We propose Adaptive Diffusion Denoised Smoothing, a method for certifying the predictions of a vision model against adversarial examples, while adapting to the input. Our key insight is to reinterpret a guided denoising diffusion model as a long sequence of adaptive Gaussian Differentially Private (GDP) mechanisms refining a pure noise sample into an image. We show that these adaptive mechanisms can be composed through a GDP privacy filter to analyze the end-to-end robustness of the guided denoising process, yielding a provable certification that extends the adaptive randomized smoothing analysis. We demonstrate that our design, under a specific guiding strategy, can improve both certified accuracy and standard accuracy on ImageNet for an ℓ2 threat model.
@article{arxiv.2507.08163,
title = {Adaptive Diffusion Denoised Smoothing : Certified Robustness via Randomized Smoothing with Differentially Private Guided Denoising Diffusion},
author = {Frederick Shpilevskiy and Saiyue Lyu and Krishnamurthy Dj Dvijotham and Mathias Lécuyer and Pierre-André Noël},
journal= {arXiv preprint arXiv:2507.08163},
year = {2025}
}