KAIST Image and Video Systems lab

Notice

Hit	373
Subject	[Image and Vision Computing] Adversarial anchor-guided feature refinement for adversarial defense (Hakmin Lee) is accepted in ELSEVIER Image and Vision Computing
Name	관리자
Date	2023-06-22
Title: Adversarial anchor-guided feature refinement for adversarial defense Authors: Hakmin Lee and Yong Man Ro Abstract: Adversarial training (AT), which is known as a robust training method for defending against adversarial examples, usually loses the performance of models for clean examples due to the feature distribution discrepancy between clean and adversarial. In this paper, we propose a novel Adversarial Anchor-guided Feature Refinement (AAFR) defense method aimed at reducing the discrepancy and delivering reliable performances for both clean and adversarial examples. We devise adversarial anchor that detects whether the feature comes from clean or adversarial example. Then, we use adversarial anchor to refine the feature to reduce the discrepancy. As a result, the proposed method substantially achieves adversarial robustness while preserving the performance for clean examples. The effectiveness of the proposed method is verified with comprehensive experiments on CIFAR-10, CIFAR-100, and Tiny ImageNet datasets. IMAGE VIDEO SYSTEM (IVY.) KOREA ADVANCED INSTITUTE OF SCIENCE AND TECHNOLOGY (KAIST), Image and Vision Computing