arXiv 2403.093462024-12-28cs.CVcs.AI

B-AVIBench：面向黑盒对抗视觉指令的大规模视觉语言模型鲁棒性评估

B-AVIBench: Towards Evaluating the Robustness of Large Vision-Language Model on Black-box Adversarial Visual-Instructions

National Key Laboratory of Human-Machine Hybrid Augmented Intelligence（人机混合增强智能国家重点实验室）；National Engineering Research Center for Visual Information and Applications（视觉信息与应用国家工程研究中心）；Institute of Artificial Intelligence and Robotics（人工智能与机器人研究院）；Xi’an Jiaotong University（西安交通大学）；Shanghai Artificial Intelligence Laboratory（上海人工智能实验室）；Osaka University（大阪大学）

Hao Zhang, Wenqi Shao, Hong Liu, Yongqiang Ma, Ping Luo, Yu Qiao, Nanning Zheng, Kaipeng Zhang

备注：: Accepted by IEEE Transactions on Information Forensics & Security
DOI：: 10.1109/TIFS.2024.3520306

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

本文提出B-AVIBench框架，系统评估大规模视觉语言模型面对黑盒对抗视觉指令的鲁棒性，揭示其脆弱性与固有偏见。

摘要：

大规模视觉语言模型（LVLMs）在良好响应用户视觉指令方面已取得显著进展。然而，这些包含图像和文本的指令容易受到有意或无意的攻击。尽管LVLMs面对此类威胁的鲁棒性至关重要，但当前该领域的研究仍然有限。为弥补这一不足，我们提出了B-AVIBench框架，用于分析LVLMs在面对各种黑盒对抗视觉指令（B-AVIs）时的鲁棒性，包括四种基于图像的B-AVIs、十种基于文本的B-AVIs以及九种内容偏见B-AVIs（如性别、暴力、文化和种族偏见等）。我们生成了316K个B-AVIs，涵盖五类多模态能力（十个任务）和内容偏见。随后，我们对14个开源LVLMs进行了全面评估以衡量其性能。B-AVIBench还可作为从业者评估LVLMs抵御B-AVIs鲁棒性的便捷工具。我们的发现和大量实验结果表明了LVLMs的脆弱性，并强调即使在GeminiProVision和GPT-4V等先进的闭源LVLMs中也存在固有偏见。这凸显了增强LVLMs鲁棒性、安全性和公平性的重要性。源代码和基准可在https://github.com/zhanghao5201/B-AVIBench获取。

英文摘要：

Large Vision-Language Models (LVLMs) have shown significant progress in responding well to visual-instructions from users. However, these instructions, encompassing images and text, are susceptible to both intentional and inadvertent attacks. Despite the critical importance of LVLMs' robustness against such threats, current research in this area remains limited. To bridge this gap, we introduce B-AVIBench, a framework designed to analyze the robustness of LVLMs when facing various Black-box Adversarial Visual-Instructions (B-AVIs), including four types of image-based B-AVIs, ten types of text-based B-AVIs, and nine types of content bias B-AVIs (such as gender, violence, cultural, and racial biases, among others). We generate 316K B-AVIs encompassing five categories of multimodal capabilities (ten tasks) and content bias. We then conduct a comprehensive evaluation involving 14 open-source LVLMs to assess their performance. B-AVIBench also serves as a convenient tool for practitioners to evaluate the robustness of LVLMs against B-AVIs. Our findings and extensive experimental results shed light on the vulnerabilities of LVLMs, and highlight that inherent biases exist even in advanced closed-source LVLMs like GeminiProVision and GPT-4V. This underscores the importance of enhancing the robustness, security, and fairness of LVLMs. The source code and benchmark are available at https://github.com/zhanghao5201/B-AVIBench.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。