Adaptive Smoothness-weighted Adversarial Training for Multiple Perturbations with Its Stability Analysis

• Published in: arXiv.org
• Main Authors: Xiao, Jiancong, Qin, Zeyu, Fan, Yanbo, Wu, Baoyuan, Wang, Jue, Zhi-Quan Luo
• Format: Paper
• Language: English
• Published: Ithaca Cornell University Library, arXiv.org 02.10.2022
• Subjects: Algorithms; Perturbation; Smoothness; Stability analysis; Training

Adversarial Training (AT) has been demonstrated as one of the most effective methods against adversarial examples. While most existing works focus on AT with a single type of perturbation e.g., the \(\ell_\infty\) attacks), DNNs are facing threats from different types of adversarial examples. Theref...