 6.12334549e-01
-9.37544167e-01 1.81322455e-01 -8.44444871e-01 -9.23047960e-02
-3.38772088e-01 -1.15278208e+00 1.67881656e+00 3.99666488e-01
-3.73983920e-01 -5.90764701e-01 2.20334947e-01 2.63174683e-01
8.51425007e-02 6.62926376e-01 1.79075390e-01 -5.25950968e-01
-1.00882757e+00 -9.80872661e-03 -5.37584662e-01 1.89827122e-02
-2.33061582e-01 -2.21105441e-02 -7.83061087e-01 -1.35455906e-01
-5.96329272e-01 -5.10577857e-01 1.06304586e+00 5.55514216e-01
1.24628329e+00 -1.91046685e-01 -4.13811952e-01 6.81485832e-01
1.22150981e+00 5.11090234e-02 5.80063343e-01 1.66820168e-01
7.51481652e-01 1.80971786e-01 6.52107537e-01 6.09062195e-01
5.55811644e-01 6.29533470e-01 6.36563420e-01 -6.08672760e-02
-3.91824961e-01 -5.64867198e-01 6.51317537e-01 2.88551658e-01
1.60451889e-01 -6.65626466e-01 -5.44449031e-01 1.09197581e+00
-2.28949213e+00 -1.43198621e+00 -1.38093177e-02 1.65472007e+00
5.99828660e-01 -1.48503780e-02 2.00641230e-01 -5.52009821e-01
1.22024131e+00 5.50555766e-01 -5.08163154e-01 -9.34402645e-02
-1.54215679e-01 -1.08781278e-01 2.35656649e-01 1.98637605e-01
-1.39698482e+00 9.46638942e-01 6.16994047e+00 2.96559215e-01
-9.82272506e-01 1.98087752e-01 6.33701801e-01 -5.14411449e-01
-7.32517242e-02 -1.73940793e-01 -7.03678071e-01 7.65501320e-01
1.30076873e+00 -3.20317209e-01 2.30142996e-02 6.64877176e-01
4.05848473e-01 -1.29401237e-02 -1.53983963e+00 1.21451521e+00
4.04457748e-01 -1.98425436e+00 3.95878971e-01 -2.49981910e-01
8.12621057e-01 4.07154709e-01 -4.34290230e-01 4.02103543e-01
-4.77328263e-02 -8.07456732e-01 1.17339718e+00 5.36721170e-01
7.86287725e-01 -5.02559543e-01 5.88991344e-01 -2.91420311e-01
-1.60163474e+00 1.44033492e-01 -7.60699138e-02 -2.49343645e-02
1.02483582e+00 2.66701430e-01 -8.34398746e-01 2.79380172e-01
6.93931043e-01 1.28787935e+00 -4.01308984e-01 1.48063326e+00
-1.35441229e-01 7.38099754e-01 -2.54491836e-01 3.68577361e-01
5.81031561e-01 3.62366199e-01 4.91068751e-01 1.57681060e+00
4.91030961e-01 2.30060279e-01 2.26410255e-01 8.06768239e-01
-5.07438958e-01 -4.49699193e-01 -3.39366734e-01 -2.69109249e-01
3.27743739e-01 1.13533676e+00 -6.38970435e-01 -8.06903541e-01
-5.27717829e-01 1.42298639e+00 1.13379881e-01 4.79976863e-01
-1.42345214e+00 -2.87171125e-01 1.03209651e+00 -4.92253166e-04
7.91014612e-01 -6.92875534e-02 3.33816171e-01 -1.47734213e+00
7.06908777e-02 -4.44455534e-01 6.83122516e-01 -1.32315803e+00
-1.07476020e+00 8.01629901e-01 1.06321514e-01 -1.24074101e+00
-4.33082134e-01 -1.18703462e-01 -7.84567237e-01 5.97545207e-01
-1.45175958e+00 -1.13688326e+00 -8.06214392e-01 8.07502329e-01
1.03868258e+00 1.84985936e-01 5.58664024e-01 5.46713591e-01
-5.76259851e-01 2.10836604e-01 -3.76741648e-01 5.08237481e-01
7.61333406e-01 -9.02437985e-01 8.07834685e-01 9.55880284e-01
2.92420089e-01 -1.14511490e-01 7.11552203e-01 -6.86135888e-01
-1.07866466e+00 -1.75608492e+00 1.14484239e+00 -4.85646635e-01
5.77368259e-01 -5.66164613e-01 -9.17428851e-01 1.10427725e+00
3.59154552e-01 4.96837437e-01 5.31958640e-01 -4.79926765e-01
-4.44279760e-01 2.03286827e-01 -7.10866630e-01 3.50234628e-01
1.34677207e+00 -7.40300953e-01 -9.39233959e-01 5.07560849e-01
1.21728694e+00 -4.37469780e-01 -7.43184328e-01 9.86435935e-02
2.67854303e-01 -6.28437340e-01 9.82799768e-01 -8.38236809e-01
8.75740767e-01 -5.45435965e-01 -2.94399150e-02 -9.82461393e-01
-4.20802027e-01 -8.46877694e-01 -4.75973070e-01 1.07706869e+00
3.68796438e-01 3.98440957e-01 8.46698284e-01 3.50898802e-01
-4.56691206e-01 -3.08212906e-01 -7.46834219e-01 -4.35065836e-01
-7.47596800e-01 -4.80123401e-01 5.10868967e-01 6.21976078e-01
-2.88857296e-02 4.50793117e-01 -7.30848849e-01 4.30637181e-01
5.12983501e-01 3.32461178e-01 5.46536267e-01 -8.90531719e-01
-1.31483123e-01 3.75918150e-02 -6.81756139e-01 -1.30656028e+00
3.68521333e-01 -6.98939323e-01 4.91541684e-01 -1.85865211e+00
5.07419527e-01 3.20316851e-01 -2.98528224e-01 5.56206763e-01
-2.06453413e-01 6.90591276e-01 2.80908138e-01 1.90063789e-01
-1.45725060e+00 6.28931522e-01 9.20177042e-01 -8.29016492e-02
-6.33428842e-02 -6.40303016e-01 -3.94807070e-01 3.65909070e-01
3.31130832e-01 -4.84928906e-01 -3.20631564e-01 -7.33859539e-01
-1.99913964e-01 6.05616808e-01 6.87337577e-01 -1.19884515e+00
1.45786777e-01 -2.33411729e-01 4.87140536e-01 -7.49616325e-01
5.40893435e-01 -5.92942119e-01 2.88883090e-01 3.31367068e-02
-5.98032534e-01 1.04782991e-01 2.55757034e-01 8.44026983e-01
-6.25041008e-01 3.86420161e-01 5.17189443e-01 2.81187668e-02
-1.34404182e+00 9.03778374e-01 -4.89534646e-01 2.66990811e-01
1.51721799e+00 -7.29589313e-02 -3.58934104e-01 -5.11972606e-01
-9.64196324e-01 5.81829607e-01 3.55573028e-01 9.15931344e-01
6.13569558e-01 -1.57363677e+00 -9.78450179e-01 -2.68743839e-02
4.96572584e-01 5.69584742e-02 5.01435459e-01 3.70658934e-01
-5.34716368e-01 5.58312833e-01 -4.41812217e-01 -7.60740876e-01
-1.09417140e+00 5.92958272e-01 8.57288018e-02 -1.69717655e-01
-6.27818465e-01 9.63716030e-01 5.14388442e-01 6.30290210e-01
3.76790643e-01 -4.92214173e-01 -1.85148984e-01 1.45792067e-01
9.32341516e-01 -2.17772886e-01 -3.26996714e-01 -1.01846886e+00
-4.47938025e-01 3.29506576e-01 -1.46051913e-01 -2.78033931e-02
1.37016284e+00 -2.65608370e-01 3.05238754e-01 1.21014984e-02
1.49585235e+00 -6.81734800e-01 -2.06623220e+00 -9.18207467e-02
-2.15102762e-01 -2.95205206e-01 -5.30740432e-02 -5.97374439e-01
-1.02616036e+00 6.39353752e-01 2.11216033e-01 -1.35323808e-01
1.08800185e+00 4.49270993e-01 1.19040835e+00 4.59590182e-02
1.50584370e-01 -5.95110476e-01 3.25818628e-01 4.26922619e-01
1.02490377e+00 -1.25272012e+00 -4.51392263e-01 -2.27490410e-01
-9.23576415e-01 8.96578908e-01 8.23747158e-01 -1.93455622e-01
2.17884868e-01 2.15359889e-02 -2.96347141e-01 -1.69982076e-01
-1.16975141e+00 -2.66115814e-01 2.71048635e-01 4.78669673e-01
2.49317393e-01 -2.70626605e-01 1.49666846e-01 8.00074041e-01
3.82738322e-01 5.03141880e-01 6.96337461e-01 8.30355704e-01
-4.02841687e-01 -5.32594502e-01 -1.60812303e-01 2.93234140e-01
-3.06238741e-01 -1.64201900e-01 -1.71249062e-01 3.37980807e-01
5.56330271e-02 7.47711062e-01 7.75887728e-01 -2.61702478e-01
2.17893004e-01 4.02038246e-02 5.01631461e-02 -7.39361227e-01
-4.55430388e-01 -6.02690205e-02 1.54721364e-01 -1.07648325e+00
-5.26692390e-01 -7.18807459e-01 -1.52688158e+00 1.11056492e-01
4.61755067e-01 4.99601513e-01 3.82213473e-01 7.45654821e-01
9.05184209e-01 6.46370173e-01 4.07000333e-01 -1.31650448e+00
1.85190275e-01 -7.28596985e-01 -1.79049373e-01 7.43960440e-01
6.78038955e-01 -4.69194263e-01 -2.08595037e-01 7.93669403e-01]
|
[10.47044563293457, 0.6698200702667236]
|
899c4283-a93d-41b6-a639-4059c9386be1
|
uniparma-semeval-2021-task-5-toxic-spans
|
2103.09645
| null |
https://arxiv.org/abs/2103.09645v2
|
https://arxiv.org/pdf/2103.09645v2.pdf
|
UniParma at SemEval-2021 Task 5: Toxic Spans Detection Using CharacterBERT and Bag-of-Words Model
|
With the ever-increasing availability of digital information, toxic content is also on the rise. Therefore, the detection of this type of language is of paramount importance. We tackle this problem utilizing a combination of a state-of-the-art pre-trained language model (CharacterBERT) and a traditional bag-of-words technique. Since the content is full of toxic words that have not been written according to their dictionary spelling, attendance to individual characters is crucial. Therefore, we use CharacterBERT to extract features based on the word characters. It consists of a CharacterCNN module that learns character embeddings from the context. These are, then, fed into the well-known BERT architecture. The bag-of-words method, on the other hand, further improves upon that by making sure that some frequently used toxic words get labeled accordingly. With a 4 percent difference from the first team, our system ranked 36th in the competition. The code is available for further re-search and reproduction of the results.
|
['Andrea Prati', 'Leonardo Rossi', 'Akbar Karimi']
|
2021-03-17
| null |
https://aclanthology.org/2021.semeval-1.25
|
https://aclanthology.org/2021.semeval-1.25.pdf
|
semeval-2021
|
['toxic-spans-detection']
|
['natural-language-processing']
|
[ 7.66694322e-02 -3.07923943e-01 -1.27494290e-01 -2.10150005e-03
-6.37781024e-01 -6.36427999e-01 8.67690682e-01 6.67671919e-01
-1.15788138e+00 5.65348923e-01 2.59337544e-01 -3.52408290e-01
2.64141887e-01 -8.46745312e-01 -5.20656168e-01 -6.20700896e-01
2.92885989e-01 5.98372519e-01 4.23379928e-01 -4.03725773e-01
7.04227626e-01 5.93703270e-01 -1.30021298e+00 4.53699142e-01
6.76984787e-01 8.80140603e-01 3.13831925e-01 5.44107258e-01
-5.40976942e-01 6.91053569e-01 -6.98722243e-01 -6.20287001e-01
1.84308458e-02 -2.00147793e-01 -4.79692519e-01 -2.10990384e-01
1.95565313e-01 -1.23026311e-01 -3.24367642e-01 1.00681806e+00
3.44821304e-01 9.55110788e-02 7.07084119e-01 -4.80209321e-01
-7.04735160e-01 7.15408921e-01 -2.76276588e-01 3.20380867e-01
2.28932425e-01 3.79754603e-02 1.25663745e+00 -9.98198271e-01
5.81938982e-01 1.02213299e+00 3.65868479e-01 6.06365561e-01
-8.18688273e-01 -5.51964164e-01 1.79896012e-01 3.05029035e-01
-1.35985339e+00 -1.31011754e-01 4.70702946e-01 -6.19088411e-01
1.22271228e+00 1.42178521e-01 6.14593089e-01 1.38006210e+00
1.21314406e-01 8.12647283e-01 9.78059709e-01 -8.60253215e-01
3.21066916e-01 5.56509435e-01 9.90805179e-02 5.16254604e-01
3.79342347e-01 -1.71415791e-01 -4.23777014e-01 -1.28821493e-03
1.06718682e-01 -3.32385376e-02 -1.62602067e-01 -1.73835829e-01
-7.57426143e-01 9.93365526e-01 1.81185558e-01 8.84404898e-01
-3.97531837e-01 1.34141088e-01 6.07014239e-01 -4.10420820e-02
4.44448262e-01 7.09414423e-01 -3.22706431e-01 -4.38980907e-01
-1.08146715e+00 1.90855935e-01 8.94393802e-01 6.85706556e-01
5.32366991e-01 -2.15523884e-01 -1.06416181e-01 8.00123394e-01
2.34798193e-01 2.05515921e-01 8.86957169e-01 1.13785616e-03
4.07269478e-01 8.17893028e-01 -6.15642481e-02 -9.21248972e-01
-1.57621026e-01 -5.62727869e-01 -4.85417485e-01 5.46958968e-02
3.33833069e-01 1.15598388e-01 -9.52888906e-01 1.26958764e+00
-8.56533870e-02 -1.87148675e-01 -1.35696337e-01 5.86274445e-01
4.16961223e-01 8.44627023e-01 1.80180743e-01 4.70999954e-03
1.45877910e+00 -8.70545447e-01 -6.17205799e-01 -3.35194528e-01
5.76995909e-01 -7.50481188e-01 1.01845431e+00 7.09557235e-01
-7.02197194e-01 -2.41525352e-01 -1.27497673e+00 -3.40705253e-02
-1.13339078e+00 -1.25123963e-01 2.22212285e-01 8.55981767e-01
-8.19508672e-01 6.41745329e-01 -5.51661313e-01 -5.17164409e-01
2.68910050e-01 1.48700044e-01 -2.58397996e-01 -2.15014189e-01
-1.38865197e+00 1.23260450e+00 6.36542857e-01 -2.62302607e-01
-7.38192379e-01 -4.12930936e-01 -7.67305672e-01 -1.17537841e-01
4.67218876e-01 1.10745400e-01 1.06280839e+00 -8.49780023e-01
-1.29364014e+00 9.29927766e-01 1.01863787e-01 -3.25860441e-01
6.72324479e-01 -4.38979894e-01 -5.49454927e-01 1.31568670e-01
-1.06680028e-01 2.43140548e-01 8.39569449e-01 -9.36706662e-01
-6.52212620e-01 -5.42017817e-02 -1.10415354e-01 -1.06072746e-01
-1.02172589e+00 2.82569706e-01 -6.97704136e-01 -7.12376595e-01
-4.76075143e-01 -7.13607669e-01 -5.91161139e-02 -2.11074501e-01
-4.09196049e-01 -4.47682589e-01 4.42565888e-01 -6.44550860e-01
1.68478322e+00 -2.27393889e+00 5.01912907e-02 3.48678142e-01
1.29667908e-01 5.44497311e-01 -1.91098422e-01 9.00741160e-01
-2.08450913e-01 3.03955972e-01 -2.55097747e-01 -5.38002968e-01
1.49151832e-01 1.07568547e-01 -3.46849918e-01 4.93550569e-01
2.71024495e-01 6.22403324e-01 -1.04493797e+00 -4.01365012e-01
1.23511210e-01 4.53968614e-01 -3.09127331e-01 1.09932594e-01
-3.74670088e-01 -7.06819445e-02 -5.04118621e-01 3.68893147e-01
4.86845881e-01 8.83228332e-02 4.26259860e-02 2.14736789e-01
-5.01834095e-01 4.07106191e-01 -9.38530505e-01 1.52898109e+00
-6.11964405e-01 4.39618886e-01 -3.18896472e-01 -6.83910310e-01
8.21901858e-01 2.69092560e-01 2.66424119e-01 -7.51468301e-01
3.90657306e-01 5.41036248e-01 3.67064290e-02 -4.45170403e-01
7.70165443e-01 -6.95718527e-02 -1.77987456e-01 3.19848835e-01
-5.37734553e-02 -2.87860278e-02 7.22784042e-01 4.32095945e-01
1.25791264e+00 2.38289349e-02 4.35280204e-01 -1.37768716e-01
8.43179643e-01 -1.44268960e-01 4.57497761e-02 6.11190856e-01
7.15678260e-02 5.42060971e-01 5.16257286e-01 -2.73186713e-01
-1.12287092e+00 -6.37503624e-01 -2.92680468e-02 1.12546074e+00
-4.08452749e-01 -8.34747374e-01 -7.13507175e-01 -6.72433972e-01
1.06778275e-02 1.13488352e+00 -6.95683181e-01 -1.56245053e-01
-5.93635857e-01 -4.45926785e-01 5.26993215e-01 3.43614250e-01
-3.73167358e-02 -1.40785289e+00 -8.08884859e-01 4.36367810e-01
9.83541906e-02 -8.87668967e-01 -4.49569494e-01 6.32835925e-01
-3.19070935e-01 -7.46033251e-01 -8.76872897e-01 -5.28240442e-01
4.13168639e-01 -2.35308632e-01 9.69519377e-01 2.98191220e-01
-3.70104373e-01 1.68621913e-01 -8.80507410e-01 -6.25292957e-01
-7.74439514e-01 2.37439424e-01 -8.22192281e-02 1.30566120e-01
6.65339947e-01 -3.36737782e-01 -2.98417330e-01 -2.83414602e-01
-1.27134144e+00 -2.85698742e-01 7.82221973e-01 5.99303842e-01
2.44249582e-01 -1.04242176e-01 1.67970836e-01 -9.12083030e-01
7.23535359e-01 -4.28363591e-01 -5.35658181e-01 7.79357627e-02
-5.21826208e-01 1.73688829e-01 9.33401883e-01 -5.53748727e-01
-3.81483763e-01 3.48644592e-02 -3.65652144e-01 -1.27333820e-01
-1.42525032e-01 5.87647915e-01 -2.30848074e-01 1.95347309e-01
5.12545049e-01 3.69621754e-01 -3.91981333e-01 -7.44227946e-01
5.46892405e-01 1.02390850e+00 2.98034996e-01 -2.43126839e-01
8.27133238e-01 1.08157806e-01 -4.64238942e-01 -9.68169272e-01
-6.79490089e-01 -8.74202967e-01 -5.70463955e-01 -1.05072014e-01
8.66910934e-01 -7.34994471e-01 -2.11025089e-01 5.29110312e-01
-1.26471210e+00 -8.57709572e-02 -1.65134057e-01 3.76037002e-01
-1.98713735e-01 7.86268353e-01 -6.33985937e-01 -1.06049216e+00
-2.56931722e-01 -8.30724895e-01 7.48273253e-01 -1.22829884e-01
-4.41560775e-01 -7.99078763e-01 2.38781750e-01 5.43870963e-02
4.50314313e-01 -6.65301457e-02 1.18248832e+00 -1.13377118e+00
-9.97503325e-02 -7.59629130e-01 -4.13355120e-02 5.42767763e-01
-6.17936924e-02 -5.52470349e-02 -8.88538718e-01 -1.52321562e-01
-4.27299067e-02 -3.00185621e-01 1.16039491e+00 -1.37572527e-01
1.08447123e+00 -1.76744476e-01 -8.63038003e-02 1.87715530e-01
1.45128536e+00 2.41365120e-01 7.71628976e-01 8.41563702e-01
6.08762622e-01 5.83761752e-01 3.87523979e-01 6.89697325e-01
7.81385452e-02 6.48919165e-01 5.54236531e-01 2.21273795e-01
6.88333884e-02 -2.53527105e-01 6.62061989e-01 8.89304399e-01
3.38880658e-01 -6.48041785e-01 -1.03283954e+00 6.98142648e-01
-1.55535400e+00 -7.79710770e-01 -2.49740541e-01 2.21446681e+00
1.03683591e+00 5.02143562e-01 8.40916112e-02 3.46996874e-01
4.55885261e-01 1.33423552e-01 -2.38831162e-01 -7.96747208e-01
-8.10286775e-02 3.47240418e-01 7.62772501e-01 2.57523507e-01
-1.05876708e+00 1.11993062e+00 5.22776270e+00 1.28260481e+00
-1.14809620e+00 -2.34756470e-02 3.01759601e-01 1.32611647e-01
-3.13848466e-01 -1.96414784e-01 -9.53970313e-01 7.40969241e-01
1.09848869e+00 -6.52603135e-02 3.91241550e-01 7.83290207e-01
1.12782262e-01 -3.09800059e-01 -1.01622140e+00 9.75179553e-01
6.18893683e-01 -9.34422672e-01 2.10994005e-01 2.09511265e-01
3.47828060e-01 -1.88382268e-02 9.28263441e-02 4.15548384e-01
1.15169168e-01 -9.77186203e-01 1.14585412e+00 4.52150077e-01
6.68181181e-01 -9.34663355e-01 7.79332578e-01 5.44879079e-01
-8.39626670e-01 -2.10034192e-01 -4.35673654e-01 7.11827353e-02
8.79869461e-02 8.09528589e-01 -4.90531713e-01 4.37354118e-01
4.60036308e-01 6.34184837e-01 -7.51965582e-01 1.25931108e+00
-5.01477003e-01 6.10111177e-01 -1.84347942e-01 -5.58957160e-01
6.21744037e-01 -1.15853876e-01 3.97980243e-01 1.71212709e+00
3.07804376e-01 -2.17330381e-01 1.83108285e-01 7.01821208e-01
-1.48839951e-01 5.39873779e-01 -4.98072654e-01 -6.93624675e-01
3.45248505e-02 1.42289269e+00 -7.62718797e-01 -5.39745569e-01
-4.20758307e-01 1.21970654e+00 2.72706240e-01 -1.40164018e-01
-6.06575429e-01 -6.96663439e-01 3.76767069e-01 5.15117422e-02
6.76174104e-01 -3.39152277e-01 -6.39242977e-02 -8.85271966e-01
5.80644533e-02 -9.49448049e-01 1.31235078e-01 -4.24581647e-01
-1.40313244e+00 8.77059042e-01 -2.35191941e-01 -1.10854030e+00
-2.49316785e-02 -1.10555148e+00 -3.10648143e-01 9.16196942e-01
-1.61623943e+00 -8.54030430e-01 1.42922461e-01 2.34750882e-01
6.49629593e-01 -1.80019349e-01 1.04553342e+00 4.86290425e-01
-6.19079530e-01 6.34144604e-01 4.01187629e-01 3.02095145e-01
7.41290987e-01 -1.26681900e+00 5.12724876e-01 8.13409150e-01
2.68299460e-01 5.52066147e-01 7.49122918e-01 -8.19470167e-01
-1.19052136e+00 -9.53513563e-01 1.54190564e+00 -5.75141191e-01
1.25523460e+00 -8.48348200e-01 -8.74947488e-01 1.17816553e-01
2.84523129e-01 -3.62335414e-01 7.79491305e-01 -2.52150625e-01
-6.92746520e-01 1.32038206e-01 -7.54613459e-01 6.19253039e-01
6.37024224e-01 -5.88660061e-01 -6.94508672e-01 3.92162949e-01
4.78522331e-01 3.22975852e-02 -3.04488391e-01 -3.65892440e-01
3.85718524e-01 -6.44777179e-01 5.37231147e-01 -5.59178650e-01
5.95865250e-01 -2.10059360e-02 -2.19691932e-01 -1.30868912e+00
-2.34944984e-01 -3.46829534e-01 4.93325815e-02 1.26770270e+00
5.77500641e-01 -4.16318983e-01 6.13896847e-01 3.21777225e-01
-1.56440675e-01 -6.90352440e-01 -7.97785461e-01 -1.09029949e+00
3.35496694e-01 -4.58726376e-01 9.27315652e-02 6.93519533e-01
2.62701571e-01 4.58066076e-01 -3.74871999e-01 -4.46759790e-01
2.05914631e-01 -3.11535448e-01 2.73308873e-01 -1.14369261e+00
-3.81772608e-01 -6.75770342e-01 -4.25608158e-01 -9.44874525e-01
1.05331406e-01 -1.05107903e+00 3.22191417e-01 -1.46217275e+00
2.53355831e-01 -6.58200160e-02 -4.79977280e-01 4.41913426e-01
-1.45766899e-01 3.55416983e-01 4.40475971e-01 -1.18608512e-01
-6.57446682e-01 4.94920403e-01 7.30357349e-01 -2.16997907e-01
8.12020451e-02 -3.79890501e-01 -6.28899574e-01 5.56640029e-01
8.27537179e-01 -7.02309191e-01 1.32302061e-01 -2.54353762e-01
5.63481212e-01 -8.17220151e-01 -1.70002997e-01 -1.15360391e+00
9.47849005e-02 -9.09739174e-03 2.71045744e-01 -6.20007038e-01
4.64168578e-01 -9.17019904e-01 -3.51457328e-01 7.45909333e-01
-3.40756714e-01 3.78484249e-01 2.57834345e-01 4.49088156e-01
6.52126828e-03 -8.03479254e-01 7.01194227e-01 -3.81826937e-01
-7.64685690e-01 1.27399281e-01 -1.02717912e+00 -1.05746128e-01
8.71489286e-01 -3.24438453e-01 -8.19324479e-02 -2.43788451e-01
-2.35975131e-01 -9.19323713e-02 4.43527728e-01 4.65352267e-01
4.96719480e-01 -9.52454329e-01 -7.32247353e-01 -4.63764556e-02
5.03724098e-01 -5.43331742e-01 -1.93818063e-01 6.41181409e-01
-7.01354563e-01 4.07720387e-01 9.46042016e-02 1.40508503e-01
-1.09914386e+00 8.42200577e-01 5.07718977e-03 -5.42073250e-01
-5.65129936e-01 8.74287903e-01 -2.20441028e-01 -4.44150008e-02
4.02894884e-01 -1.88796446e-01 -6.06744468e-01 4.98909801e-01
9.77273166e-01 2.20746070e-01 3.26422215e-01 -7.68954456e-01
-4.81070071e-01 3.71686280e-01 -3.63843530e-01 -2.62573063e-01
1.47851825e+00 3.28139216e-01 -2.57972211e-01 7.94935346e-01
1.26531780e+00 4.13680464e-01 -6.84987068e-01 -1.00870021e-01
4.83940631e-01 -3.12809408e-01 -1.01769567e-01 -1.01561046e+00
-6.10100210e-01 1.07943261e+00 4.82711762e-01 2.49573976e-01
7.88332164e-01 -2.83815295e-01 8.03443968e-01 3.86067659e-01
2.24037766e-01 -1.36999714e+00 2.20610350e-01 1.02197921e+00
7.78241694e-01 -8.62336814e-01 -7.40088057e-03 -1.37746423e-01
-4.22533661e-01 1.29417741e+00 2.74037808e-01 -2.46383816e-01
4.38804775e-01 1.99845687e-01 1.48255840e-01 4.28963974e-02
-6.35423660e-01 -3.39107245e-01 1.35749668e-01 4.26510036e-01
4.51213241e-01 -2.73046084e-02 -7.54388511e-01 5.77700853e-01
-1.76598296e-01 -3.75889450e-01 6.36887491e-01 8.74246120e-01
-6.93959951e-01 -1.53784800e+00 -2.77345896e-01 3.23172450e-01
-7.05142856e-01 -6.19799912e-01 -7.84018934e-01 5.69398463e-01
2.11295247e-01 1.03064740e+00 -9.47469100e-02 -3.04948747e-01
4.84199584e-01 3.01990300e-01 1.72438666e-01 -9.21363890e-01
-9.99164760e-01 3.35088298e-02 7.97051862e-02 -4.04408067e-01
1.11676022e-01 -6.25779927e-01 -1.07641351e+00 -1.91783085e-01
-3.92746270e-01 2.63037056e-01 1.02477121e+00 8.10094416e-01
-7.51083866e-02 3.93368125e-01 4.70467657e-01 -6.29097402e-01
-8.23792994e-01 -1.03692806e+00 -6.77377522e-01 4.49135691e-01
8.10297802e-02 -3.19123358e-01 -2.54124880e-01 -2.71912694e-01]
|
[9.59294605255127, 10.0322904586792]
|
d00379c7-84fd-4048-8857-688791c67d66
|
orchnet-a-robust-global-feature-aggregation
|
2303.00477
| null |
https://arxiv.org/abs/2303.00477v1
|
https://arxiv.org/pdf/2303.00477v1.pdf
|
ORCHNet: A Robust Global Feature Aggregation approach for 3D LiDAR-based Place recognition in Orchards
|
Robust and reliable place recognition and loop closure detection in agricultural environments is still an open problem. In particular, orchards are a difficult case study due to structural similarity across the entire field. In this work, we address the place recognition problem in orchards resorting to 3D LiDAR data, which is considered a key modality for robustness. Hence, we propose ORCHNet, a deep-learning-based approach that maps 3D-LiDAR scans to global descriptors. Specifically, this work proposes a new global feature aggregation approach, which fuses multiple aggregation methods into a robust global descriptor. ORCHNet is evaluated on real-world data collected in orchards, comprising data from the summer and autumn seasons. To assess the robustness, We compare ORCHNet with state-of-the-art aggregation approaches on data from the same season and across seasons. Moreover, we additionally evaluate the proposed approach as part of a localization framework, where ORCHNet is used as a loop closure detector. The empirical results indicate that, on the place recognition task, ORCHNet outperforms the remaining approaches, and is also more robust across seasons. As for the localization, the edge cases where the path goes through the trees are solved when integrating ORCHNet as a loop detector, showing the potential applicability of the proposed approach in this task. The code and dataset will be publicly available at:\url{https://github.com/Cybonic/ORCHNet.git}
|
['U. J. Nunes', 'C. Premebida', 'C. Liu', 'M. J. Coombes', 'P. Conde', 'L. Garrote', 'T. Barros']
|
2023-03-01
| null | null | null | null |
['loop-closure-detection']
|
['computer-vision']
|
[-1.00865178e-01 -3.86972755e-01 -1.58137009e-01 -2.03119814e-01
-6.39794588e-01 -8.13904226e-01 6.57552719e-01 6.08590961e-01
-4.36274320e-01 5.01366615e-01 -3.34915876e-01 -2.39126459e-01
-3.96046132e-01 -1.08045435e+00 -7.66804516e-01 -6.80139959e-01
-2.09292203e-01 1.63936540e-01 4.40328211e-01 -2.09694088e-01
1.57989115e-01 1.17989719e+00 -1.87642956e+00 -2.47233719e-01
9.52244341e-01 1.21313226e+00 4.66989070e-01 2.61656404e-01
1.68544412e-01 -7.95928538e-02 -2.90589035e-01 -5.25605455e-02
3.74772638e-01 1.87923476e-01 -2.26302892e-01 2.30527250e-03
7.61916697e-01 -3.44073594e-01 1.07764147e-01 9.47436392e-01
5.37235200e-01 6.45170659e-02 4.28184241e-01 -1.25573361e+00
-2.19543070e-01 1.53770506e-01 -5.78499734e-01 -1.21268190e-01
1.21800452e-01 5.92314005e-02 1.02306151e+00 -1.11720240e+00
5.48201799e-01 1.00009263e+00 9.08649266e-01 -3.15546095e-01
-1.38506937e+00 -6.37432694e-01 3.03803593e-01 2.19648451e-01
-1.94080341e+00 -2.50865489e-01 5.06770670e-01 -3.00564706e-01
5.66393793e-01 3.01332265e-01 5.46161592e-01 7.23391771e-01
2.95676053e-01 7.71665692e-01 1.02075863e+00 -1.99720830e-01
4.14756835e-01 -1.22045942e-01 -2.37594284e-02 5.89043140e-01
6.35259330e-01 3.05252910e-01 -5.16324222e-01 -2.07870658e-02
6.32178068e-01 1.99084565e-01 -6.93945885e-02 -9.07234430e-01
-9.33347523e-01 8.46832097e-01 9.90702808e-01 3.25370580e-01
-6.66657865e-01 -3.96478325e-02 2.35897347e-01 -1.19626611e-01
5.35507977e-01 2.29409352e-01 -2.37590134e-01 1.95177406e-01
-1.16451991e+00 6.20883107e-01 7.72861183e-01 8.85006368e-01
1.07544935e+00 -1.99305817e-01 3.79717499e-02 6.38782084e-01
3.61000329e-01 6.80830479e-01 -1.03880882e-01 -7.18970180e-01
2.69453049e-01 9.39261913e-01 7.01472312e-02 -1.24593687e+00
-5.54747343e-01 -5.22388935e-01 -7.51657546e-01 3.96398544e-01
2.92477459e-01 3.04756686e-02 -7.18319535e-01 1.51068532e+00
5.09570479e-01 2.98124969e-01 -9.02857184e-02 8.55741024e-01
8.33419621e-01 3.89818817e-01 3.58934924e-02 2.18260467e-01
1.25320208e+00 -6.05737031e-01 -5.59269845e-01 -3.71333688e-01
5.14395773e-01 -8.95121753e-01 5.86933374e-01 2.13979766e-01
-4.73418832e-01 -5.86582303e-01 -1.21914279e+00 5.11882566e-02
-1.05237961e+00 8.09042454e-01 6.05854392e-01 4.00182724e-01
-1.11890292e+00 6.08808160e-01 -9.49369371e-01 -1.00358081e+00
4.21586603e-01 2.99422681e-01 -5.12256384e-01 -8.80053490e-02
-8.33680749e-01 1.06984401e+00 4.85482723e-01 5.60390890e-01
-8.77612472e-01 -4.68526632e-01 -1.08034670e+00 -2.64668893e-02
3.78646016e-01 -7.18992725e-02 8.81855011e-01 -9.71446931e-02
-1.01751983e+00 7.76727378e-01 -2.24355474e-01 -5.65065563e-01
5.53255677e-01 -5.48363328e-01 -3.00974727e-01 -1.47996679e-01
5.29884398e-01 6.28408909e-01 4.32700634e-01 -1.43747008e+00
-8.76264095e-01 -6.83595598e-01 -6.36643637e-03 1.02324545e-01
-7.97116309e-02 -3.80171895e-01 -2.22471431e-01 -4.85025197e-01
5.28711438e-01 -1.23713720e+00 -1.89136609e-01 4.41702962e-01
-3.31274629e-01 -6.56352043e-02 1.15937126e+00 -4.36345428e-01
1.15108669e+00 -2.20594001e+00 -1.24728002e-01 3.32633883e-01
-8.52170140e-02 3.21131378e-01 -2.15848818e-01 7.07645833e-01
1.49824008e-01 -3.39439325e-02 -6.33374393e-01 -3.04702252e-01
8.31715316e-02 4.04928565e-01 -1.24146678e-01 7.32181072e-01
4.16013330e-01 7.39520848e-01 -6.46605074e-01 -2.48058766e-01
6.06401503e-01 5.61430275e-01 -9.47541073e-02 -2.73294803e-02
4.36638482e-02 2.05960482e-01 -4.61631358e-01 1.14249504e+00
1.34043598e+00 3.04558754e-01 2.99825184e-02 -1.79133639e-01
-7.10323930e-01 -8.83869529e-02 -1.53091311e+00 1.70728016e+00
-5.61769962e-01 5.21843135e-01 3.05686176e-01 -8.47485781e-01
1.33211112e+00 -6.47516325e-02 4.22517031e-01 -6.03580773e-01
-8.71639922e-02 4.99794900e-01 -2.39743680e-01 -2.26250254e-02
7.39520609e-01 4.09949332e-01 -1.76914632e-01 8.05176944e-02
2.12470200e-02 -4.52923626e-01 3.60326618e-01 -2.08179429e-01
9.91881311e-01 4.86666739e-01 5.43941379e-01 -3.86205077e-01
5.83866656e-01 2.06219584e-01 3.63755375e-01 7.66668200e-01
-2.75389284e-01 5.71329832e-01 2.71631718e-01 -3.45655054e-01
-5.93759954e-01 -1.18122876e+00 -6.23939335e-01 9.44463909e-01
4.23975110e-01 -3.71263951e-01 -4.08227801e-01 -5.30732214e-01
6.78341150e-01 5.23930311e-01 -5.90886772e-01 4.12590280e-02
-2.58844435e-01 -6.82117641e-01 5.92361569e-01 5.91786742e-01
7.23464906e-01 -9.35376942e-01 -9.54517603e-01 3.03324848e-01
-1.14897322e-02 -1.23503125e+00 1.42579675e-01 5.38700759e-01
-8.27470601e-01 -1.04845321e+00 -4.09035951e-01 -3.88632059e-01
2.75800496e-01 5.37236452e-01 7.66505539e-01 -1.39244035e-01
-4.50851798e-01 2.00308472e-01 -3.61935943e-01 -5.44384539e-01
-2.86505781e-02 5.46191394e-01 -6.02689870e-02 -1.10198572e-01
5.07153451e-01 -6.91759706e-01 -5.48029840e-01 4.25886810e-01
-8.41442525e-01 -6.13821864e-01 7.76003957e-01 6.03804767e-01
7.19436884e-01 -1.00153096e-01 2.34507963e-01 -4.28585619e-01
3.97669435e-01 -6.30192876e-01 -1.20773542e+00 1.91304207e-01
-3.81885827e-01 -2.57387221e-01 3.58559340e-01 1.11864068e-01
-5.77492058e-01 6.62953258e-01 -5.14679141e-02 -1.52468592e-01
-6.52295053e-01 7.56370723e-01 -3.25198799e-01 -4.12011147e-01
6.16856039e-01 1.92803368e-01 1.71192344e-02 -6.22635901e-01
3.50199491e-01 5.48731685e-01 5.38657129e-01 -3.66788775e-01
9.95148540e-01 6.33872569e-01 2.51846105e-01 -1.11810911e+00
-5.66446781e-01 -7.29371488e-01 -1.14753604e+00 -1.74250096e-01
6.12240374e-01 -1.16921103e+00 -6.29475892e-01 4.02470350e-01
-1.20354807e+00 -7.03315362e-02 -1.20601721e-01 4.21265781e-01
-3.98376942e-01 2.20045388e-01 1.78914648e-02 -7.99167395e-01
-1.67709500e-01 -1.09847689e+00 1.59640467e+00 2.91488707e-01
-2.64397431e-02 -8.12746644e-01 1.49677306e-01 -1.62924975e-01
2.39389747e-01 7.17453182e-01 3.49168748e-01 -5.45517504e-01
-8.52506757e-01 -5.72145343e-01 -3.81821185e-01 2.39794523e-01
2.95787424e-01 2.52707541e-01 -1.12800980e+00 -3.00527900e-01
-3.59008014e-01 4.42484021e-03 8.55175138e-01 2.53032237e-01
8.00533354e-01 2.96747953e-01 -4.50860381e-01 6.73250496e-01
1.78117919e+00 -5.42405248e-02 4.08317775e-01 5.75592160e-01
5.03658056e-01 5.59834301e-01 1.14096105e+00 7.16236532e-01
4.07052338e-01 7.96036661e-01 9.95211244e-01 -3.93967897e-01
2.01878369e-01 -2.30256185e-01 2.19510868e-01 1.80216819e-01
-5.22576421e-02 -1.23436153e-01 -1.05975425e+00 8.05648506e-01
-2.12146449e+00 -7.54574120e-01 -2.40851313e-01 2.22527027e+00
1.52789265e-01 -1.29893944e-01 -1.37937188e-01 1.23819910e-01
6.61321402e-01 4.29468066e-01 -4.42210197e-01 -2.55798638e-01
-2.75627762e-01 4.53953445e-01 1.12921906e+00 2.99288064e-01
-1.71958423e+00 1.19809246e+00 5.36000299e+00 7.49565661e-01
-1.33171976e+00 1.27155622e-02 -9.03826505e-02 3.21955264e-01
3.43124688e-01 1.02259368e-01 -9.53294814e-01 2.07585081e-01
5.01766741e-01 4.98181693e-02 1.07902020e-01 1.06802905e+00
4.37915623e-01 -6.34386539e-01 -9.21030700e-01 4.86925721e-01
-1.32088035e-01 -1.11851740e+00 -2.50210971e-01 2.63797939e-01
4.29191887e-01 2.97019958e-01 -4.66927961e-02 1.33936614e-01
2.62424260e-01 -7.95650482e-01 8.11688244e-01 2.36254185e-01
4.56235915e-01 -6.52771652e-01 8.96044075e-01 3.45580339e-01
-1.82401633e+00 -8.50833952e-02 -4.36531544e-01 3.34379822e-02
7.73558095e-02 6.29517317e-01 -8.24164093e-01 9.74087059e-01
9.44057107e-01 1.06793714e+00 -1.08388448e+00 1.44627643e+00
-3.62546176e-01 2.54528731e-01 -8.54167581e-01 2.55866289e-01
5.20270824e-01 -3.60634327e-01 5.42787910e-01 1.19443142e+00
7.00273931e-01 -2.67760336e-01 4.36918259e-01 1.07246184e+00
3.04732680e-01 1.83855787e-01 -1.20400977e+00 2.00969726e-01
6.25325322e-01 1.54652798e+00 -7.93140411e-01 2.36275390e-01
-1.32366061e-01 6.63149178e-01 1.11969374e-01 1.43337607e-01
-7.77562141e-01 -5.26055276e-01 7.61927724e-01 8.86512250e-02
5.82109392e-01 -5.02645612e-01 -2.50583202e-01 -7.38262951e-01
2.41223037e-01 -4.81339008e-01 7.51018301e-02 -6.39978290e-01
-9.10577655e-01 3.33485663e-01 1.32693619e-01 -1.55874002e+00
-3.10984366e-02 -6.91387951e-01 -6.67333782e-01 7.76399553e-01
-1.81751525e+00 -1.50395381e+00 -8.28580379e-01 2.97544062e-01
2.12264538e-01 1.05409898e-01 1.01195812e+00 3.80770504e-01
-4.89524841e-01 2.69618899e-01 3.51049811e-01 -1.70459226e-01
5.50464869e-01 -1.05471122e+00 4.00983393e-01 1.20868146e+00
9.27774757e-02 5.26329100e-01 5.19058347e-01 -6.33051991e-01
-1.25409222e+00 -1.40041196e+00 7.85096228e-01 -3.41290295e-01
6.76406860e-01 -5.63349843e-01 -9.01277065e-01 5.38183510e-01
-3.86725888e-02 2.42743433e-01 3.28155845e-01 -9.36421379e-02
1.81108173e-02 -4.28763837e-01 -1.19097865e+00 3.39551181e-01
1.07285070e+00 -5.92861950e-01 -1.62442520e-01 1.68800890e-01
2.34212130e-01 -4.20873255e-01 -9.85157669e-01 7.74792969e-01
6.17174327e-01 -1.02515912e+00 9.31573987e-01 3.12233239e-01
2.20605597e-01 -6.98893726e-01 -5.46639085e-01 -1.35263634e+00
-1.27655387e-01 -8.40060562e-02 2.24550694e-01 1.39912891e+00
2.13809296e-01 -7.63685286e-01 7.09586680e-01 -1.70227721e-01
-4.60178629e-02 -4.28804100e-01 -1.08758318e+00 -9.84717131e-01
-1.27763644e-01 -4.73336667e-01 7.66789436e-01 6.95827425e-01
-5.71302891e-01 -1.68858007e-01 4.76012081e-02 7.32169986e-01
5.22130847e-01 2.77943760e-01 9.51870441e-01 -1.51029027e+00
4.67772305e-01 -3.21679682e-01 -7.47504592e-01 -6.93647742e-01
2.55119473e-01 -9.14406717e-01 2.69182682e-01 -1.56849408e+00
-5.26105464e-01 -5.60891926e-01 -8.55363458e-02 6.33746803e-01
1.46370873e-01 3.37655932e-01 4.58620548e-01 1.11569278e-01
-2.17026442e-01 7.10975170e-01 7.21107662e-01 -2.10183874e-01
-2.72351205e-01 3.18138629e-01 -3.29077095e-01 5.54338336e-01
8.82864296e-01 -3.91549915e-01 -3.27481329e-02 -3.39500517e-01
-1.95149615e-01 -3.73962343e-01 7.39318550e-01 -1.37482131e+00
2.67592877e-01 7.28836805e-02 2.45716721e-01 -1.24969316e+00
4.98130083e-01 -1.23896253e+00 1.26043350e-01 5.17564893e-01
1.59377411e-01 2.98810840e-01 6.39637768e-01 4.15234387e-01
-3.96415919e-01 -1.60956264e-01 7.10639238e-01 7.16261864e-02
-1.15822387e+00 1.80227473e-01 -3.19472492e-01 -5.21915138e-01
1.20645797e+00 -3.01575840e-01 -2.74641901e-01 8.73147044e-03
-6.67583883e-01 3.97451013e-01 5.00666499e-01 6.50955737e-01
6.01860702e-01 -1.28765345e+00 -6.85588479e-01 4.21747029e-01
6.01452768e-01 1.77474096e-01 -1.02313116e-01 9.42542434e-01
-8.64311457e-01 4.73884434e-01 -3.36475968e-01 -1.14234996e+00
-1.21237075e+00 1.64168894e-01 2.67559022e-01 3.47601972e-03
-2.65975505e-01 3.92371118e-01 -6.76079988e-02 -8.96394074e-01
1.95809916e-01 -6.04144573e-01 -3.06749463e-01 5.43193817e-01
4.24770415e-02 3.35835695e-01 4.47658241e-01 -9.23720062e-01
-7.44879127e-01 9.56552505e-01 2.63365149e-01 1.23913653e-01
1.50561857e+00 -4.98193540e-02 -2.16933146e-01 2.89100468e-01
9.27528024e-01 -8.66975784e-02 -1.06073070e+00 -8.48712176e-02
4.74216849e-01 -5.70249677e-01 9.73077267e-02 -4.84924257e-01
-1.02504957e+00 8.53759348e-01 1.25058496e+00 1.10490903e-01
9.63162780e-01 -2.08400965e-01 1.85888410e-01 4.41441238e-01
5.06085753e-01 -8.39051723e-01 -5.27807951e-01 5.39376259e-01
9.64238405e-01 -1.39815283e+00 2.49105021e-01 -4.78526562e-01
-3.10340911e-01 9.59192812e-01 5.03564835e-01 -2.54469067e-01
8.74513149e-01 2.40016162e-01 9.72960070e-02 -2.18264475e-01
-2.38682047e-01 -7.27472007e-01 -9.80915688e-03 6.99320555e-01
2.74753779e-01 2.23961398e-01 -3.36958319e-01 1.31842315e-01
-1.33736506e-01 1.20591089e-01 2.82123387e-01 1.27330959e+00
-4.03629810e-01 -1.07545936e+00 -6.12085164e-01 9.75371674e-02
1.29743725e-01 5.29020354e-02 -5.29452920e-01 1.13659072e+00
5.10894477e-01 7.34543025e-01 -2.19286736e-02 -2.31733367e-01
7.99671650e-01 -1.78133965e-01 4.05314192e-02 -6.09849393e-01
-6.21128500e-01 -9.90256816e-02 -2.73291439e-01 -7.84714878e-01
-4.24371749e-01 -8.58502150e-01 -8.97571266e-01 -2.80751914e-01
-5.27908146e-01 -1.17555290e-01 9.79939342e-01 6.39222324e-01
4.05413538e-01 2.74127722e-01 6.20482385e-01 -1.24007571e+00
-4.86282706e-01 -7.45767593e-01 -7.92503417e-01 -9.70723107e-02
3.80269468e-01 -9.99890387e-01 -2.55620629e-01 -4.33071166e-01]
|
[7.555147647857666, -2.002532720565796]
|
fabbd9c1-62c6-4011-9f38-2c4e6b8c8292
|
mucpad-a-multi-domain-chinese-predicate
|
2205.06703
| null |
https://arxiv.org/abs/2205.06703v1
|
https://arxiv.org/pdf/2205.06703v1.pdf
|
MuCPAD: A Multi-Domain Chinese Predicate-Argument Dataset
|
During the past decade, neural network models have made tremendous progress on in-domain semantic role labeling (SRL). However, performance drops dramatically under the out-of-domain setting. In order to facilitate research on cross-domain SRL, this paper presents MuCPAD, a multi-domain Chinese predicate-argument dataset, which consists of 30,897 sentences and 92,051 predicates from six different domains. MuCPAD exhibits three important features. 1) Based on a frame-free annotation methodology, we avoid writing complex frames for new predicates. 2) We explicitly annotate omitted core arguments to recover more complete semantic structure, considering that omission of content words is ubiquitous in multi-domain Chinese texts. 3) We compile 53 pages of annotation guidelines and adopt strict double annotation for improving data quality. This paper describes in detail the annotation methodology and annotation process of MuCPAD, and presents in-depth data analysis. We also give benchmark results on cross-domain SRL based on MuCPAD.
|
['Min Zhang', 'Zhenghua Li', 'Qingrong Xia', 'Chen Gong', 'Haoping Yang', 'Yahui Liu']
|
2022-05-13
| null |
https://aclanthology.org/2022.naacl-main.123
|
https://aclanthology.org/2022.naacl-main.123.pdf
|
naacl-2022-7
|
['semantic-role-labeling']
|
['natural-language-processing']
|
[ 4.94264036e-01 3.95318210e-01 -7.64250219e-01 -6.23027205e-01
-7.70137668e-01 -9.73516285e-01 4.47649002e-01 2.96298206e-01
-5.51628530e-01 1.26902616e+00 6.95210397e-01 -2.42732778e-01
-1.63635641e-01 -6.07461095e-01 -5.88779032e-01 -1.69967934e-01
3.95548999e-01 5.99550724e-01 4.81192589e-01 -4.04378116e-01
-1.64810978e-02 -4.84896041e-02 -1.31262541e+00 8.46660674e-01
8.71894717e-01 7.84836769e-01 2.99110085e-01 9.72093120e-02
-4.93887156e-01 1.30134964e+00 -9.53125000e-01 -7.06167936e-01
-2.03872740e-01 -3.15003812e-01 -1.48541045e+00 -2.01118231e-01
6.82685450e-02 1.69591591e-01 -1.32204771e-01 9.90185738e-01
2.71031648e-01 2.66299732e-02 2.64795989e-01 -1.23261297e+00
-1.05807781e+00 1.33510101e+00 -1.14762947e-01 1.11134812e-01
4.73930210e-01 -7.00458229e-01 1.59943438e+00 -5.02782941e-01
9.54619527e-01 1.47458410e+00 6.49224043e-01 8.13393593e-01
-8.74044359e-01 -7.03896344e-01 3.30535114e-01 2.94833302e-01
-1.14771438e+00 -2.92254925e-01 9.87441659e-01 -1.28751263e-01
1.02793455e+00 1.29429862e-01 2.84658730e-01 1.20081723e+00
-4.44438279e-01 1.09753263e+00 7.91198552e-01 -6.83089614e-01
-2.45430786e-02 -7.61844665e-02 4.45813924e-01 2.99328685e-01
3.25286776e-01 -5.21176875e-01 -4.20446277e-01 -2.53639102e-01
6.06943369e-01 -6.40309155e-01 -1.39309421e-01 -1.52962089e-01
-1.09753454e+00 9.59800363e-01 -1.23678401e-01 7.62464702e-01
2.20717818e-01 -7.65705705e-02 8.31413090e-01 2.92265594e-01
5.76150656e-01 5.69773316e-01 -1.14891553e+00 -3.89193594e-01
-2.60747194e-01 3.55205745e-01 8.31646860e-01 1.19356656e+00
2.89385110e-01 -1.25688076e-01 -2.17880048e-02 1.47190964e+00
1.02404840e-01 1.76088378e-01 5.25386631e-01 -1.40783966e+00
9.42756414e-01 7.97844052e-01 2.56649077e-01 -7.60427713e-01
-3.55957180e-01 -4.53166217e-02 -3.44616026e-01 -6.59358144e-01
4.30967063e-01 -4.37854491e-02 -2.86489964e-01 1.87629163e+00
1.47754783e-02 -3.52815300e-01 5.73644757e-01 6.20584130e-01
1.30243909e+00 3.63933086e-01 5.45146525e-01 3.23284827e-02
1.70762086e+00 -9.11407173e-01 -1.04048121e+00 -3.55425835e-01
9.92431819e-01 -4.77881283e-01 1.50854015e+00 1.39392331e-01
-9.61454511e-01 -5.03783703e-01 -1.06789100e+00 -5.03935635e-01
-6.17303729e-01 4.21206683e-01 6.14167392e-01 5.14574945e-01
-4.79043394e-01 3.38534206e-01 -5.04919171e-01 -2.14539140e-01
4.45423365e-01 -1.77221056e-02 -4.10668701e-01 -5.01969866e-02
-2.00331879e+00 7.49685645e-01 1.12244236e+00 -4.57623452e-01
-3.81845832e-01 -7.33684301e-01 -1.19044459e+00 -5.09493612e-02
7.92529345e-01 -8.87421072e-02 1.58408594e+00 -9.84791815e-01
-1.17213178e+00 1.31893826e+00 -3.58632267e-01 -5.62407017e-01
2.89912641e-01 -5.97074211e-01 -7.81703055e-01 1.79905936e-01
5.54405272e-01 6.80774570e-01 3.28065157e-01 -1.10828507e+00
-1.05542123e+00 -4.18316312e-02 5.17784476e-01 2.95063257e-01
-2.44292200e-01 6.74090624e-01 -4.79387552e-01 -9.83943224e-01
-1.11188263e-01 -6.22466683e-01 1.63080037e-01 -5.20951748e-01
-1.64335519e-01 -8.66732121e-01 9.37249482e-01 -5.54762781e-01
1.39715993e+00 -2.15073466e+00 -2.54346341e-01 -6.03340685e-01
1.23793622e-02 3.89147341e-01 3.19010615e-02 3.80722582e-01
-3.24922830e-01 3.41325194e-01 -4.40508455e-01 -2.41532758e-01
1.14287764e-01 8.41605723e-01 -6.12282097e-01 8.94512385e-02
4.06290323e-01 8.31246376e-01 -1.09160721e+00 -6.95742905e-01
-7.99436867e-02 4.72002886e-02 -4.56811100e-01 -2.14085340e-01
-6.26415133e-01 4.84302580e-01 -5.03213227e-01 8.19390237e-01
5.33797681e-01 -3.51276934e-01 8.02078784e-01 -6.45859390e-02
-4.78996299e-02 1.02578318e+00 -8.92344952e-01 1.73570895e+00
-4.00956690e-01 5.65424323e-01 -1.60629541e-01 -1.11821306e+00
8.99740160e-01 5.46423376e-01 5.86557329e-01 -7.59089708e-01
-9.05705523e-03 3.97693723e-01 -8.20061415e-02 -2.39745751e-01
6.93610191e-01 -4.35023103e-03 -6.74846351e-01 1.86988696e-01
-1.16357937e-01 -5.09168096e-02 7.25322723e-01 -7.40958229e-02
8.95959616e-01 3.01901251e-01 7.20944405e-01 -4.61602658e-01
1.03912807e+00 4.41317707e-01 1.11361134e+00 5.05034149e-01
-2.48273462e-01 3.16332966e-01 9.06961083e-01 -4.70917732e-01
-8.45565736e-01 -7.20031142e-01 -3.88144553e-01 1.35188341e+00
1.74403265e-01 -5.36930203e-01 -5.01501203e-01 -1.16685843e+00
4.96574212e-03 8.60206783e-01 -3.46656233e-01 3.73049200e-01
-1.21602964e+00 -7.20054388e-01 1.15270472e+00 7.77321517e-01
8.90264273e-01 -1.36615956e+00 -4.30919915e-01 6.34581923e-01
-6.93744659e-01 -1.74344289e+00 4.52890247e-02 1.84825018e-01
-4.34053540e-01 -1.50961232e+00 -2.95163840e-01 -1.28204560e+00
1.89212814e-01 3.73681597e-02 1.55489314e+00 7.33413324e-02
2.21617907e-01 -4.79718037e-02 -8.10651124e-01 -4.94095027e-01
-5.32056451e-01 1.39393091e-01 -2.10838392e-01 -6.87392354e-01
7.79230595e-01 -1.33085623e-01 1.31486475e-01 4.23419446e-01
-6.89992189e-01 -1.25493318e-01 -6.09985217e-02 8.47894967e-01
6.67358100e-01 4.54858929e-01 8.45531285e-01 -1.20335376e+00
6.89349532e-01 -3.71777058e-01 -7.71564960e-01 3.43123853e-01
-6.85759783e-02 -8.24971721e-02 6.68986976e-01 -1.74776345e-01
-1.57844186e+00 -1.97196156e-01 -4.15895760e-01 3.30312282e-01
-2.27204874e-01 4.43641782e-01 -6.02383852e-01 5.90611637e-01
5.26655436e-01 -1.38365895e-01 -2.73373991e-01 -8.27575088e-01
3.45668554e-01 6.97412670e-01 6.02091372e-01 -1.26815355e+00
3.68346155e-01 3.90375346e-01 -4.24739510e-01 -7.42177784e-01
-1.60482597e+00 -3.62784028e-01 -7.27838755e-01 4.06311572e-01
9.11785841e-01 -1.12558913e+00 -6.50193393e-01 4.11811262e-01
-1.31132901e+00 -5.86891353e-01 -2.60162264e-01 1.21335797e-01
-4.00029063e-01 5.94756186e-01 -8.32031786e-01 -3.84546250e-01
-1.51637211e-01 -7.64219463e-01 8.87759507e-01 1.73305608e-02
-4.72677976e-01 -1.25440669e+00 -2.31499791e-01 5.12717783e-01
-4.84782234e-02 5.15428223e-02 1.27179027e+00 -1.06758845e+00
8.29696655e-02 3.62052798e-01 -4.70524818e-01 6.06562972e-01
4.31942582e-01 -3.86659771e-01 -7.62123883e-01 8.12990740e-02
-1.27303660e-01 -4.23917264e-01 4.92733747e-01 1.23501725e-01
1.50715888e+00 -2.31404528e-01 -3.53944153e-01 3.16391401e-02
1.22839105e+00 5.61404705e-01 4.41409469e-01 8.16937208e-01
6.74797475e-01 7.94523954e-01 1.30289030e+00 1.89407870e-01
6.73200190e-01 5.51098168e-01 1.48565874e-01 1.38906851e-01
-2.31196195e-01 -4.02002901e-01 1.86882839e-01 4.73043799e-01
2.40554512e-01 -5.78173995e-01 -1.04514420e+00 7.77493179e-01
-1.87868500e+00 -7.65106559e-01 -2.46316776e-01 1.59146011e+00
1.25664496e+00 3.94220293e-01 5.06532900e-02 3.15787941e-01
7.93830931e-01 3.43236625e-01 -1.76112771e-01 -2.09474578e-01
-6.86248481e-01 3.21553588e-01 4.19445008e-01 3.60199541e-01
-1.51793921e+00 1.59779060e+00 6.18188620e+00 9.65554357e-01
-5.62708616e-01 3.45776230e-01 3.49162728e-01 5.35891294e-01
-2.51838714e-01 9.20882747e-02 -1.38915563e+00 5.73825479e-01
5.62237203e-01 -9.23161656e-02 -1.28974468e-01 1.14392996e+00
-8.28095302e-02 1.27552196e-01 -8.89645576e-01 7.62819052e-01
-1.45183235e-01 -1.65768385e+00 -1.19286245e-02 -3.42041194e-01
4.62356716e-01 2.98194494e-02 -3.60207707e-01 5.13671696e-01
6.88906431e-01 -8.10123742e-01 8.62325311e-01 -3.75736475e-01
9.60251331e-01 -7.63047814e-01 8.68728578e-01 2.57325828e-01
-1.28201103e+00 -1.85111701e-01 -3.83638114e-01 -2.42565632e-01
2.88783073e-01 3.90929043e-01 -5.35402954e-01 6.64103329e-01
8.77613544e-01 1.14440608e+00 -3.53789985e-01 2.56176502e-01
-8.65227520e-01 7.55219996e-01 -1.95481293e-02 4.36793156e-02
3.32805425e-01 1.38048260e-02 3.81146073e-01 1.34912741e+00
-1.27896592e-01 1.41952246e-01 2.07339242e-01 5.83196402e-01
-2.09077954e-01 -9.41143837e-03 -3.35801125e-01 -3.04802507e-01
9.51151371e-01 7.71160305e-01 -6.59549356e-01 -5.16197681e-01
-6.22800767e-01 6.76039398e-01 2.79156804e-01 1.11757122e-01
-1.03654468e+00 -3.72742653e-01 9.96084869e-01 -2.79915258e-02
2.20944807e-01 -2.16629729e-01 -3.55185062e-01 -9.75168169e-01
1.49897650e-01 -1.01586747e+00 9.66401219e-01 -6.59045577e-01
-1.43242311e+00 5.03000975e-01 4.18967098e-01 -1.06849253e+00
-1.00713156e-01 -9.09388065e-01 3.97414491e-02 5.79084694e-01
-1.85069001e+00 -1.23495901e+00 1.46696210e-01 4.77455318e-01
8.70843351e-01 -1.80345654e-01 8.13843429e-01 6.73452199e-01
-3.25190276e-01 5.19337416e-01 -2.72955596e-01 6.88745737e-01
7.59469748e-01 -1.20241284e+00 5.80508530e-01 7.13187516e-01
-2.36918628e-02 6.91211700e-01 4.85156178e-01 -8.44732642e-01
-5.10152221e-01 -1.01528490e+00 1.50950491e+00 -7.38708854e-01
9.23392773e-01 -3.90129358e-01 -1.18516779e+00 9.11032319e-01
1.16372548e-01 -8.92190784e-02 7.03377664e-01 1.30715355e-01
-2.95267940e-01 2.25779235e-01 -1.09995484e+00 4.91160363e-01
1.46482110e+00 -6.57856762e-01 -1.32553899e+00 3.79188955e-01
1.38216865e+00 -6.44788742e-01 -9.85915780e-01 6.34591639e-01
1.77894071e-01 -4.70243484e-01 9.84149277e-01 -7.59818494e-01
4.56580698e-01 -5.13351202e-01 -4.24893618e-01 -8.19284499e-01
-1.10882215e-01 -3.08542699e-01 4.60433774e-02 1.53847015e+00
5.60709596e-01 -6.20711505e-01 6.99504673e-01 4.32250798e-01
-5.38876235e-01 -2.99352109e-01 -9.16736484e-01 -9.34593678e-01
1.69220164e-01 -6.82168305e-01 8.25048387e-01 1.23174667e+00
2.05183290e-02 4.76652354e-01 -1.06372647e-01 -1.02909051e-01
2.67455280e-01 8.54974538e-02 2.67672449e-01 -1.40492833e+00
-2.08421588e-01 -1.01349659e-01 2.20911339e-01 -1.15046537e+00
9.14450884e-01 -7.77875602e-01 -2.35615045e-01 -1.45320141e+00
-9.27313641e-02 -8.83101046e-01 -1.61658645e-01 1.12081313e+00
-5.78288175e-02 2.77017634e-02 -1.27743885e-01 1.88254938e-01
-6.88162208e-01 4.27149802e-01 1.05823946e+00 6.19681142e-02
-1.14547297e-01 -2.92534053e-01 -9.58931327e-01 1.06221092e+00
1.02605700e+00 -6.58705413e-01 -5.82673252e-01 -7.41986811e-01
3.43575358e-01 -2.48552412e-01 4.21637930e-02 -6.78173363e-01
-3.84246707e-01 -3.40839028e-01 1.04399554e-01 -6.91109955e-01
3.35642308e-01 -7.47662544e-01 -3.09685081e-01 2.55825520e-01
-4.37793463e-01 -5.42350113e-03 2.41951063e-01 2.00032055e-01
-7.30836987e-01 -5.63720167e-01 7.23591745e-01 -4.34920430e-01
-1.43828356e+00 -2.81589001e-01 -3.11970264e-01 7.43272424e-01
8.07254791e-01 -1.57282218e-01 -7.20371187e-01 7.86833093e-02
-5.47754884e-01 2.70515352e-01 4.23710316e-01 8.61023545e-01
1.79032385e-01 -1.21185899e+00 -5.76067984e-01 -5.31764664e-02
3.20273578e-01 1.76664323e-01 -2.29572654e-02 1.49733827e-01
-5.18202960e-01 7.65956759e-01 4.77804756e-03 -2.42708698e-01
-1.44218922e+00 2.90512055e-01 -3.44782099e-02 -5.40052533e-01
-6.89203441e-01 8.51389706e-01 2.84780767e-02 -7.38591075e-01
5.68153933e-02 -2.39413619e-01 -7.29489565e-01 1.59849852e-01
3.76925170e-01 9.39555839e-02 2.41565591e-04 -6.63157523e-01
-5.63949466e-01 2.08468542e-01 -7.39601720e-03 -5.77358380e-02
1.30700278e+00 -2.09212109e-01 -4.09013867e-01 2.93937773e-01
9.29453611e-01 2.51068234e-01 -9.90640700e-01 -4.61203933e-01
7.70709336e-01 -2.71422654e-01 -4.85390484e-01 -9.55747485e-01
-4.87571985e-01 3.50483686e-01 -1.41907647e-01 7.44925737e-02
9.78406429e-01 3.77383858e-01 9.39995825e-01 4.34924543e-01
3.15448135e-01 -1.72594249e+00 1.66008039e-03 1.34436905e+00
7.78329849e-01 -1.09240830e+00 -1.70851052e-01 -9.33565140e-01
-8.51655006e-01 8.63955915e-01 7.53149271e-01 2.08731368e-01
3.80404353e-01 2.16822669e-01 8.77896175e-02 -2.29377031e-01
-5.56878090e-01 -1.23149142e-01 -5.97362183e-02 7.10378587e-01
8.20390582e-01 8.41827616e-02 -6.85920298e-01 1.00715256e+00
-4.48270589e-01 -1.45923615e-01 4.31990445e-01 1.17834520e+00
-2.79194206e-01 -1.64728022e+00 -1.06554531e-01 -8.43641236e-02
-9.41498280e-01 -2.60541409e-01 -4.09662187e-01 1.30183232e+00
3.35783064e-01 1.10992801e+00 1.29971862e-01 2.94692218e-01
2.46090963e-01 2.03543007e-01 2.42324084e-01 -8.80410314e-01
-2.60427177e-01 -3.48545194e-01 1.08813787e+00 -3.83965850e-01
-9.10911918e-01 -5.92244506e-01 -1.79806685e+00 1.35464221e-01
9.15886238e-02 3.81418020e-01 4.37200487e-01 1.08451629e+00
2.53086448e-01 5.14297724e-01 -1.18993334e-01 -1.37826512e-02
-1.60534680e-01 -8.09945226e-01 -3.63559961e-01 4.49634761e-01
-7.78798461e-02 -1.00287306e+00 -1.65295631e-01 8.08076411e-02]
|
[10.254103660583496, 9.321586608886719]
|
a9332e96-ffb2-4b33-a7d1-806056aa3e51
|
personalization-disentanglement-for-federated
|
2306.0357
| null |
https://arxiv.org/abs/2306.03570v1
|
https://arxiv.org/pdf/2306.03570v1.pdf
|
Personalization Disentanglement for Federated Learning
|
Personalized federated learning (PFL) jointly trains a variety of local models through balancing between knowledge sharing across clients and model personalization per client. This paper addresses PFL via explicit disentangling latent representations into two parts to capture the shared knowledge and client-specific personalization, which leads to more reliable and effective PFL. The disentanglement is achieved by a novel Federated Dual Variational Autoencoder (FedDVA), which employs two encoders to infer the two types of representations. FedDVA can produce a better understanding of the trade-off between global knowledge sharing and local personalization in PFL. Moreover, it can be integrated with existing FL methods and turn them into personalized models for heterogeneous downstream tasks. Extensive experiments validate the advantages caused by disentanglement and show that models trained with disentangled representations substantially outperform those vanilla methods.
|
['Guodong Long', 'Peng Yan']
|
2023-06-06
| null | null | null | null |
['personalized-federated-learning', 'disentanglement']
|
['methodology', 'methodology']
|
[-4.80145276e-01 2.56862849e-01 -6.30212605e-01 -5.27755380e-01
-7.98423529e-01 -5.97290695e-01 7.29156256e-01 -5.88897645e-01
-4.55190316e-02 7.97793329e-01 8.46182942e-01 -2.99584139e-02
-2.82033682e-01 -7.71415412e-01 -6.07552946e-01 -8.89475286e-01
2.38726616e-01 6.09451115e-01 -3.44952136e-01 2.17347685e-02
-5.21068394e-01 3.77159476e-01 -1.23600459e+00 5.49180031e-01
7.70397782e-01 7.70331502e-01 1.17562369e-01 4.15092438e-01
-2.17258513e-01 7.85865486e-01 -6.41084090e-02 -9.18994129e-01
2.63196856e-01 -3.46709639e-02 -9.54737902e-01 -1.07754670e-01
1.22514777e-01 -6.16736531e-01 -6.91645145e-01 6.81330323e-01
4.22339827e-01 2.38992244e-01 6.84160948e-01 -1.35844886e+00
-1.30078781e+00 1.15718246e+00 -3.58227134e-01 9.38811079e-02
-2.14217708e-01 3.70426029e-01 1.61420548e+00 -5.94793081e-01
6.01911724e-01 1.33746445e+00 5.26149988e-01 6.74462080e-01
-1.61365569e+00 -7.63865888e-01 3.34171921e-01 3.23612660e-01
-1.10625601e+00 -5.10732234e-01 6.25951827e-01 -6.66012406e-01
9.50213075e-01 9.31123793e-02 3.63625944e-01 1.72966504e+00
-1.77764386e-01 1.09168077e+00 7.07149386e-01 7.01260790e-02
-1.71516894e-03 5.16708970e-01 2.92331994e-01 6.84865296e-01
3.01125616e-01 2.78677315e-01 -6.79225922e-01 -4.06703025e-01
7.97442436e-01 5.10332227e-01 -5.95896900e-01 -7.92842865e-01
-9.17965770e-01 1.03857768e+00 3.45465660e-01 1.93238899e-01
-4.39816207e-01 2.11827353e-01 4.89897102e-01 3.37711304e-01
4.89293903e-01 5.38167715e-01 -9.38304901e-01 -2.71114353e-02
-8.47240329e-01 2.16499195e-01 7.59189308e-01 8.34957421e-01
1.12139630e+00 -5.14952727e-02 -8.47713768e-01 7.71468699e-01
5.79873383e-01 2.29307171e-02 7.06498086e-01 -1.11558938e+00
4.31840122e-01 6.68605745e-01 -7.48443007e-02 -4.37444270e-01
9.56716910e-02 -5.55806875e-01 -4.14887995e-01 -3.89359402e-03
-5.27097434e-02 -3.23502690e-01 -6.11376882e-01 2.17174673e+00
2.02530250e-01 3.78915846e-01 2.01188698e-01 6.31746590e-01
7.15449035e-01 4.60446388e-01 2.46029571e-01 2.01726407e-02
1.15653968e+00 -1.46305668e+00 -6.93509281e-01 5.97748123e-02
5.27048409e-01 -1.91557765e-01 7.27522969e-01 3.28328982e-02
-9.73426700e-01 -2.89037973e-01 -8.36358190e-01 -3.23008388e-01
-1.52431592e-01 -1.04244016e-01 9.80826914e-01 5.65130293e-01
-1.14246392e+00 7.33227730e-01 -1.02817619e+00 -2.02097729e-01
9.22384024e-01 4.39185262e-01 -4.88813818e-01 2.07156967e-02
-1.21164703e+00 7.94223309e-01 1.42370671e-01 -2.02130407e-01
-1.16636217e+00 -1.10848808e+00 -6.44945920e-01 7.40752101e-01
3.33531857e-01 -1.27556729e+00 1.51222312e+00 -8.99484515e-01
-1.78248465e+00 5.88575065e-01 -1.75925300e-01 -2.77112573e-01
6.34581447e-01 -2.12941900e-01 -2.12943137e-01 -3.79585445e-01
-9.98421833e-02 9.54568014e-02 8.29903126e-01 -1.14851320e+00
-4.76588011e-01 -5.59127629e-01 7.42188394e-02 1.58315420e-01
-5.93315721e-01 -4.12436217e-01 -6.06573462e-01 -3.51301312e-01
-6.00579202e-01 -4.84880894e-01 -3.67296673e-02 -9.56633538e-02
-2.75795847e-01 -5.52040577e-01 8.29425633e-01 -6.49204612e-01
1.09432721e+00 -1.97623849e+00 6.24744713e-01 -1.40809134e-01
7.62113035e-01 2.35810801e-01 -3.62045676e-01 6.84153259e-01
4.37998958e-02 1.25490785e-01 3.69902998e-01 -9.44596529e-01
2.92411327e-01 5.15007436e-01 -3.83533061e-01 2.89104789e-01
-4.50050905e-02 1.39753604e+00 -8.19531441e-01 -2.87740052e-01
-1.76689684e-01 7.80417144e-01 -8.74694824e-01 6.39429033e-01
-4.30424511e-01 4.14762795e-01 -8.61701846e-01 3.61885488e-01
7.02185035e-01 -4.91525620e-01 7.09738672e-01 -4.05861199e-01
1.64448872e-01 3.17463696e-01 -6.41937077e-01 1.76227736e+00
-6.50614321e-01 4.35667545e-01 2.35994264e-01 -7.63309956e-01
5.06899118e-01 5.93260884e-01 5.11835039e-01 -4.54313844e-01
2.70834386e-01 -3.78502637e-01 -6.66690528e-01 -7.38557398e-01
3.49706501e-01 -1.27919599e-01 1.82049736e-01 9.45901513e-01
6.29667699e-01 7.88554013e-01 -4.15780365e-01 4.67627138e-01
9.57194448e-01 3.57174784e-01 -2.92346487e-03 -1.70981079e-01
1.42107069e-01 -5.99065661e-01 8.89404833e-01 6.97099984e-01
-3.80198926e-01 2.25575566e-01 5.80898583e-01 -2.71395147e-01
-6.60955369e-01 -1.29515219e+00 1.86621577e-01 1.57416308e+00
-1.41496018e-01 -3.65059167e-01 -4.61946070e-01 -1.04227698e+00
5.12028098e-01 9.10367668e-01 -8.61591101e-01 -5.43984532e-01
-2.72871763e-01 -4.36298311e-01 4.60854232e-01 8.10363412e-01
1.29693881e-01 -6.41592383e-01 3.82791534e-02 -4.49490137e-02
-2.08006829e-01 -7.53348708e-01 -6.73165441e-01 -2.24629287e-02
-7.45110750e-01 -1.03247297e+00 -6.98628604e-01 -2.99433529e-01
1.00092977e-01 4.15147394e-01 1.11258435e+00 -3.75037223e-01
1.44064620e-01 6.28695130e-01 -2.57131577e-01 6.98477700e-02
-2.12386578e-01 4.42946851e-01 1.62010521e-01 4.35385048e-01
3.77442718e-01 -9.56184506e-01 -6.99964166e-01 5.23874126e-02
-6.61590993e-01 1.13220520e-01 6.20595634e-01 1.01294911e+00
3.08647513e-01 -4.53565747e-01 6.48183584e-01 -1.31646323e+00
8.40289831e-01 -1.01529121e+00 -3.05570751e-01 7.16953695e-01
-1.10367477e+00 6.64678931e-01 5.07083714e-01 -4.19816107e-01
-1.82114470e+00 -3.80230814e-01 7.80052878e-03 -9.28684235e-01
1.94613665e-01 2.89665163e-01 -4.83598977e-01 2.80785650e-01
4.32386339e-01 8.67656711e-03 3.77194658e-02 -9.03145254e-01
1.25953007e+00 6.30795658e-01 2.22388089e-01 -6.98797882e-01
4.97130573e-01 4.70457494e-01 -7.54005849e-01 -6.96776360e-02
-8.22852433e-01 -2.32188717e-01 -3.52424473e-01 2.34026834e-01
8.95708680e-01 -1.09249723e+00 -1.05867147e+00 1.01733454e-01
-8.78973067e-01 -3.36405754e-01 -5.41158080e-01 3.96223664e-01
-5.79178631e-01 2.24878937e-01 -7.06654727e-01 -5.18924236e-01
-5.70794284e-01 -1.22119999e+00 8.30156684e-01 3.32665324e-01
-7.24924132e-02 -1.30900407e+00 5.88441253e-01 7.49693751e-01
6.15901053e-01 -2.20581621e-01 1.14060020e+00 -9.54401791e-01
-8.31981719e-01 -1.68183222e-02 -2.99251854e-01 3.27511936e-01
2.19448805e-01 -1.65381670e-01 -1.38736689e+00 -3.57990712e-01
-2.99965024e-01 -4.17690963e-01 1.22445333e+00 1.61515549e-01
1.16791832e+00 -7.41983116e-01 -6.23715281e-01 1.03255284e+00
1.28894448e+00 -3.37642521e-01 2.40721777e-01 -4.85864058e-02
9.32115376e-01 3.15108329e-01 -2.54390180e-01 5.64012527e-01
9.00718451e-01 7.94106960e-01 5.07577717e-01 3.91424835e-01
-1.25729769e-01 -5.44828594e-01 3.47432047e-01 7.01055884e-01
-1.22105844e-01 -2.00172856e-01 -3.86058688e-01 5.29199421e-01
-2.26513052e+00 -1.08672905e+00 3.61744344e-01 1.74827206e+00
8.48337412e-01 -6.54775560e-01 -2.89466581e-03 -6.14227831e-01
5.12461483e-01 4.69377875e-01 -8.45177293e-01 -3.98271233e-01
1.29710659e-01 1.14486434e-01 3.54919165e-01 5.75082064e-01
-5.87826610e-01 9.61855829e-01 6.03657246e+00 6.57889903e-01
-8.19447458e-01 8.11414123e-01 3.68592650e-01 -5.33093214e-01
-9.87879455e-01 -4.42257300e-02 -7.90613651e-01 3.51271272e-01
8.43206286e-01 -3.83709669e-01 7.32241750e-01 9.96422291e-01
-2.14478727e-02 7.83959806e-01 -1.51831686e+00 8.10269952e-01
-1.03502654e-01 -1.65344226e+00 2.53751665e-01 2.66506046e-01
9.85983074e-01 4.25384641e-01 3.08841825e-01 8.17378998e-01
1.18515909e+00 -9.42455113e-01 4.75456595e-01 9.96385872e-01
4.09799337e-01 -6.62891686e-01 4.08239454e-01 3.57957095e-01
-8.68913770e-01 -3.29994291e-01 -2.01823384e-01 3.13538849e-01
2.09387794e-01 2.62508124e-01 -7.92825043e-01 7.99988747e-01
6.21735692e-01 6.95974290e-01 -9.10467431e-02 5.48406243e-01
-1.88291445e-01 4.59398538e-01 2.03812957e-01 3.41502786e-01
-2.04181038e-02 -1.95545137e-01 3.13016981e-01 9.41631138e-01
8.02107900e-02 -1.27876714e-01 -1.56637743e-01 1.41712236e+00
-4.61058438e-01 -1.53678104e-01 -2.93468148e-01 -3.68335932e-01
4.88619477e-01 1.20925093e+00 3.91551852e-01 -3.35581273e-01
-5.28032064e-01 1.35672021e+00 1.20327926e+00 9.73901033e-01
-6.74080789e-01 1.78438023e-01 1.66512597e+00 -2.10617736e-01
6.22271955e-01 7.54526556e-02 -7.45475963e-02 -1.83791065e+00
-4.78666246e-01 -5.04429519e-01 5.98087966e-01 -4.69341338e-01
-1.82189405e+00 5.74719667e-01 -2.69910008e-01 -4.23040301e-01
-2.75672823e-01 -3.27984363e-01 -8.46055567e-01 1.29046047e+00
-1.72492778e+00 -1.67123818e+00 -6.66825399e-02 8.81545424e-01
3.90599191e-01 -2.80728489e-01 9.85963404e-01 3.03415745e-01
-1.14798415e+00 1.13302922e+00 6.08473659e-01 -1.94844365e-01
7.06435025e-01 -9.23049510e-01 1.38940424e-01 4.48702902e-01
1.52548835e-01 7.95422196e-01 3.66571337e-01 -4.10024405e-01
-1.44966269e+00 -1.42762744e+00 8.41553628e-01 -8.85554016e-01
5.50102472e-01 -2.26701483e-01 -9.00899112e-01 1.35373247e+00
4.12105620e-01 -6.61899596e-02 1.09430873e+00 6.41535282e-01
-9.09341455e-01 -1.69281155e-01 -1.00057638e+00 4.26294565e-01
1.06697106e+00 -8.85145962e-01 -4.29236263e-01 1.92947522e-01
1.25296068e+00 8.17532390e-02 -9.89481390e-01 3.76288742e-02
7.36199260e-01 -1.11536574e+00 8.65083992e-01 -1.29727292e+00
3.98494601e-01 4.29898590e-01 -2.91718394e-01 -1.39458406e+00
-8.35130990e-01 -5.96719027e-01 -8.33831012e-01 1.53176200e+00
3.64884645e-01 -8.75560462e-01 8.56238484e-01 8.77822280e-01
9.55195278e-02 -1.02669764e+00 -6.16794586e-01 -5.15211642e-01
2.29120269e-01 -3.49406391e-01 1.19094861e+00 1.25919080e+00
1.04595140e-01 3.77137572e-01 -5.78966379e-01 2.09696278e-01
5.99172115e-01 2.74631858e-01 5.75015008e-01 -1.28787494e+00
-8.73204231e-01 -7.41495013e-01 -3.50988097e-02 -1.18730569e+00
6.93801880e-01 -1.35145211e+00 -6.21133387e-01 -1.56994188e+00
5.85104048e-01 -3.45948964e-01 -6.51714563e-01 7.69910097e-01
-2.81350821e-01 -3.43110800e-01 -8.20459519e-03 4.18986201e-01
-5.01955450e-01 1.13046753e+00 1.07351589e+00 -1.28276512e-01
-3.76879722e-01 1.36735380e-01 -1.45611155e+00 2.90733486e-01
4.93652552e-01 -2.97994763e-01 -6.05526745e-01 -9.13849294e-01
1.00227639e-01 1.35417208e-01 2.05821931e-01 8.08702335e-02
1.22647531e-01 -3.33355993e-01 9.05521289e-02 1.15690939e-01
3.61738533e-01 -7.80174553e-01 1.89254493e-01 -1.32244244e-01
-5.03678441e-01 -3.76027018e-01 -2.06714258e-01 9.89166260e-01
1.59498155e-01 1.84982970e-01 5.97179413e-01 -1.32192627e-01
-5.43939114e-01 8.40847492e-01 -2.63967570e-02 -2.35810369e-01
1.04073536e+00 1.32976383e-01 -4.12453264e-01 -4.80211914e-01
-1.14444697e+00 5.53458929e-01 2.04651147e-01 5.94534338e-01
2.34373435e-01 -1.54017818e+00 -5.07066309e-01 5.20911217e-01
-2.76806764e-03 -3.00290197e-01 7.24622130e-01 6.98727608e-01
-2.81129424e-02 4.50771660e-01 -4.91939811e-03 -1.02294706e-01
-9.14695859e-01 5.70062935e-01 5.43176711e-01 -6.12526596e-01
-4.74561214e-01 1.23893011e+00 4.07095253e-01 -6.94603264e-01
3.15235585e-01 9.23567191e-02 -1.85865045e-01 1.30121395e-01
3.90707910e-01 5.92436492e-01 -1.28776163e-01 -5.28983653e-01
-1.15663670e-01 -5.88850491e-03 -5.45054853e-01 1.03311680e-01
1.64309776e+00 -3.24892789e-01 -3.67308781e-02 1.81711018e-01
1.47658169e+00 -1.39617860e-01 -1.71458352e+00 -7.32361019e-01
-4.04104710e-01 -5.39601147e-01 2.18133420e-01 -8.44548166e-01
-1.49094725e+00 8.88262510e-01 3.59344572e-01 -5.12963161e-02
7.19242990e-01 4.70532835e-01 8.56613398e-01 2.12550849e-01
2.24925652e-01 -5.15968740e-01 -3.19200642e-02 1.91240802e-01
7.25094318e-01 -1.06356001e+00 -2.85865307e-01 2.59779636e-02
-9.49309170e-01 6.69858217e-01 7.17148423e-01 6.90339953e-02
8.22535932e-01 -6.59350827e-02 2.55094916e-02 -1.56611040e-01
-1.32470632e+00 -3.58881545e-03 3.54331225e-01 5.68455875e-01
2.12152466e-01 2.97285885e-01 2.53224950e-02 1.47708702e+00
4.27082777e-02 5.83111010e-02 -7.03623369e-02 3.96718264e-01
1.45198181e-02 -1.41469121e+00 2.85064608e-01 4.13373113e-01
-3.02334726e-01 3.23618650e-02 -1.28217638e-01 3.03552032e-01
1.59495488e-01 6.67809486e-01 -1.79796815e-01 -6.82764173e-01
2.17875242e-01 4.08050358e-01 4.02352542e-01 -6.47899568e-01
-7.05842435e-01 -2.37469301e-01 -8.82641524e-02 -9.28729773e-01
4.53710109e-02 -3.78076971e-01 -6.97311759e-01 -6.14981413e-01
-4.45098311e-01 2.03233212e-01 2.39095047e-01 9.89813030e-01
1.04909968e+00 6.48740888e-01 6.90197289e-01 -9.48202789e-01
-1.09518898e+00 -6.86614752e-01 -7.80593693e-01 3.49322796e-01
4.84238267e-01 -6.52229726e-01 -2.85307914e-01 -1.22664645e-01]
|
[5.825687408447266, 6.287322998046875]
|
783fb678-6f93-4421-a35d-35b09987a196
|
information-prebuilt-recurrent-reconstruction
|
2112.05755
| null |
https://arxiv.org/abs/2112.05755v4
|
https://arxiv.org/pdf/2112.05755v4.pdf
|
Information Prebuilt Recurrent Reconstruction Network for Video Super-Resolution
|
The video super-resolution (VSR) method based on the recurrent convolutional network has strong temporal modeling capability for video sequences. However, the temporal receptive field of different recurrent units in the unidirectional recurrent network is unbalanced. Earlier reconstruction frames receive less spatio-temporal information, resulting in fuzziness or artifacts. Although the bidirectional recurrent network can alleviate this problem, it requires more memory space and fails to perform many tasks with low latency requirements. To solve the above problems, we propose an end-to-end information prebuilt recurrent reconstruction network (IPRRN), consisting of an information prebuilt network (IPNet) and a recurrent reconstruction network (RRNet). By integrating sufficient information from the front of the video to build the hidden state needed for the initially recurrent unit to help restore the earlier frames, the information prebuilt network balances the input information difference at different time steps. In addition, we demonstrate an efficient recurrent reconstruction network, which outperforms the existing unidirectional recurrent schemes in all aspects. Many experiments have verified the effectiveness of the network we propose, which can effectively achieve better quantitative and qualitative evaluation performance compared to the existing state-of-the-art methods.
|
['Ming Yu', 'Shuyun Wang', 'Gang Yan', 'Yingchun Guo', 'Cuihong Xue']
|
2021-12-10
| null | null | null | null |
['video-super-resolution']
|
['computer-vision']
|
[ 1.85214639e-01 -2.14088380e-01 -1.99256435e-01 -4.28871587e-02
-3.41613084e-01 1.88497994e-02 2.55935550e-01 -7.30577648e-01
-1.56757861e-01 6.54502094e-01 6.01830423e-01 5.88507392e-03
1.26145735e-01 -7.02684164e-01 -5.37352681e-01 -8.99414062e-01
2.36314952e-01 -3.50557536e-01 7.52111256e-01 -2.66509235e-01
2.82932132e-01 2.29739755e-01 -1.55408573e+00 7.96393931e-01
5.23137093e-01 8.71119499e-01 7.49476552e-01 3.97128463e-01
-1.15934305e-01 1.67713737e+00 -3.54035079e-01 3.20704162e-01
-7.87593201e-02 -6.85599267e-01 -6.03837073e-01 -1.60885513e-01
-1.04162246e-01 -6.26481891e-01 -9.43135083e-01 9.64029670e-01
5.22617877e-01 3.47400665e-01 -3.37312482e-02 -5.21163046e-01
-7.68947423e-01 8.00581872e-01 -5.71689188e-01 6.51558161e-01
4.33112204e-01 -8.60101432e-02 5.32273233e-01 -8.62280250e-01
8.96262228e-01 1.15314007e+00 6.85926318e-01 8.03606331e-01
-8.08483720e-01 -8.99317622e-01 4.94480252e-01 4.06820148e-01
-1.30053973e+00 -5.30464590e-01 7.67426074e-01 -1.24040358e-01
9.03412640e-01 1.68938592e-01 6.51188970e-01 1.15510750e+00
2.73879707e-01 8.02614689e-01 7.85770714e-01 1.35056093e-01
-2.28627965e-01 -2.61138856e-01 7.88000748e-02 5.39707780e-01
-2.62847543e-01 2.93380260e-01 -5.81930995e-01 5.04814625e-01
1.53417444e+00 5.35884738e-01 -5.79865038e-01 2.40570560e-01
-1.37277377e+00 3.01771939e-01 6.58898830e-01 7.06321597e-01
-5.08170784e-01 1.68060631e-01 5.32246530e-01 4.05980647e-01
4.15193081e-01 -2.66240746e-01 -4.52477336e-02 -1.07456319e-01
-1.00782061e+00 -2.41096109e-01 8.46051574e-02 7.87789345e-01
6.55909002e-01 5.08718669e-01 -3.61395031e-01 7.71189868e-01
2.23578617e-01 -6.09805770e-02 8.48627508e-01 -1.14635313e+00
4.73443180e-01 5.10399938e-01 -2.60712928e-03 -1.12524319e+00
-3.75336379e-01 -4.88596201e-01 -1.37386847e+00 -2.45289169e-02
-2.74163723e-01 5.81043176e-02 -9.64317560e-01 1.64906371e+00
-4.07185107e-02 6.91502333e-01 2.42445439e-01 1.23876202e+00
1.40575397e+00 1.16623592e+00 -2.28771701e-01 -7.17068434e-01
9.98085082e-01 -1.21349132e+00 -1.18121159e+00 2.43449528e-02
1.99691474e-01 -6.54588461e-01 6.28842115e-01 2.08635882e-01
-1.31720304e+00 -9.66816127e-01 -1.09963226e+00 -2.66548216e-01
1.63968220e-01 1.97523117e-01 3.06304157e-01 -2.92321831e-01
-1.25038946e+00 6.63571358e-01 -7.84969211e-01 -5.36004864e-02
5.07233106e-02 1.01944938e-01 -4.73145276e-01 -1.95522271e-02
-1.36021566e+00 6.02891207e-01 3.03715169e-01 6.72162294e-01
-1.06051588e+00 -5.19928098e-01 -8.47602248e-01 -3.57413776e-02
2.51697391e-01 -4.39428121e-01 1.05612612e+00 -9.86605048e-01
-1.53957558e+00 3.08467180e-01 -4.48763430e-01 -3.92560035e-01
4.30919766e-01 -3.58725078e-02 -5.98034441e-01 1.68146297e-01
-6.42343005e-03 3.87006313e-01 6.92256629e-01 -1.06268573e+00
-6.35266721e-01 -5.49423024e-02 1.96815608e-03 4.17033941e-01
9.51964036e-02 8.60621631e-02 -7.56905258e-01 -9.55063462e-01
5.88165402e-01 -5.42923510e-01 -2.52957255e-01 -3.65809679e-01
-2.73823161e-02 3.25363502e-02 1.12673330e+00 -9.31141019e-01
1.77520001e+00 -2.42651892e+00 3.75119239e-01 -2.88874567e-01
4.35904920e-01 3.97652000e-01 -3.31437439e-02 1.62962407e-01
-4.32200998e-01 1.16534404e-01 7.23642260e-02 -2.21670449e-01
-8.21758568e-01 2.15112910e-01 -5.40799797e-01 2.71982223e-01
-1.48734137e-01 7.75216579e-01 -9.19046581e-01 -5.22796929e-01
3.45709443e-01 8.02003503e-01 -2.76372284e-01 3.33592832e-01
3.20506506e-02 6.13906741e-01 -4.17377263e-01 3.38068724e-01
5.49095035e-01 -4.70453352e-01 5.57997748e-02 -3.62395078e-01
-5.39771855e-01 5.11128485e-01 -1.16877377e+00 1.80216658e+00
-3.47879261e-01 7.68980205e-01 -2.15217650e-01 -7.36179829e-01
9.82403874e-01 6.97435677e-01 5.18964469e-01 -1.18703818e+00
2.57421304e-02 -2.47435048e-02 -3.12969625e-01 -5.77965200e-01
6.80464625e-01 -6.84405863e-02 3.43560189e-01 4.27874655e-01
-1.06907383e-01 8.95015180e-01 -1.99900679e-02 2.24230602e-01
1.14735079e+00 4.01985884e-01 -1.01372264e-01 2.12548956e-01
7.03090847e-01 -5.09380400e-01 1.18786883e+00 3.72777939e-01
-2.00385422e-01 9.75791812e-01 2.43678778e-01 -7.55822301e-01
-1.10458922e+00 -8.65855217e-01 2.81702340e-01 7.30980337e-01
5.65302789e-01 -4.44652319e-01 -3.83553088e-01 -3.54450166e-01
-7.79635906e-01 2.33519375e-01 -6.18177176e-01 -2.53059417e-01
-9.47772264e-01 -3.39080095e-01 4.01936948e-01 7.23229408e-01
1.12842524e+00 -1.39897394e+00 -5.95752120e-01 4.31799591e-01
-7.78145373e-01 -1.04819071e+00 -5.63847959e-01 -3.07581425e-01
-1.04980803e+00 -7.95183897e-01 -7.41118312e-01 -9.22062457e-01
5.22625506e-01 7.17830479e-01 9.00703847e-01 2.45145246e-01
1.64709195e-01 -2.96502262e-01 -4.92095172e-01 5.46861053e-01
-1.73951864e-01 -1.75291970e-01 -1.68789342e-01 -1.25722971e-03
5.78668900e-02 -6.76748514e-01 -7.50530064e-01 5.68532407e-01
-1.05115986e+00 6.75276935e-01 5.18519938e-01 8.81876588e-01
8.05655241e-01 3.03692520e-01 5.20976305e-01 -7.05215156e-01
2.30203167e-01 -3.20564866e-01 -4.46106732e-01 2.47217014e-01
-3.17697823e-01 4.39584069e-02 6.98159456e-01 -4.85037684e-01
-1.41628253e+00 -2.01532260e-01 -9.70681235e-02 -9.24614668e-01
4.58574206e-01 5.28485477e-01 1.29733577e-01 1.60571024e-01
2.17157289e-01 6.79340422e-01 -6.18745498e-02 -5.19232810e-01
-9.62322056e-02 4.23573345e-01 6.80004060e-01 1.36863410e-01
3.23265016e-01 4.40510303e-01 -4.17105615e-01 -5.50238907e-01
-7.09347069e-01 -3.33730102e-01 -2.91268975e-01 -5.20874143e-01
9.53822017e-01 -1.35148311e+00 -7.45986342e-01 5.18029690e-01
-1.31061041e+00 -1.61320567e-01 -2.18980491e-01 5.26470006e-01
-4.74919021e-01 3.51257443e-01 -1.15303457e+00 -7.16766059e-01
-2.91083395e-01 -1.20478964e+00 6.96619749e-01 3.78692031e-01
3.27982217e-01 -7.60469854e-01 -1.04599064e-02 1.50339350e-01
5.04312813e-01 1.52313471e-01 4.24917966e-01 2.21004292e-01
-1.03566229e+00 2.65732497e-01 -4.17846560e-01 2.03471735e-01
1.41400129e-01 1.06799334e-01 -7.83959329e-01 -1.47443905e-01
2.81023741e-01 7.83007368e-02 1.27400768e+00 4.48289126e-01
1.05869639e+00 -3.99649292e-01 -1.10664405e-01 6.88856602e-01
1.50494802e+00 4.28831130e-01 1.30851257e+00 2.84296542e-01
8.86460602e-01 2.52938688e-01 6.00370586e-01 2.53417820e-01
4.43942755e-01 5.44240892e-01 3.01294118e-01 -1.45211861e-01
-4.26765501e-01 -4.30825502e-01 7.63803840e-01 1.30062819e+00
-6.00608706e-01 1.16186962e-02 -3.25203717e-01 4.46050614e-01
-2.32845807e+00 -1.54829574e+00 8.01321343e-02 2.05466485e+00
6.84046865e-01 9.40878615e-02 -7.93095753e-02 1.03346422e-01
9.27406073e-01 5.77488005e-01 -5.30023754e-01 -2.44560391e-01
-2.66339451e-01 -4.13910300e-01 3.00821066e-01 3.53708118e-01
-6.84987187e-01 9.66006219e-01 6.67195559e+00 7.99695134e-01
-1.21559763e+00 8.52940306e-02 5.52392244e-01 -8.91891420e-02
-4.02294189e-01 -6.43557310e-02 -7.31118262e-01 5.23965955e-01
8.65275919e-01 -1.95609089e-02 6.96915209e-01 4.58255172e-01
6.34775519e-01 9.79942307e-02 -6.90664828e-01 1.25047922e+00
-2.66693681e-02 -1.76512194e+00 1.93005294e-01 -2.86793739e-01
6.59171462e-01 6.71345145e-02 5.20089120e-02 2.92329401e-01
9.19535831e-02 -9.97689843e-01 5.94579518e-01 9.10774529e-01
9.03951466e-01 -7.49518812e-01 8.78455222e-01 3.28844935e-01
-1.87154579e+00 -1.53664693e-01 -4.91606385e-01 -1.48274943e-01
2.28322342e-01 5.49441040e-01 -8.05111751e-02 8.18836033e-01
9.69329178e-01 1.42015588e+00 -2.79481471e-01 5.95836163e-01
-1.99856728e-01 3.42179537e-01 1.90674104e-02 4.65874493e-01
2.49715611e-01 -2.56472319e-01 5.38066685e-01 9.25414264e-01
2.64484942e-01 3.90988022e-01 -5.68458773e-02 7.43984520e-01
-9.73310173e-02 -3.14527243e-01 -6.01383746e-01 -4.50739749e-02
3.33705097e-01 1.05694604e+00 -5.62731922e-01 -4.58424509e-01
-4.29740518e-01 1.04925597e+00 2.39187390e-01 6.31358385e-01
-9.41398799e-01 -2.04339728e-01 4.15374488e-01 1.04718126e-01
2.82887012e-01 -1.06510766e-01 -3.67056802e-02 -1.54637194e+00
2.22242683e-01 -7.27787673e-01 4.76182312e-01 -1.16883075e+00
-8.47582817e-01 9.43395853e-01 -3.67851585e-01 -1.72065878e+00
-4.29325610e-01 9.98439044e-02 -4.93681967e-01 8.70611668e-01
-1.61182427e+00 -8.82736862e-01 -4.68994588e-01 8.45291913e-01
9.07892764e-01 -1.31204814e-01 4.43428248e-01 7.04541028e-01
-8.20029318e-01 3.44095975e-01 -1.90878630e-01 2.44548395e-01
4.56008852e-01 -5.65040708e-01 1.61070809e-01 1.22799194e+00
-3.12859654e-01 6.89060450e-01 5.14444709e-01 -7.04030931e-01
-1.27725089e+00 -1.13280034e+00 7.94697702e-01 7.01898187e-02
4.33283657e-01 9.18040276e-02 -1.25333071e+00 9.09635484e-01
4.08421867e-02 1.43572465e-01 1.57671168e-01 -2.27175385e-01
-4.10969675e-01 -1.81063637e-01 -8.78567696e-01 6.54425383e-01
1.23211181e+00 -7.80189455e-01 -5.34523249e-01 -1.76057234e-01
1.25537026e+00 -6.07787371e-01 -7.44882882e-01 6.39999449e-01
6.29921734e-01 -1.28554988e+00 8.74168217e-01 -1.56512022e-01
8.36771190e-01 -7.89744377e-01 -5.28843030e-02 -7.66193330e-01
-7.78848708e-01 -5.87208092e-01 -2.48010978e-01 1.18195283e+00
1.42740265e-01 -4.76136863e-01 6.00915670e-01 2.17612401e-01
-2.20871419e-01 -7.67444789e-01 -9.02589202e-01 -4.98483688e-01
-5.56016326e-01 -2.35549912e-01 3.96352798e-01 7.90137291e-01
-1.60334289e-01 5.09655833e-01 -8.12600374e-01 1.48173094e-01
3.28700215e-01 1.80012271e-01 2.08921790e-01 -6.82622850e-01
-3.20057850e-03 -3.09939742e-01 -3.87067109e-01 -1.39523816e+00
-9.95821580e-02 -4.60342169e-01 7.29584917e-02 -1.79748380e+00
3.82001340e-01 -3.14327866e-01 -5.80388606e-01 3.80879641e-01
-4.79351245e-02 2.88446069e-01 2.11850569e-01 5.70722044e-01
-8.02362263e-01 6.27541900e-01 1.59790432e+00 8.33358616e-02
-3.36285532e-01 -1.45276457e-01 -4.71512616e-01 6.11469746e-01
4.56789762e-01 -1.11348912e-01 -5.31770051e-01 -6.35144114e-01
2.28151679e-01 6.92166746e-01 3.12620521e-01 -9.92414832e-01
5.58835149e-01 4.63449769e-02 5.20898283e-01 -1.10607672e+00
3.25130761e-01 -7.84787476e-01 7.26051092e-01 3.57602388e-01
-4.32675004e-01 3.39652121e-01 -3.02980971e-02 5.44413507e-01
-5.60845315e-01 1.93787977e-01 8.21619868e-01 -3.19483250e-01
-9.57597017e-01 5.40733814e-01 -3.95541877e-01 -3.73275876e-01
7.32254624e-01 -4.50294167e-01 -4.41165090e-01 -3.41827005e-01
-6.84923589e-01 2.16320038e-01 3.84345651e-01 5.18069506e-01
1.12898374e+00 -1.53718960e+00 -6.62183106e-01 2.68908679e-01
-3.00442636e-01 1.75771847e-01 8.91337395e-01 7.86013126e-01
-5.92149079e-01 2.66084909e-01 -3.45205426e-01 -5.89233816e-01
-1.23093748e+00 7.94465780e-01 3.56688231e-01 -5.21250725e-01
-1.22071171e+00 5.57900429e-01 4.34489757e-01 6.74039721e-02
2.87141174e-01 -2.24451602e-01 -7.60078251e-01 3.70156649e-03
9.95346546e-01 4.50784653e-01 -2.08788663e-01 -6.84406161e-01
-7.63721541e-02 6.63879156e-01 -2.45868713e-01 -1.36797562e-01
1.40916145e+00 -5.17872155e-01 -2.78213769e-01 7.65649557e-01
1.09303308e+00 -3.87128681e-01 -1.35699904e+00 -4.33123648e-01
-5.39995015e-01 -4.99419570e-01 2.00581431e-01 -3.44596058e-01
-1.52390587e+00 7.22048461e-01 6.46482170e-01 6.74374998e-02
1.54922926e+00 -4.22595114e-01 1.07776797e+00 1.15697943e-01
4.07069504e-01 -9.79016900e-01 1.00644499e-01 6.30736768e-01
8.33105803e-01 -9.87487137e-01 -3.47669013e-02 -2.57108420e-01
-6.89302087e-01 1.19941998e+00 5.96750021e-01 -3.07673901e-01
5.96188068e-01 2.42588207e-01 -7.94752128e-03 -3.21191475e-02
-1.22434652e+00 -9.09085013e-03 2.05412768e-02 2.04852432e-01
4.02365863e-01 -3.61804903e-01 -3.11689824e-01 4.34058487e-01
2.41505206e-01 4.12478328e-01 5.63197613e-01 7.30399430e-01
-3.83380353e-01 -6.24172449e-01 -2.35819388e-02 1.44011930e-01
-4.90228981e-01 -9.10722092e-02 1.72294870e-01 3.94703954e-01
-4.65678424e-02 7.66314685e-01 1.47435218e-01 -8.20637703e-01
1.73845291e-01 -5.27724087e-01 1.93178281e-01 -2.45307729e-01
-5.65915525e-01 4.50869799e-01 -1.51636437e-01 -9.62451339e-01
-8.21124017e-01 -4.04462576e-01 -1.48334050e+00 -4.33307797e-01
-1.23110771e-01 1.82229318e-02 1.33215517e-01 8.74109030e-01
4.41507369e-01 1.17753494e+00 8.69660258e-01 -8.20485234e-01
1.96912184e-01 -8.42650056e-01 -4.42064822e-01 8.89092758e-02
6.24428332e-01 -4.70728457e-01 -2.89119512e-01 2.37547830e-01]
|
[11.046378135681152, -1.8380577564239502]
|
97cb6c22-0e56-4206-9d38-bd09fe4aff13
|
deepilluminance-contextual-illuminance
|
1905.04791
| null |
https://arxiv.org/abs/1905.04791v2
|
https://arxiv.org/pdf/1905.04791v2.pdf
|
DeepIlluminance: Contextual Illuminance Estimation via Deep Neural Networks
|
Computational color constancy refers to the estimation of the scene illumination and makes the perceived color relatively stable under varying illumination. In the past few years, deep Convolutional Neural Networks (CNNs) have delivered superior performance in illuminant estimation. Several representative methods formulate it as a multi-label prediction problem by learning the local appearance of image patches using CNNs. However, these approaches inevitably make incorrect estimations for the ambiguous patches affected by their neighborhood contexts. Inaccurate local estimates are likely to bring in degraded performance when combining into a global prediction. To address the above issues, we propose a contextual deep network for patch-based illuminant estimation equipped with refinement. First, the contextual net with a center-surround architecture extracts local contextual features from image patches, and generates initial illuminant estimates and the corresponding color corrected patches. The patches are sampled based on the observation that pixels with large color differences describe the illumination well. Then, the refinement net integrates the input patches with the corrected patches in conjunction with the use of intermediate features to improve the performance. To train such a network with numerous parameters, we propose a stage-wise training strategy, in which the features and the predicted illuminant from previous stages are provided to the next learning stage with more finer estimates recovered. Experiments show that our approach obtains competitive performance on two illuminant estimation benchmarks.
|
['Jun Zhang', 'Meng Wang', 'Tong Zheng', 'Shengping Zhang']
|
2019-05-12
| null | null | null | null |
['color-constancy']
|
['computer-vision']
|
[ 4.42577481e-01 -6.68902874e-01 -2.14611813e-02 -4.69377041e-01
-4.66489226e-01 -3.57048750e-01 2.76966363e-01 -1.86703220e-01
-1.61797866e-01 7.49888957e-01 2.21338272e-02 1.26108944e-01
2.46660396e-01 -6.76952362e-01 -8.41190755e-01 -1.14731431e+00
4.61393893e-01 -3.43806893e-01 1.23559423e-01 4.48473580e-02
4.30803120e-01 3.47295403e-01 -1.78961289e+00 4.01388079e-01
1.05461884e+00 1.57449853e+00 1.71688333e-01 5.49982011e-01
-2.54806161e-01 7.85726666e-01 -5.56792974e-01 -9.57801566e-02
4.73980904e-01 -4.76480216e-01 -3.16641569e-01 2.16979444e-01
7.78540432e-01 -5.28268397e-01 2.23094542e-02 1.11768150e+00
3.80622685e-01 1.02616400e-01 5.17703235e-01 -1.16305566e+00
-7.32292414e-01 -4.19360809e-02 -7.67062068e-01 -1.09638214e-01
1.39670923e-01 1.90623760e-01 9.53703821e-01 -9.90957141e-01
1.34348750e-01 9.18487549e-01 6.70457721e-01 1.47686914e-01
-1.12484324e+00 -5.98570228e-01 4.32898939e-01 2.22152278e-01
-1.37472367e+00 -3.99386704e-01 9.34148669e-01 -1.33517429e-01
5.63266456e-01 2.64053643e-01 6.81970358e-01 6.89692080e-01
2.42600843e-01 7.62036264e-01 1.32422876e+00 -4.05491620e-01
1.49345189e-01 2.22299621e-01 -2.62849063e-01 7.30446637e-01
1.15082942e-01 1.45513713e-01 -4.22910005e-01 1.48893431e-01
9.05004740e-01 1.91777244e-01 -5.79167008e-01 -2.65537977e-01
-1.03644776e+00 4.92529869e-01 1.12261510e+00 1.09313078e-01
-5.12212098e-01 2.51145571e-01 5.46536557e-02 1.47840623e-02
7.40894377e-01 2.79806405e-01 -6.15594387e-01 4.53482002e-01
-6.12417400e-01 -1.87596604e-02 2.93446541e-01 7.20845938e-01
1.47064483e+00 -8.61034319e-02 -2.68061757e-01 1.13136840e+00
7.50900060e-02 5.22521675e-01 2.82111038e-02 -9.64204311e-01
2.82817841e-01 5.42435527e-01 3.52077872e-01 -1.15013814e+00
-4.10558522e-01 -7.05621183e-01 -1.10884106e+00 3.57973546e-01
5.62747478e-01 -1.50891870e-01 -8.32910180e-01 1.65727067e+00
4.68436003e-01 3.28694135e-01 -2.60927826e-01 1.25383210e+00
4.98938054e-01 8.12874734e-01 -1.99858081e-02 -2.20898762e-01
1.06861269e+00 -1.19101012e+00 -5.01812100e-01 -2.82937378e-01
7.22476318e-02 -1.01182413e+00 9.29600179e-01 5.35270095e-01
-8.63757312e-01 -1.04249799e+00 -1.05882132e+00 -4.70497608e-02
-3.72224420e-01 3.63243967e-01 6.28400803e-01 4.58064705e-01
-1.16647422e+00 5.32534301e-01 -2.20781952e-01 3.17415409e-02
4.38938439e-01 -5.02623729e-02 5.36392443e-02 -4.24042434e-01
-9.70979750e-01 6.13831699e-01 3.40640657e-02 5.46342731e-01
-7.62065947e-01 -6.71829879e-01 -7.61697352e-01 2.85453498e-02
2.26556540e-01 -6.25241518e-01 8.23190928e-01 -1.85917878e+00
-1.56805921e+00 3.83516163e-01 -3.85120153e-01 1.76887214e-01
3.40794533e-01 -5.68832606e-02 -3.51888746e-01 2.96669845e-02
9.09916535e-02 7.24348664e-01 1.23697984e+00 -1.60587549e+00
-8.92287016e-01 -1.36795789e-01 9.79594812e-02 3.90911877e-01
-1.79425567e-01 -2.41134763e-01 -6.94100320e-01 -3.13061923e-01
3.15196425e-01 -6.19847059e-01 -1.94824189e-01 2.75529176e-01
-4.16703552e-01 -2.34669708e-02 3.57941300e-01 -3.88061821e-01
8.25865686e-01 -2.13331914e+00 -1.41133785e-01 1.68918088e-01
5.58095686e-02 2.17282213e-02 -4.61545974e-01 -6.03355542e-02
-1.89470142e-01 -3.41445357e-01 -1.41198039e-01 -2.91489154e-01
-1.94464535e-01 -4.85315770e-02 -1.50503501e-01 6.10171318e-01
4.42975014e-01 6.79729581e-01 -9.91943002e-01 -2.83154875e-01
4.26837981e-01 6.50798917e-01 -4.20666456e-01 4.44688678e-01
-3.52001041e-01 4.72171307e-01 -1.64226577e-01 7.52940774e-01
1.09345329e+00 -2.07772315e-01 -1.65910825e-01 -6.97287261e-01
-4.87909049e-01 -6.45563304e-02 -1.13823569e+00 1.56285119e+00
-6.40046656e-01 6.93929315e-01 -2.09859777e-02 -7.16280639e-01
9.68210936e-01 4.81396727e-02 4.48599398e-01 -9.52552199e-01
2.15830877e-01 2.36051112e-01 -3.74496520e-01 -4.38690454e-01
4.50000525e-01 -9.10832360e-03 3.65879923e-01 2.29464427e-01
-2.83677310e-01 -1.11825272e-01 -2.06893861e-01 -2.65392035e-01
5.81363738e-01 4.80136245e-01 1.93095431e-01 8.33335742e-02
6.05424762e-01 -2.47028217e-01 8.33167315e-01 5.32858431e-01
-3.05524647e-01 8.73176038e-01 1.89263836e-01 -8.67747843e-01
-1.01908863e+00 -9.64153171e-01 -2.40063682e-01 1.21478641e+00
5.79473317e-01 1.22245941e-02 -6.19163394e-01 -5.92798412e-01
-1.55507505e-01 3.68655711e-01 -6.90566182e-01 -7.16382116e-02
-3.18778098e-01 -7.76005924e-01 -1.52374148e-01 5.24297178e-01
9.24389362e-01 -1.02705145e+00 -4.26566094e-01 1.38400629e-01
-3.13029051e-01 -8.64560008e-01 -4.36308891e-01 2.95171410e-01
-4.41611439e-01 -1.04673624e+00 -8.37101638e-01 -8.52103889e-01
1.03581226e+00 6.93403959e-01 1.10971832e+00 2.16664076e-01
-4.31024700e-01 5.14001250e-02 -2.51258463e-01 -2.31618285e-01
1.02519885e-01 -2.52380937e-01 -3.67045730e-01 5.13244510e-01
1.68059468e-01 -4.11096066e-01 -1.22180724e+00 4.67798680e-01
-8.05410445e-01 4.81309414e-01 8.06092322e-01 1.06067383e+00
7.26327658e-01 1.77521482e-01 2.99682051e-01 -8.13039899e-01
1.58252761e-01 -2.14980468e-01 -6.26118422e-01 2.83146352e-01
-4.63844776e-01 -1.70860633e-01 8.71281803e-01 -2.51886100e-01
-1.47519684e+00 2.43151337e-01 4.68657725e-02 -2.88382351e-01
-3.77586544e-01 2.66072005e-01 -2.04545394e-01 -4.07799482e-01
7.04198062e-01 3.50868523e-01 -3.88809443e-01 -2.36175939e-01
3.35239589e-01 5.03791153e-01 4.90459412e-01 -3.80082399e-01
6.92302465e-01 5.01048267e-01 -3.58779952e-02 -3.71906966e-01
-1.24115407e+00 -4.65103209e-01 -6.01770878e-01 -5.25069475e-01
6.68397427e-01 -1.04915082e+00 -6.71127379e-01 7.10958600e-01
-1.15707684e+00 -5.59555352e-01 7.69408867e-02 6.50664940e-02
-2.34206066e-01 1.52365431e-01 -4.40286487e-01 -7.50390947e-01
-1.10134862e-01 -1.05252230e+00 1.06997228e+00 5.95607698e-01
3.43310893e-01 -8.31402779e-01 6.40890971e-02 2.37447340e-02
5.83020687e-01 1.92830443e-01 8.98437679e-01 3.66567194e-01
-8.32315981e-01 -1.32317677e-01 -9.67150748e-01 6.02132559e-01
4.80988294e-01 6.67891428e-02 -1.55160213e+00 -1.66410431e-01
-1.48135051e-01 -2.53316224e-01 1.01955938e+00 6.88075483e-01
1.59432673e+00 -8.72233137e-02 -6.99756145e-02 1.00642800e+00
1.99814725e+00 2.48340368e-02 6.33005023e-01 3.27600628e-01
9.34262276e-01 4.72996205e-01 6.42472327e-01 5.01612425e-01
3.83905649e-01 3.77082735e-01 7.59127438e-01 -5.33031523e-01
-4.27481145e-01 -4.99402210e-02 6.59721643e-02 5.43900609e-01
-5.85322343e-02 -1.74494311e-01 -3.76311272e-01 4.66837168e-01
-1.66705787e+00 -6.80393159e-01 1.22366361e-01 2.17603278e+00
9.18349087e-01 -4.80688997e-02 -1.84502944e-01 -1.59872070e-01
8.23183596e-01 3.42511863e-01 -8.43373775e-01 -1.65388629e-01
-3.84927243e-01 1.08798426e-02 5.16262233e-01 4.35887456e-01
-1.10818386e+00 7.43618548e-01 5.97816372e+00 6.45844638e-01
-1.37427509e+00 -1.25678435e-01 1.16829443e+00 1.23752780e-01
-5.08238852e-01 -4.10345867e-02 -4.69638795e-01 4.15413558e-01
3.48633200e-01 5.20249367e-01 8.35233033e-01 5.52783430e-01
1.43172681e-01 -5.22337854e-01 -8.59764218e-01 1.08907509e+00
2.09931999e-01 -1.05786431e+00 -1.02237903e-01 -3.37266386e-01
1.30202758e+00 7.52609000e-02 3.30557525e-01 6.14401558e-03
2.99734771e-02 -8.14920545e-01 7.12240756e-01 8.25715959e-01
8.70391130e-01 -7.69922316e-01 6.90607905e-01 7.24573657e-02
-1.28751266e+00 -2.07594410e-01 -9.04485047e-01 -9.58798900e-02
-2.48302266e-01 8.00646603e-01 -5.36950827e-01 5.72759449e-01
6.45576358e-01 9.91179705e-01 -5.80682933e-01 1.29234946e+00
-3.75523120e-01 2.79378861e-01 -9.23574436e-03 -8.41526315e-02
1.37241006e-01 -5.17741561e-01 2.42204405e-02 1.01563013e+00
2.37629190e-01 -1.77657217e-01 3.97871323e-02 1.05973172e+00
-9.30896029e-02 1.58759609e-01 -2.07978040e-01 5.94034672e-01
2.54161835e-01 1.62534630e+00 -6.40757501e-01 -2.55530000e-01
-6.69741094e-01 1.24007988e+00 5.70620120e-01 8.40507448e-01
-9.05300438e-01 -4.64149624e-01 6.46291614e-01 -2.41461322e-01
1.84869856e-01 8.93658996e-02 -5.15367389e-01 -1.04760695e+00
-4.02465314e-02 -5.59262693e-01 4.78940569e-02 -1.26745641e+00
-1.42715693e+00 6.99051797e-01 -6.53676152e-01 -1.43849385e+00
3.24354768e-01 -7.38016963e-01 -6.78693533e-01 1.03928423e+00
-2.17588925e+00 -1.09608626e+00 -8.65910292e-01 6.04198217e-01
4.54370797e-01 3.09302956e-01 6.86402380e-01 2.97836751e-01
-7.95898974e-01 4.20920551e-01 1.84824839e-01 -2.10086107e-02
1.01942551e+00 -1.33512819e+00 3.28970812e-02 8.18706036e-01
-1.01120181e-01 3.51587534e-01 4.58257765e-01 -2.19706148e-01
-1.08569539e+00 -1.44700038e+00 4.73165184e-01 1.16976267e-02
2.60014385e-01 -1.85179323e-01 -7.47647882e-01 2.18977496e-01
3.05965096e-01 4.58029300e-01 3.47094446e-01 1.17881171e-01
-6.17604196e-01 -7.76355267e-01 -8.42755139e-01 4.95357752e-01
7.72265911e-01 -6.58305109e-01 1.23025395e-01 3.16012979e-01
5.35516381e-01 -3.49431306e-01 -4.78702903e-01 2.31720895e-01
6.07766867e-01 -1.40006483e+00 8.93021822e-01 2.32250318e-02
5.37171841e-01 -5.61703563e-01 -5.50150871e-02 -1.58943760e+00
-5.83484709e-01 -2.80144244e-01 5.29031932e-01 1.12988102e+00
4.08642411e-01 -6.41878426e-01 5.31248748e-01 5.48919857e-01
-3.76381338e-01 -7.26034045e-01 -5.25947630e-01 -2.44667172e-01
-2.56533146e-01 -3.02402318e-01 7.55800068e-01 7.03006864e-01
-4.52387869e-01 9.77325533e-03 -4.68827009e-01 3.54458839e-01
6.67777181e-01 6.98093116e-01 6.62285030e-01 -1.03934455e+00
-7.73524269e-02 -4.83993083e-01 6.92197084e-02 -1.20171607e+00
1.04642704e-01 -4.50105190e-01 6.49278760e-01 -1.46668863e+00
4.84934330e-01 -5.94466269e-01 -7.48571873e-01 4.88413811e-01
-6.76804423e-01 7.29257584e-01 -1.04318067e-01 1.13866940e-01
-8.90016258e-01 5.74458182e-01 1.60920370e+00 -3.08943897e-01
-1.87749758e-01 1.17685115e-02 -6.88968301e-01 5.50698996e-01
8.26800764e-01 -1.08741056e-02 -2.43554547e-01 -6.54892385e-01
4.71858084e-01 -3.94995391e-01 5.28601050e-01 -1.08434224e+00
2.34152004e-01 -3.57149452e-01 1.11510766e+00 -5.10351598e-01
2.89483994e-01 -9.21523273e-01 -3.48385647e-02 1.02544762e-01
-3.25928450e-01 -4.25614089e-01 5.33561297e-02 4.94218141e-01
-2.52712965e-01 9.35636759e-02 9.56519127e-01 -4.24174704e-02
-8.16099226e-01 5.74178636e-01 -9.85299516e-03 -3.71877462e-01
6.34794891e-01 -2.77719229e-01 -4.68029648e-01 -3.11230630e-01
-1.19280055e-01 -1.05352521e-01 5.23330510e-01 1.18879624e-01
6.26258850e-01 -1.37405586e+00 -4.56659287e-01 4.58652616e-01
2.91495591e-01 1.06027514e-01 5.00494599e-01 8.25691581e-01
-5.52888453e-01 1.79761425e-01 -3.56762201e-01 -6.55657947e-01
-8.32015455e-01 5.51274359e-01 6.14569187e-01 3.93031836e-01
-1.43453300e-01 1.04632199e+00 6.21330440e-01 -1.11977905e-01
2.02102870e-01 -5.48652828e-01 -8.60172063e-02 -1.98294386e-01
5.99472523e-01 1.23580903e-01 2.12038923e-02 -3.92530829e-01
-2.56855041e-02 8.34505260e-01 2.00991422e-01 3.77907932e-01
1.32929230e+00 -5.79975188e-01 -3.91447902e-01 4.77160543e-01
1.40973890e+00 2.65608411e-02 -1.97161520e+00 -5.50551176e-01
-6.44510150e-01 -8.57608855e-01 4.30353642e-01 -1.04460895e+00
-1.47404897e+00 7.94077158e-01 6.89344525e-01 1.63207948e-01
1.76062179e+00 -4.06602234e-01 6.38348877e-01 6.01035403e-03
2.48871222e-01 -1.12346423e+00 1.91859439e-01 3.66435200e-01
6.50188506e-01 -1.46203327e+00 -6.21293262e-02 -3.87877673e-01
-4.58561659e-01 1.34138834e+00 8.75170231e-01 -7.07622319e-02
4.42136258e-01 -1.19212258e-03 3.91669065e-01 5.12500182e-02
-5.06039679e-01 -3.22934151e-01 4.80130553e-01 5.36762714e-01
3.79837662e-01 7.11876573e-03 3.14147085e-01 2.60880858e-01
2.99575388e-01 -3.31291795e-01 7.67103210e-02 2.71104306e-01
-6.79371893e-01 -6.86411858e-01 -4.71962839e-01 2.17071071e-01
-2.89187953e-02 -2.94790477e-01 -1.68193519e-01 3.62202376e-01
6.24018371e-01 9.02360678e-01 2.88028538e-01 -3.08881342e-01
1.31796733e-01 -2.98355281e-01 4.15535361e-01 -3.14005405e-01
-3.44997585e-01 4.18005258e-01 -2.78752983e-01 -8.64687920e-01
-5.45724571e-01 -4.93707031e-01 -8.30930710e-01 -1.17608137e-01
-4.44833249e-01 -1.38613924e-01 7.23832369e-01 7.82946646e-01
9.80243161e-02 7.00922668e-01 1.31511271e+00 -1.13532841e+00
1.06150704e-02 -8.56777191e-01 -7.19137311e-01 4.59520787e-01
6.78772569e-01 -4.31933641e-01 -5.14038682e-01 -3.81626636e-02]
|
[10.522788047790527, -2.5752880573272705]
|
1b9ff439-0ef3-4495-93f6-11414a5ce2c9
|
a-neuro-symbolic-approach-for-enhanced-human
|
2304.1174
| null |
https://arxiv.org/abs/2304.11740v1
|
https://arxiv.org/pdf/2304.11740v1.pdf
|
A Neuro-Symbolic Approach for Enhanced Human Motion Prediction
|
Reasoning on the context of human beings is crucial for many real-world applications especially for those deploying autonomous systems (e.g. robots). In this paper, we present a new approach for context reasoning to further advance the field of human motion prediction. We therefore propose a neuro-symbolic approach for human motion prediction (NeuroSyM), which weights differently the interactions in the neighbourhood by leveraging an intuitive technique for spatial representation called Qualitative Trajectory Calculus (QTC). The proposed approach is experimentally tested on medium and long term time horizons using two architectures from the state of art, one of which is a baseline for human motion prediction and the other is a baseline for generic multivariate time-series prediction. Six datasets of challenging crowded scenarios, collected from both fixed and mobile cameras, were used for testing. Experimental results show that the NeuroSyM approach outperforms in most cases the baseline architectures in terms of prediction accuracy.
|
['Nicola Bellotto', 'Marc Hanheide', 'Luca Castri', 'Sariah Mghames']
|
2023-04-23
| null | null | null | null |
['motion-prediction', 'time-series-prediction', 'human-motion-prediction']
|
['computer-vision', 'time-series', 'time-series']
|
[ 1.93709671e-01 1.81111798e-01 9.05358791e-02 -9.00405571e-02
1.62072092e-01 -2.01573759e-01 1.24473071e+00 1.39789507e-01
-5.70751727e-01 6.48924291e-01 5.55348516e-01 -3.04860890e-01
-2.93783873e-01 -6.65536046e-01 -3.95829558e-01 -5.20439267e-01
-4.94572580e-01 3.32506180e-01 8.23019803e-01 -5.81104338e-01
4.53802228e-01 5.97660124e-01 -1.93803382e+00 5.26799560e-01
2.68870950e-01 8.46911430e-01 4.00514126e-01 9.39666390e-01
3.46419752e-01 1.45894670e+00 -2.80905247e-01 -2.18772456e-01
2.46517897e-01 -3.44595075e-01 -8.74318719e-01 -3.36546749e-01
-5.34460843e-02 9.76360291e-02 -3.75824511e-01 5.63748598e-01
3.48438442e-01 8.11629832e-01 6.12243712e-01 -1.40750861e+00
-8.41049179e-02 3.03347170e-01 2.94281058e-02 2.96116859e-01
7.93676198e-01 4.12841052e-01 5.30101836e-01 -7.59265959e-01
9.81359720e-01 1.43716085e+00 9.30736780e-01 5.94348907e-01
-7.18223274e-01 1.42900825e-01 1.11169890e-01 1.05150545e+00
-1.17761433e+00 -2.92725652e-01 5.98791838e-01 -7.30911076e-01
1.48827648e+00 2.71959960e-01 9.59392071e-01 1.20290613e+00
4.00402308e-01 7.59847343e-01 8.87223542e-01 -3.65471333e-01
7.18549848e-01 -2.23706350e-01 -9.76893678e-02 3.98480982e-01
-3.67219038e-02 3.54803890e-01 -5.46950579e-01 -1.48124516e-01
5.95338523e-01 -1.95400462e-01 -4.33542393e-03 -4.05408531e-01
-1.76526129e+00 5.27953923e-01 3.30739200e-01 4.96852309e-01
-7.19688892e-01 4.19866860e-01 5.00255525e-01 -3.39182913e-02
2.24840119e-02 9.74566862e-02 -1.91119611e-01 -5.32897472e-01
-8.25846314e-01 8.07575762e-01 8.95068705e-01 8.60043883e-01
1.99081481e-01 -1.69341594e-01 -4.83341604e-01 2.23680913e-01
3.38466674e-01 2.12669998e-01 8.32953036e-01 -1.08604765e+00
3.66293579e-01 6.01553261e-01 5.53336203e-01 -1.25301588e+00
-9.32833970e-01 3.09425443e-01 -5.67599118e-01 3.69069844e-01
4.12571430e-01 -1.20300964e-01 -5.25651395e-01 1.35107243e+00
4.25692350e-01 7.34464824e-01 3.00429136e-01 1.02550721e+00
5.26812553e-01 7.67961323e-01 3.72622430e-01 -2.62141638e-02
1.18467772e+00 -1.15114307e+00 -5.14773965e-01 -1.60991825e-04
7.05065668e-01 -3.60995531e-01 8.51886511e-01 3.71330559e-01
-1.01430297e+00 -6.63700998e-01 -9.89825130e-01 4.63174656e-02
-5.91575742e-01 8.79993141e-02 3.63362700e-01 2.89372385e-01
-1.26875925e+00 9.36832428e-01 -1.09185195e+00 -9.09064293e-01
5.97141981e-02 2.21127957e-01 -2.94670045e-01 4.13496435e-01
-1.14958656e+00 1.23126793e+00 6.73131526e-01 1.26366362e-01
-7.22275674e-01 -6.06124774e-02 -7.31835783e-01 -1.71432778e-01
1.51025698e-01 -8.20591629e-01 1.25304067e+00 -7.24597156e-01
-1.69471300e+00 6.76174402e-01 2.72704102e-02 -1.03037369e+00
9.16626394e-01 -2.86185384e-01 -3.63608450e-01 2.06773415e-01
-1.33908761e-04 9.08894300e-01 6.06815577e-01 -9.02371228e-01
-1.00832081e+00 -2.78052166e-02 1.12652391e-01 2.26503447e-01
4.71220687e-02 -2.56709382e-02 -1.26640603e-01 -7.14940906e-01
-3.20156842e-01 -1.33833241e+00 -5.17744720e-01 -1.04150638e-01
-1.07497327e-01 -3.43143374e-01 9.18923855e-01 -7.39560604e-01
1.19855487e+00 -1.65047276e+00 5.70659995e-01 1.46804258e-01
-3.51033241e-01 4.56797183e-01 1.36871830e-01 7.64594972e-01
-2.30678301e-02 -3.00334871e-01 -3.90842497e-01 -3.80891800e-01
8.53126422e-02 3.50996315e-01 -4.58610117e-01 2.46909350e-01
1.09037101e-01 8.35933924e-01 -1.07533216e+00 -5.08670151e-01
5.24765253e-01 4.98745054e-01 -3.73460561e-01 8.62753466e-02
-5.49014509e-01 7.54972219e-01 -2.25301877e-01 2.44719565e-01
2.31717415e-02 1.85958639e-01 2.08150879e-01 3.37665796e-01
-3.03260177e-01 -2.12880313e-01 -1.17549622e+00 1.83396459e+00
-2.07539409e-01 8.71595502e-01 -7.28337467e-01 -6.47290945e-01
6.56692922e-01 4.51507032e-01 4.41450208e-01 -4.40025628e-01
3.02843638e-02 5.91803901e-03 -1.55987322e-01 -9.38963294e-01
8.90083253e-01 1.02012508e-01 -3.36470976e-02 8.17048475e-02
-2.18299344e-01 1.86756417e-01 1.48051694e-01 -1.46278247e-01
1.22300005e+00 9.21884596e-01 7.27826416e-01 -2.57028311e-01
1.12129128e+00 4.40046102e-01 4.06424940e-01 4.42047745e-01
-5.87973297e-01 3.76876622e-01 4.91784185e-01 -8.18658769e-01
-1.06362927e+00 -6.13507152e-01 4.88621384e-01 1.07541716e+00
1.72802389e-01 -3.70868236e-01 -1.04738402e+00 -3.53023440e-01
-3.36874217e-01 9.23476577e-01 -7.48424768e-01 3.99760827e-02
-1.04498088e+00 -2.34144121e-01 6.77573025e-01 8.69501889e-01
3.67809147e-01 -1.56157315e+00 -1.65187490e+00 2.11121798e-01
-2.67582566e-01 -1.36514747e+00 2.02253327e-01 -3.97204429e-01
-8.71580243e-01 -9.39627171e-01 -5.79503179e-01 -4.34688509e-01
-2.44292747e-02 2.00119331e-01 7.82611430e-01 1.03655562e-01
-3.41421142e-02 7.93965459e-01 -6.30234241e-01 -5.90635955e-01
-3.86458844e-01 -1.98025137e-01 2.01863199e-01 2.54602847e-03
3.32087576e-01 -5.16947091e-01 -6.67553484e-01 3.99316579e-01
-8.29415619e-01 1.64622396e-01 3.35310876e-01 6.14226699e-01
3.73068988e-01 7.13931620e-02 4.42677528e-01 -2.49854758e-01
5.52576244e-01 -8.09857011e-01 -5.32827497e-01 3.21123809e-01
-2.00950041e-01 5.04573733e-02 5.91117918e-01 -4.64825511e-01
-1.22401822e+00 2.27744088e-01 1.95028663e-01 -2.90531099e-01
-5.39193451e-01 5.67333341e-01 6.00031242e-02 8.50358009e-02
7.83736289e-01 -9.59605575e-02 -3.70866358e-01 -1.54355660e-01
4.20176655e-01 1.98342159e-01 8.87097716e-01 -4.43283319e-01
4.47732538e-01 7.81324327e-01 6.78661883e-01 -1.02529836e+00
1.63695320e-01 -6.05383575e-01 -1.07626224e+00 -7.16030478e-01
1.28523540e+00 -7.11789250e-01 -7.76937366e-01 4.05946225e-01
-1.38073611e+00 -6.83444858e-01 -2.69019097e-01 5.61824977e-01
-1.13591278e+00 4.01147634e-01 -2.31608093e-01 -1.12250924e+00
-1.23535218e-02 -9.08890784e-01 1.06910419e+00 1.19875148e-01
-5.91893554e-01 -1.36439145e+00 3.26740742e-01 5.55197224e-02
2.64312208e-01 8.79813850e-01 6.30885959e-01 -6.46916628e-01
-4.67708617e-01 -1.83641180e-01 2.53888816e-01 -2.46947095e-01
-5.85349917e-01 -1.37946025e-01 -9.47011828e-01 2.12336630e-01
3.71431857e-02 1.21513583e-01 6.02254152e-01 1.22565210e-01
6.56048417e-01 -1.95759341e-01 -5.61078370e-01 5.33060543e-02
1.23907113e+00 3.31452191e-01 1.09059048e+00 8.01425338e-01
6.76348448e-01 1.12702620e+00 1.16542542e+00 7.61723340e-01
7.03104079e-01 1.00597286e+00 4.80783850e-01 7.12556899e-01
-2.26495489e-02 -2.12316126e-01 5.87881505e-01 4.82058376e-01
-7.98452675e-01 -2.70444423e-01 -1.44726217e+00 7.82838047e-01
-2.60449481e+00 -1.47447610e+00 -3.64255786e-01 2.08855009e+00
5.67924455e-02 -6.35955632e-02 5.92909932e-01 2.65467495e-01
6.20949805e-01 -8.59314203e-02 -2.72870839e-01 -4.63348955e-01
8.45865235e-02 -1.27091967e-02 2.88843066e-01 4.29187626e-01
-1.24303818e+00 1.08781040e+00 5.72379446e+00 4.75977391e-01
-9.76035774e-01 1.50705412e-01 1.41273990e-01 -4.26787045e-03
3.64597768e-01 1.27781883e-01 -5.42355001e-01 3.01455110e-01
1.31454110e+00 -2.57473141e-02 3.83439422e-01 8.77495825e-01
4.78528231e-01 -3.78045917e-01 -1.30247378e+00 9.98773098e-01
2.95747444e-02 -1.24058998e+00 -1.24707147e-01 -1.74175784e-01
6.04322553e-01 -6.76225796e-02 -1.92965567e-01 1.99875355e-01
-9.94273722e-02 -1.05277836e+00 1.36219716e+00 1.32897103e+00
2.42961161e-02 -6.14955008e-01 8.02994490e-01 7.93567240e-01
-1.38961351e+00 -3.99704397e-01 -2.05441773e-01 -4.24357474e-01
5.29690564e-01 -2.19167024e-01 -9.84795868e-01 6.68815255e-01
6.06535375e-01 7.81930268e-01 -8.03832173e-01 1.03757393e+00
-1.95999760e-02 2.80503154e-01 -1.51652873e-01 -2.74834692e-01
4.60171580e-01 1.43285207e-02 9.91434634e-01 1.54182923e+00
5.85042536e-01 1.77380353e-01 -1.19559266e-01 4.51297790e-01
7.68872082e-01 -1.06237672e-01 -1.03228819e+00 3.06271046e-01
2.44169578e-01 8.99725616e-01 -1.02285206e+00 -4.01125342e-01
-2.41654739e-01 9.05657291e-01 4.49691825e-02 2.39799395e-01
-9.75377619e-01 -4.59619574e-02 3.25792253e-01 -3.44648608e-03
4.62454051e-01 -5.83983719e-01 -2.19722092e-01 -8.27825129e-01
1.39937431e-01 -5.63928246e-01 2.99514413e-01 -1.02291560e+00
-8.27928543e-01 7.17275321e-01 6.74562693e-01 -1.63756025e+00
-9.42551017e-01 -8.94709051e-01 -7.70704448e-01 4.89360094e-01
-1.13181019e+00 -1.38171160e+00 -5.03795207e-01 7.15763867e-01
3.91866565e-01 -3.95246446e-01 7.49417186e-01 -1.49150223e-01
-2.08910614e-01 -9.95894596e-02 -3.52735490e-01 -3.48755598e-01
2.27839991e-01 -1.06594217e+00 7.60259330e-01 8.92602682e-01
-5.99487051e-02 4.36662406e-01 1.01691282e+00 -8.63875031e-01
-1.15983772e+00 -1.19556522e+00 1.01606309e+00 -8.75244558e-01
6.86658204e-01 8.14879760e-02 -6.84208632e-01 7.40622044e-01
2.34292477e-01 -1.03916228e-01 5.57848096e-01 -6.39737248e-01
1.38625562e-01 3.97917420e-01 -1.16791165e+00 1.09127581e+00
1.30551457e+00 -2.53707618e-01 -8.44192564e-01 2.41952717e-01
6.04264975e-01 -2.73051322e-01 -8.31814885e-01 4.44454134e-01
7.40188181e-01 -1.38700759e+00 1.15035892e+00 -6.05087817e-01
2.92895019e-01 -5.62805355e-01 -5.63636661e-01 -9.26668644e-01
-2.36662135e-01 -7.92041659e-01 -5.10062814e-01 6.70790255e-01
-4.09524664e-02 -4.31667298e-01 7.30164707e-01 6.68546319e-01
-1.30997539e-01 -6.68891490e-01 -1.07562661e+00 -9.60058928e-01
-1.11974008e-01 -7.99486160e-01 7.49290824e-01 6.51972651e-01
2.02533007e-01 6.57235384e-02 -4.32175189e-01 3.67667973e-02
2.48947799e-01 -4.98493165e-01 8.67549598e-01 -1.09434211e+00
-2.25722402e-01 -5.78432798e-01 -1.31713879e+00 -5.52323937e-01
2.36980692e-01 -5.98037660e-01 1.99822467e-02 -1.56812847e+00
-3.19110036e-01 1.17744114e-02 8.42874646e-02 8.96309838e-02
1.43449441e-01 -1.24935687e-01 5.73767006e-01 2.97346115e-01
-7.43081272e-01 4.92386371e-01 5.97736597e-01 2.66826659e-01
-2.85997421e-01 -1.54759008e-02 2.10579038e-01 1.22991157e+00
7.95815766e-01 -1.30068898e-01 -5.90759277e-01 -2.53318042e-01
3.65481317e-01 3.92258346e-01 1.07743204e+00 -1.87214625e+00
7.03262746e-01 -3.82735372e-01 1.08712874e-01 -7.55791724e-01
6.10397577e-01 -9.49657679e-01 4.20619160e-01 6.07821882e-01
-3.39743376e-01 4.48991120e-01 3.04640949e-01 8.88546109e-01
-4.52032350e-02 -2.17214346e-01 3.47885549e-01 -2.43981183e-02
-1.48371994e+00 -1.63158193e-01 -5.90385795e-01 -3.60069275e-01
1.49365497e+00 -4.33269322e-01 -2.92009503e-01 -4.24102783e-01
-9.47307587e-01 4.18933108e-02 4.81193513e-01 5.59113383e-01
8.13364208e-01 -1.54465282e+00 -3.34112555e-01 -3.81786704e-01
2.53747553e-01 -3.61617714e-01 1.72874525e-01 1.06004536e+00
-9.23426628e-01 7.33193636e-01 -5.27024150e-01 -6.79327548e-01
-1.04548323e+00 8.05107176e-01 2.25315198e-01 -1.78880006e-01
-7.51997828e-01 3.90264481e-01 -2.28581324e-01 -2.95322329e-01
2.24566355e-01 -7.19030023e-01 -8.00224841e-01 -2.06893444e-01
5.93241692e-01 1.08374941e+00 -1.88833520e-01 -1.36217129e+00
-3.80919248e-01 5.88954806e-01 8.14302504e-01 -6.44947588e-01
1.17480552e+00 -1.43609166e-01 -8.37932825e-02 9.17238951e-01
6.24299228e-01 -5.55929959e-01 -1.35541856e+00 1.06851980e-01
7.72702277e-01 -2.01246455e-01 -5.63058913e-01 -5.46106398e-01
-2.01417834e-01 9.20783877e-01 7.84502983e-01 1.71764731e-01
1.01724279e+00 -4.45748478e-01 5.64668477e-01 9.62017000e-01
9.40291226e-01 -1.18062663e+00 6.40028110e-03 6.48179412e-01
1.13900352e+00 -9.73046303e-01 -1.11540429e-01 -5.30352220e-02
-1.14127409e+00 1.26306701e+00 3.81177217e-01 -2.34750167e-01
6.23357713e-01 -2.76293576e-01 -2.80302912e-01 -1.32816613e-01
-1.03224325e+00 -3.24481696e-01 3.41148049e-01 9.72329795e-01
1.61217287e-01 2.27225885e-01 -1.76061168e-01 5.70062041e-01
-1.64913565e-01 3.41269523e-01 4.95512307e-01 1.16689277e+00
-4.15131480e-01 -8.29275370e-01 -7.55934656e-01 -2.02309236e-01
-9.88226384e-02 3.24202836e-01 -4.81009483e-01 9.52689171e-01
4.12981838e-01 1.09716785e+00 -7.58618340e-02 -7.34871268e-01
4.77214187e-01 1.36679053e-01 4.48922634e-01 -2.36584604e-01
-6.39978647e-01 -4.45505351e-01 2.94423997e-01 -9.31656957e-01
-9.11613286e-01 -1.09565270e+00 -1.47143292e+00 -2.02017441e-01
3.07253808e-01 -3.41363698e-01 6.34290636e-01 1.05214512e+00
2.80897111e-01 3.48238796e-01 7.09417984e-02 -1.49644887e+00
-1.91688716e-01 -8.84135067e-01 -8.43651313e-03 3.97328973e-01
8.15044120e-02 -9.73180175e-01 1.85773283e-01 4.56076920e-01]
|
[7.089297771453857, 0.20028288662433624]
|
b5fb3d9a-4e52-4dbc-8723-07ae1af5ab3c
|
data-augmentation-imbalance-for-imbalanced
|
2004.13628
| null |
https://arxiv.org/abs/2004.13628v3
|
https://arxiv.org/pdf/2004.13628v3.pdf
|
Data Augmentation Imbalance For Imbalanced Attribute Classification
|
Pedestrian attribute recognition is an important multi-label classification problem. Although the convolutional neural networks are prominent in learning discriminative features from images, the data imbalance in multi-label setting for fine-grained tasks remains an open problem. In this paper, we propose a new re-sampling algorithm called: data augmentation imbalance (DAI) to explicitly enhance the ability to discriminate the fewer attributes via increasing the proportion of labels accounting for a small part. Fundamentally, by applying over-sampling and under-sampling on the multi-label dataset at the same time, the thought of robbing the rich attributes and helping the poor makes a significant contribution to DAI. Extensive empirical evidence shows that our DAI algorithm achieves state-of-the-art results, based on pedestrian attribute datasets, i.e. standard PA-100K and PETA datasets.
|
['ShengMei Shen', 'Fanhua Shang', 'Pan Zhou', 'Yang Hu', 'Xiaying Bai']
|
2020-04-19
| null | null | null | null |
['pedestrian-attribute-recognition']
|
['computer-vision']
|
[ 1.27569884e-01 -8.01555291e-02 -5.19153774e-01 -8.59461308e-01
-7.02855170e-01 -2.13793725e-01 2.61233509e-01 2.64380604e-01
-4.77137744e-01 1.08441854e+00 1.40922800e-01 8.55697244e-02
7.41812065e-02 -8.69148612e-01 -7.47420907e-01 -9.77020919e-01
3.89772594e-01 7.47058988e-01 -1.40276879e-01 1.62379846e-01
-9.72401947e-02 3.40130448e-01 -1.77556789e+00 4.29559171e-01
9.34453487e-01 1.23289311e+00 -5.33357322e-01 3.00239295e-01
-7.98811913e-02 8.08931053e-01 -3.18735331e-01 -7.52153516e-01
3.94747019e-01 -2.09867835e-01 -7.32745409e-01 1.64355040e-01
9.88936007e-01 -5.11656642e-01 -2.47839242e-01 1.05533743e+00
5.98859310e-01 -1.57013893e-01 9.06290710e-01 -1.88993049e+00
-6.25680804e-01 3.94248813e-01 -1.03591120e+00 1.03250980e-01
-1.72018215e-01 3.31501178e-02 1.15401936e+00 -7.47700393e-01
9.21688005e-02 1.50006008e+00 7.04517841e-01 6.38911724e-01
-1.35870314e+00 -1.05191326e+00 2.70683765e-01 4.41569835e-01
-1.44096935e+00 -3.07181895e-01 7.17589080e-01 -4.52282131e-01
2.76291609e-01 4.15041506e-01 3.70991349e-01 1.13805449e+00
-1.08093895e-01 9.87287819e-01 1.56002295e+00 -3.22554260e-01
-4.08974960e-02 8.82232562e-02 4.91975963e-01 4.92833644e-01
5.22611141e-01 -2.81083863e-02 -4.75805730e-01 -1.69015139e-01
3.28482896e-01 1.10670820e-01 3.66405696e-01 -3.19999635e-01
-9.43881273e-01 8.42654347e-01 3.66071165e-01 -4.78371471e-01
-3.33400875e-01 1.66211277e-01 7.80612051e-01 2.40726382e-01
7.05258489e-01 1.66505173e-01 -6.67793989e-01 1.59355879e-01
-4.32841629e-01 4.93178129e-01 3.55125964e-01 8.98633063e-01
1.08352637e+00 -2.47665569e-01 -5.45617342e-01 9.91032541e-01
8.01885501e-02 3.85382473e-01 1.87697247e-01 -7.38406301e-01
4.61499363e-01 9.26321685e-01 -6.77388832e-02 -5.93853712e-01
-4.71269518e-01 -3.91009837e-01 -1.26550162e+00 5.03983021e-01
7.22710848e-01 -1.76246032e-01 -9.19297159e-01 1.90091670e+00
4.49097216e-01 -2.12280788e-02 -3.05182815e-01 8.26160252e-01
6.18209779e-01 2.42034644e-01 8.35677743e-01 2.45379806e-02
1.64931452e+00 -1.08948457e+00 -5.95505059e-01 -4.41377103e-01
5.38252234e-01 -5.87136745e-01 1.17704308e+00 2.31680647e-01
-4.54027891e-01 -7.74720967e-01 -1.01794457e+00 -3.21063958e-02
-3.93051505e-01 2.26689428e-01 7.76687264e-01 7.19104230e-01
-5.12875855e-01 2.67579198e-01 -1.21941984e-01 -1.04872115e-01
1.23165441e+00 2.75663435e-01 -4.67450291e-01 -4.28593129e-01
-1.10760176e+00 6.91561460e-01 1.64647445e-01 -3.91913235e-01
-8.26755166e-01 -8.39718044e-01 -4.95079637e-01 -4.29768041e-02
2.53462553e-01 -6.56427205e-01 9.42351043e-01 -1.14370179e+00
-8.76133680e-01 1.24431825e+00 -5.13303801e-02 -3.21644843e-01
8.52266431e-01 -4.95042622e-01 -3.20694119e-01 -2.30576158e-01
4.81188834e-01 1.00204337e+00 6.42181695e-01 -1.35683274e+00
-1.00987339e+00 -6.49771154e-01 2.37680256e-01 1.66981146e-01
-3.83937240e-01 -1.31675750e-01 2.17189565e-01 -7.60296345e-01
-1.74480677e-01 -9.47047710e-01 -1.81550667e-01 1.08544677e-01
-6.64212525e-01 -4.68452275e-01 7.49464035e-01 -3.81217092e-01
9.94712293e-01 -2.06090903e+00 -2.77232099e-02 -6.99448436e-02
5.51702976e-01 7.00860694e-02 -1.85365871e-01 -1.86212137e-01
-3.03711802e-01 2.50868559e-01 7.29090273e-02 -3.74462962e-01
1.04677655e-01 2.96309918e-01 -1.17635161e-01 4.90375012e-01
5.18895805e-01 9.01682734e-01 -5.89156210e-01 -6.95205152e-01
1.45172447e-01 9.55340043e-02 -5.27306259e-01 1.67934075e-01
7.28241052e-04 5.46881258e-01 -5.08167982e-01 1.15149975e+00
1.00675905e+00 -1.81567967e-01 -2.49880493e-01 -3.73301655e-01
4.23392206e-02 -2.62297869e-01 -1.12895691e+00 1.22091317e+00
-1.27689138e-01 3.48953635e-01 -3.50629002e-01 -9.67202783e-01
7.73131073e-01 1.05770072e-03 4.62564468e-01 -9.92667437e-01
1.94563121e-01 1.23885095e-01 -7.74993226e-02 -3.78176749e-01
3.73933256e-01 -2.89082974e-01 -3.18360537e-01 3.86171788e-01
-9.57321450e-02 6.66976094e-01 8.15620199e-02 -4.25335893e-04
4.76298571e-01 -3.00712101e-02 4.48782235e-01 -3.31401676e-01
3.37621391e-01 -1.45293072e-01 8.78190637e-01 7.98829556e-01
-7.75695801e-01 4.35356945e-01 5.89594185e-01 -7.70920098e-01
-1.52235615e+00 -7.22795188e-01 -3.50162685e-01 1.74617982e+00
1.69918939e-01 9.96098146e-02 -8.81168544e-01 -1.00216150e+00
3.31205428e-01 4.34084475e-01 -1.01024663e+00 -3.14103454e-01
-3.98697913e-01 -1.27265143e+00 7.99279511e-01 7.04238713e-01
8.50166619e-01 -8.04215491e-01 -3.37117612e-01 2.42795032e-02
-3.71971667e-01 -1.11731184e+00 -2.80687869e-01 4.07245815e-01
-3.77559215e-01 -1.10832822e+00 -6.57360435e-01 -5.84628344e-01
6.46795154e-01 2.79216141e-01 1.26712155e+00 -6.95391232e-03
-2.12223262e-01 -2.69595414e-01 -3.69976223e-01 -6.05820775e-01
-2.29534879e-01 1.94623291e-01 1.64476797e-01 3.79322022e-01
7.91843295e-01 -5.62527359e-01 -5.43594658e-01 5.78339815e-01
-5.49018562e-01 2.33318150e-01 6.83836460e-01 9.66441751e-01
7.60228693e-01 -2.17229337e-01 9.50804770e-01 -1.03060126e+00
2.12766945e-01 -4.54060882e-01 -1.14234172e-01 2.08253950e-01
-6.93570197e-01 1.02287471e-01 5.79777062e-01 -6.38918579e-01
-7.81140447e-01 9.66412202e-02 -1.86796442e-01 -2.11180761e-01
-7.04908073e-01 -2.13669479e-01 -8.59439313e-01 -1.72741815e-01
5.54645658e-01 -1.41462103e-01 -1.07825823e-01 -4.47578877e-01
4.28492993e-01 7.20927179e-01 2.61940092e-01 -8.69801223e-01
3.58739227e-01 5.42903602e-01 2.41375148e-01 -1.71741441e-01
-1.41960490e+00 -4.06442225e-01 -8.50234628e-01 -3.86973530e-01
9.15859699e-01 -1.03408277e+00 -8.01593244e-01 9.59506989e-01
-8.58322322e-01 -1.83795959e-01 -3.92251790e-01 1.76455290e-03
-4.01017129e-01 8.43243822e-02 -4.48885292e-01 -5.04507720e-01
-2.37773150e-01 -1.31781590e+00 1.10519648e+00 4.50102746e-01
-6.07289411e-02 -3.55852962e-01 -1.30749047e-01 6.39500499e-01
2.95159608e-01 4.16739970e-01 1.13788748e+00 -9.62917149e-01
-3.66095871e-01 -2.00711161e-01 -8.97740304e-01 3.66251051e-01
-1.97968306e-03 -2.87967503e-01 -1.44717383e+00 -2.03623086e-01
-6.24370754e-01 -8.81546736e-01 1.01678526e+00 3.59630018e-01
1.48726797e+00 -2.57319272e-01 -2.96899855e-01 5.29518306e-01
1.34925091e+00 -1.21543609e-01 5.71015716e-01 6.06864870e-01
1.06187057e+00 6.71642482e-01 9.51570630e-01 6.53377175e-01
6.48304105e-01 7.22081959e-01 6.23284936e-01 -3.35582048e-01
-3.49649578e-01 4.35014302e-03 -3.73905897e-01 2.24728793e-01
-5.68422377e-02 -8.63367245e-02 -9.18226242e-01 3.50639701e-01
-1.93394959e+00 -6.88354254e-01 -2.82682180e-01 2.02239633e+00
1.02306223e+00 9.08457339e-02 4.42972004e-01 2.12763742e-01
9.58198726e-01 3.08218636e-02 -8.08598816e-01 -5.26117720e-02
-4.47758406e-01 -1.21647246e-01 7.31634796e-01 1.52925998e-01
-1.73207021e+00 7.43071437e-01 5.98443985e+00 1.17959082e+00
-6.58780158e-01 2.27127597e-01 1.48446822e+00 7.32468218e-02
1.46035999e-01 -2.85872221e-01 -1.18901038e+00 6.71123683e-01
7.47660339e-01 1.06534585e-01 7.87488073e-02 9.51230109e-01
-2.62791455e-01 -5.12881614e-02 -1.15374601e+00 9.24674511e-01
-4.69641462e-02 -9.68632281e-01 3.82889926e-01 2.65828192e-01
6.97985411e-01 -2.74972320e-01 1.09258763e-01 4.71537828e-01
4.87040430e-01 -1.15102756e+00 9.89549935e-01 4.75233614e-01
1.20073533e+00 -1.07370627e+00 9.22273993e-01 1.53892308e-01
-1.11247265e+00 -2.15980083e-01 -5.26997685e-01 -1.51583403e-01
-6.15529120e-02 8.78593624e-01 -1.13637447e-01 3.95339161e-01
7.52430975e-01 4.04827654e-01 -9.47125733e-01 9.76007581e-01
5.02036065e-02 4.83069003e-01 6.03355728e-02 2.70242453e-01
1.05626948e-01 4.25781794e-02 1.36329487e-01 1.18747652e+00
-2.42961705e-01 1.84151698e-02 5.14406741e-01 4.16543335e-01
-3.78978193e-01 -7.43972063e-02 -3.81433308e-01 3.02300811e-01
5.03925800e-01 1.34229016e+00 -4.84980643e-01 -4.69431162e-01
-4.67579663e-01 8.33034635e-01 5.86375713e-01 2.58132130e-01
-1.00410926e+00 -1.41284347e-01 1.17386317e+00 -1.16410889e-01
-5.74841499e-02 3.06980044e-01 -8.77100289e-01 -7.86177874e-01
-7.48259351e-02 -1.07490289e+00 6.20642722e-01 -3.49598736e-01
-1.83143437e+00 2.25334078e-01 -2.13061512e-01 -1.16047120e+00
3.71058255e-01 -6.08001411e-01 -5.82256429e-02 9.50281799e-01
-1.86429191e+00 -1.60680282e+00 -8.21763098e-01 3.52809638e-01
5.66575706e-01 -2.44185790e-01 8.42972517e-01 8.95116150e-01
-6.79560304e-01 1.05477428e+00 -2.65439712e-02 2.93399900e-01
1.08361495e+00 -1.23803258e+00 2.80914068e-01 4.89868879e-01
-3.94359142e-01 -3.17216627e-02 5.90917408e-01 -4.62483555e-01
-6.39266253e-01 -1.42017186e+00 6.29349947e-01 -5.14925659e-01
3.49385679e-01 -2.93510854e-01 -8.86609912e-01 6.14737809e-01
-6.54543787e-02 3.26171279e-01 1.04186523e+00 1.33525178e-01
-9.09480870e-01 -3.24824661e-01 -1.37467718e+00 2.80097365e-01
1.08434880e+00 -3.84004653e-01 -8.85703564e-02 2.58471608e-01
7.57326543e-01 -2.02478141e-01 -7.89432704e-01 5.71281910e-01
7.71619141e-01 -8.29712629e-01 1.30276000e+00 -1.06566727e+00
8.07117283e-01 -8.83115306e-02 -5.22796273e-01 -1.31610870e+00
-7.26842761e-01 3.14168990e-01 -4.08588583e-03 1.57010031e+00
4.61252294e-02 -5.03827989e-01 9.22437310e-01 5.79563200e-01
2.37782598e-02 -8.39628339e-01 -1.06830168e+00 -5.55373847e-01
3.72398525e-01 -1.33361984e-02 8.19920242e-01 1.22496235e+00
-7.04898357e-01 3.45983505e-01 -7.13312328e-01 3.79623933e-04
1.10681093e+00 -3.80891711e-02 7.71223605e-01 -1.65519547e+00
2.69488662e-01 -4.10997480e-01 -4.74849790e-01 -6.73148215e-01
3.58424425e-01 -6.89274013e-01 -1.74637020e-01 -1.10398078e+00
8.34288120e-01 -9.73786831e-01 -5.34564793e-01 6.91009223e-01
-5.60052097e-01 5.92364848e-01 5.24802841e-02 2.46553645e-01
-8.06100726e-01 5.30835509e-01 1.21661389e+00 -4.59281653e-01
4.63428676e-01 -2.18119211e-02 -1.20765936e+00 7.44263113e-01
8.33916426e-01 -5.84251106e-01 7.34646693e-02 -2.37519711e-01
3.42437206e-03 -5.63339949e-01 6.07985914e-01 -1.05504692e+00
-2.79836934e-02 -2.75925666e-01 7.70668507e-01 -5.03245413e-01
1.40129834e-01 -8.14735770e-01 -1.26482442e-01 2.58976310e-01
-5.39720356e-01 9.95555595e-02 2.30218887e-01 6.60523117e-01
-2.79487371e-02 5.85627146e-02 1.16389251e+00 -1.85136069e-02
-9.99801099e-01 6.10046208e-01 -3.24450731e-02 2.27462590e-01
1.14432085e+00 -1.72860608e-01 -7.23875046e-01 1.55380994e-01
-4.67159688e-01 3.24542135e-01 3.57095867e-01 3.30244392e-01
1.81216970e-01 -1.81805062e+00 -9.90827262e-01 2.16803908e-01
4.77037162e-01 -1.15755558e-01 5.33508301e-01 5.84607780e-01
-3.27337720e-02 1.89364165e-01 -7.75612056e-01 -5.69736004e-01
-1.46178687e+00 4.09906209e-01 2.64381349e-01 -4.30872679e-01
-2.91001469e-01 1.02594197e+00 5.04453540e-01 -4.72380698e-01
2.91455388e-01 1.96397051e-01 -3.75098914e-01 5.49626291e-01
6.41337693e-01 7.38920510e-01 5.66712394e-02 -7.72473097e-01
-4.59040880e-01 4.42912251e-01 -3.65619779e-01 4.77684200e-01
1.12906134e+00 -2.45036989e-01 -7.55416006e-02 3.85298550e-01
1.05555284e+00 -3.07326496e-01 -1.46963573e+00 -3.02354753e-01
-1.26331568e-01 -7.02723086e-01 -2.08070830e-01 -1.00982881e+00
-9.57922161e-01 8.62545013e-01 9.74267244e-01 -1.96555741e-02
1.01716518e+00 1.02032451e-02 6.34485722e-01 9.84594822e-02
3.94678295e-01 -1.18372381e+00 3.19550365e-01 3.81357610e-01
4.82750952e-01 -1.69509447e+00 2.99846977e-02 -4.39534545e-01
-7.77651668e-01 7.07967758e-01 1.13130951e+00 1.45674184e-01
3.10583085e-01 2.70543963e-01 6.73128292e-02 3.41756158e-02
-3.33489597e-01 -1.94016322e-01 1.51703637e-02 6.29830599e-01
3.28500897e-01 5.79705179e-01 -1.18570983e-01 6.85328305e-01
1.76384017e-01 -1.22421809e-01 2.72064060e-01 5.12512326e-01
-5.50282180e-01 -1.12064552e+00 -4.49751675e-01 8.80916476e-01
-4.79550034e-01 -1.08607359e-01 -1.75459638e-01 6.45872772e-01
6.87985718e-01 7.38644540e-01 2.23876134e-01 -6.02638364e-01
3.94724339e-01 2.72943437e-01 3.09303015e-01 -1.45502836e-01
-3.52724701e-01 -3.67051274e-01 2.47585222e-01 -4.73725677e-01
-4.88475174e-01 -7.14966893e-01 -7.84698665e-01 -5.95318615e-01
-1.76316381e-01 -3.01860958e-01 1.93620548e-01 9.93492424e-01
3.48719120e-01 8.23698044e-01 5.61418295e-01 -6.70970082e-01
-6.22863472e-01 -1.06848609e+00 -6.69964910e-01 8.37014854e-01
3.61779720e-01 -1.15808070e+00 -1.54224232e-01 -1.10803105e-01]
|
[14.312019348144531, 1.0243993997573853]
|
9dea9adf-dc54-4f65-93ef-54a21c382db8
|
visual-speech-aware-perceptual-3d-facial
|
2207.11094
| null |
https://arxiv.org/abs/2207.11094v1
|
https://arxiv.org/pdf/2207.11094v1.pdf
|
Visual Speech-Aware Perceptual 3D Facial Expression Reconstruction from Videos
|
The recent state of the art on monocular 3D face reconstruction from image data has made some impressive advancements, thanks to the advent of Deep Learning. However, it has mostly focused on input coming from a single RGB image, overlooking the following important factors: a) Nowadays, the vast majority of facial image data of interest do not originate from single images but rather from videos, which contain rich dynamic information. b) Furthermore, these videos typically capture individuals in some form of verbal communication (public talks, teleconferences, audiovisual human-computer interactions, interviews, monologues/dialogues in movies, etc). When existing 3D face reconstruction methods are applied in such videos, the artifacts in the reconstruction of the shape and motion of the mouth area are often severe, since they do not match well with the speech audio. To overcome the aforementioned limitations, we present the first method for visual speech-aware perceptual reconstruction of 3D mouth expressions. We do this by proposing a "lipread" loss, which guides the fitting process so that the elicited perception from the 3D reconstructed talking head resembles that of the original video footage. We demonstrate that, interestingly, the lipread loss is better suited for 3D reconstruction of mouth movements compared to traditional landmark losses, and even direct 3D supervision. Furthermore, the devised method does not rely on any text transcriptions or corresponding audio, rendering it ideal for training in unlabeled datasets. We verify the efficiency of our method through exhaustive objective evaluations on three large-scale datasets, as well as subjective evaluation with two web-based user studies.
|
['Petros Maragos', 'Anastasios Roussos', 'Athanasios Katsamanis', 'Foivos Paraperas-Papantoniou', 'George Retsinas', 'Panagiotis P. Filntisis']
|
2022-07-22
| null | null | null | null |
['3d-face-reconstruction', 'face-reconstruction']
|
['computer-vision', 'computer-vision']
|
[ 6.85041994e-02 1.90924957e-01 -9.51398239e-02 -2.93793708e-01
-6.48471594e-01 -4.11143988e-01 6.13701880e-01 -1.41314596e-01
-3.18501741e-01 4.60626394e-01 2.65776604e-01 1.23832077e-01
1.43272236e-01 -2.37139806e-01 -7.92291880e-01 -7.52926946e-01
2.09279642e-01 1.94902942e-01 -1.87051386e-01 -9.40691009e-02
-1.12672243e-03 6.78705215e-01 -1.85532486e+00 6.98805079e-02
3.80476534e-01 1.36733174e+00 1.40472159e-01 3.79895538e-01
-8.68540853e-02 4.84457701e-01 -5.37910938e-01 -6.30984545e-01
1.71085700e-01 -2.66343236e-01 -3.26787502e-01 6.01516604e-01
4.43567067e-01 -4.69323158e-01 -3.23366970e-01 9.78388071e-01
6.60434127e-01 2.07811557e-02 5.24887085e-01 -1.13413751e+00
-1.86254367e-01 -1.20777532e-01 -4.59565729e-01 -3.61724675e-01
9.73077655e-01 1.74870476e-01 7.42142975e-01 -1.12459087e+00
8.73158932e-01 1.38346028e+00 5.68428814e-01 7.39533901e-01
-1.22407687e+00 -4.84419554e-01 6.58057339e-04 1.84522301e-01
-1.51478767e+00 -1.01179421e+00 1.15412068e+00 -2.65863746e-01
4.61198419e-01 2.26412460e-01 8.77142429e-01 1.56390131e+00
-2.66737640e-01 8.53477955e-01 1.01898587e+00 -5.00679910e-01
1.60208702e-01 5.50392687e-01 -5.08976758e-01 4.50774729e-01
-3.92515689e-01 8.44513923e-02 -6.20204389e-01 3.65324616e-02
6.26874089e-01 -4.09136154e-02 -6.80946648e-01 -7.03444123e-01
-8.23041201e-01 5.41735709e-01 2.11207852e-01 3.92730236e-01
-2.36756176e-01 -2.60670543e-01 2.71368742e-01 1.86837390e-01
5.24832904e-01 -2.87397504e-01 -1.57090634e-01 -1.34208038e-01
-1.02496803e+00 6.76138476e-02 6.48234427e-01 8.69533420e-01
6.89387441e-01 -7.88393393e-02 3.22652489e-01 8.01402986e-01
5.13262093e-01 5.43218970e-01 5.26774585e-01 -8.25264156e-01
3.85348558e-01 3.84984344e-01 2.24872172e-01 -1.14281034e+00
-4.01292980e-01 -2.67293192e-02 -7.72948265e-01 1.85620442e-01
5.49483657e-01 4.76977602e-02 -4.81242150e-01 1.75166130e+00
5.85783899e-01 -1.35112284e-02 -4.46518250e-02 1.14172554e+00
9.89650130e-01 3.93491507e-01 -2.21162826e-01 -6.84915602e-01
1.14385128e+00 -4.58418161e-01 -9.64512467e-01 1.11453328e-03
1.37812868e-01 -7.94194698e-01 1.12201071e+00 5.65229893e-01
-1.21088135e+00 -7.60789037e-01 -6.04517996e-01 -7.92181119e-02
-1.01148644e-02 2.16799915e-01 1.77531660e-01 7.60486186e-01
-1.06511319e+00 3.80282223e-01 -5.71978927e-01 -4.19076234e-01
7.24926144e-02 1.79876864e-01 -8.40359509e-01 -3.93353850e-02
-9.15423751e-01 5.88728905e-01 -2.00080514e-01 4.02103096e-01
-7.31317639e-01 -3.44032019e-01 -8.49609852e-01 -8.82516801e-02
4.79655802e-01 -2.70033121e-01 1.09209013e+00 -1.28547335e+00
-1.83099103e+00 1.21333086e+00 -4.63011384e-01 -1.53402075e-01
9.12204862e-01 -2.20085472e-01 -4.48786795e-01 3.78002524e-01
-3.00075084e-01 6.08524442e-01 1.34075046e+00 -1.46912503e+00
-3.01131885e-02 -7.15263367e-01 4.86859865e-02 1.38584912e-01
-4.79489356e-01 -5.93230762e-02 -7.98859298e-01 -4.74562138e-01
3.76318134e-02 -7.73405552e-01 2.13253677e-01 4.11396652e-01
-3.29260945e-01 9.27843247e-03 7.86179245e-01 -6.35419846e-01
6.54434979e-01 -2.56525373e+00 3.32701802e-01 1.62157565e-02
-5.48480339e-02 2.86838919e-01 7.05524608e-02 3.55516940e-01
-1.32859692e-01 -1.73221141e-01 -1.12912446e-01 -9.66641068e-01
1.27038479e-01 4.64749336e-02 -2.26741269e-01 8.04922998e-01
2.18935050e-02 5.32833338e-01 -6.32657945e-01 -6.13478124e-01
4.12973195e-01 9.28937197e-01 -5.68825722e-01 4.56497759e-01
-1.16775163e-01 8.91710639e-01 -6.42835125e-02 7.06294298e-01
7.29566514e-01 8.47811252e-02 2.13118210e-01 -3.86069208e-01
2.69402098e-02 -4.99074832e-02 -1.20629358e+00 1.97433507e+00
-5.99374235e-01 5.77348888e-01 6.67551517e-01 -1.00244927e+00
8.83112013e-01 7.06466615e-01 5.59672594e-01 -6.66474342e-01
3.09765995e-01 1.51131287e-01 -5.47139883e-01 -8.35835040e-01
2.03775972e-01 -2.19373181e-01 2.54937708e-01 2.25025490e-01
1.20944038e-01 -3.45565856e-01 -1.92449450e-01 -2.17353791e-01
5.10824323e-01 2.65921593e-01 1.73240498e-01 3.54688197e-01
8.14555526e-01 -6.53253019e-01 4.11450744e-01 1.05606325e-01
-3.02517444e-01 8.33565116e-01 5.21028996e-01 -9.61963013e-02
-7.48321652e-01 -9.27100122e-01 -2.19866559e-01 5.35395324e-01
1.49686456e-01 -4.04608160e-01 -9.55231845e-01 -5.36546230e-01
-1.90331757e-01 1.95484743e-01 -4.66111749e-01 8.33087638e-02
-3.39827985e-01 -2.03140140e-01 3.92526239e-01 4.33342792e-02
3.70512128e-01 -1.11232114e+00 -4.97440338e-01 2.30085328e-02
-3.70735049e-01 -1.66850603e+00 -4.62950528e-01 -3.14277232e-01
-7.27483273e-01 -1.10871136e+00 -1.04850769e+00 -5.77068210e-01
7.03710973e-01 2.04924315e-01 7.15874732e-01 -8.07029828e-02
-2.55171180e-01 6.22268021e-01 -3.30805510e-01 7.59749115e-02
-4.44167256e-01 -4.09841925e-01 3.83172482e-01 6.11139953e-01
1.59734488e-01 -7.13521421e-01 -5.55011988e-01 5.12233794e-01
-7.16515362e-01 -2.60477901e-01 3.90656799e-01 8.02120805e-01
3.12876284e-01 -2.57797446e-02 3.15417081e-01 -3.80482435e-01
3.41932058e-01 -1.98813379e-01 -5.93831837e-01 -5.91563396e-02
-2.09982514e-01 -4.98686522e-01 7.14280486e-01 -5.67848325e-01
-1.14710522e+00 2.62214124e-01 -5.50271749e-01 -8.11526835e-01
-6.29573882e-01 -3.42812240e-02 -4.87522781e-01 -1.07291713e-01
4.95779753e-01 2.82249302e-01 3.25462967e-01 -6.19310379e-01
1.10689178e-01 8.70188773e-01 6.19911253e-01 -2.66341984e-01
6.80553734e-01 7.02078462e-01 -1.32222429e-01 -1.37213695e+00
-3.56752276e-01 -4.32648182e-01 -6.76197469e-01 -4.77705598e-01
7.49237597e-01 -8.67435217e-01 -1.24899125e+00 4.41532612e-01
-1.15545595e+00 -1.02851257e-01 -1.29552186e-01 5.60058236e-01
-7.48141587e-01 7.12609172e-01 -4.76888061e-01 -1.16870141e+00
2.12987550e-02 -1.44582748e+00 1.26352835e+00 8.16391706e-02
-1.14562720e-01 -8.27621162e-01 -2.40374550e-01 7.46349394e-01
5.13781570e-02 7.36637786e-02 7.03341246e-01 -5.59529774e-02
-3.73330146e-01 -4.01734769e-01 8.34712107e-03 5.56101084e-01
2.02236593e-01 -8.50713551e-02 -1.46528113e+00 -2.80281872e-01
2.41379440e-01 -5.67570746e-01 3.54045242e-01 3.49913657e-01
1.01749825e+00 -2.78326303e-01 -5.36617525e-02 5.71548760e-01
1.06018245e+00 -6.01620078e-02 5.35248697e-01 -2.54564732e-01
4.91733879e-01 1.17137301e+00 4.91347313e-01 5.98944902e-01
2.14445919e-01 1.23874509e+00 8.12757611e-01 -1.17149651e-01
-1.45282581e-01 -3.38311434e-01 6.05979264e-01 6.15657389e-01
-8.31966624e-02 -1.70971557e-01 -3.98142248e-01 4.17607903e-01
-1.39341843e+00 -6.60394728e-01 1.75215244e-01 2.50980496e+00
6.15489602e-01 7.46597489e-03 2.17251956e-01 5.35439730e-01
4.94737178e-01 1.25019535e-01 -4.46162462e-01 -1.47136301e-01
-1.79456905e-01 -8.06755573e-02 -7.36008063e-02 5.05652666e-01
-8.79857183e-01 6.98102415e-01 5.24772978e+00 6.94730043e-01
-1.40592313e+00 5.94103895e-02 3.01735759e-01 -1.95805892e-01
-2.45750561e-01 -2.92408556e-01 -6.08529747e-01 3.65219951e-01
5.54557621e-01 4.50238824e-01 3.73988926e-01 7.09848821e-01
6.57551706e-01 -1.32172391e-01 -1.17256820e+00 1.46133971e+00
3.27913553e-01 -8.95128429e-01 -1.63031831e-01 2.21588910e-01
2.95477033e-01 -4.66588467e-01 2.61545092e-01 -1.30031392e-01
-6.01980865e-01 -1.03592300e+00 9.17740405e-01 3.38457704e-01
9.08386588e-01 -6.73713207e-01 5.27477503e-01 4.58830833e-01
-8.20156455e-01 -7.00863227e-02 -8.80679488e-02 2.83004880e-01
3.76298517e-01 4.92035866e-01 -8.03516150e-01 4.10313368e-01
8.29254627e-01 5.83592236e-01 -1.45030767e-01 7.33104110e-01
-2.22804204e-01 2.08288074e-01 -2.63538092e-01 1.96938038e-01
-7.20106736e-02 -1.58632845e-01 8.35857928e-01 9.82595205e-01
2.69310802e-01 -2.72228830e-02 -2.22496629e-01 6.78322494e-01
-1.86809853e-01 4.63561535e-01 -8.18087339e-01 2.80843489e-02
2.98871752e-02 1.11253035e+00 -4.92075324e-01 1.79101855e-01
-5.65368831e-01 9.87494588e-01 -2.92266682e-02 5.12802303e-01
-5.68942487e-01 3.90017293e-02 6.91474795e-01 4.47176695e-01
3.23991388e-01 -2.26488605e-01 3.15107465e-01 -1.19305205e+00
3.53500366e-01 -1.02094412e+00 -1.78912833e-01 -9.71064270e-01
-8.72111738e-01 6.89933658e-01 -1.48705900e-01 -1.40571976e+00
-5.34940183e-01 -4.93182242e-01 -1.67672247e-01 4.32469487e-01
-1.49192977e+00 -9.02186573e-01 -4.91425037e-01 9.36274052e-01
5.43561757e-01 -4.48273867e-03 8.49721909e-01 6.17710471e-01
-3.80569130e-01 7.89417863e-01 -1.86160102e-01 -7.87974298e-02
8.40705037e-01 -7.46649742e-01 -9.38190520e-02 4.32135791e-01
2.85416991e-01 3.14210385e-01 8.17693830e-01 -9.95784104e-02
-1.62067163e+00 -4.27696973e-01 7.89083898e-01 -1.06191821e-01
2.58186579e-01 -7.01358438e-01 -7.84512222e-01 3.19276512e-01
-8.56161267e-02 1.36753753e-01 4.12754297e-01 -1.65331304e-01
-1.63386628e-01 -3.84036839e-01 -1.26755798e+00 4.67050731e-01
9.47771430e-01 -7.48961151e-01 -2.81531394e-01 2.39672244e-01
3.26143801e-01 -4.67611194e-01 -7.58089185e-01 2.94899076e-01
7.12438822e-01 -1.33926916e+00 9.75502670e-01 -1.06005989e-01
1.88592136e-01 -1.23373114e-01 -2.71179020e-01 -1.07443035e+00
5.64277053e-01 -7.67034352e-01 -4.19322178e-02 1.34481025e+00
-1.78807657e-02 -3.12547952e-01 9.22741830e-01 2.52771825e-01
1.36220455e-01 -4.17250931e-01 -1.16265321e+00 -4.36633080e-01
-4.41168934e-01 -5.73445559e-01 3.88276339e-01 7.30861127e-01
-9.00449604e-02 6.97603151e-02 -7.00037479e-01 1.28265014e-02
7.36161590e-01 2.81885117e-02 1.10329103e+00 -1.22932959e+00
-2.93669045e-01 -1.68937847e-01 -5.58207452e-01 -1.24118185e+00
5.00206113e-01 -4.38438028e-01 6.96481541e-02 -8.59043360e-01
-1.67705983e-01 -1.82169005e-01 4.09381241e-01 1.23374984e-01
3.87598068e-01 4.75048691e-01 1.64691359e-01 1.98768690e-01
-1.78026333e-01 8.99206281e-01 1.36216426e+00 -3.86210647e-03
-1.69632584e-01 3.28495443e-01 -1.70053616e-01 8.96566510e-01
1.68275356e-01 -9.45622846e-02 -4.30550188e-01 -1.57584459e-01
3.15105245e-02 5.20199060e-01 5.21932364e-01 -7.09197879e-01
1.62163153e-01 2.22620875e-01 3.69564831e-01 -3.23258847e-01
9.86763299e-01 -1.17523980e+00 1.66454539e-01 8.57111663e-02
-1.49534196e-01 -3.82288337e-01 1.95008889e-02 6.19486928e-01
-5.61903298e-01 -1.17861815e-01 8.33122075e-01 -1.22064129e-01
-3.90354544e-01 2.49895170e-01 -3.17667633e-01 -9.20681804e-02
7.79763818e-01 -3.70324373e-01 3.02690119e-01 -9.81254280e-01
-1.05039501e+00 -2.72724748e-01 6.87122822e-01 3.64826322e-01
7.54140615e-01 -1.12191975e+00 -4.06880647e-01 6.59266710e-01
-3.01457960e-02 1.38819115e-02 4.57093894e-01 1.03540647e+00
-2.43472442e-01 3.33502740e-01 -7.11989030e-02 -8.67844403e-01
-1.38632131e+00 7.03948617e-01 3.41919661e-01 2.74275124e-01
-8.86779964e-01 5.11254847e-01 3.21061224e-01 -3.48436147e-01
7.58886516e-01 -1.52539641e-01 -2.08408311e-01 3.27659875e-01
5.09767413e-01 1.18806183e-01 3.27125937e-01 -1.07623613e+00
-3.23964626e-01 9.45975780e-01 3.09163064e-01 -2.11077914e-01
1.17953753e+00 -4.70995575e-01 2.28861064e-01 5.12543976e-01
1.44024825e+00 2.69870371e-01 -1.41866076e+00 -2.27072731e-01
-4.36311424e-01 -6.36937559e-01 -8.99098516e-02 -1.84154257e-01
-1.19504046e+00 1.22443390e+00 6.11163318e-01 2.47767150e-01
1.28975391e+00 3.78153399e-02 6.96520329e-01 1.09866224e-01
4.19032872e-01 -9.34231997e-01 2.85250485e-01 -4.20632772e-02
1.08872032e+00 -1.27492166e+00 -2.33707294e-01 -4.36027706e-01
-5.67025185e-01 1.14014137e+00 1.79645896e-01 3.10711294e-01
7.94704676e-01 1.92289464e-02 2.32315540e-01 7.28282481e-02
-4.63467211e-01 -1.49118453e-01 1.59685954e-01 6.11501455e-01
2.63484329e-01 -2.87987411e-01 7.11554065e-02 2.95998991e-01
-1.96892381e-01 3.05917521e-04 4.24859613e-01 4.29194897e-01
1.08750753e-01 -9.38009381e-01 -5.58302402e-01 -4.13580716e-01
-4.75290179e-01 2.35728085e-01 -3.94997358e-01 8.77729714e-01
1.53414384e-02 1.07649744e+00 -2.52108537e-02 -2.00225264e-01
5.51185608e-01 1.37373671e-01 7.42844939e-01 -1.99417531e-01
-2.51773447e-01 4.68856335e-01 -1.58712789e-01 -7.86403537e-01
-6.81097150e-01 -6.86734676e-01 -9.02510285e-01 -2.70072639e-01
-1.63265198e-01 -1.01278564e-02 9.86754417e-01 9.33170676e-01
2.23391969e-02 1.28849205e-02 8.87216926e-01 -1.42547500e+00
-3.97914946e-01 -8.38997364e-01 -7.68886924e-01 6.15267158e-01
7.65900671e-01 -8.87254119e-01 -7.28779018e-01 8.07537735e-02]
|
[13.188892364501953, -0.30633479356765747]
|
2d1a450b-9bd9-4eab-a023-2d82bda2b7b7
|
bilingually-guided-monolingual-dependency
| null | null |
https://aclanthology.org/P13-1105
|
https://aclanthology.org/P13-1105.pdf
|
Bilingually-Guided Monolingual Dependency Grammar Induction
| null |
['Yajuan L{\\"u}', 'Qun Liu', 'Kai Liu', 'Wenbin Jiang']
|
2013-08-01
| null | null | null |
acl-2013-8
|
['dependency-grammar-induction']
|
['natural-language-processing']
|
[-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01]
|
[-7.3634443283081055, 3.8467299938201904]
|
586b6928-c052-4e5b-ae3a-a49391ca4f4c
|
video-highlights-detection-and-summarization-1
|
1708.0221
| null |
http://arxiv.org/abs/1708.02210v1
|
http://arxiv.org/pdf/1708.02210v1.pdf
|
Video Highlights Detection and Summarization with Lag-Calibration based on Concept-Emotion Mapping of Crowd-sourced Time-Sync Comments
|
With the prevalence of video sharing, there are increasing demands for
automatic video digestion such as highlight detection. Recently, platforms with
crowdsourced time-sync video comments have emerged worldwide, providing a good
opportunity for highlight detection. However, this task is non-trivial: (1)
time-sync comments often lag behind their corresponding shot; (2) time-sync
comments are semantically sparse and noisy; (3) to determine which shots are
highlights is highly subjective. The present paper aims to tackle these
challenges by proposing a framework that (1) uses concept-mapped lexical-chains
for lag calibration; (2) models video highlights based on comment intensity and
combination of emotion and concept concentration of each shot; (3) summarize
each detected highlight using improved SumBasic with emotion and concept
mapping. Experiments on large real-world datasets show that our highlight
detection method and summarization method both outperform other benchmarks with
considerable margins.
|
['Chaomei Chen', 'Qing Ping']
|
2017-08-07
| null | null | null | null |
['highlight-detection']
|
['computer-vision']
|
[ 1.67067535e-02 -5.40188670e-01 -4.21630323e-01 1.50417969e-01
-8.91708910e-01 -7.17913449e-01 5.51676095e-01 7.75470316e-01
-2.96782374e-01 5.85256517e-01 7.33345330e-01 4.23143566e-01
3.32322001e-01 -3.29966217e-01 -3.91821653e-01 -3.19621533e-01
-2.27036431e-01 -1.02254532e-01 6.80063009e-01 -1.02562860e-01
5.74548244e-01 1.15863681e-01 -1.83664203e+00 4.42902535e-01
8.19879413e-01 1.19680870e+00 1.86453402e-01 6.62962794e-01
-4.98149008e-01 1.22718108e+00 -8.33035886e-01 -3.99704069e-01
1.24362640e-01 -6.95284784e-01 -1.87772393e-01 2.54859924e-01
2.87753284e-01 -1.54624104e-01 6.59610480e-02 1.08847868e+00
3.75302225e-01 2.26345196e-01 4.02429372e-01 -1.52157056e+00
-3.60912144e-01 6.77127600e-01 -8.61549675e-01 5.87179303e-01
5.63286006e-01 -2.95364559e-01 1.15130925e+00 -1.07877409e+00
1.02822077e+00 9.82741833e-01 5.51566482e-01 1.46328956e-01
-6.92813516e-01 -6.00128710e-01 3.77714127e-01 3.15496504e-01
-1.49852645e+00 -2.32840478e-01 8.50875020e-01 -7.15091944e-01
6.44175589e-01 2.62323678e-01 9.05380309e-01 1.08309269e+00
-2.65573680e-01 8.06511223e-01 8.17798436e-01 -3.81656289e-01
3.61606300e-01 2.37720996e-01 -2.31232539e-01 3.11910897e-01
-9.30277109e-02 -6.62910044e-01 -1.23169076e+00 -1.76792935e-01
2.04884961e-01 2.34430909e-01 -3.19416940e-01 5.97959869e-02
-1.44741058e+00 7.07510352e-01 -2.60219164e-02 2.65672714e-01
-6.02141380e-01 -8.87003392e-02 9.18152988e-01 1.31747529e-01
1.00619292e+00 3.62282693e-01 -1.42232198e-02 -7.07830787e-01
-1.45750046e+00 4.84874219e-01 7.77383626e-01 1.23458910e+00
7.47969389e-01 -5.89076243e-02 -4.77719992e-01 8.92858863e-01
-2.80272275e-01 3.87290806e-01 3.66390318e-01 -7.22882926e-01
4.90146548e-01 8.10133278e-01 3.79397511e-01 -1.48425210e+00
-2.49318972e-01 -8.84679481e-02 -4.87047106e-01 -2.27800891e-01
1.46709770e-01 -2.28823602e-01 -3.91762346e-01 1.35738230e+00
1.68112949e-01 5.28266490e-01 -4.29533750e-01 9.73343372e-01
8.97925735e-01 7.58073211e-01 2.43295088e-01 -6.47022963e-01
1.32491887e+00 -9.15922940e-01 -1.06320512e+00 2.14404631e-02
3.68707806e-01 -1.03170323e+00 1.06654298e+00 2.98659801e-01
-1.10431218e+00 -4.32399333e-01 -9.66612577e-01 6.05256110e-02
-4.43208396e-01 8.65437463e-03 2.23250940e-01 3.95759910e-01
-9.30774987e-01 2.11281732e-01 -2.75488824e-01 -6.04269981e-01
2.49267742e-01 -2.57341623e-01 -1.25110969e-01 1.90801814e-01
-1.03203404e+00 3.70904237e-01 3.79127860e-01 -4.53275621e-01
-4.50937659e-01 -9.82246816e-01 -6.06600463e-01 -2.82062441e-02
8.72229040e-01 -1.27574623e-01 1.10030591e+00 -1.47183287e+00
-1.18136740e+00 8.58648777e-01 -2.94846177e-01 -3.74198228e-01
6.46157801e-01 -4.27205473e-01 -6.25522196e-01 5.42349339e-01
4.70398873e-01 6.89887822e-01 9.32212770e-01 -1.17083824e+00
-1.19331706e+00 1.98088452e-01 6.56708851e-02 3.08198929e-01
-6.77865982e-01 6.34195387e-01 -7.46288657e-01 -9.29720283e-01
-2.25710452e-01 -6.39337063e-01 2.43523628e-01 3.39477020e-03
-5.29690683e-01 3.99279734e-03 1.03491664e+00 -5.51121712e-01
1.78274930e+00 -2.27631330e+00 -7.49394223e-02 -1.48781473e-02
4.16088015e-01 4.10439819e-02 1.74590021e-01 9.14227605e-01
2.68106967e-01 2.09556460e-01 2.41616756e-01 -2.62387335e-01
-1.88980311e-01 -3.22800130e-01 -5.73238373e-01 1.86519966e-01
1.59469098e-01 5.93754947e-01 -1.24338138e+00 -8.89360964e-01
7.60777146e-02 2.40188152e-01 -4.26986784e-01 2.07881182e-01
-3.05726588e-01 2.32281193e-01 -2.36492112e-01 8.47777367e-01
3.54953468e-01 -3.68196368e-01 5.71362535e-03 -2.08792344e-01
-6.63459003e-01 -2.05653727e-01 -1.22078729e+00 1.36205566e+00
-1.52844889e-02 9.78125930e-01 -2.68327206e-01 -4.44917947e-01
8.97422075e-01 3.24438691e-01 8.73808742e-01 -6.52096748e-01
5.41912355e-02 5.80807738e-02 -6.19230390e-01 -6.83133721e-01
1.09641993e+00 3.35052274e-02 -1.81503177e-01 3.67019504e-01
-2.43485466e-01 2.82263458e-01 4.36501145e-01 5.71336091e-01
1.12587917e+00 -1.04269885e-01 4.75523412e-01 -4.53477502e-02
4.27180171e-01 9.25311521e-02 7.05317438e-01 4.32114482e-01
-3.81189406e-01 9.91452634e-01 8.45707655e-01 -3.18721682e-01
-9.29428935e-01 -8.31906140e-01 5.38221240e-01 1.29809952e+00
3.99545103e-01 -8.15069616e-01 -6.50735199e-01 -4.35186356e-01
-8.20671692e-02 3.39702070e-01 -6.06153369e-01 3.74039024e-01
-3.42993766e-01 -1.89951882e-01 2.85607874e-01 4.40697491e-01
3.04375708e-01 -9.90563512e-01 -8.45970750e-01 3.37199926e-01
-6.99828148e-01 -1.37749887e+00 -8.85665476e-01 -1.73221081e-01
-3.04890513e-01 -1.19790423e+00 -1.02181053e+00 -4.71853197e-01
5.18422782e-01 9.27391112e-01 1.16092277e+00 2.22975731e-01
-9.00677890e-02 5.29758155e-01 -1.08735573e+00 -6.76093757e-01
2.86168251e-02 -7.58214742e-02 -6.39318302e-02 3.61865282e-01
5.90330720e-01 -2.62632310e-01 -7.23179102e-01 3.09057206e-01
-9.35569644e-01 7.71937668e-02 5.87824173e-02 3.08158159e-01
6.89604044e-01 2.64455155e-02 7.75618613e-01 -9.03097987e-01
7.84859121e-01 -8.08835804e-01 -3.26690465e-01 9.39911529e-02
-4.09989089e-01 -7.58813918e-01 4.44206238e-01 -5.49699903e-01
-9.70170915e-01 -2.45748654e-01 4.82276022e-01 -6.85355842e-01
1.92937598e-01 5.63683271e-01 2.49988139e-01 4.22341406e-01
6.24060810e-01 -8.92061070e-02 -2.93459415e-01 -2.92911530e-01
2.57285863e-01 6.43642902e-01 5.03360987e-01 -2.05186918e-01
6.07608020e-01 8.81622136e-01 -3.38013172e-01 -1.08373678e+00
-9.17272031e-01 -1.09122765e+00 -4.59734797e-01 -8.91336441e-01
8.81344914e-01 -1.41457760e+00 -3.39185774e-01 2.79839307e-01
-1.08269906e+00 -7.68135414e-02 -2.18971074e-01 3.08378726e-01
-2.67130405e-01 5.88170588e-01 -5.16228139e-01 -1.01350069e+00
-2.32387409e-01 -7.79229403e-01 1.07415092e+00 4.68807131e-01
-5.91096580e-01 -6.91210926e-01 6.76432112e-03 6.15401343e-02
2.73945808e-01 5.94424307e-01 3.27597588e-01 -5.28764784e-01
-4.46456254e-01 -3.93027842e-01 -4.48011488e-01 7.78151527e-02
1.54907233e-03 4.92417157e-01 -9.26069975e-01 3.10657751e-02
-4.48403060e-01 -1.64670780e-01 6.25826716e-01 4.74487901e-01
9.74274099e-01 -2.56389767e-01 -2.84967154e-01 1.82799846e-01
1.38010466e+00 6.32108673e-02 5.11478245e-01 4.77238774e-01
5.79385042e-01 7.69604862e-01 1.00577617e+00 1.08587503e+00
4.93483722e-01 6.44079208e-01 3.67182642e-01 8.94624218e-02
-5.16930297e-02 -4.58856463e-01 4.13959891e-01 1.13078344e+00
-1.78905413e-01 -5.28650343e-01 -7.68828809e-01 8.30055535e-01
-2.28019023e+00 -1.26735258e+00 -2.86911041e-01 2.02685165e+00
4.60845917e-01 2.50307232e-01 7.33199894e-01 1.82513088e-01
1.01423192e+00 4.58506018e-01 -2.26765990e-01 -4.97369170e-02
-3.34176451e-01 -3.36887419e-01 4.11052972e-01 -9.35458019e-02
-1.05191720e+00 7.65824080e-01 5.81138515e+00 7.58384526e-01
-1.07879174e+00 3.49353433e-01 5.50861120e-01 -3.99121374e-01
-1.96911812e-01 -7.00646788e-02 -6.39972746e-01 1.01566827e+00
6.83276892e-01 -4.68700916e-01 9.40669104e-02 8.13253403e-01
4.73292798e-01 -5.02491355e-01 -8.31958592e-01 1.23262465e+00
5.07460177e-01 -1.44518089e+00 -1.09303169e-01 -3.30048919e-01
1.05728841e+00 -2.57925123e-01 -2.51379907e-01 2.26174131e-01
-1.65721506e-01 -5.02722800e-01 1.16679621e+00 5.87123096e-01
8.32087278e-01 -7.62688994e-01 5.78012526e-01 4.59068827e-02
-1.45986891e+00 -1.32574635e-02 -1.88676000e-01 -8.66633430e-02
3.82977754e-01 8.03410947e-01 -4.98442143e-01 3.15867662e-01
8.61073017e-01 1.08706486e+00 -4.29963619e-01 1.40823507e+00
-2.01992646e-01 5.56350350e-01 3.00909076e-02 -3.01400900e-01
1.02360658e-01 -4.87785526e-02 6.63196862e-01 1.78901875e+00
6.22500122e-01 -5.26191816e-02 2.84468800e-01 6.07269108e-01
-2.38312662e-01 5.68653107e-01 -3.03578019e-01 -1.09441668e-01
8.89502764e-01 1.38951945e+00 -1.41205704e+00 -3.18626314e-01
-6.75154209e-01 9.36802208e-01 5.67315752e-03 3.89114290e-01
-9.43364084e-01 -2.65400529e-01 6.56191289e-01 2.74882317e-01
3.66110951e-01 1.47159882e-02 -3.29025723e-02 -1.13336289e+00
1.47511169e-01 -5.54800749e-01 3.60928744e-01 -9.66976345e-01
-1.27748096e+00 4.18512821e-01 -4.92909737e-02 -1.77758515e+00
2.77806027e-03 -5.73601834e-02 -7.68974662e-01 4.15889442e-01
-1.58979905e+00 -7.46740818e-01 -6.89172328e-01 4.49882060e-01
9.06282842e-01 1.07713658e-02 4.38513696e-01 4.75915253e-01
-3.57119620e-01 2.07852110e-01 -7.60071650e-02 -8.38871822e-02
1.14884257e+00 -1.14119124e+00 2.97279153e-02 9.00975525e-01
4.74067479e-02 7.56573826e-02 9.00365949e-01 -8.24971020e-01
-1.15802932e+00 -1.13157845e+00 9.86383319e-01 -2.95479625e-01
9.72334743e-01 -2.78040320e-01 -7.66099215e-01 2.12723345e-01
2.62525499e-01 -1.52630255e-01 9.58163321e-01 5.58404028e-02
-4.21547085e-01 -1.18942544e-01 -7.87852943e-01 6.95194960e-01
8.83776486e-01 -7.73108423e-01 -2.24685743e-01 4.69938129e-01
6.42607927e-01 -3.63931000e-01 -4.20248359e-01 -1.63229406e-01
5.32873333e-01 -1.13093925e+00 6.34864450e-01 -1.08350508e-01
7.22316504e-01 -4.25236732e-01 2.93237325e-02 -1.22375298e+00
-3.31875645e-02 -8.11291754e-01 -2.23768502e-01 1.61024046e+00
9.02888402e-02 9.60951895e-02 5.57926118e-01 3.21077794e-01
-6.80123046e-02 -3.40897292e-01 -6.50202811e-01 -6.52378321e-01
-8.15354288e-01 -4.92085904e-01 2.23507673e-01 1.04711068e+00
3.88436288e-01 6.82787001e-02 -7.24714100e-01 -1.10483304e-01
1.86497003e-01 -6.74193650e-02 7.44213104e-01 -1.43935275e+00
2.65386820e-01 -6.38020337e-01 -4.20716554e-01 -6.96305037e-01
-1.81054533e-01 -3.81254256e-01 1.79947808e-01 -1.44104886e+00
4.87593323e-01 -3.38468663e-02 -2.37414896e-01 5.90062588e-02
-3.48971486e-01 3.00591797e-01 3.29540610e-01 5.60684741e-01
-1.31894350e+00 1.29575163e-01 7.06905127e-01 1.26890734e-01
-2.68044740e-01 -3.09233516e-01 -6.00669920e-01 6.93004787e-01
4.28737581e-01 -3.49395752e-01 -4.09442186e-01 9.33461562e-02
9.46606815e-01 1.54735250e-02 9.38765779e-02 -1.13792109e+00
4.03194785e-01 -1.75731540e-01 -4.97088656e-02 -8.71924043e-01
2.14993715e-01 -7.50544369e-01 2.42995359e-02 -1.79273281e-02
-3.62417191e-01 4.04192269e-01 7.92740807e-02 9.76383448e-01
-5.49539328e-01 -1.53818786e-01 4.73123610e-01 -6.77839369e-02
-1.07050478e+00 1.32873923e-01 -5.97768784e-01 4.48945582e-01
1.22725213e+00 -3.97887707e-01 -4.91468281e-01 -8.20256472e-01
-1.50101230e-01 2.21269280e-01 6.45829082e-01 6.48969710e-01
5.16595721e-01 -1.26455331e+00 -6.47470534e-01 -4.36726332e-01
5.78424513e-01 -2.89419591e-01 3.88803154e-01 9.09075201e-01
-5.05007088e-01 -1.13079092e-02 -7.83576146e-02 -3.40756059e-01
-1.38479114e+00 6.75565541e-01 -3.82555187e-01 6.63822815e-02
-6.46624625e-01 8.19482386e-01 1.08166307e-01 5.00113189e-01
4.01976317e-01 -1.81270704e-01 -4.56124067e-01 1.04339564e+00
1.02054381e+00 6.30685389e-01 -2.41122782e-01 -9.35985982e-01
-3.33794862e-01 3.77757311e-01 9.76183116e-02 -2.27452405e-02
1.17240512e+00 -4.86863583e-01 1.67529374e-01 8.08923483e-01
1.03526449e+00 2.24191234e-01 -1.27066398e+00 -3.78114164e-01
3.37871164e-01 -5.91934919e-01 -8.79312009e-02 -4.32233304e-01
-8.30644667e-01 5.63218951e-01 1.90898389e-01 7.42406785e-01
1.22066307e+00 -2.28623807e-01 9.17347670e-01 -1.24619611e-01
1.88274398e-01 -1.78020823e+00 5.42263865e-01 3.00567865e-01
7.36384749e-01 -1.34429252e+00 2.51421213e-01 -5.83778083e-01
-1.09921503e+00 1.24842393e+00 3.07294726e-01 -4.66637947e-02
5.68920135e-01 1.41345412e-01 1.46959364e-01 -3.65817606e-01
-7.57012904e-01 -5.30352712e-01 3.67458135e-01 3.52694720e-01
3.88344735e-01 8.27238932e-02 -4.73044068e-01 5.17966330e-01
1.28210008e-01 3.39867361e-02 7.98005581e-01 9.20104980e-01
-6.36310577e-01 -6.35111272e-01 -3.25479805e-01 4.28957731e-01
-7.26576388e-01 3.00026741e-02 -5.19725442e-01 7.06315696e-01
1.23123206e-01 1.10487556e+00 6.93337172e-02 -3.90532672e-01
3.05024087e-01 2.15881541e-02 -2.83579916e-01 -6.20158374e-01
-7.88188577e-01 4.35205847e-01 1.32080436e-01 -5.21621108e-01
-8.37441683e-01 -7.83349395e-01 -1.02798378e+00 -2.06167862e-01
-2.94490874e-01 1.61343649e-01 6.29533410e-01 6.39270902e-01
5.00005007e-01 5.86862803e-01 6.13676548e-01 -8.57905626e-01
7.42803812e-02 -8.28631699e-01 -7.26445615e-01 7.06179321e-01
1.44221738e-01 -7.78566241e-01 -5.11251509e-01 5.07845104e-01]
|
[10.097009658813477, 0.4384283721446991]
|
ff9d7062-65f3-4dd9-b807-9263e1f16dca
|
h4vdm-h-264-video-device-matching
|
2210.11549
| null |
https://arxiv.org/abs/2210.11549v1
|
https://arxiv.org/pdf/2210.11549v1.pdf
|
H4VDM: H.264 Video Device Matching
|
Methods that can determine if two given video sequences are captured by the same device (e.g., mobile telephone or digital camera) can be used in many forensics tasks. In this paper we refer to this as "video device matching". In open-set video forensics scenarios it is easier to determine if two video sequences were captured with the same device than identifying the specific device. In this paper, we propose a technique for open-set video device matching. Given two H.264 compressed video sequences, our method can determine if they are captured by the same device, even if our method has never encountered the device in training. We denote our proposed technique as H.264 Video Device Matching (H4VDM). H4VDM uses H.264 compression information extracted from video sequences to make decisions. It is more robust against artifacts that alter camera sensor fingerprints, and it can be used to analyze relatively small fragments of the H.264 sequence. We trained and tested our method on a publicly available video forensics dataset consisting of 35 devices, where our proposed method demonstrated good performance.
|
['Edward J. Delp', 'Stefano Tubaro', 'Paolo Bestagini', 'Ziyue Xiang']
|
2022-10-20
| null | null | null | null |
['video-forensics']
|
['computer-vision']
|
[ 5.33708513e-01 -5.28099835e-01 -1.13443211e-01 1.02994025e-01
-8.43963683e-01 -1.10492694e+00 2.45906994e-01 3.05163693e-02
-1.22703180e-01 3.18098009e-01 -1.35376856e-01 -6.15179002e-01
1.61170319e-01 -5.43932319e-01 -9.85895216e-01 -5.90208650e-01
1.22422338e-01 -1.14030354e-01 5.82272589e-01 6.06771648e-01
6.36895835e-01 4.67444927e-01 -1.27676940e+00 3.74550879e-01
-2.77162157e-02 1.17281079e+00 3.57646137e-01 8.01731110e-01
3.10876310e-01 1.02483070e+00 -9.03252602e-01 -6.59021556e-01
6.27182424e-01 -3.88696402e-01 -6.59750819e-01 2.51703709e-01
4.89283890e-01 -8.72659385e-01 -5.38358986e-01 1.42232347e+00
3.83741200e-01 -1.19448826e-01 3.98592919e-01 -1.50826836e+00
-1.51458904e-01 3.05195898e-01 -8.64883125e-01 7.30730057e-01
9.19248462e-01 -6.08741567e-02 3.80490959e-01 -3.60852718e-01
7.12266386e-01 9.95333552e-01 7.86724865e-01 3.05497020e-01
-7.31453955e-01 -1.05644190e+00 -5.38958073e-01 5.97137332e-01
-1.66274846e+00 -6.17319405e-01 9.65309262e-01 -5.76644361e-01
5.02935648e-01 3.67432535e-01 3.36381376e-01 1.24056447e+00
3.14805239e-01 6.86213970e-01 8.91448915e-01 -3.97657812e-01
1.96697786e-01 7.97957033e-02 2.26459667e-01 5.41000724e-01
6.24803007e-01 -2.16605398e-03 -4.05857503e-01 -5.59233546e-01
6.64352953e-01 2.68180817e-01 -5.37040889e-01 -3.56029533e-02
-1.03892434e+00 5.47962725e-01 -4.63865489e-01 2.12760925e-01
-3.61600161e-01 -1.40804695e-02 6.69207454e-01 4.51693803e-01
-2.55758375e-01 1.11117855e-01 1.40060797e-01 -4.77048248e-01
-1.14939058e+00 1.94188785e-02 6.94927156e-01 1.02390587e+00
5.43511927e-01 -3.63158315e-01 3.07686239e-01 3.32459986e-01
1.60102472e-01 6.86431229e-01 3.31360877e-01 -1.23582017e+00
7.07468569e-01 -1.26722064e-02 1.65994167e-01 -1.51190078e+00
4.23621118e-01 5.48461199e-01 -3.50312680e-01 -7.53462166e-02
4.41425860e-01 6.60249125e-03 -2.82832950e-01 1.16074729e+00
-1.40764946e-02 8.90777886e-01 -2.44274400e-02 6.88676775e-01
5.75421393e-01 5.05169034e-01 -3.61256599e-01 -3.86430115e-01
1.47664225e+00 -4.31618482e-01 -7.85414696e-01 2.21829683e-01
3.85128766e-01 -1.00556839e+00 3.85954201e-01 6.59983754e-01
-8.60594094e-01 -5.14636576e-01 -1.14042974e+00 4.66017902e-01
-7.53067210e-02 -2.87887096e-01 -1.72125101e-02 1.10858250e+00
-6.37236297e-01 5.72435021e-01 -6.70948207e-01 -4.08077002e-01
3.46199989e-01 4.21074688e-01 -5.22726178e-01 -2.27766201e-01
-9.14679945e-01 1.47108123e-01 2.22533897e-01 -4.87884462e-01
-9.92969811e-01 -2.22046837e-01 -7.61399150e-01 -1.68212011e-01
6.01783454e-01 -7.76747391e-02 1.15109169e+00 -6.51755154e-01
-1.00912225e+00 1.05658615e+00 -2.30521411e-01 -6.19625151e-01
3.89351100e-01 -1.68383598e-01 -9.45887029e-01 8.18293393e-01
3.03871632e-01 -1.63357388e-02 1.45314527e+00 -1.12991011e+00
-5.43882728e-01 -4.36993271e-01 -3.93179283e-02 -6.26670003e-01
-3.00388277e-01 6.11958086e-01 -6.96296692e-01 -6.00886762e-01
-1.17577270e-01 -9.89282310e-01 5.41932702e-01 -1.83391362e-01
-7.36008942e-01 3.09228957e-01 1.47659278e+00 -9.96024787e-01
1.54849851e+00 -2.36021972e+00 -4.86371994e-01 4.09387559e-01
1.73981220e-01 7.16766059e-01 2.59657949e-01 5.04874885e-01
-3.67195122e-02 3.28617394e-01 -1.57384738e-01 -3.94312404e-02
-4.40648049e-01 2.57304031e-02 -6.32023156e-01 7.35759854e-01
-3.90826464e-01 2.88924068e-01 -7.71659017e-01 -6.54691815e-01
4.08547878e-01 2.34635249e-01 -7.43808001e-02 3.14375401e-01
4.84903723e-01 4.88146096e-01 -3.74314755e-01 9.40382898e-01
9.98289764e-01 -1.13963246e-01 3.04755241e-01 -4.32777137e-01
9.81936306e-02 -1.77745745e-01 -1.29754174e+00 1.33799875e+00
-6.59671500e-02 1.03704321e+00 -1.14975892e-01 -6.35479450e-01
9.45491314e-01 6.04829907e-01 4.20235574e-01 -6.04180694e-01
2.01842234e-01 1.37138978e-01 -2.97154874e-01 -1.10060227e+00
4.27552909e-01 3.85745585e-01 8.45915079e-02 6.25498235e-01
-1.94837585e-01 4.79987890e-01 -4.22972674e-03 -7.87972212e-02
1.79572105e+00 -1.38982251e-01 4.95114356e-01 1.65134430e-01
7.48420775e-01 -4.23018605e-01 5.82806289e-01 9.46871817e-01
-5.01369119e-01 6.27386034e-01 2.62064427e-01 -1.54317066e-01
-1.04565156e+00 -1.02531791e+00 9.83277336e-02 3.19960177e-01
5.65174639e-01 -7.69447625e-01 -8.55181813e-01 -6.25060618e-01
-1.37383550e-01 2.80509263e-01 -2.23635361e-01 2.67763939e-02
-7.31025159e-01 3.96938846e-02 1.08347571e+00 4.30497229e-01
7.89471745e-01 -6.00492120e-01 -8.25463414e-01 -6.58936203e-02
-3.16994250e-01 -1.62392461e+00 -5.75108409e-01 -5.24764657e-01
-7.50566602e-01 -1.73723900e+00 -4.51545030e-01 -7.10530639e-01
3.78902495e-01 8.03748310e-01 7.80845463e-01 2.05747575e-01
-5.76284006e-02 5.59820831e-01 -4.63459909e-01 -1.16684340e-01
-5.66658378e-01 -5.01742184e-01 2.44586572e-01 4.17981088e-01
7.16273963e-01 -4.22013760e-01 -4.24376547e-01 7.40707576e-01
-1.04809940e+00 -6.58556402e-01 3.86874168e-03 1.16276160e-01
5.20928741e-01 5.70188403e-01 -1.83474332e-01 -8.47260177e-01
5.40682852e-01 -8.00258100e-01 -4.95796949e-01 3.62480819e-01
-8.63886625e-02 -4.68842059e-01 7.18328655e-01 -6.27751052e-01
-5.23844182e-01 -8.87783617e-02 7.42719769e-02 -1.25252235e+00
-4.06931460e-01 1.78855024e-02 -4.44475919e-01 -3.02503586e-01
2.20618829e-01 7.63327554e-02 -1.61984161e-01 -5.97847641e-01
-3.76419336e-01 1.15526223e+00 8.91909420e-01 -3.51143986e-01
8.10410440e-01 7.33988643e-01 -1.32770434e-01 -1.06966507e+00
-3.20394598e-02 -7.66200900e-01 -2.42440939e-01 -4.02156591e-01
8.53742242e-01 -7.18887985e-01 -9.49400961e-01 7.69849777e-01
-1.35728621e+00 4.19009417e-01 3.77860576e-01 4.53370661e-01
-3.55025500e-01 1.14100444e+00 -4.93894041e-01 -1.07977009e+00
1.21676717e-02 -1.36115289e+00 1.15388715e+00 2.81231403e-02
-4.37452316e-01 -1.01162124e+00 6.75374689e-03 5.36749840e-01
-1.78463027e-01 4.44610834e-01 4.63007361e-01 -7.76774406e-01
-5.08496344e-01 -5.08632958e-01 -6.83331164e-03 3.76773834e-01
4.73608732e-01 8.58031660e-02 -1.11556983e+00 -4.66842264e-01
6.25952840e-01 1.69522554e-01 2.85260171e-01 1.89506173e-01
1.31214309e+00 -4.61494029e-01 -4.40560788e-01 8.31048965e-01
1.51651323e+00 8.94708931e-01 1.27554119e+00 6.78231359e-01
8.92686009e-01 4.18086320e-01 6.74007297e-01 4.78019655e-01
-2.54714966e-01 8.93658340e-01 2.45952025e-01 5.08574128e-01
1.64741531e-01 -3.75918180e-01 6.18573666e-01 4.57266062e-01
1.37957990e-01 -6.79940224e-01 -8.57160568e-01 3.19248140e-01
-1.60998428e+00 -1.49703610e+00 -1.75748438e-01 2.44362140e+00
-7.54858106e-02 1.36041492e-01 2.53641397e-01 7.06369102e-01
1.43504083e+00 4.41016145e-02 -4.97929096e-01 -1.38670802e-01
-9.75073725e-02 1.98547140e-01 6.87057912e-01 1.48295671e-01
-9.32450652e-01 3.30724567e-01 5.66600561e+00 9.99431610e-01
-9.78428423e-01 2.00568408e-01 3.72244298e-01 1.35873124e-01
1.50025353e-01 1.87706396e-01 -5.71274817e-01 1.19043386e+00
1.14823258e+00 2.53143817e-01 2.03642055e-01 5.58919847e-01
2.53114194e-01 -3.11769456e-01 -1.13147509e+00 1.78803551e+00
3.62354010e-01 -1.43068314e+00 -1.96703702e-01 4.70428675e-01
2.63958722e-01 -6.80872858e-01 -3.64670493e-02 -2.48248100e-01
-4.40255582e-01 -6.30797386e-01 6.99981391e-01 1.33562952e-01
8.15476537e-01 -7.15970576e-01 8.60432208e-01 1.67947739e-01
-1.41769576e+00 -3.77654620e-02 -2.47209102e-01 2.31337145e-01
3.78229827e-01 2.41025165e-01 -6.88983560e-01 4.01237071e-01
9.09497678e-01 8.49739194e-01 -5.32216549e-01 1.12823832e+00
9.83437151e-02 9.63551223e-01 -7.68152550e-02 4.58143055e-01
-1.91291183e-01 -2.25959927e-01 8.27381551e-01 8.94058526e-01
6.71715319e-01 -1.03910521e-01 -3.84413674e-02 5.28384387e-01
-1.88990235e-02 -2.14157924e-01 -1.12814784e+00 -2.63648063e-01
8.23999524e-01 6.67576492e-01 -8.13357055e-01 -1.35774747e-01
-5.50620615e-01 1.23967648e+00 -3.75409186e-01 7.76840299e-02
-1.03814387e+00 -5.44926941e-01 5.30363441e-01 3.11264366e-01
5.62868178e-01 -1.47758245e-01 3.71327311e-01 -1.09336996e+00
2.77328551e-01 -1.03373551e+00 3.57634485e-01 -8.62415373e-01
-1.13376153e+00 5.23379743e-01 2.86113154e-02 -1.82843530e+00
-3.56931120e-01 -5.22011578e-01 -6.89003110e-01 3.94691616e-01
-1.00914001e+00 -6.30978584e-01 -1.28159240e-01 1.00572026e+00
3.56957823e-01 -4.95904177e-01 4.80090767e-01 7.11006463e-01
-5.21488070e-01 5.97479403e-01 2.84086168e-01 6.48405612e-01
7.05276608e-01 -5.13063550e-01 4.79148328e-01 1.27233231e+00
1.64282918e-01 8.56411517e-01 5.71022093e-01 -9.89579201e-01
-1.65368092e+00 -7.56658435e-01 4.54880983e-01 -4.13002312e-01
6.25885546e-01 -2.40507588e-01 -9.14034069e-01 6.75248265e-01
1.41643748e-01 -1.87143579e-01 1.15893948e+00 -6.93869174e-01
-3.96977842e-01 4.96640541e-02 -1.47463381e+00 7.30299726e-02
7.75483787e-01 -1.04340672e+00 -5.15102804e-01 1.19662397e-01
3.43111426e-01 -2.02830583e-01 -8.62099946e-01 2.56162900e-02
7.02719390e-01 -1.27429700e+00 9.09071386e-01 -2.93903053e-01
2.42040142e-01 -4.60874557e-01 -7.21892655e-01 -2.55301744e-01
3.36472243e-01 -1.08591139e+00 -6.96475983e-01 1.54362047e+00
-4.54721361e-01 -3.47201139e-01 6.53775752e-01 4.09294277e-01
4.32858974e-01 -2.35141203e-01 -7.59852111e-01 -1.28144395e+00
-7.85331070e-01 -7.18450129e-01 8.61783981e-01 9.80189204e-01
-1.80946782e-01 -1.58020347e-01 -8.36168647e-01 4.94982928e-01
8.61926079e-01 -1.69782061e-02 9.01118696e-01 -9.20649946e-01
-6.82582438e-01 7.11413398e-02 -9.92942810e-01 -1.03676713e+00
1.36991516e-01 -4.11133975e-01 -4.90248203e-01 -6.31492496e-01
3.89778346e-01 -1.55408293e-01 -6.07066564e-02 -1.16888233e-01
2.16645703e-01 5.27328134e-01 4.39825565e-01 7.77163208e-01
-4.64863479e-01 -3.29667836e-01 3.92852455e-01 -3.67673635e-02
3.15109603e-02 2.27469310e-01 -3.41412961e-01 6.53482318e-01
8.12234282e-01 -6.43320501e-01 -3.56946260e-01 -3.26719761e-01
-5.45307882e-02 6.15795612e-01 5.23658931e-01 -1.35824203e+00
7.30269924e-02 1.58180460e-01 3.51262167e-02 -4.54622746e-01
1.36885121e-01 -1.34792519e+00 6.55990303e-01 5.27812481e-01
1.15976892e-01 3.83712560e-01 1.19348511e-01 7.54608035e-01
-2.62941748e-01 -7.57554054e-01 3.62568796e-01 -7.34356791e-02
-9.29829717e-01 9.97718796e-02 -6.41853273e-01 -1.96038708e-01
1.23998463e+00 -8.99098456e-01 -3.15531731e-01 -4.02763397e-01
-2.92633712e-01 -4.04771179e-01 1.02848697e+00 3.81798327e-01
8.71259689e-01 -1.23286033e+00 2.11777054e-02 1.30684301e-01
2.82472614e-02 -8.22896123e-01 7.43660331e-02 5.10832250e-01
-8.03025365e-01 9.36052054e-02 -2.47876182e-01 -7.30817020e-01
-1.88788521e+00 9.74422157e-01 -3.52415815e-03 1.39795333e-01
-5.09014547e-01 2.09935397e-01 -2.34692082e-01 3.54042470e-01
1.14753723e-01 9.15508494e-02 3.92482467e-02 -1.15991347e-01
1.12202799e+00 8.60890269e-01 5.56145646e-02 -1.10996425e+00
-6.41746700e-01 8.65199625e-01 1.09515443e-01 -1.00312509e-01
7.87678599e-01 -2.64082521e-01 1.34097949e-01 2.77948350e-01
1.80699408e+00 4.46310490e-01 -9.49345231e-01 -1.30939279e-02
-1.48438334e-01 -1.11134255e+00 -1.84384346e-01 1.00887138e-02
-1.32946837e+00 8.02216172e-01 9.12939906e-01 3.67291451e-01
1.06355739e+00 -2.55028784e-01 1.46775949e+00 1.53631374e-01
9.58464563e-01 -7.59500802e-01 -1.59231812e-01 -8.11491832e-02
-6.76215217e-02 -1.18741143e+00 3.02499495e-02 -5.72512925e-01
-5.58363736e-01 1.27335846e+00 1.93055883e-01 -8.14255849e-02
4.68346536e-01 4.21627790e-01 -2.45321780e-01 -1.50575653e-01
1.75099060e-01 2.91311771e-01 -1.03087850e-01 8.24733436e-01
-6.01124093e-02 -1.89816639e-01 1.83054090e-01 2.17758521e-01
-3.64443846e-02 1.04201131e-01 7.68664777e-01 1.47270417e+00
-1.73783660e-01 -1.20206726e+00 -9.59681928e-01 2.92896032e-01
-8.14196169e-01 2.54142612e-01 -5.41906238e-01 5.77715218e-01
3.34382087e-01 1.36131728e+00 1.99843198e-01 -9.56196606e-01
1.73425656e-02 -1.82422370e-01 5.47316134e-01 -1.58766434e-01
-4.82873529e-01 -1.49385422e-01 -1.94823354e-01 -8.09168816e-01
-8.54298532e-01 -1.04100251e+00 -9.23005104e-01 -7.74037898e-01
-1.76811531e-01 1.27953842e-01 5.12191594e-01 8.24194908e-01
2.56231099e-01 -6.21113703e-02 7.74520218e-01 -5.13322294e-01
-6.22537658e-02 -1.99477062e-01 -7.98053741e-01 7.47213542e-01
5.19415200e-01 -6.22528255e-01 -3.61821830e-01 4.58122641e-01]
|
[12.408952713012695, 1.0055556297302246]
|
7e52db81-e2b4-4ed1-8735-c8a00243f4cc
|
text2chart-a-multi-staged-chart-generator
|
2104.04584
| null |
https://arxiv.org/abs/2104.04584v1
|
https://arxiv.org/pdf/2104.04584v1.pdf
|
Text2Chart: A Multi-Staged Chart Generator from Natural Language Text
|
Generation of scientific visualization from analytical natural language text is a challenging task. In this paper, we propose Text2Chart, a multi-staged chart generator method. Text2Chart takes natural language text as input and produce visualization as two-dimensional charts. Text2Chart approaches the problem in three stages. Firstly, it identifies the axis elements of a chart from the given text known as x and y entities. Then it finds a mapping of x-entities with its corresponding y-entities. Next, it generates a chart type suitable for the given text: bar, line or pie. Combination of these three stages is capable of generating visualization from the given analytical text. We have also constructed a dataset for this problem. Experiments show that Text2Chart achieves best performances with BERT based encodings with LSTM models in the first stage to label x and y entities, Random Forest classifier for the mapping stage and fastText embedding with LSTM for the chart type prediction. In our experiments, all the stages show satisfactory results and effectiveness considering formation of charts from analytical text, achieving a commendable overall performance.
|
['Swakkhar Shatabda', 'Md. Saddam Hossain Mukta', 'Farhana Meem', 'Tamim Bin Zakir', 'Riyasaat Ahmed Rahul', 'Annysha Huzzat', 'Hasin Kawsar Jahan', 'Md. Mahinur Rashid']
|
2021-04-09
| null | null | null | null |
['type-prediction']
|
['computer-code']
|
[ 2.14530155e-02 2.60994345e-01 -1.26684438e-02 -2.64631808e-01
-6.15372241e-01 -4.85089332e-01 8.72181535e-01 4.63472724e-01
8.52584168e-02 6.93430841e-01 3.42987776e-01 -6.39005780e-01
9.83362943e-02 -9.62920070e-01 -6.33098423e-01 -4.64201003e-01
-2.36661389e-01 4.96735036e-01 -1.93000078e-01 2.88191557e-01
5.80985129e-01 6.18204594e-01 -1.44637346e+00 6.79307699e-01
8.91770661e-01 9.90227640e-01 2.05493346e-01 1.07779849e+00
-1.11360323e+00 1.15408981e+00 -9.09485161e-01 -2.48512179e-01
-1.99230373e-01 -4.99929518e-01 -7.12278366e-01 -1.88870460e-01
1.83019891e-01 -6.09185919e-02 3.03277850e-01 6.89015090e-01
4.41022217e-01 -7.97172859e-02 1.30065560e+00 -1.45984030e+00
-9.51231658e-01 9.33656573e-01 -6.88679755e-01 -2.59488404e-01
3.96555096e-01 -3.14420670e-01 9.67650771e-01 -1.10469270e+00
1.01016605e+00 1.41046572e+00 3.88088048e-01 3.88719589e-01
-1.17589152e+00 -5.26284039e-01 -8.65603462e-02 8.77548754e-02
-1.03847706e+00 -9.29468498e-02 7.72923350e-01 -1.02030444e+00
7.95855761e-01 3.15770119e-01 5.71393013e-01 8.23020577e-01
4.39536989e-01 7.55977511e-01 1.00930297e+00 -7.98583865e-01
4.66014594e-01 3.12175930e-01 3.15809369e-01 7.89017022e-01
-6.85268641e-02 -3.01654160e-01 -5.90179086e-01 3.41550931e-02
7.18084395e-01 -1.65714264e-01 1.52623832e-01 -2.85197139e-01
-1.46221912e+00 7.94305682e-01 6.39530599e-01 4.09920722e-01
-3.67570728e-01 5.14894119e-03 7.27444291e-01 2.71157771e-01
3.30679566e-01 6.51525438e-01 -2.58183386e-02 6.85330927e-02
-9.53829527e-01 2.99595952e-01 7.71725535e-01 1.46150541e+00
5.05957544e-01 2.93644458e-01 -6.06734931e-01 4.04272884e-01
1.96656615e-01 3.94513458e-01 5.63430429e-01 -1.57276809e-01
8.22372794e-01 1.19684839e+00 1.15157373e-01 -1.26864123e+00
-5.94489872e-01 -1.22050829e-01 -9.24069464e-01 5.63820779e-01
2.94080585e-01 -3.33252668e-01 -1.21073544e+00 1.13312757e+00
2.03681797e-01 -5.07593989e-01 4.61709797e-01 5.06255805e-01
1.27841091e+00 1.20186198e+00 2.76721567e-01 9.73641798e-02
1.66358697e+00 -1.01338649e+00 -9.23070431e-01 2.65292078e-01
7.60887444e-01 -7.24180937e-01 1.08599794e+00 1.67879358e-01
-7.43442774e-01 -7.36841559e-01 -1.20725310e+00 -3.22676510e-01
-9.51679111e-01 9.17483211e-01 3.35934818e-01 4.22851205e-01
-1.00042796e+00 3.55959773e-01 -5.16824841e-01 -3.38346452e-01
1.88499689e-01 -9.75032076e-02 -2.99191177e-01 5.97750902e-01
-9.55734372e-01 7.48912632e-01 5.11183083e-01 3.25167120e-01
-3.39671880e-01 -6.20040059e-01 -9.43232477e-01 3.28344196e-01
-1.20961256e-01 -5.59435725e-01 9.22331154e-01 -6.59282267e-01
-1.21006787e+00 6.94167256e-01 -3.59716475e-01 -5.58211744e-01
7.15800822e-01 -1.65106282e-01 -5.85119009e-01 6.06745370e-02
4.17211391e-02 8.50922346e-01 7.95616746e-01 -1.35890794e+00
-6.57221198e-01 -2.70405978e-01 -3.02512378e-01 1.84313785e-02
-3.32241893e-01 -1.03246793e-01 -1.13813229e-01 -6.66926026e-01
-9.53204855e-02 -4.89625275e-01 3.57571468e-02 1.22153372e-01
-1.12839091e+00 -4.71069396e-01 8.67942333e-01 -9.21435535e-01
1.39674318e+00 -1.81047654e+00 -7.35581964e-02 4.30095077e-01
4.53282386e-01 4.86945324e-02 2.26134256e-01 8.20584297e-01
-2.05407798e-01 3.86366636e-01 -1.28336966e-01 -1.39803171e-01
2.03820303e-01 -3.99668664e-01 -6.29918337e-01 -1.92673698e-01
4.77586478e-01 7.55456746e-01 -5.72075367e-01 -9.72346961e-01
1.82371184e-01 4.93126988e-01 7.35479370e-02 5.26987672e-01
-4.53251690e-01 1.25197887e-01 -6.62192822e-01 4.00838643e-01
4.92752701e-01 -3.11086923e-01 7.38906860e-02 -2.29935408e-01
-4.63455111e-01 -1.34952381e-01 -9.70013618e-01 1.38322628e+00
-4.90869313e-01 1.30916715e+00 -5.85001409e-01 -6.62594914e-01
1.83957565e+00 3.19195181e-01 2.75605433e-02 -5.01355529e-01
1.25770211e-01 1.28426045e-01 -2.00619429e-01 -8.87583137e-01
7.09906995e-01 2.18248397e-01 6.12747297e-02 3.64515692e-01
-4.09515901e-03 4.95494194e-02 3.30684930e-01 2.04670429e-01
6.83991671e-01 4.26276654e-01 2.57958561e-01 -1.21699601e-01
5.91513991e-01 3.31750482e-01 -1.79094434e-01 4.40048903e-01
3.87538612e-01 4.41607237e-01 1.09997296e+00 -7.21688807e-01
-1.27225900e+00 -8.07338178e-01 1.16442606e-01 7.66005099e-01
-5.15742540e-01 -4.71282482e-01 -7.88041949e-01 -3.79540116e-01
-8.36045519e-02 1.16499567e+00 -1.09507823e+00 3.04255515e-01
-4.78819191e-01 -2.11929306e-01 3.67301345e-01 4.83833522e-01
2.57806033e-01 -1.52584565e+00 -1.01255465e+00 9.64024290e-03
3.17068219e-01 -5.87796748e-01 -1.18748389e-01 4.35613334e-01
-5.68356395e-01 -9.19988930e-01 -1.06811309e+00 -1.10548019e+00
9.87942398e-01 -3.61748397e-01 8.96814704e-01 -2.67690301e-01
-3.20371091e-01 -4.51139092e-01 -3.77687335e-01 -5.66331923e-01
-5.39983451e-01 2.45449603e-01 -6.06417298e-01 7.46788010e-02
1.92248777e-01 -2.73986638e-01 -3.33359241e-01 -2.15522066e-01
-8.80391300e-01 6.35124207e-01 5.16263902e-01 6.51615441e-01
3.71805519e-01 -2.21964791e-01 6.20282292e-01 -1.05789161e+00
9.59661782e-01 -3.94227713e-01 -4.64583457e-01 4.79028076e-01
-6.00736141e-01 5.42493939e-01 1.17760956e+00 -1.71501845e-01
-1.05400980e+00 1.60030156e-01 1.28335238e-01 -4.34018463e-01
-1.43127516e-01 8.23717117e-01 -3.54965702e-02 6.67202652e-01
8.38956118e-01 1.77205577e-01 -2.63856053e-01 -5.38561046e-01
7.46084809e-01 7.42558539e-01 4.37820792e-01 -3.98575187e-01
6.45592749e-01 3.01843211e-02 7.17745870e-02 -7.72753477e-01
-1.82136610e-01 1.47958741e-01 -1.01179385e+00 -4.11661357e-01
1.14863551e+00 -4.88185376e-01 -7.63470888e-01 4.20963671e-03
-1.46755028e+00 -7.92671219e-02 -1.90673649e-01 2.05111101e-01
-3.01928341e-01 -2.11652070e-01 -1.15905933e-01 -9.70710397e-01
-7.88195014e-01 -9.92872655e-01 8.94518316e-01 4.25550312e-01
-3.76318157e-01 -1.04380465e+00 -1.43550053e-01 -4.30552691e-01
1.26853839e-01 1.08982074e+00 1.39972496e+00 -8.65543127e-01
-3.74038547e-01 -3.76090765e-01 -4.83800799e-01 -2.32624531e-01
1.50706276e-01 7.52453685e-01 -9.52446163e-01 1.49454296e-01
-5.49780786e-01 -2.34782875e-01 7.17573583e-01 3.01193088e-01
1.03578925e+00 -5.96218228e-01 -4.54075664e-01 7.04306543e-01
1.29761708e+00 9.35130060e-01 3.68043840e-01 5.43277085e-01
9.81994748e-01 9.22034860e-01 4.25008416e-01 3.17109078e-01
4.59183186e-01 1.92329198e-01 -4.65287641e-02 -4.25184160e-01
6.47529552e-04 -8.42733324e-01 1.49151251e-01 8.06811750e-01
4.00501281e-01 -4.56194013e-01 -1.13790810e+00 3.88190776e-01
-1.70657063e+00 -8.84264946e-01 -3.77501845e-01 1.88910127e+00
6.11917436e-01 1.96984008e-01 1.82445258e-01 3.40086520e-01
8.83145094e-01 2.41538912e-01 -6.27988994e-01 -1.04992306e+00
6.68558991e-03 -8.20353404e-02 1.39736995e-01 1.55866265e-01
-7.43691266e-01 7.29265511e-01 5.56294537e+00 5.59081197e-01
-1.56897664e+00 -4.76461440e-01 9.94359672e-01 2.50199020e-01
-5.46831429e-01 -5.46990298e-02 -5.87254286e-01 4.09233958e-01
9.36650813e-01 -5.92526674e-01 -2.12139618e-02 8.56609702e-01
2.07244948e-01 1.89189419e-01 -1.26268888e+00 9.21688199e-01
-4.80746031e-02 -1.66599965e+00 6.45573795e-01 -4.98562515e-01
3.45261514e-01 -6.58835411e-01 -2.01729402e-01 1.19289532e-01
2.49144867e-01 -1.30814528e+00 1.03995585e+00 7.58870304e-01
1.13395858e+00 -9.08340335e-01 2.57592559e-01 7.77688101e-02
-1.17322695e+00 1.18135452e-01 -3.25183898e-01 2.37250909e-01
2.41107810e-02 3.36616904e-01 -1.27026641e+00 6.49558961e-01
4.28412855e-01 5.98598599e-01 -8.79010201e-01 8.95274341e-01
-2.23798081e-01 4.83713657e-01 4.98164110e-02 -8.09014261e-01
2.62500763e-01 -2.79552162e-01 3.49131733e-01 1.52256095e+00
7.06047058e-01 -3.15893799e-01 -2.19436914e-01 1.24106514e+00
-5.45659401e-02 6.51843488e-01 -7.67408073e-01 -3.02910239e-01
5.31890213e-01 1.29422033e+00 -1.01680601e+00 -7.44518340e-01
1.07976004e-01 4.75552052e-01 3.01602155e-01 5.43182552e-01
-6.01516187e-01 -1.26729500e+00 -2.95568168e-01 -1.07234694e-01
1.15841180e-01 -5.47375791e-02 -8.70328188e-01 -8.12453389e-01
1.55384853e-01 -4.67459738e-01 3.82325023e-01 -1.41753018e+00
-5.62466025e-01 1.37157989e+00 -1.65344119e-01 -1.46483016e+00
-4.94668216e-01 -7.03003883e-01 -9.47139919e-01 1.11079597e+00
-9.78791058e-01 -1.31587672e+00 -3.45514715e-01 1.67960092e-01
5.52249551e-01 -3.53944719e-01 9.84532654e-01 -2.43148178e-01
-6.45304620e-01 3.14801306e-01 4.62363005e-01 3.73003095e-01
4.91125315e-01 -1.72542202e+00 9.64111447e-01 7.51537919e-01
3.06761712e-01 6.14760518e-01 8.25975835e-01 -6.46672428e-01
-1.31484699e+00 -1.00730360e+00 1.05133808e+00 2.36873496e-02
6.52846992e-01 -8.91158223e-01 -8.06698501e-01 5.39515376e-01
5.09277463e-01 -4.87760335e-01 5.29514432e-01 -6.96540251e-02
-2.72713304e-01 4.13928460e-03 -8.31507981e-01 7.70174026e-01
4.00264144e-01 -3.16909432e-01 -4.59765285e-01 2.52509862e-01
7.57180214e-01 -4.38951582e-01 -8.83386135e-01 -2.38378122e-01
6.31919920e-01 -7.43585408e-01 5.10828972e-01 -5.60045898e-01
1.09883153e+00 -3.90161395e-01 3.97569120e-01 -1.50433517e+00
-4.87217074e-03 -6.93351150e-01 -1.09432535e-02 1.50512266e+00
9.86556470e-01 -3.04239482e-01 6.52693033e-01 5.23973942e-01
-4.16767001e-02 -6.69071317e-01 -4.68376577e-01 -3.17853302e-01
2.22584372e-03 -7.91794881e-02 9.48205471e-01 1.04075181e+00
1.77245677e-01 6.87267005e-01 -3.65195304e-01 -3.83623004e-01
3.32761019e-01 5.00354946e-01 1.04741633e+00 -1.13586760e+00
4.15768176e-01 -5.18008411e-01 -3.08738858e-01 -5.92472732e-01
-1.69991583e-01 -1.11815572e+00 -1.38221413e-01 -2.12949038e+00
-2.62506276e-01 -3.24788481e-01 2.33657174e-02 4.29821312e-01
-6.54152259e-02 -5.62725902e-01 4.74470645e-01 1.63401455e-01
6.41018227e-02 5.10079503e-01 1.26417828e+00 -2.83330917e-01
-5.52274764e-01 -1.86374396e-01 -5.27866900e-01 2.74560153e-01
8.30081582e-01 -2.37088621e-01 -4.73532796e-01 -3.81875515e-01
2.39417166e-01 6.49108529e-01 3.28968763e-02 -1.10756874e+00
1.77039802e-01 -4.43686582e-02 1.09351921e+00 -1.20308185e+00
6.75919233e-03 -6.48148179e-01 9.16040763e-02 3.46428812e-01
-9.56500351e-01 7.29926884e-01 2.83907533e-01 2.15645462e-01
-1.85059592e-01 -1.58010781e-01 4.26976770e-01 2.48587847e-01
-3.22990000e-01 -2.26432607e-01 -3.43242913e-01 -3.29138219e-01
8.17928195e-01 -3.41178447e-01 -5.87027729e-01 -2.13983208e-01
-7.15061903e-01 3.48395199e-01 1.14346959e-01 6.71770275e-01
6.64357245e-01 -1.38207614e+00 -6.95352018e-01 2.38585785e-01
2.09037915e-01 8.32350478e-02 -8.44342750e-04 2.12549746e-01
-1.08791125e+00 5.41284740e-01 -6.83725655e-01 -4.17331040e-01
-1.16486883e+00 6.84064031e-01 -1.79142430e-01 -2.70783991e-01
-8.74606967e-01 4.33861524e-01 1.01004526e-01 -2.21934229e-01
2.22679660e-01 -9.50134277e-01 -8.84216428e-01 4.99293208e-01
6.78525746e-01 3.56910497e-01 -1.53941527e-01 -5.57761729e-01
6.21467382e-02 3.96028459e-01 -2.03300808e-02 -3.59780878e-01
1.22794735e+00 3.93244594e-01 -2.24788748e-02 9.50044632e-01
1.35090625e+00 -1.39458273e-02 -9.78332937e-01 2.15954140e-01
5.29506624e-01 -1.43413514e-01 -3.45498323e-01 -9.23771083e-01
-8.84037673e-01 1.25691903e+00 4.47542161e-01 6.21220350e-01
8.50965440e-01 -4.64233845e-01 4.04224694e-01 4.02665913e-01
-2.93553799e-01 -7.06580102e-01 -3.00540447e-01 1.34084791e-01
1.26846755e+00 -7.55752385e-01 -5.06046414e-02 -3.12825367e-02
-8.39267910e-01 1.77823210e+00 6.42845333e-01 -1.03734657e-01
4.10332471e-01 1.87277094e-01 3.34590256e-01 -3.56830150e-01
-9.01510656e-01 1.94317684e-01 4.94322091e-01 5.98998308e-01
8.37205946e-01 1.73751593e-01 -3.18673611e-01 2.00896651e-01
-8.15442204e-01 -2.49454356e-03 6.60151184e-01 6.77038252e-01
-3.03562522e-01 -6.41793132e-01 -5.07673562e-01 8.11082602e-01
-1.59394473e-01 1.72432400e-02 -6.31439090e-01 8.93873334e-01
-2.67010719e-01 6.21983826e-01 3.78841668e-01 -4.58616525e-01
2.98073620e-01 3.73197943e-01 -1.36629432e-01 -2.35656664e-01
-6.19817078e-01 -7.21366554e-02 2.97841150e-02 5.24546653e-02
1.57213937e-02 -5.05819082e-01 -1.55167949e+00 -1.03123724e-01
8.11322853e-02 5.33646882e-01 1.00127780e+00 4.59081143e-01
4.92078215e-01 7.84974396e-01 6.09088242e-01 -6.44464433e-01
-1.68422498e-02 -1.16834295e+00 -4.05695468e-01 2.51903266e-01
5.82288265e-01 -3.92148048e-01 5.74060977e-02 6.46132886e-01]
|
[11.320622444152832, 2.0398199558258057]
|
5a807448-d3db-46db-a1cb-9733ce90a6cb
|
a-curriculum-domain-adaptation-approach-to
|
1812.09953
| null |
http://arxiv.org/abs/1812.09953v3
|
http://arxiv.org/pdf/1812.09953v3.pdf
|
A Curriculum Domain Adaptation Approach to the Semantic Segmentation of Urban Scenes
|
During the last half decade, convolutional neural networks (CNNs) have
triumphed over semantic segmentation, which is one of the core tasks in many
applications such as autonomous driving and augmented reality. However, to
train CNNs requires a considerable amount of data, which is difficult to
collect and laborious to annotate. Recent advances in computer graphics make it
possible to train CNNs on photo-realistic synthetic imagery with
computer-generated annotations. Despite this, the domain mismatch between the
real images and the synthetic data hinders the models' performance. Hence, we
propose a curriculum-style learning approach to minimizing the domain gap in
urban scene semantic segmentation. The curriculum domain adaptation solves easy
tasks first to infer necessary properties about the target domain; in
particular, the first task is to learn global label distributions over images
and local distributions over landmark superpixels. These are easy to estimate
because images of urban scenes have strong idiosyncrasies (e.g., the size and
spatial relations of buildings, streets, cars, etc.). We then train a
segmentation network, while regularizing its predictions in the target domain
to follow those inferred properties. In experiments, our method outperforms the
baselines on two datasets and two backbone networks. We also report extensive
ablation studies about our approach.
|
['Hassan Foroosh', 'Philip David', 'Yang Zhang', 'Boqing Gong']
|
2018-12-24
| null | null | null | null |
['synthetic-to-real-translation']
|
['computer-vision']
|
[ 4.13802266e-01 2.85687506e-01 -2.51731962e-01 -8.08511972e-01
-6.02316141e-01 -6.49488389e-01 4.90670860e-01 -1.96034297e-01
-5.21501243e-01 7.83998489e-01 -1.28765747e-01 -2.77656704e-01
2.73233682e-01 -8.99386525e-01 -1.12972665e+00 -4.89930451e-01
3.05457801e-01 7.72175550e-01 5.57515919e-01 -1.59297749e-01
7.66549185e-02 3.03641349e-01 -1.60139251e+00 4.37389985e-02
1.18764317e+00 9.57841098e-01 5.76309025e-01 3.24674994e-01
-1.70781374e-01 6.02467418e-01 -4.07449514e-01 -2.72845238e-01
3.14752817e-01 -3.10933083e-01 -8.44732285e-01 5.46861768e-01
6.88809752e-01 -3.04311186e-01 -3.31256777e-01 1.36317813e+00
-8.74726102e-02 1.88178465e-01 6.82125270e-01 -1.32870615e+00
-5.38185954e-01 4.22855794e-01 -6.49239242e-01 -2.63697430e-02
-2.14370027e-01 3.12340051e-01 9.93436694e-01 -5.11450887e-01
6.68384612e-01 1.07139432e+00 5.48722923e-01 3.38003099e-01
-1.23734677e+00 -6.98046923e-01 5.17086327e-01 8.96459222e-02
-1.34125340e+00 -1.96944401e-01 8.65726531e-01 -5.12552738e-01
3.63989681e-01 1.80701762e-02 5.69580197e-01 1.19804800e+00
-2.24188447e-01 9.90513861e-01 1.03479147e+00 -8.31546113e-02
3.47825378e-01 3.15296352e-01 -9.41816345e-02 6.37416303e-01
2.03582495e-01 -7.36014023e-02 -1.40839800e-01 3.19139808e-01
1.03538764e+00 -1.47500753e-01 -7.14078769e-02 -6.04462504e-01
-1.17160845e+00 8.65430713e-01 6.81706131e-01 7.45878294e-02
-5.15153594e-02 2.39291355e-01 2.08658174e-01 -1.55817986e-01
5.27906656e-01 4.46438313e-01 -5.98966718e-01 2.23135203e-01
-8.71937573e-01 3.57374758e-01 4.70830411e-01 1.29183030e+00
1.17583311e+00 -5.03992960e-02 1.60032406e-01 9.05924678e-01
2.24680156e-01 4.51908141e-01 1.99290067e-01 -1.25482929e+00
5.90465188e-01 5.25459170e-01 1.48689955e-01 -9.43627775e-01
-4.01291102e-01 -4.81089264e-01 -9.28821564e-01 2.92726653e-03
8.82726014e-01 -1.82712317e-01 -1.19545507e+00 1.91059875e+00
3.27025533e-01 4.90016699e-01 -1.69707641e-01 1.09149122e+00
5.65071166e-01 5.76525986e-01 3.43727320e-01 3.89080971e-01
1.20810723e+00 -1.15420520e+00 -2.88507760e-01 -9.27975476e-01
6.34712696e-01 -4.74335909e-01 1.28808904e+00 9.34082717e-02
-7.38427937e-01 -7.17137992e-01 -9.31763351e-01 -1.60543367e-01
-3.47623318e-01 1.96550950e-01 7.83454120e-01 4.68257368e-01
-8.17812443e-01 4.67725843e-01 -7.56990314e-01 -3.68668109e-01
8.40459466e-01 1.99229807e-01 -1.55130297e-01 -1.11031979e-01
-1.17212927e+00 6.70888543e-01 6.04165554e-01 -1.19323435e-04
-9.25807357e-01 -7.65701115e-01 -1.09725595e+00 -3.61835659e-02
5.24597526e-01 -6.96145773e-01 1.27028930e+00 -1.50475156e+00
-1.19062483e+00 1.13947797e+00 4.66737151e-02 -4.07993257e-01
6.33383811e-01 -9.35651883e-02 -1.39395103e-01 -1.36735186e-01
4.76689339e-01 1.17671776e+00 6.02074802e-01 -1.46580720e+00
-9.61644232e-01 -2.79068261e-01 2.91608363e-01 2.99475342e-01
1.09068483e-01 -3.93514156e-01 -7.52596915e-01 -4.23133045e-01
1.97703809e-01 -1.12870896e+00 -5.85642159e-01 2.20021397e-01
-6.75164402e-01 -1.45904616e-01 7.16097176e-01 -4.21965361e-01
6.14472449e-01 -2.19922209e+00 -1.40339151e-01 2.36823738e-01
2.48571113e-02 8.19310993e-02 -1.90668985e-01 -1.93353400e-01
-4.05470170e-02 -4.27186228e-02 -6.81954384e-01 -2.93827176e-01
6.64630011e-02 5.73458076e-01 -3.97910088e-01 4.43179578e-01
2.16604173e-01 9.07403648e-01 -1.06291795e+00 -6.06003582e-01
2.41738573e-01 2.47340783e-01 -4.88190114e-01 2.08363324e-01
-8.10258031e-01 8.30852449e-01 -5.18619001e-01 4.27065700e-01
8.09950531e-01 -4.52330351e-01 2.09854484e-01 -7.53570944e-02
2.02043559e-02 3.89739841e-01 -1.13753808e+00 1.95587254e+00
-4.79883909e-01 8.11345339e-01 -3.98664363e-02 -1.32471502e+00
7.98162162e-01 -1.37957111e-01 3.72512907e-01 -8.57052863e-01
9.32406485e-02 1.22639336e-01 -2.00328872e-01 -4.68105465e-01
4.10631686e-01 1.09700691e-02 -2.17854530e-01 3.24784786e-01
-1.66640669e-01 -5.44226289e-01 1.96757987e-01 7.67236948e-02
6.51436090e-01 4.54849839e-01 -2.45019216e-02 -2.81488121e-01
2.69581974e-01 4.45879191e-01 8.57596636e-01 5.01290262e-01
-1.34872735e-01 8.72697532e-01 5.18410087e-01 -6.92854941e-01
-1.23799181e+00 -9.74496841e-01 -1.01835161e-01 9.80897903e-01
6.26210272e-01 7.24136159e-02 -1.05050135e+00 -8.18718553e-01
-1.27928466e-01 7.69641578e-01 -6.09799445e-01 2.48308247e-03
-7.19368041e-01 -6.27437651e-01 3.81353378e-01 7.99275815e-01
9.37774777e-01 -9.96322811e-01 -5.36395848e-01 1.25887021e-01
-4.71257716e-01 -1.54206431e+00 -4.78620261e-01 -1.33441272e-03
-7.11791515e-01 -1.10855293e+00 -4.78767872e-01 -9.84099329e-01
9.23221052e-01 3.54165107e-01 1.44470692e+00 1.24545671e-01
-2.24560440e-01 -8.21778253e-02 1.14284577e-02 -4.40194607e-01
-5.29073142e-02 2.67637014e-01 -4.06422019e-01 2.93228086e-02
3.23484868e-01 -5.15840352e-01 -7.09196985e-01 5.16939938e-01
-8.87313664e-01 5.35521984e-01 4.89898473e-01 6.27291083e-01
8.33974540e-01 1.36529088e-01 2.14906529e-01 -1.41260171e+00
6.54326379e-02 -4.58405674e-01 -8.97235036e-01 1.98494211e-01
-2.05246478e-01 -1.74880717e-02 5.54458082e-01 -3.14441651e-01
-1.20527303e+00 4.81806904e-01 -1.51413873e-01 -1.41392186e-01
-5.73577523e-01 2.74372250e-01 -4.08254534e-01 2.25546822e-01
6.30010188e-01 1.15523785e-01 -3.21896851e-01 -1.53401464e-01
3.56284946e-01 4.41047102e-01 7.19972491e-01 -8.10580730e-01
8.74588847e-01 7.04145074e-01 -2.64011025e-01 -7.93113530e-01
-1.34966218e+00 -3.41910571e-01 -7.62180388e-01 -9.37811062e-02
1.01502264e+00 -1.11609292e+00 -2.55389363e-01 4.55546886e-01
-1.12920392e+00 -8.81574810e-01 -2.60859191e-01 3.00032914e-01
-7.24543154e-01 7.30508268e-02 -2.30221570e-01 -4.03918236e-01
3.29530030e-01 -1.27965844e+00 1.26169848e+00 3.92370462e-01
-1.57528877e-01 -1.07292604e+00 -1.73748881e-01 5.50121069e-01
1.47833228e-01 4.29591328e-01 8.40673149e-01 -2.52032429e-01
-1.05163860e+00 2.21708268e-01 -7.30466008e-01 3.49865496e-01
3.48674469e-02 -9.44093838e-02 -1.12872672e+00 5.92192411e-02
-3.08704227e-01 -3.79424751e-01 8.40182722e-01 4.93394881e-01
1.70850956e+00 3.53323109e-02 -4.89602119e-01 9.28345203e-01
1.36237323e+00 2.36040577e-02 7.31743097e-01 3.78927737e-01
8.86122525e-01 8.70735645e-01 8.44205737e-01 1.80760063e-02
6.82514548e-01 5.63800931e-01 6.43466175e-01 -3.85857195e-01
-2.09919348e-01 -4.42210108e-01 -1.43936515e-01 3.12357664e-01
1.72391117e-01 -1.34021387e-01 -1.00145912e+00 8.45564187e-01
-1.96320009e+00 -5.74944198e-01 -2.01198861e-01 1.88788640e+00
8.61795425e-01 3.18361253e-01 -8.58804211e-02 -4.27782565e-01
6.93860173e-01 1.88613266e-01 -9.29258704e-01 4.51439321e-02
-7.10722283e-02 3.93897817e-02 8.02345216e-01 5.03921092e-01
-1.38974881e+00 1.37476707e+00 5.87183285e+00 7.95512736e-01
-1.13709950e+00 2.39942893e-02 1.09757257e+00 3.34987491e-01
-4.28273827e-01 5.74912913e-02 -7.37328231e-01 5.77368021e-01
5.61452687e-01 2.74631172e-01 3.00621569e-01 1.06369257e+00
1.31918222e-01 -2.77384400e-01 -1.07173443e+00 7.73239911e-01
-2.24400252e-01 -1.33558989e+00 -1.55780077e-01 -2.06283070e-02
1.09842992e+00 3.16283047e-01 1.02778926e-01 2.52876371e-01
7.67571747e-01 -1.12233877e+00 7.47912049e-01 1.56428039e-01
9.18203115e-01 -6.47409678e-01 5.59878290e-01 6.11512303e-01
-1.07946515e+00 1.24081023e-01 -5.02104044e-01 4.11329307e-02
7.39953741e-02 6.23307049e-01 -7.73078561e-01 8.51723775e-02
7.04732418e-01 7.24677384e-01 -4.43980992e-01 8.80404115e-01
-6.24656975e-01 5.34364760e-01 -3.56899828e-01 2.66159117e-01
5.80926776e-01 -5.12352526e-01 2.27837916e-02 1.04742956e+00
9.59701538e-02 1.04404136e-03 3.73607099e-01 1.30167341e+00
-2.66910315e-01 -2.91202784e-01 -6.23267233e-01 1.84525073e-01
3.85027140e-01 1.24547160e+00 -1.08884633e+00 -3.12762052e-01
-4.18763220e-01 8.30994964e-01 3.97464484e-01 6.70760572e-01
-1.04816592e+00 -1.59288794e-01 9.31716204e-01 2.45476678e-01
1.67190731e-01 -3.16097796e-01 -6.79964662e-01 -1.08261406e+00
-4.40365858e-02 -7.14656293e-01 1.52089270e-02 -8.17037702e-01
-1.12182701e+00 4.63382214e-01 -1.58708841e-02 -1.11924410e+00
-1.50680050e-01 -5.49144864e-01 -4.98109311e-01 8.47816467e-01
-1.83862734e+00 -1.19268262e+00 -5.21148443e-01 4.90236968e-01
7.53032565e-01 2.36137554e-01 5.30965149e-01 3.00031632e-01
-5.26093364e-01 2.41117716e-01 -2.54716501e-02 5.61579585e-01
6.10030830e-01 -1.30851674e+00 7.80191362e-01 8.25773835e-01
1.47910535e-01 2.49350563e-01 5.91625571e-01 -4.99240667e-01
-6.46755159e-01 -1.55435431e+00 5.89938521e-01 -3.87687653e-01
5.42825699e-01 -6.08047903e-01 -9.26151574e-01 9.24442589e-01
8.24890435e-02 1.80935040e-01 3.32010597e-01 6.81061298e-02
-4.30019587e-01 -1.52802691e-01 -8.96799445e-01 6.23178840e-01
1.21895099e+00 -5.04295647e-01 -2.33410031e-01 5.73085368e-01
7.37662911e-01 -7.39256024e-01 -4.01574105e-01 3.41582149e-01
2.47725472e-01 -1.04279101e+00 1.04997027e+00 -4.95692551e-01
5.44317067e-01 -4.52364743e-01 -4.59650345e-02 -1.45570123e+00
-7.34511167e-02 -8.05784240e-02 5.06518126e-01 9.85823870e-01
5.45861602e-01 -4.43087012e-01 1.22844839e+00 8.38956833e-01
-2.50151604e-01 -5.02654433e-01 -5.39093554e-01 -6.49788260e-01
6.93725646e-02 -4.88916993e-01 7.21610427e-01 1.07651281e+00
-7.22888350e-01 3.19034815e-01 -3.17133032e-02 4.18520153e-01
6.58674061e-01 3.14177722e-01 9.56478775e-01 -1.28793490e+00
-1.38302281e-01 -3.45292032e-01 -2.04968929e-01 -1.44694090e+00
4.72848088e-01 -5.89079142e-01 3.41108829e-01 -1.51657975e+00
1.03884444e-01 -9.74693060e-01 8.96905586e-02 3.51190746e-01
-1.78239122e-01 3.50144029e-01 -1.03412971e-01 8.93535987e-02
-7.23207414e-01 4.15496379e-01 1.57669735e+00 -4.54434603e-01
-4.36028652e-02 7.41111711e-02 -7.21607685e-01 1.02414560e+00
9.55016792e-01 -4.33048666e-01 -7.62694716e-01 -8.41302574e-01
2.92072654e-01 -2.19764635e-01 5.38561702e-01 -9.76145625e-01
1.83584869e-01 -4.90921915e-01 3.54947776e-01 -5.36928296e-01
2.74623632e-01 -8.38225901e-01 2.75344159e-02 1.15895569e-01
-3.25657457e-01 -3.96261573e-01 1.55787989e-01 5.57554126e-01
-3.32352757e-01 -1.42813087e-01 9.38227654e-01 -3.83145273e-01
-1.13875282e+00 5.42227030e-01 -1.14857256e-01 4.04773384e-01
1.13454866e+00 -2.98254251e-01 -2.04332799e-01 -3.44425738e-01
-5.20873308e-01 4.79766160e-01 6.58770144e-01 4.94745791e-01
3.62607330e-01 -1.06634736e+00 -5.13533294e-01 1.98369935e-01
2.27096975e-01 9.74488556e-01 2.20873028e-01 4.99657989e-01
-7.68945813e-01 1.68180346e-01 -2.52659470e-01 -8.74760270e-01
-8.97748053e-01 3.46061081e-01 4.20511335e-01 -1.52209029e-01
-5.08378744e-01 9.52344298e-01 8.11099887e-01 -7.35048413e-01
2.39467263e-01 -4.58679616e-01 -3.62808518e-02 -2.84472764e-01
3.11287493e-01 -9.29304138e-02 -1.62366837e-01 -5.16732335e-01
-7.14188069e-02 5.76592565e-01 -4.61741798e-02 1.01656228e-01
1.32814074e+00 -2.66741037e-01 5.49776815e-02 2.20673114e-01
1.04551387e+00 -5.26004255e-01 -1.92845476e+00 -4.60692018e-01
6.78825974e-02 -5.18894374e-01 1.63587406e-02 -6.23370409e-01
-1.31535304e+00 1.01988506e+00 2.05508441e-01 1.16049107e-02
9.80946422e-01 1.32272705e-01 1.00629818e+00 1.96404040e-01
4.37804997e-01 -1.40152037e+00 -2.30210088e-02 5.33209383e-01
4.03860718e-01 -1.53462720e+00 -2.56946623e-01 -7.85169542e-01
-6.99490070e-01 8.14310193e-01 9.59481359e-01 -1.84524208e-01
5.47799826e-01 6.36776090e-02 2.01354638e-01 -1.76453978e-01
-3.15090537e-01 -4.52261955e-01 1.54658467e-01 8.52980494e-01
1.77698821e-01 2.24726796e-01 2.67812431e-01 2.18258828e-01
-2.78605103e-01 -2.24454120e-01 3.43279779e-01 5.60138762e-01
-5.17439425e-01 -9.88563240e-01 -2.93599337e-01 2.75955737e-01
-3.45233269e-02 7.52732679e-02 -2.75422394e-01 9.63219941e-01
5.44791341e-01 7.38064826e-01 3.94091398e-01 -9.70232338e-02
2.84867674e-01 -3.17386359e-01 2.90225178e-01 -7.63060391e-01
-2.05566338e-03 -1.33563012e-01 7.19598904e-02 -6.59794927e-01
-5.77262104e-01 -6.88772500e-01 -1.35671508e+00 -1.33590311e-01
-6.86974823e-02 -1.06733434e-01 8.06082308e-01 1.15645051e+00
1.31315455e-01 5.77710092e-01 4.06927586e-01 -8.08968902e-01
-1.91206932e-01 -6.88126504e-01 -6.28828883e-01 6.49042785e-01
1.36781141e-01 -6.92308426e-01 4.71068136e-02 3.72410864e-01]
|
[9.641218185424805, 0.7312387228012085]
|
b69d4f4a-12ce-403a-8eaf-04e11b48b9e8
|
stemm-self-learning-with-speech-text-manifold
|
2203.10426
| null |
https://arxiv.org/abs/2203.10426v1
|
https://arxiv.org/pdf/2203.10426v1.pdf
|
STEMM: Self-learning with Speech-text Manifold Mixup for Speech Translation
|
How to learn a better speech representation for end-to-end speech-to-text translation (ST) with limited labeled data? Existing techniques often attempt to transfer powerful machine translation (MT) capabilities to ST, but neglect the representation discrepancy across modalities. In this paper, we propose the Speech-TExt Manifold Mixup (STEMM) method to calibrate such discrepancy. Specifically, we mix up the representation sequences of different modalities, and take both unimodal speech sequences and multimodal mixed sequences as input to the translation model in parallel, and regularize their output predictions with a self-learning framework. Experiments on MuST-C speech translation benchmark and further analysis show that our method effectively alleviates the cross-modal representation discrepancy, and achieves significant improvements over a strong baseline on eight translation directions.
|
['Mingxuan Wang', 'Yang Feng', 'Lei LI', 'Rong Ye', 'Qingkai Fang']
|
2022-03-20
| null |
https://aclanthology.org/2022.acl-long.486
|
https://aclanthology.org/2022.acl-long.486.pdf
|
acl-2022-5
|
['speech-to-text-translation']
|
['natural-language-processing']
|
[ 5.79823852e-01 1.23317994e-01 -5.51403821e-01 -5.62367141e-01
-1.54448736e+00 -6.74055457e-01 1.04405057e+00 -7.26373553e-01
1.22928591e-02 5.47703445e-01 8.84013295e-01 -7.95009017e-01
7.69186556e-01 -1.08816341e-01 -7.51307189e-01 -5.36604047e-01
7.37256229e-01 8.75193000e-01 -3.18890423e-01 -3.89809459e-01
-4.21113074e-01 -1.88516349e-01 -9.20584321e-01 8.37616146e-01
1.02588773e+00 5.37751079e-01 1.58644229e-01 4.84190166e-01
-5.17322719e-01 7.17898071e-01 -3.85680109e-01 -7.19586551e-01
2.09989935e-01 -1.04527164e+00 -1.03723538e+00 3.77512157e-01
4.01306629e-01 -2.06565291e-01 -6.69296563e-01 9.09790993e-01
7.27431417e-01 1.28055200e-01 8.49894702e-01 -1.08915889e+00
-1.09862006e+00 9.51116383e-01 -5.58198094e-01 -7.61494488e-02
3.52843106e-01 1.11711547e-01 8.79881859e-01 -1.47330642e+00
7.49263585e-01 1.53700638e+00 1.43607944e-01 9.44700122e-01
-1.46348822e+00 -5.01186788e-01 3.55494544e-02 -1.77017450e-02
-1.00647557e+00 -1.08789229e+00 6.24366045e-01 -2.75918782e-01
9.75908101e-01 3.89210165e-01 5.13892286e-02 1.86214983e+00
-2.33604461e-01 1.23979151e+00 1.08188593e+00 -6.03081882e-01
-1.09812155e-01 1.33626670e-01 -3.13711405e-01 3.81289005e-01
-5.07886589e-01 4.90085408e-03 -8.16445231e-01 -1.69065185e-02
4.57691848e-01 -1.72390774e-01 -1.82400778e-01 -1.41198397e-01
-1.81685257e+00 6.63082480e-01 3.16445231e-02 3.46632957e-01
-1.04472592e-01 -1.19915798e-01 5.86850047e-01 7.51750112e-01
7.11586118e-01 -1.38785318e-03 -3.73800218e-01 -3.16704333e-01
-8.60571563e-01 -3.21456373e-01 5.71181417e-01 1.26322937e+00
6.86758816e-01 2.33913377e-01 -4.50504005e-01 1.28615057e+00
2.63514847e-01 1.15934026e+00 8.59477639e-01 -6.33188963e-01
1.35550153e+00 2.91191250e-01 -3.21367830e-01 -2.93326914e-01
2.12773271e-02 -2.41878703e-01 -1.06950915e+00 -4.73553956e-01
8.20084959e-02 -4.35750902e-01 -1.17389011e+00 1.88039804e+00
1.66569903e-01 -1.08461060e-01 4.97030258e-01 1.02271879e+00
8.44222844e-01 1.08159089e+00 -1.26928883e-02 -3.11126173e-01
1.00907207e+00 -1.40231025e+00 -9.26824689e-01 -4.75751489e-01
8.31998825e-01 -1.11353958e+00 1.37594485e+00 -2.90916085e-01
-1.17805862e+00 -4.80355650e-01 -6.76279545e-01 -2.67423540e-01
-3.94689552e-02 5.84902108e-01 1.74333259e-01 2.99129397e-01
-1.02609897e+00 2.72006124e-01 -9.80336070e-01 -4.58865315e-01
9.66251418e-02 2.08276600e-01 -4.74340528e-01 -2.48452947e-01
-1.22977042e+00 1.11419022e+00 1.49863690e-01 -4.66391444e-02
-8.49682152e-01 -3.14797759e-01 -9.28877652e-01 -1.48418978e-01
1.22222841e-01 -1.05965877e+00 1.40818846e+00 -1.45502830e+00
-2.08988619e+00 8.61623764e-01 -6.86232090e-01 -7.55929798e-02
5.16446173e-01 7.09920947e-04 -5.84786713e-01 -3.71986553e-02
6.77964464e-02 8.78375828e-01 1.09761202e+00 -1.27400899e+00
-2.84260869e-01 -1.74780250e-01 -4.15329218e-01 6.12160981e-01
-4.71683204e-01 2.88051367e-01 -5.09973705e-01 -8.07263553e-01
3.09556156e-01 -1.15289629e+00 -9.79076400e-02 -4.81241077e-01
-5.24929702e-01 -6.27544373e-02 4.96038973e-01 -8.30942929e-01
9.91566300e-01 -2.07326746e+00 9.63531852e-01 -2.23171756e-01
-2.16195583e-01 1.16536662e-01 -6.83372617e-01 7.14537263e-01
-1.50728896e-01 -1.68501616e-01 -3.77205461e-01 -1.03602481e+00
2.46040538e-01 4.19989288e-01 -6.72368348e-01 2.32912943e-01
2.27193505e-01 1.26935315e+00 -7.11650014e-01 -5.44610620e-01
4.48422432e-02 4.88292068e-01 -2.21016362e-01 3.82490247e-01
-2.43429422e-01 7.65273929e-01 -1.31466508e-01 6.81883991e-01
4.92922693e-01 -3.36626828e-01 5.29780865e-01 -9.79905501e-02
3.07690620e-01 8.01155090e-01 -4.20863003e-01 2.13469934e+00
-6.22758210e-01 6.78434312e-01 -6.29841760e-02 -8.45115840e-01
8.27311516e-01 6.64409041e-01 3.38858575e-01 -7.51502752e-01
7.44712353e-02 4.44444656e-01 -5.82956858e-02 -4.48227406e-01
3.75889361e-01 -6.09283268e-01 -2.20280159e-02 7.02211380e-01
4.13910717e-01 -1.13625430e-01 -3.02508682e-01 1.78515986e-01
7.30745077e-01 3.25755060e-01 -1.15937352e-01 2.12221533e-01
4.14950609e-01 2.71106865e-02 3.75738919e-01 2.00509623e-01
-3.22583306e-04 8.89878929e-01 2.19832212e-01 -2.82273795e-02
-1.10216916e+00 -1.17943263e+00 3.54241669e-01 1.53437543e+00
-7.90561512e-02 -4.06121820e-01 -7.89551437e-01 -1.02509332e+00
-2.94420093e-01 7.42035031e-01 -2.89336830e-01 -2.73702532e-01
-6.66637242e-01 -6.32677734e-01 6.70823753e-01 5.04826665e-01
2.10781813e-01 -6.58903182e-01 5.75615108e-01 1.81971490e-02
-1.04476535e+00 -1.27471721e+00 -9.41561520e-01 1.80502981e-02
-9.90124226e-01 -1.73901558e-01 -1.15869927e+00 -1.04093075e+00
6.60194516e-01 4.77156937e-01 1.07197225e+00 -4.11400020e-01
6.42635226e-01 1.34676054e-01 -3.88401240e-01 1.70992538e-01
-1.07714272e+00 3.99112463e-01 2.27884501e-01 1.65785223e-01
1.79641560e-01 -3.86545032e-01 -2.03149512e-01 4.77493197e-01
-6.23958945e-01 5.06692529e-01 7.88289189e-01 1.16985822e+00
4.61897045e-01 -9.37864780e-01 5.95666051e-01 -4.59645361e-01
5.40896595e-01 -5.88724673e-01 -6.13854602e-02 4.98611450e-01
-4.19310510e-01 1.06300689e-01 5.46403110e-01 -7.43036091e-01
-1.08267629e+00 2.21415922e-01 -1.21647380e-01 -8.82090211e-01
6.37278557e-02 6.02413774e-01 -4.84057128e-01 2.89806634e-01
5.60120881e-01 7.18983471e-01 3.49843293e-01 -4.98086035e-01
8.79033864e-01 1.04040623e+00 4.28804100e-01 -7.97596157e-01
9.39224422e-01 1.38146669e-01 -3.32674086e-01 -4.57011700e-01
-6.19623780e-01 -2.49511525e-01 -6.62756741e-01 7.21672773e-02
6.95215225e-01 -1.16216028e+00 3.16860490e-02 8.89262408e-02
-1.35090542e+00 -4.07689810e-01 -9.12506506e-02 7.12879956e-01
-7.65678763e-01 4.04003620e-01 -8.48557174e-01 -5.87514758e-01
-3.56631339e-01 -1.27508509e+00 1.50051582e+00 -3.79287750e-01
-1.82675913e-01 -9.35303390e-01 2.16497898e-01 8.75545382e-01
4.71985817e-01 -4.51894939e-01 7.87504613e-01 -5.89627445e-01
-4.52823162e-01 1.64263532e-01 -1.24568418e-01 4.66379970e-01
3.71980667e-01 -3.61378312e-01 -9.84696865e-01 -3.70863616e-01
-1.37967676e-01 -4.13651854e-01 9.35940385e-01 -4.66061644e-02
5.31450093e-01 -5.57406306e-01 -1.95855543e-01 6.35134697e-01
6.48377061e-01 -2.51273721e-01 5.12432516e-01 -1.30974740e-01
9.17541146e-01 7.13592291e-01 3.77255499e-01 -1.17678143e-01
6.60332322e-01 8.95240426e-01 6.95433468e-02 -3.01561654e-01
-5.30197918e-01 -6.12830460e-01 1.10147917e+00 1.76636958e+00
2.20210269e-01 -3.58413219e-01 -8.84091556e-01 4.15566266e-01
-1.89839149e+00 -8.57718110e-01 2.22116947e-01 1.87672842e+00
1.13856506e+00 -3.23093325e-01 3.04886878e-01 -4.25123304e-01
7.57913649e-01 1.88773423e-01 -3.80991042e-01 -1.27226979e-01
-4.91207212e-01 -1.45959094e-01 2.71819770e-01 6.25876546e-01
-8.44659448e-01 1.39434206e+00 7.04281616e+00 1.02501607e+00
-1.26284313e+00 6.01392508e-01 5.35886705e-01 -1.68093771e-01
-7.61683226e-01 1.39153004e-01 -4.45151329e-01 4.47713286e-01
1.33533037e+00 -1.08712517e-01 7.19753921e-01 5.05587935e-01
7.74741620e-02 7.74109900e-01 -1.26112318e+00 1.11084259e+00
3.30763251e-01 -1.33794284e+00 4.25779998e-01 6.38261586e-02
9.55511272e-01 4.48868752e-01 3.24598014e-01 5.47591686e-01
3.55811894e-01 -9.60297763e-01 7.28072464e-01 9.81765836e-02
1.19934618e+00 -2.96097457e-01 4.74456698e-01 4.63305652e-01
-9.03823018e-01 2.46549651e-01 -1.79532051e-01 2.19932228e-01
3.82161736e-01 1.94706246e-02 -1.07895243e+00 7.48974442e-01
-1.09523550e-01 7.98032761e-01 -2.52602398e-01 1.80497497e-01
-1.64995670e-01 8.86065185e-01 -1.09347120e-01 8.12992826e-02
2.92371273e-01 -3.44234943e-01 6.62539601e-01 1.37359846e+00
5.02117634e-01 -2.48590142e-01 2.06619889e-01 8.13610673e-01
-4.23990726e-01 2.64638424e-01 -6.72755301e-01 -5.19100666e-01
4.17448997e-01 8.53983402e-01 -1.39131337e-01 -6.40169919e-01
-6.54822469e-01 1.59420168e+00 3.30264330e-01 8.45467389e-01
-5.84334850e-01 2.09871814e-01 8.36927414e-01 -1.29861832e-01
8.00028294e-02 -3.58946741e-01 -3.99623811e-01 -1.86053658e+00
1.01344585e-01 -1.35475326e+00 1.94601819e-01 -7.56463230e-01
-1.51637995e+00 7.80962825e-01 -3.02490801e-01 -1.67752898e+00
-6.47516847e-01 -3.46586913e-01 -3.36477280e-01 1.10650849e+00
-1.28530061e+00 -1.80299473e+00 4.70987231e-01 7.14608014e-01
1.00037766e+00 -6.14178717e-01 8.16201210e-01 4.09158140e-01
-4.92486477e-01 9.56417263e-01 3.06910992e-01 2.31326073e-01
9.94826972e-01 -9.82828140e-01 7.25298226e-01 8.83166790e-01
5.12971699e-01 6.60998702e-01 4.76483256e-01 -5.51224709e-01
-1.87437820e+00 -1.07129502e+00 1.19141364e+00 -8.02723527e-01
7.45346248e-01 -5.33479691e-01 -8.39546025e-01 9.83916163e-01
6.41390979e-01 -1.01454772e-01 7.46378124e-01 1.22884944e-01
-7.30304539e-01 1.87156543e-01 -6.15892887e-01 8.82597327e-01
1.10460007e+00 -1.20880580e+00 -7.52943873e-01 5.38729906e-01
1.15085006e+00 -5.67228973e-01 -7.24333823e-01 2.78080821e-01
3.06403458e-01 -2.89764076e-01 8.39574814e-01 -1.06037509e+00
6.79335952e-01 -2.24780127e-01 -6.34467304e-01 -1.72039223e+00
-1.24913417e-01 -1.07835734e+00 -1.64957255e-01 1.32888186e+00
8.70473087e-01 -5.37483037e-01 4.86960977e-01 1.87006146e-01
-5.25421977e-01 -5.62542081e-01 -1.24563193e+00 -8.71285617e-01
5.51315486e-01 -2.35962674e-01 7.80364394e-01 1.29477954e+00
5.61328053e-01 1.03528595e+00 -7.95251846e-01 4.46422286e-02
3.63688618e-01 2.18938917e-01 9.30778861e-01 -4.43780810e-01
-5.55396676e-01 -6.24544501e-01 -4.62964885e-02 -1.72175288e+00
6.09921098e-01 -1.53047788e+00 8.13138857e-02 -1.49660039e+00
3.45855564e-01 2.79220492e-02 -6.27980679e-02 5.47783136e-01
-3.53668779e-01 -1.46099483e-03 2.96187490e-01 6.43279135e-01
-4.60705668e-01 1.11939096e+00 1.48177397e+00 -3.84945720e-01
3.16851251e-02 -1.10848643e-01 -5.50587058e-01 1.78314880e-01
4.52851385e-01 -2.78965384e-01 -4.35285091e-01 -1.05367625e+00
-9.67870876e-02 4.94439751e-01 -1.30138025e-01 -2.91359305e-01
1.47014588e-01 -3.82573694e-01 1.77007049e-01 -5.77928185e-01
6.75116897e-01 -6.45042300e-01 2.26431675e-02 -8.42935964e-02
-7.24000156e-01 4.35507074e-02 -3.33549902e-02 3.25818688e-01
-4.71662909e-01 2.87926584e-01 4.82201576e-01 1.70205772e-01
-9.96386334e-02 3.01105201e-01 -2.78704882e-01 4.09425721e-02
3.82174283e-01 2.54819334e-01 -6.26912951e-01 -7.47629166e-01
-7.60343194e-01 2.94072211e-01 4.92694527e-01 8.26102078e-01
5.97911477e-01 -2.02281713e+00 -1.10469460e+00 2.73688436e-01
3.05152833e-01 -5.27101636e-01 1.86577793e-02 1.13006973e+00
1.63190722e-01 3.19217354e-01 1.03461638e-01 -8.09146047e-01
-1.10220659e+00 4.58731741e-01 4.10462201e-01 1.06289983e-01
-5.30555725e-01 6.92879379e-01 2.07910299e-01 -1.05085301e+00
1.38420627e-01 -7.45477453e-02 3.30306560e-01 -1.10999703e-01
3.74557883e-01 1.41631201e-01 -2.23917365e-02 -1.15560400e+00
-2.54112035e-01 4.29155827e-01 -4.59885299e-02 -9.13236737e-01
1.00549662e+00 -6.77978992e-01 2.16762088e-02 7.12040901e-01
1.36808896e+00 -5.14214560e-02 -1.02270555e+00 -8.15822124e-01
-3.79778668e-02 -4.62578952e-01 -2.74144679e-01 -8.67871881e-01
-9.67716515e-01 1.19490552e+00 2.94570208e-01 -2.67450809e-01
8.68589580e-01 3.30387354e-01 1.11356127e+00 4.65080172e-01
6.19631000e-02 -1.04038501e+00 1.99643448e-01 8.19334626e-01
1.03792596e+00 -1.48409784e+00 -5.69591999e-01 -3.45352650e-01
-1.28357577e+00 9.38471973e-01 3.81157249e-01 5.15026927e-01
1.46596402e-01 7.72766620e-02 6.15687251e-01 2.78798670e-01
-1.04347730e+00 -2.33175948e-01 5.32695115e-01 6.57548070e-01
7.95628667e-01 3.54081213e-01 -4.49520070e-03 4.48324025e-01
-1.57440111e-01 -2.66411990e-01 6.28013238e-02 6.36381328e-01
-1.37201503e-01 -1.37092102e+00 -3.97067636e-01 -6.24848716e-02
-1.28252834e-01 -3.70329559e-01 -8.32188249e-01 2.05646127e-01
-4.03724253e-01 1.14658642e+00 -1.75004259e-01 -8.22348714e-01
2.36570746e-01 5.61226308e-01 4.46144193e-01 -5.65548301e-01
-4.19968694e-01 5.93863845e-01 2.79660434e-01 -4.04702276e-01
-1.59218147e-01 -7.37129033e-01 -1.09939742e+00 -2.52693415e-01
-1.66572601e-01 9.49183130e-04 7.47731328e-01 1.07925701e+00
6.55352235e-01 2.77034998e-01 8.88695061e-01 -8.80964637e-01
-1.08093762e+00 -1.32556081e+00 3.61300111e-02 5.04031181e-01
5.87359428e-01 -3.66828978e-01 -4.48600680e-01 2.74152845e-01]
|
[14.504891395568848, 7.225879192352295]
|
9a1b0252-e5b6-4bcc-b24f-6300195bac29
|
end-to-end-neural-bridging-resolution
| null | null |
https://aclanthology.org/2022.coling-1.64
|
https://aclanthology.org/2022.coling-1.64.pdf
|
End-to-End Neural Bridging Resolution
|
The state of bridging resolution research is rather unsatisfactory: not only are state-of-the-art resolvers evaluated in unrealistic settings, but the neural models underlying these resolvers are weaker than those used for entity coreference resolution. In light of these problems, we evaluate bridging resolvers in an end-to-end setting, strengthen them with better encoders, and attempt to gain a better understanding of them via perturbation experiments and a manual analysis of their outputs.
|
['Vincent Ng', 'Yufang Hou', 'Hideo Kobayashi']
| null | null | null | null |
coling-2022-10
|
['coreference-resolution']
|
['natural-language-processing']
|
[ 2.50816911e-01 7.47154236e-01 -5.79000354e-01 -4.18965250e-01
-1.11086893e+00 -5.96203387e-01 4.82653379e-01 2.45257184e-01
-6.25435174e-01 1.19246256e+00 8.26192677e-01 -4.12872881e-01
-2.79407889e-01 -6.41187072e-01 -9.14256573e-01 3.24131250e-02
8.28515664e-02 1.10918427e+00 2.52842963e-01 -6.39059067e-01
1.81923911e-01 1.09303117e-01 -1.12405169e+00 6.07615173e-01
7.65877724e-01 2.58933634e-01 -2.94398338e-01 4.12864447e-01
-1.91004127e-02 9.53263223e-01 -7.16892779e-01 -9.50442374e-01
-6.38944358e-02 -3.30863178e-01 -1.34950840e+00 -9.43907499e-01
4.71092969e-01 -1.98295847e-01 -6.37460709e-01 1.16788208e+00
7.14790821e-01 3.79720926e-02 1.79130435e-01 -8.89410198e-01
-1.07843637e+00 1.40525568e+00 -3.67384881e-01 7.29551733e-01
4.56878752e-01 -8.44660029e-02 1.41594899e+00 -5.63763320e-01
1.12615418e+00 1.45123041e+00 8.82382631e-01 7.40430772e-01
-1.49317169e+00 -7.58963287e-01 3.43743600e-02 4.23734188e-01
-1.28896630e+00 -9.10133064e-01 4.65219647e-01 1.27458319e-01
1.48893595e+00 1.84815407e-01 1.34312674e-01 1.63705540e+00
-1.58207223e-01 5.50543010e-01 5.54575205e-01 -3.81342083e-01
-1.25816092e-01 -2.83883274e-01 2.87760943e-01 4.51742083e-01
4.58982825e-01 4.15499449e-01 -6.76995933e-01 -2.18155265e-01
5.73970139e-01 -7.52080142e-01 -6.75794899e-01 -2.18298480e-01
-9.73018408e-01 1.01549852e+00 5.54093063e-01 4.37912136e-01
-1.26169920e-01 -2.20615659e-02 5.04675627e-01 4.65871900e-01
2.25651935e-01 9.98403728e-01 -6.57192707e-01 -2.19030142e-01
-8.91392887e-01 4.84040111e-01 1.04409361e+00 8.01491261e-01
4.22275305e-01 -4.05960649e-01 -1.28840478e-02 6.67512119e-01
3.15170251e-02 -5.34584187e-02 3.89895678e-01 -1.17583430e+00
9.20405149e-01 3.98631364e-01 2.44067833e-01 -7.55095482e-01
-7.22479641e-01 -4.69597816e-01 -3.90443563e-01 -1.08142219e-01
6.46026552e-01 -3.46349955e-01 -4.93451983e-01 2.40649152e+00
-1.74764037e-01 1.72523513e-01 3.87237191e-01 8.15405667e-01
9.11185086e-01 1.68696478e-01 2.10919097e-01 -1.39107049e-01
1.17421126e+00 -7.39683330e-01 -8.91062319e-01 -5.34957588e-01
7.81787217e-01 -3.61938477e-01 8.38113010e-01 4.82696630e-02
-1.56408656e+00 -2.69579083e-01 -1.14470327e+00 -3.36540312e-01
-1.64091632e-01 -2.61079282e-01 7.42927134e-01 1.88833401e-01
-9.38107491e-01 8.11165333e-01 -8.14019918e-01 -5.18509567e-01
2.06495240e-01 3.90989810e-01 -7.07587242e-01 1.65924475e-01
-2.02953625e+00 1.72484112e+00 5.54568648e-01 5.23057170e-02
-4.05310512e-01 -9.18078959e-01 -1.05711997e+00 2.21853748e-01
5.10966301e-01 -8.38205814e-01 1.50036156e+00 -5.51971853e-01
-9.76005852e-01 1.07752562e+00 -8.64013657e-02 -6.34969771e-01
3.80305916e-01 -3.70498180e-01 -6.52767599e-01 -5.01441620e-02
5.69332354e-02 5.28397620e-01 -9.44978967e-02 -1.15385056e+00
-5.51184714e-01 -2.28588834e-01 3.49544764e-01 2.09795669e-01
3.02398294e-01 9.38227475e-02 -1.20118864e-01 -3.35819125e-01
-2.11355150e-01 -7.84169436e-01 -1.19474046e-01 -6.09113038e-01
-4.97410387e-01 -2.73405612e-01 1.95005476e-01 -5.69396555e-01
1.31399691e+00 -1.87261963e+00 4.85902160e-01 -2.49754131e-01
1.87233135e-01 5.09754241e-01 -3.10139149e-01 5.15704691e-01
-5.22813261e-01 4.62308675e-01 -3.62925902e-02 -3.34482938e-01
2.07604587e-01 1.43516317e-01 -3.76299292e-01 2.54433304e-01
3.47374797e-01 9.84341443e-01 -9.23212051e-01 -4.55868751e-01
-2.53123909e-01 4.09876436e-01 -8.10319006e-01 5.29188178e-02
-1.67745829e-01 6.13346184e-03 -2.79902518e-01 2.83578247e-01
3.82434607e-01 -9.67230201e-02 5.37224829e-01 -4.26420897e-01
-1.59998015e-01 1.02423048e+00 -9.29237664e-01 2.15648031e+00
-2.90682968e-02 6.29029214e-01 1.58109739e-01 -1.03248918e+00
6.13403738e-01 5.03846228e-01 1.98723361e-01 -8.85516465e-01
7.35422298e-02 4.50160056e-02 3.59246820e-01 -5.98974943e-01
5.75715482e-01 -4.48377997e-01 -2.26064101e-01 4.17472601e-01
1.58979669e-01 4.39057201e-01 2.04669043e-01 3.55966479e-01
1.43988526e+00 1.86294436e-01 1.90155078e-02 -3.29288282e-02
3.79999094e-02 1.07586406e-01 8.96832466e-01 8.69698524e-01
-2.35370204e-01 6.95509255e-01 5.27342856e-01 -3.82608801e-01
-8.52404356e-01 -1.14537585e+00 -2.47393981e-01 8.79077256e-01
1.63834229e-01 -5.90743959e-01 -8.57609451e-01 -7.03375578e-01
-1.01756817e-02 1.02405679e+00 -8.08613718e-01 -5.55535257e-01
-1.00192833e+00 -7.68438280e-01 1.26107168e+00 8.33614767e-01
4.54723425e-02 -1.24599326e+00 -6.15761817e-01 4.96041685e-01
-7.19241381e-01 -1.12079537e+00 -2.15306640e-01 3.64463478e-01
-6.59220815e-01 -1.50575972e+00 9.95376334e-02 -7.46790171e-01
-5.61816581e-02 -2.27167368e-01 1.82447684e+00 2.77197748e-01
-6.67407224e-03 -1.80419609e-01 -1.46412537e-01 -8.17420781e-02
-5.08859634e-01 7.07436144e-01 8.90882835e-02 -8.42425883e-01
9.27778661e-01 -7.83861816e-01 -2.08537295e-01 8.69787261e-02
-6.96270645e-01 -3.14659953e-01 6.63721800e-01 9.11271989e-01
1.12495899e-01 -2.88391918e-01 8.62054408e-01 -1.15345275e+00
1.10819829e+00 -4.63633388e-01 -3.11917990e-01 4.77556616e-01
-7.05967665e-01 4.61497515e-01 3.04804891e-01 -1.87519237e-01
-9.93849754e-01 -4.94295925e-01 -2.18871102e-01 -3.15342396e-01
-2.86440458e-02 8.32518876e-01 -1.40619472e-01 4.95685726e-01
8.73323202e-01 -5.05321264e-01 -3.88793617e-01 -7.75830925e-01
6.30491853e-01 3.78673881e-01 1.20461905e+00 -9.42068577e-01
5.68243921e-01 -9.50963497e-02 -5.39108336e-01 -1.62148312e-01
-1.20703220e+00 -1.16513728e-03 -5.95773041e-01 6.01807773e-01
6.87464714e-01 -9.57470119e-01 -6.95484936e-01 -2.88828105e-01
-1.38297880e+00 -9.23487321e-02 -9.86930057e-02 2.88504273e-01
-4.36813921e-01 1.00484990e-01 -9.26238060e-01 -1.84670791e-01
-3.99965435e-01 -1.17651331e+00 6.05459929e-01 3.24160904e-01
-9.40211833e-01 -8.62118244e-01 3.86202574e-01 5.35640717e-01
4.62008625e-01 1.73239652e-02 1.43174314e+00 -9.09966469e-01
-3.25725317e-01 -1.31191658e-02 -2.91961133e-01 -2.98620135e-01
-2.16533467e-01 -7.03399405e-02 -7.72672415e-01 -1.78445682e-01
-4.70019668e-01 -5.39142847e-01 1.00150406e+00 1.77683160e-01
6.13266289e-01 -1.16641700e-01 -7.08703816e-01 5.46918094e-01
1.23458397e+00 -1.82645500e-01 7.26409614e-01 6.60652995e-01
1.92893207e-01 5.28799713e-01 2.98222899e-01 -6.98457137e-02
6.84763730e-01 9.36923027e-01 5.18317223e-01 6.23960532e-02
-2.71503091e-01 -5.13099849e-01 1.04924319e-02 2.04524875e-01
-3.78387123e-02 -2.63735294e-01 -9.67248142e-01 8.28507006e-01
-2.05757213e+00 -1.39141953e+00 9.36784148e-02 1.91443288e+00
1.34876382e+00 5.56379855e-01 -1.61534980e-01 -4.87618856e-02
7.42237628e-01 2.23411962e-01 -3.93624276e-01 -5.29924572e-01
-2.71335810e-01 4.86057758e-01 2.30278507e-01 8.60089064e-01
-1.04467976e+00 1.13154376e+00 7.72718287e+00 9.30548366e-03
-8.91661584e-01 -9.14728567e-02 1.21163689e-01 -3.46615523e-01
-3.52105170e-01 2.88499504e-01 -7.61385262e-01 2.23338559e-01
1.11781228e+00 1.43463925e-01 6.86266601e-01 3.10805470e-01
-2.55937934e-01 1.77802965e-01 -1.57923400e+00 7.31725633e-01
-7.99186453e-02 -1.64313924e+00 -1.68155506e-01 -2.20590636e-01
1.67127147e-01 4.63486761e-01 -1.94143891e-01 6.63579464e-01
8.82409334e-01 -1.45795190e+00 6.13411129e-01 3.07519227e-01
5.55271924e-01 -8.19715798e-01 8.29464912e-01 1.89794645e-01
-5.71104765e-01 1.52310869e-02 -5.04006565e-01 -1.68734238e-01
3.57149273e-01 1.76890299e-01 -4.81125206e-01 5.61911285e-01
4.49578524e-01 4.15468544e-01 -5.31996965e-01 6.48226500e-01
-6.50005102e-01 5.48090160e-01 -2.04858661e-01 3.91838640e-01
1.36903301e-01 3.32901657e-01 5.40591419e-01 1.37861288e+00
-1.84610128e-01 3.28503609e-01 -2.07795035e-02 1.09058523e+00
-5.83916605e-01 -5.15292227e-01 -3.76016647e-01 -9.56573233e-04
1.25480354e+00 9.47376430e-01 5.59430383e-02 -7.07083121e-02
-2.98509270e-01 6.21113837e-01 1.08908415e+00 3.11782807e-01
-9.82306063e-01 -3.40721607e-01 9.52973366e-01 -9.21509564e-02
8.75766873e-02 1.63124859e-01 -2.06319571e-01 -1.28871644e+00
-1.29077122e-01 -1.21673298e+00 7.46693075e-01 -7.64176905e-01
-1.18631566e+00 5.55332839e-01 5.11949696e-02 -2.99714267e-01
-5.48227727e-01 -2.76404917e-01 -6.52125776e-01 8.73682797e-01
-1.64041889e+00 -9.18248177e-01 2.05780461e-01 3.53823513e-01
1.30472168e-01 1.30729899e-01 1.29121208e+00 5.44528604e-01
-8.28974247e-01 1.01270211e+00 -3.70648772e-01 5.78850925e-01
9.79450703e-01 -1.16309607e+00 7.32713044e-01 7.03330219e-01
1.41892806e-01 9.55704689e-01 1.11334956e+00 -4.38372135e-01
-1.26491249e+00 -5.58768570e-01 1.33690166e+00 -9.80308831e-01
6.92228496e-01 -1.14652723e-01 -1.18621075e+00 1.10565948e+00
5.11637270e-01 -2.59775203e-02 5.99140227e-01 1.12217200e+00
-7.34205961e-01 3.71699125e-01 -1.00140893e+00 5.76379955e-01
1.44983482e+00 -6.05629027e-01 -1.42249596e+00 -2.18979836e-01
7.50599325e-01 -6.82401597e-01 -1.01200163e+00 6.41010284e-01
4.45634574e-01 -8.81396115e-01 1.22851527e+00 -1.37730026e+00
6.34074390e-01 -1.47319749e-01 -4.15892988e-01 -1.45986879e+00
-5.86077809e-01 -4.73609775e-01 -1.17704518e-01 1.32106090e+00
7.78161049e-01 -1.47633851e-01 7.37579048e-01 6.84069633e-01
-2.51815289e-01 -5.61639011e-01 -1.02697015e+00 -3.60794306e-01
4.95548636e-01 -1.21057697e-01 6.60144091e-01 1.43048477e+00
6.59689784e-01 1.06372178e+00 -2.66482830e-01 3.29773664e-01
5.75598538e-01 1.77231550e-01 4.18668777e-01 -1.43198192e+00
-2.59384066e-01 -5.78780830e-01 6.21628249e-03 -6.11783326e-01
6.48311734e-01 -8.38769436e-01 -3.36557589e-02 -1.54613340e+00
3.24903905e-01 -3.60457808e-01 -5.20098150e-01 7.28619099e-01
-3.46843570e-01 1.66046485e-01 1.51754186e-01 9.27708745e-02
-6.88951671e-01 3.44822735e-01 5.43653309e-01 -1.91053763e-01
-6.01592765e-04 -6.16134346e-01 -1.32947731e+00 6.61884606e-01
4.58383560e-01 -5.92744648e-01 -2.69909292e-01 -9.98991966e-01
5.38799226e-01 4.21529621e-01 1.47768512e-01 -5.52617252e-01
2.41587400e-01 1.11150503e-01 2.21485659e-01 -3.64167541e-01
2.54523069e-01 -3.55584741e-01 1.79110423e-01 2.61946172e-01
-8.69103670e-01 3.23867470e-01 2.35407546e-01 2.01567695e-01
-1.68443084e-01 -2.72613078e-01 8.19701970e-01 -1.60062760e-01
-5.97377717e-01 -7.70367384e-02 7.99557492e-02 7.42532730e-01
3.57701987e-01 2.15998650e-01 -8.36308718e-01 -2.59176582e-01
-8.48097801e-01 3.25929254e-01 5.57730913e-01 7.02061117e-01
1.39755785e-01 -1.26883113e+00 -1.05698287e+00 -1.18153274e-01
1.28162250e-01 7.76515082e-02 -1.97781608e-01 8.08606923e-01
-1.65392607e-01 5.44025123e-01 -1.67694852e-01 -2.84396112e-02
-1.06348169e+00 6.49399102e-01 6.29245460e-01 -6.06479704e-01
-7.91302979e-01 7.98084736e-01 -1.11651756e-01 -5.49174428e-01
4.93967772e-01 -1.65525019e-01 -2.33588338e-01 8.74676704e-02
6.43654287e-01 5.69984727e-02 2.35494316e-01 -3.62840801e-01
-6.92251861e-01 1.13242924e-01 -4.35960621e-01 -1.95880327e-02
1.42999029e+00 8.46070051e-02 2.76521474e-01 -9.92678404e-02
7.79433250e-01 8.11975077e-02 -8.80575418e-01 -3.69229168e-01
4.01032418e-01 -1.59863830e-01 -5.80255054e-02 -1.27598250e+00
-9.51954842e-01 7.05095530e-01 1.48125350e-01 -7.47453198e-02
7.03376353e-01 2.07194816e-02 6.08805001e-01 6.16811216e-01
2.79513955e-01 -9.71017659e-01 -5.22789061e-01 7.67472684e-01
8.52266014e-01 -1.13545096e+00 -4.80900593e-02 -3.98895480e-02
-3.40698391e-01 7.84213066e-01 5.83051264e-01 -6.41071349e-02
1.98849618e-01 5.59813440e-01 1.63514391e-01 -2.92527109e-01
-1.17675149e+00 -1.41341351e-02 -2.42594838e-01 5.22231102e-01
9.27518368e-01 -2.69886732e-01 -4.05641168e-01 1.14155734e+00
-4.02454942e-01 6.54665753e-02 5.29740155e-01 6.00882471e-01
-2.56238669e-01 -1.20916855e+00 -1.18235618e-01 -1.12754673e-01
-8.42581272e-01 -3.99052769e-01 -7.81006396e-01 1.02501929e+00
-6.13566041e-02 1.03286016e+00 -1.05045289e-02 -4.04898763e-01
9.22006130e-01 2.24389240e-01 7.71687210e-01 -5.43949962e-01
-5.50908267e-01 -5.25558829e-01 8.42342198e-01 -7.45051086e-01
-3.26558769e-01 -7.50959635e-01 -1.42504370e+00 -7.04340518e-01
-4.11874145e-01 4.89778191e-01 2.91482657e-01 1.09452963e+00
4.85881656e-01 5.04457653e-01 -2.65360475e-02 -5.14435291e-01
-8.18906605e-01 -1.12492263e+00 -5.83930612e-02 7.75730908e-01
3.19661289e-01 -7.07536995e-01 -1.01487182e-01 -4.03977782e-01]
|
[9.32451057434082, 9.51655387878418]
|
d4c22668-8a15-4490-b127-84cf1f010382
|
reinforcement-learning-with-analogical
|
1712.1007
| null |
http://arxiv.org/abs/1712.10070v1
|
http://arxiv.org/pdf/1712.10070v1.pdf
|
Reinforcement Learning with Analogical Similarity to Guide Schema Induction and Attention
|
Research in analogical reasoning suggests that higher-order cognitive
functions such as abstract reasoning, far transfer, and creativity are founded
on recognizing structural similarities among relational systems. Here we
integrate theories of analogy with the computational framework of reinforcement
learning (RL). We propose a psychology theory that is a computational synergy
between analogy and RL, in which analogical comparison provides the RL learning
algorithm with a measure of relational similarity, and RL provides feedback
signals that can drive analogical learning. Simulation results support the
power of this approach.
|
['Matt Jones', 'James M. Foster']
|
2017-12-28
| null | null | null | null |
['analogical-similarity']
|
['reasoning']
|
[-1.13393798e-01 2.24721938e-01 -1.03673056e-01 -2.18396991e-01
4.87242728e-01 -5.63164592e-01 7.14963794e-01 3.32825512e-01
-2.30517164e-01 5.49780488e-01 1.59514174e-01 -4.13701981e-01
-8.11584711e-01 -1.22727728e+00 -4.98218983e-01 -4.82248031e-02
-9.85984504e-02 2.57577062e-01 -4.64668348e-02 -8.65040898e-01
1.03330004e+00 9.05831099e-01 -1.27761960e+00 2.83464223e-01
1.13556266e+00 6.27708375e-01 3.27870071e-01 4.72958118e-01
-2.05946043e-01 1.35806608e+00 -3.91708881e-01 -2.92656720e-01
2.41627216e-01 -9.41647291e-01 -1.16134977e+00 -5.92834532e-01
-1.52740493e-01 -4.84156638e-01 -5.65164030e-01 7.03958511e-01
1.87723562e-01 4.43032771e-01 6.38831377e-01 -1.10443258e+00
-1.79780698e+00 7.95491278e-01 1.90539181e-01 3.47764760e-01
1.11153233e+00 3.94793928e-01 9.56261277e-01 -6.11845732e-01
2.74009436e-01 1.59608161e+00 6.83936059e-01 4.67715770e-01
-1.23124313e+00 -5.92935085e-01 -2.12116241e-01 5.52080333e-01
-1.11923003e+00 2.88513064e-01 8.06871414e-01 -2.75038511e-01
1.10285664e+00 8.31876844e-02 1.54282904e+00 3.01563859e-01
6.62205458e-01 3.62159431e-01 1.61022270e+00 -7.08579063e-01
4.22813058e-01 1.90770179e-01 5.97715564e-02 5.24676681e-01
1.53594986e-01 9.86095846e-01 -7.24109948e-01 -5.26797026e-02
1.53482842e+00 1.59284040e-01 5.47532178e-02 -6.13028370e-03
-7.27074683e-01 8.58146012e-01 1.02820718e+00 5.96559405e-01
-6.46902800e-01 3.35465848e-01 5.97495288e-02 1.03215110e+00
-5.95498741e-01 1.48233283e+00 -5.48574477e-02 4.72861111e-01
-1.79365560e-01 1.78436339e-01 5.07862508e-01 5.89838326e-01
7.90578961e-01 3.73296678e-01 -2.08566517e-01 4.33761001e-01
4.56410140e-01 5.28873920e-01 8.07355464e-01 -1.39893138e+00
-2.58461982e-01 6.56418502e-01 -3.79309058e-01 -9.21520114e-01
-4.40909803e-01 -3.79396498e-01 6.91352710e-02 4.75436628e-01
3.85575518e-02 2.54610568e-01 -4.05866839e-02 1.86156905e+00
1.24795072e-01 3.61829028e-02 6.32168829e-01 8.98109674e-01
8.35696101e-01 6.00504518e-01 2.06763268e-01 -1.46861196e-01
1.17429531e+00 -6.52011454e-01 -7.01276898e-01 -2.90469348e-01
5.30918181e-01 -4.45496351e-01 1.60949302e+00 1.78167999e-01
-1.55747223e+00 -9.15932894e-01 -1.29015911e+00 -1.70585930e-01
-4.72453415e-01 -6.82165265e-01 1.18374002e+00 5.51208794e-01
-1.37889743e+00 1.19917846e+00 4.46581654e-02 -5.57545602e-01
2.86506295e-01 1.86098024e-01 5.24769761e-02 9.99754220e-02
-1.53650677e+00 1.56579912e+00 7.31538355e-01 -8.37318525e-02
-3.70484799e-01 -7.02387750e-01 -6.03585780e-01 2.28572935e-01
-5.48460567e-03 -1.09007096e+00 1.12933874e+00 -1.29157877e+00
-1.78281724e+00 8.53021204e-01 3.87700021e-01 -2.28603989e-01
-3.34432244e-01 -1.70405135e-01 -4.96978581e-01 5.35099626e-01
-1.85354263e-01 6.97900176e-01 2.67690241e-01 -1.23205864e+00
-1.17603503e-01 -3.83189261e-01 5.09024978e-01 6.51486754e-01
-1.13447338e-01 -2.55147725e-01 6.98172629e-01 -6.33864641e-01
3.65789294e-01 -6.78968489e-01 -8.41725990e-02 6.39520139e-02
4.65480596e-01 -5.70386946e-01 1.34804800e-01 -4.10409331e-01
7.59417415e-01 -2.07263684e+00 -3.74694052e-03 4.29660976e-01
2.58504122e-01 -1.79601684e-01 -4.11569715e-01 8.92397463e-01
-2.91909575e-01 -5.51557168e-02 4.40706998e-01 1.18826532e+00
2.13015929e-01 1.94939956e-01 -6.17992699e-01 -1.33753922e-02
1.46457434e-01 1.57922888e+00 -1.22415352e+00 -1.56648010e-01
-1.93766672e-02 -1.10377774e-01 -7.31443405e-01 6.73291445e-01
-6.87151849e-02 3.09691727e-01 -3.79887134e-01 3.14109802e-01
2.86881030e-01 -5.31980544e-02 5.03084064e-01 8.94279778e-02
8.33092108e-02 5.58094084e-01 -7.42977858e-01 1.52522779e+00
-4.44543660e-01 6.41064405e-01 -9.98891175e-01 -8.81896675e-01
1.36074042e+00 3.25494766e-01 -3.23004425e-01 -1.36023343e+00
2.26632521e-01 1.64756298e-01 8.67975771e-01 -6.56590462e-01
2.79025614e-01 -5.82192361e-01 3.10619354e-01 9.69367445e-01
7.45849730e-03 -1.09987521e+00 -1.71135202e-01 2.47910842e-01
6.90215647e-01 5.18451452e-01 8.47293496e-01 -5.22127867e-01
3.87742341e-01 -1.19740158e-01 -9.55644697e-02 8.13904345e-01
7.89233893e-02 -4.25721765e-01 1.98367178e-01 -4.30427283e-01
-9.00641382e-01 -1.61797154e+00 3.71083245e-02 1.16744232e+00
3.72243434e-01 -2.31559239e-02 -5.55285394e-01 2.11333722e-01
3.37992944e-02 1.46896005e+00 -6.20452881e-01 -9.12774801e-01
-4.63866174e-01 1.97170749e-02 3.52808297e-01 8.13265383e-01
3.59477967e-01 -1.88305604e+00 -9.69343781e-01 1.39834538e-01
3.51387829e-01 -4.49963629e-01 1.50783807e-01 1.56172857e-01
-1.45434713e+00 -8.06138694e-01 2.36810833e-01 -9.27121460e-01
3.76937926e-01 4.67766464e-01 1.26891828e+00 7.21426427e-01
-3.20953995e-01 8.89317930e-01 -2.33131111e-01 -2.99162179e-01
-6.77159965e-01 -7.92946815e-01 7.05901459e-02 -1.08306146e+00
4.67740059e-01 -1.17000592e+00 -5.72383642e-01 3.76987383e-02
-8.59938562e-01 -2.06674844e-01 8.24208856e-01 7.22770572e-01
1.45108119e-01 -1.49953380e-01 1.19587576e+00 -1.92923397e-01
1.54538417e+00 -5.38308918e-01 -2.20820844e-01 4.59368825e-01
-8.98109198e-01 2.67898500e-01 6.11242712e-01 -7.37316787e-01
-1.20117128e+00 -7.31763840e-01 4.26998168e-01 -1.53925270e-01
9.98160318e-02 5.80551207e-01 1.41868606e-01 -3.00550431e-01
1.06776381e+00 6.58066690e-01 2.42010012e-01 2.72820234e-01
8.54558885e-01 2.21822202e-01 3.85119498e-01 -1.17348528e+00
5.65200269e-01 -1.31409436e-01 1.94373995e-01 -4.12708879e-01
-5.32399058e-01 4.97196406e-01 -3.14632207e-01 -3.89757961e-01
6.93850398e-01 -5.68456709e-01 -1.51665699e+00 -3.67339283e-01
-9.53855753e-01 -2.32675850e-01 -9.89493966e-01 7.44046807e-01
-1.23136640e+00 7.49118850e-02 -7.80447602e-01 -8.22924912e-01
-2.24998102e-01 -7.02457786e-01 -3.33218393e-03 6.43540740e-01
-7.66809523e-01 -9.74413812e-01 3.02808844e-02 2.03095041e-02
5.37911117e-01 -2.48114899e-01 1.51731896e+00 -8.94367456e-01
-5.82428336e-01 1.37123570e-01 2.73964088e-02 -3.34109478e-02
4.27663550e-02 -3.62560570e-01 -7.07198799e-01 2.87534267e-01
4.34841663e-01 -9.63762164e-01 1.14256859e-01 4.55447361e-02
1.00817347e+00 -1.07833378e-01 9.83842909e-02 2.19003424e-01
1.41070855e+00 9.76493418e-01 1.01736784e+00 1.75759584e-01
-2.91062389e-02 8.88353348e-01 6.72454774e-01 2.55905360e-01
3.73343587e-01 2.17773333e-01 -2.07947418e-01 3.41104180e-01
1.30804718e-01 -5.98054171e-01 1.04098648e-01 8.17878425e-01
-4.00484681e-01 5.63596368e-01 -6.82510495e-01 2.93403864e-03
-1.50188625e+00 -1.48489368e+00 2.74902582e-01 2.05618429e+00
1.27947092e+00 1.24920756e-01 -5.75109869e-02 3.65640409e-02
2.77017027e-01 -5.26173174e-01 -7.90928781e-01 -1.20038462e+00
-1.10498760e-02 6.89511955e-01 -4.18339729e-01 4.34523702e-01
9.41681117e-02 1.35112059e+00 8.06361198e+00 4.77577746e-01
-6.18879914e-01 -3.20991457e-01 2.81309664e-01 3.89720023e-01
-5.91864765e-01 8.61464441e-02 3.47787626e-02 -3.18521321e-01
9.21720147e-01 -5.83302438e-01 1.06470323e+00 5.64346552e-01
6.98688477e-02 -1.27546579e-01 -1.77979445e+00 6.66240692e-01
-1.24297805e-01 -1.32511270e+00 5.93133390e-01 -2.31057346e-01
4.42291319e-01 -9.79386389e-01 1.80666313e-01 5.04985094e-01
5.88793635e-01 -1.43236399e+00 5.64998627e-01 8.32482398e-01
2.60052413e-01 -6.63445830e-01 2.31406286e-01 6.74792528e-02
-9.79023397e-01 -5.56331038e-01 -6.56635344e-01 -7.38264501e-01
-3.79940093e-01 -3.99353981e-01 -6.89776659e-01 2.38109142e-01
4.33785975e-01 4.07958716e-01 -3.14855307e-01 6.41883552e-01
-8.38165879e-01 1.10124201e-02 3.13761175e-01 -5.45487940e-01
-1.74391970e-01 -6.16823614e-01 -1.10169098e-01 5.68693459e-01
1.26775041e-01 9.33671236e-01 -1.58854827e-01 1.40054703e+00
3.30897033e-01 2.33441591e-01 -8.52367520e-01 -9.25157368e-02
1.14160609e+00 7.82066643e-01 -7.93720484e-01 -4.49507445e-01
-2.51982838e-01 6.76552653e-01 4.91847694e-01 2.31248036e-01
-6.70427918e-01 -5.38935736e-02 3.73877525e-01 -2.19056591e-01
-1.78294890e-02 -1.09977469e-01 -6.50532961e-01 -5.06134331e-01
-4.99704510e-01 -1.00737965e+00 1.31563872e-01 -1.37242246e+00
-1.47565866e+00 1.79040998e-01 -7.55849332e-02 -7.84167171e-01
-3.51418912e-01 -6.89489245e-01 -8.69618475e-01 9.31630254e-01
-9.25161779e-01 -7.72989273e-01 -2.10551366e-01 9.07075286e-01
1.76320449e-01 -2.31849566e-01 1.13512123e+00 -6.64353251e-01
2.62932807e-01 4.38354045e-01 -4.32734072e-01 -2.11403519e-01
1.01044573e-01 -1.15446520e+00 1.58362404e-01 -6.55204207e-02
2.89487280e-02 9.89783943e-01 5.51730335e-01 -6.01438165e-01
-1.63460243e+00 -6.35856166e-02 4.96761620e-01 -2.23221868e-01
8.45626831e-01 1.25250995e-01 -1.08307922e+00 5.69024086e-01
4.46126997e-01 -4.69016373e-01 1.19080007e+00 -5.99903241e-02
-7.41862357e-01 -1.19027227e-01 -1.26728964e+00 1.09253800e+00
1.32866752e+00 -8.87247503e-01 -1.71202028e+00 2.65152395e-01
1.02757287e+00 2.83742368e-01 -1.20368659e+00 -1.43227652e-01
8.21248055e-01 -9.55584228e-01 1.46768868e+00 -8.80854845e-01
6.81408644e-01 -8.90176147e-02 -3.06775600e-01 -1.45531344e+00
-8.93833280e-01 -1.24731198e-01 1.31081581e-01 5.25509477e-01
2.64089145e-02 -1.19490123e+00 4.99617383e-02 5.04227042e-01
8.16433132e-03 -4.80074435e-01 -5.64804494e-01 -1.02874637e+00
5.99992812e-01 4.40624356e-02 7.75910914e-01 1.28809655e+00
9.12570238e-01 6.48393750e-01 3.39982331e-01 -1.05091877e-01
3.12762856e-01 2.75013477e-01 2.26600319e-01 -1.59124255e+00
-7.37861574e-01 -8.50421965e-01 -2.64960438e-01 -6.62809372e-01
3.61261636e-01 -1.19149220e+00 -4.98790681e-01 -1.41999984e+00
-2.82647181e-03 -1.88134521e-01 -3.07554901e-01 2.24263087e-01
-9.16578248e-02 -3.27040970e-01 4.99517500e-01 2.59603560e-01
1.24308700e-03 5.85597992e-01 1.60723698e+00 4.26065236e-01
-4.43919152e-01 -6.01420283e-01 -1.16096759e+00 7.00412333e-01
1.15698147e+00 -2.62948900e-01 -1.01737034e+00 -2.79696118e-02
5.92283189e-01 5.87675750e-01 5.07717729e-01 -9.77148652e-01
2.27126926e-01 -6.38659239e-01 8.56260419e-01 8.90531465e-02
2.14866713e-01 -9.39596474e-01 -4.33443934e-02 1.29798150e+00
-7.64059186e-01 6.71831846e-01 3.89445662e-01 1.54793739e-01
2.79561579e-01 -4.62305069e-01 6.77559555e-01 -1.57615110e-01
-6.06567442e-01 -5.93332946e-01 -5.62583983e-01 1.14593044e-01
1.00248921e+00 -4.61286098e-01 -4.49132770e-01 -3.71582359e-01
-6.92567110e-01 -1.14529341e-01 1.93516150e-01 3.33490819e-01
1.06018913e+00 -1.46836150e+00 -5.11455894e-01 -1.14656799e-01
-1.14190154e-01 -9.41633821e-01 -9.53554362e-02 4.23363417e-01
-5.25623024e-01 3.07224542e-01 -9.81465042e-01 2.66692340e-02
-5.99671006e-01 1.16346705e+00 6.19618714e-01 2.35015973e-01
-4.95115012e-01 7.12114215e-01 2.40482390e-01 -3.47878307e-01
-2.77118921e-01 -1.27187967e-01 -5.75838208e-01 -3.03459942e-01
5.96629798e-01 3.73145610e-01 -6.92070842e-01 1.27210036e-01
-1.47219360e-01 6.94435358e-01 2.48862445e-01 -4.18691605e-01
8.90756726e-01 1.35754049e-01 -5.08758008e-01 6.68762267e-01
4.05224800e-01 -3.41515958e-01 -5.70932388e-01 -3.73960823e-01
7.90511891e-02 -4.22382623e-01 -4.29579437e-01 -9.44724441e-01
-4.11655307e-01 8.56622040e-01 5.96154749e-01 2.52123326e-01
1.24489045e+00 3.35274816e-01 6.37196302e-02 1.06506133e+00
2.84354329e-01 -1.28691339e+00 9.72873747e-01 2.88458765e-01
1.38101900e+00 -5.81110060e-01 3.57211262e-01 -2.30899304e-01
-5.46670198e-01 1.49018967e+00 9.76744294e-01 -7.54550040e-01
4.31764007e-01 -2.81943567e-02 -3.90718609e-01 -2.77679980e-01
-9.06677961e-01 -1.45865262e-01 1.95054129e-01 8.54856372e-01
7.80245423e-01 -1.27663296e-02 -7.16271162e-01 5.62347293e-01
-8.24710965e-01 1.43005610e-01 3.73774469e-01 9.74398673e-01
-7.82123923e-01 -7.75902152e-01 -6.11231267e-01 2.19436958e-02
3.38041455e-01 -5.05307078e-01 -5.47926784e-01 8.87299597e-01
-3.43972892e-02 9.30861115e-01 3.09885859e-01 -2.54598230e-01
4.55497444e-01 3.77206691e-02 1.26529837e+00 -2.57014751e-01
-1.06544161e+00 -4.50796574e-01 -4.11182076e-01 -7.99727738e-01
-3.38963151e-01 -4.64072049e-01 -2.03358865e+00 -5.33192933e-01
1.45248592e-01 2.20653147e-01 3.62190664e-01 1.09812534e+00
9.82197523e-02 5.62418580e-01 5.68096995e-01 -4.17752773e-01
-8.07406425e-01 -7.24258363e-01 -5.64096987e-01 3.17376196e-01
-3.41306537e-01 -5.95391512e-01 -2.03454524e-01 -3.37298393e-01]
|
[10.598616600036621, 2.3063225746154785]
|
32817920-0c95-47a4-8b9c-6f366a52d69b
|
pag-net-progressive-attention-guided-depth
|
1911.09878
| null |
https://arxiv.org/abs/1911.09878v1
|
https://arxiv.org/pdf/1911.09878v1.pdf
|
PAG-Net: Progressive Attention Guided Depth Super-resolution Network
|
In this paper, we propose a novel method for the challenging problem of guided depth map super-resolution, called PAGNet. It is based on residual dense networks and involves the attention mechanism to suppress the texture copying problem arises due to improper guidance by RGB images. The attention module mainly involves providing the spatial attention to guidance image based on the depth features. We evaluate the proposed trained models on test dataset and provide comparisons with the state-of-the-art depth super-resolution methods.
|
['Sankaraganesh Jonna', 'Arpit Bansal', 'Rajiv R. Sahay']
|
2019-11-22
| null | null | null | null |
['depth-map-super-resolution']
|
['computer-vision']
|
[ 6.27858758e-01 5.41414201e-01 2.27736875e-01 -1.94401562e-01
-4.42654163e-01 3.50200266e-01 4.96834338e-01 -5.76718748e-01
-3.44481319e-01 8.54070365e-01 7.35937953e-01 3.24734837e-01
-1.69503763e-02 -1.26346219e+00 -6.02056384e-01 -6.13812983e-01
2.67257720e-01 3.17808509e-01 7.73631573e-01 -4.66777146e-01
7.09539533e-01 4.27748471e-01 -1.63144171e+00 4.32177305e-01
1.00530529e+00 1.00154448e+00 5.69683373e-01 4.39180136e-01
-4.14208025e-01 8.15932214e-01 -4.49420661e-01 1.89725205e-01
1.76506892e-01 -3.26120168e-01 -8.47643971e-01 -1.15354232e-01
6.05393171e-01 -1.09870923e+00 -6.76192999e-01 1.14268851e+00
6.34542584e-01 3.06494176e-01 3.96685630e-01 -5.51304638e-01
-1.05587089e+00 1.09357201e-01 -8.42786014e-01 8.22588086e-01
4.82164800e-01 -4.10294533e-01 3.70155632e-01 -9.72310603e-01
7.55469263e-01 1.64248145e+00 2.65358627e-01 8.63320231e-01
-8.33253503e-01 -4.36091810e-01 3.25260043e-01 3.49059135e-01
-1.09940541e+00 -3.37927938e-01 7.44742513e-01 -1.69163034e-03
1.08214271e+00 -2.61486948e-01 4.30794388e-01 9.15150583e-01
3.10339123e-01 5.38393974e-01 1.08713448e+00 -3.02877188e-01
1.28223836e-01 -2.26414785e-01 -9.42342728e-02 6.90506577e-01
-6.65366277e-03 4.05062556e-01 -7.67990291e-01 3.09843868e-01
1.74219012e+00 2.67068204e-02 -4.97104943e-01 -9.92612392e-02
-7.82179713e-01 6.40334606e-01 7.49141276e-01 1.39072090e-01
-3.94059151e-01 1.32594369e-02 -6.97261989e-02 -2.54197329e-01
9.31647897e-01 1.07941985e-01 -2.21908063e-01 2.66896307e-01
-8.49574208e-01 1.71784282e-01 6.52762055e-02 8.89393210e-01
9.35434282e-01 2.20789477e-01 -1.77787244e-01 7.53863931e-01
2.76175767e-01 4.42936011e-02 5.60548961e-01 -1.06167305e+00
4.70246047e-01 5.62614083e-01 1.08226903e-01 -8.25620532e-01
-4.32906032e-01 -2.25415975e-01 -1.06618118e+00 6.27217948e-01
-1.66067496e-01 7.65973181e-02 -1.34262490e+00 1.38162231e+00
2.59021759e-01 2.55807012e-01 1.76505804e-01 1.33572018e+00
1.48418784e+00 6.50659740e-01 -2.41007641e-01 -1.30397424e-01
9.03935075e-01 -1.27064133e+00 -1.01716232e+00 -5.75196862e-01
-1.48216799e-01 -4.62350428e-01 8.93005550e-01 4.43353355e-01
-1.33597958e+00 -5.76260567e-01 -1.25935757e+00 -6.60584092e-01
-3.23127538e-01 -2.98834413e-01 6.26974046e-01 1.24412671e-01
-1.48118591e+00 8.71403694e-01 -6.40758455e-01 -4.78135318e-01
5.97710490e-01 4.59184140e-01 -5.02949893e-01 -3.69071811e-01
-1.09899998e+00 7.51523554e-01 1.95807144e-01 2.68806517e-01
-9.88852859e-01 -4.59223032e-01 -7.88065910e-01 -1.05359986e-01
1.12247773e-01 -6.78421199e-01 1.03444576e+00 -6.27220333e-01
-1.64484692e+00 9.10973907e-01 -2.82601744e-01 -6.85552806e-02
2.90307909e-01 -5.22412777e-01 -6.86918572e-02 4.34503198e-01
3.13719928e-01 7.52695858e-01 9.36413825e-01 -1.21940076e+00
-8.83526146e-01 -8.07431161e-01 1.34738907e-01 5.73330104e-01
8.58999044e-02 -2.26582244e-01 -5.57401240e-01 -6.35894835e-01
7.62573063e-01 -2.43546233e-01 -2.56640077e-01 -1.63228616e-01
-2.35635087e-01 1.30630329e-01 9.61059570e-01 -7.93590963e-01
9.21180665e-01 -2.00331783e+00 4.70549226e-01 -1.65046439e-01
4.67752695e-01 -8.44544638e-03 9.56829935e-02 -4.96256948e-02
1.38278663e-01 2.78206825e-01 -1.75381377e-01 -5.28354585e-01
-6.21638715e-01 1.20622635e-01 -1.95566237e-01 4.06791449e-01
-7.48891011e-02 6.01509035e-01 -7.09324479e-01 -5.62081635e-01
3.28307569e-01 7.59277344e-01 -3.89639407e-01 5.88741779e-01
-1.03313006e-01 6.67307794e-01 -4.43874985e-01 7.36763299e-01
1.22102571e+00 -3.03944387e-02 -3.90346557e-01 -3.24748874e-01
-2.75963932e-01 4.14928466e-01 -1.01850152e+00 2.22285056e+00
-2.30115518e-01 3.26172501e-01 2.97461182e-01 -4.21900272e-01
9.83474433e-01 -5.19003533e-02 1.50261804e-01 -1.25484240e+00
1.07498899e-01 -9.29705799e-03 -5.49889207e-01 -4.18755800e-01
8.41173589e-01 -1.54723480e-01 4.94230717e-01 3.90046000e-01
3.15027200e-02 -1.82970643e-01 -3.92259091e-01 2.05310076e-01
9.89237845e-01 4.93012697e-01 -1.66633487e-01 -1.44873306e-01
6.34362936e-01 -2.90127069e-01 5.74143410e-01 4.43068773e-01
-2.42918804e-01 1.10986185e+00 1.82938814e-01 -4.25547093e-01
-1.07391632e+00 -1.06204808e+00 -2.19794258e-01 8.45748425e-01
6.54648840e-01 -5.12982383e-02 -9.09095526e-01 -4.50164407e-01
-3.63918811e-01 6.17951378e-02 -1.10902488e+00 9.86484578e-04
-5.68540692e-01 -6.41762853e-01 -3.56256515e-02 7.74806857e-01
1.27807140e+00 -1.25401878e+00 -4.64551687e-01 2.06008762e-01
-1.48356020e-01 -1.03914237e+00 -1.65673465e-01 1.61531970e-01
-1.22153926e+00 -9.23291445e-01 -9.66275930e-01 -7.48508513e-01
6.77871108e-01 5.57703018e-01 1.08190536e+00 4.50516865e-02
-7.63091072e-02 -8.85055363e-02 -3.19812238e-01 -3.84528451e-02
3.61742973e-01 1.53654113e-01 -3.01214725e-01 -2.59061962e-01
2.03422815e-01 -9.38605964e-01 -1.07346356e+00 2.81835645e-01
-8.17077935e-01 2.90195495e-01 5.72792590e-01 6.67579710e-01
1.08701038e+00 1.75433591e-01 3.47551525e-01 -1.00218058e+00
4.40303832e-01 -2.32573256e-01 -4.85247523e-01 -2.18074977e-01
-5.40718019e-01 3.75844389e-01 1.61263555e-01 8.86306986e-02
-1.61688197e+00 -2.35489756e-01 -1.45442158e-01 -4.01597232e-01
-7.01669604e-02 -9.23222154e-02 -3.11917752e-01 -5.24690330e-01
6.08081520e-01 3.09477955e-01 -3.22099596e-01 -8.41518521e-01
1.58438072e-01 3.73477429e-01 8.40166628e-01 -2.79917330e-01
4.76676166e-01 9.71517026e-01 3.40473838e-02 -6.27205253e-01
-9.86745536e-01 -3.65682989e-02 -7.96521127e-01 -1.56405848e-02
1.21450078e+00 -9.49554622e-01 -1.30983561e-01 5.29817939e-01
-1.24592519e+00 -4.18006480e-01 9.04131010e-02 2.23443776e-01
-6.68276548e-01 1.94023401e-01 -8.08469236e-01 -6.90954983e-01
-4.96929735e-01 -8.78445268e-01 1.27341521e+00 5.99887490e-01
4.00874972e-01 -7.68927574e-01 1.85336813e-01 2.96862394e-01
6.98090971e-01 3.32047254e-01 8.21715474e-01 2.20404580e-01
-1.14024758e+00 4.72895235e-01 -7.72846162e-01 1.20737962e-01
2.62536407e-01 -4.47196335e-01 -1.20267999e+00 -1.07335515e-01
1.97542440e-02 -8.93980116e-02 1.27040136e+00 5.71233094e-01
1.31005788e+00 8.49260017e-02 -3.08188230e-01 1.27177846e+00
1.68038476e+00 1.37626901e-01 1.32264233e+00 7.55691171e-01
9.83756185e-01 4.78375137e-01 7.73382425e-01 2.88931072e-01
2.91394114e-01 4.49135602e-01 9.37099516e-01 -3.27395350e-01
-2.84219056e-01 -4.08185720e-01 -1.69821694e-01 2.87005991e-01
-5.96559703e-01 3.41388583e-02 -3.92503828e-01 4.22376305e-01
-1.74155986e+00 -7.53529191e-01 -4.89163436e-02 1.94261754e+00
4.10031289e-01 3.21062505e-01 -3.74486893e-01 -1.03487005e-03
7.24126875e-01 4.49297905e-01 -5.66215277e-01 -4.76885855e-01
-4.11132663e-01 5.67688823e-01 5.12332678e-01 7.54254818e-01
-1.01807749e+00 1.17340088e+00 7.38242912e+00 5.94142199e-01
-9.51896489e-01 2.78716207e-01 6.06504738e-01 -1.82799846e-01
-3.79850298e-01 -3.32686543e-01 -8.93462539e-01 1.76127031e-01
3.77598554e-01 4.29087132e-02 5.57392001e-01 7.16228902e-01
1.96178287e-01 -5.41696966e-01 -8.70506227e-01 1.16316116e+00
2.55154133e-01 -1.24958110e+00 3.09520990e-01 1.76319256e-02
9.69037473e-01 2.96136942e-02 1.69452667e-01 -4.43714224e-02
-2.67952383e-02 -1.29979765e+00 3.63632619e-01 7.53453553e-01
1.07642376e+00 -9.86559153e-01 8.82293105e-01 -4.15185690e-02
-1.22572196e+00 -6.12278581e-02 -9.04325068e-01 -2.17603505e-01
1.41982019e-01 4.10406291e-01 -1.24052189e-01 5.32377303e-01
1.08765471e+00 9.60049927e-01 -4.97434974e-01 7.77761877e-01
-5.36309898e-01 -1.76946923e-01 1.18398676e-02 6.17408454e-01
1.85826272e-01 -4.84219454e-02 2.64074743e-01 7.97457218e-01
4.57112938e-01 5.36024988e-01 -4.93310869e-01 1.13629794e+00
4.68239747e-02 -3.59913595e-02 -6.38085783e-01 4.96556908e-01
2.35392541e-01 9.82421279e-01 -4.63658661e-01 -2.72864193e-01
-4.27384973e-01 1.38349926e+00 5.89446187e-01 5.77623129e-01
-5.04963398e-01 -2.65944362e-01 4.34719086e-01 4.96507585e-01
4.01719987e-01 5.77185787e-02 -5.65778971e-01 -9.54884708e-01
-1.76964119e-01 -4.50750411e-01 2.27152839e-01 -1.44696093e+00
-1.00892162e+00 1.03277481e+00 -1.48110181e-01 -8.52922857e-01
9.86191258e-03 -3.98278207e-01 -6.47291422e-01 1.44127297e+00
-1.92277098e+00 -7.91361511e-01 -1.01484370e+00 9.43226218e-01
6.01860404e-01 -1.34921940e-02 5.88373363e-01 3.24202955e-01
-4.73261297e-01 3.00152570e-01 -1.96083680e-01 -3.05895418e-01
6.54998064e-01 -1.06300616e+00 4.27990854e-01 7.54914403e-01
-6.50120616e-01 2.97720641e-01 4.37023342e-01 -8.31479549e-01
-9.43444133e-01 -1.00806570e+00 4.70069051e-01 -1.36537135e-01
4.94170524e-02 -1.84169114e-01 -1.06206203e+00 7.46721983e-01
1.62241235e-01 2.12706849e-01 -6.59926385e-02 -2.12621406e-01
4.03959909e-03 -8.89791362e-03 -1.55513048e+00 1.66934460e-01
1.50245810e+00 -5.92360973e-01 -6.19571745e-01 -1.39120221e-01
9.50030923e-01 -8.35541010e-01 -5.90384960e-01 6.10786498e-01
2.98630655e-01 -1.53612387e+00 1.02863848e+00 -5.77692278e-02
9.44452941e-01 -3.24104935e-01 -2.80301571e-01 -1.11267304e+00
-8.76950800e-01 -8.92022625e-02 -2.25855947e-01 9.42852616e-01
-2.46267803e-02 -3.57896626e-01 1.22023368e+00 5.52092195e-01
-3.44776124e-01 -8.12894106e-01 -9.53792095e-01 -2.01209217e-01
-1.80562824e-01 1.65102661e-01 6.73058569e-01 6.57540619e-01
-3.76406342e-01 3.47696871e-01 -3.63734931e-01 3.47643614e-01
1.00641358e+00 -2.23585039e-01 3.76899630e-01 -1.10230339e+00
-1.60688087e-01 -2.69266039e-01 -4.29404110e-01 -1.26324511e+00
-1.24640360e-01 -3.24487776e-01 3.78982984e-02 -2.09452629e+00
2.53851414e-01 -8.18575472e-02 -2.69354939e-01 9.89907756e-02
-1.79793000e-01 5.42978942e-01 -2.62705952e-01 5.21337539e-02
-5.32493651e-01 8.44147563e-01 1.69839203e+00 5.62690236e-02
-3.94793272e-01 -5.47410727e-01 -8.44784200e-01 6.85191393e-01
7.15156734e-01 -6.84067905e-02 -6.01547241e-01 -8.24988008e-01
1.09723233e-01 1.97427765e-01 2.61322200e-01 -1.03086209e+00
1.00863099e-01 -9.35739353e-02 7.25066364e-01 -9.83894587e-01
5.55355012e-01 -5.49326479e-01 -2.32989609e-01 2.25997400e-02
-8.26688260e-02 -4.48279530e-02 3.04386735e-01 5.21579683e-01
-3.12606126e-01 2.32652873e-01 9.37767088e-01 -5.46770990e-01
-1.06877387e+00 6.22930288e-01 -9.46808532e-02 -1.30958892e-02
5.92083573e-01 -4.23763573e-01 -7.00886965e-01 -4.07742471e-01
-1.01212013e+00 8.85815099e-02 5.77018917e-01 1.33864507e-01
1.30588567e+00 -1.52560496e+00 -5.84815204e-01 4.37571734e-01
-1.79164544e-01 6.39730990e-01 6.56380832e-01 2.47421175e-01
-7.78235316e-01 4.34508204e-01 -6.86413407e-01 -3.19514602e-01
-9.75590467e-01 6.22938216e-01 6.77338481e-01 -1.24474712e-01
-9.87638593e-01 1.10507774e+00 8.95668983e-01 1.23886820e-02
3.69406700e-01 -2.63290793e-01 -6.31603360e-01 -5.54857254e-01
9.64712143e-01 5.05660474e-01 -9.59644094e-03 -3.82988364e-01
-4.79355335e-01 1.07068717e+00 -2.20302165e-01 -2.53659844e-01
1.49646485e+00 -5.48108816e-01 -2.90340036e-01 1.44960567e-01
7.69622505e-01 -3.48830402e-01 -1.66070080e+00 -2.42811218e-01
-7.03891933e-01 -9.12595332e-01 5.72073877e-01 -8.09053659e-01
-1.38967633e+00 9.97167885e-01 9.53277051e-01 -2.52863586e-01
1.28395367e+00 -1.30635813e-01 7.66715467e-01 -1.85631767e-01
6.58206463e-01 -1.03423417e+00 3.18163931e-01 5.91013670e-01
1.15044057e+00 -1.40912700e+00 1.72423467e-01 -5.24134874e-01
-2.05918059e-01 8.84046137e-01 1.29572177e+00 -5.24842322e-01
7.11738050e-01 1.98206782e-01 2.96354350e-02 -4.01351929e-01
-5.13918459e-01 -1.36895940e-01 -8.55579972e-02 1.04353237e+00
3.24380815e-01 -4.76686120e-01 -2.58010805e-01 5.89285493e-01
-7.87701830e-02 2.14444369e-01 6.37604117e-01 6.94596589e-01
-8.31426203e-01 -7.06769288e-01 -5.15656054e-01 2.29878843e-01
-3.74129623e-01 -3.24703455e-01 -3.91834527e-01 6.10069573e-01
3.96963179e-01 7.65037596e-01 3.32602829e-01 -4.78748739e-01
4.03743237e-01 -4.44054157e-01 7.72117853e-01 -7.70038426e-01
-8.71398076e-02 2.13301759e-02 -1.07708886e-01 -1.10733569e+00
-5.62399268e-01 -1.86796471e-01 -1.46203458e+00 -4.75891143e-01
1.14782058e-01 -2.80003220e-01 4.04795974e-01 7.58823872e-01
3.04917783e-01 9.82297897e-01 4.37001109e-01 -1.51997054e+00
3.33791226e-01 -1.17868531e+00 -1.02878988e+00 1.01237826e-01
7.20569789e-01 -1.01589191e+00 -5.58575332e-01 -4.62964118e-01]
|
[9.873373031616211, -2.4120566844940186]
|
4c5b11ab-8591-46dc-86b7-9ab84d62fef3
|
chordmixer-a-scalable-neural-attention-model
|
2206.05852
| null |
https://arxiv.org/abs/2206.05852v2
|
https://arxiv.org/pdf/2206.05852v2.pdf
|
ChordMixer: A Scalable Neural Attention Model for Sequences with Different Lengths
|
Sequential data naturally have different lengths in many domains, with some very long sequences. As an important modeling tool, neural attention should capture long-range interaction in such sequences. However, most existing neural attention models admit only short sequences, or they have to employ chunking or padding to enforce a constant input length. Here we propose a simple neural network building block called ChordMixer which can model the attention for long sequences with variable lengths. Each ChordMixer block consists of a position-wise rotation layer without learnable parameters and an element-wise MLP layer. Repeatedly applying such blocks forms an effective network backbone that mixes the input signals towards the learning targets. We have tested ChordMixer on the synthetic adding problem, long document classification, and DNA sequence-based taxonomy classification. The experiment results show that our method substantially outperforms other neural attention models.
|
['Zhirong Yang', 'Lei Cheng', 'Tong Yu', 'Ruslan Khalitov']
|
2022-06-12
| null | null | null | null |
['document-classification', 'long-range-modeling']
|
['natural-language-processing', 'natural-language-processing']
|
[ 3.89181823e-01 5.23357876e-02 -4.56501395e-01 -3.03350449e-01
-4.78251994e-01 -5.33227384e-01 4.71444190e-01 -4.46299836e-02
-5.58059454e-01 6.69139206e-01 2.88344443e-01 -4.29638028e-01
1.02256713e-02 -5.24482131e-01 -9.70434010e-01 -9.67240632e-01
-2.00720772e-01 6.64053559e-01 2.77419865e-01 -1.16143383e-01
4.22771484e-01 2.40092918e-01 -1.16091335e+00 5.38553238e-01
7.87263393e-01 6.39579535e-01 4.95160669e-01 1.07729709e+00
-3.22285593e-01 1.06217051e+00 -7.70365477e-01 -1.04058802e-01
6.46108091e-02 -6.08885288e-01 -9.19018626e-01 -3.75329614e-01
1.68758333e-01 -1.05499998e-01 -3.34241182e-01 8.41374218e-01
6.22969329e-01 2.19473407e-01 7.03933179e-01 -1.14885676e+00
-9.87115204e-01 1.26155400e+00 -5.29158711e-01 2.89846838e-01
1.28709152e-02 3.47043216e-01 1.13150513e+00 -5.49957514e-01
4.93063778e-01 1.25365531e+00 8.10743093e-01 5.91373086e-01
-9.77009475e-01 -5.77280939e-01 2.82927573e-01 4.78870153e-01
-1.07116187e+00 -6.32870495e-02 7.10132420e-01 -4.52286988e-01
1.24400485e+00 2.10897326e-01 5.48769653e-01 1.28920460e+00
3.19925129e-01 9.85513747e-01 2.82520980e-01 -2.75753498e-01
5.26317256e-03 -5.06648421e-01 6.33443832e-01 2.87876368e-01
-1.34206206e-01 -8.24683532e-02 -1.59983844e-01 -1.33721486e-01
7.78370917e-01 1.11128107e-01 -2.98633546e-01 7.02471212e-02
-1.34474063e+00 8.24671030e-01 6.36031210e-01 3.71641070e-01
-3.95166993e-01 3.19461018e-01 5.99108040e-01 2.67414868e-01
1.61955375e-02 7.74582446e-01 -5.85262656e-01 -1.41159832e-01
-6.38668358e-01 4.22148854e-01 6.25191092e-01 9.12739336e-01
3.97195548e-01 -3.78331877e-02 -3.54565978e-01 1.05023575e+00
9.90968756e-03 1.03550442e-01 1.00176299e+00 -7.21077383e-01
3.37637901e-01 2.62695104e-01 5.47875836e-03 -8.62911701e-01
-5.95175028e-01 -4.68812257e-01 -1.02162039e+00 -1.64180353e-01
5.92732668e-01 -1.78464353e-01 -8.77998412e-01 1.92816162e+00
-2.73555610e-02 3.05180430e-01 -7.66168348e-03 9.25037742e-01
6.81748748e-01 1.07728493e+00 6.57891482e-02 -1.95689961e-01
1.22201204e+00 -1.50776660e+00 -7.34995127e-01 -2.57824808e-01
6.01273596e-01 -5.14360607e-01 1.19942093e+00 4.64218855e-01
-1.08556974e+00 -7.00881243e-01 -1.08886230e+00 -4.15843576e-01
-3.29351604e-01 -3.25358868e-01 3.81394178e-01 2.61294782e-01
-6.06230974e-01 7.12539434e-01 -5.12204766e-01 -1.60066769e-01
1.14879534e-01 3.27660680e-01 2.25139293e-03 9.14109498e-02
-1.42996705e+00 7.68130720e-01 6.51355922e-01 1.40670687e-01
-9.08566892e-01 -6.41395628e-01 -7.36650527e-01 4.52218920e-01
1.37761310e-01 -5.02122462e-01 1.38912261e+00 -1.12429357e+00
-1.68097067e+00 5.46102107e-01 8.62755477e-02 -5.92298031e-01
2.39821836e-01 -3.86915833e-01 -3.34685706e-02 -2.17543215e-01
-3.54777992e-01 8.54693830e-01 6.46973610e-01 -7.30788410e-01
-2.61786133e-01 -1.13148980e-01 -2.09774122e-01 2.24808484e-01
-2.00969324e-01 1.67146713e-01 -4.44177806e-01 -1.01408887e+00
-4.45988566e-01 -9.20806527e-01 -5.61405599e-01 -4.70017105e-01
-4.07905877e-01 -3.62603128e-01 8.03157032e-01 -5.68394065e-01
1.31492341e+00 -2.00382447e+00 5.97286582e-01 -9.10557061e-02
2.38425080e-02 4.57152009e-01 -6.46815717e-01 5.58533967e-01
-2.93991804e-01 9.13931206e-02 -5.00391960e-01 2.29162406e-02
-2.80873299e-01 2.83458292e-01 -4.01238769e-01 2.72053242e-01
1.83013633e-01 1.19824147e+00 -1.04025984e+00 -2.91362286e-01
-2.18058571e-01 3.18627596e-01 -7.33436346e-01 5.23928881e-01
-8.13405395e-01 2.58722901e-01 -9.63401496e-02 2.83078223e-01
3.52986574e-01 -5.90731382e-01 4.28639323e-01 1.75298944e-01
7.31771290e-02 2.68145829e-01 -7.01592743e-01 1.82393265e+00
-1.43726379e-01 7.48598754e-01 -2.00297669e-01 -1.12727940e+00
9.38515961e-01 2.46151254e-01 3.75207454e-01 -4.05735880e-01
2.87393630e-01 -2.31053084e-02 6.61723852e-01 -7.58720160e-01
4.47095037e-01 1.73747420e-01 -1.24001145e-01 6.56287313e-01
-4.49650697e-02 1.68725580e-01 -7.68579915e-02 3.28536406e-02
1.18936229e+00 -3.44203003e-02 3.29053074e-01 -1.81190401e-01
4.64391112e-01 -3.77136111e-01 6.68306470e-01 8.59180987e-01
-2.05793068e-01 6.52208447e-01 9.23500896e-01 -6.24255419e-01
-1.42950964e+00 -4.69631821e-01 1.47090957e-01 1.67549479e+00
-8.00576434e-02 -3.18137646e-01 -9.02325928e-01 -6.19155467e-01
-1.63637504e-01 4.41357046e-01 -8.82786572e-01 -3.14016283e-01
-8.78864467e-01 -8.08793604e-01 9.26775992e-01 7.38317132e-01
1.12105429e-01 -1.69672585e+00 -5.24299383e-01 4.26652819e-01
-1.88613981e-01 -6.59503937e-01 -9.11400259e-01 5.51801860e-01
-6.04835391e-01 -1.00038803e+00 -9.56169069e-01 -1.01559997e+00
2.67835110e-01 -7.05254376e-02 8.90841067e-01 2.20335543e-01
-1.45985484e-01 -4.67325479e-01 -3.41080070e-01 -4.86184180e-01
-2.84309596e-01 7.73375928e-01 -6.48747683e-02 -1.54387370e-01
4.13488477e-01 -7.18534172e-01 -4.08952624e-01 2.76071966e-01
-9.15535808e-01 3.31918858e-02 7.16110229e-01 1.25083458e+00
1.58439308e-01 -5.60060620e-01 9.68849301e-01 -9.60917234e-01
8.84118199e-01 -7.78484881e-01 -4.79824483e-01 1.99339435e-01
-2.84566104e-01 3.21777821e-01 9.96603012e-01 -9.52168286e-01
-5.80956399e-01 -5.11530191e-02 -3.73036385e-01 -4.34705973e-01
-1.36535898e-01 7.28237510e-01 -3.43522102e-01 3.36027920e-01
7.77156770e-01 4.44600672e-01 1.86171293e-01 -4.85759050e-01
4.02704507e-01 9.85829055e-01 5.85153222e-01 -6.03244960e-01
8.80707726e-02 -1.07463539e-01 -3.63274157e-01 -9.43508685e-01
-5.35491705e-01 -1.95004344e-01 -6.47869110e-01 1.53933793e-01
8.55969906e-01 -3.66560251e-01 -9.29180443e-01 8.60545754e-01
-1.44800198e+00 -8.16838264e-01 8.92763287e-02 2.75267065e-01
-5.90089679e-01 3.80713314e-01 -1.08809006e+00 -5.41479409e-01
-5.13833344e-01 -1.18817902e+00 6.83253050e-01 1.45710230e-01
-4.57578540e-01 -6.18676305e-01 3.62538427e-01 3.21167558e-02
4.14362997e-01 -4.69039194e-02 1.16141224e+00 -9.74755943e-01
-4.92311984e-01 6.79403916e-02 -6.24707155e-02 9.93938744e-02
-1.08053304e-01 1.82227999e-01 -8.86420250e-01 -1.39365107e-01
-3.40584368e-01 -5.53537965e-01 1.00510788e+00 5.02308428e-01
1.77297568e+00 -3.37267756e-01 -3.23435247e-01 8.40317011e-01
9.18061197e-01 6.34906769e-01 7.13027120e-01 3.27482492e-01
8.95968199e-01 4.49036390e-01 3.41500431e-01 1.53180063e-01
-3.31505015e-02 4.93858308e-01 5.32421768e-01 1.87922835e-01
3.19210023e-01 -2.45773289e-02 2.80742973e-01 1.08507967e+00
-8.72610658e-02 -5.78121543e-01 -9.36717331e-01 5.57876348e-01
-2.01501822e+00 -1.30070913e+00 7.82302022e-02 1.86756253e+00
1.03476715e+00 1.69533268e-01 2.55974114e-01 1.29313141e-01
5.65853775e-01 3.14582080e-01 -8.76506150e-01 -5.77306211e-01
9.03772376e-03 3.20821628e-02 3.98198038e-01 5.08872688e-01
-9.59246337e-01 9.65119481e-01 7.23539114e+00 7.41597414e-01
-1.32497036e+00 -2.25562260e-01 4.94533390e-01 -1.88067466e-01
-1.66638926e-01 -5.00824332e-01 -7.45804071e-01 7.37561166e-01
1.20912468e+00 3.75254750e-02 3.03590417e-01 8.75346959e-01
-6.84523359e-02 4.58836198e-01 -1.14500749e+00 6.89861834e-01
-5.56333028e-02 -1.52611744e+00 1.30924791e-01 -1.04481228e-01
4.76914674e-01 2.29871899e-01 3.98356207e-02 4.31484759e-01
4.50661242e-01 -1.41410065e+00 4.45499033e-01 4.55112070e-01
7.15696096e-01 -9.52922165e-01 7.07056046e-01 7.04050541e-01
-8.90492439e-01 -2.70344853e-01 -5.57279050e-01 -2.55630553e-01
7.70227835e-02 1.14343621e-01 -8.80313933e-01 1.14548743e-01
4.41041440e-01 8.12277734e-01 -2.28026897e-01 1.02662981e+00
-1.99432328e-01 7.57308245e-01 -4.41132933e-02 -4.95184392e-01
4.73019987e-01 -1.77115738e-01 2.97210604e-01 1.64017892e+00
4.02462035e-02 2.02624634e-01 5.73520213e-02 4.63176668e-01
-2.09384635e-01 6.39826357e-02 -5.83468914e-01 -2.42534742e-01
5.78098118e-01 8.62601638e-01 -5.96569061e-01 -4.52322096e-01
-3.87700617e-01 8.96953166e-01 7.17988729e-01 5.22145510e-01
-1.10162926e+00 -8.64503205e-01 9.59381104e-01 -1.73933774e-01
4.34843749e-01 -2.67104656e-01 -2.71106929e-01 -8.87800992e-01
-4.40928191e-01 -1.31790924e+00 5.11082470e-01 -7.11991608e-01
-1.20205235e+00 6.93103671e-01 -3.75198752e-01 -7.86958635e-01
-2.19062701e-01 -6.66862428e-01 -7.77431071e-01 9.20055568e-01
-9.82060611e-01 -8.80738020e-01 -8.48030001e-02 3.63820136e-01
8.09821963e-01 -1.70409471e-01 7.75087714e-01 2.97375232e-01
-7.89754152e-01 8.38676035e-01 2.76059449e-01 4.29102421e-01
6.35473073e-01 -1.15253043e+00 8.15151691e-01 4.83009160e-01
-1.13212308e-02 9.61824894e-01 6.22726798e-01 -5.83548427e-01
-1.04764462e+00 -1.15235019e+00 7.38479078e-01 -3.58980268e-01
3.79288942e-01 -5.75705707e-01 -1.57857847e+00 1.08691156e+00
6.45313144e-01 -2.01253459e-01 9.31783319e-01 1.65704489e-01
-5.11123717e-01 2.83500433e-01 -5.61126411e-01 6.82303369e-01
1.11064672e+00 -4.23688859e-01 -7.68886209e-01 2.30969250e-01
1.26439869e+00 -3.47485185e-01 -6.69224441e-01 2.56444454e-01
8.54502618e-01 -6.24855101e-01 9.45832908e-01 -1.38668060e+00
7.30913818e-01 -2.42820963e-01 5.43122776e-02 -1.38083589e+00
-8.95689309e-01 -5.65469086e-01 -2.96659201e-01 8.02765667e-01
4.29865301e-01 -3.35016757e-01 8.99170458e-01 1.57827765e-01
-3.06306571e-01 -8.76718521e-01 -4.85317618e-01 -7.86923110e-01
3.28245431e-01 -2.93348253e-01 8.21923375e-01 1.22407699e+00
2.55502880e-01 7.39549518e-01 -7.54858196e-01 6.92174351e-03
3.41487527e-01 9.57345366e-02 7.70922363e-01 -1.01425254e+00
-6.68179810e-01 -8.42423737e-01 -7.80829936e-02 -1.55607963e+00
1.85781971e-01 -8.48385394e-01 3.31969827e-01 -1.02361298e+00
1.93227053e-01 -3.45414072e-01 -4.80266869e-01 3.92293602e-01
-5.43403387e-01 -1.89691097e-01 -3.85535620e-02 3.25467512e-02
-4.63610888e-01 6.44938290e-01 1.14340830e+00 -2.00709745e-01
-1.66764945e-01 -1.79750755e-01 -5.04939377e-01 5.73630750e-01
9.27522242e-01 -4.25244093e-01 -3.38062853e-01 -5.46409488e-01
1.87032402e-01 1.26978280e-02 -2.08905473e-01 -7.05957055e-01
3.19131196e-01 -5.87477326e-01 3.52421224e-01 -6.93552196e-01
1.66115507e-01 -4.91469622e-01 3.63937728e-02 7.01312184e-01
-9.10467625e-01 1.21894464e-01 2.65664645e-02 4.98915523e-01
-1.43890386e-03 -2.61875659e-01 8.65612924e-01 -2.02126712e-01
-3.75473291e-01 4.12936389e-01 -7.10473299e-01 -4.31317929e-03
9.38243210e-01 8.37884173e-02 -3.63116115e-01 -2.46400997e-01
-5.30137002e-01 4.30641741e-01 2.54251778e-01 6.92491591e-01
4.17250961e-01 -1.26302302e+00 -6.74180686e-01 3.23233753e-01
-1.13735721e-01 2.18166724e-01 1.68334231e-01 3.84502351e-01
-6.26445055e-01 5.96077025e-01 -5.67156076e-01 -5.94023347e-01
-1.26343966e+00 1.19255662e+00 4.10710603e-01 -3.76996905e-01
-3.19862992e-01 1.25296509e+00 3.60347271e-01 -7.44350374e-01
5.77128589e-01 -5.34734011e-01 -3.99107307e-01 3.97268683e-02
7.75415599e-01 2.21713841e-01 -4.02998209e-01 -2.12369621e-01
-1.11314319e-01 5.28848886e-01 -3.76784354e-01 1.25531793e-01
1.23060012e+00 2.59891033e-01 -1.34020701e-01 5.71134865e-01
1.22105694e+00 -2.84728259e-01 -1.26862907e+00 -2.54150748e-01
2.75041133e-01 -6.95473328e-02 -5.42723358e-01 -4.22033846e-01
-7.99294531e-01 1.08031976e+00 -7.70219490e-02 1.29447326e-01
9.02839720e-01 -2.22187102e-01 9.10341501e-01 6.56963110e-01
-4.63677458e-02 -7.60066330e-01 2.32969895e-01 1.14260900e+00
1.04662764e+00 -1.08023846e+00 -5.01051843e-01 6.65810406e-02
-4.57085311e-01 1.03901494e+00 9.58713770e-01 -5.37590906e-02
3.73834550e-01 4.35677230e-01 1.57344773e-01 6.15098774e-02
-1.27612805e+00 8.96425769e-02 -2.63972804e-02 5.71494997e-01
7.01831698e-01 -1.11638501e-01 -3.41428965e-01 7.97005296e-01
-9.08477753e-02 -9.03423727e-02 4.25303102e-01 7.04543829e-01
-6.76567018e-01 -1.15153587e+00 -4.18677002e-01 3.34505707e-01
-5.60714364e-01 -2.02913091e-01 -4.16920781e-01 4.95446354e-01
-1.14324316e-02 3.25375557e-01 2.48344779e-01 -4.84881371e-01
-2.54729036e-02 4.17761475e-01 2.90708214e-01 -5.14445066e-01
-9.26094830e-01 3.42607126e-02 -1.76173031e-01 -3.26560587e-01
8.19854811e-02 -4.72313911e-01 -1.33120871e+00 -1.69478163e-01
-1.03545390e-01 2.89714038e-01 2.11948037e-01 7.10191488e-01
3.93529624e-01 8.60019386e-01 2.89647132e-01 -8.27732325e-01
-6.11831248e-01 -1.41593277e+00 -2.66146868e-01 4.39924955e-01
6.49834633e-01 -2.35492468e-01 4.64685634e-02 2.21221939e-01]
|
[10.807540893554688, 6.863028049468994]
|
2cda35db-be90-43a5-8d1e-3b74134a854a
|
multiple-manifolds-metric-learning-with
|
1805.11918
| null |
http://arxiv.org/abs/1805.11918v1
|
http://arxiv.org/pdf/1805.11918v1.pdf
|
Multiple Manifolds Metric Learning with Application to Image Set Classification
|
In image set classification, a considerable advance has been made by modeling
the original image sets by second order statistics or linear subspace, which
typically lie on the Riemannian manifold. Specifically, they are Symmetric
Positive Definite (SPD) manifold and Grassmann manifold respectively, and some
algorithms have been developed on them for classification tasks. Motivated by
the inability of existing methods to extract discriminatory features for data
on Riemannian manifolds, we propose a novel algorithm which combines multiple
manifolds as the features of the original image sets. In order to fuse these
manifolds, the well-studied Riemannian kernels have been utilized to map the
original Riemannian spaces into high dimensional Hilbert spaces. A metric
Learning method has been devised to embed these kernel spaces into a lower
dimensional common subspace for classification. The state-of-the-art results
achieved on three datasets corresponding to two different classification tasks,
namely face recognition and object categorization, demonstrate the
effectiveness of the proposed method.
|
['Kai-Xuan Chen', 'Xiao-Jun Wu', 'Rui Wang', 'Josef Kittler']
|
2018-05-30
| null | null | null | null |
['object-categorization']
|
['computer-vision']
|
[ 1.32938363e-02 -2.48514667e-01 -2.87257023e-02 -5.75802863e-01
-3.67397904e-01 -3.47927451e-01 5.80534339e-01 -5.39860010e-01
-2.65607595e-01 1.84802458e-01 -6.97234720e-02 -7.88514391e-02
-4.46675658e-01 -3.62682611e-01 -1.42208695e-01 -9.67759609e-01
-1.87489450e-01 -2.13397592e-01 -1.15244955e-01 1.69516150e-02
4.24521118e-01 6.35389924e-01 -1.71975529e+00 -2.12251484e-01
9.04279888e-01 9.43805099e-01 4.32439148e-02 4.35282081e-01
7.58785754e-02 4.72416937e-01 -5.78161515e-02 -3.98709118e-01
2.58333057e-01 -4.82700735e-01 -8.40105653e-01 6.79519296e-01
2.60955721e-01 4.62468714e-03 -8.41838598e-01 1.40546608e+00
2.91191563e-02 1.25816062e-01 1.12114656e+00 -1.44353640e+00
-1.01183045e+00 -4.36574630e-02 -3.73568654e-01 1.02009200e-01
9.18678101e-03 -2.91354984e-01 8.52130234e-01 -1.37482226e+00
2.28448525e-01 1.40359938e+00 2.43556961e-01 4.91248459e-01
-1.14995110e+00 -2.27033302e-01 -2.73718715e-01 4.38592792e-01
-1.78949571e+00 -3.01357895e-01 9.90401626e-01 -6.99430227e-01
2.48411953e-01 5.34650624e-01 1.48996443e-01 4.55350667e-01
3.25199395e-01 6.11169040e-01 1.17669034e+00 -2.45174468e-01
-4.08341624e-02 4.70304102e-01 3.05995762e-01 8.91505480e-01
2.80952871e-01 -2.32611418e-01 -1.71050951e-01 -7.75995329e-02
6.81594610e-01 3.32950681e-01 -3.70947182e-01 -6.62099004e-01
-1.32314885e+00 8.43313098e-01 3.77874792e-01 6.43569052e-01
-2.32226074e-01 -4.33990031e-01 1.80355668e-01 2.83944577e-01
4.51296777e-01 4.83425036e-02 7.90132284e-02 1.11637898e-01
-2.86896735e-01 -3.31649810e-01 7.14775383e-01 1.01865447e+00
9.47206378e-01 -1.42102405e-01 1.02203719e-01 8.24243844e-01
7.24694848e-01 5.47406733e-01 3.93383712e-01 -5.63064933e-01
4.54458952e-01 9.93962824e-01 -6.37796596e-02 -1.58323252e+00
-3.47563207e-01 -4.51513566e-02 -1.06853426e+00 2.27417257e-02
1.72414064e-01 3.31328154e-01 -4.55860376e-01 1.48476458e+00
4.09656584e-01 3.61433297e-01 4.09814090e-01 1.11368525e+00
5.11634767e-01 3.51961941e-01 -1.79961100e-01 2.10155755e-01
1.18458033e+00 -4.94726926e-01 -5.99032998e-01 4.44770932e-01
9.84794557e-01 -7.41176307e-01 9.48769331e-01 1.39569715e-01
-6.01889968e-01 -6.35130823e-01 -1.46967924e+00 6.83533773e-02
-5.54178417e-01 6.50473118e-01 5.49465239e-01 9.48701918e-01
-8.31615627e-01 9.29314554e-01 -6.86286151e-01 -3.16444695e-01
3.23670924e-01 2.89409786e-01 -7.91094482e-01 -3.10205162e-01
-9.97521996e-01 8.10686767e-01 1.49661437e-01 5.30538797e-01
-6.61416352e-01 -1.80171415e-01 -9.47372496e-01 -3.16231221e-01
-1.71123669e-01 -1.78982034e-01 4.83717293e-01 -6.26430690e-01
-1.48139679e+00 1.16205680e+00 1.22177906e-01 2.60948062e-01
1.56183094e-01 1.08728306e-02 -8.25628161e-01 2.40728363e-01
-1.47506803e-01 5.35292812e-02 1.17489636e+00 -1.01908779e+00
-4.45991337e-01 -9.11824226e-01 1.50184736e-01 2.91157991e-01
-1.06588459e+00 6.78850859e-02 -1.24334125e-02 -4.98765111e-01
5.82597494e-01 -1.20057023e+00 7.89940059e-02 -2.43028998e-01
-3.84996742e-01 -3.93197060e-01 1.10707593e+00 -6.68188155e-01
1.18219399e+00 -2.38939166e+00 8.92018259e-01 1.35310322e-01
8.42627957e-02 1.49392620e-01 -2.47908726e-01 1.53401345e-01
-3.20784807e-01 1.62804663e-01 -3.46456587e-01 -4.45117682e-01
5.34948744e-02 2.40759566e-01 -1.81390837e-01 9.87033308e-01
5.65682709e-01 3.58471006e-01 -7.78120279e-01 -3.84226531e-01
2.67447472e-01 5.38851202e-01 -2.46976018e-01 3.80303293e-01
6.70168340e-01 7.01232493e-01 -6.39693499e-01 2.11159706e-01
1.14913797e+00 1.40543543e-02 -3.38107273e-02 -3.58921826e-01
-1.47891454e-02 -1.94009580e-02 -1.37665880e+00 1.76368272e+00
-1.91314042e-01 3.08736861e-01 3.61312442e-02 -1.37675762e+00
9.70574975e-01 1.95292085e-01 5.34672558e-01 4.84581627e-02
3.03966582e-01 3.38467628e-01 1.39646932e-01 -6.35456920e-01
2.47561887e-01 1.13414042e-01 1.52095571e-01 3.89321119e-01
2.06927702e-01 -1.18764117e-01 1.25417501e-01 -4.21807468e-02
8.14464688e-01 -2.30461173e-02 -1.46794811e-01 -7.98657894e-01
1.38793159e+00 -4.22185451e-01 4.24395949e-01 -7.09656812e-03
4.95429244e-03 7.12650359e-01 1.88485548e-01 -1.97910786e-01
-6.80857241e-01 -1.14230156e+00 -7.02158928e-01 5.06695926e-01
2.55300492e-01 -2.27894858e-01 -1.03517246e+00 -8.80186498e-01
-5.55324554e-02 2.80414939e-01 -5.61460078e-01 -7.80475199e-01
-2.15248957e-01 -8.41307282e-01 3.86463672e-01 2.12017775e-01
8.02741110e-01 -3.73697072e-01 5.63666560e-02 -1.32515892e-01
2.37271860e-01 -1.13620043e+00 -8.51457894e-01 -3.49844158e-01
-1.11629593e+00 -1.42733335e+00 -5.99758506e-01 -8.99090469e-01
1.08224559e+00 9.04449284e-01 4.51518953e-01 9.01009962e-02
-5.83020329e-01 8.49126399e-01 -3.69748682e-01 2.59658266e-02
-3.93540502e-01 -1.89178184e-01 8.26256812e-01 9.46964562e-01
5.95175683e-01 -4.60535496e-01 -4.98928994e-01 8.83206248e-01
-1.40716779e+00 -2.92370319e-01 5.81813395e-01 7.78577387e-01
1.49277285e-01 1.09357111e-01 5.11902809e-01 -3.47692996e-01
3.17212939e-01 -5.27592719e-01 -5.41477859e-01 2.18860626e-01
-3.90413940e-01 2.98131406e-01 4.85062927e-01 -1.89777046e-01
-7.38453805e-01 -1.56353548e-01 4.29910153e-01 -6.45915568e-01
-6.82409033e-02 4.98306692e-01 -5.64073622e-01 -5.12062073e-01
3.60551894e-01 3.87765169e-01 4.77172494e-01 -7.84859002e-01
4.89755213e-01 1.15866959e+00 1.12898842e-01 -3.90282452e-01
1.03131902e+00 6.87827587e-01 2.83379048e-01 -1.04778290e+00
-7.97811210e-01 -5.93024731e-01 -1.42387831e+00 -7.73476362e-02
1.06838942e+00 -5.79254091e-01 -4.89258379e-01 7.89164960e-01
-9.35221255e-01 4.82107759e-01 1.67397276e-01 9.32166517e-01
-6.94097400e-01 8.71391177e-01 -4.80396241e-01 -7.17852294e-01
-3.78001481e-02 -1.09768987e+00 7.55604506e-01 2.31988236e-01
3.95016402e-01 -1.35231721e+00 -5.92983440e-02 2.40804270e-01
9.22580585e-02 -6.12325408e-02 9.92118895e-01 -4.69171852e-01
-4.35306609e-01 -3.96451890e-01 -2.11115777e-01 1.07171035e+00
8.11139882e-01 -1.03179738e-01 -7.66100883e-01 -5.24177670e-01
5.82514942e-01 -3.11813429e-02 5.52152812e-01 -2.18570560e-01
1.10750866e+00 -1.39298871e-01 -2.97219038e-01 5.63690603e-01
1.16621411e+00 1.22176804e-01 5.95455945e-01 -3.06257047e-02
9.10029411e-01 8.66107166e-01 7.20338881e-01 2.61601418e-01
2.97210455e-01 5.55986285e-01 3.55598360e-01 9.14162844e-02
3.16981524e-01 8.50440338e-02 4.13152754e-01 1.38640082e+00
-1.91538796e-01 4.64189589e-01 -5.44277191e-01 3.19183916e-01
-1.80623400e+00 -6.36554778e-01 -2.04011604e-01 2.65604901e+00
4.63098317e-01 -2.27590442e-01 -1.35100484e-01 2.90436238e-01
9.96727169e-01 4.20411080e-02 -4.07246262e-01 -2.97089685e-02
-2.25493759e-01 -2.79350400e-01 2.12221265e-01 1.44561946e-01
-1.45532620e+00 5.99843264e-01 5.71509504e+00 7.30784833e-01
-1.13347232e+00 -4.94238874e-03 4.11552787e-01 5.59680939e-01
7.83940479e-02 -3.01488806e-02 -5.75992703e-01 2.78283089e-01
9.10844147e-01 -2.08885893e-01 4.33685094e-01 8.32131803e-01
4.96521406e-02 1.35414347e-01 -1.40713406e+00 1.51185787e+00
4.09965158e-01 -8.20033610e-01 3.17684561e-01 4.01207179e-01
6.28821611e-01 -5.99587679e-01 6.55429184e-01 2.40446553e-01
-3.21691036e-01 -1.05756187e+00 1.67995080e-01 7.90938020e-01
7.51627505e-01 -7.25890279e-01 6.10992134e-01 2.40599528e-01
-1.15598297e+00 8.36685672e-02 -8.13255131e-01 2.52071247e-02
-4.51647371e-01 3.23092043e-01 -5.10411084e-01 9.47078228e-01
5.06821990e-01 1.19690406e+00 -8.23530555e-01 9.03536499e-01
2.36560434e-01 3.27423722e-01 -1.86390672e-02 1.16400318e-02
3.25003207e-01 -1.24387538e+00 8.16035926e-01 6.92666113e-01
3.82941753e-01 1.90695316e-01 -2.41810102e-02 8.90382111e-01
-1.80275328e-02 4.04356182e-01 -9.16756451e-01 -4.20603156e-01
-1.60743937e-01 1.80902922e+00 -3.35558861e-01 -2.00913861e-01
-7.10791826e-01 1.10884750e+00 2.05120206e-01 1.90439373e-01
-7.62789488e-01 -7.04757094e-01 1.12898004e+00 -2.38114655e-01
1.02315441e-01 -6.59201920e-01 1.16369218e-01 -1.53812754e+00
1.54216379e-01 -4.36743647e-01 2.74375498e-01 -2.30627716e-01
-1.37150037e+00 5.77083468e-01 -1.42885670e-01 -1.75521505e+00
8.64622071e-02 -1.15173995e+00 -5.57530642e-01 8.97138417e-01
-1.14692199e+00 -8.70094240e-01 -2.24470153e-01 8.97136867e-01
4.58171889e-02 -4.68228936e-01 8.71106923e-01 5.08142591e-01
-7.15127885e-01 4.08506811e-01 7.22451210e-01 3.98251563e-01
4.08683568e-01 -1.33348286e+00 -2.35579997e-01 6.33107543e-01
2.50257969e-01 7.80797660e-01 1.73548415e-01 -1.24450866e-02
-1.94843376e+00 -1.16399169e+00 3.33848029e-01 -5.45869648e-01
7.05832839e-01 -3.57108295e-01 -1.03824544e+00 4.50845510e-01
-2.54552327e-02 8.95940214e-02 9.24451590e-01 -2.81211615e-01
-4.35856432e-01 -2.82492697e-01 -1.08104706e+00 6.71899855e-01
1.12148583e+00 -7.79747903e-01 -5.69401681e-01 6.10550404e-01
2.26051480e-01 1.45613387e-01 -1.33137834e+00 4.06988025e-01
1.20153166e-01 -6.88627899e-01 6.63029373e-01 -1.18605793e+00
-3.69459987e-02 -6.57268941e-01 -5.35019815e-01 -1.30716324e+00
-3.51508468e-01 -4.24172223e-01 1.46305531e-01 1.27461600e+00
-1.02637567e-01 -8.08914244e-01 5.33378780e-01 5.68610966e-01
-1.41519547e-01 -6.46926701e-01 -1.08151698e+00 -1.00075161e+00
2.28888635e-02 -2.01511439e-02 3.06213796e-01 9.25269961e-01
1.77523762e-01 4.00579751e-01 -2.03123167e-01 3.75994802e-01
9.12759840e-01 1.11513890e-01 6.21515810e-01 -1.39265692e+00
2.38228127e-01 -3.45866263e-01 -1.25745869e+00 -9.58923280e-01
5.19712865e-01 -1.32489967e+00 -2.47198492e-01 -8.78685176e-01
3.11130136e-01 -3.59743059e-01 -5.22525907e-01 -2.10093632e-01
-1.28614157e-01 2.46455669e-01 -1.20779909e-02 4.21513170e-01
-4.16966468e-01 1.05878079e+00 1.11725068e+00 -2.39664152e-01
3.21527496e-02 1.06213763e-01 -4.07646209e-01 7.55566597e-01
5.31282663e-01 -8.68749171e-02 -3.00591022e-01 -2.81513065e-01
-5.31248033e-01 -3.00220191e-01 2.84260094e-01 -1.22974873e+00
-7.03369826e-02 -2.70378087e-02 6.62332103e-02 -2.83638984e-01
5.19010663e-01 -9.88666117e-01 6.45385757e-02 3.32046092e-01
-9.72033143e-02 -6.90794885e-02 -1.30044892e-01 6.41773880e-01
-5.40898442e-01 -4.14081067e-01 1.15551031e+00 3.37882787e-01
-6.43200338e-01 6.67706847e-01 -5.01583032e-02 -2.67773300e-01
1.29691494e+00 -7.39535987e-02 7.63319433e-02 -9.31320488e-02
-7.70222783e-01 -3.32035720e-02 3.07036251e-01 9.00826931e-01
9.23687637e-01 -1.97551799e+00 -6.29106581e-01 2.44648248e-01
2.76031315e-01 -5.55342913e-01 3.23796391e-01 1.23879147e+00
-2.84036428e-01 6.81895316e-01 -3.13973993e-01 -8.54913473e-01
-1.25844669e+00 7.38659680e-01 5.24419963e-01 1.73632815e-01
-1.90405622e-01 3.91346574e-01 3.91448617e-01 -6.39906049e-01
-1.51976526e-01 -9.17593017e-02 -5.02503693e-01 -7.48505145e-02
5.82221091e-01 5.22033453e-01 3.83251198e-02 -1.26312590e+00
-4.82243210e-01 9.18996751e-01 -1.47918547e-02 -1.17603429e-02
1.11619031e+00 -3.39182675e-01 -4.75652754e-01 6.24741435e-01
2.03068542e+00 -3.66478711e-01 -9.29196417e-01 -4.55358446e-01
3.65089208e-01 -8.55501771e-01 -2.41197217e-02 3.33301634e-01
-9.47985768e-01 1.18482912e+00 7.85207212e-01 4.05318767e-01
6.78577304e-01 -1.23711653e-01 3.17378014e-01 3.96596700e-01
6.16367877e-01 -1.01629734e+00 1.73223719e-01 4.48732257e-01
1.18294311e+00 -1.22740936e+00 -3.08932215e-01 -7.05317974e-01
-3.83879125e-01 1.40644729e+00 4.54803884e-01 -3.01389217e-01
1.12750053e+00 -7.61889458e-01 -1.14006974e-01 1.63096357e-02
-1.85994839e-04 -1.37815028e-01 7.06426501e-01 4.61316735e-01
3.20974410e-01 7.85813183e-02 -4.49533135e-01 4.63565052e-01
6.42330870e-02 -4.41442579e-01 4.49575156e-01 6.80570781e-01
-3.47512931e-01 -8.98746550e-01 -4.86708820e-01 3.86321932e-01
-2.67268538e-01 4.13809776e-01 -1.75394937e-01 5.07766366e-01
-3.13298851e-01 1.21031523e+00 -1.20624132e-01 -7.92350829e-01
2.75066882e-01 -5.36364727e-02 5.26016235e-01 -6.60979748e-01
2.92243659e-01 -4.40341264e-01 -4.49033290e-01 -5.32075822e-01
-5.24163067e-01 -7.26159811e-01 -1.11172581e+00 -2.68246774e-02
-2.47847438e-01 4.89007026e-01 7.57714510e-01 8.82915616e-01
5.30410051e-01 -6.73775524e-02 1.40283728e+00 -9.52094436e-01
-1.12987840e+00 -1.10674119e+00 -1.28085041e+00 8.71798813e-01
-4.86341119e-02 -1.20867813e+00 -7.33745515e-01 -4.90372665e-02]
|
[7.9232707023620605, 4.077700138092041]
|
b41ccf89-afd1-4164-9104-84f4916f8acc
|
dynamic-bandits-with-an-auto-regressive
|
2210.16386
| null |
https://arxiv.org/abs/2210.16386v2
|
https://arxiv.org/pdf/2210.16386v2.pdf
|
Dynamic Bandits with an Auto-Regressive Temporal Structure
|
Multi-armed bandit (MAB) problems are mainly studied under two extreme settings known as stochastic and adversarial. These two settings, however, do not capture realistic environments such as search engines and marketing and advertising, in which rewards stochastically change in time. Motivated by that, we introduce and study a dynamic MAB problem with stochastic temporal structure, where the expected reward of each arm is governed by an auto-regressive (AR) model. Due to the dynamic nature of the rewards, simple "explore and commit" policies fail, as all arms have to be explored continuously over time. We formalize this by characterizing a per-round regret lower bound, where the regret is measured against a strong (dynamic) benchmark. We then present an algorithm whose per-round regret almost matches our regret lower bound. Our algorithm relies on two mechanisms: (i) alternating between recently pulled arms and unpulled arms with potential, and (ii) restarting. These mechanisms enable the algorithm to dynamically adapt to changes and discard irrelevant past information at a suitable rate. In numerical studies, we further demonstrate the strength of our algorithm under non-stationary settings.
|
['Djallel Bouneffouf', 'Negin Golrezaei', 'Qinyi Chen']
|
2022-10-28
| null | null | null | null |
['marketing']
|
['miscellaneous']
|
[ 1.31531417e-01 -2.45831907e-02 -5.63934326e-01 -1.42809704e-01
-7.07958579e-01 -1.18278193e+00 4.95464593e-01 8.93203095e-02
-7.23948598e-01 1.11127996e+00 -2.90079862e-02 -5.26070893e-01
-5.57675362e-01 -8.20367038e-01 -1.14846051e+00 -8.68912876e-01
-2.66569614e-01 6.05735421e-01 -1.06708203e-02 -2.31494203e-01
1.84908852e-01 5.19987643e-01 -1.05289495e+00 1.31858140e-01
5.89040279e-01 1.18588817e+00 -2.38958076e-02 8.18719387e-01
-1.96153179e-01 8.20370615e-01 -3.73374790e-01 -6.44749343e-01
5.59991181e-01 -3.53892058e-01 -7.33854473e-01 8.38828832e-03
-4.69892323e-01 -4.52584088e-01 -3.20166230e-01 1.07144356e+00
2.46132091e-01 2.47271001e-01 3.16516399e-01 -1.02716327e+00
-4.49744612e-01 1.10977089e+00 -8.28877985e-01 5.51007926e-01
1.55248418e-01 9.47033688e-02 1.20136321e+00 -8.38279575e-02
4.38744068e-01 1.13565004e+00 2.14323938e-01 6.91258848e-01
-1.35380638e+00 -3.55863243e-01 8.57968807e-01 -1.75692543e-01
-5.23067415e-01 -3.13715845e-01 8.72109890e-01 -1.18911959e-01
2.49372080e-01 6.35677874e-01 6.92942858e-01 1.23179877e+00
2.54001707e-01 1.21782708e+00 1.28601623e+00 -4.00434881e-01
6.96916699e-01 3.58018056e-02 9.74164829e-02 1.53965890e-01
3.05859208e-01 5.35337389e-01 -3.13809335e-01 -4.75053072e-01
5.55692852e-01 3.58641863e-01 1.85330142e-03 -5.35773754e-01
-8.79570484e-01 1.02507734e+00 2.58088738e-01 3.38948779e-02
-8.89053166e-01 4.37409699e-01 1.85719907e-01 8.32957149e-01
4.13353890e-01 2.25475982e-01 -5.42973459e-01 -2.34923258e-01
-5.74326336e-01 3.89881313e-01 8.69886696e-01 8.80627215e-01
2.76403159e-01 -2.68714428e-01 -4.86318290e-01 4.91186649e-01
-1.29400298e-01 5.42516768e-01 2.70110309e-01 -8.61607552e-01
7.72553265e-01 -7.49870855e-03 1.05892837e+00 -4.95559841e-01
-3.52696568e-01 -8.41718793e-01 -6.15602076e-01 1.05801590e-01
5.47667623e-01 -3.23642313e-01 -7.87394226e-01 1.91491377e+00
2.66477972e-01 -2.64768094e-01 -2.16803387e-01 9.07445550e-01
-2.34784335e-01 3.87268335e-01 -2.25581348e-01 -8.32468569e-01
8.63723814e-01 -8.98899317e-01 -5.99463820e-01 -3.90191227e-01
3.57575893e-01 -4.39755172e-01 7.97681212e-01 5.54888070e-01
-1.58446991e+00 2.76798993e-01 -6.15365505e-01 7.10718632e-01
1.54872881e-02 -6.16013288e-01 7.18629301e-01 8.51666152e-01
-5.81714988e-01 5.96435130e-01 -1.00932848e+00 2.29101881e-01
4.38268244e-01 3.84316206e-01 1.88862294e-01 4.88352440e-02
-9.27724659e-01 4.38686579e-01 -7.49672353e-02 3.95269394e-01
-1.14229524e+00 -4.41947192e-01 -1.27785191e-01 -5.60355447e-02
9.78224218e-01 -8.03398669e-01 1.71652603e+00 -1.42771995e+00
-1.73243463e+00 4.71885830e-01 -5.15748747e-02 -7.59016871e-01
1.22430706e+00 -2.98424512e-01 -6.71558380e-02 -1.89362824e-01
3.12904194e-02 -1.61848411e-01 7.44034052e-01 -1.14758575e+00
-8.47301424e-01 -4.98305827e-01 6.20559156e-01 4.28559929e-02
-5.12392037e-02 -1.08151481e-01 -8.84052068e-02 -6.91496372e-01
3.88851650e-02 -1.23668754e+00 -7.07835078e-01 -5.20528078e-01
-4.34754074e-01 2.78037369e-01 -3.19529325e-02 -4.86291498e-02
1.12189639e+00 -1.83767128e+00 1.93877324e-01 3.45936984e-01
-1.94874093e-01 -2.34659895e-01 -2.11839944e-01 6.33164942e-01
1.66638643e-01 2.77799964e-01 1.41609639e-01 -1.51735529e-01
1.51882008e-01 2.91710764e-01 -7.11612105e-01 5.65995038e-01
-3.24112624e-01 9.85270500e-01 -1.00412881e+00 1.67207971e-01
-2.44934291e-01 -4.78593439e-01 -5.37124157e-01 -6.95681050e-02
-5.38860083e-01 4.36480522e-01 -1.11330068e+00 6.97701514e-01
5.96497118e-01 -2.91104764e-01 1.40037671e-01 6.37846768e-01
-8.09648409e-02 2.03342736e-02 -1.22441125e+00 1.14748967e+00
-5.26281416e-01 -5.34346215e-02 4.26799506e-01 -1.06794751e+00
2.56819665e-01 -8.11941107e-04 4.04610157e-01 -8.12368751e-01
2.24054724e-01 2.56648213e-01 -1.91148221e-01 -1.88768864e-01
5.01859367e-01 -3.38434905e-01 -1.96910575e-01 6.42569363e-01
-7.42828488e-01 3.93326670e-01 1.61830068e-01 1.75048441e-01
1.40953660e+00 -2.36812994e-01 -2.24401932e-02 -2.76116077e-02
2.55800366e-01 -2.10465677e-02 7.32506633e-01 1.57594132e+00
-2.23641500e-01 5.17197959e-02 8.50320578e-01 -5.54731786e-01
-7.56904900e-01 -1.08348978e+00 2.92250842e-01 1.36900210e+00
3.22041005e-01 4.63259608e-01 -2.33921215e-01 -7.62196481e-01
5.30372262e-01 7.54245400e-01 -9.66761112e-01 3.24116722e-02
-3.31378013e-01 -8.90841007e-01 -1.82238445e-01 2.12542266e-01
1.31334081e-01 -8.83287668e-01 -6.48712158e-01 6.78092539e-01
1.84706375e-01 -6.88697755e-01 -5.31395197e-01 2.88801908e-01
-9.80805874e-01 -9.14824247e-01 -7.78870821e-01 -7.72125646e-02
4.83046055e-01 3.65970671e-01 1.01507211e+00 -3.62593085e-01
1.11174800e-01 6.68197811e-01 -4.13006395e-01 -4.30331826e-01
-2.41799966e-01 4.17402983e-02 1.74292121e-02 2.76072353e-01
-1.57231078e-01 -4.35762465e-01 -1.03070855e+00 3.69949698e-01
-1.03666663e+00 -1.45971999e-01 7.88974643e-01 9.71617818e-01
7.63625681e-01 -1.00838207e-01 7.86728680e-01 -1.05745673e+00
7.55644321e-01 -6.74028397e-01 -1.17013562e+00 4.62938398e-01
-5.59765339e-01 2.24799186e-01 6.08945251e-01 -8.65563810e-01
-1.02252686e+00 7.99636841e-02 4.88353014e-01 -3.69019151e-01
4.49272364e-01 5.13233304e-01 2.77854174e-01 -8.91883895e-02
2.92574704e-01 3.03265691e-01 -8.45921785e-02 -4.47694153e-01
5.04975438e-01 3.66086185e-01 2.16484100e-01 -9.31081772e-01
6.68485343e-01 7.69523680e-01 2.27740675e-01 -4.09672000e-02
-9.68970418e-01 -4.18695621e-02 1.19875334e-01 -1.66069075e-01
-1.51535757e-02 -6.72115862e-01 -1.39468563e+00 1.73562497e-01
-8.09097111e-01 -4.08846587e-01 -6.06469810e-01 5.24535596e-01
-9.43314075e-01 1.10930823e-01 -6.45940065e-01 -1.63287270e+00
-1.29824728e-01 -8.52718115e-01 5.24197400e-01 4.09448713e-01
4.40394431e-01 -6.37195826e-01 1.81929231e-01 2.68246382e-01
5.55106997e-01 3.20501387e-01 5.94569802e-01 -6.18017256e-01
-8.75039279e-01 -3.77650112e-01 2.40725473e-01 -4.32819799e-02
-6.85905963e-02 -4.55416799e-01 -4.04409438e-01 -5.02112746e-01
1.08103447e-01 -2.60028034e-01 7.21124351e-01 6.26740277e-01
1.25473368e+00 -8.26328218e-01 -5.23152351e-01 2.09526092e-01
1.24007857e+00 4.80365813e-01 1.34241059e-01 7.43091345e-01
-2.14639649e-01 4.69997615e-01 7.91230679e-01 8.84937704e-01
3.66291143e-02 4.73192155e-01 1.00825930e+00 3.96759450e-01
8.15676570e-01 -1.02648675e-01 3.63194078e-01 8.98079798e-02
-1.43628657e-01 -4.03040946e-01 -4.83639181e-01 6.27603948e-01
-2.17582726e+00 -1.13482404e+00 1.85508117e-01 2.65801430e+00
9.33549464e-01 5.37061512e-01 6.04034603e-01 -3.12321216e-01
7.55710721e-01 4.49598916e-02 -1.12707031e+00 -6.31252289e-01
-2.18996868e-01 5.15340343e-02 1.22454584e+00 4.26761091e-01
-9.13200021e-01 5.50287187e-01 6.56175804e+00 7.20441461e-01
-7.93153286e-01 1.86672017e-01 1.06922603e+00 -9.59201276e-01
-4.60506022e-01 2.24912912e-02 -5.33092439e-01 7.27864742e-01
9.02118385e-01 -4.27174896e-01 1.03030443e+00 9.76698577e-01
2.38934666e-01 -7.82871321e-02 -1.04745674e+00 5.77645481e-01
-7.29007423e-01 -1.43008590e+00 -3.43219459e-01 2.18013540e-01
9.02047396e-01 -1.07345087e-02 3.86157453e-01 3.24219346e-01
1.09849262e+00 -7.17713952e-01 1.00984502e+00 7.62808084e-01
3.10654998e-01 -1.08702385e+00 4.81197625e-01 5.76798975e-01
-6.87635899e-01 -7.21503437e-01 -1.22339927e-01 -1.46364003e-01
2.70462662e-01 6.22525454e-01 -4.70973402e-02 6.79151416e-01
4.04726297e-01 -9.02152359e-02 1.74943909e-01 1.08670270e+00
2.85150879e-03 5.56495905e-01 -4.90281105e-01 -4.43650335e-01
6.30424917e-01 -4.24790472e-01 7.22858012e-01 7.09341168e-01
1.75207317e-01 1.35599941e-01 1.05378591e-01 7.02757835e-01
-7.99018517e-02 -9.83487368e-02 -1.99108154e-01 -2.09703922e-01
4.17839676e-01 9.08748388e-01 -6.84832156e-01 2.97493283e-02
-3.61955076e-01 8.51179242e-01 1.70438707e-01 5.46877086e-01
-9.14897621e-01 -1.82896182e-01 5.78298926e-01 -1.62920002e-02
5.84569454e-01 5.34115322e-02 -1.52821302e-01 -9.51161802e-01
3.08634281e-01 -5.63390315e-01 6.39151335e-01 -2.62385607e-01
-1.50291526e+00 2.37759203e-01 -2.52650470e-01 -9.07254457e-01
-2.94940710e-01 -1.34695783e-01 -3.67147982e-01 6.55618370e-01
-1.56487930e+00 -7.45579123e-01 4.48243707e-01 4.56341058e-01
4.41388369e-01 1.11469015e-01 2.22710744e-01 -8.23317096e-02
-6.59845769e-01 6.37522578e-01 9.10529673e-01 -3.38125259e-01
1.03637278e-01 -1.18332756e+00 8.39577839e-02 6.17019534e-01
-1.07276358e-01 4.63667840e-01 9.62143660e-01 -5.08638263e-01
-1.89404035e+00 -9.52022433e-01 2.19143420e-01 -3.24867189e-01
1.12328315e+00 -1.97120801e-01 -3.56725186e-01 8.30682456e-01
-2.20948935e-01 2.58633673e-01 2.59003699e-01 2.32628301e-01
-1.69391260e-01 -4.54153419e-01 -1.22004485e+00 5.70725322e-01
1.13328052e+00 5.65033220e-02 -2.22436756e-01 5.53052843e-01
7.70324171e-01 -4.70417917e-01 -3.94448131e-01 2.05833137e-01
7.78527796e-01 -9.41729426e-01 7.18946576e-01 -1.21778357e+00
2.21295401e-01 2.65565366e-01 -1.53977692e-01 -1.20904195e+00
-3.59109581e-01 -1.46557081e+00 -4.19202119e-01 8.58828068e-01
5.40465593e-01 -1.00134015e+00 1.11298800e+00 6.81228042e-01
4.37015057e-01 -8.41544628e-01 -1.43329406e+00 -1.37170815e+00
2.46966675e-01 -3.05998176e-01 7.17168570e-01 3.96284103e-01
-1.63818792e-01 -1.73640430e-01 -5.82305133e-01 -1.09234815e-02
6.41571999e-01 6.92152321e-01 5.55789411e-01 -8.57418954e-01
-7.51629353e-01 -7.10277259e-01 2.52769262e-01 -1.29913938e+00
-8.22839886e-02 -3.63505125e-01 -6.53875768e-02 -9.54424083e-01
4.98729199e-01 -7.44464278e-01 -8.12299609e-01 1.51640385e-01
2.15104315e-03 -3.84521157e-01 1.79938078e-01 2.67583907e-01
-8.62023175e-01 5.91167271e-01 1.12830400e+00 -2.12392882e-01
-5.62566161e-01 8.77204835e-01 -8.77154648e-01 2.95175523e-01
6.56927526e-01 -6.16127610e-01 -2.36364499e-01 -2.12238684e-01
7.83651054e-01 7.23039508e-01 1.68496743e-01 -2.35678002e-01
1.46737337e-01 -9.25228536e-01 -8.57778415e-02 -6.00817680e-01
1.02064453e-01 -8.20956171e-01 3.30146223e-01 8.21146011e-01
-6.93727493e-01 1.13451883e-01 -1.04279444e-01 1.35076630e+00
2.89078176e-01 -2.35995039e-01 5.78759432e-01 -1.82480380e-01
9.41176564e-02 5.06320953e-01 -3.39239508e-01 -6.74033887e-04
1.19204354e+00 7.90589154e-02 -3.82937074e-01 -7.71235526e-01
-9.31269944e-01 6.16692543e-01 1.75590351e-01 1.79727808e-01
1.32597730e-01 -1.15199161e+00 -4.98481721e-01 -3.11613232e-01
-1.54229224e-01 -2.21639857e-01 4.31843966e-01 9.52499270e-01
1.63024070e-03 4.79859412e-01 1.72591999e-01 -1.97623357e-01
-8.01315486e-01 1.14677095e+00 2.83019781e-01 -6.17122650e-01
-2.51862794e-01 7.26863146e-01 6.89177141e-02 2.21949235e-01
3.57007653e-01 -1.03123046e-01 2.73231149e-01 6.84338957e-02
4.69622642e-01 4.40515697e-01 -1.29056945e-01 4.32252809e-02
-1.83300704e-01 -3.18220407e-02 -5.44473290e-01 -5.26444018e-01
1.52551425e+00 -4.07636136e-01 1.33921161e-01 5.14564276e-01
4.71746236e-01 1.92108676e-01 -1.35242641e+00 -5.09974241e-01
2.03867808e-01 -7.19527602e-01 -3.76826972e-02 -9.63698030e-01
-1.14279139e+00 2.90146083e-01 3.91648948e-01 8.71465325e-01
1.10840464e+00 -1.88344326e-02 6.35132909e-01 3.87006313e-01
6.16614342e-01 -1.20785642e+00 -5.21799661e-02 3.83470535e-01
7.48791158e-01 -9.96467948e-01 -2.03683391e-01 4.62278835e-02
-5.79566061e-01 8.45845520e-01 4.14782278e-02 -1.00548573e-01
4.27892178e-01 9.99153703e-02 -4.07167614e-01 1.76008329e-01
-1.08414185e+00 -2.80029535e-01 -2.16381222e-01 9.06546116e-02
-1.60512134e-01 3.42644185e-01 -6.05451047e-01 9.89808679e-01
1.38228154e-02 -3.60487401e-02 5.99295557e-01 1.23728001e+00
-4.84664321e-01 -1.12441468e+00 -4.84652936e-01 6.25998855e-01
-9.80664372e-01 2.69232422e-01 -3.10667843e-01 5.75397432e-01
-6.51449680e-01 1.07102454e+00 -4.14425433e-02 1.29903331e-01
3.72833312e-01 -2.86438733e-01 5.08044183e-01 -1.18547738e-01
-6.15013242e-01 2.18700007e-01 -6.18743077e-02 -5.79681516e-01
-1.80050746e-01 -7.31079578e-01 -5.72275698e-01 -5.48663616e-01
-3.62421662e-01 3.16362053e-01 7.04298973e-01 8.98092628e-01
3.86863768e-01 3.80504876e-01 1.38384402e+00 -3.75538707e-01
-1.59578454e+00 -6.28514230e-01 -8.16526592e-01 1.36291817e-01
6.45087659e-01 -5.79302967e-01 -5.81843019e-01 -5.41857719e-01]
|
[4.504827499389648, 3.2714035511016846]
|
dfc51034-97cf-4c60-8fad-547fdd931116
|
timeline-summarization-based-on-event-graph
| null | null |
https://aclanthology.org/2021.emnlp-main.519
|
https://aclanthology.org/2021.emnlp-main.519.pdf
|
Timeline Summarization based on Event Graph Compression via Time-Aware Optimal Transport
|
Timeline Summarization identifies major events from a news collection and describes them following temporal order, with key dates tagged. Previous methods generally generate summaries separately for each date after they determine the key dates of events. These methods overlook the events’ intra-structures (arguments) and inter-structures (event-event connections). Following a different route, we propose to represent the news articles as an event-graph, thus the summarization becomes compressing the whole graph to its salient sub-graph. The key hypothesis is that the events connected through shared arguments and temporal order depict the skeleton of a timeline, containing events that are semantically related, temporally coherent and structurally salient in the global event graph. A time-aware optimal transport distance is then introduced for learning the compression model in an unsupervised manner. We show that our approach significantly improves on the state of the art on three real-world datasets, including two public standard benchmarks and our newly collected Timeline100 dataset.
|
['Kathleen McKeown', 'Heng Ji', 'Tian Gao', 'Lingfei Wu', 'Mo Yu', 'Tengfei Ma', 'Manling Li']
| null | null | null | null |
emnlp-2021-11
|
['timeline-summarization']
|
['natural-language-processing']
|
[ 2.45648965e-01 2.95679599e-01 -4.20494378e-01 -3.00687432e-01
-8.68945062e-01 -7.03796387e-01 1.02021706e+00 1.24605155e+00
-1.72619164e-01 6.28361821e-01 1.38080335e+00 1.73845425e-01
-2.79533893e-01 -9.64406788e-01 -8.54723573e-01 -2.90761501e-01
-6.82585120e-01 5.14880419e-01 5.13597310e-01 2.29355264e-02
5.23550332e-01 2.46178731e-01 -1.31140900e+00 5.65816522e-01
7.19956398e-01 6.15337074e-01 -2.78766491e-02 5.84988713e-01
-4.60555375e-01 9.60355341e-01 -9.74598646e-01 -3.56399328e-01
-1.61943004e-01 -8.39619279e-01 -1.00199008e+00 2.07729071e-01
2.79930323e-01 -1.47837535e-01 -7.71832705e-01 8.56129527e-01
2.77893096e-02 5.12712657e-01 6.35451317e-01 -1.07260442e+00
-1.95616871e-01 1.27075207e+00 -1.02328956e+00 5.82504272e-01
5.51318526e-01 -5.49565911e-01 1.41312289e+00 -2.74457216e-01
1.22787678e+00 1.03716791e+00 5.81921101e-01 -8.88288468e-02
-9.92066503e-01 -1.90158952e-02 4.38896328e-01 2.28742152e-01
-1.04902840e+00 -1.63704351e-01 8.78388226e-01 -1.16323702e-01
1.19228446e+00 5.63940108e-01 9.21873748e-01 8.74146581e-01
4.82327610e-01 8.76393139e-01 3.94413084e-01 -2.31608137e-01
3.25729311e-01 -7.57282734e-01 4.81764585e-01 3.27762514e-01
2.09666237e-01 -3.86140317e-01 -9.72872198e-01 -3.92416030e-01
1.66197032e-01 2.35362768e-01 -3.65006775e-01 -1.24330431e-01
-1.64928651e+00 6.29775286e-01 3.52329016e-01 3.95254433e-01
-7.63714373e-01 1.52021363e-01 9.88341331e-01 1.64919093e-01
8.61300945e-01 2.21660107e-01 -2.36887723e-01 -4.30742092e-02
-1.47520733e+00 3.69292438e-01 1.00084841e+00 1.04326165e+00
5.00438511e-01 -2.86730468e-01 -4.78478670e-01 4.37157393e-01
-2.30717853e-01 -1.01023875e-01 3.20697367e-01 -5.22520900e-01
8.68368626e-01 7.77852893e-01 5.05950265e-02 -1.37813222e+00
-6.21659875e-01 -7.36120522e-01 -7.75327623e-01 -7.07222760e-01
1.84838742e-01 -4.86242138e-02 -5.91408789e-01 1.50822794e+00
4.53399897e-01 6.72442198e-01 2.31500082e-02 4.64810342e-01
9.51011837e-01 1.46491385e+00 4.28093486e-02 -8.63322675e-01
1.56442857e+00 -9.27222133e-01 -8.60178590e-01 3.26094367e-02
2.96924770e-01 -5.78152478e-01 7.01270282e-01 7.91548789e-02
-1.32407641e+00 -1.31972730e-01 -1.05062008e+00 -1.12863012e-01
-3.11827362e-01 -1.65509626e-01 4.52098459e-01 -2.10757762e-01
-5.81789613e-01 9.59679127e-01 -8.26274335e-01 -6.62562609e-01
1.78366870e-01 -3.49517584e-01 -9.89825204e-02 2.76439756e-01
-1.10300946e+00 4.84502673e-01 8.73410881e-01 -4.38220888e-01
-8.34707737e-01 -8.61626923e-01 -7.46918261e-01 3.40098053e-01
5.92850804e-01 -7.68674791e-01 1.26810563e+00 -4.63597447e-01
-7.63571143e-01 7.87163258e-01 -4.04337943e-01 -1.09024239e+00
3.87331992e-01 -2.16380134e-01 -6.73279405e-01 6.43563330e-01
3.00743908e-01 3.10328960e-01 6.98310375e-01 -1.08889806e+00
-1.18396413e+00 -1.95331261e-01 -2.95890011e-02 3.17222655e-01
-3.05572301e-01 2.47516319e-01 -7.66940355e-01 -1.11333072e+00
2.78063059e-01 -3.99538398e-01 -5.83478622e-02 -6.20074153e-01
-7.17004299e-01 -4.01841909e-01 8.23163688e-01 -7.88935900e-01
2.07033086e+00 -1.95177460e+00 3.70897561e-01 8.30387175e-02
3.67513686e-01 -5.69129884e-01 2.31172562e-01 1.10120952e+00
-1.01826742e-01 -2.27753185e-02 -4.62851167e-01 -3.11243802e-01
-5.06832749e-02 6.08864576e-02 -8.78977895e-01 5.40532112e-01
-2.19745055e-01 8.81937087e-01 -1.32265115e+00 -6.49664402e-01
-1.06245838e-01 6.59422651e-02 -2.39581004e-01 -4.96201999e-02
-4.62002397e-01 1.74004316e-01 -4.55590904e-01 2.07718655e-01
2.72246689e-01 -2.91258723e-01 1.26067884e-02 -4.99137014e-01
-4.75746900e-01 6.00961030e-01 -1.08078480e+00 1.93020511e+00
1.58940330e-01 7.54693627e-01 -5.57989478e-01 -1.10576034e+00
7.61867762e-01 2.81426102e-01 9.22509193e-01 -4.58864421e-01
-1.04998693e-01 -1.47144422e-01 -7.71901190e-01 -2.38582015e-01
1.17209125e+00 2.69599676e-01 -5.06040215e-01 6.49739265e-01
-1.08768716e-01 7.37880617e-02 6.72576964e-01 8.24823439e-01
1.14784801e+00 -2.32370049e-02 7.10237622e-01 -2.59867996e-01
6.35184795e-02 4.22108620e-01 4.30260986e-01 6.56767368e-01
2.00169772e-01 8.49751294e-01 1.00919700e+00 -5.40979981e-01
-1.07684290e+00 -1.07483721e+00 1.40519813e-01 8.36203933e-01
2.55028188e-01 -1.14198017e+00 -6.05952203e-01 -7.29400158e-01
-2.36550927e-01 1.08481026e+00 -7.65548527e-01 -2.17318349e-02
-8.59283745e-01 -6.25403881e-01 3.30239654e-01 2.91317493e-01
1.63405031e-01 -9.08270538e-01 -8.77991021e-01 6.54308915e-01
-6.81243300e-01 -9.50272143e-01 -8.58018517e-01 9.97627992e-03
-9.00900483e-01 -9.43160176e-01 -7.44019806e-01 -8.00420642e-01
6.48271978e-01 4.24827665e-01 1.34511209e+00 -3.53564501e-01
2.14827750e-02 3.58459085e-01 -5.13172567e-01 -4.24664736e-01
-1.96206734e-01 1.67856336e-01 -4.27415580e-01 9.24480557e-02
-1.18531197e-01 -7.45264769e-01 -6.71289384e-01 -2.76722945e-03
-1.02128744e+00 3.66786808e-01 1.41572831e-02 3.03116411e-01
9.86405134e-01 5.17811537e-01 6.32162690e-01 -9.64078784e-01
6.04294121e-01 -9.31219757e-01 -2.88822830e-01 3.64594609e-01
-2.15996146e-01 2.40002796e-02 6.86357081e-01 -1.80654079e-01
-1.07109320e+00 -3.07396859e-01 4.62493807e-01 6.55121133e-02
3.47109474e-02 9.83018219e-01 3.22304934e-01 9.67626452e-01
5.34874737e-01 4.70405966e-01 -8.99442375e-01 -3.74157846e-01
5.28471589e-01 1.05111785e-01 9.65892494e-01 -5.50889075e-01
5.23472428e-01 1.00130713e+00 -4.42277938e-02 -7.54651368e-01
-1.02733660e+00 -7.51348436e-01 -3.93442541e-01 -4.43251073e-01
6.99953794e-01 -8.64873767e-01 -1.29439831e-01 1.03151947e-01
-1.33335221e+00 2.49368995e-01 -7.60363996e-01 5.39303362e-01
-5.40568471e-01 6.00404501e-01 -6.46954775e-01 -3.30416203e-01
-4.76134837e-01 -3.20316285e-01 1.06689239e+00 3.05298209e-01
-6.47860646e-01 -8.63352895e-01 2.68946439e-01 -3.28528821e-01
-3.42499427e-02 9.05665636e-01 9.00880635e-01 -8.50068390e-01
-3.43137503e-01 -2.89676696e-01 -8.49227682e-02 -6.00041986e-01
2.03236833e-01 2.76500672e-01 -4.16505188e-01 -1.15542255e-01
-7.92457983e-02 2.94041336e-01 1.25181520e+00 7.63840675e-01
9.89065528e-01 -6.95693374e-01 -6.41389847e-01 3.84918720e-01
1.30373693e+00 1.28808141e-01 4.69110012e-01 3.41471732e-01
4.97912437e-01 8.36035430e-01 4.22406882e-01 6.97340369e-01
6.17500007e-01 3.04917306e-01 3.21114212e-01 1.39057353e-01
-1.10475466e-01 -6.65964961e-01 4.81038928e-01 9.38228071e-01
-1.32612064e-02 -7.11671948e-01 -7.03274190e-01 1.01815987e+00
-2.13928485e+00 -1.49090123e+00 -4.53087151e-01 2.09993982e+00
7.33453274e-01 4.07899261e-01 3.85503173e-01 1.35834098e-01
9.71976280e-01 8.93753648e-01 -2.55975425e-01 -1.51625261e-01
-3.88228893e-01 -2.13037565e-01 4.10663605e-01 2.03890741e-01
-1.23714590e+00 5.49301445e-01 6.08797979e+00 7.45021284e-01
-7.10068762e-01 -7.83641860e-02 6.39676392e-01 -3.28509301e-01
-5.93203902e-01 2.36662850e-01 -6.50932670e-01 4.04767871e-01
1.15696383e+00 -9.53414202e-01 -7.17844889e-02 5.06851673e-01
4.89881039e-01 -1.82779074e-01 -1.07030571e+00 6.26290202e-01
1.10625386e-01 -2.03010583e+00 3.71595263e-01 -3.72956634e-01
9.88493264e-01 -1.54132530e-01 -4.75094974e-01 -5.32248542e-02
1.29935965e-01 -2.25457013e-01 1.12931037e+00 5.86890161e-01
4.53276813e-01 -1.06367421e+00 3.41638416e-01 1.43682778e-01
-1.82258868e+00 3.05458844e-01 -1.12777263e-01 1.83406904e-01
6.67208433e-01 8.34018648e-01 -4.73990738e-01 1.24339926e+00
6.96184814e-01 1.34930634e+00 -3.07013601e-01 1.23071742e+00
-2.33764187e-01 9.70407426e-01 -3.07294369e-01 6.48597628e-02
3.76120746e-01 -2.69787818e-01 1.18473101e+00 1.72152150e+00
4.19214249e-01 1.97225049e-01 1.49528906e-01 3.53190750e-01
-2.53959924e-01 1.25253946e-01 -3.86999756e-01 -1.02339216e-01
5.18188179e-01 1.03418922e+00 -1.59476125e+00 -6.36596382e-01
-3.48071426e-01 8.82234991e-01 8.83981362e-02 2.92400360e-01
-1.10704565e+00 -4.82591510e-01 1.43381432e-02 8.82022530e-02
3.28010321e-01 -1.38843670e-01 -1.65728614e-01 -1.28267121e+00
2.00301111e-01 -4.25349951e-01 1.17884481e+00 -6.57032013e-01
-1.00920141e+00 5.61481535e-01 4.66335267e-01 -1.45233595e+00
-1.07912630e-01 5.98285913e-01 -9.77277339e-01 1.40300378e-01
-1.32223570e+00 -8.53608429e-01 -1.20541289e-01 5.12021959e-01
9.42545652e-01 1.77024260e-01 4.80055124e-01 6.91461489e-02
-5.03672242e-01 -4.61731441e-02 2.10377827e-01 -8.70313346e-02
5.36127925e-01 -1.29121804e+00 7.58288622e-01 1.27237082e+00
4.15924937e-01 3.17769468e-01 1.05933952e+00 -8.84396017e-01
-1.22490072e+00 -1.28720891e+00 1.42717695e+00 6.88107461e-02
1.01461112e+00 -1.03801154e-01 -8.67718756e-01 7.96862006e-01
7.43872643e-01 -5.69773018e-01 3.06420535e-01 -1.36909083e-01
-3.00682634e-01 -1.43102139e-01 -7.63528705e-01 6.80920064e-01
1.29210567e+00 -2.28182167e-01 -1.20278049e+00 6.66162789e-01
1.10320306e+00 -4.68322515e-01 -6.79454565e-01 -3.56038241e-03
4.67051044e-02 -6.57029808e-01 8.31331193e-01 -6.54589891e-01
8.60220790e-01 -4.27529126e-01 1.59150496e-01 -1.40377355e+00
-2.27118477e-01 -1.25279772e+00 -5.80289364e-01 1.62355447e+00
2.56995201e-01 -2.64088690e-01 4.54611838e-01 -2.57144660e-01
-4.67796445e-01 -3.65312696e-01 -9.66889858e-01 -5.15076697e-01
-6.47337377e-01 -3.24810386e-01 6.88755095e-01 1.05927253e+00
3.04375678e-01 4.31025565e-01 -4.38691229e-01 1.78630054e-01
7.09589303e-01 6.76085651e-01 3.84447306e-01 -1.29465926e+00
1.10142156e-01 -5.74772656e-01 -6.00823350e-02 -9.72355902e-01
-1.05760425e-01 -9.50313509e-01 -1.04956761e-01 -2.07603359e+00
3.73956800e-01 2.14782119e-01 -1.18763901e-01 1.45202324e-01
-1.85850143e-01 -4.86695975e-01 -4.82900590e-02 4.94399309e-01
-1.05414605e+00 4.89624888e-01 8.93618107e-01 -2.41629079e-01
-3.55969638e-01 -1.46469921e-01 -5.45393169e-01 7.37344325e-01
6.64613128e-01 -8.24063182e-01 -5.29929698e-01 -2.97963321e-01
5.02069175e-01 4.46829736e-01 7.26755559e-02 -9.37439263e-01
6.56638026e-01 -2.52198756e-01 2.17398582e-03 -1.33244872e+00
-1.78718716e-01 -7.11823285e-01 5.02869427e-01 3.58146459e-01
-7.12280095e-01 3.63162220e-01 1.66423753e-01 1.01608801e+00
-4.82612520e-01 3.10202707e-02 2.96620786e-01 9.42892116e-03
-7.90607393e-01 3.38184714e-01 -2.32126310e-01 3.10749382e-01
1.14245570e+00 -3.36903296e-02 -5.85593998e-01 -4.96020854e-01
-6.46702349e-01 3.42777669e-01 1.75694317e-01 2.09401429e-01
5.82273304e-01 -1.22569287e+00 -1.11682642e+00 -5.74831188e-01
1.75098464e-01 1.49786711e-01 3.57749403e-01 7.24447906e-01
-5.70158660e-01 1.72543034e-01 5.78391738e-02 -4.66852486e-01
-1.07691109e+00 6.71231925e-01 -2.18603790e-01 -6.02723658e-01
-1.46031857e+00 4.89289939e-01 1.60401136e-01 5.35618126e-01
4.66368467e-01 -7.33561456e-01 -5.45792103e-01 7.78626561e-01
8.27211320e-01 6.04145229e-01 -9.04774144e-02 -5.32445371e-01
-1.01606444e-01 4.08515006e-01 -7.81786591e-02 -3.27436887e-02
1.74920285e+00 -4.39705253e-01 -2.46423334e-01 6.36713505e-01
1.10652912e+00 2.68094569e-01 -1.08942986e+00 -4.04368997e-01
4.70947027e-01 -9.36563909e-02 -2.15814665e-01 -1.73483253e-01
-8.82741332e-01 8.40231255e-02 -4.26675826e-01 8.92675340e-01
1.34655774e+00 4.89817679e-01 1.12269068e+00 9.26130712e-02
6.23517781e-02 -1.07595468e+00 -6.77428320e-02 5.22958279e-01
1.13231885e+00 -4.54666018e-01 4.64026153e-01 -5.48581243e-01
-7.76252925e-01 1.29993331e+00 -2.60032147e-01 -3.11056226e-01
4.43251252e-01 6.44663274e-02 -7.16030180e-01 -5.65178335e-01
-8.68315637e-01 -3.13388146e-02 5.08849084e-01 9.09590945e-02
2.41233319e-01 9.10115167e-02 -7.66050220e-01 6.04983687e-01
-3.66999209e-01 -4.17594552e-01 6.53092682e-01 9.69550312e-01
-5.70759654e-01 -4.79499966e-01 -2.59438753e-01 5.07924438e-01
-6.05527699e-01 4.04381054e-03 -3.32960546e-01 6.55369520e-01
-3.06417227e-01 8.63368511e-01 4.90879953e-01 -1.77729465e-02
6.40759706e-01 -5.20414151e-02 2.74570566e-02 -5.29125750e-01
-7.18815088e-01 4.56109107e-01 3.94735932e-01 -4.96965647e-01
-6.64108574e-01 -1.03173566e+00 -1.71114898e+00 -4.02847290e-01
2.06178799e-01 4.80895847e-01 3.15754771e-01 6.91268086e-01
4.45241392e-01 9.52444792e-01 7.55609691e-01 -6.38803065e-01
7.15933219e-02 -5.62879443e-01 -3.96329522e-01 5.78888714e-01
3.69664490e-01 -1.11120261e-01 -3.32649618e-01 5.06239533e-01]
|
[12.570459365844727, 9.515050888061523]
|
188a45be-49dd-43ad-96ce-5c8ca0f1dbcd
|
hand-segmentation-for-hand-object-interaction
|
1603.02345
| null |
http://arxiv.org/abs/1603.02345v3
|
http://arxiv.org/pdf/1603.02345v3.pdf
|
Hand Segmentation for Hand-Object Interaction from Depth map
|
Hand segmentation for hand-object interaction is a necessary preprocessing
step in many applications such as augmented reality, medical application, and
human-robot interaction. However, typical methods are based on color
information which is not robust to objects with skin color, skin pigment
difference, and light condition variations. Thus, we propose hand segmentation
method for hand-object interaction using only a depth map. It is challenging
because of the small depth difference between a hand and objects during an
interaction. To overcome this challenge, we propose the two-stage random
decision forest (RDF) method consisting of detecting hands and segmenting
hands. To validate the proposed method, we demonstrate results on the publicly
available dataset of hand segmentation for hand-object interaction. The
proposed method achieves high accuracy in short processing time comparing to
the other state-of-the-art methods.
|
['Hung-Shuo Tai', 'Kar-Han Tan', 'Truong Q. Nguyen', 'Byeongkeun Kang', 'Nan Jiang', 'Daniel Tretter']
|
2016-03-08
| null | null | null | null |
['hand-segmentation']
|
['computer-vision']
|
[ 3.17493618e-01 -3.04726064e-01 4.28539664e-02 -1.58910915e-01
-2.58467346e-02 -5.68072319e-01 2.28187561e-01 -1.83056921e-01
-5.21567822e-01 6.95514202e-01 -2.37598553e-01 -2.90454179e-01
3.44975293e-02 -6.16303384e-01 -6.46435171e-02 -6.40437543e-01
4.01708513e-01 6.96916580e-01 7.39093721e-01 1.47498056e-01
3.44911695e-01 8.18461299e-01 -1.48500335e+00 1.77283973e-01
7.70676196e-01 8.08605015e-01 4.24978435e-01 6.03444159e-01
-1.86601043e-01 3.07940751e-01 -4.90093768e-01 -1.11304760e-01
3.50004524e-01 -5.59063137e-01 -9.16106999e-01 3.36927712e-01
-2.96948217e-02 -6.77919984e-01 5.44255674e-02 9.46500421e-01
9.90903139e-01 2.62379020e-01 8.47442091e-01 -1.10584927e+00
-1.31109893e-01 2.34993324e-01 -9.86387372e-01 -1.51608944e-01
5.58708489e-01 5.69215156e-02 3.77897680e-01 -6.77234471e-01
7.39024937e-01 1.23690474e+00 2.96942472e-01 6.79177582e-01
-8.22338104e-01 -5.85776567e-01 3.93503994e-01 1.90564796e-01
-1.38247192e+00 5.99644519e-02 7.06042051e-01 -7.09423184e-01
7.54527986e-01 1.99881524e-01 6.48354232e-01 5.92401206e-01
-1.26653686e-01 1.00470877e+00 1.42942107e+00 -6.30627155e-01
1.75611719e-01 1.27390563e-01 3.64946842e-01 8.99196148e-01
1.70526430e-01 -1.21582329e-01 -1.14494562e-01 9.21415165e-02
1.01128805e+00 2.35276222e-01 -2.39573300e-01 -1.92822427e-01
-1.29697061e+00 1.55953482e-01 4.74473059e-01 2.62962908e-01
-5.62186658e-01 -2.69237697e-01 1.57371774e-01 -2.71459103e-01
-1.23992106e-02 -2.18709990e-01 -3.52131903e-01 -4.99481037e-02
-8.84922028e-01 3.21841091e-01 8.58412504e-01 8.22840989e-01
2.21671805e-01 -4.72922742e-01 -3.19387674e-01 5.77222526e-01
4.97033298e-01 3.75652462e-01 4.58661020e-02 -4.12326008e-01
3.15263391e-01 6.27456009e-01 3.52317661e-01 -6.20341182e-01
-5.73492169e-01 2.56230474e-01 -7.66741276e-01 7.36000597e-01
9.08915281e-01 -1.24328487e-01 -1.31957829e+00 8.98434222e-01
6.99532032e-01 -2.83187360e-01 -2.26711318e-01 1.28749466e+00
1.03883386e+00 2.64778197e-01 8.26288760e-02 -3.01147193e-01
1.53983867e+00 -9.92790937e-01 -8.32687378e-01 -6.83982298e-02
1.10776342e-01 -9.28250968e-01 1.13699317e+00 7.68665314e-01
-7.40562439e-01 -4.64493245e-01 -7.81938076e-01 -1.47368396e-02
-2.95844674e-01 3.20064962e-01 8.08534980e-01 1.01749861e+00
-4.76202190e-01 4.57222074e-01 -7.64839292e-01 -6.94694698e-01
2.83742964e-01 5.67188919e-01 -1.89619660e-01 -1.93805486e-01
-5.38439870e-01 7.65417814e-01 4.01687771e-01 5.61321139e-01
-3.06306183e-01 2.23518535e-02 -2.84688413e-01 -3.49852413e-01
4.69265223e-01 -3.96433979e-01 1.17730904e+00 -5.77834070e-01
-1.96147883e+00 8.84703398e-01 -2.78040975e-01 2.31027082e-01
8.92924070e-01 -3.17910135e-01 4.20081839e-02 -4.91847433e-02
-3.20272774e-01 4.09725726e-01 6.97108924e-01 -1.42394698e+00
-5.50350666e-01 -7.79377341e-01 5.11742309e-02 2.16590881e-01
3.00032705e-01 2.75818467e-01 -6.54395521e-01 -3.59030873e-01
4.67277169e-01 -1.04156363e+00 -1.47277325e-01 3.60591449e-02
-8.24404776e-01 -2.74999619e-01 9.53470707e-01 -9.92180943e-01
9.92976785e-01 -1.91421366e+00 9.56452712e-02 4.14625078e-01
-5.42969592e-02 4.37976658e-01 2.52118289e-01 2.90981587e-02
3.19879532e-01 -1.73969030e-01 -2.20254734e-01 -1.72429323e-01
-2.22690538e-01 -3.24975364e-02 2.00346440e-01 3.78018469e-01
-3.25411201e-01 5.23681462e-01 -7.97172010e-01 -9.96743143e-01
3.95622879e-01 6.00294411e-01 -1.35533169e-01 3.54806483e-01
-2.16078550e-01 9.23877060e-01 -5.11530757e-01 1.14548516e+00
8.16658854e-01 3.04308772e-01 1.96682841e-01 -2.86987990e-01
-2.19336584e-01 -1.77370220e-01 -1.47428644e+00 1.60592055e+00
-2.02226251e-01 2.23881364e-01 1.85013860e-01 -3.48328173e-01
7.11193204e-01 4.84133810e-01 2.55727708e-01 -1.91733718e-01
5.78343093e-01 2.58367777e-01 1.49138689e-01 -7.05947697e-01
1.20173454e-01 -7.93581922e-03 6.09261692e-01 6.09470069e-01
-3.39662313e-01 -2.66859770e-01 5.73822930e-02 -2.55421370e-01
8.12357843e-01 4.63599056e-01 2.70360112e-01 9.61584672e-02
4.51173604e-01 -1.49530187e-01 3.24211001e-01 4.46325213e-01
-4.08010572e-01 7.27734685e-01 2.29382902e-01 -2.52248138e-01
-5.42110622e-01 -8.85469913e-01 2.28905454e-02 8.14405918e-01
4.06037658e-01 1.82046384e-01 -1.04146266e+00 -7.60125101e-01
-1.65343806e-01 2.87332118e-01 -3.27504098e-01 5.78289688e-01
-4.56391752e-01 -6.40873432e-01 3.62188406e-02 7.94420540e-01
8.76947284e-01 -1.39903605e+00 -8.48507285e-01 1.39650479e-01
-4.19374183e-02 -1.12220752e+00 -3.32499772e-01 -7.02450871e-02
-9.17087555e-01 -1.21597564e+00 -1.15555668e+00 -8.08221102e-01
8.96010280e-01 1.79492921e-01 7.36442387e-01 1.30489781e-01
-6.87318146e-01 2.14464501e-01 -4.82520819e-01 -5.39818466e-01
6.59544319e-02 1.52115002e-01 -3.70171368e-02 -1.94793925e-01
2.28367254e-01 -2.30622903e-01 -8.73251617e-01 5.82609653e-01
-5.35596967e-01 8.63060355e-02 6.97589457e-01 5.33434987e-01
5.86912632e-01 1.50190130e-01 8.35811421e-02 -8.54297936e-01
5.62938809e-01 8.93687829e-02 -5.14614761e-01 4.44820255e-01
-3.46222073e-01 -1.18772350e-01 -1.59779992e-02 -5.92608690e-01
-1.34354758e+00 7.83938646e-01 -4.29827571e-02 -1.16260182e-02
-5.44365644e-01 1.11452103e-01 -5.21747410e-01 -1.58360794e-01
4.79713708e-01 -6.20184019e-02 -9.87768546e-02 -6.74922049e-01
1.54371172e-01 1.17342389e+00 4.30660784e-01 -4.87524748e-01
4.52169061e-01 4.50512469e-01 1.09170660e-01 -8.69532645e-01
-3.49274278e-01 -5.54702103e-01 -1.27137983e+00 -3.81207228e-01
1.16014063e+00 -2.54942477e-01 -1.02990758e+00 8.62366915e-01
-1.44271410e+00 -4.56485301e-01 2.93706656e-01 6.16436124e-01
-2.08572045e-01 3.93729270e-01 -4.74500597e-01 -1.42907143e+00
-4.32257801e-01 -1.16780126e+00 9.82899725e-01 4.23530340e-01
-6.13068752e-02 -5.33970892e-01 -1.78316712e-01 3.61543477e-01
2.32193805e-02 3.62608314e-01 7.60564685e-01 -2.37171873e-01
-7.59476960e-01 -4.02692288e-01 -3.77960116e-01 6.89299256e-02
4.56212670e-01 1.20039120e-01 -1.03365433e+00 -5.43984643e-04
-4.39094365e-01 -4.84846309e-02 6.98932350e-01 4.79651272e-01
1.06763673e+00 1.71423987e-01 -6.41642570e-01 8.50501955e-02
1.15418220e+00 4.21881437e-01 6.95137143e-01 1.73370332e-01
8.88734698e-01 8.72474670e-01 9.40092742e-01 4.42882359e-01
4.40955684e-02 4.96504515e-01 1.60614610e-01 -2.69043237e-01
-1.29387245e-01 1.73512861e-01 -2.11360022e-01 3.10519516e-01
-8.85550380e-01 -3.57442468e-01 -9.83242929e-01 3.07048112e-01
-1.73122275e+00 -6.28786564e-01 -4.12173867e-01 2.18994761e+00
6.22523367e-01 2.59784013e-02 5.33450425e-01 6.25542819e-01
7.94589460e-01 -4.19158667e-01 -4.91756260e-01 -9.38168988e-02
3.38083595e-01 3.52819562e-01 4.74305302e-01 3.65882337e-01
-1.05879974e+00 9.63093638e-01 6.36792421e+00 3.22777122e-01
-1.02162683e+00 1.54467210e-01 5.77188656e-02 1.75011262e-01
2.75422603e-01 -4.14883345e-01 -7.16349483e-01 2.16097787e-01
-1.76652879e-01 5.43268919e-01 5.59872568e-01 8.21324766e-01
2.04065852e-02 -6.67289972e-01 -1.06690156e+00 1.25800574e+00
4.38524410e-02 -4.32874143e-01 -3.55948418e-01 -2.23593846e-01
3.96541417e-01 -5.44048846e-01 -2.25972533e-01 -1.69475719e-01
-1.72863121e-03 -9.20723498e-01 4.90429819e-01 3.32737654e-01
7.56415129e-01 -5.11487126e-01 8.80352318e-01 3.34773749e-01
-1.29361403e+00 2.07900077e-01 6.38765395e-02 -1.48055479e-01
2.83330768e-01 5.63363552e-01 -1.00780785e+00 2.02050164e-01
7.09119678e-01 1.81546658e-02 -2.81621963e-01 1.19985962e+00
-3.02626431e-01 1.32532761e-01 -3.90033364e-01 -9.70433578e-02
-3.19243282e-01 -1.74019173e-01 2.78246969e-01 1.17030919e+00
6.73315376e-02 5.61484814e-01 2.85085738e-01 7.54599512e-01
3.56299609e-01 2.29818821e-01 -3.44211876e-01 -1.08377911e-01
2.31364608e-01 1.13404953e+00 -1.54200304e+00 -2.26970717e-01
-3.25575948e-01 1.65062714e+00 -2.15311304e-01 4.29988146e-01
-4.52029467e-01 -6.78422511e-01 2.75572598e-01 3.38630199e-01
7.51562119e-02 -4.72720385e-01 -4.87552017e-01 -7.62974679e-01
2.77784944e-01 -6.38879955e-01 1.56861052e-01 -6.20152593e-01
-9.07607317e-01 7.08605528e-01 -6.02261797e-02 -1.01972747e+00
3.39947045e-02 -8.89911294e-01 -4.27584112e-01 9.59663570e-01
-1.17845786e+00 -1.45346594e+00 -1.02599740e+00 6.64233863e-01
5.79212666e-01 1.30624287e-02 7.95948327e-01 1.47009015e-01
-5.61112106e-01 3.36081654e-01 -3.69921058e-01 1.53507531e-01
7.31839657e-01 -1.21675587e+00 1.90339342e-01 6.56038821e-01
-2.60584980e-01 3.99960190e-01 4.97865945e-01 -9.11763668e-01
-1.21437359e+00 -5.13087571e-01 4.21842575e-01 -4.16724265e-01
-1.16652906e-01 -1.53080195e-01 -6.05445683e-01 5.46254039e-01
-1.26908310e-02 1.19955108e-01 4.31752264e-01 7.88243189e-02
-1.39402775e-02 1.81579024e-01 -1.66329479e+00 4.22660410e-01
1.18340969e+00 -2.48070210e-01 -3.47203851e-01 2.93804020e-01
-9.21799894e-03 -7.57614315e-01 -4.83437866e-01 4.01729524e-01
1.12135565e+00 -1.10227621e+00 8.89100730e-01 -1.36886716e-01
-1.69153467e-01 -5.88804066e-01 5.89980073e-02 -8.10637355e-01
-8.84707570e-02 -3.71065646e-01 5.18083125e-02 1.20285416e+00
9.69769135e-02 -4.29613382e-01 9.78691459e-01 8.88461947e-01
3.32034916e-01 -4.95459408e-01 -5.42283356e-01 -6.44070446e-01
-4.52060342e-01 -1.50924891e-01 3.97228688e-01 5.30720472e-01
3.40282284e-02 6.81509525e-02 -7.24047981e-03 2.64668763e-01
8.44413400e-01 2.18315482e-01 8.49049687e-01 -1.65025425e+00
-2.22280338e-01 -2.55852699e-01 -3.69367182e-01 -8.59812140e-01
5.94463712e-03 -3.47052187e-01 4.46286917e-01 -2.00125766e+00
4.20595706e-01 -5.92808366e-01 -2.96044983e-02 4.49990422e-01
-3.63863617e-01 2.79712230e-01 1.19515814e-01 2.65281081e-01
-2.42338449e-01 -1.03992641e-01 1.38253570e+00 -4.77533899e-02
-7.35936224e-01 3.11836571e-01 5.20196036e-02 8.23474884e-01
6.30356133e-01 -2.21862480e-01 -1.86443642e-01 -2.24598318e-01
-3.17431182e-01 -1.83028802e-02 3.74991328e-01 -1.05374706e+00
1.24302991e-01 -3.59431028e-01 4.79345232e-01 -8.63604605e-01
1.01133019e-01 -1.03973544e+00 6.78186268e-02 6.74351811e-01
1.95453197e-01 -3.53431582e-01 -8.74000415e-02 2.98455894e-01
9.77618843e-02 -1.98586285e-01 8.30101192e-01 -3.97971839e-01
-6.94005311e-01 1.12511314e-01 -2.81296879e-01 -5.36123276e-01
1.23481190e+00 -5.13351500e-01 7.71418437e-02 2.28653233e-02
-7.75884807e-01 1.53307458e-02 3.47895294e-01 2.27800146e-01
5.49619794e-01 -9.47986245e-01 -4.03662950e-01 2.83483088e-01
-2.25355461e-01 4.02494609e-01 1.34522036e-01 9.24086094e-01
-8.47314119e-01 2.76489586e-01 -4.78635371e-01 -5.43157697e-01
-1.84209955e+00 4.27585602e-01 2.04958603e-01 1.15801081e-01
-4.68247175e-01 9.20952201e-01 -1.09052978e-01 -3.39838356e-01
5.42074561e-01 -6.56832218e-01 -1.19439140e-01 -1.69512495e-01
4.83645290e-01 8.78076077e-01 8.21791813e-02 -4.38710123e-01
-6.52196825e-01 1.00323415e+00 1.33539483e-01 -2.82971799e-01
6.44271016e-01 -3.95597629e-02 -2.03122094e-01 3.39477956e-01
5.79567492e-01 2.84702629e-02 -9.77050781e-01 -1.37948878e-02
-2.21555322e-01 -7.62856781e-01 7.48312473e-03 -1.01486647e+00
-9.59271848e-01 1.09557319e+00 1.16859782e+00 -1.74729049e-01
1.09193289e+00 -5.14879227e-02 8.00030470e-01 4.65135664e-01
7.36941874e-01 -1.08897424e+00 -5.19419694e-03 1.87218025e-01
8.63455117e-01 -1.32965362e+00 1.41059563e-01 -9.57640767e-01
-7.22396135e-01 1.13802314e+00 7.80985534e-01 2.62119979e-01
9.18313146e-01 4.35610920e-01 3.09835672e-01 6.21773116e-02
2.42969483e-01 -6.61196053e-01 2.33202666e-01 7.72312045e-01
5.62119782e-01 1.74647540e-01 -3.98662120e-01 4.95208740e-01
1.51294276e-01 5.60434163e-01 -5.07244989e-02 1.36552179e+00
-3.21553677e-01 -1.14251256e+00 -6.47946119e-01 3.29635620e-01
-4.22270179e-01 3.13525558e-01 -6.14206016e-01 5.99065661e-01
3.50054204e-01 9.20095086e-01 -1.50365964e-01 -4.56572711e-01
4.49805081e-01 1.62629247e-01 1.05481029e+00 -5.96933067e-01
-4.57880199e-01 9.58500355e-02 -3.37629110e-01 -3.92451823e-01
-6.31103635e-01 -4.66537207e-01 -1.56063187e+00 -1.24633566e-01
-6.18681014e-01 -3.40743572e-01 9.72528160e-01 9.81105328e-01
-4.97958027e-02 3.42133284e-01 1.93326682e-01 -1.38477051e+00
-3.84517647e-02 -1.10831392e+00 -9.90346789e-01 1.43995672e-01
1.39147952e-01 -1.01873410e+00 3.80563214e-02 1.69146478e-01]
|
[6.558807849884033, -0.5581557750701904]
|
9ceaf47c-dca9-4a59-9f15-f2a89523ecc8
|
cose-co-sentence-conditioned-generative
| null | null |
https://openreview.net/forum?id=1yieqYLUIXj
|
https://openreview.net/pdf?id=1yieqYLUIXj
|
CoSe-Co: Sentence Conditioned Generative CommonSense Contextualizer for Language Models
|
Pre-trained Language Models (PTLMs) have been shown to perform well on natural language reasoning tasks requiring commonsense. Prior work has leveraged structured commonsense present in knowledge graphs (KGs) to assist PTLMs. Some of these methods use KGs as separate static modules which limits knowledge coverage since KGs are finite, sparse, and noisy. Other methods have attempted to obtain generalized and scalable commonsense by training PTLMs on KGs. Since they are trained on symbolic KG phrases, applying them on natural language text during inference leads to input distribution shift. To this end, we propose a task agnostic sentence-conditioned generative CommonSense Contextualizer (CoSe-Co), which is trained to generate contextually relevant commonsense inferences given a natural language input. We devise a method to create semantically related sentence-commonsense pairs to train CoSe-Co. We observe commonsense inferences generated by CoSe-Co contain novel concepts that are relevant to the entire sentence context. We evaluate CoSe-Co on multi-choice QA and open-ended commonsense reasoning tasks on the CSQA, ARC, QASC, and OBQA datasets. CoSe-Co outperforms state-of-the-art methods in both these settings, while being task-agnostic, and performs especially well in low data regimes showing it is more robust and generalises better.
|
['Balaji Krishnamurthy', 'Jivat Neet Kaur', 'Sumit Bhatia', 'Milan Aggarwal', 'Rachit Bansal']
|
2021-09-03
| null | null | null |
akbc-workshop-cskb-2021-10
|
['novel-concepts']
|
['reasoning']
|
[ 6.17135108e-01 5.85343063e-01 3.55212204e-02 -3.57125700e-01
-9.92549896e-01 -7.49146521e-01 8.82046223e-01 2.30272368e-01
-3.91290456e-01 8.94481182e-01 6.67683780e-01 -4.92325485e-01
-1.22694127e-01 -1.05065978e+00 -8.92968535e-01 -2.16430739e-01
3.88236970e-01 7.98905373e-01 2.80884832e-01 -7.27409840e-01
3.45725656e-01 1.06842846e-01 -1.11828804e+00 6.59298360e-01
1.46914363e+00 4.51141179e-01 2.67543674e-01 5.47022402e-01
-1.87424675e-01 1.38015616e+00 -7.29205906e-01 -9.71089959e-01
-1.39262244e-01 -9.20330226e-01 -1.44540119e+00 -3.86202186e-01
6.39302075e-01 1.11191772e-01 -1.83625728e-01 1.15479624e+00
4.26401049e-01 5.59867859e-01 9.56172645e-01 -9.94919419e-01
-1.40112317e+00 1.31523037e+00 1.01596013e-01 4.93510514e-01
6.40967727e-01 3.80031526e-01 1.27006328e+00 -6.35617316e-01
9.09082055e-01 1.75608790e+00 5.19874334e-01 7.62935638e-01
-1.67498839e+00 -4.12430972e-01 -2.44167402e-01 7.50470161e-01
-1.02312362e+00 -3.97985697e-01 9.24354196e-01 -4.13591638e-02
1.69171751e+00 3.10197651e-01 4.83226508e-01 1.44431531e+00
1.83568254e-01 8.11507821e-01 1.48286712e+00 -7.87499368e-01
4.62176383e-01 -6.55947551e-02 1.19648933e-01 6.48390949e-01
1.11024693e-01 -2.37977982e-01 -6.31151080e-01 -2.53634632e-01
3.82718176e-01 -5.27202487e-01 -3.14543098e-01 5.16381636e-02
-1.23985577e+00 1.13396001e+00 6.27471030e-01 4.34569329e-01
-4.36031044e-01 2.42865026e-01 4.82231826e-01 5.85469425e-01
3.36988688e-01 9.83629763e-01 -4.09116060e-01 -2.49318019e-01
-8.57013345e-01 5.45226634e-01 8.74587297e-01 7.52655149e-01
4.78072524e-01 -2.13119853e-02 -6.44333065e-01 9.27466452e-01
-1.81125868e-02 6.24502718e-01 6.04329109e-01 -9.98347700e-01
6.59468889e-01 5.99326313e-01 -4.71087307e-01 -9.86402869e-01
-1.89159423e-01 -1.98269427e-01 -4.46077824e-01 -3.83982241e-01
1.58726051e-01 6.19080178e-02 -1.07343221e+00 2.08252811e+00
2.01079585e-02 2.66137049e-02 5.99810004e-01 5.77130437e-01
8.68615389e-01 4.99328732e-01 4.16453540e-01 7.58065432e-02
1.25107455e+00 -5.71432650e-01 -6.58819199e-01 -6.48176908e-01
7.15988457e-01 -2.96271771e-01 1.77736807e+00 1.83839053e-01
-1.06139135e+00 -1.14287071e-01 -7.24386513e-01 -5.25501966e-01
-6.65982246e-01 -3.43814254e-01 5.73571086e-01 3.88763458e-01
-1.09300077e+00 5.16650736e-01 -4.53641653e-01 -3.53248537e-01
6.14321589e-01 -3.07887614e-01 4.83660698e-02 -4.38996971e-01
-1.86597717e+00 1.61348307e+00 9.74045932e-01 -3.43015283e-01
-1.00438213e+00 -7.34900057e-01 -1.23096740e+00 1.79462850e-01
6.22729778e-01 -1.29805136e+00 1.19479620e+00 -7.25954354e-01
-1.43896794e+00 1.02046084e+00 -1.93580508e-01 -8.06763470e-01
1.61797002e-01 -1.10840477e-01 -4.12681669e-01 2.59003073e-01
4.65493470e-01 6.53083026e-01 5.77109635e-01 -1.18974793e+00
-1.33727238e-01 -1.48114175e-01 3.55635226e-01 3.33224654e-01
-2.31342074e-02 -5.73383123e-02 2.38489091e-01 -6.88416481e-01
-1.52156293e-01 -6.45350039e-01 1.83834918e-02 -6.86674893e-01
-7.08223522e-01 -6.08661413e-01 4.71144646e-01 -7.81962097e-01
1.12076962e+00 -1.83204627e+00 4.47517663e-01 6.48184866e-02
-3.42447571e-02 9.67089832e-02 -5.01412638e-02 4.74020541e-01
1.24210522e-01 8.63139704e-02 -6.22077048e-01 -1.76881865e-01
5.19417524e-01 6.12360716e-01 -7.94328570e-01 -3.32015514e-01
5.13998747e-01 1.41283548e+00 -1.39141572e+00 -7.42503941e-01
7.51154423e-02 6.38477132e-02 -6.28850639e-01 -1.18155703e-01
-7.51257300e-01 -3.89535204e-02 -1.35659017e-02 6.51596367e-01
2.11353436e-01 -3.48530620e-01 1.00913964e-01 -9.79534313e-02
6.72168016e-01 8.70448947e-01 -5.93068719e-01 1.81008255e+00
-5.63383579e-01 5.83733737e-01 -7.62039244e-01 -9.03587699e-01
5.07496774e-01 1.22902364e-01 -5.40145516e-01 -5.96228242e-01
-2.95557119e-02 2.65136659e-01 1.13667689e-01 -6.72527492e-01
5.15191138e-01 -1.08514524e+00 -2.41607890e-01 2.00727463e-01
4.74078745e-01 -9.03990746e-01 4.26471323e-01 8.64908576e-01
1.34699261e+00 -6.61955588e-03 5.42439401e-01 -2.62381583e-01
3.91382813e-01 3.05160642e-01 1.57847375e-01 9.48697984e-01
-3.12761664e-02 1.49524376e-01 5.11256576e-01 2.61434495e-01
-7.55458832e-01 -1.48047841e+00 -6.11702316e-02 1.22136414e+00
-9.53245014e-02 -2.08088458e-01 -6.09243393e-01 -7.82506526e-01
9.40806940e-02 2.02994061e+00 -7.04589248e-01 -4.29024905e-01
-3.87598664e-01 -3.93343121e-01 1.04569995e+00 6.33712709e-01
5.65487444e-01 -1.68070936e+00 -5.39571583e-01 4.26023602e-01
-6.48734093e-01 -1.25323260e+00 -6.88718036e-02 2.22309828e-01
-6.59241617e-01 -9.86697674e-01 -2.84810781e-01 -5.41027546e-01
2.56041765e-01 -2.50631362e-01 1.61374474e+00 -3.23374778e-01
-2.17471465e-01 4.67647791e-01 -5.20487785e-01 -5.44780195e-01
-7.99396753e-01 -1.58336505e-01 -2.36585915e-01 -6.00289345e-01
7.46355772e-01 -7.48864532e-01 -3.86108761e-03 -5.37431896e-01
-1.11944199e+00 -2.48643681e-02 4.62156206e-01 9.15385306e-01
3.43023807e-01 -2.85922084e-02 9.24231887e-01 -9.48837996e-01
1.33120382e+00 -7.21727312e-01 8.72099698e-02 4.46012825e-01
-5.08321524e-01 3.18448275e-01 8.00158203e-01 -1.91431135e-01
-1.34900963e+00 -6.52267098e-01 -6.13646246e-02 -3.36278200e-01
-2.11475134e-01 8.94197166e-01 -6.67359531e-02 3.77266139e-01
1.26974869e+00 4.61343229e-01 -3.15273672e-01 1.33911874e-02
8.23218584e-01 4.46486890e-01 7.54834592e-01 -1.13388598e+00
5.82754552e-01 3.61653149e-01 -4.67203632e-02 -6.11652315e-01
-1.38360941e+00 -2.02306479e-01 -2.86019444e-01 1.84797436e-01
8.75218630e-01 -6.53448403e-01 -2.71732479e-01 -1.18686497e-01
-1.28353322e+00 -5.65063357e-01 -6.69552326e-01 1.70529649e-01
-8.25677633e-01 4.59783524e-01 -5.99741876e-01 -6.33594930e-01
-6.03347063e-01 -5.50317466e-01 9.22618389e-01 4.19931114e-02
-5.66096544e-01 -1.43373954e+00 1.96260363e-01 5.19570291e-01
3.72699440e-01 5.04452646e-01 1.50844431e+00 -8.57553959e-01
-1.64465174e-01 2.88918674e-01 -2.32405528e-01 5.74170291e-01
8.43069404e-02 -3.60120803e-01 -9.54329431e-01 1.19060598e-01
-1.10902004e-01 -1.03559244e+00 1.13021672e+00 1.40931264e-01
1.01513731e+00 -5.44284999e-01 -1.14446104e-01 1.39901817e-01
1.36996162e+00 -1.27453059e-01 7.69303858e-01 3.82271230e-01
3.84721816e-01 3.61219049e-01 3.66059601e-01 -4.46407087e-02
8.21840227e-01 1.29125997e-01 1.66849554e-01 3.00228715e-01
-3.45152676e-01 -3.38548601e-01 6.52858555e-01 5.55109322e-01
-1.22863360e-01 -1.33792892e-01 -1.15445995e+00 1.15016544e+00
-1.68211961e+00 -1.46625924e+00 1.48388911e-02 1.42060971e+00
1.78605354e+00 2.14972273e-01 -2.96396315e-01 -1.44991148e-02
2.37459317e-01 9.15101394e-02 -5.74172676e-01 -7.44680405e-01
-5.26691556e-01 8.94266665e-01 -1.57531887e-01 6.89833343e-01
-7.49110401e-01 1.43959904e+00 5.45647526e+00 1.02848864e+00
-5.19114017e-01 1.66065350e-01 7.61447549e-02 -8.29896182e-02
-7.94213712e-01 1.77674338e-01 -3.13893169e-01 2.66783714e-01
8.89105201e-01 -3.70402336e-01 6.26335084e-01 8.41337562e-01
-2.84852803e-01 -2.83954918e-01 -1.26864398e+00 7.10411668e-01
3.96616340e-01 -1.45484161e+00 5.85333407e-01 -5.10957956e-01
8.44324052e-01 7.89872557e-02 -7.11202398e-02 9.39289331e-01
8.74643087e-01 -1.08913040e+00 5.62922239e-01 5.24012625e-01
6.32517993e-01 -6.17663622e-01 7.22896457e-01 3.71400177e-01
-5.79961240e-01 6.51071742e-02 -5.10624290e-01 -1.05257489e-01
2.85718709e-01 6.71031535e-01 -1.12341022e+00 6.76848292e-01
4.39679831e-01 4.01423007e-01 -8.08459282e-01 2.73891270e-01
-9.26429808e-01 9.54293489e-01 -1.84274316e-01 -2.20147803e-01
4.55460846e-01 2.61213571e-01 6.22317612e-01 1.69539654e+00
-1.18159749e-01 4.79039431e-01 -6.29441999e-03 1.63387060e+00
-3.08732986e-01 -7.67505020e-02 -6.27367258e-01 -3.61412466e-01
5.32461226e-01 8.93195927e-01 -3.98039192e-01 -8.53013456e-01
2.43807584e-02 1.17755687e+00 7.41965711e-01 3.62610519e-01
-6.92579031e-01 -6.12525761e-01 2.77965367e-01 -2.17257082e-01
9.21866745e-02 8.84262621e-02 -3.56950760e-01 -1.32728732e+00
-9.70973074e-02 -9.61117148e-01 7.44024396e-01 -1.22467756e+00
-1.94904852e+00 3.82400155e-01 3.19043785e-01 -2.90128291e-01
-5.59148848e-01 -6.56349063e-01 -6.64546609e-01 9.29744542e-01
-1.66781271e+00 -1.17804718e+00 -7.85296857e-02 9.63730454e-01
6.04012847e-01 -5.00711100e-03 1.00952065e+00 -4.51551557e-01
6.64763972e-02 4.24827129e-01 -3.96189719e-01 2.14125454e-01
5.79761565e-01 -1.63591290e+00 3.39950353e-01 9.38404441e-01
2.62963414e-01 1.08732843e+00 9.15113807e-01 -1.05017066e+00
-9.42713916e-01 -9.70917821e-01 1.11752069e+00 -1.04035425e+00
9.76526201e-01 -8.14878792e-02 -1.05369997e+00 1.04910946e+00
5.63850462e-01 -2.51132071e-01 8.30262005e-01 4.31964427e-01
-7.56364167e-01 5.06749034e-01 -1.27964783e+00 7.20202744e-01
1.25501418e+00 -9.83450353e-01 -1.90190756e+00 5.47832251e-01
9.74883080e-01 -4.52689320e-01 -7.28013456e-01 1.76104471e-01
-1.36447072e-01 -6.77620232e-01 8.58105004e-01 -1.13541353e+00
1.04789436e+00 -1.41091749e-01 -4.10647869e-01 -1.91477275e+00
-3.04678291e-01 -3.63252103e-01 -3.00557137e-01 1.05169249e+00
5.01228929e-01 -7.20105290e-01 1.50412098e-01 5.77285528e-01
-3.19951922e-01 -5.52422285e-01 -1.02740622e+00 -1.00404787e+00
3.79066110e-01 -4.84623253e-01 4.30293769e-01 1.42274272e+00
4.98456955e-01 8.92396212e-01 3.18165690e-01 -8.44744891e-02
6.33247972e-01 2.76909888e-01 3.43246788e-01 -8.44698668e-01
-5.67107141e-01 -4.25680965e-01 -3.05632472e-01 -4.67242479e-01
6.57767951e-01 -1.47283113e+00 6.03209250e-02 -1.99283826e+00
5.39923549e-01 -1.52486876e-01 -1.53309554e-01 9.28763628e-01
-7.19477654e-01 -1.03881590e-01 2.83839434e-01 -2.34938517e-01
-5.15635550e-01 5.88243902e-01 1.34743369e+00 -2.42403343e-01
-7.15660453e-02 -7.53693640e-01 -8.61468017e-01 6.54736102e-01
9.48268890e-01 -3.35070878e-01 -6.97645724e-01 -2.95988768e-01
5.72747707e-01 -3.22287500e-01 9.90137041e-01 -7.39053011e-01
7.78575242e-02 -4.53038871e-01 1.84053525e-01 -3.08576047e-01
2.65995651e-01 -2.48296991e-01 -2.57481843e-01 1.96241707e-01
-5.85049570e-01 -1.65895000e-01 3.70314837e-01 4.36428785e-01
-1.37939855e-01 -3.71651471e-01 5.32242537e-01 -4.99438643e-01
-9.13782656e-01 -5.07845938e-01 8.02747905e-02 9.81883287e-01
6.28612459e-01 -8.23462903e-02 -7.45270371e-01 -3.20403159e-01
-8.01864564e-01 -4.13627066e-02 2.33666033e-01 1.77047700e-01
7.79675126e-01 -1.16880560e+00 -9.29628611e-01 -4.06751692e-01
3.91906679e-01 1.27269194e-01 2.14467376e-01 6.40306830e-01
-1.86137468e-01 6.08094692e-01 -3.30321081e-02 -2.37589821e-01
-9.08710659e-01 6.84730709e-01 2.33823434e-01 -3.16497892e-01
-5.02625823e-01 1.26213133e+00 -1.68130621e-01 -6.12540126e-01
-3.09580833e-01 -7.05956101e-01 4.98406775e-02 -1.83402181e-01
1.99625015e-01 2.00495258e-01 -6.59104884e-02 -2.95286059e-01
-3.92866671e-01 -8.22055712e-02 -7.45890364e-02 -2.78633177e-01
1.25423622e+00 1.33200616e-01 -4.40417171e-01 5.76274157e-01
7.21800029e-01 -5.15238754e-02 -5.68600953e-01 -4.66496050e-01
1.74613968e-01 -2.12972164e-01 -1.12875074e-01 -1.51290298e+00
-1.58751056e-01 8.47998381e-01 -2.17324898e-01 2.10080490e-01
9.72514868e-01 5.71979761e-01 6.91277325e-01 7.03247368e-01
4.07233059e-01 -1.18855429e+00 3.10972482e-01 9.36516285e-01
1.27030087e+00 -1.00714099e+00 -1.44427508e-01 -3.08688194e-01
-9.83278096e-01 9.06349480e-01 5.41153789e-01 -2.37957686e-02
-3.14388461e-02 -1.00994751e-01 -3.27245504e-01 -5.79050064e-01
-7.72982657e-01 -4.48269188e-01 3.94711792e-01 6.83771551e-01
2.94128746e-01 3.34194958e-01 -1.86040208e-01 6.70350730e-01
-7.80427277e-01 5.06963767e-02 4.77284938e-01 9.84383345e-01
-4.22297031e-01 -6.10660434e-01 -3.38928849e-01 5.77128232e-01
-8.21557269e-02 -9.07931089e-01 -8.70651543e-01 6.58203661e-01
-3.25841717e-02 1.00735915e+00 -3.84008080e-01 4.83026356e-03
3.19824100e-01 7.01926887e-01 1.01835811e+00 -8.94786000e-01
-4.86294299e-01 -7.72868454e-01 5.18597901e-01 -4.40601707e-01
-3.80891263e-01 -6.48442268e-01 -1.67077088e+00 -2.71564335e-01
-1.38411701e-01 -9.78535786e-02 1.55923873e-01 1.37410092e+00
3.80201519e-01 7.23295629e-01 -1.81588978e-01 -2.11153701e-01
-1.01129162e+00 -1.17120516e+00 -2.95074910e-01 8.76903057e-01
1.23270601e-01 -6.03397369e-01 -3.72408807e-01 2.60528892e-01]
|
[10.427983283996582, 8.056236267089844]
|
d68ac372-a764-427c-bdfc-5b0434249378
|
an-entity-guided-text-summarization-framework
|
2302.03205
| null |
https://arxiv.org/abs/2302.03205v1
|
https://arxiv.org/pdf/2302.03205v1.pdf
|
An entity-guided text summarization framework with relational heterogeneous graph neural network
|
Two crucial issues for text summarization to generate faithful summaries are to make use of knowledge beyond text and to make use of cross-sentence relations in text. Intuitive ways for the two issues are Knowledge Graph (KG) and Graph Neural Network (GNN) respectively. Entities are semantic units in text and in KG. This paper focuses on both issues by leveraging entities mentioned in text to connect GNN and KG for summarization. Firstly, entities are leveraged to construct a sentence-entity graph with weighted multi-type edges to model sentence relations, and a relational heterogeneous GNN for summarization is proposed to calculate node encodings. Secondly, entities are leveraged to link the graph to KG to collect knowledge. Thirdly, entities guide a two-step summarization framework defining a multi-task selector to select salient sentences and entities, and using an entity-focused abstractor to compress the sentences. GNN is connected with KG by constructing sentence-entity graphs where entity-entity edges are built based on KG, initializing entity embeddings on KG, and training entity embeddings using entity-entity edges. The relational heterogeneous GNN utilizes both edge weights and edge types in GNN to calculate graphs with weighted multi-type edges. Experiments show the proposed method outperforms extractive baselines including the HGNN-based HGNNSum and abstractive baselines including the entity-driven SENECA on CNN/DM, and outperforms most baselines on NYT50. Experiments on sub-datasets show the density of sentence-entity edges greatly influences the performance of the proposed method. The greater the density, the better the performance. Ablations show effectiveness of the method.
|
['Jingqiang Chen']
|
2023-02-07
| null | null | null | null |
['entity-embeddings']
|
['methodology']
|
[ 1.14466861e-01 7.57762372e-01 -4.59212929e-01 -2.07086325e-01
-5.01184106e-01 -1.71267882e-01 2.97533780e-01 5.83749354e-01
-5.00052810e-01 7.09822536e-01 1.23693788e+00 -5.35928197e-02
-1.34948090e-01 -1.33982623e+00 -7.30639756e-01 -2.61934727e-01
-2.18182504e-01 2.93100715e-01 1.75706461e-01 -4.84120011e-01
3.83367270e-01 3.23369764e-02 -9.87724483e-01 3.22138131e-01
1.31558895e+00 6.78402305e-01 1.65988937e-01 9.12606299e-01
-6.59655690e-01 9.28466618e-01 -9.07567441e-01 -7.42313802e-01
-1.26212433e-01 -4.34242457e-01 -9.92441773e-01 -1.60461977e-01
3.56384337e-01 -1.84843317e-01 -6.85131490e-01 1.00797153e+00
9.20807183e-01 4.64678228e-01 8.79546285e-01 -9.18559432e-01
-1.02673483e+00 1.49139202e+00 -6.23215973e-01 3.34009618e-01
3.96188855e-01 -2.22220391e-01 1.52297091e+00 -8.61260355e-01
9.10425782e-01 1.29459989e+00 8.14193130e-01 3.95218611e-01
-8.13095689e-01 -2.61786819e-01 1.54503837e-01 1.62523553e-01
-1.13421357e+00 -3.46224487e-01 8.65578353e-01 1.80800319e-01
1.62772179e+00 3.61794859e-01 6.16301239e-01 9.86042440e-01
2.08582819e-01 9.28362846e-01 -9.27574001e-03 -2.04304010e-01
8.52316469e-02 -9.77562442e-02 5.69573879e-01 7.69267917e-01
8.67403746e-01 -8.14175308e-01 -7.29341269e-01 -1.94710623e-02
2.16290414e-01 -1.52850792e-01 -5.18100500e-01 8.15228447e-02
-1.01578534e+00 8.27568114e-01 9.03598607e-01 2.12893933e-01
-6.67594969e-01 5.19592017e-02 9.18039799e-01 8.09455961e-02
7.18442261e-01 8.45160902e-01 -3.20299089e-01 9.42706242e-02
-1.07030797e+00 2.12427437e-01 1.10422921e+00 1.22776830e+00
6.45703197e-01 1.74771756e-01 -7.88191319e-01 9.56454039e-01
-2.00865325e-02 3.03027987e-01 7.74058521e-01 -4.70791578e-01
1.23871052e+00 1.17168474e+00 -6.52154922e-01 -1.42403424e+00
-4.42275941e-01 -7.32919514e-01 -1.23617029e+00 -8.58937442e-01
-4.53285336e-01 -4.43802297e-01 -8.40120494e-01 1.53682148e+00
1.20777585e-01 1.51647091e-01 6.05308115e-01 4.57374871e-01
1.89283943e+00 7.81257629e-01 7.02584758e-02 -4.93741557e-02
1.25479448e+00 -1.21595073e+00 -7.98183501e-01 -2.36846372e-01
9.28655386e-01 -2.51435101e-01 7.56513655e-01 -3.30236882e-01
-1.26847088e+00 -3.58933181e-01 -1.06813633e+00 -5.78921258e-01
-5.92695832e-01 1.74590454e-01 3.72236013e-01 1.23574339e-01
-1.28257763e+00 9.18143272e-01 -5.06998003e-01 -6.29255176e-01
6.00216448e-01 2.14880735e-01 -3.45911294e-01 1.68069780e-01
-1.61568689e+00 9.29006279e-01 1.07869554e+00 1.17585160e-01
-3.01204592e-01 -7.31935859e-01 -1.47718692e+00 6.09280169e-01
2.24436194e-01 -1.45934844e+00 8.53678167e-01 -3.14079672e-01
-1.04715168e+00 4.03338701e-01 -2.47480437e-01 -8.28617811e-01
-3.59668471e-02 -5.69150262e-02 -4.36850578e-01 4.57082570e-01
2.68746287e-01 6.55494988e-01 3.07834685e-01 -1.06744766e+00
-6.10833883e-01 -1.51801601e-01 9.22490284e-02 8.34751964e-01
-6.96080625e-01 -3.97408754e-01 -5.76377273e-01 -7.43373215e-01
5.98320365e-02 -3.77206683e-01 -1.73765182e-01 -1.00327694e+00
-1.25217283e+00 -5.65079808e-01 7.09687293e-01 -1.04366124e+00
1.92353880e+00 -1.68028617e+00 4.57021147e-01 3.88955511e-02
6.82179213e-01 2.19130769e-01 -3.62184942e-01 8.68510425e-01
7.02773128e-03 3.49568278e-01 -2.28577450e-01 -5.72956622e-01
3.46652828e-02 6.83343187e-02 -4.43342894e-01 -1.98744267e-01
3.06145698e-01 1.42250681e+00 -1.16204488e+00 -7.93576896e-01
-2.61200309e-01 1.82797700e-01 -5.56202233e-01 1.28765717e-01
-1.35569975e-01 -4.72175062e-01 -6.96288526e-01 3.66840661e-01
4.16308045e-01 -1.97824046e-01 1.08284272e-01 -6.21213734e-01
2.88775235e-01 7.52765357e-01 -8.39842021e-01 1.60582840e+00
-2.76935339e-01 3.83496583e-01 -4.33409631e-01 -1.09021997e+00
9.00119126e-01 7.63506964e-02 1.56260699e-01 -3.40371370e-01
1.22506551e-01 -8.66414979e-03 -3.38789135e-01 -5.86329103e-01
1.33377635e+00 2.68451691e-01 -3.71918201e-01 3.56986284e-01
4.80940282e-01 -2.71196246e-01 6.24744236e-01 1.27640033e+00
1.49387813e+00 -1.71928331e-01 3.23093057e-01 -1.07559137e-01
3.57242763e-01 4.26236279e-02 4.78669614e-01 6.06947601e-01
3.97710174e-01 5.51453292e-01 8.74512970e-01 4.37712148e-02
-7.38412321e-01 -9.07744348e-01 4.09656137e-01 8.25910330e-01
1.15406252e-01 -1.06645989e+00 -6.84157848e-01 -9.45278049e-01
4.66896929e-02 1.12483990e+00 -7.28179634e-01 -7.14438796e-01
-6.11536026e-01 -7.83341408e-01 7.32295871e-01 7.86916792e-01
7.32759833e-01 -1.20811391e+00 -6.43008947e-03 2.10192189e-01
-4.41177279e-01 -1.06409383e+00 -6.55195773e-01 -5.44498600e-02
-9.39302742e-01 -7.72393823e-01 -5.62463582e-01 -9.19751883e-01
8.82149398e-01 2.16038302e-01 1.48661017e+00 -1.50420684e-02
5.19008934e-02 4.81162727e-01 -5.71251631e-01 -4.47052717e-01
-2.98953325e-01 7.95383632e-01 -1.60065129e-01 -4.03710663e-01
2.46970698e-01 -7.14258611e-01 -5.10637522e-01 -4.83227938e-01
-9.22858298e-01 2.36615270e-01 8.38409901e-01 5.87815642e-01
4.30597872e-01 1.32498190e-01 9.72411215e-01 -1.32331109e+00
1.30915332e+00 -7.00636864e-01 3.57106030e-01 5.37392259e-01
-4.80709612e-01 2.61123449e-01 8.52911353e-01 -3.93222086e-02
-1.02027237e+00 -6.31192863e-01 -2.93003190e-02 -1.16910219e-01
5.66209376e-01 1.14324236e+00 -1.60601646e-01 5.84070265e-01
6.29083276e-01 3.59182298e-01 -4.49509174e-01 -1.56107575e-01
8.29195917e-01 6.54641330e-01 6.04208648e-01 -4.20785159e-01
6.63529098e-01 -2.84293527e-03 -1.15506597e-01 -8.37584198e-01
-1.02362871e+00 -4.34280038e-01 -4.08089519e-01 1.30180672e-01
1.01552749e+00 -9.54416931e-01 -2.10197225e-01 1.64905459e-01
-1.27689266e+00 7.89281726e-02 -8.02443802e-01 2.74583995e-01
-7.45538026e-02 5.98045170e-01 -8.87705266e-01 -2.57874459e-01
-1.24779642e+00 -5.19663870e-01 1.06986427e+00 5.86080909e-01
-3.17693055e-01 -1.40221381e+00 -8.52714351e-04 2.47629195e-01
4.04880852e-01 3.35661829e-01 1.02190709e+00 -1.09416068e+00
-3.22567701e-01 -2.36765251e-01 -5.02479374e-01 3.96703005e-01
2.38574624e-01 -5.37542477e-02 -5.49773514e-01 -1.67839319e-01
-4.27967161e-01 -1.88152373e-01 1.53797185e+00 5.39092183e-01
9.55817819e-01 -7.03245461e-01 -4.67346907e-01 5.33926547e-01
1.26119590e+00 -4.82286125e-01 6.51598275e-01 8.38531181e-02
1.23105800e+00 4.82339799e-01 1.16371527e-01 3.02562743e-01
1.18285525e+00 1.08407140e-02 1.82309374e-01 -4.90826517e-02
-5.14087617e-01 -6.90217972e-01 5.52565753e-01 1.64641392e+00
-7.79908448e-02 -6.52308404e-01 -5.79844892e-01 7.22370207e-01
-1.95075905e+00 -1.10659289e+00 -2.27984622e-01 1.69695878e+00
1.07678854e+00 1.24149859e-01 -1.52237982e-01 -2.18729213e-01
8.38069558e-01 5.30781031e-01 -4.72295582e-01 -4.61056828e-01
-4.51795697e-01 1.71965003e-01 5.20559132e-01 3.68182600e-01
-8.58583927e-01 1.12922537e+00 4.73070192e+00 8.86842489e-01
-7.17615128e-01 -2.18629748e-01 4.73618597e-01 5.15499385e-04
-6.77959919e-01 -3.77585813e-02 -1.05147982e+00 3.15024674e-01
8.61760855e-01 -8.45694542e-01 -2.61803251e-03 6.25212789e-01
4.43729982e-02 5.82627114e-03 -8.88117850e-01 6.07722580e-01
5.95926702e-01 -1.68237138e+00 7.03654885e-01 -4.18403655e-01
8.85703027e-01 4.84112538e-02 -3.49650234e-01 8.20403516e-01
6.84147596e-01 -7.67097116e-01 2.32648090e-01 5.16894281e-01
5.62892437e-01 -7.37088621e-01 9.31609750e-01 1.84007555e-01
-1.47289443e+00 2.07420081e-01 -7.39295006e-01 3.88238519e-01
2.90657312e-01 8.40329766e-01 -1.13363647e+00 1.41426170e+00
4.95552540e-01 1.13665318e+00 -7.05103815e-01 8.03942263e-01
-5.89724123e-01 6.53858125e-01 -7.92058930e-02 -2.59287894e-01
3.15401405e-01 -2.01533958e-01 8.81142020e-01 1.72669220e+00
3.48781109e-01 8.59608129e-02 8.90313461e-02 8.03464592e-01
-8.60526085e-01 2.35096589e-01 -5.93768179e-01 -2.10775718e-01
6.33048177e-01 1.59714067e+00 -6.17957354e-01 -7.27424383e-01
-2.18927726e-01 1.04346538e+00 7.49897063e-01 4.87285614e-01
-6.19397879e-01 -1.10379469e+00 1.35581851e-01 -8.07795525e-02
3.64380121e-01 2.38864273e-02 -2.70601064e-01 -1.43643034e+00
1.26936272e-01 -3.65706563e-01 8.29825759e-01 -7.89896786e-01
-1.27960277e+00 6.95346177e-01 6.71249405e-02 -7.49445319e-01
-1.31039470e-01 8.45544785e-02 -1.11967278e+00 9.34440434e-01
-1.59936237e+00 -1.16455150e+00 -2.74231285e-01 2.88928807e-01
5.71925223e-01 -9.38869193e-02 6.02391303e-01 -3.55999218e-03
-9.58512962e-01 5.23389161e-01 -2.49371201e-01 4.72677290e-01
5.47375202e-01 -1.63109040e+00 8.28018606e-01 1.06351280e+00
3.27419341e-02 8.54089200e-01 4.68541563e-01 -1.11507618e+00
-1.23681974e+00 -1.70065284e+00 1.23346007e+00 -3.19454670e-01
5.96050680e-01 -8.70514289e-02 -9.09583986e-01 8.46230567e-01
6.22225881e-01 -2.62861431e-01 6.43744111e-01 2.36541331e-01
-1.74234480e-01 8.77325162e-02 -8.83954883e-01 7.41863132e-01
1.30275404e+00 -2.44239762e-01 -1.11756492e+00 3.98489654e-01
1.34294438e+00 -5.47375023e-01 -9.23439145e-01 4.28310245e-01
-4.32287008e-02 -4.21515107e-01 8.88420820e-01 -8.60647798e-01
9.09133613e-01 -1.64088279e-01 1.52509004e-01 -1.98836195e+00
-3.91482502e-01 -4.16700274e-01 -5.98638356e-01 1.76208210e+00
8.10474813e-01 -6.56030834e-01 7.08000422e-01 3.46760273e-01
-6.78802907e-01 -9.81223583e-01 -4.60068583e-01 -5.42365491e-01
-1.32297486e-01 9.91975144e-02 6.08254433e-01 9.70086873e-01
3.34918886e-01 1.13415456e+00 9.14780721e-02 7.23849908e-02
3.45864862e-01 1.17848754e-01 6.30981088e-01 -9.41023290e-01
2.71298941e-02 -4.86925453e-01 -2.46282488e-01 -1.02907407e+00
3.54201466e-01 -1.47069383e+00 -3.21353137e-01 -2.59074283e+00
4.38068539e-01 -9.53054875e-02 -6.39195889e-02 3.85606050e-01
-7.17904985e-01 -3.33508313e-01 7.20143244e-02 -4.21367586e-02
-9.86457646e-01 9.21112716e-01 1.13116586e+00 -3.19372296e-01
-4.55076844e-01 -2.40314841e-01 -1.19686627e+00 5.55894792e-01
7.39156246e-01 -3.70526344e-01 -7.18681335e-01 -5.68648577e-01
5.16401529e-01 -5.09196669e-02 -9.04002413e-02 -8.32769871e-01
7.49296188e-01 3.19705695e-01 3.34870875e-01 -7.46461809e-01
1.63602866e-02 -2.19795272e-01 -2.87387699e-01 4.48048487e-02
-5.71190596e-01 1.61563084e-01 -2.04441845e-02 6.45479679e-01
-3.75507742e-01 -3.21365535e-01 2.57376015e-01 -1.72153041e-01
-5.66108167e-01 3.81869733e-01 1.85745791e-01 5.76611042e-01
6.19211316e-01 -1.96560621e-01 -7.58553386e-01 -5.23109436e-01
-5.15024185e-01 6.66560769e-01 8.99039581e-02 3.76594216e-01
7.75751293e-01 -1.30764294e+00 -9.89051223e-01 -2.60413349e-01
-3.57673094e-02 6.01314902e-01 4.64498669e-01 7.01734424e-01
-4.60901290e-01 1.98901638e-01 2.37249807e-01 -7.33610615e-02
-1.17141402e+00 2.14768052e-01 1.78165492e-02 -8.58855724e-01
-7.58394897e-01 1.06240344e+00 -1.06442399e-01 -4.79399621e-01
5.65673262e-02 -6.43161118e-01 -7.21435726e-01 4.65237141e-01
3.98405164e-01 5.01737833e-01 1.33783460e-01 -5.57028711e-01
-1.42519116e-01 1.95214108e-01 -3.18216205e-01 2.56512076e-01
1.61798644e+00 -1.33935198e-01 -5.03010154e-01 8.12782422e-02
1.21368384e+00 1.72359347e-01 -5.41012883e-01 -3.74860048e-01
3.06651480e-02 1.75545052e-01 8.30328390e-02 -5.72603762e-01
-1.09247375e+00 6.13372684e-01 -6.22148871e-01 4.50316161e-01
1.12907374e+00 -9.51983593e-03 1.30993056e+00 5.31321228e-01
-1.54350713e-01 -1.09651661e+00 6.99478537e-02 7.34115005e-01
1.00928605e+00 -6.91960037e-01 4.03754950e-01 -5.69777668e-01
-1.06201291e+00 1.14392447e+00 5.85300505e-01 -2.40239084e-01
2.48892397e-01 -9.70436167e-03 -5.71701169e-01 -4.56647724e-01
-7.53333330e-01 -2.75525868e-01 6.18304014e-01 3.62207651e-01
2.92525411e-01 7.44108483e-02 -3.45288515e-01 1.05883920e+00
-5.92103601e-01 -3.01049620e-01 7.43928552e-01 6.72458827e-01
-5.95910549e-01 -7.50391483e-01 3.79995465e-01 1.10360765e+00
-4.27523702e-01 -5.56429982e-01 -5.80026269e-01 4.89051282e-01
-2.45641947e-01 7.59640455e-01 -3.43484394e-02 -5.37398815e-01
6.56698704e-01 -9.72485989e-02 1.26101360e-01 -1.08174431e+00
-7.37244070e-01 -6.75274849e-01 7.15827942e-01 -1.87247232e-01
-1.50303930e-01 -4.58171695e-01 -1.54911232e+00 -4.14580375e-01
-4.17676359e-01 3.27580720e-01 4.15757507e-01 6.59106195e-01
8.74578059e-01 1.01922631e+00 5.05454063e-01 -6.45454586e-01
-4.59447801e-01 -1.25111079e+00 -5.19211948e-01 4.42250669e-01
-2.92682480e-02 -7.01120794e-02 -4.51560259e-01 -1.76060528e-01]
|
[12.575108528137207, 9.536314010620117]
|
87a27118-d0a5-46a9-bd4b-9e1524a79ea9
|
high-performance-offline-handwritten-chinese
|
1505.04925
| null |
http://arxiv.org/abs/1505.04925v1
|
http://arxiv.org/pdf/1505.04925v1.pdf
|
High Performance Offline Handwritten Chinese Character Recognition Using GoogLeNet and Directional Feature Maps
|
Just like its great success in solving many computer vision problems, the
convolutional neural networks (CNN) provided new end-to-end approach to
handwritten Chinese character recognition (HCCR) with very promising results in
recent years. However, previous CNNs so far proposed for HCCR were neither deep
enough nor slim enough. We show in this paper that, a deeper architecture can
benefit HCCR a lot to achieve higher performance, meanwhile can be designed
with less parameters. We also show that the traditional feature extraction
methods, such as Gabor or gradient feature maps, are still useful for enhancing
the performance of CNN. We design a streamlined version of GoogLeNet [13],
which was original proposed for image classification in recent years with very
deep architecture, for HCCR (denoted as HCCR-GoogLeNet). The HCCR-GoogLeNet we
used is 19 layers deep but involves with only 7.26 million parameters.
Experiments were conducted using the ICDAR 2013 offline HCCR competition
dataset. It has been shown that with the proper incorporation with traditional
directional feature maps, the proposed single and ensemble HCCR-GoogLeNet
models achieve new state of the art recognition accuracy of 96.35% and 96.74%,
respectively, outperforming previous best result with significant gap.
|
['Lianwen Jin', 'Zhuoyao Zhong', 'Zecheng Xie']
|
2015-05-19
| null | null | null | null |
['offline-handwritten-chinese-character', 'offline-handwritten-chinese-character']
|
['computer-vision', 'natural-language-processing']
|
[-3.23518276e-01 -5.16946316e-01 1.00391224e-01 -3.61676633e-01
-4.27918106e-01 -3.81071121e-01 5.18133700e-01 -4.24368560e-01
-6.98908985e-01 6.32242858e-01 -2.72616819e-02 -4.75823343e-01
1.53680460e-03 -8.05537283e-01 -6.38481259e-01 -5.73947310e-01
-2.92238835e-02 2.46728316e-01 2.05282673e-01 -6.00529373e-01
5.19034386e-01 7.68334389e-01 -9.75494146e-01 3.18974972e-01
7.65024722e-01 1.21123743e+00 2.68243194e-01 9.23745990e-01
1.19524203e-01 1.16125309e+00 -9.33009207e-01 -7.40469992e-01
2.05129489e-01 3.74183357e-02 -6.69305503e-01 -2.42234558e-01
3.09797406e-01 -4.38206285e-01 -9.90559995e-01 7.78048098e-01
7.20987260e-01 -2.39994354e-03 3.98287773e-01 -5.74829161e-01
-1.47362578e+00 5.79980433e-01 -3.49986672e-01 1.65781066e-01
3.07572279e-02 5.48888221e-02 6.08957350e-01 -1.14839113e+00
4.57741499e-01 7.88429916e-01 1.00350356e+00 7.78730273e-01
-3.24314922e-01 -4.90083754e-01 -5.54293543e-02 3.33852381e-01
-1.39933932e+00 -1.56334545e-02 5.13929188e-01 3.96790691e-02
1.27763736e+00 4.01599258e-01 6.28580153e-01 1.07674742e+00
1.32246688e-01 1.19667888e+00 1.07235193e+00 -2.78298914e-01
-1.25710323e-01 -1.76525682e-01 4.16566640e-01 8.25784326e-01
3.35751101e-02 -1.01274580e-01 -2.77543008e-01 2.91210562e-01
8.91839862e-01 1.52934954e-01 -2.60422468e-01 4.39209223e-01
-1.33544481e+00 7.14922309e-01 8.17493796e-01 6.12990856e-01
-3.92117828e-01 4.18737829e-01 6.04054034e-01 4.09259170e-01
2.41699949e-01 4.86948460e-01 -5.49241960e-01 -4.58295822e-01
-7.41732955e-01 1.14571713e-01 6.71678305e-01 1.08265150e+00
3.27457339e-01 5.97716212e-01 -3.36995304e-01 1.12694883e+00
-4.81051952e-02 4.47403908e-01 9.51884389e-01 -2.88208961e-01
4.75497276e-01 5.52108884e-01 -2.34458238e-01 -1.29765522e+00
-2.46775329e-01 -7.74997830e-01 -1.45289600e+00 -1.71957508e-01
2.96685517e-01 -1.43070996e-01 -1.34640014e+00 9.42416430e-01
-5.80415606e-01 1.06873326e-01 3.03533286e-01 1.23084068e+00
1.14528263e+00 7.67749846e-01 -2.17631787e-01 7.73576677e-01
1.26171899e+00 -1.31714845e+00 -5.18204212e-01 -7.18979836e-02
6.17852628e-01 -8.34195137e-01 9.75229084e-01 7.05475509e-01
-8.29642534e-01 -6.73117578e-01 -1.27713704e+00 -1.55253634e-01
-6.82084084e-01 8.80577743e-01 9.46937978e-01 9.53944027e-01
-1.31277466e+00 6.77697718e-01 -5.96522093e-01 -2.37850845e-01
5.91179788e-01 3.09114963e-01 -5.55246174e-01 -4.17873055e-01
-1.15331089e+00 1.02870166e+00 5.22048533e-01 7.19815910e-01
-1.08426201e+00 -2.23426551e-01 -3.92821908e-01 1.13709345e-01
2.17326984e-01 1.16413042e-01 1.02769148e+00 -8.96921337e-01
-1.81705403e+00 5.60148895e-01 3.84786397e-01 -7.89054632e-01
7.45079935e-01 -2.73959905e-01 -4.10406351e-01 -7.88852107e-03
-5.43559134e-01 5.01751781e-01 2.65112281e-01 -6.10319316e-01
-3.95298809e-01 -2.96513587e-01 -1.74290657e-01 -4.40730061e-03
-5.08892894e-01 1.40350088e-01 -8.10770571e-01 -7.80337870e-01
1.34527937e-01 -8.28222513e-01 -2.40583614e-01 -3.44761461e-01
-4.05326426e-01 -3.37829381e-01 1.07655203e+00 -9.89413500e-01
7.31340051e-01 -2.03193212e+00 -1.93439931e-01 5.67932762e-02
1.68080792e-01 9.99700844e-01 -3.69893759e-01 2.75832355e-01
1.49366602e-01 1.94340259e-01 -3.29931229e-02 -1.85293078e-01
-2.08195060e-01 1.56233935e-02 -3.95065963e-01 6.39947057e-01
2.69262850e-01 1.20587599e+00 -6.36723578e-01 9.26184431e-02
2.94204891e-01 6.94631636e-01 -2.04058900e-01 1.07728086e-01
2.30216458e-02 -1.24668583e-01 -3.77312005e-01 9.65085387e-01
9.48976517e-01 -3.16198081e-01 -6.00703247e-02 -1.18485726e-01
-2.75558263e-01 -1.62477970e-01 -6.43509805e-01 1.44906020e+00
-2.91591763e-01 1.17781603e+00 -3.29684466e-01 -1.29919744e+00
1.66623962e+00 2.18869835e-01 3.25312912e-02 -9.33022499e-01
1.94447622e-01 7.20277548e-01 1.29645139e-01 -1.71839580e-01
8.80213380e-01 5.68762243e-01 6.62459284e-02 -2.38682240e-01
2.28282332e-01 1.91688478e-01 -8.13934356e-02 1.94456000e-02
9.91282165e-01 -1.03665613e-01 -8.74634087e-02 -2.83698380e-01
5.58159053e-01 2.45320261e-01 2.47426093e-01 1.17036808e+00
-1.15502700e-01 1.08533335e+00 3.06909323e-01 -8.79191458e-01
-1.30078137e+00 -4.81285959e-01 -9.28071141e-02 5.88881493e-01
-1.39985187e-02 -1.24085031e-01 -5.24852157e-01 -3.53611797e-01
-2.76798993e-01 1.12171128e-01 -6.01640821e-01 5.10584656e-03
-9.78381932e-01 -9.93336976e-01 1.45091951e+00 7.68915951e-01
1.61202168e+00 -1.08237028e+00 -3.90937924e-01 3.37007910e-01
2.99458593e-01 -1.11620176e+00 -2.69102007e-01 1.37432396e-01
-7.74774909e-01 -8.86608541e-01 -1.57194924e+00 -1.10800183e+00
3.43059748e-01 2.78852314e-01 8.79499257e-01 4.62711841e-01
-4.43945020e-01 -1.02101706e-01 -8.92115653e-01 -1.27198160e-01
-9.14474949e-02 3.99615765e-01 -3.91025215e-01 -1.88374341e-01
3.03439528e-01 4.41054106e-02 -6.54914021e-01 3.29004675e-01
-7.50157416e-01 9.86323804e-02 8.64372075e-01 1.26833141e+00
5.45950346e-02 -2.41406754e-01 6.19627118e-01 -6.73808575e-01
7.36122310e-01 -1.64803803e-01 -5.06309390e-01 4.53132451e-01
-5.07827938e-01 -2.61459649e-01 1.03639925e+00 -4.20716047e-01
-7.86930382e-01 -2.47508138e-01 -5.04978597e-01 -4.68750596e-01
-1.29706711e-01 8.35377276e-01 2.59028614e-01 -5.51161289e-01
5.81581831e-01 9.42019939e-01 -1.82333410e-01 -5.59245050e-01
4.13364992e-02 1.02429223e+00 7.08156049e-01 -4.84874457e-01
1.77336961e-01 1.49105340e-01 -3.25219929e-01 -9.63835359e-01
-2.66572505e-01 -2.72112697e-01 -2.32420713e-01 -8.16071555e-02
8.32682669e-01 -1.10882235e+00 -1.04235625e+00 1.47268367e+00
-1.25860977e+00 -3.33891600e-01 4.43454593e-01 4.49403614e-01
-7.91698471e-02 3.79180908e-01 -1.01179612e+00 -6.35817647e-01
-6.57793343e-01 -1.22891355e+00 7.00884998e-01 3.93346816e-01
7.59828806e-01 -9.36918020e-01 -2.96542317e-01 1.07898161e-01
1.13858044e+00 1.06105849e-01 3.91541719e-01 -9.00829673e-01
-7.28593528e-01 -6.38605118e-01 -6.79310083e-01 8.22849929e-01
-1.33921325e-01 1.05860896e-01 -8.75928640e-01 -4.39398408e-01
-3.72431785e-01 -4.56187278e-01 1.27741098e+00 1.21467464e-01
1.57194674e+00 -1.99950844e-01 6.50857985e-02 9.59488750e-01
1.71431863e+00 5.81761241e-01 1.35618317e+00 6.31600440e-01
8.76183629e-01 -1.13946117e-01 3.01376343e-01 3.05842340e-01
1.80811003e-01 5.41906059e-01 2.84855276e-01 -3.43015909e-01
-1.85649186e-01 -1.42886899e-02 1.74270764e-01 1.11957264e+00
-6.26802444e-01 -3.04854214e-01 -1.26869965e+00 3.95732641e-01
-1.69329262e+00 -6.41775310e-01 -1.07968226e-03 1.74000025e+00
5.43653846e-01 -1.14932582e-02 -3.35065961e-01 -3.66861857e-02
6.44655764e-01 1.73508525e-01 -4.39686507e-01 -4.87668693e-01
-7.81201661e-01 4.22845423e-01 9.45667565e-01 1.31293431e-01
-1.16982937e+00 1.26537514e+00 6.00637674e+00 1.07738614e+00
-1.54756010e+00 -1.49050459e-01 1.00636733e+00 6.11957431e-01
2.78113753e-01 -2.27384388e-01 -8.56784105e-01 2.82172978e-01
7.34794199e-01 1.59590110e-01 3.76081526e-01 1.16268373e+00
-3.10140908e-01 2.90566862e-01 -5.09281218e-01 1.24291074e+00
2.57703096e-01 -1.85156918e+00 1.42283499e-01 -1.32547051e-01
8.26016903e-01 3.68598700e-01 2.79967457e-01 5.55096388e-01
3.02173287e-01 -1.50407362e+00 6.64861977e-01 5.19861579e-01
9.25194740e-01 -8.09130490e-01 1.39320886e+00 1.89456701e-01
-1.02216589e+00 -1.32571772e-01 -8.92241955e-01 -3.81465591e-02
-4.23883975e-01 3.98609966e-01 -6.19311929e-01 6.92443728e-01
6.71351969e-01 1.05869031e+00 -9.59845781e-01 1.20456243e+00
7.34951720e-02 6.39431953e-01 2.86316276e-02 -7.20380068e-01
8.95818353e-01 -5.61236851e-02 1.20576061e-01 1.48418498e+00
4.24798429e-01 2.27700789e-02 -2.38577962e-01 6.13262653e-01
-3.28410923e-01 -3.74840908e-02 -4.11349058e-01 -1.40183896e-01
2.71909356e-01 1.14466202e+00 -6.93564713e-01 -5.27129591e-01
-2.17016384e-01 1.11424148e+00 2.79295892e-01 3.09852213e-01
-8.59776855e-01 -9.53861296e-01 2.36915141e-01 -5.99199235e-01
4.39633518e-01 -4.26930845e-01 -2.33221471e-01 -1.50095332e+00
-1.49334380e-02 -1.21582389e+00 -5.69307730e-02 -7.17535794e-01
-1.19967079e+00 1.09888017e+00 -6.60766959e-01 -1.06041121e+00
2.43832543e-01 -1.51744866e+00 -5.87911904e-01 1.01923752e+00
-1.46394372e+00 -1.20929277e+00 -4.90464807e-01 6.41904593e-01
7.74077117e-01 -7.23197520e-01 7.77712405e-01 4.65317756e-01
-6.74526155e-01 9.69063342e-01 5.36453545e-01 9.65518951e-01
2.88630098e-01 -1.16865575e+00 7.26084352e-01 7.57911742e-01
-6.57042637e-02 6.43451571e-01 1.65783018e-01 -4.09317195e-01
-1.85377622e+00 -1.06697261e+00 4.70775515e-01 1.75793190e-02
3.52278203e-01 -4.21412081e-01 -8.04857373e-01 3.99509519e-01
1.97561517e-01 1.73218146e-01 1.25244036e-01 -1.93148151e-01
-4.60589677e-01 -1.13424107e-01 -9.70513344e-01 5.10425985e-01
6.36757791e-01 -4.28708732e-01 -1.78054869e-01 2.40451485e-01
6.91685259e-01 -7.50480831e-01 -7.28746057e-01 3.55830997e-01
6.06057048e-01 -9.25643861e-01 8.82565260e-01 -6.63156450e-01
6.19584918e-01 -9.23600197e-02 -4.39891785e-01 -1.12452173e+00
-2.50278205e-01 -2.37957001e-01 2.14378923e-01 7.93310523e-01
5.13441384e-01 -7.44334519e-01 9.17169869e-01 2.54109949e-01
-5.73854506e-01 -1.06102049e+00 -8.99521112e-01 -9.58522916e-01
3.05452228e-01 -2.45437920e-01 5.74474156e-01 8.68769944e-01
-4.95302647e-01 -1.49433255e-01 -7.58335769e-01 -7.03460425e-02
2.68528223e-01 -2.04522073e-01 5.71755469e-01 -7.68085480e-01
-7.75125474e-02 -5.04675984e-01 -9.21163499e-01 -1.27084792e+00
-2.74274260e-01 -6.99453056e-01 -1.00947611e-01 -1.41517627e+00
2.57509023e-01 -5.89718223e-01 -3.60453993e-01 5.78040421e-01
1.33855581e-01 5.30557573e-01 5.20051777e-01 2.50393599e-01
-4.13509756e-01 6.35520101e-01 1.39415228e+00 -3.71336102e-01
1.87926814e-01 -2.07761869e-01 -3.85609895e-01 3.48724574e-01
8.93401802e-01 9.92611423e-02 2.59157389e-01 -8.75035822e-01
-7.57589936e-03 -3.41579542e-02 4.67122346e-01 -1.30969000e+00
4.73297626e-01 4.14621502e-01 7.77694345e-01 -6.32471204e-01
2.42874295e-01 -3.72882873e-01 -1.22903764e-01 6.50341809e-01
-3.87360692e-01 3.29750776e-01 2.30796695e-01 2.18674123e-01
-7.09783614e-01 -1.04963228e-01 7.04458475e-01 -2.75023490e-01
-1.13497162e+00 4.39780653e-01 -3.59512806e-01 -4.40656096e-01
6.70833170e-01 -3.14233184e-01 -6.74769819e-01 -1.90189153e-01
-4.20568705e-01 -1.95190355e-01 4.31036111e-03 5.46481848e-01
1.07756412e+00 -1.34230518e+00 -9.79525030e-01 -1.57689780e-01
-2.04141796e-01 -1.74794689e-01 3.57347131e-01 5.69650173e-01
-1.35860491e+00 9.74921405e-01 -4.31234598e-01 -4.62176830e-01
-9.23029482e-01 1.67027429e-01 5.89055955e-01 -3.11635822e-01
-7.83885777e-01 8.99447858e-01 -4.31590289e-01 -6.23562932e-01
4.09754962e-01 -2.47814655e-01 -3.05597633e-01 -4.37078387e-01
5.33516765e-01 4.28248763e-01 4.97502297e-01 -4.76981580e-01
-3.92764419e-01 6.35487974e-01 -3.65288526e-01 1.91768527e-01
1.62164807e+00 5.72633028e-01 -1.82962924e-01 -7.99432322e-02
1.42929530e+00 -4.85334873e-01 -1.00944340e+00 -1.98594481e-02
-1.11295216e-01 -5.23184359e-01 6.82669058e-02 -1.13907242e+00
-1.57203102e+00 1.02689052e+00 7.43263543e-01 4.77588437e-02
1.20095646e+00 -4.71255600e-01 7.21159101e-01 1.06498873e+00
4.84211802e-01 -9.33419287e-01 2.13893443e-01 1.13250589e+00
9.15179372e-01 -1.27210462e+00 -3.17642212e-01 8.82681906e-02
-6.29542053e-01 1.64456451e+00 8.32152784e-01 -7.29247689e-01
4.69990253e-01 1.58007488e-01 6.56328723e-02 -1.57564774e-01
-5.50684750e-01 2.24309847e-01 1.23518631e-01 4.48975086e-01
5.43072343e-01 2.10681945e-01 -2.56506085e-01 4.77455914e-01
-6.97398335e-02 8.11613798e-02 6.70232236e-01 8.48002017e-01
-2.85511911e-01 -8.15240443e-01 -1.95520312e-01 4.62676883e-01
-6.64282262e-01 -3.75898063e-01 -3.04794371e-01 9.32920218e-01
-1.94464102e-01 5.36837697e-01 2.24133264e-02 -7.79925883e-01
2.96259165e-01 -4.69717801e-01 2.76046157e-01 -8.70425776e-02
-7.69041657e-01 -4.60834414e-01 -6.75919577e-02 -3.17836583e-01
-9.76407826e-02 1.18959574e-02 -7.19274938e-01 -5.99461615e-01
-4.69483316e-01 5.52104153e-02 1.00264585e+00 6.59899235e-01
3.38280737e-01 4.85242724e-01 5.70782661e-01 -6.06414318e-01
-8.87911558e-01 -1.26516831e+00 -6.39826715e-01 -8.84998962e-02
2.27437660e-01 -7.02747107e-02 -7.18065398e-03 -3.26237559e-01]
|
[11.777482986450195, 2.58612322807312]
|
c1c5921c-9eec-4ecb-bac1-e9366fbadba6
|
multi-temporal-land-cover-classification-with
|
1802.0208
| null |
http://arxiv.org/abs/1802.02080v4
|
http://arxiv.org/pdf/1802.02080v4.pdf
|
Multi-Temporal Land Cover Classification with Sequential Recurrent Encoders
|
Earth observation (EO) sensors deliver data with daily or weekly temporal
resolution. Most land use and land cover (LULC) approaches, however, expect
cloud-free and mono-temporal observations. The increasing temporal capabilities
of today's sensors enables the use of temporal, along with spectral and spatial
features. Domains, such as speech recognition or neural machine translation,
work with inherently temporal data and, today, achieve impressive results using
sequential encoder-decoder structures. Inspired by these sequence-to-sequence
models, we adapt an encoder structure with convolutional recurrent layers in
order to approximate a phenological model for vegetation classes based on a
temporal sequence of Sentinel 2 (S2) images. In our experiments, we visualize
internal activations over a sequence of cloudy and non-cloudy images and find
several recurrent cells, which reduce the input activity for cloudy
observations. Hence, we assume that our network has learned cloud-filtering
schemes solely from input data, which could alleviate the need for tedious
cloud-filtering as a preprocessing step for many EO approaches. Moreover, using
unfiltered temporal series of top-of-atmosphere (TOA) reflectance data, we
achieved in our experiments state-of-the-art classification accuracies on a
large number of crop classes with minimal preprocessing compared to other
classification approaches.
|
['Marco Körner', 'Marc Rußwurm']
|
2018-02-06
|
multi-temporal-land-cover-classification-with-1
|
https://www.mdpi.com/2220-9964/7/4/129
|
https://www.mdpi.com/2220-9964/7/4/129
|
international-journal-of-geo-information-2018
|
['unet-segmentation']
|
['computer-vision']
|
[ 5.06613135e-01 -5.18326819e-01 1.48168936e-01 -5.00908554e-01
-3.75095546e-01 -7.21255004e-01 6.27628744e-01 1.32747322e-01
-3.90686303e-01 6.60113990e-01 -1.64128706e-01 -8.02643239e-01
1.38183296e-01 -1.14100444e+00 -6.45080447e-01 -7.50764906e-01
-4.24348444e-01 -2.33145341e-01 -1.84974223e-02 -5.44637561e-01
-2.55189806e-01 6.15140557e-01 -2.19412160e+00 7.11882591e-01
1.10195529e+00 1.22771132e+00 5.84376216e-01 8.56113434e-01
-4.38445330e-01 6.49931967e-01 -2.76510596e-01 1.31307051e-01
2.50773668e-01 -4.40263301e-01 -5.22110999e-01 9.96680036e-02
4.14510399e-01 -1.27636999e-01 -1.63824216e-01 1.07132375e+00
1.92179278e-01 -3.54112200e-02 3.62498671e-01 -8.56468320e-01
-4.87252295e-01 2.32806563e-01 -3.83350164e-01 2.54206032e-01
-2.19840720e-01 2.26189882e-01 8.57239246e-01 -8.25146914e-01
5.55954814e-01 7.83978462e-01 7.49780118e-01 4.03726339e-01
-1.31374812e+00 -4.43118781e-01 5.42739987e-01 1.23201236e-01
-1.28260899e+00 -4.84155536e-01 4.24313694e-01 -5.58741987e-01
1.33777881e+00 5.82739949e-01 1.17451000e+00 9.27553356e-01
1.46194115e-01 6.75131202e-01 1.23262334e+00 -2.27006063e-01
1.16715379e-01 -2.14768961e-01 -7.10099861e-02 3.28421295e-01
4.61130477e-02 3.76824796e-01 -2.46299863e-01 1.01833068e-01
4.33937311e-01 4.72718447e-01 -2.65426755e-01 -4.70202081e-02
-1.00482869e+00 6.41469300e-01 9.70896721e-01 3.02267849e-01
-8.01047564e-01 3.95602584e-01 4.27885145e-01 7.52097368e-01
9.01515007e-01 1.51926875e-01 -7.37320721e-01 1.59701124e-01
-1.10959446e+00 1.25426158e-01 3.47022861e-01 9.53377843e-01
9.69357371e-01 4.42561269e-01 2.47182459e-01 5.07244945e-01
1.56670347e-01 1.09679437e+00 3.38112473e-01 -6.10621095e-01
2.99498767e-01 3.88754010e-01 3.20117801e-01 -7.27479160e-01
-3.29601496e-01 -5.76827466e-01 -1.33726084e+00 1.49367794e-01
-1.00577630e-01 -7.88306668e-02 -1.30551648e+00 1.39490187e+00
-3.62053886e-02 7.11418986e-02 3.42870027e-01 9.49270129e-01
3.45241785e-01 9.23258603e-01 9.37164426e-02 -4.89863753e-01
1.32977176e+00 -6.01969898e-01 -7.49251127e-01 -4.38525468e-01
8.24460208e-01 -4.23201054e-01 9.78384733e-01 3.72604541e-02
-3.92108917e-01 -4.53149319e-01 -8.40540707e-01 3.06558728e-01
-1.07702339e+00 2.91274250e-01 9.55696702e-01 4.91952360e-01
-1.11638033e+00 8.13604057e-01 -1.14017224e+00 -4.38489974e-01
2.98829406e-01 -1.15853183e-01 -1.84401780e-01 8.54296759e-02
-1.36939728e+00 7.53901362e-01 3.00283819e-01 7.89434135e-01
-9.10381317e-01 -5.78679562e-01 -7.59318888e-01 2.67873913e-01
4.36182059e-02 -4.22778189e-01 9.11191523e-01 -1.52673388e+00
-1.30230808e+00 8.04352522e-01 -3.67498517e-01 -8.08539808e-01
1.16963290e-01 -1.37417674e-01 -7.23894358e-01 7.33005926e-02
-1.03568308e-01 6.85880005e-01 1.12810123e+00 -1.10838890e+00
-7.15676248e-01 -2.78494358e-01 -1.80267066e-01 1.00466171e-02
-3.29285473e-01 2.18579359e-02 2.35689595e-01 -5.58235109e-01
3.05235386e-01 -9.76526082e-01 -4.48686540e-01 2.18631446e-01
3.56153667e-01 2.98047513e-01 8.14714551e-01 -7.41804957e-01
1.02221608e+00 -2.35219002e+00 -5.24054132e-02 6.66081207e-03
-3.99086103e-02 3.56836051e-01 -5.33349335e-01 5.03873885e-01
-4.13908809e-01 2.88109362e-01 -7.21825182e-01 -2.38200381e-01
-3.41778070e-01 4.99863416e-01 -7.76421964e-01 4.40156072e-01
6.39021158e-01 8.50445867e-01 -1.07586300e+00 7.06579862e-03
3.02393883e-01 5.36759377e-01 -2.25860924e-01 3.23184878e-01
-6.79193139e-01 2.67978251e-01 -1.31715089e-01 5.66924155e-01
9.12052214e-01 -2.03014240e-01 1.98412865e-01 3.19784492e-01
-6.78069711e-01 4.34831768e-01 -5.32828867e-01 1.66363645e+00
-8.00234199e-01 1.04460275e+00 1.41330451e-01 -9.20100272e-01
9.79038239e-01 4.53359544e-01 3.75041306e-01 -9.33529913e-01
-2.79315859e-01 4.84715879e-01 -6.82561006e-03 -5.14162481e-01
5.76060832e-01 -8.98627043e-02 4.20942128e-01 1.65641513e-02
-1.21357717e-01 -2.06094295e-01 -1.14183262e-01 -2.92252332e-01
8.10781956e-01 3.42917651e-01 -4.33707573e-02 -2.69374788e-01
3.52657139e-01 3.32950532e-01 3.15613031e-01 5.81083536e-01
-1.00741558e-01 7.07209587e-01 2.62149423e-01 -9.26553369e-01
-1.21090794e+00 -6.61501110e-01 -2.08226398e-01 9.06617284e-01
-3.85912418e-01 -2.86702961e-01 -2.56407112e-01 -1.07566491e-01
-4.46079038e-02 5.09934902e-01 -6.33042514e-01 1.38927400e-01
-3.44006568e-01 -1.09639883e+00 5.24909616e-01 4.26307797e-01
5.90829670e-01 -1.11571550e+00 -1.03650391e+00 3.96128774e-01
-2.69261777e-01 -1.08989513e+00 2.65372932e-01 6.25590265e-01
-1.18365574e+00 -6.27462089e-01 -6.47470415e-01 -1.92479357e-01
3.61560822e-01 6.37404740e-01 1.16892862e+00 8.49161446e-02
-2.73801833e-01 -8.57844651e-02 -6.07034087e-01 -4.33153391e-01
-2.50467926e-01 2.33988389e-01 -8.29330552e-03 1.70956627e-01
4.82071787e-01 -7.56363928e-01 -5.34930468e-01 -1.80377841e-01
-1.07248104e+00 2.38089472e-01 4.35571522e-01 8.67353559e-01
4.55843329e-01 -1.38667732e-01 1.72250897e-01 -5.73888719e-01
7.69115686e-02 -4.30161864e-01 -1.15358460e+00 2.61301398e-01
-6.02350175e-01 1.98599696e-02 5.75344682e-01 -1.59532875e-01
-8.38157713e-01 3.79536271e-01 -1.57702193e-01 -7.79899359e-01
-2.77841538e-01 1.00118268e+00 3.29312772e-01 8.70317966e-02
8.30318570e-01 7.39760339e-01 -1.33260339e-01 -3.50006640e-01
2.99928218e-01 9.43947673e-01 4.81520265e-01 -1.12529509e-01
6.76787615e-01 6.71839118e-01 -1.59004435e-01 -1.19391835e+00
-7.43304551e-01 -6.04206741e-01 -8.58079672e-01 -3.05763781e-01
8.25922966e-01 -1.33125103e+00 -6.60085678e-01 5.63027382e-01
-1.47608566e+00 -4.12806451e-01 -1.87964901e-01 3.71886879e-01
-1.59317106e-01 2.22685561e-02 -1.95407465e-01 -1.30350673e+00
-5.10675251e-01 -6.17073119e-01 1.16032267e+00 -1.14451759e-01
3.50761116e-01 -6.23007298e-01 -1.21185638e-01 -4.38686401e-01
9.16099966e-01 3.56454939e-01 6.14670992e-01 2.44956881e-01
-8.65792394e-01 -9.61926579e-02 -4.83019650e-01 4.58699107e-01
4.15396273e-01 2.31146410e-01 -1.44939971e+00 -3.36694628e-01
1.18971542e-02 3.42235789e-02 1.47702575e+00 3.49697739e-01
1.22089708e+00 -2.54419357e-01 -5.43908663e-02 9.56655145e-01
1.65068495e+00 -9.29202512e-02 8.35684001e-01 2.99700648e-01
5.99834979e-01 6.96286976e-01 5.20134032e-01 4.56372172e-01
9.96915102e-02 2.54817784e-01 9.72962201e-01 -3.07829976e-01
2.11908236e-01 7.60792270e-02 5.34537613e-01 8.03600371e-01
-3.75821680e-01 -2.68640041e-01 -1.13034070e+00 9.76345539e-01
-1.88950598e+00 -1.21040583e+00 -3.42694491e-01 2.20331216e+00
6.04050577e-01 -1.60554692e-01 -2.74639666e-01 -6.29995540e-02
3.33201021e-01 6.23121619e-01 -5.74672759e-01 -3.73820156e-01
-5.34040868e-01 1.36779085e-01 1.12696147e+00 3.34238082e-01
-1.38690400e+00 1.11142695e+00 6.27432346e+00 2.30187044e-01
-1.65361464e+00 1.16514869e-01 2.84271479e-01 -6.59468919e-02
-5.17528355e-01 -2.44432855e-02 -4.53946263e-01 3.42314571e-01
1.55946600e+00 2.24489748e-01 7.82481372e-01 5.88442385e-01
4.84698445e-01 1.21020675e-01 -6.64718091e-01 7.42933750e-01
-5.59506714e-01 -1.53478289e+00 -9.63991731e-02 -1.03930824e-01
6.63748324e-01 9.27179158e-01 -1.00589275e-01 1.68097004e-01
2.97311008e-01 -1.05015218e+00 5.74241698e-01 6.70623779e-01
1.04905951e+00 -3.09042037e-01 7.59324193e-01 3.43059361e-01
-1.47467017e+00 -1.15021162e-01 -6.40262246e-01 -3.49119127e-01
-3.21143448e-01 8.10678482e-01 -2.79845715e-01 8.58992219e-01
1.09405005e+00 1.15617526e+00 -4.12089288e-01 6.83933854e-01
-1.39338779e-03 6.58644378e-01 -6.80010319e-01 3.73558663e-02
6.61262512e-01 -3.60077769e-01 1.65021896e-01 1.31709599e+00
4.80406910e-01 3.43516141e-01 -1.87634602e-01 7.78669536e-01
8.97841379e-02 -1.02972545e-01 -1.04719317e+00 -3.78407478e-01
2.35366613e-01 1.01638556e+00 -5.21325648e-01 -3.71452808e-01
-5.38865030e-01 1.05530560e+00 -2.66266591e-03 5.08550704e-01
-5.00625789e-01 -2.92314708e-01 1.19253063e+00 -5.18140905e-02
5.98535359e-01 -5.40106356e-01 -7.75574818e-02 -1.45933068e+00
1.37172058e-01 -7.57496893e-01 -1.73096324e-03 -9.93756771e-01
-9.42240298e-01 8.28244686e-01 -3.08747500e-01 -1.71394980e+00
-2.75350899e-01 -6.68024898e-01 -2.83980161e-01 1.26714289e+00
-2.15381145e+00 -1.12500584e+00 -5.97266257e-01 2.09744900e-01
3.75282735e-01 6.48287907e-02 1.24729896e+00 3.33287060e-01
-1.00155473e-01 -1.75665349e-01 4.99661416e-01 -4.84954156e-02
2.86198676e-01 -1.04907727e+00 8.83040428e-01 1.09971237e+00
2.33017877e-02 3.20397913e-01 5.78748763e-01 -4.90749389e-01
-1.45178318e+00 -1.52354610e+00 1.10394895e+00 -4.69823778e-02
7.84086049e-01 -4.98332381e-01 -1.14627957e+00 6.54091597e-01
2.36708954e-01 4.63293046e-01 1.68019652e-01 4.46035266e-02
-4.49493974e-01 -4.04399455e-01 -6.16246104e-01 3.06067586e-01
9.62044179e-01 -1.06101227e+00 -7.95011669e-02 6.45162702e-01
9.12156224e-01 -3.48955035e-01 -6.27456605e-01 4.22267288e-01
6.14014268e-01 -8.61263692e-01 6.28207624e-01 -7.64608026e-01
6.09572589e-01 -5.57811320e-01 -4.75245804e-01 -1.42382073e+00
-3.65646571e-01 -3.82349819e-01 2.11102694e-01 6.37158275e-01
2.95284152e-01 -7.15015829e-01 4.94300008e-01 -1.80091280e-02
-2.63138175e-01 -2.75404423e-01 -8.57949018e-01 -8.90624940e-01
5.96508980e-02 -5.06657600e-01 8.70612264e-01 1.12132013e+00
-6.06641114e-01 -1.11594565e-01 -3.74542922e-01 7.58494675e-01
2.52883255e-01 6.53526306e-01 3.37025970e-01 -1.35875964e+00
2.02331513e-01 -2.31516063e-01 -2.44270265e-01 -8.65338624e-01
-4.06922810e-02 -8.42786074e-01 1.16900556e-01 -1.15772641e+00
-3.74734461e-01 -3.75255167e-01 -3.06687087e-01 7.57623136e-01
3.81619669e-02 1.18405081e-01 1.47173464e-01 3.40973139e-01
1.29620522e-01 8.03229153e-01 8.70397985e-01 -4.05039966e-01
-1.14568301e-01 -5.76585159e-02 1.22509159e-01 2.61409163e-01
7.39998162e-01 -6.22941613e-01 4.85500647e-03 -9.78250921e-01
6.38537943e-01 1.44545302e-01 5.73628485e-01 -8.75741303e-01
-3.04193795e-02 -3.10833365e-01 2.99295455e-01 -9.19474721e-01
2.12320879e-01 -1.11440730e+00 3.55450779e-01 6.28041446e-01
-8.19675475e-02 1.22318514e-01 4.99812007e-01 3.72972965e-01
-5.12130499e-01 1.51720658e-01 6.25339389e-01 -3.14518332e-01
-9.33943272e-01 3.71844798e-01 -7.85520852e-01 -7.13752449e-01
3.81339490e-01 -1.75611433e-02 -2.90402621e-01 -1.61312267e-01
-7.82943904e-01 3.63436013e-01 3.39959413e-01 5.19930661e-01
5.03684521e-01 -1.00299764e+00 -9.91596043e-01 5.96256077e-01
4.86316681e-01 1.04703093e-02 1.84367478e-01 6.75126016e-01
-7.74961531e-01 7.15983570e-01 -3.42164636e-01 -1.04880905e+00
-1.06123781e+00 5.10605335e-01 8.28902781e-01 -2.31689867e-02
-6.91753983e-01 5.64783931e-01 -1.32636756e-01 -6.85559988e-01
-3.04016054e-01 -8.21283340e-01 -4.58839163e-02 3.94154578e-01
2.96749234e-01 -3.14769268e-01 4.47066396e-01 -3.45240057e-01
-4.10306692e-01 3.06010514e-01 3.96218836e-01 -9.07545257e-03
1.68810880e+00 -1.15282461e-01 -2.56493747e-01 8.58351707e-01
9.84354913e-01 -6.95427239e-01 -1.20816839e+00 -3.37511063e-01
5.11889644e-02 -5.81530452e-01 2.95903057e-01 -5.18020689e-01
-1.25309241e+00 1.37002230e+00 1.16023326e+00 5.46190381e-01
1.50628614e+00 -6.04218721e-01 3.76041353e-01 8.24267149e-01
2.16295913e-01 -7.41117954e-01 -8.55414987e-01 7.63710320e-01
7.65887678e-01 -1.52914238e+00 -1.95220470e-01 -2.16022506e-01
-3.19196016e-01 1.18462467e+00 1.37821615e-01 4.69657779e-02
7.31627166e-01 2.84707695e-01 3.36418748e-01 -1.81407407e-01
-1.26429772e+00 -7.82553732e-01 -1.27106473e-01 5.52434027e-01
5.59327483e-01 5.37982106e-01 1.43702418e-01 -1.05857700e-01
-8.03237334e-02 2.41701141e-01 4.67607141e-01 8.76483083e-01
-4.29344475e-01 -7.31668532e-01 -2.71495342e-01 5.70238233e-01
1.63118076e-02 -5.90911746e-01 3.13775390e-02 3.82762969e-01
-1.11100405e-01 6.80287957e-01 4.61938262e-01 -2.52274424e-01
1.40651345e-01 1.40802890e-01 1.45795822e-01 -5.02426684e-01
-7.64260173e-01 1.16165362e-01 4.14834823e-03 -6.20932400e-01
-7.57918000e-01 -9.24887717e-01 -8.14427257e-01 -4.61124033e-01
-2.84903616e-01 -1.04496971e-01 1.00598443e+00 7.97654033e-01
4.76602197e-01 7.24399447e-01 9.50548530e-01 -1.00309968e+00
-2.57254601e-01 -1.13766325e+00 -6.97651267e-01 5.07983416e-02
1.00635529e+00 -1.58190876e-01 -1.69188038e-01 2.10736036e-01]
|
[9.598726272583008, -1.6166291236877441]
|
f1969ed7-84ed-49c3-bcd7-59adc99b564b
|
on-gibbs-sampling-architecture-for-labeled
|
2306.15135
| null |
https://arxiv.org/abs/2306.15135v1
|
https://arxiv.org/pdf/2306.15135v1.pdf
|
On Gibbs Sampling Architecture for Labeled Random Finite Sets Multi-Object Tracking
|
Gibbs sampling is one of the most popular Markov chain Monte Carlo algorithms because of its simplicity, scalability, and wide applicability within many fields of statistics, science, and engineering. In the labeled random finite sets literature, Gibbs sampling procedures have recently been applied to efficiently truncate the single-sensor and multi-sensor $\delta$-generalized labeled multi-Bernoulli posterior density as well as the multi-sensor adaptive labeled multi-Bernoulli birth distribution. However, only a limited discussion has been provided regarding key Gibbs sampler architecture details including the Markov chain Monte Carlo sample generation technique and early termination criteria. This paper begins with a brief background on Markov chain Monte Carlo methods and a review of the Gibbs sampler implementations proposed for labeled random finite sets filters. Next, we propose a short chain, multi-simulation sample generation technique that is well suited for these applications and enables a parallel processing implementation. Additionally, we present two heuristic early termination criteria that achieve similar sampling performance with substantially fewer Markov chain observations. Finally, the benefits of the proposed Gibbs samplers are demonstrated via two Monte Carlo simulations.
|
['Pramod K. Varshney', 'Donald J. Bucci Jr.', 'Anthony Trezza']
|
2023-06-27
| null | null | null | null |
['object-tracking', 'multi-object-tracking']
|
['computer-vision', 'computer-vision']
|
[ 5.55542648e-01 -1.35449454e-01 -4.11300845e-02 -4.83977765e-01
-1.04147053e+00 -1.03686035e-01 6.42038286e-01 4.15484458e-02
-4.11750585e-01 1.07112610e+00 -1.65959284e-01 -1.34266168e-01
7.83370957e-02 -1.07023203e+00 -3.59894991e-01 -7.84276426e-01
-1.83365852e-01 7.75531828e-01 4.70304757e-01 5.86994886e-01
8.29493552e-02 5.35422683e-01 -1.57103229e+00 -3.59457225e-01
4.98762429e-01 7.33318985e-01 3.99334073e-01 1.02221620e+00
1.57038286e-01 3.48337471e-01 -5.04801333e-01 5.55215292e-02
-1.80756330e-01 -6.33115411e-01 -2.64453858e-01 -1.07068673e-01
-1.43248856e-01 -4.58721608e-01 3.14860195e-01 1.01568079e+00
7.30445385e-01 3.35659295e-01 8.07924986e-01 -9.53857720e-01
3.27815861e-01 8.76563907e-01 -5.91960728e-01 -1.47082046e-01
6.43454552e-01 -4.11182158e-02 4.66647655e-01 -8.36687624e-01
1.74651295e-01 1.34305692e+00 8.85406017e-01 3.04699749e-01
-1.11500132e+00 -8.41571093e-01 -2.03816399e-01 -9.96192023e-02
-1.70253682e+00 -3.15009207e-01 3.36828411e-01 -2.90283799e-01
8.15224230e-01 1.69741869e-01 1.14375114e+00 1.03112161e+00
5.36557376e-01 8.51883829e-01 1.26805079e+00 -6.24594748e-01
9.51228142e-01 -2.41755828e-01 1.65691063e-01 2.62751698e-01
7.96356320e-01 4.97867316e-01 -4.61992472e-01 -6.43546343e-01
6.91414237e-01 2.08434403e-01 2.67395377e-01 -1.71219721e-01
-1.01894033e+00 8.97705436e-01 -4.07417327e-01 -1.81655139e-01
-5.97671270e-01 6.68276906e-01 4.10992391e-02 -5.22594810e-01
2.51274258e-01 -3.54161203e-01 -3.49753648e-02 -3.09943885e-01
-1.49630010e+00 4.85909522e-01 1.04995620e+00 1.31344354e+00
9.49166834e-01 2.46802419e-01 -1.78519428e-01 3.69145960e-01
1.13588333e+00 1.47881377e+00 -1.23489507e-01 -7.86893308e-01
-2.17371389e-01 -3.20545226e-01 5.57489872e-01 -2.48610109e-01
-1.77066609e-01 -3.11807599e-02 -7.52757788e-01 1.45107388e-01
1.32915169e-01 -5.48903823e-01 -1.04913330e+00 1.23542130e+00
5.95959902e-01 3.96615505e-01 -1.09351046e-01 4.32957768e-01
3.79657507e-01 6.09584630e-01 2.79590487e-01 -6.36002898e-01
1.44844127e+00 -3.31496038e-02 -9.51545119e-01 -2.84157813e-01
-2.80299455e-01 -1.07826161e+00 5.15414476e-01 3.27892452e-01
-1.13141108e+00 -2.92529583e-01 -9.87583995e-01 6.17453039e-01
-4.66170497e-02 -3.74010772e-01 6.48409367e-01 1.40418434e+00
-5.86008608e-01 4.06910807e-01 -1.56555223e+00 -6.21208727e-01
3.18743944e-01 8.04070234e-02 7.08860576e-01 -4.42079864e-02
-9.64056551e-01 5.41704655e-01 3.52778286e-01 -1.31454334e-01
-1.24655998e+00 -1.49909794e-01 -5.86277008e-01 -8.64657015e-02
1.05583139e-01 -6.56094551e-01 1.47966492e+00 1.04370937e-01
-1.77531791e+00 2.05657154e-01 -5.69517374e-01 -6.79901421e-01
3.55129540e-01 -3.04820329e-01 -4.52126682e-01 4.54827733e-02
1.45660728e-01 4.17639196e-01 7.29666412e-01 -1.05526423e+00
-7.71418810e-01 -7.13747442e-02 -8.04716706e-01 -1.98491458e-02
3.85340959e-01 -3.83810177e-02 -2.11023360e-01 -4.50033337e-01
1.91977218e-01 -8.74601185e-01 -7.00170934e-01 -5.69011927e-01
-4.26470160e-01 4.64093350e-02 4.56643730e-01 -2.56501255e-03
1.20592105e+00 -1.78437507e+00 -9.28356111e-01 6.80187225e-01
-3.84165078e-01 -1.64003938e-01 4.36088413e-01 7.96894550e-01
3.99557620e-01 -3.30149710e-01 -3.56402159e-01 -2.67035246e-01
-1.37004048e-01 1.77431837e-01 -2.34605417e-01 7.12147772e-01
-4.07018930e-01 4.29940045e-01 -1.18531871e+00 -7.34999180e-01
6.35922492e-01 5.19828200e-01 -3.87503475e-01 1.02260023e-01
-1.96571425e-01 4.69039708e-01 -5.05382895e-01 7.13533878e-01
9.12210643e-01 -2.73177505e-01 3.46951067e-01 -5.84568828e-03
-2.22886398e-01 5.59124090e-02 -1.83243346e+00 1.07635891e+00
1.10461535e-02 2.77206749e-01 -6.21484146e-02 -4.17414695e-01
1.09563386e+00 4.64133441e-01 7.47373164e-01 -1.76050402e-02
3.13352197e-01 4.05450642e-01 -6.35406494e-01 -1.01155631e-01
9.23005164e-01 -7.71370590e-01 -2.00253338e-01 7.59851158e-01
-1.48802057e-01 -6.00570321e-01 2.19959363e-01 4.50244220e-03
9.07996416e-01 2.49024868e-01 9.56510961e-01 -2.21378922e-01
1.06197566e-01 -1.57221481e-01 5.38164139e-01 1.45130825e+00
-2.32502833e-01 6.33495688e-01 -4.96334434e-01 -1.97393261e-02
-8.90826643e-01 -1.49917138e+00 -4.05604273e-01 7.50016570e-01
3.26538801e-01 -2.23588124e-01 -6.16210461e-01 4.86322567e-02
-1.38313994e-01 8.41674924e-01 -1.95766747e-01 3.35986674e-01
-3.81702125e-01 -1.24931324e+00 4.88167048e-01 4.96104568e-01
5.02228141e-01 -9.06763434e-01 -1.06472182e+00 6.78989589e-01
-2.86543667e-01 -9.07103062e-01 -8.65290966e-03 2.17790276e-01
-1.27972579e+00 -1.02061605e+00 -5.60185492e-01 -2.69927412e-01
3.26616913e-01 2.31268212e-01 9.34306622e-01 -6.05027974e-01
-3.05204391e-01 7.00994313e-01 -4.62986261e-01 -8.34829569e-01
-4.85252440e-01 -1.99933529e-01 5.02407774e-02 -2.15028137e-01
8.87991965e-01 -5.14791250e-01 -6.80238128e-01 2.98227489e-01
-6.64617598e-01 -2.74677694e-01 6.34958267e-01 4.81565118e-01
8.17118287e-01 -1.07456781e-01 2.55290449e-01 -8.76121700e-01
4.69048291e-01 -3.63809466e-01 -8.18778336e-01 -5.82353882e-02
-6.56061351e-01 -9.05592665e-02 -1.93016790e-02 -1.85410783e-01
-1.28636694e+00 2.46733129e-01 -3.07903260e-01 2.03502715e-01
-2.64546305e-01 3.23242158e-01 -6.84211254e-02 1.61995113e-01
6.53093755e-01 7.42635950e-02 -1.06498487e-01 -4.44056541e-01
3.67565125e-01 7.62026727e-01 3.33632201e-01 -5.13764322e-01
2.51169831e-01 1.17750680e+00 1.89013958e-01 -1.10573971e+00
-5.74272037e-01 -7.39275753e-01 -1.14893302e-01 -3.73969376e-01
8.40120673e-01 -9.43852484e-01 -8.53748918e-01 6.73301637e-01
-8.64726126e-01 -1.89755544e-01 -8.06207776e-01 1.07182169e+00
-9.18965161e-01 5.90666413e-01 -3.14981192e-01 -1.61849594e+00
-3.96955222e-01 -9.45648074e-01 1.13572943e+00 4.36051369e-01
-3.35501909e-01 -1.00160635e+00 5.03071606e-01 -2.63840556e-01
2.78376698e-01 3.82088691e-01 8.11057836e-02 -4.34880704e-01
-7.41043150e-01 -5.61776221e-01 1.27968907e-01 -6.85680583e-02
-2.41929088e-02 6.34484515e-02 -9.18729424e-01 -5.99757433e-01
-1.34656966e-01 1.00708805e-01 5.64371228e-01 1.09186876e+00
4.47894394e-01 1.14480726e-01 -9.16930854e-01 2.11185887e-01
1.68574798e+00 2.56561458e-01 7.53583014e-01 -6.72478899e-02
7.87606761e-02 -1.79042414e-01 9.75777686e-01 1.28638232e+00
1.37962669e-01 2.00613793e-02 4.15282637e-01 2.88170755e-01
8.62285271e-02 -2.88346827e-01 3.42729539e-01 7.21745253e-01
1.93441696e-02 -4.17300701e-01 -6.75458431e-01 7.66683996e-01
-1.48498225e+00 -1.36867714e+00 -4.83089328e-01 2.52635479e+00
7.75629401e-01 1.02593169e-01 2.28688642e-01 1.70333758e-01
1.16607332e+00 1.28460258e-01 -4.44571137e-01 1.44676575e-02
1.87145606e-01 5.73729038e-01 1.03419113e+00 7.56936610e-01
-1.07357812e+00 7.24180639e-01 8.10439968e+00 8.87275815e-01
-2.85669416e-01 2.37632662e-01 2.32632589e-02 6.85476810e-02
-1.71614632e-01 4.62378055e-01 -1.60745931e+00 5.54575026e-01
1.26501524e+00 1.48825143e-02 -6.75109103e-02 7.35335112e-01
5.62245727e-01 -1.08090961e+00 -8.25011194e-01 8.38051677e-01
-1.94476292e-01 -1.09674323e+00 6.98672608e-02 1.01357177e-01
8.80241096e-01 2.42020071e-01 -4.22986358e-01 -1.02351360e-01
1.14272332e+00 -6.28894985e-01 7.40427732e-01 6.92939937e-01
8.50096762e-01 -6.35890901e-01 6.58389747e-01 3.46893460e-01
-1.46247852e+00 3.12138826e-01 -4.41408873e-01 -1.34956151e-01
1.03395569e+00 1.37673771e+00 -1.17272151e+00 4.12327558e-01
8.06316197e-01 2.68758327e-01 9.81579944e-02 1.40896666e+00
-1.28618807e-01 1.03801084e+00 -1.00563538e+00 -7.13949084e-01
-3.00410204e-02 -1.32964730e-01 7.49916375e-01 1.37214863e+00
5.70638835e-01 3.05728409e-02 2.91079283e-01 5.24956048e-01
6.68893814e-01 -3.34875196e-01 -2.58085966e-01 3.15196514e-01
1.18312657e+00 7.88940310e-01 -1.05630934e+00 -6.78993046e-01
-2.89401233e-01 1.55068845e-01 -4.88872766e-01 4.69641417e-01
-8.08546364e-01 -6.29463017e-01 1.93121389e-01 1.69363946e-01
6.44303858e-01 -4.02835339e-01 -5.40565848e-01 -8.18613827e-01
-7.80332625e-01 -3.69285405e-01 4.01321650e-01 -4.89377081e-01
-1.09909201e+00 6.38352036e-02 7.63105989e-01 -1.21381009e+00
-7.11954832e-01 -1.85768604e-01 -3.67319793e-01 8.73427272e-01
-1.13065410e+00 -7.58150160e-01 -4.05390948e-01 5.48359156e-01
3.34555715e-01 1.51347175e-01 7.87145138e-01 2.30539721e-02
-1.74000952e-03 1.11590378e-01 5.16424358e-01 -1.77559450e-01
2.57566601e-01 -9.27194476e-01 5.72387040e-01 1.01466286e+00
-4.06951070e-01 7.61362970e-01 1.27582598e+00 -1.24912298e+00
-1.06738842e+00 -9.38449621e-01 6.34350479e-01 2.94517651e-02
3.47715884e-01 -1.30378112e-01 -3.39648843e-01 7.08502829e-01
1.25496373e-01 -3.62115890e-01 9.16598737e-01 -1.18982449e-01
4.36386526e-01 3.68347135e-03 -1.27052796e+00 4.19382304e-01
4.45724517e-01 -1.23409890e-01 -3.42367560e-01 1.98827267e-01
-1.19335011e-01 -1.35856822e-01 -9.44338620e-01 4.63816345e-01
9.15768266e-01 -9.87497449e-01 8.85557830e-01 4.74167585e-01
-6.80579424e-01 -5.48362434e-01 -4.24015403e-01 -6.73378825e-01
-2.18366906e-01 -1.10364318e+00 -1.91590235e-01 9.69704628e-01
2.44614482e-02 -6.32978022e-01 1.05095971e+00 1.95027545e-01
2.22378448e-01 -2.33906165e-01 -1.07304692e+00 -7.18556643e-01
-4.39419687e-01 -8.78863275e-01 6.57912731e-01 1.08340301e-01
-3.00758958e-01 5.30933328e-02 -4.08038437e-01 1.70856550e-01
1.45890391e+00 1.04419865e-01 9.12761211e-01 -1.21128130e+00
-3.48470688e-01 1.57930747e-01 -7.78213190e-03 -1.56362927e+00
-7.05900967e-01 -2.35467479e-01 8.12197506e-01 -1.64895725e+00
3.86577755e-01 -5.39537251e-01 2.19228864e-01 -1.01915002e-01
-2.34675065e-01 2.88549066e-01 -2.38071576e-01 2.92597443e-01
-6.00411713e-01 4.54040468e-01 7.71883070e-01 4.04868335e-01
-1.10606544e-01 5.54728925e-01 -7.25021958e-02 7.83514559e-01
8.34272265e-01 -7.93639421e-01 -2.81495929e-01 1.54891789e-01
2.64590085e-01 1.32173240e-01 2.60527939e-01 -1.28246546e+00
2.63144732e-01 -2.92384684e-01 2.92788923e-01 -1.48067153e+00
4.51778144e-01 -7.02118337e-01 6.93711102e-01 6.79385722e-01
1.33333281e-01 -2.61897594e-01 -2.76186377e-01 1.12693810e+00
1.34811312e-01 -5.40975630e-01 1.36259043e+00 -4.04703110e-01
-3.49877447e-01 1.98086977e-01 -1.24767673e+00 -1.52540840e-02
1.18464124e+00 -7.42585897e-01 4.34371889e-01 -4.80818838e-01
-9.72835124e-01 8.20473069e-04 4.60656554e-01 -3.82461965e-01
6.25264764e-01 -1.16601169e+00 -5.75382710e-01 3.63098025e-01
-1.11459322e-01 1.17949523e-01 1.27041072e-01 6.86497569e-01
-6.29638612e-01 2.29688674e-01 2.21227050e-01 -9.86050248e-01
-9.38357770e-01 -1.07463822e-02 2.12334041e-02 -3.58959317e-01
-4.14055049e-01 7.45519757e-01 -4.11921352e-01 -1.06117576e-01
1.64252356e-01 -4.80066687e-01 1.66212440e-01 -1.68329060e-01
5.58727622e-01 9.95736659e-01 -1.85166284e-01 -2.04448566e-01
-5.10982811e-01 5.84333777e-01 1.93423435e-01 -8.80163252e-01
8.00446570e-01 -6.04722738e-01 2.03610331e-01 9.05604899e-01
5.24492860e-01 -7.52812028e-02 -1.38631177e+00 -2.35207640e-02
-1.74198866e-01 -2.53987908e-01 -2.97877431e-01 -3.82511973e-01
-3.27826649e-01 8.20487082e-01 7.78243124e-01 2.66937405e-01
8.52448642e-01 -1.20452732e-01 7.66611993e-01 1.75243974e-01
8.22472215e-01 -1.34295404e+00 -2.29918867e-01 4.58693087e-01
-4.25297096e-02 -8.26533556e-01 6.57028139e-01 -3.26127857e-01
-2.19759896e-01 7.73461998e-01 -1.69050664e-01 -3.40598106e-01
1.38823617e+00 5.61202109e-01 -1.99535191e-01 -1.19940184e-01
-3.63964260e-01 -3.45156908e-01 -4.29579854e-01 8.78199399e-01
4.53343481e-01 4.56197143e-01 -3.82178873e-01 1.45248666e-01
-2.01470897e-01 3.49074334e-01 5.94969869e-01 1.46848226e+00
-1.06535172e+00 -9.46445942e-01 -9.14591134e-01 7.37944484e-01
-4.60656404e-01 -1.04347371e-01 4.07456428e-01 5.69846988e-01
-2.98280120e-01 1.34086668e+00 1.66748598e-01 4.19666618e-02
1.55024901e-02 -2.26470921e-02 4.96792108e-01 -6.01530731e-01
-1.42375275e-01 6.86913669e-01 9.50275138e-02 -4.56569791e-01
-9.24669981e-01 -1.44610417e+00 -1.24899685e+00 -4.39732701e-01
-7.51620650e-01 4.32531863e-01 8.43590260e-01 8.90114903e-01
1.78447723e-01 1.92049742e-01 2.95612395e-01 -1.00966120e+00
-5.72249353e-01 -1.26885056e+00 -1.09304726e+00 -1.77534014e-01
1.54090914e-04 -6.81901455e-01 -2.80488551e-01 2.33481959e-01]
|
[6.7456769943237305, 3.939074993133545]
|
46175c77-5f53-4b60-b0a1-907a1a02de51
|
trig-transformer-based-text-recognizer-with
|
2111.08314
| null |
https://arxiv.org/abs/2111.08314v1
|
https://arxiv.org/pdf/2111.08314v1.pdf
|
TRIG: Transformer-Based Text Recognizer with Initial Embedding Guidance
|
Scene text recognition (STR) is an important bridge between images and text, attracting abundant research attention. While convolutional neural networks (CNNS) have achieved remarkable progress in this task, most of the existing works need an extra module (context modeling module) to help CNN to capture global dependencies to solve the inductive bias and strengthen the relationship between text features. Recently, the transformer has been proposed as a promising network for global context modeling by self-attention mechanism, but one of the main shortcomings, when applied to recognition, is the efficiency. We propose a 1-D split to address the challenges of complexity and replace the CNN with the transformer encoder to reduce the need for a context modeling module. Furthermore, recent methods use a frozen initial embedding to guide the decoder to decode the features to text, leading to a loss of accuracy. We propose to use a learnable initial embedding learned from the transformer encoder to make it adaptive to different input images. Above all, we introduce a novel architecture for text recognition, named TRansformer-based text recognizer with Initial embedding Guidance (TRIG), composed of three stages (transformation, feature extraction, and prediction). Extensive experiments show that our approach can achieve state-of-the-art on text recognition benchmarks.
|
['Shugong Xu', 'Runze Ma', 'Zhiwei Jia', 'Yue Tao']
|
2021-11-16
| null | null | null | null |
['scene-text-recognition']
|
['computer-vision']
|
[ 5.99934518e-01 -3.81688565e-01 -1.87001318e-01 -5.98954618e-01
-2.54395515e-01 -6.44437969e-02 7.43521512e-01 -1.63974985e-01
-3.43789876e-01 9.62426066e-02 2.37895250e-01 -1.90325037e-01
3.35844427e-01 -7.43892014e-01 -6.64220929e-01 -7.00367093e-01
9.40429807e-01 1.11677133e-01 4.32053894e-01 -1.01288088e-01
5.20009696e-01 2.04593793e-01 -1.38775587e+00 6.93424404e-01
1.02072394e+00 1.22509396e+00 5.18126786e-01 4.01468396e-01
-7.23259568e-01 1.13705456e+00 -3.96739572e-01 -5.22028804e-01
-1.96994916e-02 -5.51173270e-01 -6.18967354e-01 2.84082264e-01
3.70130002e-01 -4.19444293e-01 -7.83933461e-01 8.01733196e-01
4.69318479e-01 -1.17196711e-02 5.38791001e-01 -9.35397863e-01
-1.00107300e+00 6.86428070e-01 -5.50299525e-01 5.47622368e-02
1.40589610e-01 4.62790541e-02 9.45736468e-01 -1.30584586e+00
3.31451327e-01 1.15614712e+00 6.54826403e-01 5.29432416e-01
-8.31069410e-01 -5.56157053e-01 6.24647796e-01 6.34569347e-01
-1.41084564e+00 -5.30493617e-01 9.92337525e-01 -2.02544436e-01
1.11578155e+00 1.36053801e-01 5.56527495e-01 1.34730148e+00
2.93137759e-01 1.26839912e+00 7.51706541e-01 -5.34540594e-01
-4.29602712e-02 4.91154455e-02 2.09070638e-01 6.84381723e-01
-6.78358376e-02 -4.41480786e-01 -6.63595438e-01 3.42845619e-01
7.19453335e-01 3.57908338e-01 -3.60549182e-01 -7.09901154e-02
-1.04531026e+00 5.38841605e-01 5.44465959e-01 3.76953959e-01
-8.78212377e-02 1.07916631e-01 4.87947017e-01 4.02808696e-01
3.75991762e-01 -1.45405710e-01 -2.82932222e-01 -8.58104229e-02
-7.63774633e-01 -3.16860229e-01 5.73025405e-01 9.44347978e-01
6.93360090e-01 2.15692192e-01 -3.72155517e-01 1.04032981e+00
3.41350436e-01 4.51922983e-01 8.26195300e-01 2.18569964e-01
1.04248750e+00 1.29625797e+00 -4.92685199e-01 -1.13446081e+00
-7.76826516e-02 -4.69666630e-01 -1.23745906e+00 -4.89870310e-01
3.47267091e-02 1.26720086e-01 -1.00215864e+00 1.19497275e+00
5.92637248e-02 3.78911972e-01 -8.97038355e-02 7.57445097e-01
8.15690339e-01 8.98887932e-01 -6.07118197e-02 4.98100854e-02
1.25388455e+00 -1.39094758e+00 -7.15118587e-01 -5.38139760e-01
7.52506256e-01 -8.33954036e-01 1.21270657e+00 1.58433452e-01
-6.11355901e-01 -6.95205629e-01 -1.12942493e+00 -4.89079237e-01
-6.01424932e-01 5.28931737e-01 1.89097628e-01 4.75220621e-01
-7.28691936e-01 3.60454828e-01 -8.37863207e-01 -5.57266295e-01
4.39721078e-01 4.23713416e-01 -1.93333805e-01 -3.15796584e-01
-1.00192225e+00 7.40890443e-01 4.53403711e-01 5.68699718e-01
-5.87276757e-01 -1.08809747e-01 -8.52358818e-01 3.15250546e-01
4.62140381e-01 -4.55408424e-01 8.40676010e-01 -1.40976739e+00
-1.84117353e+00 4.64695066e-01 -3.95724118e-01 -1.88946277e-01
3.82631063e-01 -2.27622584e-01 -3.97661954e-01 -4.76238579e-02
-3.32022041e-01 4.09047961e-01 1.20959330e+00 -8.06673348e-01
-6.17912531e-01 -2.48372197e-01 -1.81988969e-01 3.31349760e-01
-8.53928745e-01 1.67261899e-01 -8.09567511e-01 -8.94691646e-01
2.68870980e-01 -6.96827590e-01 1.57130621e-02 -2.65285037e-02
-4.11427349e-01 -4.77145195e-01 1.34118092e+00 -6.21502638e-01
1.41516781e+00 -2.16227651e+00 1.43005967e-01 -1.94397941e-01
-2.52074227e-02 4.27922577e-01 -3.28597337e-01 4.61308122e-01
4.41487059e-02 3.06517221e-02 -1.45675018e-01 -6.33336186e-01
2.86734533e-02 3.00154001e-01 -7.37163842e-01 2.52372533e-01
5.41889071e-01 1.27512205e+00 -5.08805335e-01 -4.97182220e-01
4.30994868e-01 6.46411538e-01 -4.19283122e-01 2.49781027e-01
-3.08591247e-01 1.08298220e-01 -6.13436520e-01 6.18063092e-01
6.56081080e-01 -4.31814969e-01 2.26132840e-01 -3.42462242e-01
-7.74606466e-02 4.23906356e-01 -1.05325043e+00 1.56155503e+00
-3.51462930e-01 6.55986547e-01 -2.41406262e-01 -1.34857619e+00
1.20656407e+00 2.09736004e-01 1.49156913e-01 -7.49846220e-01
4.15358663e-01 1.03005178e-01 -2.95345426e-01 -6.22433722e-01
5.02912760e-01 1.31383166e-01 1.81965515e-01 3.43578666e-01
1.90137625e-02 2.67660320e-01 -2.87475407e-01 -1.30230099e-01
9.56314981e-01 2.31454775e-01 3.56012024e-02 3.38769443e-02
9.67792273e-01 -2.18010604e-01 5.06457329e-01 4.14828688e-01
6.68776631e-02 7.14471579e-01 4.04567778e-01 -7.81840086e-01
-8.18462908e-01 -3.21444869e-01 1.35838808e-02 1.11370265e+00
2.38159448e-01 -5.99059641e-01 -5.88176370e-01 -9.59235251e-01
-2.64987081e-01 3.55424672e-01 -7.67850816e-01 -2.34895647e-01
-8.07270527e-01 -6.27353013e-01 5.15514135e-01 7.99434900e-01
1.08608246e+00 -9.81606841e-01 -2.43289933e-01 1.96745247e-01
-2.72398233e-01 -1.35609567e+00 -7.48327136e-01 3.36382359e-01
-9.37679291e-01 -8.20262015e-01 -6.33981526e-01 -1.07695222e+00
8.65458369e-01 4.37968403e-01 4.73380566e-01 3.62342656e-01
-5.14686666e-02 2.74808705e-01 -6.17095709e-01 -1.58911850e-02
-6.71646819e-02 3.45115930e-01 -3.92997891e-01 5.40986955e-01
4.67140198e-01 -3.54049534e-01 -4.66161549e-01 6.44287944e-01
-1.20143199e+00 5.03619909e-01 7.26714253e-01 1.19508052e+00
6.00223839e-01 1.59940105e-02 4.12520170e-01 -6.80073082e-01
3.72025937e-01 -9.04136449e-02 -4.01211262e-01 5.83192647e-01
-5.54895163e-01 1.21544495e-01 1.14181530e+00 -5.24768949e-01
-1.25091422e+00 1.55103922e-01 -1.00348346e-01 -4.99395013e-01
-5.09117842e-02 7.00091600e-01 -3.69816899e-01 -1.84453085e-01
1.73874363e-01 9.40854609e-01 -2.52183914e-01 -6.46819234e-01
1.20383017e-01 9.62780952e-01 1.08822890e-01 -4.97180909e-01
6.03258371e-01 3.49185169e-01 -1.34328097e-01 -8.54529738e-01
-9.34808135e-01 -1.64252877e-01 -8.78400207e-01 6.14223555e-02
7.64546692e-01 -8.75317395e-01 -4.95559692e-01 6.77012384e-01
-1.18780518e+00 -4.30881947e-01 7.44258193e-03 2.30095670e-01
-2.67020226e-01 6.84553623e-01 -5.40145278e-01 -4.58894789e-01
-5.50239444e-01 -1.26165998e+00 1.16207266e+00 2.32301265e-01
4.89547729e-01 -9.63477194e-01 -2.73082197e-01 2.58106321e-01
5.94810426e-01 -2.74759263e-01 8.33479822e-01 -5.78866780e-01
-8.19790959e-01 -1.79332823e-01 -5.60252607e-01 4.80812311e-01
2.23080099e-01 -6.89189658e-02 -9.45218265e-01 -2.81486124e-01
1.33797944e-01 -4.00134444e-01 1.11503279e+00 -5.86727709e-02
1.45353639e+00 -4.08162743e-01 -3.95071685e-01 7.60274112e-01
1.33405030e+00 4.73033972e-02 8.60273778e-01 3.20035815e-01
1.00120723e+00 2.59009480e-01 2.57258087e-01 3.71852696e-01
5.76504886e-01 6.99633837e-01 1.91363126e-01 1.55363858e-01
-2.87511021e-01 -5.24205923e-01 6.84791267e-01 1.35231125e+00
2.07363769e-01 -3.95992815e-01 -8.90182734e-01 2.91729510e-01
-2.14559150e+00 -6.74099207e-01 1.55820623e-02 1.92443562e+00
6.65777802e-01 1.87700212e-01 -4.24803048e-01 7.74043426e-02
7.83102334e-01 2.81968623e-01 -5.76277375e-01 -2.53178686e-01
-2.27778673e-01 -3.56512666e-02 1.50435001e-01 2.71263033e-01
-9.50856209e-01 1.30951583e+00 5.61923504e+00 1.10561180e+00
-1.57399631e+00 -1.83327943e-01 6.02148414e-01 2.56828487e-01
-7.92914033e-02 1.04908444e-01 -1.02147317e+00 4.73090738e-01
6.21904671e-01 3.24707888e-02 3.45435143e-01 7.68845856e-01
-1.54312525e-03 2.98040271e-01 -1.16874361e+00 1.10550773e+00
5.13919413e-01 -1.23387504e+00 6.52402520e-01 -2.70996600e-01
6.06168926e-01 6.29438926e-03 1.16246538e-02 4.95879591e-01
-2.73759991e-01 -9.37735856e-01 5.75205147e-01 5.45563340e-01
8.93651545e-01 -5.91793001e-01 7.62986720e-01 4.47269320e-01
-1.60560155e+00 -1.39813572e-01 -7.41575062e-01 4.85681593e-02
-2.59411365e-01 6.08422279e-01 -7.23183513e-01 7.18681037e-01
4.35504377e-01 1.24802661e+00 -8.61376226e-01 6.84399724e-01
-2.69093245e-01 6.18090093e-01 -3.15899640e-01 -3.32461774e-01
3.58327121e-01 -7.44799972e-02 2.63305604e-01 1.34797120e+00
2.31900498e-01 5.88646904e-03 2.76690394e-01 8.03822160e-01
-3.10162067e-01 3.84379447e-01 -4.70978141e-01 -1.17391489e-01
1.77719191e-01 1.07573318e+00 -7.42188931e-01 -4.34880465e-01
-6.16678178e-01 1.21822262e+00 4.20651466e-01 3.08316350e-01
-7.51091063e-01 -6.90111041e-01 1.20424636e-01 -1.36317164e-01
6.27892494e-01 -2.94626653e-01 -4.46721673e-01 -1.65666318e+00
4.16665435e-01 -9.15035486e-01 2.71520317e-01 -7.91308939e-01
-1.07532465e+00 5.51043391e-01 -5.32705367e-01 -1.27256358e+00
2.34333694e-01 -8.16415608e-01 -6.34485722e-01 7.86695480e-01
-1.85181010e+00 -1.51311243e+00 -5.12379348e-01 8.50330114e-01
1.08013046e+00 -1.96434245e-01 4.56924051e-01 2.63938487e-01
-1.06547177e+00 9.77571487e-01 7.11624324e-02 4.33176577e-01
7.47939944e-01 -7.35895932e-01 4.96554255e-01 8.54846358e-01
1.07555948e-01 5.58315575e-01 -3.59963030e-02 -7.03231394e-01
-1.99274170e+00 -1.36779177e+00 9.98969913e-01 -3.89090747e-01
5.76657295e-01 -6.10048652e-01 -1.09958100e+00 8.76828432e-01
3.07984501e-01 9.90944058e-02 4.05240715e-01 -3.36504169e-02
-5.53050637e-01 -3.88652503e-01 -6.29639328e-01 9.13577259e-01
1.01838529e+00 -6.07370436e-01 -6.25755012e-01 -4.22739517e-03
7.28233814e-01 -3.82162601e-01 -5.92211962e-01 2.73189902e-01
5.28882861e-01 -8.09172273e-01 4.73144174e-01 -2.08822817e-01
7.15028465e-01 -3.33370745e-01 -2.13801548e-01 -8.20829272e-01
-2.98146248e-01 -4.31362212e-01 -1.33143395e-01 1.38247705e+00
1.86886951e-01 -7.96287835e-01 7.87934542e-01 3.15122187e-01
-2.77006984e-01 -9.60653484e-01 -1.05170548e+00 -6.04806840e-01
-3.65519635e-02 -4.70827609e-01 6.29049361e-01 1.16176164e+00
1.53913125e-01 6.55571461e-01 -4.45805460e-01 2.50273049e-02
1.34316787e-01 2.07361296e-01 7.51512527e-01 -8.94795358e-01
-7.61071977e-04 -6.03956342e-01 -3.67230296e-01 -1.95814288e+00
9.98676196e-02 -1.01464057e+00 7.36431479e-02 -1.48865569e+00
3.92263353e-01 -2.86778957e-01 -3.17231864e-01 8.54434371e-01
-2.95893937e-01 -1.06733859e-01 2.67040461e-01 3.17954302e-01
-8.07040870e-01 1.09334040e+00 1.31312835e+00 -4.49087828e-01
-3.38348225e-02 -2.96835572e-01 -6.37082398e-01 5.14528096e-01
5.62095761e-01 -3.56153309e-01 -2.47349232e-01 -1.01746106e+00
3.14953148e-01 -2.52395540e-01 2.43306249e-01 -9.51413453e-01
7.84898758e-01 -1.76583499e-01 5.41508734e-01 -9.13796961e-01
1.27879471e-01 -1.14862812e+00 -2.73724139e-01 1.76892370e-01
-4.56304133e-01 -3.40139605e-02 2.43978471e-01 5.81478596e-01
-4.19353813e-01 -2.78087050e-01 6.35496557e-01 2.84411311e-01
-6.19847775e-01 4.33048338e-01 -2.89561898e-01 -2.41045132e-01
5.85069001e-01 -3.47076118e-01 -4.72961724e-01 -6.29327372e-02
-2.12695941e-01 3.09583426e-01 2.76839346e-01 5.59003949e-01
9.48236585e-01 -1.30894637e+00 -6.48812354e-01 4.08299834e-01
1.93655521e-01 1.16850883e-01 2.59439290e-01 8.64419281e-01
-2.30761811e-01 6.65241718e-01 1.87019750e-01 -7.36229181e-01
-1.03644621e+00 4.69508857e-01 3.66513133e-01 -5.03334939e-01
-9.24059272e-01 6.38362288e-01 3.44856858e-01 -3.50397557e-01
4.01320219e-01 -6.33101881e-01 -1.70388758e-01 -1.52488098e-01
5.99610269e-01 9.61895194e-03 1.33124933e-01 -4.64454025e-01
-2.06995472e-01 1.01073480e+00 -5.26692569e-01 4.33517098e-01
1.26873076e+00 -2.61585504e-01 -1.23269811e-01 3.47202629e-01
1.12198162e+00 -2.84267783e-01 -1.31104064e+00 -7.04009652e-01
-7.89488852e-02 -3.50949913e-01 1.65503591e-01 -6.95309341e-01
-1.17163765e+00 1.25244331e+00 3.86160702e-01 1.15195043e-01
1.31556678e+00 -4.81305212e-01 9.30939674e-01 9.15678918e-01
-5.98304048e-02 -1.12358546e+00 3.97480220e-01 1.02953124e+00
1.00714505e+00 -1.23115373e+00 -1.33572206e-01 -3.77437145e-01
-5.80396652e-01 1.59520316e+00 8.23133111e-01 -1.02105603e-01
6.73032343e-01 2.09619537e-01 -1.68925777e-01 9.16707665e-02
-9.83670294e-01 -4.45894487e-02 2.69659370e-01 2.60774285e-01
4.43893850e-01 -4.23577100e-01 -4.93451692e-02 8.47508073e-01
3.25628638e-01 2.00787202e-01 2.67562985e-01 9.92249727e-01
-4.62635607e-01 -1.27740657e+00 -1.17352143e-01 5.36347628e-01
-2.35265791e-01 -3.82256985e-01 -5.04906952e-01 4.97424006e-01
-4.02477458e-02 6.56292856e-01 4.30730060e-02 -6.22986197e-01
3.88846159e-01 3.67643148e-01 3.31664443e-01 -6.01342618e-01
-6.46265030e-01 1.12597413e-01 -2.85853922e-01 -2.61243671e-01
-2.75513172e-01 -4.34708536e-01 -1.03128123e+00 -1.37948021e-01
-7.10497618e-01 -1.71704218e-01 5.66945612e-01 1.12797880e+00
5.49884498e-01 6.01250947e-01 9.93164122e-01 -5.41140974e-01
-3.83356065e-01 -1.04679668e+00 -3.05185020e-01 1.17696509e-01
1.52471960e-01 -3.61174464e-01 -2.92012244e-01 3.45339596e-01]
|
[11.855624198913574, 2.160369873046875]
|
86e3a04d-ef47-4a1c-a4c1-0b0df7ed18db
|
diagnostic-test-accuracy-dta-of-artificial
|
2306.07999
| null |
https://arxiv.org/abs/2306.07999v2
|
https://arxiv.org/pdf/2306.07999v2.pdf
|
Artificial intelligence in digital pathology: a diagnostic test accuracy systematic review and meta-analysis
|
Ensuring diagnostic performance of AI models before clinical use is key to the safe and successful adoption of these technologies. Studies reporting AI applied to digital pathology images for diagnostic purposes have rapidly increased in number in recent years. The aim of this work is to provide an overview of the diagnostic accuracy of AI in digital pathology images from all areas of pathology. This systematic review and meta-analysis included diagnostic accuracy studies using any type of artificial intelligence applied to whole slide images (WSIs) in any disease type. The reference standard was diagnosis through histopathological assessment and / or immunohistochemistry. Searches were conducted in PubMed, EMBASE and CENTRAL in June 2022. We identified 2976 studies, of which 100 were included in the review and 48 in the full meta-analysis. Risk of bias and concerns of applicability were assessed using the QUADAS-2 tool. Data extraction was conducted by two investigators and meta-analysis was performed using a bivariate random effects model. 100 studies were identified for inclusion, equating to over 152,000 whole slide images (WSIs) and representing many disease types. Of these, 48 studies were included in the meta-analysis. These studies reported a mean sensitivity of 96.3% (CI 94.1-97.7) and mean specificity of 93.3% (CI 90.5-95.4) for AI. There was substantial heterogeneity in study design and all 100 studies identified for inclusion had at least one area at high or unclear risk of bias. This review provides a broad overview of AI performance across applications in whole slide imaging. However, there is huge variability in study design and available performance data, with details around the conduct of the study and make up of the datasets frequently missing. Overall, AI offers good accuracy when applied to WSIs but requires more rigorous evaluation of its performance.
|
['Deborah D Stocken', 'Emily L Clarke', 'Darren Treanor', 'Henschel Freduah-Agyemang', 'Caroline Cartlidge', 'Gillian Matthews', 'Charlotte Jennings', 'Clare McGenity']
|
2023-06-13
| null | null | null | null |
['whole-slide-images', 'specificity']
|
['computer-vision', 'natural-language-processing']
|
[ 4.91206616e-01 -3.61913294e-02 -7.21156240e-01 3.12904306e-02
-1.15733755e+00 -6.14352345e-01 2.46920273e-01 3.94019544e-01
-5.57012200e-01 4.70189989e-01 -1.16105941e-04 -6.38741136e-01
-7.02531099e-01 -4.90367502e-01 -4.54049021e-01 -1.00288594e+00
-1.21259451e-01 7.19788194e-01 1.23541161e-01 7.26910710e-01
2.86378533e-01 7.22491801e-01 -9.84375656e-01 3.48941565e-01
6.06487989e-01 6.47330642e-01 1.37201503e-01 1.17197585e+00
-2.09732667e-01 8.35253775e-01 -7.85599113e-01 -4.06612903e-01
1.18022941e-01 -3.14810187e-01 -8.22805464e-01 -5.78832626e-03
2.97027886e-01 -3.80750477e-01 -1.59084737e-01 5.51676869e-01
7.81408370e-01 -6.89037919e-01 1.04975402e+00 -1.07979310e+00
-5.31072259e-01 3.83070529e-01 -7.54408538e-01 7.13023782e-01
1.35451376e-01 6.26442730e-01 5.49785197e-01 -4.82633084e-01
1.01099169e+00 7.01491058e-01 7.98476756e-01 3.54489774e-01
-9.69853878e-01 -9.22947466e-01 -5.80849349e-01 1.04970252e-02
-1.29536295e+00 -2.17688054e-01 8.09238926e-02 -6.68518782e-01
1.11493957e+00 5.76643169e-01 1.01826239e+00 3.48873794e-01
7.78167725e-01 3.54072958e-01 1.29036534e+00 -5.46908140e-01
-4.17638905e-02 2.09065318e-01 -2.22152956e-02 3.27063084e-01
8.43030095e-01 -1.64887398e-01 8.53382498e-02 -5.66450119e-01
6.27005339e-01 1.15893027e-02 -5.37975915e-02 1.26546741e-01
-1.37189388e+00 5.97041249e-01 3.51654649e-01 5.14806390e-01
-2.10184231e-01 -1.34772539e-01 8.41492772e-01 1.63588934e-02
3.68443340e-01 4.62742865e-01 1.14593692e-01 9.50634256e-02
-9.34700370e-01 -7.70975351e-02 5.43751299e-01 3.68093312e-01
-2.63388723e-01 -2.11820096e-01 2.35600621e-02 9.32301104e-01
1.49733439e-01 7.33855009e-01 8.32749844e-01 -1.00203693e+00
2.15071719e-02 7.41145730e-01 -2.11090878e-01 -8.51720154e-01
-8.93687904e-01 -1.47098243e-01 -1.06035817e+00 2.19543040e-01
3.10452312e-01 -5.34430631e-02 -1.31940961e+00 1.08804739e+00
8.90116245e-02 -3.48288834e-01 6.03390671e-02 7.80974984e-01
9.61089075e-01 4.82997984e-01 5.95989883e-01 -2.57372260e-01
1.81706858e+00 -4.93734896e-01 -8.41521382e-01 8.32362324e-02
1.00877154e+00 -5.34788489e-01 7.74323404e-01 4.92761493e-01
-1.29254234e+00 1.82608053e-01 -1.12784386e+00 -1.61730107e-02
-4.94931996e-01 1.73036978e-01 5.26477396e-01 6.59938216e-01
-1.19084978e+00 -1.86657637e-01 -1.00230527e+00 -7.33486831e-01
7.54521251e-01 7.30019748e-01 -6.53465450e-01 -4.52679507e-02
-7.26139367e-01 1.29127240e+00 1.82038516e-01 -1.44815460e-01
-2.76593268e-01 -1.09919643e+00 -4.71941292e-01 -4.44616169e-01
1.18942790e-01 -9.47383046e-01 1.02038693e+00 -9.76177096e-01
-8.63450706e-01 1.47962427e+00 -3.02707762e-01 -4.38457787e-01
2.93409169e-01 6.80226505e-01 -4.24381763e-01 5.71323156e-01
2.80390501e-01 5.57257712e-01 -1.41688928e-01 -8.45771670e-01
-1.00100255e+00 -6.61035180e-01 -3.78139734e-01 3.86636376e-01
-1.36716977e-01 3.18890631e-01 -5.36822200e-01 -6.82601035e-01
-2.72990286e-01 -1.03124881e+00 -3.95697594e-01 2.26725340e-01
5.48284575e-02 -6.38086498e-02 5.69393039e-01 -6.97745621e-01
1.34408391e+00 -1.93559241e+00 -4.52172786e-01 3.16394538e-01
4.06012446e-01 3.97172511e-01 1.79521650e-01 2.17204630e-01
-1.13108121e-01 6.82791233e-01 -2.40312945e-02 7.08353966e-02
-3.90695244e-01 -1.05220169e-01 3.84345621e-01 9.52714086e-01
1.73487008e-01 1.15355659e+00 -5.50182939e-01 -8.73596549e-01
4.19823110e-01 4.71678197e-01 -3.01134754e-02 -4.66379225e-02
6.35277927e-01 -2.75072992e-01 -2.68504649e-01 9.34686303e-01
5.30077517e-01 -6.04969919e-01 2.95314491e-01 -1.06561095e-01
3.14352810e-02 1.30589619e-01 -8.07980061e-01 1.23948216e+00
-3.50102395e-01 7.24934042e-01 2.74406403e-01 -5.86894453e-01
5.63632488e-01 5.62709033e-01 7.26070821e-01 -5.44299901e-01
8.63922536e-02 4.23416883e-01 5.78437626e-01 -6.08732224e-01
-2.41500899e-01 -4.65309978e-01 1.71500772e-01 5.30748904e-01
-6.88431561e-02 -3.68293464e-01 2.70303249e-01 2.29813755e-01
1.43210936e+00 -6.83541894e-01 7.26933241e-01 -2.14159191e-01
3.82629067e-01 6.68252885e-01 2.83801913e-01 8.61228228e-01
-4.33899224e-01 6.35550380e-01 4.97103840e-01 -3.43001515e-01
-1.09528184e+00 -9.33889568e-01 -6.85079277e-01 1.37332454e-01
-3.69085997e-01 -5.43453284e-02 -4.80630666e-01 -4.81174976e-01
2.72786394e-02 1.47653013e-01 -9.35793519e-01 -2.89522409e-02
-1.32272497e-01 -1.15773535e+00 8.80364776e-01 6.01515710e-01
1.94241300e-01 -8.85731101e-01 -1.05502117e+00 1.12166069e-01
1.82033762e-01 -7.82505155e-01 -2.59774681e-02 -1.65794179e-01
-9.97881114e-01 -1.16116476e+00 -1.23523128e+00 -8.10504436e-01
7.39088714e-01 5.99221177e-02 7.86749423e-01 3.95347476e-01
-9.41991925e-01 3.91817361e-01 1.37090325e-01 -9.33248222e-01
-6.03013992e-01 -6.89093769e-02 -1.85112670e-01 -8.31182837e-01
6.67524874e-01 9.13027599e-02 -7.72165418e-01 -4.32689860e-03
-1.19022501e+00 -2.22387359e-01 1.40574849e+00 8.87176335e-01
7.85773695e-01 -1.23631604e-01 5.34391701e-01 -1.01159871e+00
3.04684192e-01 -4.92840528e-01 -1.99363202e-01 2.67288029e-01
-6.53081357e-01 -7.79151380e-01 1.33810505e-01 -4.42928135e-01
-7.60581374e-01 -3.18445563e-01 1.76361799e-01 -5.73937260e-02
-3.16521078e-01 8.04770052e-01 4.49853122e-01 -1.33284584e-01
5.42782903e-01 -3.20386201e-01 8.00241768e-01 1.65801466e-01
-5.69925547e-01 8.86934042e-01 2.51289487e-01 -2.51860172e-03
2.03005761e-01 7.16142356e-01 3.53729695e-01 -7.51384437e-01
-2.49915570e-01 -7.05303669e-01 -1.39676929e-01 4.19723988e-02
5.97339749e-01 -7.46604860e-01 -6.57253325e-01 4.74357486e-01
-3.06264311e-01 -4.66004848e-01 -1.43311173e-01 8.14648807e-01
-1.33574128e-01 9.93720070e-03 -9.64924574e-01 -5.34427583e-01
-1.01087391e+00 -1.66360593e+00 7.88386464e-01 2.67400950e-01
-6.56392395e-01 -1.02643991e+00 2.15133861e-01 1.10824004e-01
4.95735466e-01 3.64296168e-01 8.09933126e-01 -8.10965896e-01
2.37216651e-01 -7.81709433e-01 -5.36793768e-01 -4.64578062e-01
7.02163696e-01 6.64128065e-01 -6.08774304e-01 -5.03245555e-02
-2.30921715e-01 3.60169709e-02 4.45376456e-01 1.18779325e+00
1.02743936e+00 -3.03450406e-01 -8.94271851e-01 2.78379023e-01
1.58355892e+00 1.00876141e+00 6.95506752e-01 7.05610096e-01
1.61063343e-01 6.95974827e-01 6.21945739e-01 -1.02403127e-01
1.93611726e-01 4.77097154e-01 -2.34053452e-02 -2.39283875e-01
-2.64468759e-01 3.17210346e-01 -2.85803109e-01 2.07093656e-01
-1.96119398e-02 -2.00323805e-01 -1.48222840e+00 1.04469502e+00
-9.49065149e-01 -7.46714950e-01 -3.43306959e-01 2.05111003e+00
5.54242551e-01 2.36187860e-01 4.23752278e-01 1.23982474e-01
5.83706737e-01 -4.17288989e-01 -4.09869909e-01 -7.46147096e-01
-7.69942924e-02 -2.26005800e-02 7.16365814e-01 2.19008401e-01
-7.54253089e-01 2.87749916e-02 7.38435888e+00 7.28724837e-01
-1.28520691e+00 -1.84249341e-01 1.10999775e+00 -3.13967884e-01
6.73663840e-02 -2.59510607e-01 -4.70207721e-01 3.87392014e-01
1.37008333e+00 -4.64834630e-01 -4.97828513e-01 2.59343296e-01
3.00404072e-01 -7.30106950e-01 -4.87784386e-01 4.57200587e-01
6.67534173e-02 -1.64103603e+00 2.90600141e-03 5.25020599e-01
4.85515982e-01 -1.09842427e-01 4.80493963e-01 -2.61085361e-01
-4.73276188e-04 -1.07794213e+00 9.70345829e-03 2.73077041e-01
1.62576342e+00 -4.51583773e-01 1.32203567e+00 -2.40894720e-01
-6.06498182e-01 1.35987654e-01 -1.10948972e-01 2.14577496e-01
-9.43981260e-02 5.78346968e-01 -1.47345519e+00 4.29115832e-01
9.61389601e-01 4.72396225e-01 -5.52446365e-01 1.13123512e+00
4.00851637e-01 9.96982276e-01 -5.78236461e-01 -1.83134377e-01
2.60950923e-01 9.24921408e-02 2.44379684e-01 1.52576923e+00
1.71468019e-01 5.36561668e-01 -6.21501923e-01 2.48490885e-01
3.13096195e-01 2.21972853e-01 -6.38253629e-01 -3.60155046e-01
3.83235335e-01 1.34708321e+00 -1.22729814e+00 -5.38233161e-01
-6.76214933e-01 3.25000584e-01 -3.93289298e-01 5.03943190e-02
-4.16592538e-01 -5.39836466e-01 1.85407668e-01 3.66389573e-01
-3.74241471e-01 6.96265876e-01 -6.55755639e-01 -2.60635048e-01
-3.04554582e-01 -9.98197436e-01 1.11926925e+00 -8.33840787e-01
-1.10622418e+00 9.63182673e-02 2.23002702e-01 -1.04372883e+00
-3.21212858e-01 -7.11463988e-01 -3.85077417e-01 1.01493442e+00
-8.12528372e-01 -1.21908653e+00 -4.90828343e-02 -5.22185825e-02
2.38457441e-01 1.39450312e-01 9.31904972e-01 -9.90610719e-02
-4.16889787e-01 6.98864043e-01 1.91232145e-01 2.23099172e-01
1.16456437e+00 -1.25046265e+00 -2.92678803e-01 1.99245811e-01
-8.95037532e-01 7.94662058e-01 3.05076420e-01 -8.75097573e-01
-1.26875043e+00 -5.71884692e-01 7.77661204e-01 -3.06776106e-01
6.55375183e-01 4.18303937e-01 -8.28727901e-01 7.95351446e-01
2.71630973e-01 -9.82429236e-02 1.30601132e+00 -4.33159620e-01
1.36972696e-01 1.30314201e-01 -1.70087540e+00 6.79428458e-01
3.20494801e-01 -1.69445369e-02 -3.68284643e-01 -9.05927569e-02
1.11490138e-01 -6.60633087e-01 -1.41638148e+00 7.20210195e-01
1.14607513e+00 -4.66314524e-01 9.54786122e-01 -1.91362023e-01
4.93461698e-01 -1.91412672e-01 5.38672447e-01 -8.65982115e-01
-4.35883701e-01 1.96067691e-01 3.10957581e-01 9.27801192e-01
4.52304184e-01 -7.83609509e-01 8.54199111e-01 7.51081049e-01
-1.40319318e-02 -1.01708651e+00 -9.55816865e-01 -8.46182629e-02
4.12772804e-01 8.34929477e-03 2.89970726e-01 8.05521965e-01
4.30605739e-01 -3.08028102e-01 6.55571103e-01 -5.08158654e-02
5.15662253e-01 -2.36156657e-01 4.32499737e-01 -7.97049344e-01
2.92479008e-01 -8.52641940e-01 -8.15760612e-01 3.17132175e-01
-4.34513181e-01 -7.61052191e-01 -6.20362103e-01 -2.19818687e+00
8.15102577e-01 -3.99459869e-01 -1.86289176e-01 4.48633343e-01
-1.92280903e-01 4.99758571e-01 -1.12884752e-01 5.38819373e-01
-5.42029515e-02 -5.86363077e-01 1.07553172e+00 -4.21230078e-01
-7.96679780e-02 -2.63566345e-01 -8.64417255e-01 5.54147959e-01
1.00201118e+00 -3.94803613e-01 -2.11505309e-01 1.42003894e-01
8.53426084e-02 -1.21306945e-02 2.09683299e-01 -8.83090854e-01
3.59654963e-01 -2.29407057e-01 7.37742364e-01 -7.08254933e-01
7.36253560e-02 -8.97667646e-01 7.37278163e-01 9.76241589e-01
-2.60287017e-01 1.92944735e-01 6.30436897e-01 1.34305745e-01
-1.74155310e-01 -3.13701928e-01 6.21346295e-01 -2.00992465e-01
-4.34958905e-01 -2.64143683e-02 -1.02274847e+00 -1.88714638e-01
1.34951627e+00 -7.44316876e-01 -5.33667803e-01 -2.28332847e-01
-5.45005083e-01 3.01626623e-01 7.22402215e-01 -1.01369627e-01
3.83817106e-01 -1.09865236e+00 -7.20534682e-01 -5.59029758e-01
3.53738427e-01 1.20457590e-01 7.57087827e-01 1.58139658e+00
-9.81856883e-01 8.37769568e-01 -2.76130974e-01 -6.22896135e-01
-1.38402200e+00 6.69873536e-01 5.05085051e-01 -6.44296587e-01
-5.37525594e-01 4.00521576e-01 2.09605500e-01 -2.12428570e-01
9.92194097e-03 -2.38604724e-01 -1.20289102e-01 1.64312139e-01
9.04651761e-01 3.80976319e-01 2.98382372e-01 -5.88422120e-01
-7.45137691e-01 7.93217957e-01 -3.89845252e-01 -2.93078750e-01
1.05185163e+00 -1.56488121e-01 -1.69131279e-01 7.14603603e-01
1.36973202e+00 -1.62893072e-01 -2.77628392e-01 3.67498845e-01
-3.26629251e-01 -7.13493451e-02 -3.93698625e-02 -1.15921474e+00
-8.21514428e-01 5.96487522e-01 6.59352362e-01 -8.27993602e-02
1.27510142e+00 1.99998822e-02 6.59422651e-02 -3.26445192e-01
7.59146363e-02 -8.11160088e-01 -4.49149072e-01 -2.56113440e-01
6.25913501e-01 -1.07455552e+00 5.11432528e-01 -2.71066725e-01
-8.49562347e-01 1.10053253e+00 4.16262448e-01 -2.45923433e-03
5.74168503e-01 6.97057664e-01 2.22508878e-01 -5.09304583e-01
-8.24503064e-01 5.23273051e-01 2.54481196e-01 6.72340453e-01
7.24683106e-01 -6.60489826e-03 -1.03475273e+00 6.19858384e-01
1.77868932e-01 4.82027650e-01 6.42083824e-01 1.23107481e+00
1.26291990e-01 -4.69009310e-01 -6.30297482e-01 1.17439950e+00
-1.14494777e+00 4.29288715e-01 -6.34102583e-01 1.32276893e+00
-2.33356774e-01 5.88393271e-01 3.45234394e-01 1.07053593e-01
5.12997396e-02 -2.41650134e-01 2.09212795e-01 -2.02532008e-01
-8.13606441e-01 1.56806409e-01 1.83159541e-02 -1.42272636e-01
-7.99941242e-01 -8.18736851e-01 -1.35065401e+00 -3.26849401e-01
-5.26289999e-01 1.91327348e-01 7.16143489e-01 6.51688397e-01
2.35246405e-01 5.10662317e-01 -2.04603048e-03 -2.70279646e-01
5.00194691e-02 -8.74925554e-01 -5.38091004e-01 -1.31715626e-01
2.89637417e-01 -4.95610148e-01 -6.31925166e-01 -7.07329344e-03]
|
[15.14658260345459, -3.051274299621582]
|
37924a1a-c547-49d2-8002-6de0ee7ac7a5
|
heart-murmur-detection-from-phonocardiogram
| null | null |
https://www.medrxiv.org/content/10.1101/2022.08.11.22278688v1
|
https://www.medrxiv.org/content/10.1101/2022.08.11.22278688v1.full.pdf
|
Heart Murmur Detection from Phonocardiogram Recordings: The George B. Moody PhysioNet Challenge 2022
|
Objective Cardiac auscultation is an accessible diagnostic screening tool that can help to identify patients with heart murmurs for follow-up diagnostic screening and treatment, especially in resource-constrained environments. However, experts are needed to interpret the heart sound recordings, limiting the accessibility of auscultation for cardiac care. The George B. Moody PhysioNet Challenge 2022 invites teams to develop automated approaches for detecting abnormal heart function from multi-location phonocardiogram (PCG) recordings of heart sounds.
Approach For the Challenge, we sourced 5272 PCG recordings from 1568 pediatric patients in rural Brazil. We required the Challenge participants to submit the complete code for training and running their models, improving the transparency, reproducibility, and utility of the diagnostic algorithms. We devised a cost-based evaluation metric that captures the costs of screening, treatment, and diagnostic errors, allowing us to investigate the benefits of algorithmic pre-screening and facilitate the development of more clinically relevant algorithms.
Main results So far, over 80 teams have submitted over 600 algorithms during the course of the Challenge, representing a diversity of approaches in academia and industry. We will update this manuscript to share an analysis of the Challenge after the end of the Challenge.
Significance The use of heart sound recordings for both heart murmur detection and clinical outcome identification allowed us to explore the potential of automated approaches to provide accessible pre-screening of less-resourced populations. The submission of working, open-source algorithms and the use of novel evaluation metrics supported the reproducibility, generalizability, and relevance of the researched conducted during the Challenge.
|
['Gari D. Clifford', 'Ali Bahrami Rad', 'Reza Sameni', 'Miguel T. Coimbra', 'Sandra Mattos', 'ASHISH SHARMA', 'Nadi Sadr', 'Erick A. Perez Alday', 'Annie Gu', 'Francesco Renna', 'Jorge Oliveira', 'Andoni Elola', 'Yashar Kiarashi', 'Matthew A. Reyna']
|
2022-08-16
| null | null | null |
computing-in-cardiology-2022-8
|
['predict-clinical-outcome', 'classify-murmurs']
|
['time-series', 'time-series']
|
[ 3.90641928e-01 1.23395629e-01 2.65379310e-01 -2.22786695e-01
-1.00994229e+00 -7.24453628e-01 -5.42005599e-01 5.80357671e-01
-1.06582545e-01 4.47682500e-01 3.24105650e-01 -7.82435954e-01
-4.90068167e-01 -2.91330457e-01 -2.65299708e-01 -2.07308218e-01
-3.85140032e-01 6.23857439e-01 1.95919931e-01 5.76783538e-01
1.62739828e-01 3.16423297e-01 -1.18218184e+00 1.48060560e-01
8.97372484e-01 4.60789174e-01 1.54658049e-01 1.22162116e+00
4.49426800e-01 6.71111047e-01 -7.52068460e-01 -1.86911583e-01
-1.31211802e-02 -9.33863819e-01 -7.79300451e-01 -1.81537271e-01
5.21869123e-01 -6.14217281e-01 4.08198625e-01 1.99553862e-01
1.45787787e+00 -2.15734899e-01 2.39578277e-01 -5.77890933e-01
-8.49510357e-02 6.18241727e-01 1.80786088e-01 6.69330716e-01
4.49117988e-01 4.22243804e-01 7.27855384e-01 -4.33660209e-01
4.79808390e-01 3.56753021e-01 1.29949498e+00 5.24347246e-01
-9.69634533e-01 -3.63617420e-01 -5.39562762e-01 -3.03754266e-02
-1.02050018e+00 -2.15614483e-01 2.24186063e-01 -7.47228801e-01
7.48307526e-01 5.75934649e-01 1.21765351e+00 2.62497187e-01
-3.36422145e-01 6.12072237e-02 9.40206647e-01 -3.92816454e-01
5.03451861e-02 -4.22639884e-02 2.72564795e-02 5.77260077e-01
6.65389180e-01 -1.47066787e-01 -1.97060063e-01 -6.33381009e-01
5.84551811e-01 -2.15508893e-01 -6.02868736e-01 1.99761447e-02
-1.17558718e+00 5.22568583e-01 -4.20635432e-01 1.89273551e-01
-3.80582958e-01 -2.79319644e-01 5.45355320e-01 1.32475272e-01
2.53542066e-01 7.24765480e-01 -4.20443445e-01 -9.27625537e-01
-9.58203375e-01 -4.37644608e-02 8.74490023e-01 8.12072381e-02
-1.20851867e-01 -1.61388457e-01 -2.91388243e-01 9.29743707e-01
1.04837477e-01 6.12414837e-01 3.34091872e-01 -1.38383067e+00
1.87242165e-01 3.87737960e-01 1.44810989e-01 -8.79356921e-01
-7.29359865e-01 -5.70385158e-01 -5.22295654e-01 -3.13201845e-01
4.37689781e-01 -6.64078951e-01 -3.38052601e-01 1.21348894e+00
3.51978749e-01 2.23168746e-01 -1.74546555e-01 9.14444625e-01
7.10327506e-01 -1.59916617e-02 7.90438652e-02 -2.86630988e-01
1.18040836e+00 -3.65271837e-01 -2.38250092e-01 2.43948072e-01
1.04842269e+00 -7.39848971e-01 9.18578327e-01 5.04904568e-01
-1.30836976e+00 -4.24713314e-01 -6.43613994e-01 6.01559579e-01
2.17629388e-01 6.84057251e-02 1.35511935e-01 1.22630310e+00
-8.74882638e-01 8.36329818e-01 -1.05754161e+00 -2.87316650e-01
3.65908742e-01 2.68825650e-01 -5.37676066e-02 9.38569233e-02
-9.62882161e-01 7.74090111e-01 -1.44251421e-01 8.96050185e-02
-3.50281477e-01 -9.99160469e-01 -5.32921970e-01 -4.10084091e-02
-2.04200554e-03 -7.17350423e-01 1.14944792e+00 -3.83580297e-01
-1.05939698e+00 9.25675809e-01 1.13544166e-01 -1.85742021e-01
5.97779155e-01 -2.16518998e-01 -3.21841747e-01 6.59575582e-01
1.95000798e-01 1.34540558e-01 4.57485169e-01 -1.37627035e-01
-7.47665465e-01 -9.69980210e-02 -5.25160611e-01 2.16418579e-01
-1.28911078e-01 3.40206832e-01 -2.48147883e-02 -5.51023781e-01
1.23979218e-01 -6.86426163e-01 1.25838537e-02 -2.78110147e-01
7.06389025e-02 3.48186493e-01 5.05729795e-01 -1.20290029e+00
1.40092242e+00 -2.09671831e+00 -3.75576466e-01 1.57323748e-01
5.07305026e-01 6.11751199e-01 2.75666326e-01 3.41006219e-01
-1.22499600e-01 4.59788203e-01 -2.44063005e-01 8.01250525e-03
-3.03895324e-01 -2.10251566e-02 7.94945359e-02 4.09199744e-01
1.85415536e-01 7.77831554e-01 -9.29454386e-01 -3.95007312e-01
4.54938889e-01 5.08150578e-01 -5.92837453e-01 3.78414929e-01
4.07226086e-01 7.01387584e-01 -2.96953339e-02 4.68295842e-01
1.88794300e-01 -5.43455005e-01 4.38846320e-01 2.90435225e-01
-9.96155739e-02 5.46174943e-01 -1.25360096e+00 9.82810616e-01
-3.17836814e-02 6.56647325e-01 3.02923433e-02 -7.39290893e-01
7.68084168e-01 7.67771363e-01 6.01668179e-01 -6.21814542e-02
-5.53597622e-02 6.28451765e-01 6.05730355e-01 -7.97751486e-01
-4.66161489e-01 -8.82850364e-02 5.03506660e-01 8.88663232e-01
-5.05738914e-01 -3.89844418e-01 1.66359097e-01 -1.49306625e-01
1.44707060e+00 -1.05069488e-01 1.71530455e-01 -1.72393277e-01
2.93693304e-01 -1.87751099e-01 4.27652597e-01 1.10950279e+00
-4.62148845e-01 1.17722821e+00 2.88055003e-01 -5.68926871e-01
-6.17369533e-01 -7.47317970e-01 -3.60009253e-01 6.92516565e-01
-8.36174905e-01 -2.55365580e-01 -5.82265437e-01 -4.04349357e-01
-1.84203506e-01 1.49216905e-01 -2.84824669e-01 1.83675867e-02
-6.70527041e-01 -8.83786678e-01 1.00985920e+00 6.47075653e-01
7.36583918e-02 -8.73389602e-01 -1.51790929e+00 6.57486558e-01
-4.54100698e-01 -6.90128207e-01 -3.35325032e-01 -1.64571345e-01
-9.35745716e-01 -1.34905851e+00 -1.14585900e+00 -5.98812521e-01
1.46812469e-01 -3.89602840e-01 9.62555230e-01 3.73731732e-01
-1.00211406e+00 6.65116012e-01 -4.19980317e-01 -5.96458793e-01
-7.44281709e-01 1.23039328e-01 -2.23445728e-01 -3.50810826e-01
-2.71123070e-02 -5.62286496e-01 -7.90445268e-01 2.22121835e-01
-3.48415464e-01 -1.64499700e-01 2.36047566e-01 6.46765411e-01
3.85134637e-01 -5.35327852e-01 7.16609180e-01 -7.45832026e-01
8.21950078e-01 -2.62302190e-01 -4.10679549e-01 1.91563860e-01
-8.85280371e-01 -4.75795388e-01 2.30951920e-01 -3.40630591e-01
-2.66284078e-01 -3.59303206e-01 -1.50384754e-01 -2.86099594e-02
-1.99285910e-01 4.23241079e-01 7.28790700e-01 -3.08580212e-02
7.63165236e-01 -8.28914642e-02 2.66514689e-01 -5.42248130e-01
-4.79366481e-01 7.48171449e-01 5.89708447e-01 -4.65406299e-01
2.99887091e-01 4.56105731e-02 2.14963078e-01 -7.54678607e-01
-5.67546189e-01 -5.43685973e-01 -1.70086756e-01 -2.51116544e-01
7.56835759e-01 -6.11769378e-01 -5.63857257e-01 4.38959241e-01
-7.39421725e-01 -4.50279087e-01 -5.47695637e-01 9.35577691e-01
-9.83426496e-02 5.72092474e-01 -4.77141768e-01 -8.99325192e-01
-7.52469778e-01 -9.40356970e-01 6.84371114e-01 1.46234110e-02
-7.74429739e-01 -1.06306982e+00 6.28242552e-01 5.37506282e-01
6.21722221e-01 5.40964663e-01 8.47235501e-01 -7.65760839e-01
-7.66005367e-03 -4.15515691e-01 -5.49245588e-02 5.35388589e-01
3.96998227e-01 2.37894580e-01 -8.99455309e-01 -7.66974613e-02
-1.88236356e-01 -1.63054273e-01 2.88536489e-01 6.53772593e-01
1.03564489e+00 1.74172521e-01 1.42767161e-01 4.56311315e-01
7.45296538e-01 2.53169507e-01 2.19645903e-01 3.50018255e-02
4.07721758e-01 6.39312625e-01 3.18119168e-01 5.01415610e-01
5.05722106e-01 1.76728427e-01 -9.47827771e-02 -3.23691934e-01
-6.71112388e-02 1.47332624e-02 -8.33330601e-02 8.46925557e-01
-4.74110723e-01 2.88197309e-01 -1.58351171e+00 6.76120400e-01
-1.33023417e+00 -6.78424835e-01 -4.00362134e-01 2.36671710e+00
7.63412356e-01 3.07378382e-03 5.27377665e-01 3.80641192e-01
8.39571416e-01 -3.64059150e-01 -1.71976805e-01 -5.98495543e-01
1.30556554e-01 6.41758084e-01 9.28118601e-02 3.93865049e-01
-7.38089144e-01 -8.71883556e-02 7.46230173e+00 -6.73137233e-02
-1.33712089e+00 1.16996072e-01 8.50854635e-01 -8.81109852e-03
-1.76434666e-02 -1.27680227e-01 -2.92810112e-01 5.28101146e-01
1.19497216e+00 8.20831656e-02 1.84998259e-01 4.51752275e-01
4.94177282e-01 -5.44902459e-02 -8.57049048e-01 7.53926814e-01
-4.52894671e-03 -1.44804156e+00 -7.92950392e-01 -3.38666677e-01
4.00364816e-01 2.98333496e-01 -1.44819915e-01 -1.42549440e-01
-4.14618820e-01 -8.70073318e-01 1.75769776e-01 4.83857632e-01
1.22587359e+00 -8.67002681e-02 8.29593420e-01 4.02845182e-02
-8.70051563e-01 -9.27459449e-02 2.83845812e-01 -4.01363939e-01
9.43339169e-02 5.28283834e-01 -1.31653106e+00 2.88321763e-01
6.54242277e-01 2.84459859e-01 -5.20001471e-01 1.38329756e+00
-1.35465153e-02 1.37069428e+00 -6.79349124e-01 1.32557616e-01
-3.94116193e-01 1.06751882e-01 5.59962392e-01 1.17880452e+00
3.45414966e-01 2.27192014e-01 -5.16656823e-02 6.81594849e-01
3.81461501e-01 4.93367612e-01 -3.00384820e-01 -2.05956519e-01
4.63547617e-01 9.73375022e-01 -7.26247132e-01 -3.94366145e-01
-4.78497624e-01 7.33142346e-02 -9.39656049e-02 7.45125562e-02
-5.57453394e-01 -4.64672744e-01 2.87339184e-02 6.83772743e-01
4.10385467e-02 4.83903378e-01 -6.79634452e-01 -6.24837697e-01
7.55861849e-02 -1.22595119e+00 6.79805875e-01 -6.56792521e-01
-6.30393922e-01 2.12547883e-01 2.17030868e-02 -9.59475756e-01
-5.82157850e-01 -1.93905890e-01 -8.65731478e-01 1.17985857e+00
-1.01382816e+00 -3.59719753e-01 -3.81988913e-01 -8.55298191e-02
3.52999382e-02 8.36327299e-02 1.06157637e+00 3.46162975e-01
-4.96119171e-01 3.71271640e-01 -3.15473974e-01 5.56292161e-02
7.04966784e-01 -1.23853040e+00 1.15052171e-01 5.43802679e-01
-3.39131325e-01 7.22041190e-01 3.07284623e-01 -7.98749447e-01
-7.40573406e-01 -7.17708051e-01 1.03838742e+00 -6.39982998e-01
3.05899441e-01 4.76163745e-01 -7.60283768e-01 2.26279855e-01
-4.26960349e-01 -1.08808398e-01 1.04997134e+00 -6.68103471e-02
3.04096192e-01 -4.27281857e-02 -1.01025772e+00 9.82443839e-02
6.09611869e-01 -5.14017820e-01 -7.06959665e-01 9.15213227e-02
1.27962589e-01 -5.20069420e-01 -1.28352761e+00 6.92000508e-01
8.54358435e-01 -9.04775441e-01 5.57540894e-01 -2.68148631e-01
1.31598353e-01 -1.06365187e-02 4.46356058e-01 -7.31171608e-01
6.81074150e-03 -1.20873392e+00 1.38090327e-01 7.73575962e-01
4.80092376e-01 -9.47510421e-01 5.44731796e-01 8.15407038e-01
-1.49980381e-01 -8.29429746e-01 -5.89284480e-01 -1.54124916e-01
-1.02945954e-01 -5.56589365e-01 3.83271933e-01 8.00217211e-01
1.33028090e-01 -1.31470829e-01 3.56290024e-03 1.38368696e-01
3.81320536e-01 1.15129158e-01 4.21625197e-01 -1.35172129e+00
-8.50453675e-01 -2.63483852e-01 -3.18228424e-01 -1.85867429e-01
-7.66876340e-01 -6.53618693e-01 -3.00780535e-01 -1.58782184e+00
1.98108822e-01 -5.62941670e-01 -1.55569851e-01 5.64464271e-01
-6.32118583e-01 5.41042089e-01 1.81175005e-02 1.48803309e-01
-2.06441239e-01 -5.52673936e-01 1.07824266e+00 5.15889227e-01
-4.67004925e-01 2.71422058e-01 -5.97221732e-01 6.23473763e-01
8.67663085e-01 -5.74041724e-01 -5.15342951e-01 -2.04261377e-01
3.60893309e-01 3.23434114e-01 5.84434092e-01 -1.06839848e+00
-2.41023302e-01 1.05186701e-01 1.99071184e-01 -2.96677589e-01
-2.45034266e-02 -3.39137703e-01 4.37023371e-01 1.01531672e+00
-9.81708020e-02 2.31744438e-01 2.78978705e-01 -1.79206267e-01
-1.13487458e-02 -3.53624195e-01 5.53767383e-01 -1.74503356e-01
1.48689568e-01 -6.02822239e-03 -6.27205014e-01 9.13546801e-01
6.07464492e-01 -3.96614283e-01 -2.48267114e-01 -5.98076820e-01
-1.04883134e+00 1.82458796e-02 2.57068127e-01 -1.72028065e-01
4.03014183e-01 -3.84813517e-01 -1.00878823e+00 3.02092135e-01
-2.16898054e-01 6.55217990e-02 2.78283715e-01 1.54464650e+00
-1.30597007e+00 2.59752780e-01 -1.69558767e-02 -6.93963706e-01
-1.42105699e+00 -1.75772116e-01 6.52399778e-01 -1.17921382e-01
-8.10862482e-01 6.85341239e-01 -3.95048499e-01 -1.31503388e-01
3.35838258e-01 -5.45095980e-01 1.91552863e-02 -8.58589038e-02
5.52005291e-01 8.36067438e-01 1.93190232e-01 4.12912704e-02
-3.66210759e-01 3.90531629e-01 3.94371092e-01 -1.52673349e-01
1.05772936e+00 -1.18547559e-01 -7.74557739e-02 3.55464011e-01
4.93872404e-01 1.31019264e-01 -5.99416137e-01 5.14500320e-01
-2.26378918e-01 -2.00586066e-01 -1.07196473e-01 -1.00618076e+00
-5.77875853e-01 1.09149480e+00 8.80567074e-01 5.09359121e-01
1.21538424e+00 -3.03608984e-01 4.54321802e-01 8.53782594e-02
-2.01892689e-01 -9.47795153e-01 -2.65066892e-01 -7.47072473e-02
3.85246456e-01 -7.10370064e-01 -1.89768448e-01 -2.12378740e-01
-5.46627104e-01 1.02835989e+00 2.36632004e-01 4.13053364e-01
4.57231015e-01 1.93623453e-01 5.44515967e-01 -2.28114963e-01
-5.93459487e-01 2.02787101e-01 2.22708121e-01 8.11947227e-01
6.18085921e-01 3.46596777e-01 -6.47608459e-01 5.67348540e-01
-1.68388058e-02 2.43192226e-01 5.24107039e-01 9.39422965e-01
-3.62025619e-01 -1.10253310e+00 -3.90326321e-01 8.12483132e-01
-9.55194056e-01 -2.54926324e-01 -4.05077487e-01 4.08535838e-01
3.04089516e-01 9.40465987e-01 -2.54870296e-01 1.71084568e-01
3.22479665e-01 4.70577657e-01 4.05855685e-01 -9.00803864e-01
-9.17651474e-01 2.17262730e-01 4.17950034e-01 -1.17247321e-01
-2.76507020e-01 -9.90168452e-01 -9.12596881e-01 1.63673058e-01
-2.36179829e-01 3.37032914e-01 4.85602409e-01 6.64632857e-01
7.02284396e-01 5.13769686e-01 4.05681878e-01 -1.03671789e-01
-8.01706314e-01 -9.34417129e-01 -3.88813078e-01 2.48810947e-01
4.19935316e-01 -1.72463745e-01 -4.64458257e-01 1.24536239e-01]
|
[14.301526069641113, 3.2638609409332275]
|
5844b90a-7c35-4d52-bd64-30c4491aa22e
|
protoinfomax-prototypical-networks-with
|
2108.12229
| null |
https://arxiv.org/abs/2108.12229v5
|
https://arxiv.org/pdf/2108.12229v5.pdf
|
ProtoInfoMax: Prototypical Networks with Mutual Information Maximization for Out-of-Domain Detection
|
The ability to detect Out-of-Domain (OOD) inputs has been a critical requirement in many real-world NLP applications. For example, intent classification in dialogue systems. The reason is that the inclusion of unsupported OOD inputs may lead to catastrophic failure of systems. However, it remains an empirical question whether current methods can tackle such problems reliably in a realistic scenario where zero OOD training data is available. In this study, we propose ProtoInfoMax, a new architecture that extends Prototypical Networks to simultaneously process in-domain and OOD sentences via Mutual Information Maximization (InfoMax) objective. Experimental results show that our proposed method can substantially improve performance up to 20% for OOD detection in low resource settings of text classification. We also show that ProtoInfoMax is less prone to typical overconfidence errors of Neural Networks, leading to more reliable prediction results.
|
['Mykola Pechenizkiy', 'Vlado Menkovski', 'Meng Fang', "Iftitahu Ni'mah"]
|
2021-08-27
| null |
https://aclanthology.org/2021.findings-emnlp.138
|
https://aclanthology.org/2021.findings-emnlp.138.pdf
|
findings-emnlp-2021-11
|
['zero-shot-out-of-domain-detection', 'few-shot-text-classification']
|
['natural-language-processing', 'natural-language-processing']
|
[ 2.60249972e-01 6.79970741e-01 -1.99766174e-01 -7.97576487e-01
-2.67200500e-01 -4.50047404e-01 8.44948947e-01 1.68407559e-01
-6.00383162e-01 8.35951149e-01 2.74284631e-01 -6.79161966e-01
2.06480492e-02 -5.00986993e-01 -2.02620968e-01 4.48157340e-02
9.59591940e-02 7.19219923e-01 -1.99051318e-03 -3.11245978e-01
4.08315092e-01 1.52273372e-01 -1.34584272e+00 5.18601596e-01
1.26428008e+00 9.11259294e-01 1.67789668e-01 8.91069591e-01
-2.19742924e-01 9.47384179e-01 -9.52082336e-01 -5.63498080e-01
1.32868305e-01 -1.49927691e-01 -1.14564610e+00 -8.19452778e-02
3.81035179e-01 -6.71204746e-01 -2.43699804e-01 1.02797556e+00
4.30193096e-01 2.32693374e-01 9.92820799e-01 -1.10386169e+00
-4.99741346e-01 7.92050123e-01 -4.76470450e-03 2.41981059e-01
4.87035692e-01 2.33028427e-01 1.17102790e+00 -8.27901602e-01
4.21250612e-01 1.34839404e+00 4.61629808e-01 5.85715234e-01
-1.29141974e+00 -4.72290546e-01 -1.03935215e-03 -1.91799060e-01
-7.55371571e-01 -5.74656427e-01 3.84213448e-01 -2.55721837e-01
1.56576586e+00 3.01253885e-01 -1.04573388e-02 1.41590500e+00
3.56746197e-01 1.15420544e+00 1.09480834e+00 -5.59999526e-01
2.83429116e-01 7.12590337e-01 5.81538439e-01 3.68578762e-01
3.01541269e-01 -1.07392222e-01 -6.66665614e-01 -3.76858145e-01
2.85293698e-01 -3.68590713e-01 -1.76644608e-01 1.18496187e-01
-9.25881088e-01 1.12272966e+00 1.35405153e-01 3.62782687e-01
-4.26368505e-01 -5.33235312e-01 2.76853204e-01 8.21837902e-01
6.74125373e-01 1.09564972e+00 -7.24720478e-01 -3.45985264e-01
-7.45292723e-01 2.31434360e-01 1.57537198e+00 9.52466607e-01
3.60546738e-01 2.58825738e-02 -1.63834065e-01 1.11432338e+00
2.92456478e-01 2.25138083e-01 9.34379458e-01 -9.68810558e-01
7.17548847e-01 7.16008961e-01 2.00062215e-01 -9.70138609e-01
-6.47890568e-01 -3.66169631e-01 -8.66065621e-01 -1.63739119e-02
4.70448643e-01 -5.32499015e-01 -5.90299308e-01 1.66087043e+00
1.68866485e-01 -5.38189232e-01 4.87409562e-01 7.18503654e-01
7.85324275e-01 5.58398783e-01 -5.94824832e-03 -2.68215775e-01
1.40023577e+00 -6.65941119e-01 -8.88792515e-01 -8.22015226e-01
1.08911204e+00 -7.45578825e-01 1.02769542e+00 4.29129153e-01
-7.87068665e-01 -2.97863752e-01 -1.06621432e+00 1.22270212e-01
-2.82126278e-01 -1.08464807e-01 8.21738064e-01 8.89146268e-01
-8.23517859e-01 5.53605139e-01 -3.17781299e-01 -3.99950325e-01
3.86575274e-02 4.34043497e-01 -4.27192062e-01 2.36885771e-02
-1.73086429e+00 1.29687560e+00 7.21401513e-01 -1.37483001e-01
-3.62926602e-01 -3.79869491e-01 -8.59207392e-01 7.24377260e-02
4.94065434e-01 -3.38412046e-01 1.62430370e+00 -9.57276464e-01
-1.51130605e+00 6.62378609e-01 -7.41481483e-02 -7.25894749e-01
6.26897633e-01 -4.85276699e-01 -2.93303818e-01 -2.17620477e-01
8.15964937e-02 6.65863991e-01 6.11810207e-01 -7.52330899e-01
-6.30655289e-01 -1.54950693e-01 3.29153575e-02 5.50988317e-01
-8.37022960e-01 -1.04523242e-01 2.42909119e-02 -4.89089787e-01
-7.81794935e-02 -7.32812047e-01 -3.97232212e-02 -4.57971424e-01
-7.01269925e-01 -8.11560333e-01 3.96063745e-01 -4.86732900e-01
1.18408549e+00 -1.84716296e+00 -3.00481856e-01 7.39244325e-03
2.33932599e-01 4.39298600e-01 2.54111784e-03 4.87501442e-01
7.16243684e-02 1.89394951e-01 -1.82893738e-01 -1.92564428e-01
1.88005269e-01 1.80155843e-01 -2.60659397e-01 5.07987328e-02
3.80443245e-01 4.38641965e-01 -7.17886627e-01 -4.81602550e-01
3.25352773e-02 -5.67156263e-02 -6.11858130e-01 5.13550580e-01
-3.20664048e-01 -1.22252680e-01 -3.05054516e-01 2.89396793e-01
3.97713184e-01 -4.11227345e-01 3.71218592e-01 3.26727331e-01
4.60675061e-02 8.78481627e-01 -1.09007037e+00 1.11509526e+00
-4.70306486e-01 8.67991745e-01 -1.54628962e-01 -8.64548326e-01
9.65149283e-01 4.70255435e-01 -1.57385235e-04 -6.93262696e-01
2.85124600e-01 1.78213462e-01 4.14625943e-01 -5.01734972e-01
8.87301385e-01 -6.13937378e-02 -1.82135582e-01 7.73585200e-01
3.28505218e-01 -5.47687560e-02 1.46196634e-01 3.84304225e-01
1.11836398e+00 -7.45135128e-01 5.39102495e-01 -3.14452589e-01
2.31608748e-01 1.27145395e-01 4.25316453e-01 1.39463341e+00
-3.23154479e-01 1.83594048e-01 6.91854894e-01 -1.76371679e-01
-8.33055377e-01 -6.37583792e-01 -3.92479956e-01 1.22119558e+00
-2.20872492e-01 -1.42117783e-01 -6.56073391e-01 -1.06154931e+00
8.45929757e-02 1.41230714e+00 -3.90359193e-01 -2.43301541e-01
-1.78162366e-01 -8.33227098e-01 8.08181942e-01 2.73063302e-01
4.87760127e-01 -1.09865546e+00 -3.72943640e-01 3.70415360e-01
-2.74683863e-01 -1.13161755e+00 -5.08593842e-02 5.67064822e-01
-8.93344700e-01 -8.21504354e-01 -6.64673328e-01 -6.14459276e-01
4.15737778e-01 1.18779153e-01 1.20890725e+00 4.82904166e-03
2.07325548e-01 1.03322595e-01 -3.14578205e-01 -5.09278119e-01
-1.07485044e+00 4.67517108e-01 5.16786158e-01 -2.99228132e-01
9.75814998e-01 -1.64800555e-01 -1.62876949e-01 4.26653713e-01
-7.98722327e-01 9.27358419e-02 6.26862526e-01 1.14472473e+00
-3.82900178e-01 6.01731353e-02 9.18600440e-01 -1.21716154e+00
1.40800631e+00 -6.59671128e-01 -3.18794668e-01 7.03408644e-02
-9.87015247e-01 2.97360033e-01 5.55507243e-01 -4.52652514e-01
-1.31610703e+00 -2.49886408e-01 -2.98955679e-01 1.14293925e-01
-5.46895325e-01 6.66999042e-01 1.55671626e-01 3.36638570e-01
9.79430497e-01 2.98527461e-02 1.69612631e-01 -4.20058668e-01
3.16713657e-03 1.43359697e+00 1.12457611e-01 -2.62582719e-01
3.26904476e-01 -1.53300837e-01 -5.98429918e-01 -1.19857085e+00
-9.24597442e-01 -5.44275939e-01 -3.67620826e-01 -9.70404074e-02
5.39385319e-01 -7.75717199e-01 -7.57362008e-01 2.58910388e-01
-1.21986485e+00 -5.05457699e-01 2.78610915e-01 6.30205452e-01
-3.68603379e-01 4.28750634e-01 -7.47120798e-01 -1.13018930e+00
-3.76250118e-01 -9.77325082e-01 4.50896472e-01 3.86249185e-01
-8.92564833e-01 -1.06380641e+00 -1.28088370e-01 4.59273964e-01
3.74258637e-01 -6.06122732e-01 9.18349445e-01 -1.74912870e+00
9.65536311e-02 -4.08872068e-01 -1.19282462e-01 5.43465436e-01
1.14287674e-01 -4.39453959e-01 -1.26704156e+00 -1.07720338e-01
1.83663338e-01 -8.22553575e-01 7.24743426e-01 7.95395076e-02
6.84790909e-01 -5.77820063e-01 -1.30394548e-01 -1.99108124e-01
7.56321073e-01 1.64928496e-01 2.73379415e-01 1.53660804e-01
7.45869651e-02 1.09638071e+00 8.21622550e-01 5.35408795e-01
3.81858289e-01 4.38086748e-01 1.11023374e-01 1.44512802e-01
3.10702920e-01 -2.02920288e-01 3.51415187e-01 6.71326280e-01
4.22202557e-01 -7.25545824e-01 -1.06379390e+00 4.06081408e-01
-1.61401451e+00 -7.43327260e-01 -7.24088848e-02 2.13399196e+00
1.12573767e+00 5.98364830e-01 -1.13369517e-01 -1.50832105e-02
6.95557773e-01 -7.65897557e-02 -5.93366265e-01 -7.21983850e-01
5.18304221e-02 -2.17362210e-01 3.13043892e-01 5.65962553e-01
-1.15786612e+00 9.00352776e-01 6.74342155e+00 5.58427513e-01
-7.52246976e-01 4.97443601e-02 7.89310038e-01 1.62563309e-01
-4.22460139e-02 -3.08840156e-01 -1.07613039e+00 4.32808727e-01
1.10770762e+00 -4.62119393e-02 5.21386974e-02 1.14424753e+00
-8.45566839e-02 -2.97215253e-01 -1.23804760e+00 4.59447652e-01
1.31679088e-01 -9.09301758e-01 -1.40882611e-01 1.36478931e-01
5.73137403e-01 1.44733176e-01 -1.61779821e-01 7.21374929e-01
7.50471413e-01 -9.40395951e-01 1.95803359e-01 7.35126287e-02
4.26406771e-01 -5.31045973e-01 1.03323936e+00 1.05762124e+00
-1.35759786e-01 -2.62291003e-02 -5.67836165e-01 -2.78818458e-01
-6.15558922e-02 6.65681183e-01 -1.66298747e+00 1.04724832e-01
4.08832610e-01 1.23913296e-01 -4.25191313e-01 7.31982648e-01
-1.42560437e-01 7.63848484e-01 -5.53728998e-01 -6.64318502e-01
3.98492873e-01 4.20985937e-01 5.55371642e-01 1.25864017e+00
-1.12006329e-01 -2.27966867e-02 -2.82701887e-02 6.60282373e-01
-2.92475671e-01 1.09427817e-01 -9.66210783e-01 -2.72084981e-01
5.28604805e-01 9.65711474e-01 -4.86350834e-01 -4.84858930e-01
-3.42332363e-01 9.82415795e-01 5.07007837e-01 1.34386286e-01
-2.63616204e-01 -6.34678185e-01 7.00610936e-01 -3.53147268e-01
-1.45311773e-01 9.73980948e-02 -3.56549352e-01 -1.17621672e+00
-1.42582253e-01 -1.16713750e+00 5.83622396e-01 -3.88503730e-01
-1.60469353e+00 6.27931714e-01 -1.62632510e-01 -8.31038833e-01
-8.33671808e-01 -9.64573503e-01 -5.60142279e-01 7.55951583e-01
-1.29449379e+00 -3.42611969e-01 3.78459692e-02 1.92632303e-01
1.00678909e+00 -3.08202893e-01 1.02069998e+00 -3.86707112e-02
-5.13348103e-01 7.43328094e-01 1.97229609e-01 3.35180640e-01
1.06200039e+00 -1.44940996e+00 4.49845880e-01 5.25924623e-01
-2.41901856e-02 7.28870630e-01 1.01481318e+00 -8.62123430e-01
-1.03395724e+00 -6.41963243e-01 1.38982213e+00 -5.12463510e-01
4.25565273e-01 -4.43028152e-01 -1.03059244e+00 6.87775314e-01
3.48873734e-01 -5.36621094e-01 8.52046847e-01 5.18818319e-01
-2.34115437e-01 4.08745617e-01 -1.10689974e+00 7.42096961e-01
6.25643492e-01 -4.52891678e-01 -1.21974409e+00 5.80434203e-01
7.21374333e-01 -3.27547193e-01 -9.22945321e-01 2.71704614e-01
4.49086905e-01 -1.00426233e+00 7.04509020e-01 -8.85093033e-01
5.68329811e-01 5.22273123e-01 -3.45171355e-02 -1.48817706e+00
-2.89390096e-03 -4.84450042e-01 -2.00998083e-01 1.21559012e+00
8.24736357e-01 -8.41841042e-01 6.54149830e-01 1.23893607e+00
-4.53306548e-03 -4.52166051e-01 -8.44562531e-01 -7.23378062e-01
1.34666950e-01 -6.01596832e-01 6.38409480e-02 1.29443955e+00
6.42832696e-01 7.22491264e-01 -4.84796047e-01 5.71509302e-02
4.22038466e-01 -1.53274000e-01 6.45674884e-01 -1.45483124e+00
-3.39408994e-01 -3.32243025e-01 -6.30872548e-02 -1.35012805e+00
4.14859295e-01 -7.30228126e-01 2.94850230e-01 -1.22576940e+00
-7.31589124e-02 -4.78289604e-01 -3.93566154e-02 3.19252938e-01
-3.51208121e-01 9.59056523e-03 4.18176688e-02 7.68711865e-02
-4.77682799e-01 5.45248985e-01 6.48618817e-01 -2.31710710e-02
-3.66555899e-01 2.52139986e-01 -7.67104685e-01 8.94214034e-01
9.86960113e-01 -7.01476455e-01 -3.30270261e-01 -2.53198057e-01
8.42031464e-02 1.29286259e-01 -1.43250689e-01 -8.28052819e-01
2.69292533e-01 -1.95861310e-01 4.00271684e-01 -5.03616571e-01
3.25044423e-01 -6.21637583e-01 -6.26032531e-01 5.57031631e-01
-9.67412949e-01 -2.57025659e-01 -5.62257785e-03 5.16303658e-01
-5.85895628e-02 -8.18902552e-01 4.73456502e-01 -2.70121694e-01
-4.75021243e-01 -3.91370833e-01 -1.04535353e+00 3.84939551e-01
6.54193103e-01 -5.41041270e-02 -4.31657106e-01 -6.90206766e-01
-3.50965500e-01 4.30972517e-01 -1.56309884e-02 5.91346264e-01
4.21423733e-01 -7.67093062e-01 -6.93681419e-01 2.44532123e-01
2.31078386e-01 -3.40644479e-01 3.56750330e-03 6.80863738e-01
-1.61463127e-01 8.78408968e-01 6.20921813e-02 -5.14856398e-01
-1.23075151e+00 1.56011451e-02 2.57957637e-01 -4.67974335e-01
-3.73921305e-01 6.95444703e-01 2.55333632e-01 -9.00610924e-01
4.75946426e-01 -1.16226003e-02 -3.97034943e-01 1.40197039e-01
5.89085281e-01 7.50333741e-02 1.54630333e-01 -1.81013227e-01
-1.38129443e-01 -6.27797186e-01 -7.17480242e-01 -2.64207870e-01
1.10362601e+00 -1.13464653e-01 3.84077579e-01 5.32068431e-01
8.56857896e-01 -3.87016922e-01 -8.54139149e-01 -4.33074355e-01
2.94240117e-01 -3.05285245e-01 2.06784829e-01 -1.02820802e+00
-2.15917617e-01 9.24434841e-01 2.79336423e-01 7.16029704e-01
4.89306539e-01 -2.17232049e-01 5.30790389e-01 1.27628994e+00
1.55942574e-01 -1.44240546e+00 9.79335234e-02 9.48135316e-01
6.17503881e-01 -1.65424204e+00 -2.43025571e-01 -1.86414480e-01
-9.93694186e-01 1.12817705e+00 9.81030643e-01 3.23491395e-01
4.22454596e-01 9.18675661e-02 1.24164402e-01 -6.60631955e-02
-1.21209908e+00 1.37784570e-01 2.54614055e-01 3.76742780e-01
7.33563304e-01 -3.58742997e-02 -2.24392429e-01 6.93804562e-01
-1.95348412e-01 -3.13084811e-01 6.19360745e-01 9.51724410e-01
-6.30664825e-01 -7.65188456e-01 -2.03834996e-01 1.03599298e+00
-4.55743223e-01 -3.56815904e-01 -7.62809098e-01 6.96924806e-01
-6.68775022e-01 1.17815447e+00 2.85973102e-01 -5.13125360e-01
1.33788690e-01 6.19539499e-01 -2.62208909e-01 -7.65828907e-01
-7.22732484e-01 -3.16070735e-01 6.90978289e-01 -1.99688122e-01
-5.32340072e-02 -4.00809824e-01 -8.96691501e-01 -2.54462004e-01
-7.85104632e-01 2.52327263e-01 5.23921371e-01 1.18460476e+00
4.35777605e-01 2.47783676e-01 5.00104666e-01 -2.50854880e-01
-1.12401247e+00 -1.61433625e+00 -4.34541047e-01 3.75310183e-01
1.54390365e-01 -6.19448364e-01 -6.17659271e-01 -3.97817284e-01]
|
[12.665203094482422, 7.91787576675415]
|
dd8fa725-8ca7-4d6d-91a5-d5253950d0d0
|
exploring-the-compositional-generalization-in
|
2306.0448
| null |
https://arxiv.org/abs/2306.04480v1
|
https://arxiv.org/pdf/2306.04480v1.pdf
|
Exploring the Compositional Generalization in Context Dependent Text-to-SQL Parsing
|
In the context-dependent Text-to-SQL task, the generated SQL statements are refined iteratively based on the user input utterance from each interaction. The input text from each interaction can be viewed as component modifications to the previous SQL statements, which could be further extracted as the modification patterns. Since these modification patterns could also be combined with other SQL statements, the models are supposed to have the compositional generalization to these novel combinations. This work is the first exploration of compositional generalization in context-dependent Text-to-SQL scenarios. To facilitate related studies, we constructed two challenging benchmarks named \textsc{CoSQL-CG} and \textsc{SParC-CG} by recombining the modification patterns and existing SQL statements. The following experiments show that all current models struggle on our proposed benchmarks. Furthermore, we found that better aligning the previous SQL statements with the input utterance could give models better compositional generalization ability. Based on these observations, we propose a method named \texttt{p-align} to improve the compositional generalization of Text-to-SQL models. Further experiments validate the effectiveness of our method. Source code and data are available.
|
['Lijie Wen', 'Yawen Yang', 'Fukun Ma', 'Shuang Li', 'Xuming Hu', 'Wei Liu', 'Aiwei Liu']
|
2023-05-29
| null | null | null | null |
['text-to-sql']
|
['computer-code']
|
[ 4.21045184e-01 2.01212108e-01 -6.32558614e-02 -1.07323670e+00
-6.86149180e-01 -7.39742339e-01 5.59328079e-01 1.15828171e-01
-1.21232565e-03 4.13982987e-01 4.69301552e-01 -4.53194112e-01
-7.67830312e-02 -7.79377401e-01 -8.35020304e-01 -1.78376976e-02
7.79839382e-02 6.34171188e-01 5.27088642e-01 -6.72257245e-01
2.86840409e-01 2.05941990e-01 -1.81123006e+00 1.22471595e+00
1.02487195e+00 7.53683269e-01 3.13039184e-01 5.41567922e-01
-7.34935522e-01 7.31849253e-01 -6.34880722e-01 -7.16986716e-01
1.18076652e-01 -4.12458122e-01 -1.03236997e+00 -4.18600142e-02
3.55814517e-01 -9.68170911e-02 3.55707109e-01 7.11338282e-01
2.17067331e-01 2.74781048e-01 1.72435194e-01 -1.15103400e+00
-4.79844868e-01 1.12021029e+00 -6.48448169e-02 -3.44767183e-01
9.24533427e-01 -7.53936470e-02 1.22919512e+00 -1.11544621e+00
7.02707112e-01 1.40080106e+00 4.35194999e-01 5.58382392e-01
-9.84605432e-01 -3.47655594e-01 6.01612866e-01 1.67742044e-01
-1.21145034e+00 -3.49565834e-01 6.57391131e-01 -5.92728481e-02
1.47413719e+00 1.10581183e+00 2.54452884e-01 8.69775057e-01
-1.53799117e-01 8.60440016e-01 6.91373348e-01 -4.69567329e-01
1.51327878e-01 3.83350492e-01 3.24253380e-01 5.92877865e-01
-1.62181646e-01 -3.33150327e-01 -6.39132679e-01 -3.52970004e-01
1.17609039e-01 1.02465525e-01 -7.55342916e-02 -2.29414254e-01
-1.10848916e+00 6.16252124e-01 -2.72293910e-02 2.84889191e-01
-4.26498763e-02 -1.98119923e-01 5.00964224e-01 3.44899774e-01
-1.30057335e-03 6.64914072e-01 -8.95285547e-01 -3.29028636e-01
-4.28030133e-01 6.78424418e-01 1.09567809e+00 1.78440046e+00
9.03499067e-01 -3.26664925e-01 -1.60463229e-01 8.05807054e-01
1.12439632e-01 3.97726446e-01 4.65500504e-01 -1.00057900e+00
9.49783087e-01 1.27368546e+00 6.57381043e-02 -7.92467594e-01
-4.54503089e-01 -2.78365046e-01 -4.99286950e-01 -5.63277364e-01
-8.09722915e-02 7.05761239e-02 -5.78958333e-01 1.47196019e+00
2.41864562e-01 -3.95545185e-01 1.40402466e-01 4.78952348e-01
8.00288141e-01 6.04194462e-01 -1.05682589e-01 -3.06750596e-01
1.29184175e+00 -8.96311939e-01 -7.61930823e-01 -1.86318889e-01
9.48390543e-01 -8.92762542e-01 1.74141002e+00 3.58020455e-01
-9.65919435e-01 -8.21355999e-01 -8.20240796e-01 5.98562416e-03
-5.57728291e-01 -2.24137958e-02 5.82692683e-01 7.21821487e-01
-7.55717874e-01 3.12131226e-01 -7.64628649e-01 -7.19373524e-01
-4.13231939e-01 8.63019377e-02 -2.08101905e-04 9.30424407e-02
-1.10753214e+00 5.97523928e-01 6.83430374e-01 -1.06825784e-01
-1.95757702e-01 -6.76718831e-01 -7.64636099e-01 1.66097358e-02
9.41169322e-01 -4.70084578e-01 1.33298016e+00 -6.21695399e-01
-1.37600863e+00 4.56394583e-01 -6.30287945e-01 -2.40514666e-01
2.25086257e-01 -3.13347995e-01 -5.19041717e-01 -4.52799559e-01
1.20254643e-01 3.22767377e-01 2.91811973e-01 -1.27368176e+00
-9.51169908e-01 -5.20791888e-01 1.96055144e-01 1.69651672e-01
-1.93609551e-01 1.68459162e-01 -9.48393345e-01 -6.87614083e-01
2.22162902e-01 -1.03469181e+00 -6.98405802e-02 -8.43381345e-01
-6.87705994e-01 -3.31440985e-01 6.64749205e-01 -4.07528609e-01
1.75643790e+00 -2.08304143e+00 1.96411148e-01 3.48799884e-01
-1.37355044e-01 -8.47383961e-02 -1.27069414e-01 1.02776730e+00
-1.21423855e-01 5.90259433e-01 -2.00976312e-01 -2.02765822e-01
2.74366140e-01 3.72746378e-01 -6.91345274e-01 -5.32686651e-01
1.54739201e-01 9.64449406e-01 -5.52751064e-01 -4.78912622e-01
-3.35740834e-03 -4.06523883e-01 -8.74962866e-01 4.07213867e-01
-9.76379931e-01 7.24483132e-02 -4.63494539e-01 6.44970059e-01
6.33103192e-01 -1.48723572e-01 5.43441296e-01 -4.19765502e-01
4.79291379e-02 3.94832581e-01 -1.15506411e+00 1.66049159e+00
-4.45517302e-01 -7.68332332e-02 -4.57870096e-01 -8.28859210e-01
1.05237615e+00 4.88379858e-02 2.95842856e-01 -7.13936210e-01
-4.73995328e-01 2.31235474e-01 -1.65652275e-01 -9.44737911e-01
7.60322690e-01 5.03700711e-02 -4.28537250e-01 3.18986237e-01
-9.23685580e-02 -3.85034174e-01 5.55453479e-01 4.32369560e-01
1.00013268e+00 3.26961011e-01 4.90885615e-01 -3.51723805e-02
7.11971164e-01 2.54217647e-02 4.86446202e-01 9.00592744e-01
3.65338624e-01 4.11579311e-01 5.86707413e-01 -4.81400460e-01
-6.15983367e-01 -1.12654591e+00 9.31201577e-02 1.76280260e+00
-4.33394220e-03 -1.31370914e+00 -5.69339752e-01 -1.06513762e+00
-2.07471475e-01 1.38420153e+00 -4.19263005e-01 3.54689769e-02
-8.04213047e-01 -4.88625884e-01 7.57820010e-01 5.40028989e-01
3.38509917e-01 -9.88378525e-01 -4.78541881e-01 2.32765719e-01
-5.31798542e-01 -1.06214893e+00 -6.08617723e-01 2.61300206e-01
-8.18548024e-01 -1.16529429e+00 1.20312184e-01 -7.74418414e-01
3.72242272e-01 -2.03068376e-01 1.19940972e+00 1.26162425e-01
1.63002610e-01 4.60954368e-01 -7.91215122e-01 -5.17905712e-01
-6.34849191e-01 2.55157828e-01 -4.00931597e-01 -4.49402891e-02
4.62112755e-01 -3.37282419e-01 6.90398291e-02 5.34830570e-01
-1.44926429e+00 4.50277776e-01 4.08781230e-01 5.38795710e-01
6.10533893e-01 1.75353184e-01 4.79657680e-01 -1.44321823e+00
7.40638316e-01 -3.07848215e-01 -3.10808510e-01 7.79139400e-01
-4.72196579e-01 4.27992105e-01 8.55765641e-01 -3.52252692e-01
-1.38802767e+00 2.34218821e-01 -1.97559729e-01 1.02702767e-01
-2.97870934e-01 1.12699175e+00 -5.49815893e-01 6.07103944e-01
7.14758277e-01 3.28140587e-01 -4.15582120e-01 -5.84787428e-01
5.63191235e-01 4.62723911e-01 5.89227140e-01 -1.13478601e+00
6.00189328e-01 9.42651778e-02 -3.42056096e-01 -7.10700929e-01
-7.10139334e-01 -2.57766098e-01 -5.01726925e-01 1.01374954e-01
6.76994860e-01 -2.71604896e-01 -5.30666709e-01 2.24249810e-01
-1.24780905e+00 -2.20329165e-01 -1.51396781e-01 -1.92196388e-02
-4.32972342e-01 5.56929410e-01 -3.70436311e-01 -7.33902752e-01
-4.74087894e-02 -1.14482355e+00 1.24894500e+00 -1.05915898e-02
-5.15273809e-01 -9.84039426e-01 -1.42195314e-01 3.70860457e-01
4.48592573e-01 6.44586757e-02 1.70929503e+00 -1.34324372e+00
-5.42326868e-01 -1.21150479e-01 2.13591769e-01 1.73349187e-01
4.68820632e-01 1.91270739e-01 -3.91348481e-01 -7.25024641e-02
3.18894416e-01 -1.75427005e-01 2.27356210e-01 -3.99654269e-01
1.45326984e+00 -6.15367889e-01 -2.57720679e-01 4.75561470e-01
1.13352537e+00 5.22848010e-01 6.31025672e-01 2.61744440e-01
5.40058613e-01 7.38011837e-01 8.05881083e-01 6.33265913e-01
6.86887622e-01 8.24884892e-01 4.19953018e-01 3.69165480e-01
2.52667755e-01 -3.80563200e-01 3.04094881e-01 9.63183284e-01
1.70569852e-01 -2.86971956e-01 -1.00800645e+00 2.78400719e-01
-1.95827878e+00 -8.13248813e-01 -2.68510431e-01 2.09959364e+00
1.12809491e+00 1.44118309e-01 -1.83045179e-01 -2.14886129e-01
4.42597687e-01 -2.78397445e-02 -3.92577589e-01 -4.20629948e-01
-7.75286555e-02 3.56752843e-01 -1.21123739e-01 4.08346385e-01
-8.80148351e-01 9.89638627e-01 5.99296093e+00 6.81987107e-01
-8.45542133e-01 -3.63656759e-01 3.00643653e-01 -1.75242230e-01
-8.43610108e-01 1.65215492e-01 -1.06503332e+00 3.91307175e-01
7.78543651e-01 -2.31138334e-01 4.74639297e-01 7.32225358e-01
8.75591114e-03 3.56471958e-03 -1.75571048e+00 5.63563704e-01
2.52456695e-01 -1.18477750e+00 6.86231136e-01 -4.14400101e-01
7.49935508e-01 -4.46272820e-01 -6.26272559e-02 8.43541861e-01
2.03924477e-01 -8.09450328e-01 5.50484717e-01 6.40496671e-01
4.94164884e-01 -4.67791647e-01 8.28322947e-01 5.23124635e-01
-1.41170478e+00 1.55202048e-02 -1.11625902e-01 3.16079140e-01
-1.51504728e-03 3.08966190e-01 -1.01250970e+00 1.08919859e+00
8.76403391e-01 5.10393977e-01 -1.22174609e+00 2.56178945e-01
8.09991881e-02 5.05079687e-01 -1.69735432e-01 -2.76794642e-01
-8.05847570e-02 -3.23699564e-02 4.23034996e-01 1.42348599e+00
4.49909598e-01 1.08200662e-01 3.06164563e-01 9.98078287e-01
1.32766113e-01 4.41133380e-01 -4.57865417e-01 -1.07435891e-02
5.13314545e-01 7.78864384e-01 -3.07264596e-01 -7.41849482e-01
-6.88741744e-01 7.87470579e-01 1.59772545e-01 6.26587689e-01
-1.02511823e+00 -6.19366765e-01 3.21814656e-01 1.99747548e-01
4.06331629e-01 8.62560272e-02 -4.19363081e-01 -1.21702063e+00
6.89911127e-01 -1.49586701e+00 6.27770901e-01 -8.18269372e-01
-1.25841415e+00 6.88036561e-01 5.01055002e-01 -9.55749273e-01
-4.59692627e-01 -3.32632005e-01 -5.27717829e-01 6.48186326e-01
-7.08972871e-01 -1.11537778e+00 -3.92366737e-01 7.72760987e-01
7.75500178e-01 -3.58791918e-01 9.35347676e-01 1.66425742e-02
-2.93643713e-01 6.77807271e-01 -3.18497568e-01 -9.27960724e-02
7.87117898e-01 -1.25107086e+00 5.12564540e-01 1.02392495e+00
-2.83606304e-03 1.38892233e+00 7.33472764e-01 -9.63169754e-01
-1.59334409e+00 -1.44122982e+00 1.09068155e+00 -5.85246146e-01
5.97022772e-01 -5.06934702e-01 -1.31147218e+00 1.26616454e+00
2.31918305e-01 -6.74795628e-01 7.99365163e-01 1.93004668e-01
-4.80306238e-01 -3.37625384e-01 -7.34077454e-01 9.23613191e-01
1.23808062e+00 -6.38629258e-01 -1.04331279e+00 1.75977856e-01
1.32315648e+00 -5.55584610e-01 -8.71454120e-01 8.62153471e-01
2.95507252e-01 -1.08461773e+00 7.31674671e-01 -9.47243392e-01
2.91863263e-01 -3.39947164e-01 -7.85438418e-01 -1.02628601e+00
1.42401129e-01 -5.76998115e-01 -7.07231238e-02 1.54280829e+00
7.51226962e-01 -5.98034441e-01 6.44327879e-01 9.07721043e-01
-4.57583994e-01 -7.29485333e-01 -6.06704652e-01 -7.78800786e-01
-1.37214795e-01 -8.46829534e-01 1.18472898e+00 7.03973889e-01
3.93706739e-01 3.63904268e-01 -1.06218681e-01 1.36486813e-01
7.75665715e-02 3.97117883e-01 1.13766944e+00 -8.24766755e-01
-6.55235529e-01 -3.90469253e-01 1.89937055e-01 -1.39602518e+00
-1.09890001e-02 -1.05464458e+00 1.19829655e-01 -1.31524420e+00
1.61697030e-01 -2.54882783e-01 1.50513783e-01 6.15586758e-01
-5.64706087e-01 -6.59364522e-01 3.46627295e-01 -4.40855511e-02
-8.08340907e-01 5.09493291e-01 8.89504135e-01 -2.47339815e-01
-5.36055923e-01 3.02506864e-01 -7.79158890e-01 4.22701150e-01
7.48865724e-01 -2.74628282e-01 -8.06027532e-01 -4.54696566e-01
6.12057924e-01 1.95693642e-01 -1.08408794e-01 -8.13859701e-01
2.82949656e-01 -6.03281856e-01 -2.36474559e-01 -8.78322005e-01
2.02535316e-01 -8.53845537e-01 4.81125563e-01 2.39907756e-01
-8.39694619e-01 3.69087338e-01 2.67690450e-01 2.14892238e-01
-4.58577871e-01 -2.87252098e-01 1.10782757e-01 -7.68330991e-02
-5.63014090e-01 -2.09868237e-01 -4.45033103e-01 2.16442600e-01
6.98425412e-01 3.71242464e-02 -4.16896999e-01 -4.39674407e-01
-6.55499816e-01 2.90040135e-01 3.08718503e-01 7.37059534e-01
7.21805453e-01 -1.33013940e+00 -5.52220285e-01 3.95749331e-01
6.13378584e-01 1.97091922e-01 -6.22553676e-02 6.26274943e-01
-4.03506219e-01 7.36272275e-01 1.84842497e-01 -5.52931845e-01
-1.37189221e+00 7.67572045e-01 1.73525602e-01 -3.73618543e-01
-5.30454665e-02 5.88331699e-01 2.97557384e-01 -1.21335471e+00
4.70617384e-01 -1.15031028e+00 8.40451643e-02 -4.51234549e-01
3.32188070e-01 2.79429853e-01 3.56470704e-01 -7.61701167e-02
-4.59917992e-01 3.32699955e-01 -2.15156913e-01 -9.94766578e-02
1.12247765e+00 -1.49706811e-01 -4.77923930e-01 6.68454766e-01
1.08246505e+00 4.13585722e-01 -4.19103682e-01 -4.38257337e-01
7.55868852e-01 -3.94633710e-01 -1.13032186e+00 -9.97887552e-01
-7.05217302e-01 3.68189275e-01 4.64626811e-02 3.84532988e-01
1.30115592e+00 1.92770898e-01 5.34709811e-01 9.65595782e-01
4.57535982e-01 -9.51447964e-01 2.22498134e-01 8.24268878e-01
1.37060702e+00 -9.02762532e-01 -1.51839226e-01 -6.28500879e-01
-9.36282277e-01 1.18537509e+00 1.02151465e+00 5.64745009e-01
3.65191370e-01 3.82157445e-01 -1.28681939e-02 -2.01074317e-01
-1.23228049e+00 1.28479913e-01 3.75153512e-01 3.98839086e-01
7.22433567e-01 4.92239259e-02 -2.86623865e-01 1.01534116e+00
-5.45268357e-01 -2.42575541e-01 3.60723615e-01 1.08127987e+00
-3.36788803e-01 -1.51370287e+00 -3.06799918e-01 6.10969603e-01
-9.31777805e-02 -1.55736566e-01 -7.73916304e-01 8.65374625e-01
1.21062227e-01 1.15914237e+00 -1.62048817e-01 -7.21139193e-01
8.46331239e-01 5.89801669e-01 9.89346355e-02 -9.81014073e-01
-8.82684886e-01 3.43267471e-02 3.23274404e-01 -8.57460797e-01
-1.55494496e-01 -6.82234168e-01 -1.64729273e+00 -1.89822331e-01
-7.16843009e-02 2.50044197e-01 3.66787583e-01 9.15351212e-01
5.16467214e-01 5.73770881e-01 5.02596438e-01 -1.02530934e-01
-6.26680434e-01 -9.86166358e-01 -4.95624058e-02 9.07165766e-01
-1.41213387e-01 1.47260074e-02 -1.29104331e-02 3.83604765e-01]
|
[9.861727714538574, 7.85487174987793]
|
64ff71b5-3dfc-492e-baeb-8ba7d999280e
|
deep-learning-for-direction-of-arrival
|
2007.13824
| null |
https://arxiv.org/abs/2007.13824v3
|
https://arxiv.org/pdf/2007.13824v3.pdf
|
Deep Learning for DOA Estimation in MIMO Radar Systems via Emulation of Large Antenna Arrays
|
We present a MUSIC-based Direction of Arrival (DOA) estimation strategy using small antenna arrays, via employing deep learning for reconstructing the signals of a virtual large antenna array. Not only does the proposed strategy deliver significantly better performance than simply plugging the incoming signals into MUSIC, but surprisingly, the performance is also better than directly using an actual large antenna array with MUSIC for high angle ranges and low test SNR values. We further analyze the best choice for the training SNR as a function of the test SNR, and observe dramatic changes in the behavior of this function for different angle ranges.
|
['Udaya Sampath K. P. Miriya Thanthrige', 'Aydin Sezgin', 'Aya Mostafa Ahmed', 'Aly El Gamal']
|
2020-07-27
| null | null | null | null |
['direction-of-arrival-estimation']
|
['audio']
|
[-7.14667365e-02 -3.20084453e-01 5.59410930e-01 5.19877747e-02
-1.06587899e+00 -9.04924273e-01 2.96359450e-01 -2.66471297e-01
-3.21657002e-01 4.56642479e-01 4.77483809e-01 -5.15082955e-01
-6.06430233e-01 -5.75422466e-01 -5.11368632e-01 -1.08831692e+00
-4.02213037e-01 1.11340396e-01 -1.32966697e-01 3.92832085e-02
-7.58916065e-02 5.59222639e-01 -1.05726659e+00 -8.27270299e-02
2.44573295e-01 1.10385084e+00 -2.94331312e-01 9.21285450e-01
3.63785177e-01 4.24938500e-01 -1.04902053e+00 -1.75806284e-01
8.21468592e-01 -6.52115583e-01 2.38571107e-01 -1.44932196e-01
4.78632778e-01 -1.91571817e-01 -4.75870669e-01 7.30544508e-01
1.16580200e+00 -1.26030505e-01 4.82218564e-01 -7.37467408e-01
3.22404981e-01 7.03241527e-01 -2.57205963e-01 3.27323318e-01
2.88379788e-01 -1.80788174e-01 8.09420049e-01 -8.14217448e-01
1.01693332e-01 6.56301677e-01 1.11980379e+00 -4.62932229e-01
-8.87531698e-01 -7.50949502e-01 -7.34832883e-01 -2.14781910e-02
-1.64135623e+00 -7.57740557e-01 8.38352084e-01 -1.24079138e-01
3.56284946e-01 2.60247648e-01 5.94266951e-01 8.67611468e-01
1.60950050e-01 1.97011068e-01 8.67406666e-01 -6.10166669e-01
5.35635650e-01 -3.42438161e-01 -1.60866961e-01 3.25565517e-01
5.27322710e-01 2.51264751e-01 -3.18276554e-01 -4.59297210e-01
8.60377967e-01 -2.87997007e-01 -4.67131257e-01 -4.39416766e-01
-1.64186823e+00 5.94040692e-01 2.73157060e-01 3.25865477e-01
-6.56250000e-01 6.99510872e-01 8.19084793e-02 7.26056039e-01
-5.50291464e-02 7.50439048e-01 -5.74474454e-01 -1.21417299e-01
-1.01379073e+00 1.07359685e-01 1.00925469e+00 4.05794203e-01
3.28888595e-01 7.32167006e-01 -2.82041848e-01 7.85415649e-01
2.60321379e-01 1.00412476e+00 1.17363907e-01 -1.20860422e+00
4.39054579e-01 -5.70317388e-01 5.36906600e-01 -1.08407080e+00
-7.91555941e-01 -1.84069955e+00 -7.65891492e-01 4.77427430e-02
8.51558089e-01 -1.14247620e+00 -5.09655237e-01 1.55404210e+00
1.10054798e-01 4.91996199e-01 2.94594258e-01 7.98724234e-01
2.62144238e-01 6.15502119e-01 -5.48574865e-01 -5.13701618e-01
8.06209743e-01 -5.82270443e-01 -7.46903002e-01 -4.44590628e-01
3.42283696e-01 -9.55248654e-01 4.27798241e-01 6.50134027e-01
-8.82217526e-01 -4.22502846e-01 -1.26786089e+00 8.11892986e-01
2.92222112e-01 3.68072122e-01 6.46479070e-01 9.75487709e-01
-1.06899619e+00 1.96343914e-01 -3.94265920e-01 6.77723214e-02
9.23800245e-02 6.73205703e-02 -2.35295407e-02 -2.89969116e-01
-7.61625469e-01 2.83771366e-01 -3.92266959e-02 3.04576844e-01
-5.29098868e-01 -5.97574413e-01 -6.13397896e-01 3.77820730e-01
8.11693370e-02 -6.64422691e-01 9.87293661e-01 -6.20540082e-01
-1.44355476e+00 4.56460342e-02 -1.43562824e-01 -5.49615562e-01
2.22595572e-01 -2.50705719e-01 -8.11435401e-01 1.38221169e-02
-1.27202615e-01 5.61135933e-02 1.02026916e+00 -1.35969234e+00
-2.53608376e-01 -1.01641141e-01 -2.67756313e-01 -1.43985404e-02
-6.46114275e-02 -4.74900186e-01 -2.19732121e-01 -9.37214315e-01
6.38519049e-01 -8.41399133e-01 -4.99178976e-01 -2.70064831e-01
-9.82348546e-02 6.04866982e-01 -3.79969482e-03 -5.05196989e-01
1.22454524e+00 -2.33722591e+00 -3.55530053e-01 7.40650952e-01
-1.34985698e-02 -2.15971041e-02 -4.37820047e-01 6.12170219e-01
-7.11425170e-02 -4.80864704e-01 1.90383121e-01 1.89184830e-01
-2.49209523e-01 -1.31126478e-01 -3.17325234e-01 5.28858483e-01
-5.90698779e-01 6.62056923e-01 -7.31297433e-01 2.11506918e-01
-4.78648953e-03 2.57853925e-01 -7.98841715e-01 8.05477351e-02
3.40245456e-01 4.93309408e-01 -2.98140466e-01 4.90472019e-01
6.78727090e-01 -3.14569741e-01 2.76397258e-01 -3.89133573e-01
-2.19067648e-01 -5.51577471e-03 -1.49275482e+00 1.41391969e+00
-7.12820292e-01 9.94492650e-01 2.95325547e-01 -9.95691359e-01
9.79383051e-01 5.45647442e-01 7.30682433e-01 -5.73232174e-01
2.49473572e-01 4.79796022e-01 4.02387232e-01 -2.10336611e-01
-1.67231232e-01 -3.11854631e-01 -9.37397555e-02 4.56202537e-01
1.72867682e-02 1.01963073e-01 -2.10659325e-01 -1.90730784e-02
1.58202434e+00 -4.53866899e-01 3.65383148e-01 1.50149474e-02
2.10897267e-01 -3.07642758e-01 3.24871570e-01 1.15892839e+00
1.51001275e-01 5.39615810e-01 3.29136431e-01 -2.50751913e-01
-7.98000455e-01 -1.12294042e+00 -2.07047224e-01 8.56296837e-01
-9.21209231e-02 -2.63771176e-01 -1.98650971e-01 -2.22001240e-01
4.33934964e-02 7.91530192e-01 -2.10209563e-01 1.01096377e-01
-5.81132710e-01 -9.92818475e-01 6.28091097e-01 3.39367867e-01
5.68648994e-01 -4.99807626e-01 -3.20700318e-01 2.68012732e-01
-1.15841188e-01 -1.28196514e+00 -1.29960999e-01 3.74910414e-01
-7.06955254e-01 -5.24612308e-01 -7.82262981e-01 -5.95022917e-01
3.27526182e-01 6.09334111e-01 9.29885685e-01 -2.78005570e-01
3.85953039e-01 7.36780226e-01 -2.93858320e-01 -5.44494510e-01
-1.15168050e-01 -2.39464641e-01 1.48596108e-01 2.38728464e-01
-4.43325371e-01 -1.17383170e+00 -7.28293180e-01 5.91844738e-01
-4.51948285e-01 -4.94161218e-01 1.20916104e+00 6.93872392e-01
2.22037494e-01 3.41439307e-01 8.56122136e-01 -2.17993721e-01
3.57564360e-01 -6.52678728e-01 -7.09428012e-01 -2.08853424e-01
-3.55356157e-01 -9.84274372e-02 7.36364901e-01 -2.26263836e-01
-6.66179240e-01 -4.78939107e-03 -4.28494185e-01 -1.24510750e-01
4.60286811e-02 6.83859408e-01 -6.46693334e-02 -3.54706466e-01
7.79574096e-01 1.77314281e-01 -4.32748675e-01 -3.86785299e-01
2.02410251e-01 5.43363273e-01 6.67264521e-01 1.03122527e-02
1.00781620e+00 5.10221779e-01 4.72662032e-01 -9.33571696e-01
-8.63944888e-01 -6.62397504e-01 -2.92496234e-01 -1.25772402e-01
5.03225476e-02 -1.01652920e+00 -6.36619747e-01 3.12593102e-01
-8.67526710e-01 -2.82560408e-01 -9.46358144e-02 1.14168966e+00
-3.93661022e-01 1.56450391e-01 -1.30448386e-01 -6.38462722e-01
-3.07911783e-01 -6.49373591e-01 7.85845697e-01 -2.30186254e-01
1.75950341e-02 -8.09811711e-01 2.96998233e-01 4.07170691e-02
7.64232278e-01 1.45411268e-01 6.32659197e-01 -5.27089477e-01
-4.60788131e-01 -7.35258579e-01 -9.31704137e-03 2.45662048e-01
-5.31373406e-03 -7.16962516e-01 -8.84075522e-01 -5.15010357e-01
1.14048451e-01 1.81601062e-01 4.02475476e-01 9.53273177e-01
6.56539142e-01 -4.13067907e-01 -1.40969515e-01 1.06477046e+00
1.63210309e+00 2.20735431e-01 6.90009415e-01 2.05652311e-01
4.47424382e-01 -2.28640750e-01 3.49950135e-01 7.98271239e-01
-1.67284995e-01 8.22240233e-01 3.05893004e-01 5.31594567e-02
-2.47839168e-01 1.89025570e-02 2.22207397e-01 9.98825550e-01
1.11405388e-01 -6.36239052e-01 -7.50512540e-01 2.77615756e-01
-1.49619031e+00 -9.13120806e-01 -2.23254949e-01 2.49025655e+00
1.99833125e-01 1.50628775e-01 4.39849049e-02 4.26286429e-01
5.88330738e-02 2.77792782e-01 -3.20416421e-01 1.04982041e-01
-3.19568902e-01 3.15280825e-01 8.89317751e-01 3.88139606e-01
-9.29712474e-01 2.53455281e-01 7.95357513e+00 7.07203209e-01
-1.29548657e+00 5.04297055e-02 -3.80825363e-02 -1.52769297e-01
-3.83102804e-01 -2.07130462e-01 -2.50558138e-01 2.83300012e-01
1.02869427e+00 5.88838905e-02 4.25810754e-01 5.60303748e-01
2.98721820e-01 5.66615649e-02 -7.17894614e-01 1.26677454e+00
-6.61791340e-02 -1.28642571e+00 -1.68798879e-01 2.10693642e-01
7.53987491e-01 2.65595466e-01 6.54513538e-02 1.65988892e-01
-6.76738322e-02 -7.19763398e-01 5.10256469e-01 5.34116685e-01
4.84773397e-01 -6.51848793e-01 1.07742310e+00 4.10950691e-01
-7.69955575e-01 -6.00208044e-01 -2.64503837e-01 -1.73558414e-01
3.31307203e-02 1.09910679e+00 -9.85543311e-01 4.17463452e-01
2.48376846e-01 2.08705202e-01 -4.11137551e-01 1.86799252e+00
-1.48219332e-01 1.33951139e+00 -6.22438550e-01 1.65207744e-01
2.13870943e-01 -9.04633477e-02 1.09957504e+00 1.10377800e+00
9.73266959e-01 7.00161010e-02 3.68350744e-02 -8.68433043e-02
3.81516316e-03 5.23751453e-02 -6.44718587e-01 2.50152826e-01
7.14764059e-01 1.14652872e+00 -5.46164870e-01 -2.13131458e-02
-2.86142677e-01 4.71747458e-01 -4.98661280e-01 8.82268012e-01
-5.31967342e-01 -3.99958402e-01 3.49366248e-01 2.84627438e-01
8.33427131e-01 -8.48421156e-01 -4.33540612e-01 -7.28986442e-01
-1.82530973e-02 -7.99279749e-01 2.23692060e-01 -8.69870663e-01
-6.76205039e-01 2.72769749e-01 -3.51505309e-01 -1.80331922e+00
-5.36131203e-01 -3.26918334e-01 -7.48398066e-01 5.39843559e-01
-9.74257946e-01 -6.53920650e-01 -1.89536840e-01 3.95062029e-01
1.86639085e-01 -4.33273435e-01 8.13641012e-01 5.91260433e-01
-8.26643314e-03 8.04975986e-01 6.96816206e-01 1.58994332e-01
5.49532831e-01 -9.06157374e-01 9.13763195e-02 8.93209994e-01
5.42005897e-01 4.11539108e-01 1.23374069e+00 -5.95891066e-02
-1.73115528e+00 -7.89897144e-01 2.52413541e-01 -4.01563823e-01
5.63189507e-01 -2.51899213e-01 -3.54546636e-01 6.01557732e-01
2.05641076e-01 2.40884066e-01 9.53198493e-01 1.83735892e-01
-9.51805338e-02 -4.40067351e-01 -7.41168737e-01 5.08796871e-01
8.22989941e-01 -3.92853133e-02 -1.84165895e-01 3.01917315e-01
4.62045461e-01 -2.40077466e-01 -6.58839643e-01 5.33337355e-01
9.38233316e-01 -1.07717609e+00 1.35381913e+00 -1.50632143e-01
-1.99172646e-01 -5.00517964e-01 -6.72845423e-01 -1.55375171e+00
-6.12792373e-01 -8.12753499e-01 -7.09916875e-02 6.74169540e-01
5.17388940e-01 -7.49442339e-01 7.95784950e-01 -5.62534988e-01
-2.05089137e-01 -5.18167078e-01 -1.03429186e+00 -8.77147257e-01
-3.29063565e-01 -8.58328700e-01 2.39929080e-01 7.07459152e-01
-4.65880781e-01 4.81826335e-01 -4.91939425e-01 8.66527796e-01
8.26128900e-01 1.72694013e-01 1.00249708e+00 -1.07922208e+00
-1.02978539e+00 -2.04756916e-01 -5.47067344e-01 -1.24367154e+00
-4.68883604e-01 -6.47292554e-01 5.82885649e-03 -1.32895386e+00
-5.94971895e-01 -7.33639300e-01 -5.03479898e-01 1.64492484e-02
1.23569809e-01 5.83704829e-01 -1.13822661e-01 8.47977102e-02
-5.17799258e-01 1.66409045e-01 1.02817953e+00 1.00885637e-01
-2.98093438e-01 8.82018924e-01 -6.90373659e-01 8.14659417e-01
5.47732294e-01 -4.92478549e-01 -2.52148062e-01 -6.49496913e-01
7.63291955e-01 6.34581506e-01 2.36033291e-01 -1.67255056e+00
1.53970242e-01 2.23896816e-01 7.51753986e-01 -5.75365245e-01
2.51679897e-01 -7.76186645e-01 4.33278888e-01 3.19157511e-01
-1.55562714e-01 -1.67030007e-01 3.54325354e-01 8.00192893e-01
-1.98480621e-01 6.78322539e-02 4.68298197e-01 3.90780479e-01
-2.53277212e-01 -2.19344079e-01 -6.38367176e-01 -1.89651966e-01
4.86695290e-01 -9.38825235e-02 1.97467640e-01 -1.13414097e+00
-6.72608078e-01 -1.90507874e-01 -1.00728162e-01 -1.47179276e-01
2.38984674e-01 -1.45364392e+00 -9.04050052e-01 4.00441408e-01
-2.08185583e-01 -5.40723503e-01 8.66214335e-02 9.43566561e-01
-3.96539420e-01 5.41785121e-01 -7.51935169e-02 -5.45213699e-01
-7.91140318e-01 1.66784838e-01 4.53304201e-01 3.78908427e-03
-3.24320048e-01 7.00677574e-01 1.20086949e-02 -2.15350628e-01
2.92147040e-01 4.64023389e-02 -7.39745796e-02 -9.95213613e-02
4.79420841e-01 5.33232629e-01 2.63582915e-01 -2.78395653e-01
-2.29986548e-01 8.07579160e-01 8.57729435e-01 -3.88654947e-01
1.20255411e+00 -2.59269744e-01 2.04352647e-01 2.78885007e-01
1.11431754e+00 1.17470074e+00 -8.99584115e-01 -2.40159824e-01
-4.07364935e-01 -4.56903994e-01 2.06286296e-01 -9.69089568e-01
-1.13032484e+00 5.73231697e-01 8.16845894e-01 1.93701729e-01
1.15749240e+00 -1.14081725e-01 6.26669466e-01 9.54610705e-01
7.47042716e-01 -3.29848528e-01 7.75254518e-02 5.11004448e-01
7.41355717e-01 -7.07150996e-01 1.05829991e-01 -1.00424342e-01
-1.49415269e-01 1.13037646e+00 -2.37907127e-01 -2.38186777e-01
8.43791485e-01 4.37432498e-01 6.37188256e-01 -2.35537559e-01
-3.16239089e-01 -1.38313353e-01 2.67460823e-01 5.70534348e-01
2.68827260e-01 4.33435589e-02 -4.99298349e-02 5.56598783e-01
-6.41115248e-01 -2.51254350e-01 5.38265467e-01 5.16741633e-01
-6.93914771e-01 -8.28231037e-01 -8.84152770e-01 6.55040026e-01
-4.75523561e-01 -2.11329356e-01 2.96616312e-02 4.78333890e-01
-8.93210177e-04 9.24659431e-01 6.88458830e-02 -4.74052995e-01
4.14586991e-01 -2.92790961e-02 3.81583601e-01 -2.10205913e-01
-2.04786807e-01 6.00875199e-01 3.82249922e-01 -3.98474425e-01
-3.11831180e-02 -6.15790248e-01 -8.69761050e-01 8.16988274e-02
-2.58498132e-01 3.87414902e-01 5.54556429e-01 9.06636834e-01
4.41426337e-01 5.54756999e-01 1.07365406e+00 -7.51588702e-01
-6.61363184e-01 -6.66698992e-01 -7.23625004e-01 -2.61428088e-01
6.69333756e-01 -3.68692338e-01 -6.79959297e-01 -5.11452913e-01]
|
[6.487947940826416, 1.327072262763977]
|
707d544a-7930-48b8-8217-780fa7fd20f4
|
tiled-sparse-coding-in-eigenspaces-for-the
|
2106.14724
| null |
https://arxiv.org/abs/2106.14724v1
|
https://arxiv.org/pdf/2106.14724v1.pdf
|
Tiled sparse coding in eigenspaces for the COVID-19 diagnosis in chest X-ray images
|
The ongoing crisis of the COVID-19 (Coronavirus disease 2019) pandemic has changed the world. According to the World Health Organization (WHO), 4 million people have died due to this disease, whereas there have been more than 180 million confirmed cases of COVID-19. The collapse of the health system in many countries has demonstrated the need of developing tools to automatize the diagnosis of the disease from medical imaging. Previous studies have used deep learning for this purpose. However, the performance of this alternative highly depends on the size of the dataset employed for training the algorithm. In this work, we propose a classification framework based on sparse coding in order to identify the pneumonia patterns associated with different pathologies. Specifically, each chest X-ray (CXR) image is partitioned into different tiles. The most relevant features extracted from PCA are then used to build the dictionary within the sparse coding procedure. Once images are transformed and reconstructed from the elements of the dictionary, classification is performed from the reconstruction errors of individual patches associated with each image. Performance is evaluated in a real scenario where simultaneously differentiation between four different pathologies: control vs bacterial pneumonia vs viral pneumonia vs COVID-19. The accuracy when identifying the presence of pneumonia is 93.85%, whereas 88.11% is obtained in the 4-class classification context. The excellent results and the pioneering use of sparse coding in this scenario evidence the applicability of this approach as an aid for clinicians in a real-world environment.
|
['Juan M Gorriz', 'Javier Ramírez', 'Andrés Ortiz', 'Juan E. Arco']
|
2021-06-28
| null | null | null | null |
['covid-19-detection']
|
['medical']
|
[ 3.60967189e-01 -4.35750514e-01 1.32710814e-01 1.19325355e-01
-6.00623369e-01 -3.34375739e-01 4.07740057e-01 2.31287137e-01
-4.38007444e-01 5.00006855e-01 3.00184369e-01 -8.70394781e-02
-2.13164449e-01 -5.84260345e-01 -2.83368260e-01 -1.04325843e+00
-8.97049308e-02 7.21808970e-01 -1.35968357e-01 2.73841262e-01
2.38184538e-02 7.10309923e-01 -1.52507055e+00 5.77158332e-01
6.35783970e-01 9.56510842e-01 7.23899424e-01 7.57340074e-01
-4.48534787e-02 4.70561206e-01 -5.21701634e-01 2.89334148e-01
2.46146798e-01 -3.93870354e-01 -4.29654837e-01 9.06249210e-02
7.71892220e-02 -3.49908561e-01 3.17251198e-02 7.21760213e-01
5.14836669e-01 -1.85206592e-01 8.78738284e-01 -7.78101981e-01
1.51932493e-01 -1.90010652e-01 -3.88963103e-01 4.50801402e-01
4.30811048e-01 5.64522929e-02 6.19231582e-01 -8.47081661e-01
7.93928266e-01 8.26524317e-01 9.07593191e-01 2.27291003e-01
-1.10294306e+00 -4.22683746e-01 -4.51370746e-01 3.30895483e-01
-1.40083003e+00 -6.47571832e-02 5.76068103e-01 -9.56090927e-01
9.41608489e-01 4.19728994e-01 8.12043309e-01 1.02750504e+00
4.53281134e-01 1.66023031e-01 1.17094040e+00 -3.04751158e-01
3.12328935e-01 1.72631890e-01 -8.11625198e-02 5.63679039e-01
6.21292174e-01 -8.49902555e-02 -1.74236335e-02 -5.21921039e-01
5.96720099e-01 5.35315156e-01 -3.88485402e-01 -4.48012173e-01
-1.29027426e+00 1.05725014e+00 2.21609756e-01 8.63008440e-01
-9.13075030e-01 -3.52157325e-01 3.86252493e-01 -1.12246461e-01
1.52441233e-01 3.54071796e-01 -1.74034491e-01 1.47463888e-01
-1.25926864e+00 9.05110687e-02 6.41417444e-01 1.88336670e-01
3.93226743e-01 5.53697795e-02 -3.52288075e-02 8.17676783e-01
3.91504735e-01 9.61290896e-01 7.21772611e-01 -7.13282645e-01
3.52003425e-01 6.26432419e-01 -5.73639795e-02 -1.46810198e+00
-7.30035722e-01 -4.80754942e-01 -1.34148264e+00 -5.14258780e-02
1.70611918e-01 -1.92772418e-01 -7.46112049e-01 1.23756003e+00
3.66170645e-01 2.54032731e-01 1.43873736e-01 8.04244101e-01
5.95543265e-01 9.19040024e-01 -1.16752431e-01 -2.81265706e-01
1.57955420e+00 -3.39224994e-01 -5.32384098e-01 1.62474185e-01
5.20527065e-01 -7.29195535e-01 4.82970893e-01 4.78605539e-01
-5.32432735e-01 -4.45232540e-01 -8.52360308e-01 6.16150022e-01
-1.45180792e-01 3.02593797e-01 3.68179679e-02 5.96538186e-01
-8.59432876e-01 1.09625779e-01 -8.42002869e-01 -8.43737721e-01
4.06609237e-01 2.94931650e-01 -3.91463101e-01 -2.89512038e-01
-8.15255225e-01 9.93652761e-01 2.34988302e-01 -9.54994187e-02
-7.81015575e-01 -6.06225967e-01 -4.79179978e-01 -2.54471414e-02
-1.83234215e-01 -8.99890542e-01 3.89058679e-01 -6.75805748e-01
-8.12185168e-01 9.83446240e-01 -1.90538377e-01 -5.21230996e-01
3.21619928e-01 2.90092379e-02 -4.45905060e-01 5.75855017e-01
2.02979624e-01 4.50437367e-01 9.24787700e-01 -1.24381268e+00
-6.84792995e-01 -3.52263272e-01 -4.06743407e-01 -2.11551972e-02
-1.36025995e-01 1.36797458e-01 -1.32206216e-01 -7.53608584e-01
8.63857046e-02 -1.25090051e+00 -2.82237381e-01 -2.98881441e-01
-8.23854953e-02 1.07803956e-01 8.85119438e-01 -9.12072539e-01
1.06410849e+00 -2.33166218e+00 9.07470414e-04 3.03420842e-01
2.13952988e-01 4.02271003e-01 8.17238390e-02 4.63303328e-01
-1.81414217e-01 -2.25222722e-01 -6.55222237e-01 6.56027943e-02
-5.55392563e-01 3.07039529e-01 -2.06423715e-01 6.72439754e-01
1.84296265e-01 4.12153214e-01 -6.57120645e-01 -5.52531004e-01
3.75620991e-01 9.70098257e-01 -6.11354172e-01 4.42672372e-01
2.96943963e-01 7.81020224e-01 -4.87096548e-01 4.69818711e-01
6.09281301e-01 -5.49053550e-01 4.00596112e-01 -2.25799233e-01
-4.07722453e-03 -1.43009260e-01 -1.02708030e+00 1.31060123e+00
-3.78514498e-01 4.96722221e-01 1.46651819e-01 -1.12789047e+00
5.90806186e-01 8.78277719e-01 1.08033550e+00 -3.89732271e-01
1.57838926e-01 2.34514594e-01 1.67439714e-01 -8.42102170e-01
-6.65491819e-02 -2.62668401e-01 3.44672501e-01 4.12051857e-01
-1.53648555e-01 -5.89552186e-02 1.16604440e-01 -1.41002193e-01
1.23525548e+00 -4.10483450e-01 4.99245346e-01 -3.78784746e-01
7.50395536e-01 3.70411098e-01 3.62103432e-01 3.91741008e-01
-8.50046277e-02 1.02485919e+00 8.59812647e-02 -6.16387606e-01
-9.96877849e-01 -1.01263320e+00 -4.86432642e-01 2.83278108e-01
-3.63571912e-01 2.74834093e-02 -8.11544716e-01 -4.26002920e-01
-1.65735289e-01 4.37260598e-01 -5.30791640e-01 2.19633400e-01
-8.23830724e-01 -1.13023257e+00 3.02982569e-01 1.11303113e-01
2.97558933e-01 -1.18545711e+00 -1.31901836e+00 2.58803189e-01
-3.93101871e-01 -1.04609072e+00 7.80096278e-02 1.06230237e-01
-1.01517451e+00 -1.34145772e+00 -9.05439734e-01 -7.76323378e-01
7.92366266e-01 2.33272687e-01 7.27216125e-01 8.21400359e-02
-7.09916174e-01 6.37192607e-01 -3.27035010e-01 -1.01903632e-01
-6.16501212e-01 -1.72579482e-01 1.22283787e-01 1.63774744e-01
1.04474589e-01 -4.10483301e-01 -7.51562417e-01 7.68108387e-03
-9.23308492e-01 -2.42713466e-01 6.87445462e-01 6.93922937e-01
6.53047204e-01 1.21277936e-01 4.65435803e-01 -7.46539772e-01
3.41799378e-01 -9.33456838e-01 -3.80201787e-01 -7.94606209e-02
-4.59151715e-01 -2.93762624e-01 5.95850945e-01 -1.83923423e-01
-7.13158309e-01 2.89534748e-01 -2.03084916e-01 -5.24743676e-01
-4.14819688e-01 4.78199452e-01 3.26780796e-01 3.09605092e-01
5.03030896e-01 3.62458199e-01 1.34943485e-01 -4.85127568e-01
-2.80510306e-01 7.60522366e-01 3.53208721e-01 -7.43227005e-02
7.04199791e-01 7.10233331e-01 1.38330787e-01 -1.22500467e+00
-3.98774117e-01 -8.46401393e-01 -5.49662292e-01 -2.00197145e-01
1.41724026e+00 -9.07797873e-01 -3.19708556e-01 3.47226262e-01
-1.13835025e+00 2.06806466e-01 -1.91156909e-01 6.62689328e-01
-3.28172147e-01 4.63579774e-01 -1.90635562e-01 -5.76601744e-01
-5.27312338e-01 -1.29420507e+00 8.42021406e-01 -2.46063203e-01
-5.09608567e-01 -9.10101950e-01 5.37528574e-01 5.55955827e-01
5.10221064e-01 4.27218318e-01 1.21049190e+00 -6.59890115e-01
-3.37868094e-01 -2.57057041e-01 -4.41418551e-02 4.82413113e-01
4.72277433e-01 -1.98058784e-01 -9.58509445e-01 -4.83201593e-01
6.30371988e-01 1.67072937e-01 6.16411746e-01 5.72737515e-01
7.43786097e-01 -2.98867792e-01 -3.85244519e-01 4.08211678e-01
1.63074565e+00 5.40883124e-01 4.11541939e-01 1.86156005e-01
6.88100159e-01 6.14610910e-01 3.45531255e-01 5.51599860e-01
1.19080506e-01 6.10517025e-01 3.49014968e-01 -3.01383138e-02
-2.36011028e-01 3.63352418e-01 2.09050044e-01 1.12022460e+00
-1.63242109e-02 -1.12389959e-01 -1.28688300e+00 6.29601777e-01
-1.30340600e+00 -1.03990209e+00 -1.24162570e-01 2.13059568e+00
3.60649943e-01 -3.33232969e-01 -5.89978397e-02 3.83762777e-01
5.96114933e-01 8.37338641e-02 -9.76278111e-02 -3.25880408e-01
-3.31374556e-02 3.39832723e-01 1.35217413e-01 1.71189189e-01
-1.17826176e+00 3.78992893e-02 6.33874083e+00 3.85478050e-01
-1.52023995e+00 2.05841795e-01 5.33483624e-01 1.29625887e-01
5.60435541e-02 -5.44611096e-01 -4.15317327e-01 6.49039984e-01
9.23460305e-01 2.10562199e-01 2.93839008e-01 6.03389978e-01
3.02777857e-01 -2.30940595e-01 -6.44000769e-01 1.16379833e+00
3.74607354e-01 -1.30396545e+00 7.55102709e-02 -3.42586986e-03
8.79782617e-01 3.22243333e-01 8.56335163e-02 -2.99575329e-01
-4.14569765e-01 -1.04172838e+00 2.05365181e-01 4.96250808e-01
7.88891435e-01 -4.78455812e-01 9.49032724e-01 3.48901451e-01
-1.02793038e+00 -6.04475401e-02 -8.92407596e-02 2.09326014e-01
8.27578455e-02 6.42725766e-01 -1.40761948e+00 3.04015011e-01
7.19974875e-01 5.05198538e-01 -3.01575989e-01 9.83505785e-01
2.29585186e-01 7.78530300e-01 -4.15753901e-01 2.49421820e-01
1.63240328e-01 -3.55455577e-01 7.20310450e-01 1.48685241e+00
6.23472393e-01 1.18891120e-01 1.59631632e-02 5.56103408e-01
3.27700108e-01 2.46103570e-01 -8.64307582e-01 9.85399634e-02
8.60909745e-02 1.12021041e+00 -9.14441943e-01 -4.59583342e-01
-5.30344665e-01 7.84456253e-01 -3.57243508e-01 2.46155322e-01
-7.07994401e-01 1.46716103e-01 6.33252382e-01 4.13679659e-01
7.83385396e-01 8.73990655e-02 -2.54741609e-01 -9.24714565e-01
-1.17042489e-01 -1.17167985e+00 3.40256035e-01 -6.11607671e-01
-1.04733646e+00 7.40102649e-01 -2.42767986e-02 -1.47830188e+00
-4.25215214e-01 -5.58473587e-01 -4.17650968e-01 7.61429012e-01
-1.23741090e+00 -6.83410108e-01 -3.31005663e-01 4.73581135e-01
4.13074732e-01 -3.76808435e-01 1.30413902e+00 3.96361172e-01
-3.97846073e-01 -1.22110747e-01 5.59473813e-01 4.88496087e-02
3.30676764e-01 -8.21706235e-01 -2.86507726e-01 7.23752856e-01
8.54787976e-03 4.99274194e-01 5.91670871e-01 -6.33750021e-01
-1.25831723e+00 -1.10978210e+00 1.01384866e+00 -2.91614056e-01
3.32993001e-01 9.52173898e-04 -7.59987116e-01 2.44636461e-01
2.24600062e-01 -7.59724621e-03 8.24668765e-01 -5.77483237e-01
-1.40551344e-01 -2.19236575e-02 -1.40418804e+00 1.38289377e-01
3.31811070e-01 -4.77618575e-01 -7.47501194e-01 3.51778686e-01
1.00452490e-01 1.39875516e-01 -8.96148086e-01 4.99972939e-01
5.20108819e-01 -1.09534264e+00 1.10827875e+00 -2.43651435e-01
5.00865400e-01 -3.96776229e-01 -4.13305223e-01 -1.03406143e+00
-3.92848879e-01 8.39846507e-02 1.84497967e-01 5.00605643e-01
6.07943423e-02 -6.56852841e-01 7.36979485e-01 -5.63377589e-02
1.97414026e-01 -8.11001182e-01 -1.08178008e+00 -4.43016469e-01
-1.97700784e-01 -2.66282052e-01 2.22975388e-01 1.11199284e+00
-4.66091216e-01 1.06466487e-01 -2.13191152e-01 3.41083467e-01
4.31981862e-01 3.16998988e-01 4.82851863e-01 -1.34216654e+00
-3.54159176e-01 -2.44685799e-01 -6.06412172e-01 -2.83988863e-01
-3.36600989e-01 -8.66629243e-01 -1.89292610e-01 -1.66366458e+00
4.81300354e-01 -5.77652633e-01 -4.48965460e-01 2.27930784e-01
9.33980849e-03 3.48037809e-01 3.99011731e-01 5.39461970e-01
-5.48753105e-02 -4.51377258e-02 9.30672586e-01 -2.12926775e-01
1.41719719e-02 -4.05691415e-02 -2.04025567e-01 7.08779573e-01
8.65564585e-01 -7.03724563e-01 -2.18348742e-01 -1.33861080e-01
7.05431029e-02 1.12597577e-01 4.03881907e-01 -1.45829546e+00
-1.60273880e-01 1.27440944e-01 4.18496042e-01 -7.53305137e-01
3.80136698e-01 -1.38830256e+00 7.55074084e-01 1.11122024e+00
2.38706112e-01 1.83390483e-01 9.07279328e-02 5.32688797e-01
-3.29963118e-01 -2.97087133e-01 8.91956985e-01 -3.06180585e-03
-2.97917038e-01 -4.58054468e-02 -8.80955696e-01 5.74425757e-02
1.18219805e+00 -2.42816389e-01 1.32980019e-01 -9.07889456e-02
-6.48062110e-01 -4.08020198e-01 3.94392878e-01 -1.97358895e-02
6.08768225e-01 -1.01130986e+00 -9.27136123e-01 4.42736208e-01
-5.60222864e-02 -1.15522183e-01 4.03510153e-01 1.32202733e+00
-9.87843812e-01 6.62464142e-01 -4.01293129e-01 -1.00995016e+00
-1.44750464e+00 8.43965948e-01 1.29856616e-01 -5.68242610e-01
-7.72501588e-01 3.51731777e-01 3.82496804e-01 4.50833961e-02
3.55501994e-02 -4.39719945e-01 -5.48788786e-01 3.87025177e-01
4.80590910e-01 5.09823740e-01 2.29272112e-01 -9.53909934e-01
-6.00915313e-01 8.65739942e-01 3.06836545e-01 1.49550319e-01
1.54805958e+00 2.34897941e-01 -2.68003345e-01 4.35727805e-01
1.45930290e+00 1.44988105e-01 -6.78684652e-01 1.76217094e-01
-1.23364106e-01 -1.78285807e-01 -1.53387398e-01 -6.36173844e-01
-1.10755050e+00 8.60456526e-01 1.13790655e+00 1.76834852e-01
1.28214014e+00 -2.51295775e-01 7.57358611e-01 2.29027554e-01
2.74957925e-01 -5.14208674e-01 -2.81520486e-01 2.81025976e-01
8.89699280e-01 -1.10041237e+00 1.29307270e-01 -1.56938300e-01
-5.40521443e-01 8.84165525e-01 -3.89207691e-01 -3.57327163e-01
9.19343054e-01 2.43790329e-01 2.66037762e-01 -3.54752213e-01
-3.94918323e-01 8.74957591e-02 1.84271693e-01 6.60007298e-01
3.34177613e-01 1.89969659e-01 -3.38258207e-01 2.83257097e-01
9.58445519e-02 -6.96865469e-02 1.20021120e-01 9.35829997e-01
-4.31010336e-01 -8.11238348e-01 -8.86551082e-01 7.05060899e-01
-5.28305352e-01 -4.11479399e-02 -1.19019493e-01 6.40482187e-01
4.45471853e-01 7.87702680e-01 6.17706478e-02 -1.25835389e-01
6.08773790e-02 2.68251032e-01 2.55551159e-01 -6.89016104e-01
-8.28118324e-01 1.08557373e-01 -1.83213949e-01 -4.55948174e-01
-7.13322580e-01 -9.88358080e-01 -1.31634164e+00 8.82956088e-02
1.69337049e-01 1.22808523e-01 6.64144695e-01 6.78962171e-01
3.59814644e-01 4.95908916e-01 6.18706942e-01 -6.65979981e-01
-2.86952406e-01 -7.26588070e-01 -5.75167775e-01 5.35197556e-01
4.67109889e-01 -6.31782472e-01 -4.83554929e-01 3.37788552e-01]
|
[15.55225944519043, -1.6935322284698486]
|
70999a78-0d3a-4de7-b20e-853f0ca88da5
|
instantaneous-physiological-estimation-using
|
2202.12368
| null |
https://arxiv.org/abs/2202.12368v1
|
https://arxiv.org/pdf/2202.12368v1.pdf
|
Instantaneous Physiological Estimation using Video Transformers
|
Video-based physiological signal estimation has been limited primarily to predicting episodic scores in windowed intervals. While these intermittent values are useful, they provide an incomplete picture of patients' physiological status and may lead to late detection of critical conditions. We propose a video Transformer for estimating instantaneous heart rate and respiration rate from face videos. Physiological signals are typically confounded by alignment errors in space and time. To overcome this, we formulated the loss in the frequency domain. We evaluated the method on the large scale Vision-for-Vitals (V4V) benchmark. It outperformed both shallow and deep learning based methods for instantaneous respiration rate estimation. In the case of heart-rate estimation, it achieved an instantaneous-MAE of 13.0 beats-per-minute.
|
['Laszlo A. Jeni', 'Conrad S. Tucker', 'Ananyananda Dasari', 'Ambareesh Revanur']
|
2022-02-24
| null | null | null | null |
['heart-rate-estimation']
|
['medical']
|
[ 1.01934627e-01 -1.58251271e-01 -8.36609378e-02 -5.37978888e-01
-6.68524683e-01 -1.11648612e-01 1.33821729e-03 7.82772526e-02
-2.76459336e-01 9.07930255e-01 2.99050122e-01 2.08757833e-01
1.90916210e-01 -7.17472583e-02 -1.46410376e-01 -7.73066342e-01
-4.28526551e-01 -1.76715747e-01 -3.36172253e-01 2.97653824e-01
-4.54542525e-02 3.01333874e-01 -1.05092359e+00 1.51221350e-01
4.76636082e-01 1.48284924e+00 -1.82145476e-01 9.29932714e-01
3.23959142e-01 1.12918234e+00 -8.53429377e-01 1.00809457e-02
1.05608758e-02 -7.49415874e-01 -3.80987644e-01 -2.57923901e-01
5.18669009e-01 -6.93354845e-01 -7.72589922e-01 4.92426813e-01
9.81437802e-01 -1.31030858e-01 2.50190109e-01 -1.37899125e+00
-4.84385639e-02 2.20261857e-01 -2.96516210e-01 8.91161442e-01
5.75561643e-01 3.98281783e-01 5.38259268e-01 -6.56873465e-01
2.20891416e-01 8.20866168e-01 1.26710463e+00 7.18651772e-01
-9.84269679e-01 -5.92916071e-01 -3.21502835e-01 5.06194949e-01
-1.39180255e+00 -8.37530196e-01 6.58564091e-01 -3.82065386e-01
1.06007695e+00 2.65727639e-01 1.08124590e+00 1.33433533e+00
3.52261364e-01 5.11804581e-01 8.34629595e-01 1.22164212e-01
1.53317023e-02 -1.45443097e-01 -4.09161523e-02 6.04490042e-01
-1.48825988e-01 1.03612885e-01 -1.00808895e+00 -1.04805954e-01
8.47215116e-01 8.14532638e-02 -8.14035118e-01 3.30916375e-01
-1.37931228e+00 4.29212809e-01 2.38638163e-01 2.40888685e-01
-7.00042248e-01 5.36888957e-01 6.52220130e-01 2.95631915e-01
5.05669773e-01 3.10782462e-01 -6.12828970e-01 -7.29434848e-01
-1.26082492e+00 1.20303724e-02 8.07546973e-01 5.65672338e-01
3.41522247e-02 3.58547449e-01 -8.28148425e-01 6.24128163e-01
1.25624202e-02 4.40360934e-01 6.61112845e-01 -1.19337869e+00
5.81272356e-02 -1.32342011e-01 2.15012774e-01 -9.94802535e-01
-9.11951661e-01 -3.19934338e-01 -9.46274281e-01 -5.35088539e-01
4.32288677e-01 -3.25923502e-01 -6.87845528e-01 1.69112921e+00
3.52486700e-01 8.10550332e-01 -1.45589888e-01 1.15722942e+00
1.04156876e+00 3.76338482e-01 9.54610407e-02 -8.41877222e-01
1.32072914e+00 -6.09962881e-01 -1.25290728e+00 -2.23925799e-01
1.65348992e-01 -3.80308747e-01 6.70908272e-01 3.28992665e-01
-1.29564834e+00 -6.33069456e-01 -8.29259634e-01 2.00218126e-01
3.47874194e-01 4.06212034e-03 4.82632339e-01 6.18732572e-01
-1.12912893e+00 9.70843613e-01 -1.04851830e+00 -1.95551246e-01
5.61293542e-01 7.85976648e-02 -1.32226467e-01 1.53455377e-01
-1.45623207e+00 7.84582138e-01 -6.34203702e-02 4.89218444e-01
-1.03360462e+00 -1.21583641e+00 -7.43192911e-01 1.49091855e-01
-2.97455825e-02 -7.53090322e-01 1.30729866e+00 -5.00957489e-01
-1.56567717e+00 7.16629446e-01 -2.87733704e-01 -7.39077270e-01
8.23360562e-01 -4.48774725e-01 -4.99535531e-01 6.93503618e-01
-3.79374146e-01 5.36955893e-01 1.09946179e+00 -2.23890543e-01
1.36177251e-02 -2.12954342e-01 -4.40289348e-01 1.87805012e-01
-2.87789911e-01 -7.08714500e-02 -2.78017342e-01 -4.72703755e-01
7.18726888e-02 -8.39827418e-01 1.55412540e-01 3.73349249e-01
9.10357758e-02 2.29652479e-01 6.41151309e-01 -9.97263849e-01
1.18917525e+00 -1.99214244e+00 -2.58795947e-01 -3.72830302e-01
5.35565972e-01 3.25888515e-01 1.10131294e-01 2.15766300e-02
-1.46037444e-01 -1.04782350e-01 -2.90885866e-02 -2.53747165e-01
-4.44816232e-01 -4.76286188e-02 -3.89969945e-02 9.45340157e-01
1.10721573e-01 9.14364934e-01 -9.44440484e-01 -5.56155443e-01
4.95916158e-01 1.02539515e+00 -4.23312068e-01 5.33832371e-01
3.40775162e-01 8.36949527e-01 4.37536128e-02 7.80992866e-01
3.91025126e-01 -1.42849222e-01 -7.27329031e-02 -5.13938844e-01
2.70345300e-01 2.50870556e-01 -5.98125041e-01 1.78108144e+00
-3.74576896e-01 1.11518061e+00 -2.64816713e-02 -9.16028202e-01
7.40822077e-01 9.72580075e-01 1.36074042e+00 -7.08938658e-01
1.83842555e-01 -1.78315103e-01 4.88002300e-02 -1.02139676e+00
-6.64312914e-02 -5.01432776e-01 3.15342277e-01 -4.97546792e-03
-7.91015550e-02 -1.11195840e-01 -1.98380515e-01 -1.56188577e-01
1.43041778e+00 3.92776802e-02 2.99785137e-01 -1.23708723e-02
2.67870098e-01 -5.56693375e-01 7.63837159e-01 7.24594116e-01
-1.00912333e+00 9.66142297e-01 4.75947142e-01 -4.91663873e-01
-5.83554208e-01 -1.10219181e+00 -3.86410594e-01 3.82761717e-01
-2.31877968e-01 -6.33844554e-01 -6.70293629e-01 -3.65778714e-01
-2.84754846e-04 3.21563512e-01 -5.52644312e-01 -4.65687752e-01
-4.59418744e-01 -8.36408675e-01 9.39686656e-01 7.52060831e-01
3.60929817e-01 -1.17224658e+00 -1.32720733e+00 5.43398678e-01
-9.03227866e-01 -1.53006160e+00 -6.30968153e-01 2.02411171e-02
-9.39444661e-01 -1.19119263e+00 -8.93337011e-01 -6.64917454e-02
3.47187277e-03 1.36509864e-02 1.29257834e+00 -5.97394556e-02
-9.12968993e-01 7.66092896e-01 -1.14138842e-01 -6.26990438e-01
1.23073207e-03 -3.69480789e-01 2.70016074e-01 -8.50046054e-03
3.86696279e-01 -6.46442056e-01 -1.20352411e+00 1.85580119e-01
-3.32612723e-01 -2.61820316e-01 -1.73756890e-02 7.10456014e-01
4.63754326e-01 -5.70140660e-01 8.14909041e-01 -3.01989168e-01
6.02837861e-01 -5.34028113e-01 -3.04701626e-01 -2.00793952e-01
-5.96589625e-01 -3.85382086e-01 5.18444836e-01 -5.77308297e-01
-5.48956096e-01 1.51724480e-02 -7.35203996e-02 -9.93072033e-01
6.79997206e-02 1.63917333e-01 4.45378453e-01 9.55900475e-02
6.37897611e-01 2.09611088e-01 1.98517725e-01 -1.87867388e-01
-2.48836786e-01 5.36853433e-01 9.62365150e-01 -1.33079141e-01
6.95643350e-02 3.39433342e-01 1.91728130e-01 -1.06325650e+00
-8.71266842e-01 -6.01638675e-01 -4.12592769e-01 -6.04203403e-01
9.79631364e-01 -1.29410720e+00 -1.03718328e+00 6.34063244e-01
-9.89700198e-01 -3.41704130e-01 -3.34398627e-01 8.10419023e-01
-6.86016679e-01 3.44269037e-01 -7.90392458e-01 -8.82309139e-01
-8.40974867e-01 -5.70822179e-01 8.93253326e-01 1.66652605e-01
-5.06085038e-01 -9.78474200e-01 5.57867177e-02 2.32065558e-01
7.69043446e-01 7.44948983e-01 1.41905412e-01 -8.04422423e-02
1.00644350e-01 -2.70820379e-01 2.71116812e-02 4.21512127e-01
3.14613879e-01 -1.71100080e-01 -1.38091207e+00 -2.80453891e-01
5.73428929e-01 -3.87802869e-01 7.45950103e-01 7.99549222e-01
1.61474693e+00 -1.98536009e-01 9.45219249e-02 9.66223240e-01
1.09474611e+00 1.37254506e-01 7.60623634e-01 -2.71069914e-01
7.46067286e-01 3.10258687e-01 4.96800423e-01 1.05337894e+00
2.02529445e-01 6.51469350e-01 4.10830140e-01 -1.50230557e-01
-9.85747650e-02 2.51296371e-01 5.22455931e-01 7.65585005e-01
-2.47812748e-01 -5.09616099e-02 -8.00626516e-01 4.64038730e-01
-1.58277929e+00 -1.19639993e+00 -2.60410786e-01 2.25182509e+00
9.69087660e-01 -1.72178745e-01 2.46012956e-01 2.13127434e-01
7.03952789e-01 2.92578697e-01 -8.83057415e-01 -3.04664522e-01
1.53324023e-01 2.19879150e-01 4.35720772e-01 2.04321384e-01
-1.00149226e+00 4.94033903e-01 7.35792685e+00 -7.54243433e-02
-1.40502083e+00 2.28238851e-01 6.35008216e-01 -7.74996340e-01
3.72391164e-01 -5.17700911e-01 -4.46680576e-01 6.86278462e-01
1.59349227e+00 -1.23037342e-02 4.23662335e-01 5.10360301e-01
6.71208560e-01 -1.47853404e-01 -1.37489045e+00 1.72066140e+00
9.28853825e-02 -9.54598129e-01 -8.10157657e-01 -2.38506168e-01
1.12979017e-01 2.39736989e-01 -1.71834871e-01 1.47921324e-01
-8.71419609e-01 -1.21202135e+00 4.30937976e-01 6.93538070e-01
1.25457072e+00 -3.46833915e-01 5.53987980e-01 6.84381090e-03
-1.03588474e+00 -4.95430231e-02 -4.09684151e-01 -1.04586240e-02
5.04640825e-02 1.00295770e+00 -8.65297079e-01 -8.65436792e-02
7.29897499e-01 1.03422081e+00 -1.87183797e-01 1.15578413e+00
-2.09925678e-02 7.64616370e-01 -4.80653137e-01 4.47911084e-01
-1.13325059e-01 2.15004221e-01 5.21670341e-01 1.29317760e+00
4.08008695e-01 5.03241777e-01 -3.23230699e-02 6.38722897e-01
-6.57669976e-02 -2.88060158e-01 -5.47547579e-01 5.05612083e-02
3.37024361e-01 1.41692066e+00 -2.83706576e-01 -4.10796911e-01
-4.00696427e-01 1.04126799e+00 -5.26635088e-02 2.39819735e-01
-1.30572593e+00 -2.69232512e-01 9.21351850e-01 1.35862887e-01
1.05515430e-02 3.69698182e-02 -6.35052621e-02 -1.32361257e+00
1.12305745e-01 -6.93632245e-01 3.31530184e-01 -8.22442651e-01
-8.28973413e-01 3.02883953e-01 -1.76512197e-01 -1.13619173e+00
-4.18561131e-01 -1.81412876e-01 -6.65111244e-01 7.65767395e-01
-1.63577759e+00 -1.58572704e-01 -9.34560180e-01 8.31487656e-01
5.15866876e-01 3.40293646e-01 9.99170899e-01 6.95260108e-01
-6.90879941e-01 7.22519219e-01 -4.50960666e-01 -4.56200987e-02
9.80688691e-01 -1.18975747e+00 2.49802917e-01 5.32915175e-01
-1.16650932e-01 2.24921390e-01 8.47430825e-01 -2.23827049e-01
-1.48836744e+00 -1.09365523e+00 7.03205407e-01 -5.51542819e-01
2.56793469e-01 -1.33525997e-01 -9.95211422e-01 2.61256069e-01
-1.32704377e-01 8.56001794e-01 7.21449912e-01 -3.27001810e-01
2.60735508e-02 -4.78359014e-01 -1.35261810e+00 9.70834284e-04
7.68244982e-01 -7.66802669e-01 -3.15400511e-01 3.76583368e-01
3.47492695e-01 -7.16752112e-01 -1.43731713e+00 3.97604614e-01
8.75853598e-01 -9.86145437e-01 1.01478553e+00 -4.68159050e-01
2.71809369e-01 1.79388359e-01 3.23442400e-01 -1.25714958e+00
-1.90303624e-01 -1.05807412e+00 -6.96801305e-01 8.57517958e-01
-2.48689488e-01 -4.38876033e-01 7.18391180e-01 7.12987304e-01
-1.13921441e-01 -7.74584472e-01 -9.10584509e-01 -5.32898247e-01
-4.71975029e-01 -5.35115302e-01 1.38976067e-01 9.22756255e-01
1.67211652e-01 1.60333082e-01 -8.08402002e-01 -6.81297481e-02
6.01317763e-01 -2.45539755e-01 1.31699622e-01 -1.10202110e+00
-1.55399173e-01 -1.14700586e-01 -4.57974881e-01 -4.58271533e-01
2.33602766e-02 -5.13805389e-01 2.63210773e-01 -1.34435642e+00
1.23717543e-02 1.26113668e-01 -9.74230587e-01 5.21833181e-01
-2.95456052e-01 5.13331592e-01 9.28484648e-03 1.42877856e-02
-3.08561414e-01 4.53275710e-01 9.74832356e-01 8.89354348e-02
-4.03435528e-01 -7.31876865e-02 -2.55334437e-01 6.13794506e-01
9.66394842e-01 -5.44938684e-01 -4.76290107e-01 -6.14795201e-02
-1.53406844e-01 8.67826223e-01 4.80339736e-01 -1.36160934e+00
-4.98367287e-02 9.31266621e-02 8.15908134e-01 -4.38353002e-01
7.26631582e-01 -5.65259755e-01 4.67273146e-02 7.51310170e-01
-3.97385538e-01 1.01454973e-01 3.08396429e-01 4.85783428e-01
-1.95553213e-01 3.87816399e-01 1.00730276e+00 -2.57021368e-01
-3.48259330e-01 6.03828549e-01 -3.97925496e-01 4.82910544e-01
7.91076481e-01 -3.04150730e-01 -2.05536962e-01 -5.75189769e-01
-7.50316560e-01 2.17433438e-01 -1.30796894e-01 3.31632346e-01
8.86244237e-01 -1.20284128e+00 -8.59846473e-01 1.64602935e-01
2.21581589e-02 -4.26322311e-01 4.03483182e-01 1.50073171e+00
-5.32354057e-01 4.00120437e-01 -4.09722716e-01 -8.82883966e-01
-1.30050278e+00 3.48099947e-01 8.57952178e-01 1.29080936e-01
-1.03479826e+00 8.37367296e-01 -1.12995245e-01 5.25439143e-01
6.33597791e-01 -4.95456219e-01 -2.28803322e-01 2.93384761e-01
8.51252496e-01 4.81857687e-01 2.10852817e-01 -4.13179368e-01
-6.77580416e-01 3.67561787e-01 3.15551251e-01 1.43797413e-01
1.10660470e+00 -3.62787306e-01 1.27900377e-01 5.80816627e-01
1.34502697e+00 -5.48084676e-01 -1.47839606e+00 8.60082209e-02
-2.68199801e-01 -4.14947510e-01 4.15504307e-01 -7.07880020e-01
-1.48840642e+00 1.20544922e+00 1.20011926e+00 -2.48148497e-02
1.36362028e+00 -6.54665828e-01 1.06433082e+00 2.56475091e-01
-1.27350599e-01 -9.82702851e-01 2.14430630e-01 1.26467630e-01
5.90025127e-01 -1.26123190e+00 8.15155655e-02 -8.91315937e-02
-6.25632524e-01 1.36408973e+00 5.02315521e-01 2.85178304e-01
6.12165451e-01 2.28870958e-01 2.74642944e-01 -1.21711254e-01
-1.06260502e+00 1.08349144e-01 1.92189053e-01 6.74429536e-01
8.71342599e-01 -1.77062035e-01 -1.30238190e-01 4.93886992e-02
3.19818109e-02 4.64289129e-01 7.37333834e-01 6.41655684e-01
-1.09918214e-01 -2.34862104e-01 -2.28992805e-01 5.55960894e-01
-1.13129151e+00 -1.86432749e-01 4.90661711e-02 3.27482790e-01
-3.14953387e-01 1.14637828e+00 2.05479264e-01 -6.77443221e-02
1.24096878e-01 4.05744940e-01 7.86655545e-01 -3.10564756e-01
-7.34189332e-01 1.25370353e-01 3.57725397e-02 -1.14699841e+00
-4.85014886e-01 -6.42476320e-01 -1.12335873e+00 -1.07396737e-01
8.15571100e-02 -9.69764888e-02 7.18255758e-01 6.26927674e-01
2.58667827e-01 9.47020471e-01 6.98671579e-01 -7.80457497e-01
-5.42841077e-01 -9.46741521e-01 -6.33142471e-01 3.47417831e-01
9.48186576e-01 -2.93002456e-01 -4.96420890e-01 1.06965490e-01]
|
[13.88534164428711, 2.84633469581604]
|
a351fbdf-5c06-4ec5-af05-b7b73a3a5aff
|
neuri-diversifying-dnn-generation-via
|
2302.02261
| null |
https://arxiv.org/abs/2302.02261v1
|
https://arxiv.org/pdf/2302.02261v1.pdf
|
NeuRI: Diversifying DNN Generation via Inductive Rule Inference
|
Deep Learning (DL) is prevalently used in various industries to improve decision-making and automate processes, driven by the ever-evolving DL libraries and compilers. The correctness of DL systems is crucial for trust in DL applications. As such, the recent wave of research has been studying the automated synthesis of test-cases (i.e., DNN models and their inputs) for fuzzing DL systems. However, existing model generators only subsume a limited number of operators, for lacking the ability to pervasively model operator constraints. To address this challenge, we propose NeuRI, a fully automated approach for generating valid and diverse DL models composed of hundreds of types of operators. NeuRI adopts a three-step process: (i) collecting valid and invalid API traces from various sources; (ii) applying inductive program synthesis over the traces to infer the constraints for constructing valid models; and (iii) performing hybrid model generation by incorporating both symbolic and concrete operators concolically. Our evaluation shows that NeuRI improves branch coverage of TensorFlow and PyTorch by 51% and 15% over the state-of-the-art. Within four months, NeuRI finds 87 new bugs for PyTorch and TensorFlow, with 64 already fixed or confirmed, and 8 high-priority bugs labeled by PyTorch, constituting 10% of all high-priority bugs of the period. Additionally, open-source developers regard error-inducing models reported by us as "high-quality" and "common in practice".
|
['Lingming Zhang', 'Yuyao Wang', 'Jinjun Peng', 'Jiawei Liu']
|
2023-02-04
| null | null | null | null |
['program-synthesis']
|
['computer-code']
|
[-1.86605096e-01 4.60874697e-04 -6.52150869e-01 -1.35768488e-01
-6.98058188e-01 -7.79974997e-01 1.66765288e-01 -1.52367324e-01
3.57172549e-01 6.22205496e-01 -1.59736171e-01 -1.09130490e+00
1.71341479e-01 -9.20643628e-01 -1.10564923e+00 1.70086354e-01
-2.28125840e-01 2.66487122e-01 4.95118529e-01 -1.55740023e-01
3.11974347e-01 8.96299034e-02 -1.77080846e+00 5.81742287e-01
1.14413106e+00 6.64827466e-01 -2.60556847e-01 7.18427598e-01
-3.60616684e-01 9.96712387e-01 -7.59109855e-01 -7.18443930e-01
1.71480373e-01 -3.65851194e-01 -8.53858113e-01 -3.72626185e-01
1.27313673e-01 -3.16838682e-01 1.84322029e-01 1.45275259e+00
1.35875165e-01 -7.47879922e-01 -1.60355404e-01 -1.69801700e+00
-1.82000607e-01 1.31985664e+00 -5.91028392e-01 -8.76950473e-02
3.01467061e-01 6.89078748e-01 1.13320327e+00 -5.28353691e-01
6.96074247e-01 1.05988646e+00 6.57403827e-01 5.59883833e-01
-1.35280764e+00 -7.18789518e-01 -2.32966423e-01 -9.31421891e-02
-1.42463398e+00 -3.31892133e-01 4.89086717e-01 -8.47341597e-01
1.63593638e+00 2.83132106e-01 5.71278572e-01 1.36151493e+00
4.94446188e-01 5.04629076e-01 9.31931973e-01 -6.20272934e-01
4.48204964e-01 1.42264947e-01 5.48522770e-01 8.96554172e-01
8.31473053e-01 2.91268319e-01 -2.78641999e-01 -6.83734953e-01
3.82593662e-01 -2.94055104e-01 -6.69677779e-02 1.53108448e-01
-1.16242421e+00 6.07462168e-01 -2.45645016e-01 3.61043453e-01
-6.44029900e-02 5.23978710e-01 7.40812480e-01 3.65123183e-01
-3.67482863e-02 8.07445228e-01 -8.54315400e-01 -8.42329204e-01
-9.85500693e-01 4.40148115e-01 1.30558693e+00 1.27719724e+00
8.85043859e-01 3.95136625e-01 -6.86766356e-02 2.78843433e-01
3.63961816e-01 4.10485089e-01 3.62392694e-01 -7.50855982e-01
5.69784522e-01 1.21194363e+00 -1.25970602e-01 -6.01454556e-01
-1.37006894e-01 -5.94580650e-01 -2.03488827e-01 4.82516676e-01
3.24708968e-01 -4.52355057e-01 -5.73886633e-01 1.53114271e+00
8.08217376e-03 2.83757150e-01 -1.08837180e-01 4.47862536e-01
3.01311016e-01 4.22149062e-01 -5.20501852e-01 1.59448534e-01
1.10153008e+00 -7.52852678e-01 -2.88410574e-01 -3.52451891e-01
1.15552878e+00 -6.11431539e-01 1.27214861e+00 9.26135719e-01
-1.08800745e+00 -3.87393951e-01 -1.36189771e+00 4.76462454e-01
-3.24976332e-02 3.92738581e-02 9.39405918e-01 8.09517324e-01
-1.00744605e+00 5.47035813e-01 -1.02875340e+00 2.90473282e-01
2.78116107e-01 4.08498168e-01 -7.46677741e-02 -4.53913882e-02
-7.83735871e-01 5.00711083e-01 3.34656268e-01 1.57468673e-02
-1.25149345e+00 -1.20430481e+00 -8.90479803e-01 1.27461061e-01
7.05044925e-01 -4.33623582e-01 1.53275394e+00 -9.18840826e-01
-1.25800395e+00 3.93935829e-01 -5.83525337e-02 -4.60902870e-01
2.69793570e-01 -2.17761263e-01 -6.89622581e-01 -6.28872871e-01
6.43786937e-02 1.40346676e-01 3.57287943e-01 -1.04735994e+00
-7.99201369e-01 4.40370627e-02 5.34312904e-01 -9.05866683e-01
-9.27090943e-02 2.31673941e-01 -3.27933937e-01 -9.45845619e-02
-4.65021133e-01 -9.54079032e-01 -1.11230351e-01 -4.35283184e-01
-7.29310870e-01 -8.77231732e-03 6.12435520e-01 -6.12823009e-01
1.54934204e+00 -2.09973407e+00 9.07888561e-02 3.58833492e-01
4.98374313e-01 2.39531934e-01 -3.75307873e-02 2.67746925e-01
-1.07819371e-01 7.10652649e-01 -1.32644624e-01 6.90376386e-02
5.60988545e-01 2.84813255e-01 -4.79447752e-01 1.31715447e-01
4.43023294e-01 8.97294104e-01 -1.09544313e+00 -2.89292395e-01
-1.15525022e-01 4.95074950e-02 -1.05660415e+00 1.11424208e-01
-1.00240862e+00 -2.15957183e-02 -2.26046115e-01 1.16551387e+00
4.68163103e-01 -9.28573161e-02 1.81162819e-01 -1.04621015e-02
-2.85416156e-01 6.45198107e-01 -1.31438851e+00 1.54879618e+00
-5.54543018e-01 6.35649741e-01 -5.13557643e-02 -2.45383650e-01
8.06416512e-01 2.51096755e-01 3.10774874e-02 -5.10404348e-01
5.94263314e-04 6.12167001e-01 4.03325558e-01 -4.73198056e-01
3.20366114e-01 2.90817857e-01 -3.76380622e-01 6.97806716e-01
6.25399174e-03 3.34715024e-02 5.09811819e-01 -2.63083708e-02
1.62731957e+00 3.62719059e-01 1.04371034e-01 -1.26503244e-01
4.24530923e-01 2.60326356e-01 9.68781233e-01 5.93867004e-01
1.99921310e-01 1.62230685e-01 1.47082686e+00 -3.92004400e-01
-9.18250322e-01 -7.03287721e-01 -1.63619090e-02 5.97584128e-01
-4.81796563e-01 -1.03847444e+00 -1.04950559e+00 -9.58894968e-01
-9.36959833e-02 1.15848422e+00 -4.26359355e-01 -4.02662516e-01
-6.34637177e-01 -4.11102057e-01 1.14811254e+00 6.47309840e-01
2.27137640e-01 -1.01072526e+00 -6.72732830e-01 2.05906078e-01
3.86088490e-02 -7.76826322e-01 -2.97712207e-01 3.64421844e-01
-5.69233537e-01 -1.32699978e+00 2.83137351e-01 -2.62815714e-01
3.94593060e-01 -4.06769872e-01 1.57471633e+00 4.32452053e-01
-4.38769281e-01 -2.01434046e-01 -3.37023199e-01 -1.20973930e-01
-1.02454162e+00 3.92236374e-02 -3.09866071e-01 -4.19451237e-01
4.74120438e-01 -5.75756013e-01 1.45744056e-01 2.34771699e-01
-9.42112684e-01 -1.91422179e-02 6.16933286e-01 7.62251496e-01
3.27115268e-01 2.84341842e-01 3.76895070e-02 -1.09568059e+00
6.45220041e-01 -5.11932552e-01 -1.33339846e+00 2.04056740e-01
-8.45149457e-01 3.22536618e-01 6.63540840e-01 -5.07332504e-01
-7.54214406e-01 -3.39134157e-01 -3.13128293e-01 -3.94667506e-01
-5.35521507e-02 1.01878238e+00 -4.05008495e-01 -1.75652698e-01
1.02933538e+00 -1.60594329e-01 -1.66330725e-01 -1.77740112e-01
6.55284747e-02 3.25341880e-01 3.78807843e-01 -1.32250190e+00
7.60093808e-01 -2.68154263e-01 -1.92522839e-01 -1.69113412e-01
-1.02498680e-01 2.35879689e-01 -2.49412462e-01 -8.02978221e-03
4.08776879e-01 -5.04209936e-01 -8.88205349e-01 4.49226886e-01
-1.28984749e+00 -7.70999849e-01 -2.95271337e-01 -9.11853183e-03
-7.64258131e-02 1.99656308e-01 -6.99869037e-01 -4.84371334e-01
-2.84818619e-01 -2.00090432e+00 9.42685306e-01 -1.27171963e-01
-6.94194257e-01 -6.86103344e-01 1.89309865e-01 2.15769326e-03
4.82395709e-01 3.30298454e-01 1.41379452e+00 -5.61884820e-01
-9.25904810e-01 -2.42618740e-01 -1.49990376e-02 6.89171493e-01
-7.40193501e-02 6.84744060e-01 -8.00075710e-01 -7.35868588e-02
-1.46818250e-01 -1.20632425e-02 -1.25636846e-01 4.61364537e-02
1.01566887e+00 -3.71705264e-01 -3.12114865e-01 6.22682273e-01
1.55626833e+00 2.51598209e-01 7.21977949e-01 2.09618598e-01
6.49330080e-01 1.54821262e-01 2.79655665e-01 5.70218682e-01
1.92059830e-01 5.65245152e-01 7.41259217e-01 2.59192765e-01
3.11398916e-02 -1.55663863e-01 7.51752019e-01 5.38493276e-01
2.13409454e-01 2.16638505e-01 -1.50039911e+00 4.87387747e-01
-1.52949750e+00 -7.11577177e-01 -5.67304015e-01 2.25532913e+00
1.26475549e+00 7.12195814e-01 -6.64037094e-03 3.63124430e-01
2.83804327e-01 -5.49389243e-01 -4.56937134e-01 -5.01770318e-01
3.53045613e-01 7.04406083e-01 4.35105711e-01 5.55387259e-01
-4.49359626e-01 8.95948470e-01 5.66143370e+00 5.75736403e-01
-1.47498357e+00 8.16807523e-03 1.69400230e-01 1.88957036e-01
-7.09435701e-01 5.74229121e-01 -1.21295893e+00 4.56907988e-01
1.36189842e+00 -4.93651897e-01 6.30221665e-01 1.34963405e+00
-4.55026254e-02 -7.73271322e-02 -1.55186093e+00 4.07149136e-01
-2.13615194e-01 -1.60499287e+00 -2.84415126e-01 2.56917566e-01
7.96473086e-01 9.27711725e-02 -2.21789017e-01 7.31744170e-01
6.46182477e-01 -1.04739571e+00 1.17055225e+00 4.30309862e-01
7.30433822e-01 -9.18949544e-01 9.79570389e-01 2.73024976e-01
-8.57375562e-01 -1.00068012e-02 5.66254295e-02 -2.78149843e-01
-1.61114201e-01 9.14403617e-01 -9.57298100e-01 4.36411649e-01
7.25342035e-01 5.12334406e-01 -9.16393518e-01 1.13149261e+00
-4.79626119e-01 1.09110284e+00 -1.68567792e-01 -8.99256766e-02
2.38778512e-03 1.67748719e-01 2.84994096e-01 1.28037858e+00
2.76620477e-01 -7.91942835e-01 9.40286070e-02 1.67616582e+00
2.22783256e-02 -4.70467925e-01 -4.38485265e-01 -3.35887372e-01
6.68171942e-01 1.08223701e+00 -4.64962572e-01 -2.06796497e-01
-6.18699133e-01 3.19299549e-01 -2.29574032e-02 2.56712466e-01
-1.21337521e+00 -4.77777779e-01 8.56705546e-01 2.04608947e-01
1.28656656e-01 -2.54106432e-01 -7.84415424e-01 -1.08877254e+00
4.00509447e-01 -1.74169326e+00 -1.03098422e-01 -5.23185968e-01
-8.84352982e-01 8.20338070e-01 2.18955025e-01 -9.02051032e-01
-6.10144019e-01 -4.98939306e-01 -5.76760411e-01 1.07249439e+00
-1.12258542e+00 -8.81413043e-01 -1.57032415e-01 1.52276456e-01
5.00161238e-02 -1.09721124e-01 7.86955476e-01 5.46324193e-01
-7.42720604e-01 9.66501355e-01 -6.10742569e-01 -9.24762338e-02
3.97299916e-01 -1.20307362e+00 8.11303616e-01 1.28164601e+00
-2.81470090e-01 1.11490393e+00 6.91148281e-01 -8.37975740e-01
-1.97190523e+00 -1.14731085e+00 8.72667611e-01 -3.31455231e-01
1.15735257e+00 -6.38935864e-01 -9.35087621e-01 1.01704288e+00
9.32187773e-03 1.70336422e-02 4.10934240e-01 1.87833324e-01
-6.94239140e-01 -2.31742248e-01 -1.00957835e+00 7.83456504e-01
7.51953423e-01 -6.24015152e-01 -8.35375339e-02 2.89179772e-01
1.10644352e+00 -7.33214855e-01 -1.01849687e+00 3.21150452e-01
3.13413948e-01 -1.22183859e+00 6.11779988e-01 -2.78578639e-01
7.15234458e-01 -5.56423008e-01 -2.95723438e-01 -9.25722361e-01
1.02153063e-01 -9.11487639e-01 -4.41415280e-01 1.47908235e+00
7.42789924e-01 -7.73538828e-01 7.26016164e-01 7.44584680e-01
-6.99650764e-01 -8.01316023e-01 -4.40647066e-01 -7.97588766e-01
-5.77590615e-02 -1.07168257e+00 1.04921305e+00 8.83228481e-01
1.44265115e-01 1.07362419e-01 -1.53001922e-03 1.45694003e-01
2.12438986e-01 -9.00549442e-02 1.00766909e+00 -1.07929456e+00
-1.07883382e+00 -7.76325285e-01 -1.73490703e-01 -3.72071832e-01
3.80642623e-01 -1.00382960e+00 -4.38862555e-02 -9.35547054e-01
2.71968786e-02 -5.31154990e-01 2.64377177e-01 1.06563389e+00
1.66343540e-01 -2.54015952e-01 -2.18838692e-01 -9.37344879e-02
-1.37951300e-01 -1.82013854e-01 6.85600817e-01 -2.35471308e-01
-1.49077877e-01 -4.04163823e-02 -7.92175770e-01 6.43348277e-01
4.22948360e-01 -4.90767300e-01 -3.90032202e-01 -4.11592454e-01
1.03602815e+00 8.37830454e-02 4.76027846e-01 -1.27294958e+00
1.48974210e-02 -1.48569539e-01 -4.51908946e-01 -4.48051810e-01
-4.24235612e-01 -6.64149284e-01 6.04233980e-01 6.59316003e-01
-6.01968691e-02 4.28161919e-01 6.67270958e-01 -2.59510148e-02
-3.46160978e-01 -7.22591162e-01 3.19494903e-01 -4.13854606e-02
-8.15801144e-01 8.91943648e-02 -3.58663261e-01 3.33150566e-01
9.30211544e-01 4.39459346e-02 -6.23538375e-01 2.75303185e-01
-2.50980467e-01 -5.21894731e-02 7.41332293e-01 3.77665609e-01
2.88953662e-01 -9.57277834e-01 -4.60025370e-01 6.70536637e-01
2.62713969e-01 3.24771367e-02 -2.95244187e-01 7.57692575e-01
-8.21882427e-01 2.32498124e-01 -3.99806490e-03 -6.01289034e-01
-8.30978513e-01 5.65915704e-01 3.49911720e-01 -4.85896081e-01
-4.27737743e-01 7.45736063e-01 -2.98682809e-01 -5.62779963e-01
1.51437178e-01 -1.14967060e+00 6.87483430e-01 -5.61996877e-01
4.72054988e-01 2.45548233e-01 5.11844456e-01 1.71321809e-01
-4.93012458e-01 -4.29076850e-02 6.07293956e-02 -8.99243429e-02
1.21775687e+00 9.24528182e-01 -7.89596617e-01 5.31505525e-01
9.45077240e-01 2.19531462e-01 -8.35339665e-01 7.85692409e-02
3.55756015e-01 -1.18016750e-01 4.64433506e-02 -1.07947648e+00
-9.30957556e-01 8.89758766e-01 -7.94312544e-03 1.04222171e-01
9.54364061e-01 -1.72941208e-01 7.09815979e-01 2.19364479e-01
9.43100631e-01 -4.20733988e-01 -1.96885198e-01 5.66216290e-01
5.32615304e-01 -5.72806060e-01 -3.39887738e-01 -3.91272515e-01
-5.42529263e-02 1.03427708e+00 8.81974876e-01 2.31883042e-02
3.38817716e-01 1.15325499e+00 -3.84817630e-01 -3.64289023e-02
-1.10162711e+00 3.64752859e-01 4.89615239e-02 4.78578627e-01
6.64204657e-01 5.39349914e-02 -2.36903027e-01 1.07440364e+00
-2.85804182e-01 4.74244535e-01 8.33101809e-01 1.23739362e+00
7.12176505e-03 -1.56631887e+00 -4.10316080e-01 5.48286319e-01
-4.66609776e-01 -3.08604985e-01 -1.85353592e-01 1.05684876e+00
4.39112097e-01 7.41314173e-01 -4.04333949e-01 -8.63526762e-01
3.85340452e-01 1.23252451e-01 4.90503758e-01 -8.78632188e-01
-9.72815335e-01 -1.96123704e-01 6.49899483e-01 -8.66547704e-01
5.46530485e-01 -6.62073433e-01 -1.21126950e+00 -5.82555115e-01
-3.76706332e-01 5.98442070e-02 5.96100867e-01 9.47372317e-01
5.56233346e-01 9.06451881e-01 1.26212031e-01 -5.02474785e-01
-7.02826619e-01 -6.58300281e-01 -1.73812717e-01 -9.93878990e-02
6.76756054e-02 -6.50683761e-01 -3.39242339e-01 1.03537679e-01]
|
[7.691824436187744, 7.644417762756348]
|
46ff50dd-ca09-4b4f-bc2e-8bd39a159b37
|
compositional-probabilistic-and-causal
|
2304.08278
| null |
https://arxiv.org/abs/2304.08278v1
|
https://arxiv.org/pdf/2304.08278v1.pdf
|
Compositional Probabilistic and Causal Inference using Tractable Circuit Models
|
Probabilistic circuits (PCs) are a class of tractable probabilistic models, which admit efficient inference routines depending on their structural properties. In this paper, we introduce md-vtrees, a novel structural formulation of (marginal) determinism in structured decomposable PCs, which generalizes previously proposed classes such as probabilistic sentential decision diagrams. Crucially, we show how mdvtrees can be used to derive tractability conditions and efficient algorithms for advanced inference queries expressed as arbitrary compositions of basic probabilistic operations, such as marginalization, multiplication and reciprocals, in a sound and generalizable manner. In particular, we derive the first polytime algorithms for causal inference queries such as backdoor adjustment on PCs. As a practical instantiation of the framework, we propose MDNets, a novel PC architecture using md-vtrees, and empirically demonstrate their application to causal inference.
|
['Marta Kwiatkowska', 'Benjie Wang']
|
2023-04-17
| null | null | null | null |
['causal-inference', 'causal-inference']
|
['knowledge-base', 'miscellaneous']
|
[ 4.37313378e-01 5.51621556e-01 -4.34281468e-01 -3.68806124e-01
-7.92640269e-01 -1.11358154e+00 6.58412695e-01 -5.81467040e-02
3.20810169e-01 9.05732870e-01 1.63823292e-01 -1.13633704e+00
-7.42317140e-01 -1.33136392e+00 -1.07923388e+00 -8.17396104e-01
-4.33777511e-01 9.46442425e-01 4.05361027e-01 1.93633899e-01
1.94773898e-01 7.30720758e-01 -1.32173443e+00 2.42349312e-01
5.99434257e-01 5.58254480e-01 -3.29744577e-01 9.48806167e-01
1.16706207e-01 7.38701940e-01 -2.44349524e-01 -6.69888139e-01
-6.79227710e-02 -2.14548960e-01 -9.88894939e-01 -6.09003186e-01
4.13464844e-01 -5.42504311e-01 -8.41059506e-01 1.05854583e+00
9.46225300e-02 -1.00893371e-01 7.19379663e-01 -1.59268332e+00
-3.44666064e-01 1.83384240e+00 -2.12751418e-01 1.30556345e-01
5.34486890e-01 -1.12980217e-01 1.70735586e+00 -3.38767618e-01
5.30124307e-01 1.90942013e+00 3.14475626e-01 2.54782706e-01
-1.85582614e+00 -5.24615109e-01 1.21673740e-01 5.31892538e-01
-1.33780539e+00 -3.18347663e-01 2.86885470e-01 -1.50593966e-01
8.27338159e-01 8.37519944e-01 2.09205851e-01 1.11715090e+00
3.20511580e-01 9.10512745e-01 1.14340007e+00 -4.97864872e-01
3.94280463e-01 -2.83473432e-01 4.73728955e-01 7.99343586e-01
7.56075263e-01 4.74198610e-01 -5.46101272e-01 -7.53960907e-01
5.63136816e-01 -1.81107149e-01 -4.37633209e-02 -3.43033940e-01
-1.12999201e+00 9.51749325e-01 -8.23099464e-02 -5.14979102e-02
3.72667879e-01 1.05937827e+00 4.44115438e-02 1.54460669e-01
-3.74262989e-01 -1.83579940e-02 -3.25387895e-01 2.90585011e-02
-5.46420872e-01 6.31558299e-01 1.38990390e+00 1.14373195e+00
7.04541445e-01 -3.50346148e-01 -8.36790562e-01 -1.11081451e-02
4.76604939e-01 1.02786362e+00 -5.42475820e-01 -1.10544944e+00
3.08824420e-01 1.51734337e-01 -4.65069264e-02 -9.08642173e-01
-5.37558310e-02 -2.18519002e-01 -5.32801926e-01 -4.10231471e-01
4.11899060e-01 1.32571533e-01 -7.42126226e-01 1.95532441e+00
3.39674532e-01 1.76107526e-01 -1.03232898e-01 3.05223554e-01
2.11849019e-01 7.02765167e-01 -1.71569139e-01 -5.45288585e-02
1.41251087e+00 -2.04843789e-01 -5.87727606e-01 3.04227322e-01
3.80100310e-01 -3.20720464e-01 7.55892158e-01 6.30629063e-01
-8.08851659e-01 1.56411469e-01 -1.02434742e+00 -2.20174000e-01
-1.16332427e-01 -5.44785559e-02 1.38903117e+00 1.24753797e+00
-1.00869191e+00 6.16427898e-01 -1.05753934e+00 3.88707034e-02
3.60110998e-01 5.17173827e-01 2.27101333e-02 -3.03517431e-01
-1.15260732e+00 2.43790865e-01 6.14271164e-01 1.67688653e-01
-1.71733987e+00 -6.67338312e-01 -6.39043987e-01 3.78380388e-01
8.38217080e-01 -8.41927052e-01 1.39106667e+00 3.42199057e-01
-1.52379966e+00 3.89629513e-01 -4.70768183e-01 -6.21405602e-01
3.48740816e-01 8.08541924e-02 -3.15579802e-01 3.58765423e-02
-1.91030338e-01 -7.66711310e-03 5.78152776e-01 -8.29312384e-01
-7.67012596e-01 -5.80645263e-01 7.28599548e-01 -5.28765976e-01
9.71144512e-02 -6.70951083e-02 -1.99533671e-01 -1.80457145e-01
2.47263029e-01 -1.06556320e+00 -5.85258603e-01 -3.25019449e-01
-1.45255220e+00 -4.23077792e-01 4.65865374e-01 9.35401171e-02
1.29079747e+00 -1.77871788e+00 2.35612631e-01 9.27894175e-01
4.30418104e-01 -5.43885887e-01 2.63432264e-01 5.08184195e-01
2.25631475e-01 3.18124503e-01 -2.84699619e-01 1.50179729e-01
7.59900808e-01 5.94966114e-01 -9.65255976e-01 3.53438020e-01
1.23263165e-01 1.07371640e+00 -9.80232120e-01 -4.87946957e-01
-4.72842809e-03 -1.02306224e-01 -7.90806830e-01 -1.42498508e-01
-8.26651573e-01 -6.23424910e-02 -6.19654834e-01 6.62859440e-01
8.20464015e-01 -2.09096000e-01 7.71569550e-01 5.99813350e-02
1.06053568e-01 7.32394755e-01 -1.52178502e+00 1.30896401e+00
-9.80828777e-02 4.39226568e-01 -3.37070614e-01 -6.31863475e-01
3.00354391e-01 6.93280026e-02 -2.52943367e-01 1.21097609e-01
6.58460632e-02 5.99266551e-02 -1.92618921e-01 4.20311987e-02
5.76449037e-01 -5.69763221e-03 -6.54875636e-01 7.33471990e-01
4.00413685e-02 -1.10407479e-01 3.17739010e-01 7.35623598e-01
1.53565824e+00 -1.22765556e-01 3.60795297e-02 -3.04544181e-01
3.61758053e-01 -9.57634971e-02 4.85675305e-01 1.46277308e+00
4.07726586e-01 5.91832586e-03 1.27947700e+00 1.36978418e-01
-4.97030884e-01 -1.99168968e+00 -1.49037525e-01 9.57387209e-01
-9.33014322e-03 -8.65873456e-01 -4.85490739e-01 -5.18853843e-01
1.50875390e-01 1.00084376e+00 -5.12136757e-01 -1.39541179e-01
-1.44625828e-01 -8.75223875e-01 1.03032410e+00 4.96834874e-01
6.94965869e-02 -2.44234905e-01 -1.37844265e-01 2.22113311e-01
3.92840430e-02 -1.06605363e+00 -8.29556957e-02 4.68116969e-01
-1.05252767e+00 -1.22943628e+00 3.73831600e-01 -1.98425865e-03
4.99793440e-01 9.09684822e-02 8.30837369e-01 -5.60161233e-01
-2.09602460e-01 1.99640945e-01 3.37048233e-01 -1.28018618e-01
-5.21695256e-01 5.67420386e-02 -4.23336439e-02 -2.27982134e-01
1.85498402e-01 -9.59011018e-01 -2.39042476e-01 1.54147372e-01
-9.50169027e-01 -1.86831519e-01 5.23432791e-01 6.18295193e-01
8.31285596e-01 4.81720328e-01 -3.55270207e-01 -1.34732676e+00
3.67474586e-01 -3.84525687e-01 -1.37673521e+00 4.63945568e-01
-4.24080014e-01 8.19854915e-01 2.77735800e-01 -1.86481535e-01
-1.15128672e+00 2.62879450e-02 2.07372636e-01 -1.61355436e-01
1.01763271e-01 6.08092070e-01 -5.47699630e-01 1.69005990e-01
5.66558897e-01 6.22045361e-02 -5.99687696e-01 -5.89857474e-02
9.17719841e-01 1.64395377e-01 8.94347787e-01 -1.43741882e+00
1.15394723e+00 8.89509976e-01 7.67426670e-01 -1.50305241e-01
-7.78519630e-01 1.15649655e-01 -1.55984908e-01 3.00960094e-01
2.87035316e-01 -6.21848643e-01 -1.42391467e+00 7.58658722e-02
-1.11971498e+00 -3.32836360e-01 -1.75596297e-01 3.54760379e-01
-5.14655530e-01 2.62983859e-01 -6.50445104e-01 -1.18635201e+00
3.78841069e-03 -9.47144687e-01 9.33062851e-01 -5.26171327e-02
-2.49203101e-01 -7.70016551e-01 1.49408683e-01 -6.09479360e-02
-1.80185124e-01 3.55393261e-01 1.49643314e+00 -5.26348233e-01
-1.44941354e+00 -3.39916497e-02 -1.68906048e-01 -2.06923604e-01
-3.68161887e-01 6.64384604e-01 -1.10342658e+00 1.71954229e-01
-5.30438721e-01 7.24332407e-02 1.08845532e+00 2.53359050e-01
1.49746013e+00 -6.58868730e-01 -1.00039589e+00 4.74176586e-01
1.48494554e+00 -1.27794430e-01 8.12902033e-01 -3.81735295e-01
5.60985506e-01 1.29641935e-01 2.52063364e-01 5.67289114e-01
4.17502075e-01 3.71817619e-01 3.30061466e-01 7.40708709e-01
2.37135291e-01 -9.61799860e-01 3.71156067e-01 3.82985145e-01
-1.31751448e-01 -2.50764966e-01 -4.88931388e-01 3.85875195e-01
-1.78041351e+00 -1.22735512e+00 -4.89217848e-01 2.25181365e+00
9.34756398e-01 2.22545817e-01 -9.63436291e-02 3.50731909e-01
7.47043490e-01 -1.48186982e-01 -3.33130538e-01 -5.98092496e-01
-1.89818457e-01 8.45598936e-01 9.76419806e-01 5.90813279e-01
-8.19584310e-01 7.94793129e-01 7.42893791e+00 1.22925246e+00
-3.57046276e-01 4.03577477e-01 5.16674936e-01 -8.23863521e-02
-1.30633867e+00 7.35655546e-01 -1.05702734e+00 1.72511518e-01
1.15526247e+00 -1.49953559e-01 6.54663265e-01 9.25679624e-01
-1.44663125e-01 -1.56009689e-01 -1.82969499e+00 6.24876678e-01
-7.56896019e-01 -1.53087950e+00 8.36454704e-02 2.21551925e-01
7.64736652e-01 -3.91376823e-01 1.76146179e-01 1.73271880e-01
1.32969546e+00 -1.03215969e+00 8.60137045e-01 1.85465708e-01
7.53678679e-01 -9.74154592e-01 2.40763500e-01 -2.04299599e-01
-9.00674641e-01 -3.49267602e-01 -4.02125001e-01 -2.05692984e-02
2.26767045e-02 1.10934889e+00 -7.25524366e-01 8.07161570e-01
4.48606998e-01 8.50634575e-02 2.48697121e-02 7.52526164e-01
-1.04364705e+00 9.72021818e-01 -8.94727588e-01 -3.16098213e-01
-4.63018157e-02 1.67434290e-02 6.15882277e-01 1.10947180e+00
3.65954459e-01 1.48766682e-01 -4.21733797e-01 1.45965135e+00
-1.96433797e-01 -7.79750168e-01 -4.98659164e-01 -2.65297174e-01
9.33647811e-01 1.05422962e+00 -5.99289954e-01 -3.02633435e-01
2.06421852e-01 5.40622413e-01 1.79450378e-01 3.12272042e-01
-1.19349360e+00 -4.42781717e-01 8.45600963e-01 -1.94549561e-01
5.90529263e-01 -1.93222106e-01 -3.57300341e-01 -1.09481931e+00
-1.02768168e-01 -5.54874718e-01 8.22183669e-01 -3.80354613e-01
-1.04113138e+00 -1.46406382e-01 3.89786988e-01 -3.57881188e-01
-2.28274137e-01 -6.32552564e-01 -6.42185569e-01 4.66698050e-01
-1.20191908e+00 -8.64565909e-01 2.41062105e-01 7.31493294e-01
-5.73792100e-01 5.71463704e-01 9.00984645e-01 -1.25892073e-01
-6.58549607e-01 8.45793247e-01 1.36484712e-01 -3.33496302e-01
5.82529306e-02 -1.52368569e+00 3.49710763e-01 1.19130063e+00
3.53654236e-01 1.01641607e+00 9.65314388e-01 -4.29032207e-01
-2.26246810e+00 -1.05801904e+00 8.63052130e-01 -5.04837275e-01
9.97352183e-01 -6.63458347e-01 -7.11747780e-02 1.09741974e+00
-3.90917391e-01 -2.84701049e-01 5.60713649e-01 7.17472494e-01
-8.30004573e-01 -4.61523980e-01 -1.09354734e+00 9.17578042e-01
1.42933857e+00 -8.63595784e-01 -5.80946922e-01 4.60130095e-01
1.05688858e+00 -4.81537133e-01 -8.38973939e-01 2.76688367e-01
7.17673540e-01 -6.90735936e-01 9.84462202e-01 -5.45532823e-01
3.76020581e-01 -6.27536297e-01 -4.32999313e-01 -6.08286679e-01
-2.72160798e-01 -1.34309471e+00 -7.77056038e-01 1.04753757e+00
5.81114233e-01 -9.94414210e-01 6.48749530e-01 5.52917302e-01
7.18696937e-02 -3.39594007e-01 -1.16760147e+00 -7.32069552e-01
-1.63461894e-01 -1.17296970e+00 1.06538689e+00 4.81104940e-01
2.43244365e-01 1.57315671e-01 -3.97685776e-03 8.40133607e-01
9.83168304e-01 6.24916911e-01 7.02588022e-01 -1.15513825e+00
-9.26198244e-01 -3.65600467e-01 -5.08970380e-01 -1.08198154e+00
1.07992999e-01 -1.00321639e+00 9.11690295e-03 -1.23360956e+00
5.72483659e-01 -7.36510992e-01 -2.21517310e-01 5.67985773e-01
-8.05602781e-03 -2.22928345e-01 -2.84647077e-01 -3.35365474e-01
-4.67671126e-01 3.41682166e-01 7.88936377e-01 -3.30701977e-01
1.27303764e-01 2.88150907e-01 -9.23547804e-01 2.67168552e-01
5.79339981e-01 -7.61893213e-01 -6.52419269e-01 -7.44283125e-02
8.41157079e-01 2.27119535e-01 6.90392792e-01 -7.69213915e-01
4.18322891e-01 -4.03969735e-01 -6.78621000e-03 -9.82243419e-01
9.20039713e-02 -3.57455730e-01 7.06748009e-01 6.91618443e-01
-4.55537915e-01 -3.40730518e-01 2.60085836e-02 9.79842544e-01
2.98768997e-01 -5.35839498e-02 2.68328525e-02 2.79758066e-01
-2.36183718e-01 4.29371834e-01 -6.49271786e-01 -1.09369621e-01
8.04746747e-01 2.01391414e-01 -5.75085521e-01 -1.48787796e-01
-6.32183075e-01 1.51769161e-01 1.26546295e-02 -2.50191450e-01
5.28169394e-01 -1.11781383e+00 -4.28176224e-01 -3.39937717e-01
2.48297937e-02 1.38884559e-01 2.39120841e-01 9.38098907e-01
-3.43994111e-01 9.76296723e-01 3.80895019e-01 -5.60489118e-01
-8.46948802e-01 5.19416869e-01 -1.78053956e-02 -4.79988545e-01
-1.16542377e-01 8.92661452e-01 2.72168279e-01 -4.14373308e-01
1.86751232e-01 -1.05321169e+00 6.99206293e-01 -5.52366734e-01
6.40531301e-01 6.14971459e-01 -9.54219475e-02 4.98086303e-01
-4.68002677e-01 -1.04222648e-01 1.21577024e-01 -4.82268751e-01
8.14345062e-01 4.98269126e-02 -7.57070482e-01 3.82255167e-01
6.66825771e-01 2.38719344e-01 -8.48908186e-01 -2.65597224e-01
1.14855729e-01 -3.02993745e-01 -2.25061804e-01 -5.46757758e-01
-8.00274551e-01 7.73675084e-01 6.20596297e-02 7.77938291e-02
8.58751655e-01 4.72647697e-01 4.46500838e-01 9.51664209e-01
8.46309781e-01 -5.33311129e-01 -7.63916016e-01 3.71061385e-01
1.99940458e-01 -2.36296743e-01 7.24038556e-02 -7.84180582e-01
3.10556322e-01 1.15336764e+00 -1.15830593e-01 7.81715736e-02
5.73147357e-01 5.26217759e-01 -1.11537957e+00 -1.17574900e-01
-1.22961605e+00 -9.18360725e-02 -1.95276573e-01 5.24071038e-01
-2.22389683e-01 8.47680569e-01 -1.52188674e-01 8.41769695e-01
-4.84689981e-01 8.46021697e-02 6.18561208e-01 6.75839722e-01
-1.06217213e-01 -1.33808267e+00 -4.18036014e-01 4.25392121e-01
-5.47044992e-01 -3.74007285e-01 -1.05457723e-01 5.99139988e-01
-1.54597387e-01 1.05014324e+00 -9.28009227e-02 -4.47967887e-01
-1.93741307e-01 -3.97400886e-01 1.03672302e+00 -4.74805444e-01
6.38335422e-02 -4.68314737e-01 3.98881823e-01 -6.57301605e-01
1.16959615e-02 -6.89004719e-01 -1.12819815e+00 -9.79937851e-01
-3.18418503e-01 1.19636402e-01 5.06373823e-01 9.44759369e-01
2.91379094e-01 4.21135485e-01 5.74343085e-01 -7.29366392e-02
-9.05905604e-01 -5.17053127e-01 -7.68942237e-01 -5.65165579e-01
-8.69617909e-02 -5.80399156e-01 -3.99910003e-01 -2.14758858e-01]
|
[7.488500118255615, 5.176295280456543]
|
77887e05-923f-461b-80f2-efe4267221d1
|
evaluating-performance-of-an-adult
|
2005.08766
| null |
https://arxiv.org/abs/2005.08766v1
|
https://arxiv.org/pdf/2005.08766v1.pdf
|
Evaluating Performance of an Adult Pornography Classifier for Child Sexual Abuse Detection
|
The information technology revolution has facilitated reaching pornographic material for everyone, including minors who are the most vulnerable in case they were abused. Accuracy and time performance are features desired by forensic tools oriented to child sexual abuse detection, whose main components may rely on image or video classifiers. In this paper, we identify which are the hardware and software requirements that may affect the performance of a forensic tool. We evaluated the adult porn classifier proposed by Yahoo, based on Deep Learning, into two different OS and four Hardware configurations, with two and four different CPU and GPU, respectively. The classification speed on Ubuntu Operating System is $~5$ and $~2$ times faster than on Windows 10, when a CPU and GPU are used, respectively. We demonstrate the superiority of a GPU-based machine rather than a CPU-based one, being $7$ to $8$ times faster. Finally, we prove that the upward and downward interpolation process conducted while resizing the input images do not influence the performance of the selected prediction model.
|
['Javier Velasco-Mata', 'Francisco Jañez-Martino', 'Roberto A. Vasco-Carofilis', 'Mhd Wesam Al-Nabki', 'Eduardo Fidalgo']
|
2020-05-18
| null | null | null | null |
['abuse-detection']
|
['natural-language-processing']
|
[-1.68933511e-01 -2.46300638e-01 -2.87450552e-01 -2.52815306e-01
-2.68723428e-01 -4.87882465e-01 -3.48795392e-02 7.08798110e-01
-7.38314211e-01 3.07673633e-01 -4.29269940e-01 -5.34860551e-01
-2.29900554e-02 -8.54559481e-01 -5.48541665e-01 -5.17512441e-01
-3.41740549e-01 2.30458498e-01 -7.58250477e-03 3.49158645e-01
5.74567795e-01 8.03180814e-01 -1.78206289e+00 4.17015046e-01
5.12949049e-01 1.11777031e+00 -4.85030204e-01 8.78121197e-01
-1.38681039e-01 7.18314052e-01 -5.14425635e-01 -1.04575670e+00
3.41729075e-01 1.73076257e-01 -5.93615711e-01 -3.90247822e-01
4.70417321e-01 -8.34797502e-01 -2.67473102e-01 9.79341865e-01
4.97948349e-01 -2.48886392e-01 5.19143879e-01 -1.31234598e+00
-5.42063639e-02 4.90732521e-01 -8.37784946e-01 6.71537340e-01
4.41114396e-01 7.98257664e-02 3.84081751e-01 -4.45980251e-01
4.32589769e-01 1.02293253e+00 8.10001254e-01 4.84147578e-01
-9.26012278e-01 -1.07257152e+00 -6.64573252e-01 5.97900867e-01
-1.36449170e+00 -4.62982535e-01 7.23730028e-01 -4.78365868e-01
8.43405485e-01 2.69723505e-01 7.14327633e-01 1.11743724e+00
3.06468219e-01 5.28278589e-01 8.76105785e-01 -5.28338552e-01
2.70026594e-01 2.19255716e-01 5.94118536e-01 7.33255267e-01
7.87760675e-01 -1.33235469e-01 -5.38136005e-01 -7.64377534e-01
2.95908272e-01 1.02671748e-02 1.42565563e-01 8.49739313e-02
-1.32942617e-01 1.11206400e+00 -2.76487261e-01 3.27160060e-01
-2.56742954e-01 5.16905934e-02 9.69179034e-01 3.34545255e-01
4.21455413e-01 2.57291168e-01 -1.38921469e-01 -7.27633655e-01
-1.14502227e+00 9.53243077e-02 8.55741501e-01 7.11967528e-01
4.98916954e-01 -9.06547680e-02 3.54614198e-01 4.36960548e-01
-3.89601439e-02 2.40316227e-01 5.33904612e-01 -8.87009859e-01
2.94469446e-01 4.70218539e-01 -1.86594442e-01 -1.37049496e+00
-4.10806239e-01 7.99156204e-02 -5.19240081e-01 -7.14728143e-03
5.20008147e-01 -6.02745749e-02 -4.00891840e-01 1.09439135e+00
4.40370262e-01 2.21028537e-01 -4.11565274e-01 3.18914145e-01
8.17573667e-01 3.45239550e-01 2.14382663e-01 -1.23239145e-01
1.45878410e+00 -5.34676552e-01 -4.16146815e-01 3.32436487e-02
9.19760287e-01 -7.43860662e-01 7.06895292e-01 7.95478463e-01
-1.17402232e+00 -2.90348232e-01 -1.12766206e+00 -5.99849932e-02
-3.66927594e-01 7.25771859e-02 8.94868255e-01 1.55865288e+00
-7.29290128e-01 1.04350495e+00 -1.19297814e+00 -2.44078398e-01
5.11161387e-01 5.17518044e-01 -4.36086446e-01 1.05811343e-01
-7.11209118e-01 7.73423672e-01 1.02355883e-01 -3.21199149e-01
-4.70845103e-01 -6.35377347e-01 -6.71587646e-01 3.36425126e-01
-2.85566241e-01 1.70440048e-01 8.69290531e-01 -7.24471271e-01
-9.56430733e-01 1.18277919e+00 9.81062129e-02 -5.97962976e-01
5.41486979e-01 -2.66018808e-01 -3.62205565e-01 6.52868509e-01
4.19926271e-02 1.62468046e-01 8.58017385e-01 -6.86360657e-01
-5.37663102e-01 -7.74061918e-01 2.81155221e-02 -2.44326398e-01
-7.66456664e-01 5.90212703e-01 -3.18266928e-01 -6.34890273e-02
-1.96033299e-01 -7.95191228e-01 6.87675998e-02 1.28553033e-01
-1.30268604e-01 -4.28555859e-03 7.61025906e-01 -9.50488091e-01
1.24655151e+00 -2.32388806e+00 -5.61381161e-01 3.02605957e-01
2.32565731e-01 5.39136946e-01 3.53850812e-01 1.74359024e-01
-1.10291801e-01 1.36779904e-01 3.18908840e-01 -3.18031907e-01
-3.11448067e-01 -1.17739983e-01 6.18390776e-02 9.76742625e-01
-3.69918257e-01 -1.42714446e-02 -5.34324884e-01 -7.74515450e-01
6.29781261e-02 5.95883608e-01 -5.53910434e-01 2.68535286e-01
6.63586199e-01 6.00922704e-02 -3.30697089e-01 5.89999914e-01
1.22002685e+00 2.30569363e-01 1.53362110e-01 8.92977864e-02
-5.31066321e-02 1.64753228e-01 -9.01069701e-01 1.25932431e+00
-4.89097059e-01 7.49448001e-01 1.81301281e-01 -9.61630404e-01
1.03962672e+00 1.65069208e-01 4.15119678e-01 -8.40585589e-01
5.22319734e-01 3.05678993e-01 -1.05663417e-02 -7.66426325e-01
6.18596435e-01 1.92338705e-01 3.62362802e-01 6.45201206e-01
-2.11358592e-02 5.27323365e-01 3.44549596e-01 1.84797235e-02
1.39117432e+00 -8.93189609e-02 4.00782339e-02 -2.97433883e-01
1.71862558e-01 -1.16914578e-01 4.69621599e-01 6.26616657e-01
-3.48969728e-01 4.43060488e-01 8.60818088e-01 -7.48177648e-01
-1.12282002e+00 -5.60332417e-01 -3.05448920e-01 1.15448403e+00
9.68720540e-02 -2.88282812e-01 -1.19059265e+00 -5.53800821e-01
-9.07334685e-02 7.66289055e-01 -5.09748995e-01 -2.19772086e-01
-7.73847222e-01 -8.47488761e-01 9.52800989e-01 4.15817916e-01
1.98396623e-01 -7.63077080e-01 -1.47519410e+00 7.74846151e-02
2.82681465e-01 -1.11365330e+00 -4.94745746e-02 -6.86082542e-02
-8.78814995e-01 -1.16994286e+00 -4.07196045e-01 -2.45540053e-01
5.40100157e-01 2.14009881e-01 6.66883826e-01 4.84004736e-01
-6.00376070e-01 2.63155073e-01 -5.98464906e-01 -3.15732986e-01
-2.55616546e-01 1.87327877e-01 1.26595661e-01 -3.47076803e-02
8.03952217e-01 -8.31117451e-01 -6.84629083e-01 -2.98275322e-01
-8.32159162e-01 -3.21742147e-03 2.55267411e-01 5.15579998e-01
-2.13240132e-01 -5.12342229e-02 -6.20647110e-02 -1.10143030e+00
4.57250684e-01 -7.84000576e-01 -4.26182926e-01 6.45683482e-02
-6.47029638e-01 -1.35643452e-01 6.21000886e-01 -4.95284259e-01
-6.45993769e-01 -7.41875097e-02 -2.58238703e-01 -7.68096209e-01
-1.19054846e-01 1.40565187e-01 1.27056494e-01 -1.32204741e-01
6.82579994e-01 1.06713213e-01 -4.14781831e-02 -6.71774745e-01
-3.27102154e-01 8.82896364e-01 2.13944554e-01 -5.78493595e-01
3.51569235e-01 4.88522887e-01 -1.40831649e-01 -9.09681380e-01
-1.26824319e-01 -4.02724385e-01 -4.77832973e-01 -1.46378562e-01
7.71155238e-01 -4.35887784e-01 -1.20420694e+00 5.77786267e-01
-1.43870044e+00 1.77995712e-01 4.45841044e-01 4.03022110e-01
-2.56228894e-01 5.58417499e-01 -8.44703317e-01 -1.01734066e+00
-9.13925052e-01 -1.19778752e+00 7.47239888e-01 4.35630739e-01
-1.51876733e-01 -6.13670290e-01 -4.47304882e-02 6.44377351e-01
1.63829803e-01 4.18135256e-01 9.67089474e-01 -8.15996587e-01
2.02679858e-01 -5.59210896e-01 -4.89108801e-01 9.24790129e-02
-4.18351889e-01 3.49794507e-01 -1.07048965e+00 -6.64066970e-01
2.49363780e-01 -7.14835078e-02 4.07181829e-01 4.71515842e-02
1.50701106e+00 -3.59970301e-01 -4.08539385e-01 8.20669711e-01
1.37517798e+00 6.40817165e-01 6.51733160e-01 6.52277589e-01
5.51833332e-01 6.35497034e-01 5.26970685e-01 6.17064416e-01
6.90522715e-02 4.57611948e-01 4.23841715e-01 8.39423835e-02
2.43383676e-01 -1.07592352e-01 2.95377046e-01 3.99789244e-01
-2.18911409e-01 1.10894233e-01 -1.20138896e+00 4.72770125e-01
-1.38255739e+00 -7.17642665e-01 -1.05826192e-01 2.59925175e+00
5.64902127e-01 2.47616112e-01 3.19367141e-01 4.53677565e-01
9.02549326e-01 -2.16806337e-01 -3.13619822e-01 -1.17747557e+00
5.82514524e-01 7.70839512e-01 7.49746442e-01 -4.95047215e-03
-8.83037269e-01 5.10351300e-01 5.78097296e+00 1.12099612e+00
-1.39259064e+00 3.58048141e-01 9.68461514e-01 -2.82175094e-01
2.15429500e-01 -6.55498803e-02 -7.48079479e-01 9.31129932e-01
1.40556550e+00 -7.77361766e-02 1.08251035e-01 1.40673745e+00
6.27414649e-03 -2.62284815e-01 -1.19376743e+00 1.27339125e+00
1.97361782e-01 -1.29324770e+00 -3.00818324e-01 4.06834573e-01
2.06174352e-03 -4.08312321e-01 -5.68946125e-03 7.66432360e-02
-2.59840220e-01 -1.11577070e+00 6.98089480e-01 -9.12356675e-02
7.92285442e-01 -1.18666065e+00 8.49955618e-01 2.37246662e-01
-7.05232084e-01 -6.25866354e-02 -4.05545801e-01 -1.28848702e-01
-4.67925519e-01 2.98183769e-01 -7.42624938e-01 9.36212093e-02
1.20298064e+00 -1.33327901e-01 -6.06542289e-01 8.04417849e-01
3.05948228e-01 7.80508041e-01 -4.36467826e-01 -1.48970142e-01
3.00220437e-02 4.95716967e-02 1.56985760e-01 1.50926018e+00
4.65864599e-01 2.77969539e-01 -3.98150146e-01 3.99194926e-01
1.31312698e-01 3.68907958e-01 -5.96620202e-01 2.09987521e-01
6.30913556e-01 1.31646919e+00 -8.22698534e-01 -1.07289650e-01
-4.26356167e-01 7.73649812e-01 2.44866148e-01 -3.66485208e-01
-1.07487547e+00 -6.83451593e-01 4.82145339e-01 6.54543221e-01
1.31397992e-01 -4.86435294e-02 -5.33410430e-01 -1.09304106e+00
-4.33761738e-02 -9.37755704e-01 4.02225465e-01 -3.51866275e-01
-7.81169057e-01 5.73721528e-01 -1.35638071e-02 -1.14141190e+00
-8.95982534e-02 -7.71238089e-01 -6.81543648e-01 4.24797714e-01
-8.92138064e-01 -9.99306798e-01 -8.70692879e-02 3.62605453e-01
6.85926452e-02 -2.79496551e-01 7.78176129e-01 7.37040818e-01
-7.82302976e-01 1.22275567e+00 1.40767679e-01 4.84354615e-01
3.13849539e-01 -6.44603550e-01 1.31982788e-01 8.52659702e-01
-1.51430652e-01 6.20827496e-01 7.69992590e-01 -3.87867957e-01
-1.62796652e+00 -5.62137902e-01 7.00005352e-01 -2.79897843e-02
3.91652584e-01 -2.59298444e-01 -8.38878095e-01 4.78887111e-01
-3.87844406e-02 -6.38634562e-02 9.92263317e-01 2.21162707e-01
-6.09142721e-01 -1.44833967e-01 -1.77302742e+00 3.30737323e-01
5.76393962e-01 -2.47974187e-01 -7.71642700e-02 1.00110270e-01
1.06145501e-01 -6.33168817e-01 -9.03837919e-01 7.50059262e-02
8.41143191e-01 -1.37592208e+00 8.36455286e-01 -5.20521700e-01
6.13516808e-01 5.34152091e-01 5.20734265e-02 -4.68198508e-01
4.86353086e-03 -2.89354235e-01 -1.46218613e-01 1.17641246e+00
1.50765821e-01 -4.87763971e-01 1.31142330e+00 9.85056221e-01
3.62498283e-01 -8.47319007e-01 -1.30538285e+00 -5.68808317e-01
2.34310225e-01 -6.35865390e-01 7.47976422e-01 9.59576249e-01
1.20898455e-01 -1.20644599e-01 -4.20754254e-01 5.64186051e-02
6.74805641e-01 -7.99494982e-02 7.75660455e-01 -9.25411046e-01
-4.34680551e-01 -3.09294671e-01 -1.00620174e+00 -3.07676941e-01
-5.38542680e-02 -3.84088755e-01 -7.37908840e-01 -4.76703316e-01
4.95884180e-01 -4.09983873e-01 1.12949647e-02 6.28505886e-01
1.57471329e-01 3.66196185e-01 2.89540976e-01 1.27799690e-01
-9.08346027e-02 -1.28077313e-01 3.79546613e-01 -2.66279969e-02
-3.31413224e-02 -4.17939462e-02 -5.10548472e-01 1.11261880e+00
5.58077753e-01 -7.69915760e-01 9.34106186e-02 -6.01466477e-01
1.31400391e-01 3.69612813e-01 -8.68079811e-02 -1.22399592e+00
2.14442313e-01 -1.76543873e-02 3.93555671e-01 -2.76210666e-01
4.84832972e-01 -6.78842425e-01 1.20099664e-01 8.80598783e-01
-4.78556976e-02 1.22414313e-01 3.18133980e-01 -1.59556225e-01
-4.82869409e-02 -8.42497468e-01 1.05159485e+00 3.94321308e-02
-3.23296934e-01 3.14615369e-01 -3.18516821e-01 -5.12758255e-01
1.18736434e+00 -5.03789663e-01 -2.36988962e-01 -3.10510933e-01
-5.87185919e-01 -3.08513314e-01 5.38040638e-01 6.37662858e-02
4.86541450e-01 -9.02431548e-01 -5.20836532e-01 1.43594071e-01
-3.00615758e-01 -7.37184823e-01 3.21663231e-01 6.49783552e-01
-1.43697298e+00 1.19745977e-01 -6.52772844e-01 -3.64021808e-01
-1.98480356e+00 6.42081797e-01 -3.99813205e-02 -2.62322813e-01
-3.91996175e-01 7.64966726e-01 -3.67350191e-01 2.08726972e-02
1.90242052e-01 1.53683826e-01 -2.69970775e-01 1.92262769e-01
8.81177902e-01 1.06692111e+00 4.10330266e-01 -4.61674839e-01
-3.74259740e-01 3.09867501e-01 -3.44106257e-01 -6.38157651e-02
1.45550084e+00 3.35246533e-01 -2.32896045e-01 -3.18591180e-03
1.32325411e+00 1.67647168e-01 -5.35612881e-01 3.33608180e-01
-4.61294763e-02 -9.96033728e-01 4.82214093e-02 -7.06952140e-02
-1.29077590e+00 1.16653800e+00 9.87459719e-01 4.64418352e-01
1.08899844e+00 -4.31593746e-01 9.83548164e-01 1.10725714e-02
6.54878855e-01 -9.44611728e-01 -2.95888871e-01 3.55303660e-02
2.20335439e-01 -1.02732241e+00 4.00326550e-01 -3.66891414e-01
-2.85803646e-01 1.44456041e+00 7.33365655e-01 -3.81070524e-01
5.13467073e-01 4.16346252e-01 -3.27553868e-01 -1.31345630e-01
-4.24477756e-01 4.33880359e-01 -2.76934296e-01 5.42923570e-01
5.10272682e-01 6.22088425e-02 -9.70506966e-01 6.39216483e-01
-4.40491587e-01 -7.53664374e-02 7.01604128e-01 9.37465191e-01
-2.51986086e-01 -1.05254889e+00 -5.80713451e-01 5.80452442e-01
-1.08617198e+00 3.11803054e-02 -9.05166194e-02 6.80235505e-01
6.53948188e-01 8.64210129e-01 4.12393093e-01 -5.30081630e-01
-1.22340657e-01 -1.74437687e-01 5.85671067e-01 -4.65850443e-01
-9.04704750e-01 -4.27509665e-01 1.45112470e-01 -6.36748254e-01
1.50172114e-01 -9.03722763e-01 -9.39150393e-01 -1.22898722e+00
-2.80382633e-01 2.19388306e-01 1.03619528e+00 7.59016097e-01
3.73467177e-01 -4.23047870e-01 7.29388952e-01 -6.70523643e-01
-2.28244498e-01 -8.45418036e-01 -6.09496117e-01 2.00590104e-01
-1.98858425e-01 -6.14148676e-01 -4.23837692e-01 -2.13467315e-01]
|
[12.814279556274414, 1.0264443159103394]
|
96fd10a9-c6c7-47ed-ac21-134f748add22
|
robust-video-background-identification-by
|
1903.02232
| null |
http://arxiv.org/abs/1903.02232v1
|
http://arxiv.org/pdf/1903.02232v1.pdf
|
Robust Video Background Identification by Dominant Rigid Motion Estimation
|
The ability to identify the static background in videos captured by a moving
camera is an important pre-requisite for many video applications (e.g. video
stabilization, stitching, and segmentation). Existing methods usually face
difficulties when the foreground objects occupy a larger area than the
background in the image. Many methods also cannot scale up to handle densely
sampled feature trajectories. In this paper, we propose an efficient
local-to-global method to identify background, based on the assumption that as
long as there is sufficient camera motion, the cumulative background features
will have the largest amount of trajectories. Our motion model at the two-frame
level is based on the epipolar geometry so that there will be no
over-segmentation problem, another issue that plagues the 2D motion
segmentation approach. Foreground objects erroneously labelled due to
intermittent motions are also taken care of by checking their global
consistency with the final estimated background motion. Lastly, by virtue of
its efficiency, our method can deal with densely sampled trajectories. It
outperforms several state-of-the-art motion segmentation methods on public
datasets, both quantitatively and qualitatively.
|
['Nianjuan Jiang', 'Xun Xu', 'Loong Fah Cheong', 'Kaimo Lin', 'Jiangbo Lu']
|
2019-03-06
| null | null | null | null |
['video-stabilization']
|
['computer-vision']
|
[ 1.18823841e-01 -5.19128144e-01 -1.29837483e-01 1.32615164e-01
-4.90173757e-01 -7.87823737e-01 3.44539076e-01 -1.59603000e-01
-4.26293731e-01 5.89579523e-01 -2.54398733e-01 -1.29096776e-01
2.95200020e-01 -5.42102098e-01 -6.57243311e-01 -1.07969189e+00
-5.23612974e-03 3.02418292e-01 9.82392550e-01 1.48073956e-01
2.23131448e-01 5.47764778e-01 -1.37123370e+00 -1.61873206e-01
6.44908488e-01 7.18707979e-01 2.20554218e-01 8.09594333e-01
-2.02878386e-01 8.21849108e-01 -5.43101490e-01 -2.14542165e-01
4.90138590e-01 -5.96910417e-01 -7.93262184e-01 7.72131145e-01
6.98619306e-01 -6.70551240e-01 -3.45370650e-01 1.33504951e+00
1.36886155e-02 3.69516730e-01 3.63244563e-01 -1.34268427e+00
3.06618392e-01 1.49364829e-01 -1.01450300e+00 6.46916747e-01
2.82370150e-01 1.50727138e-01 5.33151865e-01 -6.89596057e-01
9.55549061e-01 1.06689727e+00 5.23762703e-01 5.01320660e-01
-1.04649627e+00 -3.38332385e-01 4.23347086e-01 2.65928954e-02
-1.30159855e+00 -4.55658168e-01 7.07539737e-01 -5.41347444e-01
2.11774290e-01 4.34226334e-01 6.99491978e-01 6.06504321e-01
7.16667250e-02 9.19188678e-01 5.71120143e-01 -2.55395412e-01
1.88093275e-01 -6.75013140e-02 5.25240507e-03 5.87229550e-01
4.35811102e-01 -2.74524868e-01 -1.53604671e-01 -2.50127047e-01
1.01119196e+00 8.65143985e-02 -4.65617001e-01 -5.74636519e-01
-1.37893045e+00 5.56455255e-01 -1.22465849e-01 5.41612685e-01
-2.20392466e-01 1.78870142e-01 3.73097926e-01 -4.33643274e-02
3.98269981e-01 -2.20482461e-02 -2.46612892e-01 -3.32144231e-01
-1.52025497e+00 1.15390606e-01 6.98393166e-01 1.04200935e+00
8.23544800e-01 1.48876049e-02 1.01171784e-01 2.89601177e-01
1.19102716e-01 4.93137509e-01 2.80575126e-01 -1.23269916e+00
2.46352464e-01 4.38680500e-01 3.59214634e-01 -1.39369905e+00
6.73826039e-02 6.42231181e-02 -8.27980757e-01 2.45422609e-02
8.98844242e-01 -1.29541308e-01 -7.33219922e-01 1.43011904e+00
8.81401241e-01 6.43633425e-01 -1.07923485e-01 8.99318039e-01
3.50623429e-01 7.81734169e-01 -2.41130501e-01 -6.58381224e-01
1.08614922e+00 -1.07223594e+00 -8.72348607e-01 -2.23144695e-01
4.21785593e-01 -9.15930271e-01 3.93826842e-01 3.31986248e-01
-1.13545048e+00 -5.12607753e-01 -6.88472748e-01 3.17485631e-01
1.06721528e-01 1.05811648e-01 3.30736697e-01 6.03196084e-01
-9.58850503e-01 5.53578019e-01 -9.86627102e-01 -4.68747556e-01
1.81222379e-01 3.36957395e-01 -3.54402304e-01 -5.77001013e-02
-7.66873240e-01 5.73907733e-01 3.74952465e-01 2.15494514e-01
-7.98873901e-01 -2.61265725e-01 -7.29960084e-01 -2.32208237e-01
7.49045968e-01 -3.56052816e-01 1.06572080e+00 -1.41005480e+00
-1.25659192e+00 7.28556275e-01 -5.72744846e-01 -1.98249340e-01
1.03943229e+00 -2.02745631e-01 -1.12739213e-01 6.21144950e-01
2.03681782e-01 5.23959637e-01 1.09736156e+00 -1.27352846e+00
-9.66893196e-01 -7.13946000e-02 -2.40575060e-01 2.10728675e-01
-2.12853751e-03 1.98435768e-01 -1.10584617e+00 -5.75807810e-01
1.94748685e-01 -1.10727215e+00 -4.97698873e-01 -5.29045016e-02
-2.65166610e-01 1.16641439e-01 1.35515213e+00 -6.33886814e-01
1.39967275e+00 -2.29157424e+00 1.47220805e-01 2.03706354e-01
2.12684602e-01 4.23386276e-01 1.88795522e-01 9.18027759e-02
1.86249480e-01 -3.43375392e-02 -3.84498030e-01 -1.96508080e-01
-4.37220067e-01 6.15221411e-02 -1.35310501e-01 9.66704130e-01
6.03950489e-03 5.85810363e-01 -1.08232069e+00 -8.85196328e-01
4.27522570e-01 3.83043587e-02 -2.39413187e-01 1.14179432e-01
-1.31628528e-01 6.55318618e-01 -4.71910119e-01 5.64013839e-01
9.64379489e-01 -2.14996129e-01 9.83646736e-02 1.61556587e-01
-3.39993089e-01 -3.74665827e-01 -1.61654592e+00 1.36179709e+00
3.41022342e-01 9.07689750e-01 4.27111894e-01 -6.83236241e-01
4.14962173e-01 2.68145144e-01 1.01595402e+00 8.04807991e-02
1.88850477e-01 3.73428881e-01 -1.83948398e-01 -4.65808034e-01
6.09021962e-01 2.69676179e-01 2.94593573e-01 2.42013961e-01
-3.04115474e-01 4.71553281e-02 5.76508701e-01 3.10729176e-01
9.43823814e-01 1.80587202e-01 3.03484559e-01 -2.95293719e-01
6.88283265e-01 2.24375397e-01 9.48455930e-01 6.49020135e-01
-5.99632561e-01 8.09852719e-01 3.69695276e-01 -2.75502920e-01
-1.01525021e+00 -4.74294931e-01 6.19997866e-02 5.44175684e-01
6.70079052e-01 -1.72956020e-01 -1.18610775e+00 -6.62743568e-01
-2.82804996e-01 8.69152993e-02 -3.93650979e-01 2.57644325e-01
-7.92477727e-01 -4.46522832e-01 2.85662442e-01 3.59384030e-01
6.89208388e-01 -7.86587179e-01 -1.00386333e+00 4.62138325e-01
-4.58880305e-01 -1.47203434e+00 -8.88040245e-01 -3.18180054e-01
-9.61413860e-01 -1.25958431e+00 -9.85949755e-01 -7.46114254e-01
8.55487883e-01 1.01215613e+00 8.42160285e-01 3.87394279e-01
-1.67262092e-01 4.97215718e-01 -2.62532711e-01 2.32736349e-01
-4.91168857e-01 -2.30889857e-01 1.53878117e-02 3.00511807e-01
4.03686047e-01 1.53808501e-02 -6.42380357e-01 6.70055747e-01
-1.04339755e+00 -1.76141533e-04 2.19152585e-01 5.15718937e-01
5.86740792e-01 5.87051332e-01 -1.86780654e-02 -6.68591440e-01
-5.15214503e-02 -3.65975410e-01 -8.67236078e-01 2.49444038e-01
1.87274694e-01 -5.02471864e-01 2.54869640e-01 -7.28089452e-01
-9.07289863e-01 4.30164427e-01 2.22936779e-01 -8.28745365e-01
-3.01906168e-01 -8.92099366e-02 -4.03795302e-01 -1.61888480e-01
1.16879627e-01 2.74661183e-01 -2.66816281e-02 -1.00600705e-01
2.41838366e-01 3.62078935e-01 5.51133096e-01 -1.88318044e-01
9.14893270e-01 9.43751752e-01 3.85412835e-02 -1.29496229e+00
-4.94685352e-01 -1.12676966e+00 -9.71942663e-01 -5.25755942e-01
1.12006974e+00 -7.94398189e-01 -4.55814481e-01 7.78633535e-01
-1.23547935e+00 -2.98380256e-01 -2.97449734e-02 4.96646464e-01
-5.67541242e-01 9.85163450e-01 -5.73418796e-01 -1.07592309e+00
5.22488914e-02 -1.33837163e+00 1.05876184e+00 2.47387871e-01
-1.89233914e-01 -1.05984259e+00 -1.46583706e-01 1.21183492e-01
3.85673083e-02 5.67686439e-01 3.44898820e-01 -3.04924518e-01
-9.62040126e-01 -3.37898210e-02 -5.49256094e-02 1.79724187e-01
3.73195618e-01 6.25030577e-01 -7.32545555e-01 -3.46813679e-01
2.60010183e-01 4.36166793e-01 7.52874017e-01 8.43229830e-01
6.78553820e-01 -2.46804267e-01 -5.37647426e-01 4.37798738e-01
1.30686152e+00 3.67721468e-01 6.75940633e-01 3.96226346e-01
9.39769804e-01 6.76049411e-01 1.02699995e+00 2.44332612e-01
-3.87177132e-02 5.80698252e-01 3.43339503e-01 -7.49866217e-02
5.42414049e-03 1.76400438e-01 5.02207279e-01 7.24464118e-01
3.98773663e-02 -3.37056339e-01 -8.10483813e-01 7.33787000e-01
-2.07988405e+00 -1.27413678e+00 -6.64850056e-01 2.44335556e+00
5.59642017e-01 1.11041352e-01 3.76036912e-01 1.40790150e-01
1.24292207e+00 8.38705450e-02 -4.01915699e-01 1.08138308e-01
-1.94661081e-01 -6.08240485e-01 6.98561072e-01 5.58071434e-01
-1.53701973e+00 1.07067347e+00 5.79065228e+00 8.96618128e-01
-1.04054105e+00 -9.41320658e-02 7.15216875e-01 1.76040545e-01
8.94903243e-02 1.26925036e-01 -1.03452945e+00 7.39705563e-01
5.28404891e-01 6.87875301e-02 1.20701857e-01 7.45544136e-01
3.80812854e-01 -7.16365993e-01 -9.00825381e-01 9.52893078e-01
2.86172740e-02 -1.28191364e+00 -2.24562615e-01 1.85013637e-01
9.47761655e-01 -3.00947189e-01 -1.63098514e-01 -3.02038848e-01
-1.40391842e-01 -5.36247134e-01 8.91757786e-01 2.38347977e-01
5.62059939e-01 -6.35002077e-01 6.71912968e-01 4.55665201e-01
-1.40822315e+00 1.28060728e-01 -3.04475635e-01 2.61478543e-01
5.42734981e-01 7.02770054e-01 -4.53782290e-01 4.88460630e-01
6.36006713e-01 7.96564758e-01 -4.29966837e-01 1.27512348e+00
8.45314339e-02 4.89117503e-01 -4.94279593e-01 2.15072498e-01
4.22582179e-01 -6.28142774e-01 8.51102769e-01 1.21175969e+00
3.10807288e-01 1.59990907e-01 4.20987129e-01 4.85039771e-01
1.81029797e-01 3.74246985e-02 -6.44936800e-01 -6.30998239e-02
3.11550558e-01 1.12267661e+00 -1.45478606e+00 -5.44507742e-01
-5.63159525e-01 1.20109832e+00 -3.72758180e-01 3.98011327e-01
-9.27928209e-01 -2.06652116e-02 4.94536549e-01 1.61390007e-01
5.28971434e-01 -5.12029529e-01 1.78230003e-01 -1.27179253e+00
4.71570082e-02 -9.17275012e-01 1.86216325e-01 -3.49769384e-01
-9.69907343e-01 5.13001263e-01 1.41200302e-02 -1.53529346e+00
-2.14143708e-01 -2.94655561e-01 -6.54921412e-01 3.65322888e-01
-1.19022036e+00 -8.29972327e-01 -3.95252198e-01 6.82057559e-01
8.36126328e-01 3.39310855e-01 1.80494830e-01 4.15808320e-01
-6.97874606e-01 -7.04473397e-03 2.76326030e-01 3.53352129e-01
6.39538050e-01 -1.01091599e+00 2.39035845e-01 1.56530547e+00
7.97079280e-02 4.54212308e-01 7.70384729e-01 -9.36123967e-01
-1.23143125e+00 -1.13871503e+00 5.13188064e-01 -5.11907399e-01
6.79688275e-01 -1.51135311e-01 -1.04222393e+00 5.57086587e-01
5.43002151e-02 2.66353995e-01 3.89279932e-01 -7.06600726e-01
4.29659635e-01 2.07581207e-01 -1.01632512e+00 6.17543817e-01
8.17053914e-01 3.75450514e-02 -3.08307469e-01 1.24818213e-01
5.15006661e-01 -5.34163892e-01 -4.78280783e-01 2.07445741e-01
4.36907083e-01 -9.84994888e-01 7.55954385e-01 -2.38545060e-01
6.75476566e-02 -7.46334255e-01 -8.76666680e-02 -7.10154593e-01
-4.50827107e-02 -1.11653006e+00 -2.50099480e-01 1.38597822e+00
-1.22043729e-01 -1.99065670e-01 9.68313277e-01 6.29000127e-01
3.62851143e-01 -3.07542592e-01 -9.49573696e-01 -9.70542610e-01
-2.62637317e-01 -1.05565801e-01 2.38825679e-01 1.05796576e+00
-4.93230373e-01 -1.19583681e-01 -4.11837876e-01 2.73112863e-01
7.16294110e-01 1.74342155e-01 1.06938481e+00 -1.07867169e+00
1.90012790e-02 -4.43962574e-01 -5.89033365e-01 -1.42064631e+00
5.85601963e-02 -1.34811446e-01 3.00560802e-01 -1.20110655e+00
2.66336709e-01 -2.62327880e-01 1.78608641e-01 -1.28076687e-01
-4.73613143e-01 2.28939354e-01 2.67438143e-01 5.22783041e-01
-8.68014216e-01 1.74811408e-01 1.21730089e+00 6.27515092e-02
-3.56544882e-01 2.02428415e-01 -8.38708365e-04 1.09604669e+00
6.19402051e-01 -5.15752852e-01 -1.95820853e-01 -2.41335645e-01
-2.49954328e-01 1.86683044e-01 2.29239598e-01 -8.99078488e-01
3.29566866e-01 -4.72515076e-01 2.93155551e-01 -8.47599328e-01
2.07995668e-01 -1.13046348e+00 3.59500319e-01 4.57653910e-01
1.60484239e-01 2.39724159e-01 1.29550844e-01 8.40152860e-01
-2.69057691e-01 -4.65916216e-01 9.53271568e-01 -2.63324320e-01
-9.56981957e-01 4.64399844e-01 -8.32070529e-01 2.74080262e-02
1.50711143e+00 -6.91727340e-01 3.25279944e-02 -3.99274379e-01
-5.30624270e-01 2.41533339e-01 1.00852990e+00 1.46308705e-01
4.09533024e-01 -1.12926853e+00 -4.03745174e-01 -1.78121007e-03
-2.93262452e-01 2.01558053e-01 2.03524649e-01 1.22284317e+00
-9.29433167e-01 1.60239533e-01 1.44710585e-01 -1.00043499e+00
-1.64903092e+00 7.72712111e-01 2.74972320e-01 -9.21729505e-02
-8.25138748e-01 6.78611398e-01 2.79839188e-01 3.85882169e-01
2.97136575e-01 -4.77138638e-01 6.85059503e-02 1.31674260e-01
7.27466702e-01 6.60700619e-01 -2.08799466e-01 -1.07694829e+00
-4.27360624e-01 8.75367343e-01 1.59730598e-01 -1.87103316e-01
8.51834714e-01 -4.96578306e-01 -2.59646267e-01 4.53807443e-01
1.09593284e+00 3.90646726e-01 -1.83120656e+00 -8.64007771e-02
1.49939060e-01 -9.38995659e-01 -1.47748441e-01 8.72534364e-02
-1.30928707e+00 8.06019485e-01 3.10582310e-01 4.24715638e-01
1.09070623e+00 -2.57975221e-01 8.88160646e-01 2.19373871e-02
3.56324673e-01 -1.12465060e+00 4.51539308e-02 3.74032646e-01
1.38353422e-01 -1.34653378e+00 5.06493784e-02 -7.05231607e-01
-8.25019956e-01 1.22421551e+00 5.37056506e-01 -1.11204222e-01
3.52590054e-01 2.69983828e-01 2.57611629e-02 1.46536082e-01
-2.71539479e-01 -2.98609227e-01 1.65912107e-01 3.88009101e-01
1.58176001e-03 -2.90168226e-01 -6.53383732e-02 -1.83128372e-01
3.99705142e-01 -2.91090548e-01 8.10301960e-01 1.10171986e+00
-6.99011624e-01 -8.40823293e-01 -7.54504919e-01 6.51167110e-02
-6.57143295e-01 2.42794827e-01 -3.51925790e-01 1.01369548e+00
1.73669308e-02 1.11937559e+00 1.91176131e-01 7.16050714e-02
-1.02423377e-01 -2.41768181e-01 4.85013127e-01 -3.81746948e-01
-2.99522132e-01 7.69314110e-01 -1.87406689e-01 -6.50007069e-01
-8.41554523e-01 -9.62738335e-01 -1.30526471e+00 -4.78478283e-01
-6.44435108e-01 1.78679943e-01 2.92427450e-01 7.86282122e-01
-3.85570116e-02 2.43061543e-01 5.90884447e-01 -1.03175330e+00
-9.48710814e-02 -5.34639478e-01 -5.67738473e-01 4.95759547e-01
5.46006739e-01 -5.47558606e-01 -5.23567975e-01 5.54646850e-01]
|
[8.894689559936523, -0.8621174097061157]
|
08885234-13fa-4205-a3ac-3a57c14b344c
|
stochastic-distributed-optimization-under
|
2304.07504
| null |
https://arxiv.org/abs/2304.07504v1
|
https://arxiv.org/pdf/2304.07504v1.pdf
|
Stochastic Distributed Optimization under Average Second-order Similarity: Algorithms and Analysis
|
We study finite-sum distributed optimization problems with $n$-clients under popular $\delta$-similarity condition and $\mu$-strong convexity. We propose two new algorithms: SVRS and AccSVRS motivated by previous works. The non-accelerated SVRS method combines the techniques of gradient-sliding and variance reduction, which achieves superior communication complexity $\tilde{\gO}(n {+} \sqrt{n}\delta/\mu)$ compared to existing non-accelerated algorithms. Applying the framework proposed in Katyusha X, we also build a direct accelerated practical version named AccSVRS with totally smoothness-free $\tilde{\gO}(n {+} n^{3/4}\sqrt{\delta/\mu})$ communication complexity that improves upon existing algorithms on ill-conditioning cases. Furthermore, we show a nearly matched lower bound to verify the tightness of our AccSVRS method.
|
['Zhihua Zhang', 'Haishan Ye', 'Yuze Han', 'Dachao Lin']
|
2023-04-15
| null | null | null | null |
['distributed-optimization']
|
['methodology']
|
[ 9.18116793e-02 -1.48117468e-01 -5.27017303e-02 -3.21146876e-01
-1.39553726e+00 -4.18924451e-01 -5.07466733e-01 4.84818704e-02
-7.57548153e-01 1.09605479e+00 -2.33199358e-01 -8.11265647e-01
-6.55948997e-01 -8.42578232e-01 -6.33164048e-01 -1.00469851e+00
-7.72774756e-01 2.32531279e-01 -9.86346155e-02 -4.48169470e-01
3.40329856e-01 1.59073815e-01 -1.01036870e+00 -2.49503404e-01
9.45631206e-01 1.69651532e+00 1.63920578e-02 7.13030159e-01
8.24117213e-02 5.29266953e-01 -3.53890806e-01 -4.91738319e-01
8.02963197e-01 -6.44799709e-01 -7.04349816e-01 -1.51888669e-01
-3.19377147e-02 -2.77014315e-01 -3.97468746e-01 1.23528111e+00
5.89989066e-01 5.83598375e-01 3.62809211e-01 -1.00357950e+00
-3.98477286e-01 6.77948177e-01 -1.46075475e+00 2.13777527e-01
2.01299265e-01 -1.07021667e-01 1.05134487e+00 -7.25493491e-01
1.83980539e-01 9.85138535e-01 8.50022674e-01 1.93170905e-01
-9.79972124e-01 -1.19861245e+00 3.05057228e-01 -1.56385154e-01
-1.72172403e+00 -1.78488135e-01 3.14053655e-01 2.15689287e-01
7.11434662e-01 6.61035776e-01 1.55719608e-01 -6.91922009e-02
-2.46326119e-01 5.21180272e-01 1.24966300e+00 -5.20309925e-01
3.71700615e-01 -2.36486182e-01 1.56870008e-01 8.81001890e-01
2.65206903e-01 -2.21904423e-02 -4.37053233e-01 -6.17944956e-01
7.02265561e-01 5.90217598e-02 -3.72346729e-01 2.32226551e-01
-7.34894276e-01 1.13340509e+00 2.86093026e-01 1.19693711e-01
-2.07859099e-01 7.37471342e-01 2.29121238e-01 3.28636080e-01
5.17382085e-01 -2.34763190e-01 -6.04168177e-01 -3.60410810e-01
-1.11158788e+00 2.82834440e-01 5.89775264e-01 1.60310984e+00
6.04146004e-01 2.10155353e-01 -3.22105974e-01 7.74338365e-01
2.69339442e-01 9.43090022e-01 -2.81995922e-01 -1.33509243e+00
1.05580127e+00 6.83716461e-02 1.71216071e-01 -8.64213765e-01
-2.75032073e-01 -6.02766514e-01 -1.15950930e+00 -3.84449065e-02
5.03489077e-01 -5.48418880e-01 -2.87355423e-01 1.73905325e+00
2.85373926e-01 -1.26961095e-03 -2.67589122e-01 8.73044610e-01
1.44394115e-01 8.62893760e-01 -4.78978992e-01 -9.16563094e-01
1.23026192e+00 -8.52922142e-01 -2.63319850e-01 1.74039960e-01
7.70590723e-01 -7.59187281e-01 9.89161551e-01 5.73251784e-01
-1.62634790e+00 1.05642252e-01 -8.15012157e-01 2.36396715e-01
2.00150520e-01 -1.37533590e-01 8.33686471e-01 1.15830076e+00
-1.21013749e+00 6.26327217e-01 -6.15585864e-01 4.18905079e-01
5.56199670e-01 8.30496550e-01 2.49821860e-02 -2.55393267e-01
-5.47246873e-01 1.59934293e-02 -3.40327263e-01 1.09561250e-01
-5.48231184e-01 -6.41718447e-01 -5.85108578e-01 1.54027984e-01
6.90375209e-01 -3.13172400e-01 9.60024238e-01 -3.30802262e-01
-1.45638096e+00 3.94624919e-01 -3.85989517e-01 -2.91120231e-01
5.03121674e-01 1.53685912e-01 5.37305325e-02 2.81668603e-02
8.96128938e-02 -2.72917747e-01 3.44804376e-01 -9.41594362e-01
-7.99329937e-01 -8.58987451e-01 -1.10394314e-01 1.45522967e-01
-3.60889256e-01 4.10739988e-01 -4.12992120e-01 -6.78553641e-01
2.41506457e-01 -8.38311791e-01 -8.49309385e-01 -1.32057890e-02
-4.87401664e-01 4.19200882e-02 1.82175457e-01 -4.73638088e-01
1.46087492e+00 -2.06644893e+00 -1.22342810e-01 1.06371129e+00
3.49070311e-01 4.07332145e-02 6.04699180e-02 4.69033957e-01
4.64042686e-02 2.43066683e-01 -2.69469708e-01 -3.26158792e-01
6.35936335e-02 6.17728978e-02 3.78824286e-02 8.17439497e-01
-8.18656445e-01 2.48454973e-01 -7.39706635e-01 -3.13093185e-01
-3.75358075e-01 2.57100344e-01 -7.98528254e-01 -8.12549517e-02
1.47843361e-01 2.35213667e-01 -7.57902086e-01 6.77537203e-01
1.19914067e+00 -4.52071071e-01 2.49830902e-01 2.51012087e-01
-1.11700848e-01 -1.29057407e-01 -1.66914511e+00 1.53645313e+00
-5.60770810e-01 -6.21882491e-02 9.08659756e-01 -1.32106817e+00
6.63808703e-01 4.37521487e-02 7.01088071e-01 -7.96052039e-01
4.04531837e-01 4.95401710e-01 -4.89268661e-01 -2.37611756e-02
5.68036795e-01 -4.76877332e-01 -1.83458298e-01 6.82625413e-01
-4.87212986e-01 1.41419144e-02 1.72473237e-01 2.83569306e-01
1.46068943e+00 -3.62067670e-01 1.67279458e-03 -4.88332897e-01
4.54292804e-01 -5.16404808e-01 6.35729730e-01 9.85850751e-01
-2.27683187e-01 3.30212861e-01 7.64334142e-01 1.28665134e-01
-7.65272141e-01 -1.10655630e+00 9.25277323e-02 1.60608423e+00
2.85746843e-01 -4.86609906e-01 -4.77893740e-01 -5.27429223e-01
-4.02130857e-02 7.91649878e-01 -4.92536098e-01 4.06127661e-01
-3.18322897e-01 -1.17709398e+00 4.97637331e-01 5.51172078e-01
3.51493508e-01 -1.83595419e-01 -1.16831705e-01 9.36622396e-02
-1.02119036e-01 -6.82275712e-01 -1.06033349e+00 3.89153391e-01
-7.49023914e-01 -6.12342596e-01 -9.12063062e-01 -4.55933452e-01
9.94168878e-01 4.65298414e-01 6.61371768e-01 1.60404161e-01
-2.08951846e-01 1.76070541e-01 -3.09474170e-01 -2.77790666e-01
3.27732116e-01 -2.49116719e-01 1.77582279e-01 -1.15668602e-01
-2.23435208e-01 -7.52525568e-01 -1.13523936e+00 2.46777162e-01
-6.51436865e-01 -6.18567109e-01 3.46304446e-01 7.87491679e-01
8.93942893e-01 1.48169196e-03 4.63168919e-01 -8.64224732e-01
5.99356592e-01 -3.66704613e-01 -1.11148512e+00 -2.44384017e-02
-9.43897307e-01 9.10046548e-02 8.63949060e-01 -1.58212557e-02
-7.38499939e-01 -1.66561589e-01 -3.76526833e-01 -3.66880834e-01
6.61944032e-01 3.62917632e-01 2.51709461e-01 -4.18869704e-01
4.87125993e-01 2.61291414e-01 6.16231235e-03 -3.24157566e-01
4.78858620e-01 6.14703715e-01 2.58757710e-01 -7.67607987e-01
6.93769157e-01 6.47971332e-01 3.04169208e-01 -4.97567564e-01
-5.29094815e-01 -3.17466497e-01 4.59179938e-01 2.90077627e-01
2.26857886e-01 -8.09425771e-01 -1.71007454e+00 -1.58111826e-01
-7.32165813e-01 -4.14041996e-01 -5.27637184e-01 5.74512064e-01
-5.98902166e-01 6.52371526e-01 -6.82063818e-01 -1.49469614e+00
-7.30774403e-01 -1.00869906e+00 8.26847374e-01 1.40341684e-01
1.48595318e-01 -7.03064561e-01 -3.29779834e-01 4.93650079e-01
7.24859595e-01 -6.98796008e-03 5.85515857e-01 -5.71907401e-01
-5.62293589e-01 -3.59262645e-01 -7.38979995e-01 4.26761538e-01
-2.32749820e-01 -3.73355418e-01 -1.59722328e-01 -6.32204056e-01
1.37318805e-01 -1.54280230e-01 5.98186731e-01 5.30316114e-01
1.73949826e+00 -8.33889306e-01 -3.14869672e-01 8.83298218e-01
1.81506848e+00 1.42600670e-01 6.26211882e-01 -2.48937830e-01
8.68180618e-02 -1.27026111e-01 5.09556234e-01 1.23376179e+00
4.61813211e-01 5.32151878e-01 5.27368903e-01 -4.42833304e-02
5.00544608e-01 3.82908106e-01 3.22138041e-01 7.81445861e-01
-5.02474785e-01 -2.85711914e-01 -3.48717421e-01 4.30368066e-01
-1.59920716e+00 -7.80105531e-01 -4.20093626e-01 2.64114738e+00
9.00755763e-01 -9.17425305e-02 1.55363470e-01 7.46373907e-02
6.77571476e-01 2.75303908e-02 -1.14107333e-01 -6.05951548e-01
5.66730723e-02 9.92450953e-01 1.14521825e+00 5.24980843e-01
-5.43670416e-01 4.45454597e-01 4.92808104e+00 1.72981346e+00
-4.94919896e-01 4.24921364e-01 8.21801186e-01 -6.31659985e-01
-4.56281602e-01 8.86160061e-02 -7.19369590e-01 6.93632543e-01
7.35437632e-01 -3.00417930e-01 7.86139488e-01 9.44002092e-01
2.52440542e-01 -4.20823365e-01 -7.55965710e-01 1.23053634e+00
4.49526357e-03 -1.24895310e+00 -6.76118314e-01 3.50302160e-01
9.20145690e-01 -2.92484909e-01 3.43184918e-01 9.36497003e-02
7.88237035e-01 -1.09361982e+00 1.87413722e-01 -2.71715410e-02
1.26107800e+00 -1.04442787e+00 7.73352861e-01 1.90550417e-01
-1.46828234e+00 -2.78736562e-01 -1.36989966e-01 -1.30918562e-01
3.25808436e-01 7.18959391e-01 -1.50606737e-01 8.85124385e-01
9.06359494e-01 -3.66618782e-01 3.55707884e-01 7.62715757e-01
1.86798811e-01 4.14159775e-01 -1.00673091e+00 -2.55542725e-01
3.25648874e-01 -5.57863235e-01 4.13618237e-01 9.97808337e-01
8.10280561e-01 8.10716748e-01 3.08741748e-01 4.76486892e-01
-3.02494287e-01 4.59039032e-01 -7.32699707e-02 4.85362202e-01
5.14266431e-01 1.10519636e+00 -8.35189581e-01 -1.80490866e-01
-3.21176946e-01 8.05791855e-01 1.08965889e-01 3.47900778e-01
-1.16298985e+00 -9.42377329e-01 6.26470089e-01 -2.95352750e-02
6.92008734e-01 -3.80068183e-01 -6.87584281e-01 -7.19585836e-01
2.12030798e-01 -4.01747912e-01 6.99869633e-01 -2.29719684e-01
-1.28168666e+00 3.47857118e-01 -1.07013516e-01 -9.57927942e-01
2.07614191e-02 -3.17958057e-01 -3.73393506e-01 8.45247328e-01
-1.16007102e+00 -6.09582901e-01 -5.28886318e-02 9.46158469e-01
-3.17593031e-02 -2.59687435e-02 7.53661454e-01 5.77270925e-01
-2.50031739e-01 1.28096747e+00 8.06275010e-01 -7.37919286e-02
2.14318499e-01 -7.51811385e-01 -3.26178998e-01 7.37046957e-01
-4.14541453e-01 5.30918062e-01 6.83094919e-01 -2.61065215e-01
-1.65267384e+00 -9.69640017e-01 6.07793212e-01 2.49588847e-01
4.27111566e-01 -1.69036448e-01 -3.34537178e-01 5.22194147e-01
6.31151348e-02 2.82623321e-01 9.09614801e-01 -6.83117541e-04
-1.85102925e-01 -6.62764311e-01 -1.47621274e+00 5.77857792e-01
1.35105550e+00 -9.10471901e-02 3.53721172e-01 6.26961768e-01
4.82405871e-01 -5.87668419e-01 -1.01569116e+00 2.44226485e-01
3.17584962e-01 -1.07051933e+00 9.81091261e-01 -2.00722203e-01
5.25524653e-02 1.37625724e-01 -6.86667919e-01 -7.82005131e-01
6.41039833e-02 -1.31832623e+00 1.22958012e-02 7.47305036e-01
5.92553437e-01 -9.23073232e-01 1.06847095e+00 6.88946724e-01
-1.64902031e-01 -1.08899772e+00 -1.53476155e+00 -9.12463367e-01
2.66451538e-01 -6.99286938e-01 3.52004021e-01 6.23807967e-01
2.83306926e-01 -8.99106041e-02 -4.74576205e-01 1.03514753e-01
8.46938491e-01 -2.17261203e-02 5.72236776e-01 -3.46531123e-01
-8.99894774e-01 -4.36384469e-01 1.19159453e-01 -1.30635321e+00
-4.97748107e-01 -7.92257190e-01 -6.40048906e-02 -1.15296066e+00
2.21390679e-01 -1.03166401e+00 -2.92908818e-01 5.92337787e-01
1.09328389e-01 3.23219568e-01 1.76131174e-01 -2.45479986e-01
-9.34454799e-01 6.59499764e-01 9.47915435e-01 2.24102572e-01
-1.98549464e-01 4.07514095e-01 -1.06634116e+00 5.02398968e-01
4.41802561e-01 -4.00843441e-01 -3.41909677e-01 -4.46248174e-01
7.47186005e-01 7.58682549e-01 -6.51224554e-02 -6.28523171e-01
3.25232506e-01 -3.41971904e-01 -1.21717498e-01 -4.78865653e-01
4.58044022e-01 -7.67728150e-01 -1.34627625e-01 4.42326874e-01
-1.71050042e-01 2.16150999e-01 -7.88119808e-02 6.39648139e-01
3.61857742e-01 -8.30559060e-02 7.65930533e-01 -5.51092997e-02
1.70667693e-01 6.18560195e-01 -2.87193328e-01 2.27266341e-01
1.31026292e+00 -1.55246243e-01 -3.09902787e-01 -9.05645669e-01
-5.63023448e-01 6.14017189e-01 -7.01021850e-02 -6.11711264e-01
5.58046341e-01 -9.96113539e-01 -8.12979937e-01 -2.38760151e-02
-4.49972570e-01 2.69114554e-01 7.14582145e-01 1.44444370e+00
-6.56860650e-01 8.99551138e-02 5.18018484e-01 -2.51283258e-01
-1.04045534e+00 4.96563435e-01 1.46899093e-02 -6.14057124e-01
-1.15808114e-01 1.62842226e+00 -2.18337253e-01 -1.10681457e-02
2.47474253e-01 -2.60014802e-01 7.92482376e-01 -4.73572522e-01
3.78118604e-01 1.12117028e+00 1.16566397e-01 -8.71248022e-02
-4.37322974e-01 3.71368647e-01 -5.43050915e-02 -4.44615901e-01
1.51950669e+00 -5.43271899e-01 -4.43189025e-01 -1.85286447e-01
1.44245279e+00 4.57332551e-01 -9.11476076e-01 -2.95256287e-01
-4.69956994e-01 -6.82404935e-01 -2.76131570e-01 -4.76609558e-01
-1.42426193e+00 6.88600004e-01 7.32047021e-01 1.38851926e-01
1.36298096e+00 -1.30244652e-02 1.04259157e+00 1.55900940e-01
8.99769902e-01 -1.53597820e+00 -1.94945466e-02 5.94810426e-01
4.14040625e-01 -9.44965959e-01 4.56649095e-01 -5.74665725e-01
-4.91482079e-01 8.34350765e-01 1.30649820e-01 -3.23448360e-01
9.50546086e-01 4.23956573e-01 -5.39841473e-01 -7.94259831e-02
-3.28176647e-01 -1.22891136e-01 -2.42591217e-01 2.23577589e-01
4.18759763e-01 1.35464326e-01 -9.19804037e-01 1.10878229e+00
-2.07999572e-01 -1.51536316e-01 4.21270192e-01 1.05178630e+00
-1.68477193e-01 -1.13265789e+00 -4.00331140e-01 8.44127417e-01
-9.08260345e-01 -3.71989459e-01 4.05783504e-01 5.38789868e-01
-1.17102422e-01 1.19446504e+00 -1.49077281e-01 -2.58264661e-01
3.43640670e-02 -2.58542180e-01 5.24373651e-01 -2.25094005e-01
-6.55640125e-01 4.70504671e-01 1.73330918e-01 -7.07874954e-01
4.76039946e-02 -5.07297993e-01 -1.70952940e+00 -9.29575086e-01
-4.39864844e-01 4.85320121e-01 6.16810143e-01 9.27604854e-01
4.60659564e-01 1.76401511e-01 1.09152400e+00 -5.51154613e-01
-1.13101757e+00 -5.76053977e-01 -1.13512921e+00 -9.29152034e-03
-8.60726833e-02 -3.00502062e-01 -3.87139082e-01 -5.56532264e-01]
|
[6.323436260223389, 4.59801721572876]
|
80cad6e2-aa61-4baf-8c08-406ed0fef065
|
exemplar-based-video-colorization-with-long
|
2303.15081
| null |
https://arxiv.org/abs/2303.15081v1
|
https://arxiv.org/pdf/2303.15081v1.pdf
|
Exemplar-based Video Colorization with Long-term Spatiotemporal Dependency
|
Exemplar-based video colorization is an essential technique for applications like old movie restoration. Although recent methods perform well in still scenes or scenes with regular movement, they always lack robustness in moving scenes due to their weak ability in modeling long-term dependency both spatially and temporally, leading to color fading, color discontinuity or other artifacts. To solve this problem, we propose an exemplar-based video colorization framework with long-term spatiotemporal dependency. To enhance the long-term spatial dependency, a parallelized CNN-Transformer block and a double head non-local operation are designed. The proposed CNN-Transformer block can better incorporate long-term spatial dependency with local texture and structural features, and the double head non-local operation further leverages the performance of augmented feature. While for long-term temporal dependency enhancement, we further introduce the novel linkage subnet. The linkage subnet propagate motion information across adjacent frame blocks and help to maintain temporal continuity. Experiments demonstrate that our model outperforms recent state-of-the-art methods both quantitatively and qualitatively. Also, our model can generate more colorful, realistic and stabilized results, especially for scenes where objects change greatly and irregularly.
|
['Yue Zhang', 'Jiatong Han', 'Yu Zhang', 'Mingdao Wang', 'Xianlin Zhang', 'Xueming Li', 'Siqi Chen']
|
2023-03-27
| null | null | null | null |
['colorization']
|
['computer-vision']
|
[-1.89170927e-01 -7.57428229e-01 -1.46347880e-01 -1.80265129e-01
2.15544701e-02 -2.92394340e-01 4.80649352e-01 -2.93547183e-01
-2.30684623e-01 7.14751720e-01 2.42663622e-01 8.45710002e-03
1.53424349e-02 -8.14752042e-01 -8.37472320e-01 -6.94283903e-01
-6.90514818e-02 -4.11610544e-01 7.57178664e-01 -4.26709026e-01
6.59534037e-02 4.97021586e-01 -1.44186556e+00 3.72690350e-01
9.48473275e-01 8.59698236e-01 1.60446525e-01 3.98233384e-01
-1.03352413e-01 1.01766264e+00 -2.56043673e-01 2.91750170e-02
3.16140682e-01 -4.63012308e-01 -4.07955706e-01 1.20642096e-01
5.33296645e-01 -4.86094862e-01 -5.26256621e-01 9.64283228e-01
3.56474429e-01 4.06420022e-01 1.54922709e-01 -1.34665716e+00
-9.48394537e-01 2.15734661e-01 -1.16309917e+00 2.48656526e-01
4.02752280e-01 3.61696512e-01 6.40387893e-01 -7.92967796e-01
7.27412045e-01 1.35340691e+00 7.42364347e-01 4.73878980e-01
-1.15899074e+00 -8.21881473e-01 6.78691447e-01 4.36649442e-01
-1.21218240e+00 -2.54702419e-01 1.12928283e+00 -1.83911607e-01
5.75209796e-01 1.46525621e-01 9.11783278e-01 8.77365887e-01
2.47265458e-01 7.06217110e-01 1.07299733e+00 -1.32892206e-01
5.29863387e-02 -4.44747269e-01 -2.19963983e-01 8.35428953e-01
-1.68559644e-02 4.13576603e-01 -6.58780932e-01 2.79647052e-01
1.29791594e+00 3.39525759e-01 -5.93654871e-01 -4.67820317e-01
-1.50188172e+00 3.20668072e-01 8.04697871e-01 4.02485192e-01
-2.85323292e-01 4.90307987e-01 2.49915600e-01 1.19521812e-01
4.45072055e-01 8.07172209e-02 -3.88817340e-01 -1.70981035e-01
-1.17221844e+00 8.76980498e-02 2.03018352e-01 9.36570227e-01
7.52026021e-01 3.97293150e-01 -4.63013798e-01 9.16048825e-01
8.52719247e-02 1.86850473e-01 3.63754779e-01 -1.09518421e+00
2.68262088e-01 5.70561111e-01 2.07657367e-01 -1.20693016e+00
-5.56330979e-01 -4.83380407e-01 -1.19524777e+00 4.49077219e-01
2.77460188e-01 1.54205889e-01 -1.03806984e+00 1.94750261e+00
3.87382865e-01 6.43679857e-01 -3.49528968e-01 1.16236460e+00
6.35765195e-01 9.25307035e-01 1.03707670e-03 -4.68422741e-01
1.26659179e+00 -1.36818731e+00 -9.34590161e-01 3.61467153e-02
4.65943888e-02 -7.82810092e-01 1.20467782e+00 2.25949183e-01
-1.23841071e+00 -8.54104102e-01 -1.03960824e+00 -2.11790770e-01
-1.16814204e-01 -1.40471444e-01 6.98696375e-01 2.32084975e-01
-1.12154233e+00 4.75823402e-01 -6.98924601e-01 -2.63521761e-01
3.56820703e-01 1.24939354e-04 -4.11072433e-01 -5.63326240e-01
-1.14845598e+00 4.94929016e-01 1.82748452e-01 4.53081280e-01
-6.43235028e-01 -8.35341513e-01 -8.51565659e-01 1.57552306e-02
3.31195086e-01 -8.27619612e-01 7.40266681e-01 -1.10888374e+00
-1.54593706e+00 2.09275261e-01 -2.29553953e-01 2.42277160e-02
6.51272357e-01 -1.35335028e-01 -6.19376659e-01 1.54040590e-01
9.76060145e-03 7.99191892e-01 9.02856767e-01 -1.55781937e+00
-7.29304075e-01 -2.10008584e-02 3.39707226e-01 1.34754881e-01
-5.13954520e-01 -1.67606071e-01 -1.15946543e+00 -1.32113934e+00
2.35214069e-01 -7.80062199e-01 -1.33608282e-01 5.35220385e-01
-2.24500671e-01 2.19071701e-01 1.34457171e+00 -6.05403244e-01
1.48004973e+00 -2.23971033e+00 1.58388138e-01 -8.29674304e-02
2.49873385e-01 2.43714407e-01 -3.94549638e-01 2.87125736e-01
-2.22794279e-01 -2.82993149e-02 -2.64218986e-01 -2.20890746e-01
-3.05454820e-01 2.39656061e-01 3.92930619e-02 3.42459738e-01
2.32435629e-01 9.10442710e-01 -1.11937666e+00 -3.78885537e-01
5.51777840e-01 8.18823636e-01 -6.86853409e-01 -6.72444627e-02
-1.49275690e-01 6.73544526e-01 -3.03728461e-01 7.34980881e-01
1.06211400e+00 -1.82009786e-01 -2.13081762e-01 -6.74546182e-01
-3.38900715e-01 -3.82664025e-01 -1.27321076e+00 2.02386999e+00
-4.39634115e-01 7.63465524e-01 2.25545287e-01 -7.32248127e-01
5.82157552e-01 1.10744201e-01 6.24471605e-01 -1.10724103e+00
-8.11459497e-02 -9.84393358e-02 -1.82236910e-01 -5.44744313e-01
6.83330119e-01 -1.43511459e-01 4.63310421e-01 2.32342958e-01
-4.79724139e-01 3.03328902e-01 3.20290834e-01 3.06327254e-01
8.96769404e-01 5.38556397e-01 -2.85837263e-01 -2.62035280e-01
6.54064655e-01 -2.89201021e-01 9.92749155e-01 3.89723808e-01
-3.67274404e-01 9.84751284e-01 3.51861686e-01 -6.00788295e-01
-9.33341861e-01 -8.05203974e-01 9.30701718e-02 1.10504973e+00
8.47720742e-01 -4.14053500e-01 -4.39474165e-01 -4.70426917e-01
-1.40659973e-01 3.82979572e-01 -8.05662155e-01 -2.63031155e-01
-7.92712450e-01 -6.12262011e-01 7.30748847e-02 7.67823040e-01
9.24061954e-01 -9.51781988e-01 -3.45718324e-01 4.31682050e-01
-3.71955127e-01 -1.06293225e+00 -8.55335295e-01 -2.56719589e-01
-8.37721884e-01 -9.24967647e-01 -1.14006960e+00 -7.35770345e-01
6.33296430e-01 7.50113368e-01 1.02061212e+00 3.30802590e-01
-2.86402553e-01 1.25756219e-01 -5.93686759e-01 2.68956512e-01
1.45895347e-01 -4.72436577e-01 -2.49357656e-01 2.52514690e-01
-2.06944406e-01 -6.99484587e-01 -1.17257977e+00 5.58918893e-01
-1.40010464e+00 6.19468093e-01 2.33798310e-01 1.03731883e+00
3.99677038e-01 8.33810270e-02 3.10820520e-01 -5.68401814e-01
3.52537632e-01 -1.25182807e-01 -3.91574502e-01 3.52325082e-01
-3.18736106e-01 -5.48726097e-02 6.16074204e-01 -5.49762189e-01
-1.27656341e+00 -1.13949083e-01 -2.31938693e-03 -6.30711138e-01
1.94435671e-01 3.78701836e-01 -1.50596604e-01 -1.75843075e-01
2.26731181e-01 1.42524093e-01 -1.68065146e-01 -4.68840212e-01
4.36470419e-01 -7.89947733e-02 6.70235276e-01 -5.71985722e-01
9.50112164e-01 8.28242958e-01 -2.70636044e-02 -5.79568148e-01
-5.83790779e-01 -1.81194440e-01 -5.45105636e-01 -5.43508828e-01
9.52949226e-01 -9.37679648e-01 -6.31707728e-01 7.12125063e-01
-1.10622203e+00 -4.85654801e-01 -1.97026610e-01 2.76074320e-01
-2.71243006e-01 5.26347101e-01 -8.72292459e-01 -4.43952471e-01
-6.52131364e-02 -1.17295098e+00 8.82075489e-01 4.58171964e-01
3.21615398e-01 -1.01375258e+00 -1.72545776e-01 2.98835970e-02
6.90037847e-01 4.71033365e-01 8.82182896e-01 6.32408500e-01
-8.37760329e-01 7.15398192e-02 -6.04050338e-01 3.07295293e-01
3.12014669e-01 3.63227338e-01 -6.05440855e-01 -4.64466184e-01
-4.21803623e-01 1.77660212e-01 1.12763500e+00 5.66208780e-01
1.23523664e+00 2.71208510e-02 -1.91922590e-01 1.00495613e+00
1.50816727e+00 2.57847875e-01 8.56323600e-01 5.93683183e-01
1.17932522e+00 5.29747128e-01 5.92750728e-01 3.83496761e-01
4.21645314e-01 8.72264087e-01 4.06161100e-01 -7.11507440e-01
-6.56954110e-01 -1.07527643e-01 3.23619127e-01 7.87547648e-01
-2.84714967e-01 -1.95766196e-01 -5.70125341e-01 3.74716550e-01
-2.08935070e+00 -1.05789173e+00 -1.66722462e-01 2.00174952e+00
5.68751931e-01 -7.30422214e-02 7.40112662e-02 1.30304620e-01
7.34650493e-01 2.74307191e-01 -5.55115521e-01 2.26477608e-02
-5.42356491e-01 -1.20051563e-01 3.05883646e-01 3.75667483e-01
-8.68911862e-01 9.01941121e-01 5.68624544e+00 8.75154793e-01
-1.38204420e+00 1.58075064e-01 7.82841682e-01 -1.23601936e-01
-4.80246156e-01 1.78416148e-02 -7.60702118e-02 5.55424511e-01
-2.58874856e-02 1.26862720e-01 2.98788399e-01 3.22117537e-01
6.15035653e-01 -2.59107679e-01 -8.01612675e-01 1.15107751e+00
1.01499356e-01 -1.45798898e+00 1.89796969e-01 -1.28132716e-01
1.11939049e+00 -4.08635378e-01 1.26114234e-01 1.93982124e-02
-3.34886797e-02 -7.74667919e-01 1.14342630e+00 5.19534707e-01
8.92445207e-01 -7.85914063e-01 3.93123329e-01 -1.08046748e-01
-1.71681607e+00 -1.55548602e-01 -1.71453565e-01 8.08799565e-02
4.66025740e-01 6.05360508e-01 3.89905542e-01 7.76286364e-01
8.47939253e-01 1.24175024e+00 -7.52988517e-01 1.21798098e+00
-1.16812214e-01 1.60904855e-01 -1.05829224e-01 4.29358214e-01
4.39630389e-01 -4.00573313e-01 3.55663270e-01 1.21151686e+00
3.68880510e-01 2.03074098e-01 1.31552726e-01 7.54024029e-01
1.08911417e-01 -2.01356098e-01 -2.18458161e-01 3.31791610e-01
2.68915147e-01 1.23388052e+00 -9.38594282e-01 -3.88572365e-01
-6.59396052e-01 1.37259734e+00 8.53796229e-02 9.64538217e-01
-1.13533437e+00 -2.12204173e-01 6.78583860e-01 2.56175429e-01
3.94632578e-01 -4.56617951e-01 -1.11852147e-01 -1.41015160e+00
2.35966593e-01 -6.87679052e-01 3.19588095e-01 -1.09764624e+00
-1.26293838e+00 6.51062787e-01 -1.93829030e-01 -1.63247418e+00
1.84843138e-01 -4.22590762e-01 -7.69655347e-01 5.71122944e-01
-1.87127709e+00 -1.52887070e+00 -7.32104778e-01 1.03849483e+00
5.99920630e-01 2.25257099e-01 2.97948360e-01 7.31089294e-01
-9.15679634e-01 4.29319590e-01 2.62697995e-01 1.03413416e-02
1.02970016e+00 -8.51378500e-01 3.63083631e-02 1.19649613e+00
-1.23966478e-01 5.86744726e-01 5.52480519e-01 -5.10481060e-01
-1.50124562e+00 -1.15570748e+00 1.33480594e-01 1.12967961e-01
2.60443956e-01 -1.38392285e-01 -1.01714647e+00 2.12542653e-01
1.87571868e-01 4.86113459e-01 5.27378656e-02 -1.02007195e-01
-5.12317955e-01 -4.81235832e-01 -8.38922918e-01 7.52776742e-01
1.10303903e+00 -3.48946750e-01 -9.45442840e-02 4.46463078e-02
6.75293863e-01 -3.28943580e-01 -5.62043905e-01 5.00620246e-01
8.56526732e-01 -1.37253475e+00 1.14936721e+00 -1.94048077e-01
6.98462963e-01 -8.60194564e-01 1.58727467e-01 -1.05734468e+00
-7.06681430e-01 -7.01109171e-01 -5.42316735e-02 1.38067222e+00
1.38407335e-01 -3.41512144e-01 6.57647252e-01 5.27617455e-01
-2.24678904e-01 -7.20037758e-01 -8.63576531e-01 -8.63013864e-01
-1.00078009e-01 -4.11382645e-01 4.82027769e-01 1.20780408e+00
-2.95588046e-01 -2.30191369e-02 -8.12574863e-01 -8.27737227e-02
4.42451894e-01 2.46678948e-01 6.62524045e-01 -7.14652836e-01
-1.12864219e-01 -5.30325770e-01 -3.10200214e-01 -1.24594748e+00
-1.73196450e-01 -2.64405042e-01 1.00951873e-01 -1.72000432e+00
4.07387614e-01 -5.57023764e-01 -6.73516989e-01 4.01131779e-01
-4.15260792e-01 7.29995668e-01 3.84555399e-01 2.02483937e-01
-7.03754783e-01 8.08259368e-01 1.78151357e+00 -1.12306781e-01
-3.39255899e-01 -5.76830685e-01 -3.70270938e-01 6.52193129e-01
4.48397994e-01 -1.31773606e-01 -4.47069675e-01 -7.38412321e-01
1.33843780e-01 6.65620491e-02 4.66640711e-01 -1.21226966e+00
3.65941137e-01 -4.30056930e-01 6.63814366e-01 -6.46368742e-01
3.88172954e-01 -8.98477316e-01 3.48853052e-01 3.27441514e-01
-1.31399957e-02 3.52228373e-01 2.85381675e-01 8.42347503e-01
-5.78702509e-01 4.59751040e-01 1.01519632e+00 4.44878824e-02
-1.02312970e+00 7.50492215e-01 -2.21493647e-01 -1.62691236e-01
1.16628003e+00 -5.85785270e-01 -3.16868812e-01 -5.45934737e-01
-4.56452817e-01 2.84374237e-01 8.78547728e-01 7.15111911e-01
7.92592466e-01 -1.58648384e+00 -5.18712878e-01 3.60870808e-01
1.49653971e-01 -1.42494902e-01 8.11780393e-01 1.03405523e+00
-8.72046828e-01 3.23010087e-02 -5.67605019e-01 -6.08823657e-01
-9.56219137e-01 7.58756876e-01 2.01804817e-01 7.70134628e-02
-9.18305039e-01 9.30169702e-01 7.23806679e-01 2.35330790e-01
1.25019506e-01 -4.11156058e-01 -4.84189019e-02 -1.46288484e-01
5.68696201e-01 3.15044224e-01 -3.22925985e-01 -7.23211706e-01
-2.81183511e-01 9.72897828e-01 -8.97807255e-03 -1.34543508e-01
1.30415821e+00 -4.96295124e-01 -2.47371539e-01 2.78255612e-01
1.09402287e+00 1.55737624e-02 -1.74502301e+00 -1.69347674e-01
-4.71714109e-01 -9.52010751e-01 1.15930572e-01 -7.11586833e-01
-1.70216405e+00 8.81945074e-01 6.49899065e-01 -7.94885755e-02
1.62260342e+00 -5.14705300e-01 1.08417106e+00 -3.24199319e-01
2.17936948e-01 -1.03289497e+00 4.80623722e-01 4.69798952e-01
8.25601101e-01 -1.12698674e+00 1.21646218e-01 -6.59925520e-01
-5.27045190e-01 1.26067424e+00 8.37637365e-01 3.40885520e-02
5.80919921e-01 2.47642145e-01 7.24350587e-02 -6.93156868e-02
-6.03257358e-01 -1.08874246e-01 4.63709176e-01 6.02225542e-01
5.97853839e-01 -3.12912345e-01 -3.34255308e-01 1.39469802e-01
4.08590049e-01 -1.02320671e-01 3.58760983e-01 9.71959591e-01
-5.13210781e-02 -1.06363726e+00 -2.75710285e-01 -3.88707442e-04
-1.76711470e-01 -1.22424856e-01 1.28174797e-01 9.02644813e-01
3.54880601e-01 8.78734171e-01 1.38767853e-01 -3.83888394e-01
2.85138994e-01 -4.99626279e-01 3.86160582e-01 -7.10367486e-02
-4.24953878e-01 5.30601561e-01 -1.51839241e-01 -9.97527063e-01
-8.28954995e-01 -2.46284708e-01 -1.11677051e+00 -5.79043984e-01
-2.35486656e-01 -2.06156343e-01 4.34368789e-01 5.51749885e-01
3.17038655e-01 8.53404760e-01 5.79121709e-01 -1.07403731e+00
3.70304406e-01 -6.68920577e-01 -6.89743042e-01 7.27343321e-01
5.99062741e-01 -7.90032029e-01 -1.08734369e-01 3.54279011e-01]
|
[11.062002182006836, -1.1892063617706299]
|
13aed26d-91a7-4b48-970d-75afbd4fbdc8
|
translation-scale-and-rotation-cross-modal
|
2209.13801
| null |
https://arxiv.org/abs/2209.13801v1
|
https://arxiv.org/pdf/2209.13801v1.pdf
|
Translation, Scale and Rotation: Cross-Modal Alignment Meets RGB-Infrared Vehicle Detection
|
Integrating multispectral data in object detection, especially visible and infrared images, has received great attention in recent years. Since visible (RGB) and infrared (IR) images can provide complementary information to handle light variations, the paired images are used in many fields, such as multispectral pedestrian detection, RGB-IR crowd counting and RGB-IR salient object detection. Compared with natural RGB-IR images, we find detection in aerial RGB-IR images suffers from cross-modal weakly misalignment problems, which are manifested in the position, size and angle deviations of the same object. In this paper, we mainly address the challenge of cross-modal weakly misalignment in aerial RGB-IR images. Specifically, we firstly explain and analyze the cause of the weakly misalignment problem. Then, we propose a Translation-Scale-Rotation Alignment (TSRA) module to address the problem by calibrating the feature maps from these two modalities. The module predicts the deviation between two modality objects through an alignment process and utilizes Modality-Selection (MS) strategy to improve the performance of alignment. Finally, a two-stream feature alignment detector (TSFADet) based on the TSRA module is constructed for RGB-IR object detection in aerial images. With comprehensive experiments on the public DroneVehicle datasets, we verify that our method reduces the effect of the cross-modal misalignment and achieve robust detection results.
|
['Xingxing Wei', 'Yinyan Wang', 'Maoxun Yuan']
|
2022-09-28
| null | null | null | null |
['pedestrian-detection', 'object-detection-in-aerial-images']
|
['computer-vision', 'computer-vision']
|
[ 2.92768478e-01 -7.76382387e-01 1.23645112e-01 -7.04043508e-02
-7.22364247e-01 -6.74153686e-01 4.62088078e-01 -2.67236739e-01
-4.46075439e-01 1.99784234e-01 6.67787809e-03 -1.76358506e-01
-1.93025393e-03 -6.55108035e-01 -4.80152786e-01 -9.36418414e-01
5.38092613e-01 -3.08190465e-01 4.02109146e-01 -5.33562005e-01
4.93499897e-02 5.06486237e-01 -1.89816797e+00 2.75113702e-01
6.73500955e-01 1.15659094e+00 1.45929039e-01 6.07428849e-01
2.55185157e-01 3.75022143e-01 -5.31341136e-01 -2.07557485e-01
7.32230365e-01 -1.79842487e-01 -4.42006856e-01 1.98378578e-01
5.97743750e-01 -4.79626179e-01 -3.74587625e-01 1.13543522e+00
7.22214997e-01 8.45063180e-02 4.03971523e-01 -1.53953934e+00
-4.54515964e-01 -1.33914026e-02 -1.11437869e+00 4.87667173e-01
6.49169505e-01 4.61581439e-01 4.84516710e-01 -9.36033726e-01
3.26386206e-02 1.27886307e+00 6.58992946e-01 2.06815273e-01
-6.50429249e-01 -6.87297344e-01 6.96108043e-02 1.72826022e-01
-1.75107765e+00 -2.01969430e-01 7.87775517e-01 -3.60843509e-01
4.84144539e-01 6.07633531e-01 6.02943182e-01 7.29359448e-01
-1.60543352e-01 6.11177444e-01 1.08534443e+00 -4.45438355e-01
-3.31516564e-01 1.72615573e-01 -1.29787192e-01 6.23193979e-01
4.64215577e-01 2.77427912e-01 -5.14786899e-01 -6.58187047e-02
4.34601039e-01 4.09257472e-01 -3.24538708e-01 3.24744657e-02
-1.35072672e+00 3.66210997e-01 7.34866798e-01 8.87158886e-02
-5.17449826e-02 -1.80458367e-01 8.94776210e-02 5.64657487e-02
2.24236414e-01 -6.05555512e-02 -3.06916654e-01 3.35341692e-01
-3.87322694e-01 -8.81087705e-02 -1.05093397e-01 1.04420960e+00
6.76898479e-01 -1.53062806e-01 -2.18154788e-01 5.94799876e-01
6.37539685e-01 1.15454972e+00 3.93278658e-01 -2.60850519e-01
9.49456215e-01 8.06305051e-01 3.15135896e-01 -1.29625106e+00
-5.98138750e-01 -2.22281650e-01 -7.64134586e-01 -4.30551320e-02
4.47499871e-01 -1.20413952e-01 -5.40560663e-01 1.21003056e+00
8.47325802e-01 -3.57329287e-03 1.22031987e-01 1.47120714e+00
1.16410601e+00 4.52390432e-01 -7.54531589e-04 -2.17834130e-01
1.59572089e+00 -4.88580734e-01 -3.18664938e-01 -3.81635100e-01
4.06526685e-01 -1.01970851e+00 8.82137954e-01 1.13001503e-01
-5.89109778e-01 -8.52263331e-01 -8.94916117e-01 1.37159750e-01
-4.82044518e-01 6.83944881e-01 4.15436566e-01 7.67559886e-01
-5.65812051e-01 -4.66145314e-02 -2.57886410e-01 -4.54563558e-01
6.38991073e-02 2.64497638e-01 -3.80009145e-01 -1.52970299e-01
-1.15604365e+00 6.79247737e-01 4.42868322e-01 6.22349560e-01
-2.33449087e-01 -2.87564456e-01 -7.47488976e-01 -5.44283986e-01
5.10374010e-01 -4.24614906e-01 6.45229876e-01 -1.01612043e+00
-1.23529291e+00 9.72106278e-01 -6.97429180e-02 1.65666938e-01
4.96928334e-01 -1.06122769e-01 -8.32225442e-01 4.47145477e-02
1.29321078e-02 2.94525206e-01 9.62527931e-01 -1.31988788e+00
-9.84838545e-01 -7.08406627e-01 -8.12259614e-02 5.05169392e-01
-2.56162316e-01 4.48832184e-01 -4.70497698e-01 -3.71281773e-01
4.40889150e-01 -9.17783678e-01 1.76897347e-01 -1.62751287e-01
-5.91463804e-01 1.07564023e-02 9.51531291e-01 -5.38215816e-01
9.78417814e-01 -2.30009055e+00 -1.24336720e-01 1.60128936e-01
-2.38343161e-02 4.59791332e-01 -3.84229384e-02 -5.60264243e-03
-6.79071173e-02 -1.92104742e-01 5.66093698e-02 -8.10043234e-03
-3.61912310e-01 -9.15121064e-02 -3.30205232e-01 9.63246882e-01
2.08437800e-01 7.20173359e-01 -8.86444330e-01 -5.76892376e-01
4.73447412e-01 3.91159594e-01 2.83115730e-02 2.47764975e-01
3.57688487e-01 6.28678858e-01 -4.72002268e-01 1.25844800e+00
1.07566273e+00 1.07765324e-01 -2.74108052e-01 -8.94815743e-01
-5.22946417e-01 -5.20902634e-01 -1.41412997e+00 1.24220645e+00
-1.12290688e-01 3.99004877e-01 -1.82035550e-01 -5.76651633e-01
1.09111226e+00 1.14208804e-02 4.68243897e-01 -8.47041368e-01
3.44959497e-01 9.52436924e-02 -3.60114098e-01 -7.40363836e-01
7.83271432e-01 1.39842376e-01 -6.23491406e-02 1.47732794e-01
-3.94142836e-01 9.32755042e-03 -3.97800021e-02 -2.35782534e-01
5.32746375e-01 1.79811083e-02 1.99923292e-01 2.54692882e-01
1.00332630e+00 4.74695787e-02 4.46233153e-01 4.79970843e-01
-3.31300914e-01 7.63113618e-01 -2.43496969e-01 -5.77739418e-01
-6.30250216e-01 -8.72949183e-01 -1.34731203e-01 1.10524249e+00
9.30431545e-01 -6.61264807e-02 -2.58195341e-01 -6.01253331e-01
7.42783770e-02 1.41326770e-01 -3.83873463e-01 -2.01842353e-01
-4.91041183e-01 -1.17454398e+00 5.34482896e-01 4.79737103e-01
9.58252311e-01 -5.72904587e-01 -8.67912829e-01 -3.94164830e-01
-5.05821764e-01 -1.32327044e+00 -4.19579148e-01 -2.60083914e-01
-5.69165707e-01 -1.25696373e+00 -4.68309999e-01 -4.02927250e-01
6.22584760e-01 1.35118234e+00 6.78593338e-01 2.77136296e-01
-7.50635624e-01 6.71327651e-01 -4.33387578e-01 -4.23628658e-01
1.02079973e-01 -3.94270688e-01 2.67979026e-01 4.39138293e-01
3.99212003e-01 1.34198010e-01 -7.99509346e-01 7.89656818e-01
-9.87697780e-01 -8.01470801e-02 5.61891198e-01 5.43262303e-01
5.25196552e-01 4.02290635e-02 3.46345007e-02 6.05359487e-02
1.03388287e-01 -2.27078065e-01 -8.97682190e-01 4.34161633e-01
-2.07986146e-01 -3.80898863e-01 3.93366128e-01 -5.01743853e-01
-9.79590058e-01 4.96964365e-01 3.88249844e-01 -4.19495344e-01
-4.29122776e-01 8.40710700e-02 -3.42920274e-01 -4.83985871e-01
7.59105206e-01 2.91389823e-01 -1.79734915e-01 -1.25513434e-01
2.66846359e-01 1.02353919e+00 6.54743254e-01 -2.01353699e-01
1.20140052e+00 8.41575146e-01 2.21504830e-02 -1.02846360e+00
-9.00729954e-01 -9.49930251e-01 -6.48769379e-01 -6.32845342e-01
8.77955675e-01 -1.40034997e+00 -8.06870699e-01 5.68695545e-01
-1.17810822e+00 3.23529899e-01 1.47620082e-01 5.99689722e-01
-8.31106976e-02 5.26561677e-01 -1.97521299e-01 -1.07664549e+00
-2.80664921e-01 -1.24447083e+00 1.52903080e+00 7.11300075e-01
5.88842630e-01 -4.05152529e-01 -8.98520797e-02 5.78284919e-01
2.50224411e-01 3.02447855e-01 1.69009268e-01 -1.20747544e-01
-6.22154236e-01 -2.64572799e-01 -5.69043100e-01 -4.99266200e-02
8.34692940e-02 4.74858433e-02 -1.08097923e+00 -2.42334202e-01
-3.95337820e-01 -4.15511578e-02 7.39119411e-01 1.68451652e-01
8.62896144e-01 9.26448330e-02 -3.97205025e-01 8.67936552e-01
1.46065164e+00 -7.81718567e-02 4.69732910e-01 6.53604329e-01
1.07580066e+00 5.67900717e-01 1.21155548e+00 5.62081337e-01
4.04962927e-01 1.06103432e+00 7.21358240e-01 -4.71156299e-01
9.59191099e-02 5.01726978e-02 5.29046714e-01 1.50451928e-01
-4.06744182e-01 -1.95567384e-02 -8.86020839e-01 1.21949911e-01
-1.85480106e+00 -9.29965138e-01 -5.02607942e-01 2.37011933e+00
3.93516213e-01 -3.98334146e-01 5.01750946e-01 1.63507581e-01
1.06336129e+00 -4.38921936e-02 -3.47830325e-01 4.46267217e-01
-4.32873517e-01 -4.85627979e-01 1.06885946e+00 4.36587781e-02
-1.32115281e+00 7.97403514e-01 5.52166128e+00 6.32454515e-01
-1.30388117e+00 8.27965811e-02 3.19509625e-01 1.55855790e-01
1.63480267e-01 -2.10190415e-01 -1.09246540e+00 4.95989829e-01
1.58231646e-01 3.08635712e-01 4.90502894e-01 6.73417032e-01
1.30459324e-01 -3.30417603e-01 -5.10758936e-01 1.53561437e+00
2.64829606e-01 -6.04399204e-01 -1.67136312e-01 -5.22198081e-02
4.31081563e-01 1.20464459e-01 3.31529945e-01 -1.77326620e-01
-2.41442114e-01 -6.46647692e-01 9.45050418e-01 5.41705906e-01
9.19014871e-01 -9.32779074e-01 8.88406217e-01 2.96425760e-01
-1.76467991e+00 -4.94212598e-01 -6.37597263e-01 2.62934953e-01
-5.03491387e-02 3.72049689e-01 -5.09920716e-01 9.20861363e-01
9.84325767e-01 6.76266551e-01 -9.13359165e-01 9.82773781e-01
-1.64298460e-01 9.89138708e-03 -3.04489523e-01 1.65948540e-01
-2.37221941e-01 -4.40206885e-01 5.58546662e-01 1.04721880e+00
4.25643921e-01 3.21201533e-01 4.70905751e-01 7.18943536e-01
4.58089769e-01 -7.97345862e-02 -6.32711709e-01 2.57526994e-01
4.20792460e-01 1.51840413e+00 -6.46820724e-01 -2.88639888e-02
-4.61311489e-01 9.36270654e-01 -2.88431495e-01 4.48468328e-01
-1.15117311e+00 -2.06742913e-01 5.77357769e-01 -7.72851259e-02
1.82862818e-01 -3.40550631e-01 2.14813389e-02 -1.26662672e+00
1.49801984e-01 -7.52498150e-01 4.76291180e-01 -9.81834054e-01
-1.11904585e+00 5.93465745e-01 8.53975341e-02 -1.86889040e+00
2.73628592e-01 -7.25292444e-01 -4.67953533e-01 7.42646277e-01
-1.85235775e+00 -1.59841383e+00 -1.02291381e+00 9.63671625e-01
1.35642573e-01 -1.25002652e-01 2.96722293e-01 3.46304834e-01
-9.20319259e-01 6.44686043e-01 -1.16189569e-01 2.33538225e-01
7.15781689e-01 -7.15692222e-01 -4.48266529e-02 1.13952756e+00
-5.62502630e-02 4.18493837e-01 5.92707038e-01 -5.52610934e-01
-1.84436274e+00 -1.29324961e+00 2.46520773e-01 -4.73784953e-01
3.05676043e-01 -9.39626899e-03 -3.50251794e-01 3.03518295e-01
-2.98442215e-01 5.66775620e-01 5.10356724e-01 -4.16450322e-01
-4.38433737e-01 -4.92875963e-01 -1.16644180e+00 1.74128264e-01
9.10329878e-01 -4.53765213e-01 -3.31649810e-01 5.46215653e-01
4.84292030e-01 -6.24726117e-01 -6.90204561e-01 5.44409811e-01
5.59010506e-01 -1.16028666e+00 1.61778736e+00 -1.94772512e-01
-9.95413214e-02 -1.02539635e+00 -2.78011650e-01 -9.37311113e-01
-1.41285956e-01 -1.94530115e-01 2.38115430e-01 1.24089718e+00
-9.94853526e-02 -6.77095056e-01 2.01196492e-01 3.39808077e-01
9.54332128e-02 -6.48090243e-02 -8.28012109e-01 -8.01203370e-01
-6.99022055e-01 -2.19055131e-01 8.02276373e-01 6.97521627e-01
-2.35801756e-01 1.21731445e-01 -4.04641092e-01 8.27323318e-01
6.61409199e-01 3.99492353e-01 1.06739879e+00 -1.13174891e+00
-8.87688547e-02 -2.35621572e-01 -5.20332098e-01 -8.18971574e-01
-3.18774194e-01 -4.53205884e-01 2.22129345e-01 -1.20044410e+00
1.94026470e-01 -3.68053436e-01 -1.14271179e-01 2.56090492e-01
-5.25103450e-01 6.71475053e-01 3.49089622e-01 3.63596380e-01
-9.52856123e-01 4.59607691e-01 8.91631603e-01 -1.30330980e-01
-2.26290882e-01 2.94758826e-02 -5.32662511e-01 6.54566944e-01
6.98872328e-01 -2.00274438e-01 8.29080120e-02 -4.06056166e-01
3.82277608e-01 -1.35065556e-01 7.21202314e-01 -1.15800989e+00
3.04595947e-01 -3.04409415e-01 7.56353140e-01 -9.34637666e-01
2.22355798e-01 -1.09780610e+00 4.60071815e-03 4.07872170e-01
1.16011463e-01 3.14686149e-01 2.33317062e-01 7.01332390e-01
-1.03159122e-01 2.03419089e-01 8.08219969e-01 -1.20285578e-01
-7.82703578e-01 2.80409276e-01 -1.43012077e-01 -2.41275042e-01
1.27950597e+00 -3.62648904e-01 -7.16009200e-01 2.79497448e-02
5.70649542e-02 7.88155049e-02 4.40416247e-01 4.22090858e-01
7.62043536e-01 -1.43243623e+00 -5.69027543e-01 4.06712800e-01
6.60181105e-01 1.46980733e-01 4.32767481e-01 1.26981258e+00
-2.36611649e-01 2.59332478e-01 -7.31033683e-02 -9.48568821e-01
-1.73493123e+00 5.76678693e-01 4.58457917e-01 3.12688142e-01
-2.55890548e-01 7.63124049e-01 -3.82211618e-02 -3.98050457e-01
-2.97172437e-03 -2.05079541e-01 -4.99188274e-01 7.80253932e-02
8.22131932e-01 5.36217749e-01 1.21267095e-01 -1.31513500e+00
-6.29981995e-01 1.14387691e+00 4.31901753e-01 1.59633324e-01
9.92090583e-01 -6.64778292e-01 -6.02820218e-02 2.02654049e-01
9.15768743e-01 4.43258584e-02 -1.00654387e+00 -3.26968670e-01
-4.51611578e-01 -8.52672160e-01 1.09972499e-01 -4.57685471e-01
-1.03349137e+00 6.69073939e-01 1.07199812e+00 1.52972654e-01
1.32346177e+00 -1.21558107e-01 5.56582212e-01 2.28691146e-01
3.91751617e-01 -1.07100093e+00 1.72165632e-01 2.66487122e-01
5.80675066e-01 -1.60624206e+00 2.25932673e-01 -5.52227616e-01
-7.91664302e-01 1.32841265e+00 6.86701953e-01 2.99512863e-01
1.21897213e-01 -4.23289947e-02 1.96387023e-01 -1.50250748e-01
8.74539316e-02 -8.40751290e-01 4.67475802e-01 7.82972813e-01
8.15920681e-02 5.36401272e-02 2.43443042e-01 2.34242707e-01
1.58380508e-01 -3.73999149e-01 1.58772364e-01 8.67234230e-01
-4.90343094e-01 -5.94601870e-01 -1.23495233e+00 -6.07278273e-02
-4.12869677e-02 2.79005785e-02 -4.32699710e-01 7.10421026e-01
4.48451966e-01 1.28498900e+00 -7.51707554e-02 -9.44711864e-01
5.22366107e-01 -4.38411385e-01 4.81675565e-01 -5.51045649e-02
-6.07004464e-01 5.72945587e-02 -3.75298291e-01 -6.46789670e-01
-8.81309688e-01 -6.02839291e-01 -9.80380356e-01 -1.99540362e-01
-8.87060404e-01 -3.94967735e-01 7.31962264e-01 1.04493833e+00
2.11011216e-01 2.12845668e-01 1.04840529e+00 -9.42896485e-01
-4.64965284e-01 -6.12177789e-01 -5.77629030e-01 4.21193242e-01
4.57661778e-01 -7.69923091e-01 -5.65433323e-01 -8.77145752e-02]
|
[9.763856887817383, -1.3968803882598877]
|
5f2d590b-1fc2-450c-83a6-23b8b2141318
|
out-of-domain-intent-detection-considering
|
2305.03237
| null |
https://arxiv.org/abs/2305.03237v1
|
https://arxiv.org/pdf/2305.03237v1.pdf
|
Out-of-Domain Intent Detection Considering Multi-turn Dialogue Contexts
|
Out-of-Domain (OOD) intent detection is vital for practical dialogue systems, and it usually requires considering multi-turn dialogue contexts. However, most previous OOD intent detection approaches are limited to single dialogue turns. In this paper, we introduce a context-aware OOD intent detection (Caro) framework to model multi-turn contexts in OOD intent detection tasks. Specifically, we follow the information bottleneck principle to extract robust representations from multi-turn dialogue contexts. Two different views are constructed for each input sample and the superfluous information not related to intent detection is removed using a multi-view information bottleneck loss. Moreover, we also explore utilizing unlabeled data in Caro. A two-stage training process is introduced to mine OOD samples from these unlabeled data, and these OOD samples are used to train the resulting model with a bootstrapping approach. Comprehensive experiments demonstrate that Caro establishes state-of-the-art performances on multi-turn OOD detection tasks by improving the F1-OOD score of over $29\%$ compared to the previous best method.
|
['Yongbin Li', 'Fei Huang', 'Binyuan Hui', 'Yinhe Zheng', 'Hao Lang']
|
2023-05-05
| null | null | null | null |
['intent-detection']
|
['natural-language-processing']
|
[ 3.04467320e-01 4.91898149e-01 -3.25134277e-01 -5.55055439e-01
-8.04953098e-01 -6.63300395e-01 8.24638784e-01 7.21807629e-02
-3.07393372e-01 4.66445923e-01 6.23349547e-01 -3.39873016e-01
5.29503345e-01 -2.18607992e-01 -7.67847598e-02 -1.74934313e-01
2.82802850e-01 5.67315161e-01 3.18311378e-02 -5.53241134e-01
4.63727444e-01 -1.90329835e-01 -1.21811688e+00 7.08832622e-01
1.21689427e+00 6.44507825e-01 3.50382984e-01 9.42652762e-01
-3.84022385e-01 8.95427644e-01 -1.04703939e+00 -3.82218927e-01
-3.96985672e-02 -7.53080606e-01 -1.09882224e+00 4.32225406e-01
1.56144902e-01 -7.83477664e-01 -1.68799847e-01 5.82047462e-01
5.67435920e-01 3.60972434e-01 8.67025673e-01 -9.92777169e-01
-2.06616625e-01 2.95334905e-01 -4.58835542e-01 2.05421537e-01
8.93264830e-01 2.51312912e-01 1.12764728e+00 -1.15111351e+00
7.67402649e-01 1.59453499e+00 3.00649405e-01 8.96901548e-01
-1.11401880e+00 -1.42685130e-01 4.12541747e-01 -8.91712084e-02
-5.41383982e-01 -6.23460233e-01 8.31501365e-01 -2.16084525e-01
1.29891992e+00 2.75385767e-01 4.00569022e-01 1.36653972e+00
1.10611811e-01 1.54483068e+00 1.04191911e+00 -7.81899691e-01
1.33046776e-01 3.74621660e-01 4.34159845e-01 7.24059939e-01
-4.65072840e-01 -4.09121782e-01 -7.09849119e-01 -1.48479149e-01
3.59341770e-01 -1.34690911e-01 -1.48813605e-01 -1.99915454e-01
-9.77807224e-01 1.11683416e+00 -1.35966279e-02 8.81530792e-02
-1.73269138e-01 -6.90035939e-01 9.42162097e-01 6.99259162e-01
7.86977708e-01 4.26760375e-01 -2.89149463e-01 -5.73117733e-01
-3.68574917e-01 3.52720201e-01 1.36883807e+00 1.02103794e+00
6.10997975e-01 -3.80340189e-01 -4.87890393e-01 1.41812909e+00
2.81178355e-01 2.52429813e-01 4.57383722e-01 -8.78934681e-01
1.03498876e+00 1.00596178e+00 1.54096574e-01 -7.15502381e-01
-4.47611898e-01 2.92942673e-01 -6.94482625e-01 -1.99025467e-01
3.95169824e-01 -3.42522264e-01 -6.53567791e-01 1.48416448e+00
7.13746011e-01 -7.54138470e-01 4.56316829e-01 7.81921089e-01
9.50147450e-01 7.10975707e-01 -1.17879003e-01 -4.45892185e-01
1.55238140e+00 -1.23667967e+00 -1.08455479e+00 -5.14743268e-01
1.00307763e+00 -9.53778446e-01 1.26489139e+00 3.42589140e-01
-7.85867572e-01 -4.34378386e-01 -8.53597403e-01 -3.46318513e-01
-9.06839892e-02 2.68583447e-01 4.66877699e-01 5.51873088e-01
-5.42820096e-01 3.74342836e-02 -4.10782218e-01 -3.87003392e-01
-6.32605627e-02 -3.47354375e-02 -1.45109281e-01 -2.64994860e-01
-1.39989316e+00 9.58646178e-01 3.35973501e-01 9.84367803e-02
-1.04911482e+00 -2.77173966e-01 -1.17503631e+00 -3.66636038e-01
8.33830416e-01 -3.45621496e-01 1.69271767e+00 -7.90917277e-01
-1.87833226e+00 8.96993995e-01 -4.32654053e-01 -3.67736906e-01
6.18757784e-01 -6.56801343e-01 -1.37582317e-01 2.54107714e-01
3.28610748e-01 5.67522764e-01 1.01953518e+00 -1.14631689e+00
-7.23458946e-01 -3.38066965e-01 5.52553415e-01 9.35585141e-01
-1.95374206e-01 2.81778723e-02 -5.04306316e-01 -3.72522175e-01
5.92238009e-02 -9.62240458e-01 2.75380127e-02 -5.89706540e-01
-7.96339810e-01 -7.96053648e-01 9.53035235e-01 -7.57126331e-01
1.35865808e+00 -1.97745705e+00 2.58930296e-01 -4.63836432e-01
3.27505618e-01 3.11889172e-01 -2.94838995e-02 8.44634593e-01
3.29713821e-01 -1.39001086e-01 -1.40319750e-01 -5.38544953e-01
1.34064451e-01 -3.07345949e-02 -2.00844437e-01 -4.79105003e-02
2.78294146e-01 4.63809341e-01 -9.26999688e-01 -5.06877720e-01
4.37866718e-01 -7.80472681e-02 -5.52382231e-01 9.39957678e-01
-5.24955392e-01 3.13655704e-01 -4.55918223e-01 4.89216417e-01
4.41328287e-01 -9.78835970e-02 5.82147837e-01 8.04450586e-02
-3.19712870e-02 8.21186364e-01 -9.03352499e-01 1.80135059e+00
-8.80266488e-01 4.28084135e-01 1.63563952e-01 -7.61282325e-01
9.23196316e-01 4.63405609e-01 -6.80347607e-02 -6.15433931e-01
-5.75397611e-02 8.22232068e-02 -4.82797325e-02 -7.15184569e-01
7.07599163e-01 -1.31475747e-01 -4.93082911e-01 8.27082336e-01
3.65667522e-01 -3.08634937e-01 3.61921489e-01 5.22355855e-01
9.06972110e-01 -4.00485769e-02 7.19955981e-01 -8.29468016e-04
7.99097836e-01 -9.74893719e-02 4.03542221e-01 8.35051775e-01
-4.47979033e-01 3.19477797e-01 1.00275588e+00 -5.61838269e-01
-6.98149025e-01 -5.89429736e-01 1.47240624e-01 1.41895306e+00
9.26403329e-02 -4.62558240e-01 -9.67404187e-01 -1.50378001e+00
-2.50200629e-01 1.07283854e+00 -4.87804681e-01 -9.37547013e-02
-4.90597874e-01 -6.03977621e-01 3.41765583e-01 1.21974848e-01
6.80329740e-01 -1.26887500e+00 -4.77404505e-01 4.87045974e-01
-7.42598355e-01 -1.06151557e+00 -6.74008310e-01 2.60123134e-01
-7.94188023e-01 -1.08570135e+00 -6.72521710e-01 -7.97670126e-01
3.52595657e-01 5.21974087e-01 1.13902640e+00 -1.22122109e-01
-1.57990918e-01 3.61419976e-01 -6.22187018e-01 -3.39359552e-01
-9.81357753e-01 3.49307507e-01 4.86252382e-02 -8.73822421e-02
7.83398986e-01 5.69051178e-03 -4.18970972e-01 3.54553252e-01
-5.46033800e-01 1.90176889e-01 5.93046844e-01 1.21552503e+00
-8.36149901e-02 -5.57127178e-01 9.01910543e-01 -1.17533898e+00
1.19384909e+00 -3.03767890e-01 -6.27329126e-02 1.86513603e-01
-4.27425534e-01 -9.71562602e-03 7.31133819e-01 -2.54738510e-01
-1.66471255e+00 -2.14023143e-01 -1.66085809e-01 -1.30511433e-01
-4.84012187e-01 2.94458896e-01 -2.68643796e-01 6.73918486e-01
4.68396455e-01 3.32670867e-01 1.75517499e-01 -5.14118195e-01
4.14478779e-01 1.24246573e+00 -1.68353125e-01 -4.28319842e-01
1.78719666e-02 5.53217381e-02 -8.22657526e-01 -1.28904331e+00
-1.17227602e+00 -8.62489998e-01 -6.31348848e-01 -4.15498644e-01
7.98282444e-01 -9.36784923e-01 -5.75022936e-01 5.94137490e-01
-1.23935795e+00 -3.82717907e-01 4.53832448e-02 3.19720328e-01
-5.82787097e-01 6.18038118e-01 -7.99610734e-01 -1.34929073e+00
-4.02148128e-01 -1.15023899e+00 1.14786911e+00 2.72457361e-01
-5.70001960e-01 -9.32073414e-01 1.84006885e-01 8.54504228e-01
-1.80631623e-01 -3.64303499e-01 9.65530097e-01 -1.30092716e+00
-1.17527589e-01 -1.55508488e-01 -6.57089576e-02 3.80570710e-01
2.92129308e-01 -5.51252246e-01 -1.05430543e+00 -1.80870712e-01
4.18661296e-01 -1.12922347e+00 7.14408457e-01 8.85588229e-02
5.62635183e-01 -4.92128432e-01 -2.44822800e-01 -1.79312408e-01
8.05846989e-01 3.61622006e-01 2.79210359e-01 8.99615809e-02
5.36938965e-01 1.02775764e+00 1.23943579e+00 6.45483077e-01
6.57340884e-01 6.54359460e-01 1.63902089e-01 9.32427943e-02
1.64675206e-01 -3.51032615e-01 5.26003778e-01 1.00124013e+00
3.64757478e-01 -6.39557660e-01 -6.35240197e-01 5.50350487e-01
-1.73254156e+00 -8.05951178e-01 1.18544899e-01 1.99110067e+00
1.01737201e+00 4.21197027e-01 3.43621790e-01 -1.64124057e-01
8.29928815e-01 6.45779848e-01 -6.67921245e-01 -7.97588885e-01
4.01295066e-01 -4.12027299e-01 -3.75675887e-01 6.60040915e-01
-1.27344775e+00 1.03119302e+00 5.77139902e+00 7.26778507e-01
-5.49960434e-01 1.98354553e-02 8.67140353e-01 1.36753291e-01
-1.91174805e-01 -4.90489863e-02 -1.07009792e+00 2.75461346e-01
7.49954224e-01 1.68743327e-01 1.07383072e-01 1.02529800e+00
3.19840759e-01 -4.18023974e-01 -1.05683661e+00 6.11245751e-01
4.24301624e-01 -6.39609814e-01 -2.61936542e-02 -6.17450383e-03
4.74899977e-01 -4.56841201e-01 -4.21109974e-01 8.63929749e-01
3.57047737e-01 -5.28899789e-01 5.13569973e-02 9.68691707e-02
6.09418511e-01 -7.75943041e-01 6.96536541e-01 8.44368160e-01
-8.51504683e-01 5.32201678e-02 -2.58333772e-01 -2.45549064e-02
1.02361239e-01 5.02989292e-01 -1.42552590e+00 6.06627882e-01
2.88191468e-01 5.49654603e-01 -1.86127558e-01 2.12907687e-01
-2.82537162e-01 4.84288454e-01 -1.80235326e-01 -4.42347229e-01
3.93162280e-01 -2.19461307e-01 7.64046550e-01 1.26623392e+00
-2.32660756e-01 7.11432919e-02 4.89162594e-01 5.38193703e-01
-2.66957849e-01 1.83511183e-01 -9.00647581e-01 6.77203061e-03
4.56370205e-01 1.23716319e+00 -4.04494107e-01 -4.63856637e-01
-5.18262565e-01 1.36434817e+00 5.35073102e-01 9.76834074e-03
-5.40727794e-01 -5.40683627e-01 4.66938227e-01 -4.82754350e-01
1.20560333e-01 1.56670526e-01 1.37022689e-01 -1.30681157e+00
-7.27859931e-03 -1.35813797e+00 5.14521122e-01 -1.73775986e-01
-1.34397662e+00 4.37553674e-01 -9.36478898e-02 -1.35812211e+00
-7.29935229e-01 -4.02343333e-01 -6.43201530e-01 7.20176637e-01
-1.32757092e+00 -9.05023456e-01 -4.27532718e-02 3.58686186e-02
1.67823946e+00 -6.80851862e-02 9.97694254e-01 -2.48029679e-01
-7.36660302e-01 5.63525200e-01 -9.70842317e-03 2.92857975e-01
1.01029992e+00 -1.53374648e+00 4.52486426e-01 5.28282642e-01
-2.69879073e-01 5.96932828e-01 5.67828715e-01 -7.14846253e-01
-1.35500586e+00 -7.39669085e-01 1.10128760e+00 -4.33537930e-01
1.86323941e-01 -5.75635970e-01 -8.82271588e-01 6.53493762e-01
5.88585019e-01 -4.71352607e-01 9.41914141e-01 5.19957542e-01
-2.03276724e-01 4.07072634e-01 -1.12934959e+00 7.19445109e-01
9.20707285e-01 -5.53748548e-01 -1.28130543e+00 3.85849625e-01
8.17289412e-01 -6.39759839e-01 -7.84451544e-01 1.29559606e-01
4.79367614e-01 -1.22662878e+00 7.32822239e-01 -6.26046598e-01
5.49192607e-01 2.63636112e-01 1.52214676e-01 -1.45517421e+00
2.76699960e-01 -8.79980028e-01 -3.89549285e-01 1.39494479e+00
2.76238292e-01 -3.31746340e-01 6.33083701e-01 4.95819449e-01
-2.85868168e-01 -6.22880757e-01 -8.81077170e-01 -4.81390268e-01
-2.54956812e-01 -1.04242601e-01 5.13198748e-02 7.62703896e-01
6.59057915e-01 1.18856537e+00 -5.95467329e-01 -2.43647262e-01
4.90405947e-01 3.47696751e-01 1.11019874e+00 -7.92132854e-01
-1.20629475e-01 -2.19204035e-02 3.08658838e-01 -1.78097093e+00
1.53975621e-01 -4.10672903e-01 4.82932270e-01 -1.35982025e+00
1.46895319e-01 -1.70320749e-01 2.99876064e-01 7.20971897e-02
-7.23155379e-01 -3.27857286e-01 2.33512983e-01 2.65258551e-01
-9.81855333e-01 8.78015339e-01 1.26872277e+00 -8.02287534e-02
-6.32927358e-01 2.92559117e-01 -5.25808930e-01 8.63097668e-01
6.69106126e-01 -2.92527139e-01 -5.72488666e-01 -1.45261614e-02
-4.17154759e-01 6.80094957e-01 -2.45076165e-01 -6.52707279e-01
-2.05549031e-01 -5.92188314e-02 7.21780956e-02 -8.63494694e-01
4.82952774e-01 -3.75061125e-01 -9.49915528e-01 4.65695828e-01
-7.67017424e-01 -3.02366525e-01 -8.61518756e-02 8.84876251e-01
-2.50194609e-01 -6.85464740e-01 5.27248025e-01 -5.64641058e-01
-7.29449093e-01 -2.29831591e-01 -9.30533826e-01 7.59423971e-01
8.74615073e-01 7.49569153e-03 -2.19253093e-01 -5.98947406e-01
-5.74525654e-01 4.70748931e-01 1.63216025e-01 5.84950268e-01
5.58771670e-01 -8.98213625e-01 -6.02053404e-01 1.70439631e-01
5.21123469e-01 1.29904434e-01 4.58620250e-01 5.41637719e-01
-1.84153646e-01 6.17765963e-01 1.78405941e-01 -7.21300840e-01
-1.38233709e+00 4.33115184e-01 1.44814983e-01 -8.54522407e-01
-6.72075868e-01 7.82827318e-01 3.96196336e-01 -1.01808143e+00
3.40168059e-01 2.68357550e-03 -5.51712513e-01 4.98590678e-01
7.11277187e-01 3.00401866e-01 -5.92205822e-02 -2.99963772e-01
-2.14395285e-01 -1.04251206e-01 -7.87065983e-01 -3.36062551e-01
8.99654567e-01 -5.64265311e-01 2.99846172e-01 8.31197560e-01
1.11461055e+00 -8.40414315e-02 -1.32594192e+00 -2.57793009e-01
4.90896665e-02 -3.66329342e-01 -5.10808289e-01 -9.83546317e-01
-1.48707584e-01 9.90721881e-01 2.73766458e-01 7.41671741e-01
6.97660208e-01 -1.52707234e-01 8.26854765e-01 7.05460608e-01
1.44539416e-01 -1.38352716e+00 5.25975943e-01 9.82875466e-01
1.00728548e+00 -1.74295950e+00 -2.23238304e-01 -4.86897111e-01
-1.31292164e+00 1.19812882e+00 1.10987771e+00 8.45656618e-02
1.26773134e-01 -1.83387309e-01 3.97163242e-01 -1.62393495e-01
-1.14873052e+00 3.24800611e-02 1.22456498e-01 5.18896043e-01
6.39787555e-01 -1.53304249e-01 -3.87250125e-01 3.85886669e-01
3.13040793e-01 -3.16408992e-01 6.45964622e-01 1.12999499e+00
-5.48594177e-01 -1.07688081e+00 -1.64539531e-01 5.77908754e-01
-3.04031372e-01 -1.30615592e-01 -9.25795496e-01 8.03146064e-01
-5.63633859e-01 1.23609698e+00 -2.20557168e-01 -3.48449200e-01
4.54925120e-01 7.30044007e-01 6.03823587e-02 -1.02740419e+00
-8.70373547e-01 2.24200234e-01 7.26293266e-01 -3.68542820e-01
-3.67374182e-01 -6.77084088e-01 -9.02667046e-01 -5.97662739e-02
-5.72617948e-01 2.37284020e-01 3.50809485e-01 1.10497725e+00
4.14378077e-01 4.23085153e-01 1.10987914e+00 -7.66928613e-01
-9.05272603e-01 -1.51156032e+00 -1.95007816e-01 4.37486470e-01
3.98346782e-01 -5.85632265e-01 -5.34347594e-01 -2.29072988e-01]
|
[12.663859367370605, 7.836772441864014]
|
47cd8347-ef83-4755-bd8a-5174bc488e57
|
a-novel-distributed-representation-of-news
|
2005.11706
| null |
https://arxiv.org/abs/2005.11706v2
|
https://arxiv.org/pdf/2005.11706v2.pdf
|
A Novel Distributed Representation of News (DRNews) for Stock Market Predictions
|
In this study, a novel Distributed Representation of News (DRNews) model is developed and applied in deep learning-based stock market predictions. With the merit of integrating contextual information and cross-documental knowledge, the DRNews model creates news vectors that describe both the semantic information and potential linkages among news events through an attributed news network. Two stock market prediction tasks, namely the short-term stock movement prediction and stock crises early warning, are implemented in the framework of the attention-based Long Short Term-Memory (LSTM) network. It is suggested that DRNews substantially enhances the results of both tasks comparing with five baselines of news embedding models. Further, the attention mechanism suggests that short-term stock trend and stock market crises both receive influences from daily news with the former demonstrates more critical responses on the information related to the stock market {\em per se}, whilst the latter draws more concerns on the banking sector and economic policies.
|
['Ye Ma', 'Peiwan Wang', 'Lu Zong']
|
2020-05-24
| null | null | null | null |
['stock-market-prediction']
|
['time-series']
|
[-8.44648123e-01 2.28190362e-01 -6.83244824e-01 -1.74902275e-01
-2.92036712e-01 -3.55390221e-01 1.20344174e+00 3.14335942e-01
-3.06008905e-01 7.10414529e-01 1.50874829e+00 -3.49377126e-01
1.68082956e-02 -1.26537490e+00 -5.58118880e-01 -2.50309706e-01
-1.66570410e-01 2.57413387e-01 1.67856902e-01 -5.71156204e-01
6.74053490e-01 1.61779076e-01 -1.07789266e+00 3.10161084e-01
1.01533361e-01 1.37857962e+00 -6.74792472e-03 7.87231401e-02
-7.56010592e-01 1.87181914e+00 -5.55108488e-01 -6.42416060e-01
1.56914055e-01 -2.03424230e-01 -5.16567707e-01 -4.42560852e-01
1.90750286e-02 -4.57434058e-01 -7.61273384e-01 9.75590646e-01
4.08378720e-01 7.14646280e-02 5.54458857e-01 -7.87446916e-01
-1.53599632e+00 1.44117260e+00 -2.42709577e-01 1.05783713e+00
-6.92366287e-02 -2.04747424e-01 1.60606349e+00 -1.15487909e+00
7.31846690e-01 8.81956100e-01 9.49918151e-01 -9.92048681e-02
-7.14073956e-01 -6.83771729e-01 4.94908839e-01 2.92539775e-01
-4.31968093e-01 -1.51984915e-01 8.98032129e-01 -7.21693039e-01
1.48237777e+00 -2.12034006e-02 1.07458687e+00 1.43539393e+00
7.18165994e-01 6.92828774e-01 8.49869549e-01 7.59415934e-03
2.07366481e-01 2.68715918e-01 4.49371219e-01 1.75107419e-02
4.38075989e-01 4.80586261e-01 -8.46469104e-01 -1.92899048e-01
7.08590448e-01 4.51985627e-01 -1.40480592e-03 2.62426376e-01
-1.19902444e+00 1.23224819e+00 4.56487566e-01 7.86940634e-01
-9.51073587e-01 1.36137918e-01 6.23132110e-01 5.01932561e-01
1.35138083e+00 5.56856990e-01 -9.62459087e-01 -2.58703738e-01
-9.62766290e-01 2.61301577e-01 1.06391311e+00 3.75079602e-01
2.63597846e-01 6.91008866e-01 -1.87085420e-01 4.90828902e-01
4.16949123e-01 4.44114208e-01 9.04822052e-01 -1.24277242e-01
4.84990150e-01 4.53068286e-01 1.33851230e-01 -1.34893513e+00
-4.63814944e-01 -9.09056425e-01 -4.02501225e-01 1.19869642e-01
-7.24203512e-02 -4.37320024e-01 -5.87344468e-01 1.24704838e+00
-1.86099261e-01 7.13558257e-01 4.82574314e-01 4.31696862e-01
7.85963297e-01 1.32222331e+00 1.30575970e-01 -2.88743317e-01
1.21127188e+00 -9.28779006e-01 -9.78942871e-01 -4.73499149e-01
2.93762475e-01 -5.11464834e-01 2.98412353e-01 -3.91403228e-01
-1.18477392e+00 -2.72956580e-01 -8.96752954e-01 1.50456369e-01
-9.51394260e-01 -4.37855631e-01 2.55674273e-01 -1.22562736e-01
-9.83959913e-01 5.04939139e-01 -5.54302871e-01 -5.33756837e-02
1.49168614e-02 -2.70453393e-01 2.70409703e-01 6.04779363e-01
-1.81685519e+00 1.53898442e+00 5.54486692e-01 -8.50113481e-02
-4.16446596e-01 -9.89668548e-01 -9.57830131e-01 4.40970629e-01
-1.17036052e-01 -5.09763777e-01 1.13880646e+00 -7.38886893e-01
-1.44306719e+00 2.66760409e-01 4.74436730e-01 -1.06772804e+00
1.01722121e-01 -2.88389117e-01 -7.95629680e-01 -8.69104862e-02
1.86863601e-01 1.43874168e-01 6.60692036e-01 -3.59187812e-01
-8.40637505e-01 -6.28275126e-02 -3.23218197e-01 1.37529865e-01
-3.98366570e-01 3.61143410e-01 3.97438884e-01 -1.44314921e+00
-2.62303859e-01 -2.54159898e-01 5.46486229e-02 -6.93046987e-01
8.27975869e-02 -3.97676170e-01 6.83332503e-01 -1.04915106e+00
1.41220236e+00 -2.10241580e+00 -2.47536987e-01 9.67550501e-02
-1.84058789e-02 -1.02044113e-01 1.60321835e-02 9.36342239e-01
-3.90833616e-01 1.51796013e-01 3.54481637e-01 2.18560081e-02
4.66170400e-01 -6.77038431e-02 -1.39711034e+00 3.50467950e-01
1.43737748e-01 1.28635728e+00 -6.96085930e-01 3.38133246e-01
5.11149457e-03 3.60846758e-01 -2.28189319e-01 -1.57319531e-01
-3.26120049e-01 -2.16264799e-01 -3.39268059e-01 4.54229653e-01
2.34904826e-01 -3.08800191e-01 -2.10194975e-01 1.19144961e-01
-5.88335633e-01 9.89266276e-01 -7.24349082e-01 8.16044807e-01
-1.78349122e-01 7.36866236e-01 -5.40833473e-01 -7.86143780e-01
1.07819819e+00 7.27428019e-01 4.87058997e-01 -1.17465770e+00
2.46797432e-03 1.81970820e-01 -2.86718279e-01 -1.48500577e-01
7.33788610e-01 -4.91674304e-01 -4.38941941e-02 9.10327077e-01
3.51301506e-02 4.20446992e-01 -1.30035922e-01 1.67427748e-01
8.54320407e-01 -1.44312158e-01 3.49100173e-01 -3.82522821e-01
-2.47624382e-01 -8.20952579e-02 6.36604190e-01 3.86959463e-01
-2.58966070e-02 1.18582428e-01 6.96150064e-01 -8.87109160e-01
-1.07401764e+00 -6.66625381e-01 -1.89864933e-01 1.20303488e+00
-3.26616913e-01 -2.87065595e-01 2.70104986e-02 -4.69017178e-01
6.42708719e-01 1.26464808e+00 -9.98984933e-01 3.34575251e-02
-3.31315488e-01 -9.14074659e-01 3.21712106e-01 1.00074244e+00
2.02674687e-01 -1.47247863e+00 -5.73962152e-01 6.34879351e-01
2.67779171e-01 -7.02801943e-01 -4.25505221e-01 1.73742607e-01
-6.63867950e-01 -7.53243983e-01 -1.03960562e+00 -5.86206377e-01
-5.46019599e-02 -3.31447363e-01 1.18689573e+00 -4.46558535e-01
4.74819571e-01 3.40125412e-01 -4.21188205e-01 -7.55632401e-01
-1.99710786e-01 -1.61189258e-01 2.52633318e-02 -6.69355784e-03
5.80090702e-01 -4.87032771e-01 -4.67803389e-01 -1.81420177e-01
-7.70995796e-01 -3.48782450e-01 3.56888384e-01 8.61592472e-01
9.75657552e-02 7.34814629e-02 1.35341978e+00 -6.05722666e-01
8.73789191e-01 -1.30408072e+00 -6.41910672e-01 1.71649437e-02
-1.23176587e+00 -5.17293923e-02 -1.18464954e-01 -2.74508059e-01
-1.35933685e+00 -1.00133777e+00 -1.14502691e-01 3.31686176e-02
3.59284818e-01 1.48993707e+00 6.87597811e-01 6.47607625e-01
2.16232553e-01 4.77451831e-01 -4.35776748e-02 -4.86652166e-01
2.49420688e-01 2.58765429e-01 1.79497883e-01 1.81157038e-01
5.21516562e-01 1.62530541e-01 -7.82345116e-01 -6.74395084e-01
-9.93055642e-01 -2.86461443e-01 -8.15551635e-03 -2.46446878e-02
9.07587111e-01 -1.25460887e+00 6.65396005e-02 6.46181226e-01
-1.07157910e+00 -9.02684778e-02 -6.59545362e-01 8.78876984e-01
-2.54495502e-01 -2.98148394e-01 -1.42105818e+00 -8.96864176e-01
-4.28219765e-01 -4.60072339e-01 4.10097301e-01 -9.34543386e-02
-3.06132257e-01 -1.72055256e+00 4.85655993e-01 -1.90101817e-01
1.02273571e+00 1.84144408e-01 1.01407981e+00 -1.58787656e+00
-3.54229271e-01 -3.27050924e-01 -1.68286070e-01 2.75688320e-01
2.36604288e-01 -2.61567295e-01 -7.15524614e-01 4.14625444e-02
2.96638042e-01 -7.09042102e-02 1.16847515e+00 7.45327115e-01
-2.42517337e-01 -6.59458995e-01 1.00687407e-01 3.17476988e-01
1.19051433e+00 5.40085196e-01 4.33847874e-01 9.13097918e-01
2.47602418e-01 4.72214311e-01 2.15343133e-01 7.55294621e-01
5.90449393e-01 6.21158853e-02 9.02277082e-02 2.08968036e-02
2.54807383e-01 -4.71896112e-01 7.97747374e-01 1.11855674e+00
2.09074572e-01 5.94501980e-02 -8.66529644e-01 5.85672855e-01
-1.83006275e+00 -1.55186105e+00 1.76828638e-01 1.36124754e+00
6.98259950e-01 5.00064552e-01 1.10160016e-01 -3.52145165e-01
5.46294987e-01 9.71550345e-01 -6.11585975e-01 -1.68560073e-01
-4.59234118e-01 -1.69625338e-02 6.53727055e-01 3.16464812e-01
-1.02432394e+00 8.22970569e-01 7.23891735e+00 2.09743038e-01
-1.23080695e+00 1.49188377e-02 7.38313198e-01 -5.33206239e-02
-6.84996605e-01 -1.75503537e-01 -9.94672179e-01 7.36712098e-01
1.24698865e+00 -6.10085905e-01 -1.97848588e-01 9.42255259e-01
1.46893799e-01 2.41300300e-01 -6.97921336e-01 3.30815613e-01
7.12877735e-02 -2.07165074e+00 9.38620120e-02 1.44033805e-01
7.32216895e-01 4.91876334e-01 3.66216183e-01 5.32805502e-01
4.66337353e-01 -4.64033514e-01 1.27983022e+00 8.73738408e-01
1.74314901e-01 -8.62555206e-01 1.16174304e+00 2.75834709e-01
-1.06561291e+00 -4.19547647e-01 -4.66215193e-01 -4.14710313e-01
4.50563043e-01 5.63517928e-01 -5.42672813e-01 1.74878508e-01
5.86876988e-01 1.40851402e+00 -2.80892968e-01 5.09429634e-01
-1.47279724e-01 9.33937490e-01 1.37685195e-01 1.90686453e-02
7.42394626e-01 -2.12099880e-01 5.12290776e-01 1.15210664e+00
5.00655890e-01 2.16340587e-01 -2.56139576e-01 8.41500401e-01
-3.63318957e-02 1.07425921e-01 -9.26354527e-01 -5.05327702e-01
2.93259680e-01 5.12623012e-01 -6.70842052e-01 -5.88151753e-01
-8.66668642e-01 4.45460141e-01 2.89124865e-02 5.40035963e-01
-5.76253831e-01 -1.19731501e-01 8.89822423e-01 1.74542636e-01
8.08290720e-01 1.77295208e-01 -3.73914391e-01 -1.37346208e+00
-1.96029544e-01 -3.67728025e-01 5.75027287e-01 -7.48761892e-01
-1.75397885e+00 6.86890304e-01 -1.47747040e-01 -7.61021495e-01
-5.15703201e-01 -3.85769695e-01 -1.08347976e+00 9.03540015e-01
-1.82853198e+00 -8.88943493e-01 8.95361304e-01 2.42080644e-01
6.08827055e-01 -7.59175003e-01 5.87527394e-01 -9.97607224e-03
-4.66058940e-01 -1.88429192e-01 4.85896379e-01 4.31276321e-01
4.40312922e-01 -1.28050220e+00 1.07422864e+00 8.57989788e-01
2.40080789e-01 3.64216596e-01 4.04338568e-01 -1.27832329e+00
-6.49915099e-01 -1.16542554e+00 1.49563193e+00 -4.46367621e-01
1.64350879e+00 1.67497829e-01 -1.06394291e+00 1.19898510e+00
8.03843021e-01 -3.92608047e-01 8.77081573e-01 1.91470593e-01
-5.62043250e-01 3.87878902e-02 -6.27393961e-01 1.91064000e-01
2.13419631e-01 -6.61364079e-01 -1.53279543e+00 6.77278116e-02
1.05145049e+00 1.73856005e-01 -9.39019501e-01 -1.18578844e-01
3.39046866e-01 -8.76620233e-01 8.45122159e-01 -7.43435502e-01
4.98367310e-01 3.49691182e-01 -1.05361700e-01 -1.63824046e+00
-7.17707217e-01 -2.66077310e-01 -3.22276622e-01 1.17690277e+00
8.23594987e-01 -1.19649947e+00 3.42008740e-01 6.86113119e-01
-3.07018965e-01 -4.95666057e-01 -8.72620106e-01 -6.28038764e-01
2.88345188e-01 -3.83657694e-01 8.53959799e-01 1.46567559e+00
5.48812211e-01 1.87940612e-01 -5.40193021e-01 1.04119694e-02
2.86866929e-02 1.59657151e-01 -1.24425635e-01 -1.36369801e+00
-6.58324212e-02 -7.91961610e-01 -2.95212537e-01 -6.45759940e-01
4.82730091e-01 -8.51582289e-01 -6.02692544e-01 -1.62408733e+00
-1.56528294e-01 2.17349410e-01 -9.44716454e-01 2.10987836e-01
2.33016953e-01 -2.08847165e-01 3.54180068e-01 3.07831526e-01
-1.39639243e-01 8.25656533e-01 6.89504087e-01 -1.37282729e-01
-6.83902502e-02 1.06574126e-01 -7.71700263e-01 8.46391678e-01
7.70622551e-01 -4.93581086e-01 -1.30868509e-01 -3.41143638e-01
9.59379077e-01 3.65174353e-01 1.35405973e-01 -4.01912659e-01
4.30769116e-01 -1.44968510e-01 4.82793748e-01 -8.30494821e-01
1.87027439e-01 -5.55359066e-01 9.13503841e-02 5.50253153e-01
-6.02008760e-01 7.50877559e-01 2.63732940e-01 8.31802905e-01
-7.34108090e-01 5.47784343e-02 3.89318824e-01 -1.71178371e-01
-7.62788057e-01 1.95715740e-01 -8.45467269e-01 3.64206024e-02
8.74782085e-01 -7.59611651e-02 -6.06285512e-01 -7.98302114e-01
-7.87071764e-01 9.54706296e-02 -1.47043139e-01 6.32802486e-01
6.87838256e-01 -1.47140598e+00 -9.73922491e-01 2.75225282e-01
-1.17156096e-01 -7.28708625e-01 -5.17538786e-02 6.59057617e-01
-3.88590753e-01 7.42733002e-01 -1.73524782e-01 5.10036826e-01
-2.28917122e-01 6.95331514e-01 2.04225466e-01 -3.48390520e-01
-1.13769090e+00 7.84006655e-01 3.48684541e-03 1.84754893e-01
2.72785962e-01 -8.31180215e-01 -7.40955710e-01 1.00320470e+00
8.10700893e-01 5.58939338e-01 -2.31771216e-01 -6.95889771e-01
-1.16990402e-01 1.52194813e-01 -1.33176640e-01 -2.76221514e-01
2.08658552e+00 -1.31255686e-01 -2.45679453e-01 9.57577229e-01
8.91747773e-01 -5.72876595e-02 -1.09339666e+00 -5.52519500e-01
8.58159244e-01 8.65714177e-02 5.08606374e-01 -9.10562217e-01
-1.21648121e+00 3.70712250e-01 1.11860566e-01 7.45472908e-01
5.47184587e-01 1.54664621e-01 1.18476820e+00 2.43938982e-01
-2.45785892e-01 -1.36806631e+00 -8.41855556e-02 9.54867244e-01
1.11516941e+00 -9.23733830e-01 -1.57880813e-01 5.80000818e-01
-9.31067288e-01 1.20140541e+00 1.30240366e-01 -4.15192276e-01
1.36887801e+00 5.07538497e-01 2.86048234e-01 -5.24934351e-01
-1.23447049e+00 -6.49202988e-02 2.81536877e-01 -6.64043203e-02
2.56370872e-01 -1.77774176e-01 -4.15723845e-02 8.58617008e-01
-1.19803160e-01 -2.25674421e-01 5.71306407e-01 8.36350858e-01
-6.43830895e-01 -5.82776606e-01 -1.15711704e-01 6.08262599e-01
-9.53958809e-01 -6.19046032e-01 -3.15440744e-01 7.66790986e-01
-3.57295424e-01 4.43879128e-01 8.19372237e-01 -1.75567269e-01
1.07706077e-01 5.75149655e-01 -7.46520817e-01 -7.22235858e-01
-8.63815069e-01 2.60552943e-01 9.16787982e-02 -2.75647134e-01
-4.06057715e-01 -7.94148088e-01 -9.82910097e-01 -3.05894136e-01
-8.70720595e-02 2.95852154e-01 5.38591743e-01 1.00234056e+00
5.54460049e-01 5.38313687e-01 4.88546550e-01 -4.80343968e-01
-8.68216395e-01 -1.21027863e+00 -1.19934535e+00 5.65067567e-02
5.44931710e-01 -4.56273496e-01 -4.63263273e-01 -1.11205757e-01]
|
[4.403802871704102, 4.2868971824646]
|
3ec6ef3f-a7bb-4e21-b15e-5386263b0c84
|
maevi-motion-aware-event-based-video-frame
|
2303.02025
| null |
https://arxiv.org/abs/2303.02025v1
|
https://arxiv.org/pdf/2303.02025v1.pdf
|
MAEVI: Motion Aware Event-Based Video Frame Interpolation
|
Utilization of event-based cameras is expected to improve the visual quality of video frame interpolation solutions. We introduce a learning-based method to exploit moving region boundaries in a video sequence to increase the overall interpolation quality.Event cameras allow us to determine moving areas precisely; and hence, better video frame interpolation quality can be achieved by emphasizing these regions using an appropriate loss function. The results show a notable average \textit{PSNR} improvement of $1.3$ dB for the tested data sets, as well as subjectively more pleasing visual results with less ghosting and blurry artifacts.
|
['A. Aydin Alatan', 'Onur Selim Kılıç', 'Ahmet Akman']
|
2023-03-03
| null | null | null | null |
['video-frame-interpolation']
|
['computer-vision']
|
[ 6.81626722e-02 -4.28064793e-01 -1.60936534e-01 -2.96069413e-01
-5.48358798e-01 -3.16055976e-02 2.74963826e-01 6.75756717e-03
-3.74003202e-01 1.10258329e+00 7.13581890e-02 -3.32296104e-03
9.84086543e-02 -5.14708936e-01 -7.36986816e-01 -7.90830672e-01
-4.56226796e-01 -6.24943554e-01 6.23053730e-01 2.48938799e-01
4.50118154e-01 3.89922589e-01 -1.35627854e+00 4.30493534e-01
1.07335937e+00 1.25097990e+00 3.00885379e-01 6.50607884e-01
6.01485334e-02 1.13246787e+00 -6.23705983e-01 -2.45872557e-01
4.66293782e-01 -4.78982806e-01 -3.09846997e-01 3.03516418e-01
5.63975930e-01 -6.64962709e-01 -4.28306460e-01 1.13461030e+00
2.46944875e-01 2.75815338e-01 3.39827925e-01 -8.80896270e-01
-2.25097880e-01 1.53106719e-01 -9.00742471e-01 4.97718632e-01
5.13413727e-01 3.76755714e-01 6.59166455e-01 -9.10700917e-01
5.08826911e-01 1.05701995e+00 4.85446543e-01 2.05409586e-01
-1.17126262e+00 -7.02990830e-01 -1.87614441e-01 7.88407922e-01
-1.56909442e+00 -5.87912619e-01 1.00416148e+00 -3.96486223e-01
3.51148784e-01 4.61863935e-01 7.09527433e-01 5.44119179e-01
4.99746501e-01 6.76425695e-01 8.94673586e-01 -5.30981719e-01
1.89410403e-01 -1.72961205e-01 -3.76104355e-01 6.08672976e-01
1.70754850e-01 2.60068953e-01 -5.64954102e-01 2.90282637e-01
1.27344584e+00 1.18335880e-01 -8.06869447e-01 -2.64646828e-01
-1.28287840e+00 4.04317707e-01 4.79173690e-01 1.22819930e-01
-6.28445268e-01 2.62722135e-01 5.10796666e-01 -1.04457475e-02
4.65434641e-01 7.94003084e-02 8.85690302e-02 -1.41349405e-01
-1.34934974e+00 1.41523570e-01 2.40336493e-01 1.02050543e+00
7.54112065e-01 6.11558557e-01 -3.53725284e-01 7.81692505e-01
2.92482018e-01 4.19271916e-01 8.08323908e-04 -1.60210431e+00
5.16685307e-01 1.41942114e-01 5.48694491e-01 -1.08718431e+00
1.43703178e-01 -1.97112724e-01 -9.31124389e-01 7.45053709e-01
4.58810538e-01 -2.24932954e-01 -5.86774826e-01 1.18700612e+00
-6.02469519e-02 5.92657745e-01 -1.90479815e-01 9.79971111e-01
3.79833400e-01 1.00138581e+00 6.72894493e-02 -6.66478336e-01
1.09146726e+00 -8.13696682e-01 -1.18427026e+00 -1.10535305e-02
-1.04292899e-01 -1.02815545e+00 8.90379727e-01 4.69555438e-01
-1.61622810e+00 -9.50943232e-01 -1.29528809e+00 1.73176721e-01
4.61554557e-01 -4.57245670e-02 2.68582970e-01 5.15436113e-01
-9.54422176e-01 4.56984788e-01 -8.51490021e-01 1.70701668e-01
5.14521003e-01 2.61044472e-01 8.90177414e-02 -1.69196144e-01
-9.47160184e-01 6.86117828e-01 3.90836447e-01 1.29399195e-01
-6.68205023e-01 -9.13567483e-01 -9.01333749e-01 -8.26310553e-03
1.91818461e-01 -3.85406375e-01 9.93827045e-01 -1.19192135e+00
-1.21401942e+00 4.22601253e-01 -4.16494071e-01 -5.84876776e-01
7.53385305e-01 -3.41833174e-01 -5.89340746e-01 4.87531334e-01
-2.43179768e-01 6.84289277e-01 1.07928967e+00 -1.51712346e+00
-9.83221233e-01 1.16881825e-01 -9.50398855e-03 1.22420914e-01
-3.62747639e-01 1.54885307e-01 -5.21890461e-01 -9.55954373e-01
-1.67241991e-01 -5.11783659e-01 -2.48892367e-01 5.67727864e-01
1.68667093e-01 3.08617324e-01 1.09664810e+00 -1.10559690e+00
1.52730608e+00 -2.22230315e+00 -2.31833175e-01 -1.81361109e-01
2.05490753e-01 4.66219336e-01 2.95017749e-01 2.90153455e-02
1.05304942e-01 -2.05600843e-01 -1.85191929e-01 1.31434843e-01
-6.35661364e-01 1.22246360e-02 -3.76498960e-02 4.60067332e-01
8.58681574e-02 4.05361593e-01 -8.03521454e-01 -6.73945546e-01
7.24316418e-01 7.90545523e-01 -6.04197741e-01 2.54861563e-01
7.49278739e-02 5.57061791e-01 -8.69799778e-02 6.49677813e-01
1.02222633e+00 -9.71221402e-02 -1.04716271e-01 -4.68501985e-01
-2.75378108e-01 -1.25516459e-01 -1.32492959e+00 1.37397695e+00
-6.40595257e-01 1.34125125e+00 1.62349224e-01 -4.66649234e-01
9.19971347e-01 4.44822460e-01 6.34906530e-01 -5.77535808e-01
-9.68790576e-02 -7.86247402e-02 -5.52580357e-02 -5.68309128e-01
6.33848786e-01 8.27515423e-02 5.31432569e-01 -3.11532110e-01
-4.42188382e-01 3.32638711e-01 1.93918422e-01 -2.23824769e-01
6.94476366e-01 4.08392735e-02 2.52146244e-01 -3.76745015e-01
8.36791694e-01 -3.93268019e-01 8.63455832e-01 2.49810025e-01
-4.20372009e-01 7.28693366e-01 -1.92282274e-02 -6.12686932e-01
-1.30457902e+00 -1.31744421e+00 -2.19769225e-01 4.39496011e-01
5.80629051e-01 -6.89601824e-02 -6.74309313e-01 -5.08496948e-02
-3.71537030e-01 8.20165277e-01 -5.18522933e-02 5.72400838e-02
-1.06018686e+00 -3.46067011e-01 1.29647538e-01 6.51493907e-01
1.00844169e+00 -8.64646912e-01 -7.25809455e-01 3.33725572e-01
-3.91387045e-01 -1.32127106e+00 -6.67971671e-01 -3.98336083e-01
-1.13108206e+00 -8.07445765e-01 -1.12755311e+00 -7.53168344e-01
7.53594935e-01 3.49446654e-01 1.07808256e+00 9.26927254e-02
-2.15870991e-01 -1.67141765e-01 -1.87488347e-01 5.33705167e-02
-4.27368104e-01 -7.38312542e-01 -3.00012410e-01 2.09969759e-01
-7.07458239e-03 -3.32076520e-01 -1.09365809e+00 5.34195602e-01
-8.80999506e-01 1.62790611e-01 2.36540735e-01 7.77078867e-01
4.09438908e-01 2.64287740e-01 1.95992470e-01 -2.42877096e-01
2.28204206e-01 1.67072359e-02 -1.07800186e+00 7.13231936e-02
-4.71924007e-01 -2.50563949e-01 9.75696206e-01 -3.63563418e-01
-1.37946594e+00 -9.52220410e-02 -7.40522668e-02 -6.44845843e-01
5.97069673e-02 -1.23566695e-01 -8.18491280e-02 -3.74387681e-01
4.91117150e-01 3.51758987e-01 -2.07438052e-01 -1.24086246e-01
-9.70275551e-02 4.94760245e-01 7.90736318e-01 -2.76688963e-01
5.42509854e-01 4.54427212e-01 -4.16724682e-02 -1.03238630e+00
-5.56410626e-02 -4.52451438e-01 -3.28116417e-01 -7.36079514e-01
8.54420960e-01 -1.19943178e+00 -8.93781364e-01 3.98434371e-01
-1.25273156e+00 -9.62070301e-02 -5.17475195e-02 8.29594910e-01
-5.44142485e-01 6.73828959e-01 -7.97737181e-01 -7.45015800e-01
-2.65070945e-01 -1.41530859e+00 5.59610426e-01 4.63271022e-01
2.02342086e-02 -1.08678627e+00 -3.97724092e-01 2.64343172e-01
3.69791657e-01 3.08346748e-01 6.08645141e-01 3.98642123e-01
-1.06002975e+00 9.82513465e-03 -4.67344791e-01 5.55661201e-01
4.24851477e-01 1.47015080e-01 -7.95857489e-01 -3.68369788e-01
5.47143742e-02 5.60624719e-01 6.05859995e-01 6.31015718e-01
1.36380982e+00 -3.32389057e-01 -1.61035284e-01 7.24059105e-01
1.68314755e+00 7.89575338e-01 1.27810836e+00 1.29521742e-01
5.85030138e-01 2.81611606e-02 8.04342151e-01 7.55228817e-01
-2.91666221e-02 9.99034524e-01 2.83100069e-01 -2.95968831e-01
-5.16081512e-01 -2.71355268e-02 4.12004530e-01 5.47412992e-01
-4.48404521e-01 -2.93551952e-01 -6.81607068e-01 5.59283912e-01
-1.66899133e+00 -1.25610149e+00 -2.86609083e-01 2.38811898e+00
8.84813070e-01 9.43838134e-02 6.59986287e-02 3.38485211e-01
1.08571184e+00 3.86983931e-01 -4.23417985e-01 -2.57385284e-01
-4.57850546e-02 -1.39372703e-02 6.48973048e-01 6.97211862e-01
-9.22696829e-01 6.26927674e-01 6.71434975e+00 8.78532171e-01
-1.03585434e+00 -2.26218969e-01 8.28413129e-01 -1.69435516e-02
-1.82508290e-01 3.08475755e-02 -6.43441200e-01 9.94454205e-01
8.28303277e-01 -2.51283705e-01 3.29158336e-01 5.91919541e-01
9.64885652e-01 -3.39417279e-01 -7.62159884e-01 1.27630508e+00
-6.42118230e-02 -1.75110018e+00 4.71432582e-02 -1.27158210e-01
8.21707547e-01 -6.55911088e-01 -4.43943292e-02 -4.02726233e-01
-1.80172384e-01 -6.45823061e-01 7.39629865e-01 6.16445124e-01
8.27245474e-01 -9.59207475e-01 5.65202534e-01 2.44891942e-02
-1.37337601e+00 -7.13900030e-02 -3.65146071e-01 3.39076556e-02
5.03227413e-01 7.14695513e-01 -4.86947268e-01 4.59596753e-01
8.06298614e-01 7.33798385e-01 -3.60849887e-01 1.70388782e+00
-1.66120708e-01 6.71751976e-01 1.12353593e-01 3.12249005e-01
-1.20255813e-01 -3.01551014e-01 5.13622880e-01 1.27917027e+00
3.90771896e-01 2.12856665e-01 5.57389185e-02 5.62475562e-01
-9.91516933e-02 5.52021945e-03 -2.99149632e-01 5.80538690e-01
4.53634292e-01 8.17818224e-01 -5.38873136e-01 -3.96912307e-01
-6.17320716e-01 1.29114366e+00 -3.51048678e-01 5.31794786e-01
-1.15602720e+00 -4.71484840e-01 8.55799019e-01 4.31452185e-01
4.64765221e-01 -4.86653984e-01 -5.43876886e-01 -1.13796413e+00
1.01382203e-01 -7.11061478e-01 1.92269862e-01 -8.45115781e-01
-7.52864063e-01 4.75842834e-01 3.64908576e-03 -1.73661613e+00
-2.86465824e-01 -1.82982460e-01 -5.44396281e-01 8.36510837e-01
-1.58771622e+00 -4.97196078e-01 -4.94187236e-01 6.41474128e-01
1.04320192e+00 -6.39330298e-02 2.34919623e-01 6.84105039e-01
-2.68241435e-01 6.84645653e-01 3.51948887e-01 8.29058141e-02
6.96402133e-01 -7.94313550e-01 2.49342192e-02 1.26841402e+00
-1.17629319e-01 1.98012725e-01 8.66510212e-01 -4.48240787e-01
-1.24336958e+00 -1.02776408e+00 6.32279098e-01 1.78788692e-01
2.50272125e-01 2.65688915e-02 -1.04043937e+00 3.34853441e-01
6.87827349e-01 3.19828361e-01 2.80926704e-01 -5.57277441e-01
1.35236114e-01 -5.08408606e-01 -1.33058596e+00 8.00298035e-01
5.60289502e-01 -2.09393471e-01 -8.37365463e-02 3.75628024e-02
5.39636672e-01 -4.72413212e-01 -9.84875500e-01 3.66936088e-01
4.14333910e-01 -1.21507180e+00 1.26182449e+00 1.40877753e-01
6.11774385e-01 -6.94496930e-01 -4.53581009e-03 -8.20518732e-01
-5.00132263e-01 -7.67859101e-01 -2.09260017e-01 1.14092302e+00
1.89829811e-01 -1.06807806e-01 8.24438274e-01 5.66690803e-01
-1.01245129e-02 -4.42301184e-01 -8.69874835e-01 -8.76105487e-01
-5.13248384e-01 -1.38772458e-01 1.52815953e-01 6.83343530e-01
-1.08814217e-01 -1.87692180e-01 -8.47303033e-01 2.93382287e-01
9.86477077e-01 -1.01818927e-01 4.58065212e-01 -5.02433538e-01
-3.59015837e-02 -1.97501943e-01 -7.49219239e-01 -1.35339737e+00
-2.23890647e-01 -1.19616918e-01 8.10509175e-02 -1.25756681e+00
-1.30610317e-01 -2.23247871e-01 -3.58151942e-01 -8.23304281e-02
-4.77313548e-01 5.02524018e-01 3.77315342e-01 1.54945657e-01
-5.25468826e-01 4.31836963e-01 1.42793906e+00 -2.32541449e-02
-1.79428995e-01 1.09152541e-01 -7.84040615e-02 7.97515094e-01
6.74911737e-01 -3.28007601e-02 -3.93963814e-01 -6.33501947e-01
-4.01073515e-01 6.21449769e-01 4.24768120e-01 -1.45749795e+00
2.28262872e-01 -3.79519254e-01 8.55585635e-01 -5.51365912e-01
4.19562787e-01 -1.06277239e+00 4.12211925e-01 5.91701925e-01
-2.69964278e-01 6.61930069e-02 3.44069153e-01 4.96035308e-01
-4.20793205e-01 -2.04493552e-01 1.39168501e+00 1.26853585e-01
-1.07869172e+00 3.66556309e-02 -4.33490366e-01 -2.16422871e-01
1.23034894e+00 -6.90100074e-01 2.46742934e-01 -5.94507039e-01
-3.78008723e-01 -1.13097824e-01 6.62112653e-01 1.61095396e-01
9.91756260e-01 -1.42441046e+00 -7.81708062e-01 2.27367505e-01
-2.28661224e-01 -4.47166622e-01 3.23144615e-01 5.54432094e-01
-1.10352457e+00 7.68087953e-02 -5.57165504e-01 -7.75657654e-01
-1.40417504e+00 4.99960274e-01 5.15114777e-02 9.68413576e-02
-7.51714706e-01 6.84497118e-01 6.67011784e-03 9.28382635e-01
1.95058212e-01 -3.53255391e-01 -3.81487682e-02 -3.92657727e-01
9.21310246e-01 8.53788376e-01 -1.34366989e-01 -5.15968442e-01
-2.45016336e-01 4.09331650e-01 -1.68735728e-01 9.31177959e-02
9.30014431e-01 -4.70659077e-01 3.52745742e-01 7.55714402e-02
1.03566432e+00 1.84809238e-01 -2.14891195e+00 -6.24985248e-02
-1.18113011e-01 -1.33408117e+00 1.47959366e-01 -5.69083631e-01
-1.33644128e+00 6.49936914e-01 9.14397180e-01 3.81304584e-02
1.58150828e+00 -5.57830870e-01 1.15215814e+00 -3.85743529e-01
3.91086072e-01 -8.07631195e-01 8.25438928e-03 5.03741093e-02
6.84429765e-01 -1.23082530e+00 3.27153713e-01 -6.59693480e-01
-5.01846850e-01 1.26697743e+00 4.17903423e-01 -2.17127174e-01
3.08644623e-01 4.36926275e-01 4.85359505e-02 4.43316787e-01
-5.16295314e-01 2.21924990e-01 2.44365484e-01 6.03897333e-01
5.12308180e-01 -3.00476223e-01 -6.06227279e-01 -1.58410355e-01
4.21440423e-01 2.11440846e-01 6.31753981e-01 7.06423581e-01
-5.29033363e-01 -8.96019638e-01 -7.36759186e-01 1.26947507e-01
-6.15845978e-01 -9.64180529e-02 5.28127313e-01 5.64040124e-01
-1.11532407e-02 9.84923303e-01 1.40891314e-01 -1.25084847e-01
2.61032909e-01 -5.14572382e-01 5.97627401e-01 1.33507907e-01
-2.48352557e-01 3.23806703e-01 -8.04755092e-03 -6.47573531e-01
-5.61321378e-01 -4.54926163e-01 -1.33970463e+00 -5.93220532e-01
-9.37345177e-02 4.42998204e-03 3.84415954e-01 5.66136777e-01
1.67806208e-01 6.29111886e-01 8.39298129e-01 -7.93728709e-01
1.13325261e-01 -4.49298024e-01 -3.71014863e-01 4.14399266e-01
4.67065126e-01 -3.63769859e-01 -2.94209093e-01 8.10967386e-01]
|
[10.996702194213867, -1.7297922372817993]
|
b14aa7f0-287f-4888-ab89-0c27f93e9ce0
|
content-based-detection-of-temporal-metadata
|
2103.04736
| null |
https://arxiv.org/abs/2103.04736v2
|
https://arxiv.org/pdf/2103.04736v2.pdf
|
Content-Aware Detection of Temporal Metadata Manipulation
|
Most pictures shared online are accompanied by temporal metadata (i.e., the day and time they were taken), which makes it possible to associate an image content with real-world events. Maliciously manipulating this metadata can convey a distorted version of reality. In this work, we present the emerging problem of detecting timestamp manipulation. We propose an end-to-end approach to verify whether the purported time of capture of an outdoor image is consistent with its content and geographic location. We consider manipulations done in the hour and/or month of capture of a photograph. The central idea is the use of supervised consistency verification, in which we predict the probability that the image content, capture time, and geographical location are consistent. We also include a pair of auxiliary tasks, which can be used to explain the network decision. Our approach improves upon previous work on a large benchmark dataset, increasing the classification accuracy from 59.0% to 81.1%. We perform an ablation study that highlights the importance of various components of the method, showing what types of tampering are detectable using our approach. Finally, we demonstrate how the proposed method can be employed to estimate a possible time-of-capture in scenarios in which the timestamp is missing from the metadata.
|
['Nathan Jacobs', 'Anderson Rocha', 'Fernanda A. Andaló', 'Scott Workman', 'Tawfiq Salem', 'Rafael Padilha']
|
2021-03-08
|
content-aware-detection-of-temporal-metadata
|
https://rafaspadilha.github.io/publication/padilha22content/
|
https://rafaspadilha.github.io/publication/padilha22content/
|
ieee-transactions-on-information-forensics-8
|
['temporal-metadata-manipulation-detection']
|
['computer-vision']
|
[ 4.53363955e-01 -1.70824915e-01 -2.53732443e-01 -2.49520555e-01
-6.57343268e-01 -9.38799083e-01 8.14590156e-01 3.00445914e-01
-3.73323262e-01 5.27814567e-01 -1.86883304e-02 -2.84059137e-01
1.76933587e-01 -7.60402322e-01 -1.07492459e+00 -5.99553049e-01
-3.48714262e-01 -1.75296962e-01 4.33647335e-01 3.07547212e-01
4.48854029e-01 6.09887838e-01 -1.24008262e+00 2.64394253e-01
3.54684778e-02 1.22981727e+00 6.08261712e-02 9.00156140e-01
4.42841113e-01 9.22861874e-01 -8.33219051e-01 -6.71755850e-01
3.38721752e-01 -5.26871942e-02 -5.46228945e-01 2.17857420e-01
5.97200990e-01 -8.84587705e-01 -6.65954053e-01 1.03427267e+00
3.59103866e-02 -3.54077250e-01 3.22909921e-01 -1.86050725e+00
-4.00394827e-01 4.41525489e-01 -7.07906902e-01 3.98024887e-01
4.80178773e-01 3.50796729e-02 7.21525788e-01 -3.18442136e-01
8.69615197e-01 7.12271988e-01 8.93846571e-01 -8.98430496e-02
-8.59966636e-01 -8.01081479e-01 -9.21069533e-02 4.39577729e-01
-1.57786942e+00 -4.97837305e-01 7.13690579e-01 -4.08405602e-01
3.92145097e-01 4.72404748e-01 2.09014922e-01 1.13716447e+00
1.56466991e-01 4.83469903e-01 1.19270599e+00 -4.08039123e-01
5.75681999e-02 2.94169635e-01 -3.86915915e-02 5.04923046e-01
4.52408373e-01 -7.06699863e-02 -5.92134535e-01 -5.50262988e-01
4.00293887e-01 1.47639066e-01 -3.04141790e-01 -1.75460666e-01
-1.30289447e+00 5.46272576e-01 2.15984330e-01 1.33932233e-01
-2.57744581e-01 3.87830168e-01 3.84327263e-01 1.69178620e-01
5.20059884e-01 1.30955130e-01 -1.81785747e-01 -1.81755066e-01
-1.16762972e+00 -7.98743069e-02 8.07189822e-01 9.86366451e-01
6.14232183e-01 -3.58234763e-01 1.73979536e-01 9.76060405e-02
1.20574728e-01 5.77842414e-01 3.15797254e-02 -1.09444761e+00
4.16209757e-01 8.64817388e-03 4.83079106e-01 -1.57370830e+00
-2.28852164e-02 1.48291234e-02 -5.85421979e-01 -1.45763844e-01
5.29174387e-01 -3.36380005e-02 -3.23706657e-01 1.69794571e+00
3.45809698e-01 8.37291002e-01 -2.50175625e-01 7.94941783e-01
5.83252087e-02 5.34228086e-01 -5.14419526e-02 -1.21967487e-01
1.36574078e+00 -4.70902473e-01 -6.09628022e-01 4.55496870e-02
5.18118918e-01 -8.68695796e-01 3.71191502e-01 3.30069661e-01
-7.83820689e-01 -1.90473676e-01 -9.96723115e-01 1.99493870e-01
-5.76449752e-01 -3.17158946e-03 3.41256201e-01 7.16264129e-01
-9.59909618e-01 6.41406655e-01 -6.18377328e-01 -5.97894728e-01
1.31237030e-01 5.60259111e-02 -5.51763475e-01 -1.77672282e-01
-1.11907184e+00 6.22836709e-01 1.07039206e-01 -1.09751530e-01
-7.93222368e-01 -4.39107358e-01 -5.62783420e-01 -7.28060976e-02
4.98829633e-01 -3.48784506e-01 1.10253501e+00 -9.60173607e-01
-7.61429787e-01 1.05930817e+00 -1.95715442e-01 -7.11332560e-01
7.07112372e-01 -6.26406521e-02 -6.69337571e-01 5.16784072e-01
2.86873102e-01 3.46636087e-01 1.11916661e+00 -1.29993403e+00
-8.33654583e-01 -1.89524308e-01 1.94969311e-01 -4.31293249e-01
-4.83161062e-01 1.57335639e-01 -7.94907510e-01 -7.28781879e-01
-6.59710690e-02 -1.11723030e+00 1.61246285e-01 4.28251117e-01
-7.14330971e-01 4.13127303e-01 1.06051540e+00 -9.42383945e-01
1.23071063e+00 -2.32526016e+00 -5.01853585e-01 4.91301030e-01
1.55026525e-01 -2.45405752e-02 1.26971230e-01 8.55898738e-01
1.15365572e-02 5.02885640e-01 2.78368331e-02 -5.30767500e-01
-1.37581900e-01 1.10964794e-02 -8.63254309e-01 9.93424237e-01
-5.19925691e-02 4.32601810e-01 -9.26320016e-01 -3.63083333e-01
1.44153878e-01 3.74536216e-01 -4.42026705e-02 2.37802982e-01
-5.87845705e-02 1.45078138e-01 -8.35131183e-02 6.19592249e-01
7.92942464e-01 -3.92689377e-01 3.62470031e-01 -4.49731171e-01
-1.61000684e-01 4.01151627e-01 -1.18100095e+00 1.22940159e+00
-2.40221828e-01 9.72021043e-01 -9.18971095e-03 -4.27884370e-01
6.01694643e-01 3.58384669e-01 3.49036753e-01 -4.87121880e-01
-6.16229214e-02 -1.20168410e-01 -6.19601548e-01 -5.00481308e-01
7.59672225e-01 4.41627204e-01 -3.05493325e-01 7.98908353e-01
-6.00662351e-01 2.63187081e-01 -1.69868454e-01 4.31109458e-01
1.45019698e+00 -1.95234612e-01 4.07454133e-01 3.27300012e-01
2.26354957e-01 -3.68583240e-02 2.78303951e-01 1.05303431e+00
-2.65626490e-01 5.71066797e-01 7.22204328e-01 -3.29177082e-01
-1.29250348e+00 -8.32407773e-01 1.59666747e-01 5.24271667e-01
4.17077214e-01 -5.12015343e-01 -5.29584646e-01 -7.75389552e-01
8.28531310e-02 7.06179976e-01 -6.76070452e-01 -4.15581055e-02
-2.97013938e-01 -2.42173418e-01 8.49250257e-01 1.98270395e-01
4.73189384e-01 -3.62434506e-01 -5.59530735e-01 -1.94054708e-01
-5.71283340e-01 -1.67368472e+00 -4.40130383e-01 -4.40525085e-01
-5.02949357e-01 -1.24536061e+00 -2.00977385e-01 -1.82209074e-01
6.32449389e-01 8.90491784e-01 7.24136114e-01 4.26081419e-01
-2.27966458e-02 6.40464485e-01 -3.12429160e-01 -8.40378255e-02
-4.14376706e-01 -1.88352793e-01 3.45256291e-02 4.38650399e-01
6.91609979e-02 -6.51683152e-01 -5.09257436e-01 4.00654882e-01
-1.24765491e+00 2.44650543e-02 2.52111226e-01 1.65278360e-01
3.83050531e-01 2.86521941e-01 -1.05397636e-02 -8.17462087e-01
5.33992127e-02 -1.05422151e+00 -5.92013478e-01 2.14297593e-01
-4.05344218e-01 -4.28897947e-01 5.26087821e-01 -4.28448826e-01
-4.56499547e-01 6.71876222e-02 3.92091513e-01 -5.75334489e-01
-2.03505263e-01 2.41506562e-01 3.12341917e-02 -1.87231600e-01
7.09758997e-02 2.56017804e-01 -1.10852726e-01 -2.03462988e-01
2.16823846e-01 7.62350500e-01 7.74437189e-01 -3.12085897e-02
1.14760721e+00 1.13959455e+00 2.07478227e-03 -7.92568982e-01
-6.66072905e-01 -6.75267518e-01 -4.82002705e-01 -4.41440374e-01
4.19545323e-01 -1.07186508e+00 -7.05836356e-01 7.61415720e-01
-1.27607930e+00 1.09438626e-02 1.96191519e-01 2.87439853e-01
-2.81843156e-01 8.55332375e-01 -5.53641021e-01 -7.97456861e-01
1.27807736e-01 -6.97216809e-01 1.23189080e+00 -2.24453166e-01
-3.45314950e-01 -9.03556049e-01 -2.40207538e-01 2.89766371e-01
2.46365711e-01 4.93726790e-01 3.58161330e-01 -7.11439550e-01
-9.18737769e-01 -6.52220607e-01 -5.19535601e-01 1.34276658e-01
1.85014561e-01 3.84588122e-01 -1.12021327e+00 -3.54317129e-01
3.67104597e-02 1.13994725e-01 4.49271560e-01 -1.96795585e-03
1.25639808e+00 -7.18613386e-01 -3.58886123e-01 5.32592535e-01
1.66410112e+00 -1.48985848e-01 8.77846658e-01 5.40226877e-01
5.41920006e-01 4.11352605e-01 7.59453833e-01 8.50060880e-01
5.39552927e-01 8.89378250e-01 6.76446140e-01 1.47479311e-01
1.41192183e-01 -5.60355484e-01 4.86132950e-01 3.23209822e-01
4.87953424e-01 -5.78998983e-01 -7.46198595e-01 7.13138759e-01
-1.62766516e+00 -1.31593597e+00 -3.16155285e-01 2.39304042e+00
3.54388982e-01 2.34685257e-01 1.25376150e-01 2.08376020e-01
1.10555410e+00 4.32093859e-01 -2.00297311e-01 -4.91726510e-02
3.02336570e-02 -4.37134206e-01 1.26904535e+00 3.82576078e-01
-1.13282144e+00 5.36727130e-01 6.16216326e+00 4.87520337e-01
-1.28356004e+00 1.52862862e-01 5.73422790e-01 -1.35705248e-02
-3.26797441e-02 2.00496048e-01 -6.41084731e-01 1.00997937e+00
1.18292844e+00 2.05376018e-02 4.05900210e-01 7.09186137e-01
6.30756319e-01 -4.71702933e-01 -1.14442790e+00 7.52890885e-01
4.59671110e-01 -1.39749515e+00 -2.67818183e-01 2.98859417e-01
5.02275109e-01 -2.01327786e-01 6.54724911e-02 -3.51049364e-01
9.79675427e-02 -5.48768401e-01 9.90094900e-01 6.48013294e-01
8.28942418e-01 -3.93232971e-01 5.72505832e-01 3.74061286e-01
-1.12703168e+00 9.97793078e-02 -6.88199773e-02 -1.12335116e-01
2.93560952e-01 7.24577785e-01 -1.14712274e+00 4.48948830e-01
7.72220135e-01 8.81487012e-01 -7.43849397e-01 1.13287330e+00
-3.66188973e-01 7.93026686e-01 -5.53962588e-01 1.67861506e-01
7.06444159e-02 1.81834131e-01 6.69912636e-01 1.21158504e+00
5.64123154e-01 -1.86161458e-01 -2.78704148e-02 6.31428421e-01
-3.03073078e-01 -4.84389812e-01 -9.35017347e-01 1.49128765e-01
9.74507391e-01 1.10655868e+00 -7.08491325e-01 -3.56765360e-01
-3.07848006e-01 1.19706666e+00 4.76301908e-02 1.64282098e-01
-1.27069151e+00 -1.78475291e-01 5.82739294e-01 3.19264382e-01
6.38529778e-01 -3.42440456e-01 -1.21440150e-01 -1.04486537e+00
4.46828306e-01 -6.69229090e-01 2.30354831e-01 -1.18091214e+00
-1.15253007e+00 1.53537035e-01 -1.40467444e-02 -1.45304358e+00
-1.46961808e-01 -1.72382340e-01 -6.62541807e-01 5.13948798e-01
-1.67474484e+00 -1.08936417e+00 -4.91248101e-01 4.69631284e-01
-1.01829350e-01 2.52918690e-01 5.87273479e-01 4.77696419e-01
-5.51049531e-01 5.30663729e-01 2.78674334e-01 3.81474197e-01
1.06991327e+00 -8.41439128e-01 6.00307167e-01 1.34435570e+00
2.74873257e-01 3.62196058e-01 8.65490556e-01 -7.17070937e-01
-1.49166238e+00 -1.34765363e+00 1.11008215e+00 -4.17333901e-01
1.11354184e+00 -3.74621034e-01 -6.01628363e-01 9.21433091e-01
-1.26956090e-01 4.73955236e-02 5.70424855e-01 -4.32654619e-01
-7.55371451e-01 -2.81184465e-01 -1.20564425e+00 4.99548316e-01
7.62317479e-01 -9.56804156e-01 -2.70445704e-01 4.87319648e-01
6.36104941e-01 -1.61261037e-01 -6.84606493e-01 -8.61294046e-02
6.27771378e-01 -1.03568757e+00 8.00142646e-01 -3.08903307e-01
4.34693515e-01 -4.02875632e-01 -3.50733906e-01 -7.64610946e-01
3.99130248e-02 -7.44695544e-01 -2.15301961e-01 1.57290995e+00
-2.56545078e-02 -5.58347106e-01 5.42973280e-01 6.83492720e-01
5.24177313e-01 -7.85512477e-02 -1.10595012e+00 -8.05040121e-01
-6.70026243e-01 -7.08636403e-01 6.07918799e-01 1.19288731e+00
-2.12386549e-01 -3.46814692e-01 -8.22333395e-01 8.71736765e-01
7.10740149e-01 3.10378596e-02 9.02288020e-01 -8.26305509e-01
-1.28394157e-01 7.47997239e-02 -7.02513933e-01 -9.14454520e-01
7.42972270e-02 -4.38895702e-01 -2.48139590e-01 -9.44707453e-01
1.73366338e-01 -3.84770244e-01 -8.92097801e-02 3.38570386e-01
1.57493949e-01 4.83649671e-01 3.51063818e-01 6.26807451e-01
-6.34978652e-01 -1.84450626e-01 2.51344770e-01 -4.87359166e-02
3.81837428e-01 2.59328578e-02 -4.81419235e-01 6.08524203e-01
8.63120079e-01 -8.65220070e-01 -1.58627387e-02 -3.63578022e-01
3.27382952e-01 2.17580855e-01 1.04838634e+00 -1.06398225e+00
4.52249467e-01 -9.42272507e-03 8.65058899e-02 -5.75541079e-01
4.78222191e-01 -1.41337407e+00 4.50421929e-01 3.69443625e-01
-3.07755411e-01 3.95318091e-01 -9.93304849e-02 9.96051311e-01
-8.53077024e-02 -2.12783918e-01 4.58869338e-01 4.20169011e-02
-7.54881144e-01 1.20061800e-01 -3.60152543e-01 -1.99979112e-01
1.40757477e+00 -1.88181415e-01 -6.94722056e-01 -7.59695113e-01
-2.44220972e-01 -1.06019124e-01 9.13290858e-01 3.33807617e-01
4.94012117e-01 -1.11234725e+00 -4.64378655e-01 -1.20288707e-01
1.68047622e-01 -7.71866441e-01 1.34612210e-02 8.40964317e-01
-5.13070285e-01 1.60809249e-01 -5.96176311e-02 -5.25992155e-01
-1.51749885e+00 8.49949062e-01 6.45204866e-03 -1.28771409e-01
-3.44727874e-01 2.41169274e-01 -5.22273898e-01 3.06669891e-01
3.50685716e-01 -2.45257437e-01 3.37040991e-01 1.79410204e-02
8.13753963e-01 4.37247932e-01 -2.18175817e-02 -7.38939643e-01
-3.76503795e-01 2.25061223e-01 1.13001224e-02 -1.83538586e-01
1.24696445e+00 -5.36190927e-01 -1.35279417e-01 5.07546365e-01
1.46172726e+00 4.08064723e-01 -1.11294675e+00 -3.71084869e-01
-2.39395034e-02 -8.12573195e-01 -6.92527071e-02 -5.65482676e-01
-1.11525369e+00 5.06911039e-01 4.79273647e-01 4.94137913e-01
1.00748599e+00 -1.25228971e-01 1.01724792e+00 9.60800275e-02
6.20935857e-01 -7.44173169e-01 -3.27152193e-01 8.43894258e-02
3.53934675e-01 -1.19531488e+00 3.25647622e-01 -6.53726280e-01
-4.14492458e-01 9.40125227e-01 4.14429158e-02 -6.29545599e-02
6.56229198e-01 2.46862203e-01 -2.70431727e-01 2.34059524e-02
-7.04419255e-01 1.31188974e-01 -1.64127022e-01 4.69015926e-01
-1.51961744e-01 3.06175090e-02 -1.82920638e-02 2.31392458e-01
-1.26420051e-01 3.59547921e-02 1.02589023e+00 1.10729873e+00
-2.61072703e-02 -7.77001560e-01 -6.46540344e-01 2.15659291e-01
-7.13815391e-01 3.95467803e-02 -5.77936172e-01 8.49151850e-01
1.92104112e-02 1.15681541e+00 3.08005869e-01 -6.03580654e-01
-3.27266082e-02 -1.41916752e-01 -6.23047426e-02 -1.75048858e-01
-4.08200592e-01 -4.56128985e-01 3.60717267e-01 -8.66040885e-01
-5.45203388e-01 -9.19259012e-01 -6.09782279e-01 -9.09535944e-01
-2.19311684e-01 -5.84939532e-02 7.45199740e-01 8.33041251e-01
4.83695507e-01 -1.57036707e-01 1.08749986e+00 -6.11863315e-01
-3.55334967e-01 -4.37790990e-01 -5.27406812e-01 7.04188585e-01
6.55001044e-01 -3.02702427e-01 -7.42376208e-01 4.84833986e-01]
|
[12.341078758239746, 1.0217957496643066]
|
9010852d-bb3f-40b5-9ab2-f5a1539b9f3d
|
image-smoothing-algorithm-based-on-gradient
| null | null |
https://ieeexplore.ieee.org/document/9117646
|
https://ieeexplore.ieee.org/document/9117646
|
Image Smoothing Algorithm Based on Gradient Analysis
|
In this paper image smoothing algorithm based on gradient analysis is proposed. Our algorithm uses filtering and to achieve edge-preserving smoothing it uses two components of gradient vectors: their magnitudes (or lengths) and directions. Our method discriminates between two types of boundaries in given neighborhood: regular and irregular ones. Regular boundaries have small deviations of gradient angles and the opposite for irregular ones. To measure closeness of angles cosine of doubled difference is used. As additional measure that helps to discriminate the types of boundaries inverted gradient values were used. When gradient magnitudes are inverted bigger values refer to textures (insignificant changes in gradient) and smaller refer to strong boundaries. So textures would have bigger weights and hence they would appear smoother. We also propose to filter image of gradient magnitudes with median filter to enhance visual quality of results. The method proposed in this paper is easy to implement and compute and it gives good results in comparison with other techniques like bilateral filter.
|
['Ilia Moiseev', 'Vladimir Gudkov']
|
2020-06-16
| null | null | null | null |
['image-smoothing']
|
['computer-vision']
|
[ 2.10471243e-01 -1.36731789e-01 1.06901936e-01 -3.91176194e-01
2.16546580e-01 -5.73349595e-01 5.97710073e-01 6.34272873e-01
-6.52095914e-01 9.15585279e-01 5.53923249e-01 -2.74501443e-01
-1.40349194e-01 -1.09884286e+00 -2.61210978e-01 -6.09505773e-01
-2.09981635e-01 -1.04272023e-01 9.04225826e-01 -4.17628944e-01
7.38446653e-01 5.68397999e-01 -1.58704937e+00 2.82007068e-01
7.12816358e-01 7.15466559e-01 9.27967355e-02 9.20190156e-01
-1.48292437e-01 3.79060149e-01 -3.34255546e-01 2.20275462e-01
6.67520642e-01 -4.22672600e-01 -7.25864708e-01 1.51603043e-01
4.61594641e-01 -4.58855219e-02 2.28193924e-01 1.26809108e+00
4.43581045e-01 2.80682772e-01 9.83108521e-01 -6.54917419e-01
-6.18825436e-01 1.45237908e-01 -1.09171069e+00 4.33973819e-01
4.13248450e-01 -2.64649779e-01 3.80441248e-01 -7.11845160e-01
7.99333632e-01 1.14865708e+00 8.42472911e-01 1.24728411e-01
-1.32270026e+00 -1.66370675e-01 -1.99589521e-01 3.00373167e-01
-1.16169345e+00 -8.75703841e-02 5.59598863e-01 -4.00564790e-01
6.18833125e-01 7.10474670e-01 4.89543557e-01 -6.77623525e-02
4.68555659e-01 1.45315588e-01 1.70715582e+00 -7.50615776e-01
7.85134733e-02 2.96521187e-01 2.69120425e-01 5.28576672e-01
4.72000688e-01 -3.26484367e-02 1.32152170e-01 7.32836500e-02
8.92737627e-01 -7.15510845e-02 -4.24529791e-01 -1.45387992e-01
-1.38228452e+00 5.76918542e-01 3.87759089e-01 9.79699910e-01
-4.48417485e-01 -1.39174372e-01 3.37753713e-01 4.81530219e-01
2.14737356e-01 -6.57958388e-02 -2.53116161e-01 7.54741430e-02
-8.48116159e-01 1.92688689e-01 5.80201685e-01 3.62373650e-01
8.08326006e-01 -2.15547353e-01 -2.26463675e-02 9.19816375e-01
6.84757112e-03 2.39027128e-01 5.74894130e-01 -5.00026643e-01
1.68620482e-01 5.70578516e-01 2.81704396e-01 -1.47212291e+00
-4.80330259e-01 1.91780567e-01 -7.68912673e-01 1.17335594e+00
1.05794883e+00 -4.46609333e-02 -1.07166016e+00 1.24094558e+00
3.51956040e-01 -7.72264302e-02 -9.84581336e-02 8.72385919e-01
7.52712786e-01 7.39104867e-01 -1.82951093e-01 -1.29879266e-01
1.51727521e+00 -4.81238097e-01 -8.16516399e-01 2.55423993e-01
2.33398452e-01 -1.61374760e+00 9.08974767e-01 3.58945221e-01
-1.13445687e+00 -6.88114226e-01 -1.03593922e+00 3.86403129e-02
-8.80923033e-01 2.12282926e-01 3.38637233e-02 6.22313678e-01
-1.23520017e+00 1.00481451e+00 -5.07652462e-01 -6.76680386e-01
-2.07033142e-01 2.91264147e-01 -4.28808033e-01 6.64885104e-01
-7.62839437e-01 9.28964496e-01 2.36667186e-01 5.62596768e-02
2.78615534e-01 -2.27237001e-01 -5.21261215e-01 -1.16914943e-01
-1.17094852e-01 -1.26475781e-01 5.78422010e-01 -1.16548371e+00
-1.25366402e+00 9.86664295e-01 -3.05443794e-01 -2.08090425e-01
8.45600069e-01 1.16029963e-01 -5.64706683e-01 5.45755364e-02
-2.87595719e-01 4.30974066e-01 7.51819074e-01 -1.23236430e+00
-9.29600477e-01 -2.43753865e-01 -3.29502285e-01 1.77175790e-01
-1.09202147e-01 -1.99268341e-01 8.06810427e-03 -7.95604229e-01
6.30972743e-01 -5.76131940e-01 -4.45250012e-02 1.18257299e-01
-2.26716444e-01 -1.67796254e-01 1.21702468e+00 -9.65635478e-01
1.42455745e+00 -1.83924699e+00 -4.34729367e-01 6.88853204e-01
1.14150539e-01 2.60417700e-01 3.76130134e-01 3.79460812e-01
-1.13550946e-01 -1.40184071e-02 -9.20148492e-02 5.16539752e-01
-1.93518221e-01 1.25262327e-02 1.30419910e-01 5.73719740e-01
-5.16740531e-02 -1.38733193e-01 -5.63671172e-01 -6.81728065e-01
2.77621001e-01 7.03676105e-01 -3.24165702e-01 -1.63802296e-01
4.16528195e-01 6.50069267e-02 -2.40947887e-01 4.74831700e-01
1.20720267e+00 4.65813428e-01 -6.23051226e-02 -6.70900524e-01
-7.14430690e-01 -1.57459319e-01 -1.91265380e+00 6.18485093e-01
-3.02974671e-01 9.33497369e-01 2.85211354e-01 -9.94481981e-01
1.36519098e+00 3.34097952e-01 -1.01429336e-01 -4.76763904e-01
3.45227242e-01 3.42796743e-01 5.26960939e-02 -5.15262246e-01
7.34421670e-01 -1.67976528e-01 6.64382756e-01 1.60714462e-01
-6.75903678e-01 -1.89883277e-01 6.05816424e-01 -8.40652660e-02
6.34495258e-01 2.82842349e-02 5.18151224e-01 -8.46753359e-01
1.04388440e+00 -4.29992318e-01 5.52201271e-01 3.84210289e-01
-3.71101379e-01 6.57836556e-01 5.75961709e-01 -4.80085939e-01
-1.10065997e+00 -1.00767136e+00 -6.04469836e-01 5.99131346e-01
3.57381344e-01 4.98346835e-02 -6.24414921e-01 -3.63709778e-01
9.11957920e-02 7.12623373e-02 -7.65957057e-01 4.70313698e-01
-8.58949780e-01 -5.32980800e-01 -7.69984424e-02 2.85754740e-01
8.48291218e-01 -1.04828799e+00 -7.30408072e-01 5.35485707e-02
2.10449204e-01 -3.07065606e-01 -4.99458075e-01 -2.09328562e-01
-1.29847586e+00 -1.07208157e+00 -1.08732510e+00 -1.19816518e+00
9.92590785e-01 2.65255094e-01 8.35829973e-01 2.68108666e-01
-4.52528685e-01 -1.04503922e-01 -4.02309418e-01 -3.03863823e-01
-2.76093781e-01 -6.22569442e-01 -3.54288727e-01 7.89325219e-03
1.70072868e-01 -4.93250787e-01 -1.04261816e+00 4.49482083e-01
-8.10301423e-01 -3.99745971e-01 5.19001305e-01 6.79148376e-01
4.42784697e-01 2.13482842e-01 1.53804466e-01 -8.46833587e-01
9.34665382e-01 -3.80850099e-02 -6.99051082e-01 1.69492468e-01
-5.80779552e-01 3.06324422e-01 6.11515522e-01 -2.30751321e-01
-1.20607615e+00 -2.16357961e-01 -2.89129745e-02 5.90010107e-01
-3.79454792e-01 4.60145585e-02 3.87353718e-01 -3.88615519e-01
7.12348342e-01 -8.71132389e-02 3.62133943e-02 -6.78675532e-01
6.41043186e-02 6.49438858e-01 4.92614120e-01 -2.23442897e-01
4.79884923e-01 7.57780969e-01 2.62244105e-01 -1.21247077e+00
3.52280825e-01 -5.56933403e-01 -4.38769102e-01 -4.26344007e-01
9.50667262e-01 -2.25352403e-02 -6.65268660e-01 4.81084853e-01
-8.34164858e-01 5.08267395e-02 1.17835380e-01 5.78622222e-01
-2.37427205e-01 6.91179812e-01 -6.31717741e-01 -9.34241116e-01
-3.70765686e-01 -8.55290055e-01 4.24264878e-01 8.06511164e-01
-2.34748587e-01 -1.37792504e+00 5.51403724e-02 -3.50071371e-01
6.84231758e-01 5.27664781e-01 7.79871762e-01 1.18466787e-01
-3.32729407e-02 -3.62306654e-01 -3.33461195e-01 2.38094822e-01
4.59653169e-01 3.62532467e-01 -5.83025217e-01 -1.31786987e-01
-1.31813483e-02 3.89654577e-01 1.03218842e+00 7.19554186e-01
7.42437899e-01 -9.04904306e-02 -4.24375296e-01 3.66216153e-01
1.95689750e+00 4.58396733e-01 9.96256530e-01 7.62913287e-01
1.25163466e-01 5.63324988e-01 6.74754798e-01 3.10390711e-01
-2.83578724e-01 4.90607202e-01 -1.06818117e-02 -6.41805947e-01
-4.06066000e-01 9.77334455e-02 1.14378296e-01 5.44080734e-01
-6.15502298e-01 1.43468380e-01 -5.36898613e-01 6.57051802e-01
-1.50841534e+00 -1.10292065e+00 -7.87064254e-01 2.57894063e+00
8.70186210e-01 4.09961909e-01 1.56816125e-01 6.09842420e-01
1.17617321e+00 1.16289683e-01 1.69543460e-01 -1.08840585e+00
-2.08593711e-01 3.44356626e-01 8.11214089e-01 1.10986555e+00
-1.12279892e+00 5.34518003e-01 6.10330105e+00 7.04957128e-01
-1.43041074e+00 -2.16465384e-01 5.40743351e-01 4.87542838e-01
-1.89279810e-01 2.12152094e-01 -5.87222457e-01 6.22094572e-01
7.08770752e-02 -9.13915336e-02 2.55920496e-02 3.52150381e-01
5.14431715e-01 -1.02218318e+00 -3.02466482e-01 6.17729485e-01
-4.04784381e-01 -1.11323333e+00 -1.28526583e-01 -9.75803584e-02
7.77852714e-01 -4.23080206e-01 -6.91885576e-02 -5.02901852e-01
-2.14957464e-02 -7.11076975e-01 5.65960050e-01 7.41984606e-01
1.86390534e-01 -6.81114435e-01 9.97180104e-01 -5.70260733e-02
-1.50431287e+00 2.32772291e-01 -5.81068754e-01 -3.07481140e-01
2.64718175e-01 7.90236294e-01 -6.74511909e-01 1.82703450e-01
6.89918995e-01 3.10827106e-01 -5.03545403e-01 1.47929943e+00
9.72118974e-02 4.07036722e-01 -6.34536207e-01 -1.40455306e-01
3.15650195e-01 -9.38741148e-01 5.43152869e-01 1.65467405e+00
4.58894789e-01 -1.88124701e-01 -1.58738181e-01 4.51236963e-01
5.25226712e-01 6.19965315e-01 -5.85593939e-01 4.35788542e-01
4.54225123e-01 1.17862296e+00 -1.44189525e+00 -6.65609121e-01
-4.28996921e-01 8.70046556e-01 -2.75512934e-01 2.17973992e-01
-3.83456886e-01 -1.10913348e+00 3.46118897e-01 4.81523901e-01
3.04977328e-01 -2.43254304e-01 -7.04281032e-01 -5.52979469e-01
-2.78044678e-02 -4.53740120e-01 4.67952281e-01 -3.85540634e-01
-7.90985823e-01 6.58466101e-01 -9.27697495e-02 -1.14564574e+00
3.96509647e-01 -6.97997570e-01 -1.15168715e+00 1.14440870e+00
-1.14758527e+00 -7.02997267e-01 -4.56652254e-01 5.68146467e-01
4.60182488e-01 2.10412666e-01 2.79489547e-01 2.77221262e-01
9.74743515e-02 2.33284578e-01 4.42093521e-01 -2.57825106e-02
8.98568928e-01 -1.52967954e+00 -9.87804756e-02 1.00709093e+00
-2.17866302e-01 6.79781497e-01 1.37540078e+00 -7.60548890e-01
-5.13855875e-01 -2.09744483e-01 1.24407816e+00 3.49148095e-01
4.14937735e-01 2.70583123e-01 -9.95523751e-01 2.13802785e-01
5.65246165e-01 -2.38168612e-01 1.19261801e-01 -2.40003765e-01
1.49856284e-01 -2.40492910e-01 -1.38410509e+00 5.07566631e-01
3.48978013e-01 -5.12799919e-02 -6.41518950e-01 1.19870462e-01
-4.79781121e-01 3.52199492e-03 -7.14202404e-01 2.33697638e-01
7.51357675e-01 -1.73218262e+00 8.61368179e-01 8.63531157e-02
7.06455903e-03 -8.10085833e-01 2.72394359e-01 -1.21414912e+00
-3.46747518e-01 -4.91215706e-01 1.00193417e+00 1.15026951e+00
4.92903918e-01 -8.86021256e-01 7.08248615e-01 -8.19304660e-02
2.03635469e-01 -6.01611376e-01 -5.43176591e-01 -6.80636287e-01
-1.25269309e-01 3.94566447e-01 2.13701993e-01 8.84816408e-01
3.13510209e-01 -1.87042892e-01 -1.04641460e-01 -1.35403290e-01
5.91847956e-01 9.62276682e-02 4.32548940e-01 -1.24055171e+00
-1.85626075e-02 -7.32284725e-01 -7.61897445e-01 -6.28106534e-01
-7.23380566e-01 -5.77982366e-01 -2.60079741e-01 -1.65047002e+00
-2.12151557e-01 -5.04727244e-01 -1.42612413e-01 1.10084951e-01
-1.93047434e-01 6.82458580e-01 7.85882864e-03 -7.69287348e-02
3.79020452e-01 -2.41350606e-01 1.26294935e+00 2.33354121e-01
-4.47394490e-01 4.62647006e-02 -1.48429023e-02 1.06258941e+00
9.49792624e-01 7.39261741e-03 -1.18325859e-01 -6.06437922e-02
1.73132736e-02 -2.82840908e-01 -1.49681289e-02 -1.06991041e+00
-4.62974831e-02 -7.89706111e-02 5.33961654e-01 -5.92990100e-01
-1.20834701e-01 -7.64149904e-01 1.75297320e-01 8.42483997e-01
-1.79037973e-01 3.64975303e-01 -7.05491332e-03 3.35805714e-01
-2.94937581e-01 -5.94054997e-01 1.24984944e+00 -1.89079404e-01
-9.27111149e-01 -5.33750057e-01 -6.73481703e-01 -5.06533861e-01
1.01670992e+00 -9.53332365e-01 -6.14096634e-02 -6.12511396e-01
-1.10201478e+00 -3.06692630e-01 7.57108867e-01 -3.87565158e-02
3.66132379e-01 -1.21736801e+00 -6.05246186e-01 2.23153159e-01
-4.05242831e-01 -7.76765883e-01 3.95806991e-02 1.23288488e+00
-1.42965782e+00 6.59075901e-02 -8.15688252e-01 -1.61267638e-01
-2.05533028e+00 2.93421715e-01 1.22799434e-01 -9.15172324e-02
-6.74529850e-01 8.74490499e-01 5.77047728e-02 2.23660052e-01
1.54279828e-01 -6.17158890e-01 -7.67933488e-01 2.29421392e-01
5.97485363e-01 9.47814286e-01 -3.23387748e-03 -7.99693286e-01
-1.93216309e-01 1.38613558e+00 2.88400035e-02 -2.66603768e-01
1.01078796e+00 -2.02096239e-01 -3.92184108e-01 2.79316455e-01
1.17116404e+00 8.60943675e-01 -1.11333168e+00 2.75098294e-01
1.55237466e-01 -8.44061315e-01 1.35414079e-01 -5.85499644e-01
-8.41086447e-01 6.95883453e-01 1.11365521e+00 8.06495667e-01
1.26190925e+00 -5.30264318e-01 7.94301867e-01 -1.81558102e-01
-1.17903635e-01 -1.55546355e+00 -3.25847864e-01 2.87156224e-01
8.13884139e-01 -1.06570792e+00 2.27226436e-01 -7.38130331e-01
-4.86010492e-01 1.69364464e+00 1.07173994e-01 -5.80771804e-01
8.36811543e-01 3.31646353e-01 3.14430654e-01 -3.34583670e-02
-1.11465640e-01 -4.28048819e-01 4.72122639e-01 6.60229087e-01
1.01871252e+00 5.03716292e-03 -1.82978690e+00 -4.01995003e-01
-9.65569019e-02 5.79970516e-02 7.24310577e-01 1.17790711e+00
-1.04621875e+00 -1.28473508e+00 -8.72645795e-01 3.99735004e-01
-5.80620646e-01 1.68078661e-01 -2.09435821e-01 8.40744495e-01
4.03762728e-01 9.67307627e-01 2.80172586e-01 -1.05252601e-01
4.16505724e-01 -1.23691998e-01 6.61395669e-01 3.11856329e-01
-5.68583190e-01 3.54616225e-01 2.06591561e-01 -2.12356567e-01
-4.43165839e-01 -5.26219666e-01 -1.42469251e+00 -5.92937469e-01
-3.56804788e-01 2.67770499e-01 8.88115168e-01 4.11844552e-01
-1.18420981e-01 1.57409355e-01 3.13270599e-01 -7.28485286e-01
-2.53965855e-01 -1.00700319e+00 -8.88604760e-01 9.41893399e-01
3.03305626e-01 -5.57135880e-01 -5.65288603e-01 5.24470150e-01]
|
[11.006725311279297, -2.49760365486145]
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.