GAAT: Group Adaptive Adversarial Training to Improve the Trade-Off Between Robustness and Accuracy

Yaguan Qian, Xiaoyu Liang, Ming Kang, Bin Wang, Zhaoquan Gu, Xing Wang, Chunming Wu

Abstract: Adversarial training is by far one of the most effective methods to improve the robustness of deep neural networks against adversarial examples. However, the trade-off between robustness and accura...