Multimodal Attack

Multimodal attacks target the vulnerabilities of artificial intelligence systems that process multiple data types (e.g., images and text), aiming to manipulate their outputs through carefully crafted adversarial perturbations in one or more modalities. Current research focuses on developing effective attack strategies, often employing gradient-based optimization or evolutionary algorithms, and exploring the robustness of various model architectures, including vision-language pre-trained models and diffusion models. Understanding and mitigating these attacks is crucial for ensuring the reliability and security of increasingly prevalent multimodal AI systems across diverse applications, from image classification to text generation.

Papers

October 17, 2024

MMAD-Purify: A Precision-Optimized Framework for Efficient and Scalable Multi-Modal Attacks
Xinxin Liu, Zhongliang Guo, Siyuan Huang, Chun Pong Lau
Diffusion Model Adversarial Attack Adversarial Example Adversarial Perturbation Gradient Based Adversarial Multimodal Attack

September 26, 2024

Cross-Modality Attack Boosted by Gradient-Evolutionary Multiform Optimization
Yunpeng Gong, Qingyuan Zeng, Dejun Xu, Zhenzhong Wang, Min Jiang
Adversarial Attack Evolutionary Multitasking Multimodal Attack

August 24, 2024

Probing the Robustness of Vision-Language Pretrained Models: A Multimodal Adversarial Attack Approach
Jiwei Guan, Tianyu Ding, Longbing Cao, Lei Pan, Chen Wang, Xi Zheng
Adversarial Attack Native Robustness Adversarial Robustness Vision Language Adversarial Perturbation Adversarial Vulnerability Multimodal Attack

July 31, 2024

Vera Verto: Multimodal Hijacking Attack
Minxing Zhang, Ahmed Salem, Michael Backes, Yang Zhang
Modal Attack Multimodal Attack Hijacking Task Model Hijacking Attack

April 30, 2024

Revisiting the Adversarial Robustness of Vision Language Models: a Multimodal Perspective
Wanqi Zhou, Shuanghao Bai, Danilo P. Mandic, Qibin Zhao, Badong Chen
Adversarial Attack Vision Language Model Adversarial Robustness Contrastive Loss Multimodal Attack

April 13, 2024

Multimodal Attack Detection for Action Recognition Models
Furkan Mumcu, Yasin Yilmaz
Adversarial Learning Action Recognition Model Multimodal Attack Universal Detection

March 31, 2024

Against The Achilles' Heel: A Survey on Red Teaming for Generative Models
Lizhi Lin, Honglin Mu, Zenan Zhai, Minghan Wang, Yuxia Wang, Renxi Wang, Junjie Gao, Yixuan Zhang, Wanxiang Che, Timothy Baldwin, Xudong Han, Haonan Li
Language Model Timely Survey Generative Model Red Teaming Multimodal Attack

March 16, 2024

Improving Adversarial Transferability of Visual-Language Pre-training Models through Collaborative Multimodal Interaction
Jiyuan Fu, Zhaoyu Chen, Kaixun Jiang, Haijing Guo, Jiafeng Wang, Shuyong Gao, Wenqiang Zhang
Adversarial Attack Full Model Adversarial Robustness Multimodal Interaction Image Gradient Multimodal Attack Modality Interaction

November 29, 2023

MMA-Diffusion: MultiModal Attack on Diffusion Models
Yijun Yang, Ruiyuan Gao, Xiaosen Wang, Tsung-Yi Ho, Nan Xu, Qiang Xu
Diffusion Model Text to Image Security Vulnerability Open Source Model Multimodal Attack MMA Diffusion Leverage

May 16, 2023

Iterative Adversarial Attack on Image-guided Story Ending Generation
Youze Wang, Wenbo Hu, Richang Hong
Adversarial Text Modal Attack Multimodal Attack Multimodal Text Generation

May 7, 2023

Text-to-Image Diffusion Models can be Easily Backdoored through Multimodal Data Poisoning
Shengfang Zhai, Yinpeng Dong, Qingni Shen, Shi Pu, Yuejian Fang, Hang Su
Text to Image Diffusion Model Text to Image Synthesis Multimodal Backdoor Multimodal Attack

June 19, 2022

Towards Adversarial Attack on Vision-Language Pre-training Models
Jiaming Zhang, Qi Yi, Jitao Sang
Adversarial Attack Full Model Multimodal Attack