DiffDefense: Defending against Adversarial Attacks via Diffusion Models

Abstract

This paper presents a novel reconstruction method that leverages Diffusion Models to protect machine learning classifiers against adversarial attacks, all without requiring any modifications to the classifiers themselves. The susceptibility of machine learning models to minor input perturbations renders them vulnerable to adversarial attacks. While diffusion-based methods are typically disregarded for adversarial defense due to their slow reverse process, this paper demonstrates that our proposed method offers robustness against adversarial threats while preserving clean accuracy, speed, and plug-and-play compatibility. Code at: https://github.com/HondamunigePrasannaSilva/DiffDefence.

arXiv commentPaper published at ICIAP23

References in corpus (16)

• Explaining and Harnessing Adversarial Examples
• Denoising Diffusion Probabilistic Models
• An Introduction to Variational Autoencoders
• ZOO: Zeroth Order Optimization based Black-box Attacks to Deep Neural Networks without Training Substitute Models
• One pixel attack for fooling deep neural networks
• Transferability in Machine Learning: from Phenomena to Black-Box Attacks using Adversarial Samples
• Synthesizing Robust Adversarial Examples
• Reliable evaluation of adversarial robustness with an ensemble of diverse parameter-free attacks
• Image Super-Resolution as a Defense Against Adversarial Attacks
• Uncovering the Limits of Adversarial Training against Norm-Bounded Adversarial Examples
• Fixing Data Augmentation to Improve Adversarial Robustness
• Torchattacks: A PyTorch Repository for Adversarial Attacks
• Diffusion Models for Adversarial Purification
• Adversarial purification with Score-based generative models
• Guided Diffusion Model for Adversarial Purification
• Guided Diffusion Model for Adversarial Purification from Random Noise