Self-supervised Adversarial Training

Kejiang Chen, Hang Zhou, Yuefeng Chen +6

Recent work has demonstrated that neural networks are vulnerable to adversarial examples. To escape from the predicament, many works try to harden the model in various ways, in whi…