Adversarial Attacks on Vision-Language Multimodal Systems: A Survey
ACM Transactions on Privacy and Security · 21.11.2026
DOI: 10.1145/3871040 ↗Kurz gefasst
Fachartikel in der Zeitschrift ACM Transactions on Privacy and Security, online erschienen am 21.11.2026. Verfasst von Yaguan Qian und 11 weiteren Autor:innen. Die Arbeit nennt 200 Literaturangaben.
Abstract (Originalsprache)
With the boom of artificial intelligence, multimodal deep learning, especially large language models (LLMs), has advanced remarkably and found wide applications. However, adversarial security issues hamper further development on LLMs. This paper comprehensively reviews adversarial attacks on vision-language multimodal systems (VLMS), covering key aspects such as pretrained models, large language models, adversarial attack methodologies, defense strategies, benchmark datasets, downstream tasks, and evaluation metrics. It first details multimodal deep-learning key technologies and the features of relevant models. Then, it deeply analyzes the adversarial vulnerabilities of VLMS. Next, it systematically reviews various adversarial attack methods and defend strategies in multimodal systems, covering task-specific attacks and attacks on different models, with a detailed analysis of their principles, implementations, and effects. Finally, it looks ahead to the future of multimodal adversarial attacks, suggesting research directions like improving attack universality and compliance considerations. This review offers technical references, promotes research on multimodal adversarial attacks, and helps enhance VLMS security for reliable real-world applications. The latest papers on VLMS adversarial attacks are continuously collected in https://github.com/chilljudaoren/A-Survey-of-Adversarial-Attack.
Zitieren
APA
Qian, Y., Chen, Z., Bao, Q., Gan, Y., Lin, C., Wang, S., Yin, J., Zhu, H., Ji, S., Wang, W., Wang, B., & Gu, Z. (2026). Adversarial Attacks on Vision-Language Multimodal Systems: A Survey. ACM Transactions on Privacy and Security. https://doi.org/10.1145/3871040
Vancouver
Qian Y, Chen Z, Bao Q, Gan Y, Lin C, Wang S, et al. Adversarial Attacks on Vision-Language Multimodal Systems: A Survey. ACM Transactions on Privacy and Security. 2026. doi:10.1145/3871040
BibTeX
@article{Qian2026,
author = {Qian, Yaguan and Chen, Zhihao and Bao, Qiqi and Gan, Yan and Lin, Chenhao and Wang, Shuo and Yin, Jiali and Zhu, Hegui and Ji, Shouling and Wang, Wei and Wang, Bin and Gu, Zhaoquan},
title = {Adversarial Attacks on Vision-Language Multimodal Systems: A Survey},
journal = {ACM Transactions on Privacy and Security},
year = {2026},
doi = {10.1145/3871040}
}
EUVIMED