DIFFender: Diffusion-Based Adversarial Defense against Patch Attacks
This addresses the problem of adversarial patch attacks for real-world applications like image classification and face recognition, representing an incremental advancement in defense strategies.
The paper tackles adversarial patch attacks on deep learning models by introducing DIFFender, a defense framework that uses a text-guided diffusion model to detect and localize patches via the Adversarial Anomaly Perception phenomenon and restore images, showing good robustness in evaluations across image classification and face recognition tasks.
Adversarial attacks, particularly patch attacks, pose significant threats to the robustness and reliability of deep learning models. Developing reliable defenses against patch attacks is crucial for real-world applications. This paper introduces DIFFender, a novel defense framework that harnesses the capabilities of a text-guided diffusion model to combat patch attacks. Central to our approach is the discovery of the Adversarial Anomaly Perception (AAP) phenomenon, which empowers the diffusion model to detect and localize adversarial patches through the analysis of distributional discrepancies. DIFFender integrates dual tasks of patch localization and restoration within a single diffusion model framework, utilizing their close interaction to enhance defense efficacy. Moreover, DIFFender utilizes vision-language pre-training coupled with an efficient few-shot prompt-tuning algorithm, which streamlines the adaptation of the pre-trained diffusion model to defense tasks, thus eliminating the need for extensive retraining. Our comprehensive evaluation spans image classification and face recognition tasks, extending to real-world scenarios, where DIFFender shows good robustness against adversarial attacks. The versatility and generalizability of DIFFender are evident across a variety of settings, classifiers, and attack methodologies, marking an advancement in adversarial patch defense strategies.