"Purifying Adversarial Perturbation with Adversarially Trained Auto-encoders."

Hebi Li et al. (2019)
a service of Schloss Dagstuhl - Leibniz Center for Informatics