Towards Defending against Adversarial Examples via Attack-Invariant Features.

Dawei Zhou, Tongliang Liu, Bo Han, Nannan Wang, Chunlei Peng, Xinbo Gao

Browse the full ICML paper archive.