Threshold-Protected Searchable Sharing: Privacy Preserving Aggregated-ANN Search for Collaborative RAG
Ruoyang Rykie Guo

TL;DR
This paper introduces a privacy-preserving, threshold-based searchable sharing method compatible with HNSW indexing, enabling efficient and secure approximate nearest neighbor searches over shared data in collaborative AI environments.
Contribution
It develops a novel secure, privacy-preserving search framework with a new graph structure and a unique security analysis approach addressing leakage concerns.
Findings
Reduces search complexity from O(n^2) to O(n)
Supports dynamic insertions without compromising graph topology
Provides a new security analysis framework for leakage prevention
Abstract
LLM-powered search services have driven data integration as a significant trend. However, this trend's progress is fundamentally hindered, despite the fact that combining individual knowledge can significantly improve the relevance and quality of responses in specialized queries and make AI more professional at providing services. Two key bottlenecks are private data repositories' locality constraints and the need to maintain compatibility with mainstream search techniques, particularly Hierarchical Navigable Small World (HNSW) indexing for high-dimensional vector spaces. In this work, we develop a secure and privacy-preserving aggregated approximate nearest neighbor search (SP-ANN) with HNSW compatibility under a threshold-based searchable sharing primitive. A sharable bitgraph structure is constructed and extended to support searches and dynamical insertions over shared data…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsPrivacy-Preserving Technologies in Data · Mobile Crowdsensing and Crowdsourcing · Cryptography and Data Security
