English(EN) EntropyScan: Towards Model-level Backdoor Detection in LVLMs via Visual Attention Entropy

EntropyScan 通过视觉注意力异常检测 LVLM 后门

作者 PulseAugur 编辑部 · [1 个来源] · 2026-05-15 08:01

研究人员开发了 EntropyScan，一种检测大型视觉语言模型 (LVLM) 中后门的新方法。该方法是模型级别的，不需要了解训练数据或特定的攻击触发器。EntropyScan 通过分析 LVLM 在处理良性样本时视觉注意力分配中的异常来识别后门，这表明跨模态对齐受到干扰。该方法利用 Tsallis 熵来量化这些失真，在实验中取得了高精度。 AI

影响引入了一种检测视觉语言模型安全漏洞的新方法，这对于安全部署至关重要。

排序理由学术论文，介绍了一种检测 LVLM 后门的新方法。 [lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.CV 阅读 →

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; 学术论文，介绍了一种检测 LVLM 后门的新方法。 [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: 115 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。

报道来源 [1]

arXiv cs.CV TIER_1 English(EN) · Xilin Chen · 2026-05-15 08:01

EntropyScan：通过视觉注意力熵实现面向模型级别的LVLM后门检测

Large Vision-Language Models (LVLMs) have demonstrated remarkable capabilities across various tasks, yet they remain vulnerable to backdoor attacks. Existing defense methods predominantly focus on sample-level defense, which relies on the knowledge of training data or triggers. H…

本文如何被排名

报道来源 [1]

EntropyScan：通过视觉注意力熵实现面向模型级别的LVLM后门检测

相关实体

相关话题