samuelepapa/adversarial_augmentation

We set out to investigate how very simple data augmentations, applied at test time, can be used to defend against adversarial attacks.

Python