Adversarial Diffusion Across Modalities: A Fusion Survey of Attacks, Defenses, and Evaluation for Text, Vision, and Vision-Language Models

Abstract

Adversarial evaluation of AI systems has matured along four largely disconnected tracks: diffusion-based attacks on text and large language models (LLMs), diffusion-based attacks on image classifiers, jailbreak pipelines against vision-language models, and diffusion-based input purification defenses. Each has developed its own vocabulary, threat models, and benchmarks, with denoising diffusion models emerging as a shared generative mechanism whose recipes are now actively ported between communit

Categories

Framework mappings

Suggested from the entry's categories.

@misc{alotaibi2026adversarial,
  title = {{Adversarial Diffusion Across Modalities: A Fusion Survey of Attacks, Defenses, and Evaluation for Text, Vision, and Vision-Language Models}},
  author = {Abrar Alotaibi and Moataz Ahmed},
  year = {2026},
  month = jun,
  eprint = {2606.26566},
  archivePrefix = {arXiv},
  doi = {10.2139/ssrn.6986762},
  url = {https://arxiv.org/abs/2606.26566}
}