新攻击用单次扰动劫持MLLM · arXiv 研究

作者 PulseAugur 编辑部 · [1 个来源] · 2026-06-18 04:00

研究人员开发了一种名为语义感知劫持的新型攻击方法，该方法可以通过单次对抗性扰动来破坏多模态大语言模型（MLLM）。这种被称为语义感知通用扰动（SAUP）的技术充当语义路由器，将输入导向攻击者定义的靶点。在Qwen等模型上的实验表明，单次扰动即可实现对五个不同靶点的劫持，成功率达到66%。 AI

影响这项研究突显了MLLM的一个重大漏洞，可能影响其在自动驾驶和机器人等安全关键应用中的部署。

排序理由详细介绍针对MLLM新型攻击的研究论文。[lever_c_demoted from research: ic=1 ai=1.0]

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

本文如何被排名

Signal score: 0 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 详细介绍针对MLLM新型攻击的研究论文。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration: Single-source cluster; Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics: paper, safety; Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance: High; Clearly on-topic for AI-industry coverage.
Story freshness: 112 days old; Aged out of breaking-news scoring windows; ranking reflects the durable signal from the full source set.

完整方法见我们的编辑标准。

arXiv cs.AI TIER_1 English(EN) · Changyue Li, Jiaying Li, Youliang Yuan, Jiaming He, Zhicong Huang, Pinjia He · 2026-06-18 04:00

Semantic Router：通过单次对抗性扰动劫持 MLLM 的可行性研究

arXiv:2511.20002v3 Announce Type: replace-cross Abstract: Multimodal Large Language Models (MLLMs) are increasingly deployed in stateless systems, such as autonomous driving and robotics. This paper investigates a novel threat: Semantic-Aware Hijacking. We explore the feasibility…