Beyond Single-Point Perturbation: A Hierarchical, Manifold-Aware Approach to Diffusion Attacks

Zhijie Wang; Lin Wang; Zhenyu Wen; Cong Wang

AAAI 2026technical0 citations

Beyond Single-Point Perturbation: A Hierarchical, Manifold-Aware Approach to Diffusion Attacks

Zhijie Wang, Lin Wang, Zhenyu Wen, Cong Wang

Abstract

Latent Diffusion Models have become a powerful tool for generating high-fidelity unrestricted adversarial examples. However, the existing methods typically perturb only the initial latent or rely on prompt engineering, which is ill-suited to the iterative nature of the diffusion process, plus optimization instability due to external text prompts and cumulative drift that push the adversarial images off the data manifold. In this paper, we propose a hierarchical attack framework that operates in alignment with the model

BibTeX

@inproceedings{aaai2026_beyondsinglepoin,
  title = {Beyond Single-Point Perturbation: A Hierarchical, Manifold-Aware Approach to Diffusion Attacks},
  author = {Zhijie Wang and Lin Wang and Zhenyu Wen and Cong Wang},
  booktitle = {AAAI 2026},
  year = {2026}
}