Consistency Purification: Effective and Efficient Diffusion Purification towards Certified Robustness

TL;DR

This paper introduces Consistency Purification, a novel diffusion purification method that efficiently produces semantically aligned, on-manifold images in a single step, enhancing certified robustness and efficiency.

Contribution

The paper proposes Consistency Purification, a one-step, on-manifold image purification method using a distilled consistency model refined with LPIPS loss for semantic alignment.

Findings

Achieves state-of-the-art certified robustness.

Outperforms baseline methods in efficiency.

Generates semantically aligned purified images in one step.

Abstract

Diffusion Purification, purifying noised images with diffusion models, has been widely used for enhancing certified robustness via randomized smoothing. However, existing frameworks often grapple with the balance between efficiency and effectiveness. While the Denoising Diffusion Probabilistic Model (DDPM) offers an efficient single-step purification, it falls short in ensuring purified images reside on the data manifold. Conversely, the Stochastic Diffusion Model effectively places purified images on the data manifold but demands solving cumbersome stochastic differential equations, while its derivative, the Probability Flow Ordinary Differential Equation (PF-ODE), though solving simpler ordinary differential equations, still requires multiple computational steps. In this work, we demonstrated that an ideal purification pipeline should generate the purified images on the data manifold that are as much semantically aligned to the original images for effectiveness in one step for efficiency. Therefore, we introduced Consistency Purification, an efficiency-effectiveness Pareto superior purifier compared to the previous work. Consistency Purification employs the consistency model, a one-step generative model distilled from PF-ODE, thus can generate on-manifold purified images with a single network evaluation. However, the consistency model is designed not for purification thus it does not inherently ensure semantic alignment between purified and original images. To resolve this issue, we further refine it through Consistency Fine-tuning with LPIPS loss, which enables more aligned semantic meaning while keeping the purified images on data manifold. Our comprehensive experiments demonstrate that our Consistency Purification framework achieves state-of the-art certified robustness and efficiency compared to baseline methods.