Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Prior-Guided Adversarial Initialization for Fast Adversarial Training

About

Fast adversarial training (FAT) effectively improves the efficiency of standard adversarial training (SAT). However, initial FAT encounters catastrophic overfitting, i.e.,the robust accuracy against adversarial attacks suddenly and dramatically decreases. Though several FAT variants spare no effort to prevent overfitting, they sacrifice much calculation cost. In this paper, we explore the difference between the training processes of SAT and FAT and observe that the attack success rate of adversarial examples (AEs) of FAT gets worse gradually in the late training stage, resulting in overfitting. The AEs are generated by the fast gradient sign method (FGSM) with a zero or random initialization. Based on the observation, we propose a prior-guided FGSM initialization method to avoid overfitting after investigating several initialization strategies, improving the quality of the AEs during the whole training process. The initialization is formed by leveraging historically generated AEs without additional calculation cost. We further provide a theoretical analysis for the proposed initialization method. We also propose a simple yet effective regularizer based on the prior-guided initialization,i.e., the currently generated perturbation should not deviate too much from the prior-guided initialization. The regularizer adopts both historical and current adversarial perturbations to guide the model learning. Evaluations on four datasets demonstrate that the proposed method can prevent catastrophic overfitting and outperform state-of-the-art FAT methods. The code is released at https://github.com/jiaxiaojunQAQ/FGSM-PGI.

Xiaojun Jia, Yong Zhang, Xingxing Wei, Baoyuan Wu, Ke Ma, Jue Wang, Xiaochun Cao• 2022

Related benchmarks

TaskDatasetResultRank
Adversarial RobustnessCIFAR-10 (test)--
76
Image ClassificationCIFAR-10 (test)
Accuracy81.48
31
Adversarial RobustnessCIFAR-10
FGSM Robust Accuracy65.02
30
Adversarial Image ClassificationCIFAR-100
Clean Accuracy58.82
24
Image ClassificationCIFAR-100 WRN34-10 (test)
SA Success Rate61.95
22
Image ClassificationCIFAR-100 (test)
SA55.69
22
Image ClassificationTiny-ImageNet
Clean Accuracy45.13
22
Image ClassificationTiny ImageNet (test)
Standard Accuracy42.72
22
Image ClassificationCIFAR10 (test)
Accuracy (Natural)83.82
21
Showing 9 of 9 rows

Other info

Follow for update