arXiv:2608.25735 · RAG 检索增强
Pointing the Way, Hiding the Destination: Practical Private Dense Retrieval at Scale
指路而隐目的地：面向大规模场景的实用化私有稠密检索
Pointing the Way, Hiding the Destination: Practical Private Dense Retrieval at Scale
- 类型：arxiv
- 标识：2608.25735
- 链接：http://arxiv.org/abs/2608.25735v1
- 主分类：rag
- 形态：method
- TLDR：Hosted retrieval-augmented generation (RAG) and semantic search allow users to query valuable provider-held corpora, raising two competing demands: to hide each query and chosen result, yet reveal only the documents that the user is authorized to receive. Existing cryptographic approaches either make this costly by processing the entire corpus for every query, or sacrifice quality for efficiency by scanning a few clusters. We repurpose learned deep hashing as a private filter: a randomized binary code points the provider to a short candidate list, while encrypted reranking and oblivious key tr
- 待LLM分类：否
- 标题中文：指路而隐目的地：面向大规模场景的实用化私有稠密检索
- TLDR中文：托管式检索增强生成 (RAG) 与语义搜索让用户得以查询提供商持有的高价值语料，由此带来两个相互矛盾的需求：既要隐藏每次查询与所选结果，又要仅返回用户被授权获取的文档。现有密码学方法要么因每次查询遍历全量语料而代价高昂，要么仅扫描少量聚类以换取效率但牺牲质量。我们将习得的深度哈希复用为私有过滤器：随机化二进制码指引提供商定位到短候选列表，而加密重排序与不经意密钥 tr
- 来源文件：
- /inbox/tom/_candidates/2026-09-01-rag-retrieval-reranking-candidates.json