arXiv cs.AI· Dixi Yao, Tahseen Rabbani, Tian Li·· 2 天前AI 评分26
通过随机投影实现隐私保护的 RAG
Private Retrieval Augmented Generation via Random Projection
AI 导读
研究者提出一种基于随机投影的 RAG 数据存储隐私保护方法，对数据存储嵌入向量和查询嵌入向量施加相同投影，可证明地保持查询与检索项之间的语义相似性，同时大幅缓解数据提取攻击。在多种 RAG 架构和数据集上，该轻量方法在检索与生成性能上优于此前依赖形式化差分隐私（DP）保证的方法，在强攻击模型下经验隐私性相当。
正文 · 原文
Abstract:Retrieval-Augmented Generation (RAG) enhances the capabilities of large language models (LLMs) by querying external structured knowledge. However, it can also introduce privacy risks by leaking sensitive information from the retrieval datastore. We propose a simple method to preserve datastore privacy in RAG systems via random projection. By applying the same projection to both datastore embeddings and query embeddings, our method provably preserves semantic similarity between queries and retrieved items while substantially mitigating data extraction attacks. Across multiple RAG architectures and datasets, we show that this lightweight approach achieves superior retrieval and generation performance compared to prior methods which leverage formal differential privacy (DP) guarantees, while exhibiting comparable empirical privacy under strong attack models. Our results for the first time suggest that random projection can serve as a competitive and practical baseline for privacy-preserving RAG systems.
| Subjects: | Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR) |
| Cite as: | arXiv:2610.22297 [cs.AI] |
| (or arXiv:2610.22297v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.22297 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Dixi Yao [view email]
[v1]
Thu, 15 Oct 2026 17:45:51 UTC (204 KB)
来源：arXiv cs.AI · arxiv.org