ADBM: Adversarial diffusion bridge model for reliable adversarial purification

TL;DR

This paper introduces ADBM, a new adversarial purification method that constructs a reverse diffusion bridge to more effectively recover clean data from adversarial examples, outperforming existing diffusion-based defenses.

Contribution

The paper proposes ADBM, a novel adversarial diffusion bridge model that improves purification by directly reversing the diffusion process, addressing limitations of previous methods.

Findings

ADBM outperforms DiffPure in adversarial purification tasks.

Theoretical analysis confirms ADBM's robustness and effectiveness.

Experimental results demonstrate ADBM's superior defense capabilities.

Abstract

Recently Diffusion-based Purification (DiffPure) has been recognized as an effective defense method against adversarial examples. However, we find DiffPure which directly employs the original pre-trained diffusion models for adversarial purification, to be suboptimal. This is due to an inherent trade-off between noise purification performance and data recovery quality. Additionally, the reliability of existing evaluations for DiffPure is questionable, as they rely on weak adaptive attacks. In this work, we propose a novel Adversarial Diffusion Bridge Model, termed ADBM. ADBM directly constructs a reverse bridge from the diffused adversarial data back to its original clean examples, enhancing the purification capabilities of the original diffusion models. Through theoretical analysis and experimental validation across various scenarios, ADBM has proven to be a superior and robust defense mechanism, offering significant promise for practical applications.