Improving Visual Quality of Unrestricted Adversarial Examples with Wavelet-VAE

TL;DR

This paper introduces a wavelet-VAE approach to generate high-quality, unrestricted adversarial examples by modifying latent codes, posing new challenges to AI safety with imperceptible image alterations.

Contribution

The paper proposes a novel wavelet-VAE method for creating unrestricted adversarial examples by latent code modification, differing from traditional perturbation-based attacks.

Findings

Generates high-quality adversarial images on ImageNet

Modifications are imperceptible to humans

Effective alternative to perturbation-based attacks

Abstract

Traditional adversarial examples are typically generated by adding perturbation noise to the input image within a small matrix norm. In practice, un-restricted adversarial attack has raised great concern and presented a new threat to the AI safety. In this paper, we propose a wavelet-VAE structure to reconstruct an input image and generate adversarial examples by modifying the latent code. Different from perturbation-based attack, the modifications of the proposed method are not limited but imperceptible to human eyes. Experiments show that our method can generate high quality adversarial examples on ImageNet dataset.