NODE-AdvGAN: Improving the transferability and perceptual similarity of adversarial examples by dynamic-system-driven adversarial generative model

TL;DR

NODE-AdvGAN introduces a continuous dynamic-system approach using Neural ODEs to generate more perceptually similar and transferable adversarial examples, enhancing model robustness and black-box attack success.

Contribution

It presents a novel NODE-based adversarial generator that models the generation process as a continuous system, improving transferability and perceptual quality of adversarial examples.

Findings

Higher attack success rates compared to baselines.

Better perceptual similarity in generated adversarial examples.

Enhanced transferability in black-box scenarios.

Abstract

Understanding adversarial examples is crucial for improving model robustness, as they introduce imperceptible perturbations to deceive models. Effective adversarial examples, therefore, offer the potential to train more robust models by eliminating model singularities. We propose NODE-AdvGAN, a novel approach that treats adversarial generation as a continuous process and employs a Neural Ordinary Differential Equation (NODE) to simulate generator dynamics. By mimicking the iterative nature of traditional gradient-based methods, NODE-AdvGAN generates smoother and more precise perturbations that preserve high perceptual similarity when added to benign images. We also propose a new training strategy, NODE-AdvGAN-T, which enhances transferability in black-box attacks by tuning the noise parameters during training. Experiments demonstrate that NODE-AdvGAN and NODE-AdvGAN-T generate more effective adversarial examples that achieve higher attack success rates while preserving better perceptual quality than baseline models.