Saved in:
Bibliographic Details
Main Authors: Li, Jinmin, Gao, Kuofeng, Bai, Yang, Zhang, Jingyun, Xia, Shu-tao, Wang, Yisen
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2403.13507
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909144421761024
author Li, Jinmin
Gao, Kuofeng
Bai, Yang
Zhang, Jingyun
Xia, Shu-tao
Wang, Yisen
author_facet Li, Jinmin
Gao, Kuofeng
Bai, Yang
Zhang, Jingyun
Xia, Shu-tao
Wang, Yisen
contents Despite the remarkable performance of video-based large language models (LLMs), their adversarial threat remains unexplored. To fill this gap, we propose the first adversarial attack tailored for video-based LLMs by crafting flow-based multi-modal adversarial perturbations on a small fraction of frames within a video, dubbed FMM-Attack. Extensive experiments show that our attack can effectively induce video-based LLMs to generate incorrect answers when videos are added with imperceptible adversarial perturbations. Intriguingly, our FMM-Attack can also induce garbling in the model output, prompting video-based LLMs to hallucinate. Overall, our observations inspire a further understanding of multi-modal robustness and safety-related feature alignment across different modalities, which is of great importance for various large multi-modal models. Our code is available at https://github.com/THU-Kingmin/FMM-Attack.
format Preprint
id arxiv_https___arxiv_org_abs_2403_13507
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle FMM-Attack: A Flow-based Multi-modal Adversarial Attack on Video-based LLMs
Li, Jinmin
Gao, Kuofeng
Bai, Yang
Zhang, Jingyun
Xia, Shu-tao
Wang, Yisen
Computer Vision and Pattern Recognition
Despite the remarkable performance of video-based large language models (LLMs), their adversarial threat remains unexplored. To fill this gap, we propose the first adversarial attack tailored for video-based LLMs by crafting flow-based multi-modal adversarial perturbations on a small fraction of frames within a video, dubbed FMM-Attack. Extensive experiments show that our attack can effectively induce video-based LLMs to generate incorrect answers when videos are added with imperceptible adversarial perturbations. Intriguingly, our FMM-Attack can also induce garbling in the model output, prompting video-based LLMs to hallucinate. Overall, our observations inspire a further understanding of multi-modal robustness and safety-related feature alignment across different modalities, which is of great importance for various large multi-modal models. Our code is available at https://github.com/THU-Kingmin/FMM-Attack.
title FMM-Attack: A Flow-based Multi-modal Adversarial Attack on Video-based LLMs
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2403.13507