English(EN) DSPrompt: Dynamic Soft Prompt Defense Against M-RAG Corruption

新的 DSPrompt 防御可保护 M-RAG 免受对抗性攻击

作者 PulseAugur 编辑部 · [1 个来源] · 2026-08-18 04:00

研究人员开发了 DSPrompt，一个新颖的防御框架，旨在保护多模态检索增强生成 (M-RAG) 系统免受对抗性攻击。与现有的查询时防御不同，DSPrompt 将可学习的软提示集成到冻结检索器的视觉和文本编码器中。这种方法动态重塑嵌入语义，将恶意文档从 top-k 检索结果中推出，同时保留良性数据的效用。DSPrompt 仅带来最小的计算开销和参数增加，在多个基准测试中显著降低了攻击成功率。 AI

影响这种防御机制可以增强多模态 AI 系统在面对复杂的对抗性操纵时的安全性和可靠性。

排序理由该集群包含一篇详细介绍 AI 安全新技术的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.CL 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; 该集群包含一篇详细介绍 AI 安全新技术的学术论文。[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: 62 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.CL TIER_1 English(EN) · Chang Liu, Yuni Lai, Mingyue Cui, Cong Tian, Yunyan Zhang, Xian Wu, Kai Zhou, Bin Xiao · 2026-08-18 04:00

DSPrompt：动态软提示防御M-RAG腐败

arXiv:2608.16536v1 Announce Type: cross Abstract: Multimodal Retrieval Augmented Generation (M-RAG) is increasingly vulnerable to adversarial attacks where malicious data are crafted to produce embeddings that align with benign entries in the vector space, deceiving retrieval and…

本文如何被排名

报道来源 [1]

DSPrompt：动态软提示防御M-RAG腐败

相关实体

相关话题