YisenWang / MART

Code for ICLR2020 "Improving Adversarial Robustness Requires Revisiting Misclassified Examples"

YisenWang/MART Issues

How does MART perform on CIFAR100?
Updated 2 years ago
How many training epochs?
Closed 4 years ago3
Why not directly use CE loss on natural samples
Updated 3 years ago
the parameters of attack function
Closed 4 years ago
Result of BCE without kl divergence
Closed 4 years ago3
Replacing CE by KL in the inner maximization
Closed 4 years ago3
https://mega.nz/#!xZ8glS6J!MAnE91ND_WyfZ_8mvkuSa2YcA7q-1ehfSm-Q1fxOvvs can not be accessed in China
Closed 4 years ago2
How to test the white-box accuracy
Closed 4 years ago2
A doubt about the derivation of formulas in the paper
Closed 4 years ago6