Randomized Adversarial Training via Taylor Expansion

Gaojie Jin, Xinping Yi, Dengyu Wu +2

In recent years, there has been an explosion of research into developing more robust deep neural networks against adversarial examples. Adversarial training appears as one of the m…