MMA-Diffusion: MultiModal Attack on Diffusion Models

Yijun Yang, Ruiyuan Gao, Xiaosen Wang, Nan Xu, Qiang Xu
11/1/2023

No official code found. Best implementation: cure-lab/MMA-Diffusion (388 stars, last commit 2026-01-08).

Abstract

In recent years, Text-to-Image (T2I) models have seen remarkable advancements, gaining widespread adoption. However, this progress has inadvertently opened avenues for potential misuse, particularly in generating inappropriate or Not-Safe-For-Work (NSFW) content. Our work introduces MMA-Diffusion, a framework that presents a significant and realistic threat to the security of T2I models by effectively circumventing current defensive measures in both open-source models and commercial online services. Unlike previous approaches, MMA-Diffusion leverages both textual and visual modalities to bypass safeguards like prompt filters and post-hoc safety checkers, thus exposing and highlighting the vulnerabilities in existing defense mechanisms. Our codes are available at https://github.com/cure-lab/MMA-Diffusion.

View on arXivView PDF

Code Implementations(1)

cure-lab/MMA-DiffusionCommunity75%

[CVPR2024] MMA-Diffusion: MultiModal Attack on Diffusion Models

38856PythonMar 18, 20249 months agoNOASSERTION

Similar Papers

MMA-Diffusion: MultiModal Attack on Diffusion Models

99% match1 repo3882023

SurrogatePrompt: Bypassing the Safety Filter of Text-to-Image Models via Substitution

85% match2023

Patronus: Safeguarding Text-to-Image Models against White-Box Adversaries

85% match2025

ART: Automatic Red-teaming for Text-to-Image Models to Protect Benign Users

84% match1 repo252024

Unified Prompt Attack Against Text-to-Image Generation Models

84% match2025

From ChatGPT to ThreatGPT: Impact of Generative AI in Cybersecurity and Privacy

84% match2023

Cite this paper

@article{yang2023mmadiffusion,
  title  = {MMA-Diffusion: MultiModal Attack on Diffusion Models},
  author = {Yijun Yang and Ruiyuan Gao and Xiaosen Wang and Nan Xu and Qiang Xu},
  year   = {2023},
  eprint = {2311.17516},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/pdf/2311.17516},
  journal = {CVPR 2023 2023}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2311.17516.svg)](https://codeofpaper.com/paper/2311.17516)

Discussion