Scholarlink

A client with query q seeks to retrieve relevant documents from a remote server storing public documents D, without revealing q or the documents. Fully homomorphic encryption (FHE)-based private information retrieval (PIR) reduces communication but increases server computation. Existing solutions minimize overhead but may lead to (1) incorrect results, (2) query leakage, (3) excessive homomorphic operations, and (4) high client-side storage. We propose a two-server approach: Server 1 manages dataset attributes, while Server 2 stores data using binary Knapsack algorithm to minimize storage waste. This optimizes the above limitations while ensuring query privacy.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Private LGBTQ Searches Without a Trace

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Akshit Aggarwal,
  • Yang Li,
  • Srinibas Swain

摘要

A client with query q seeks to retrieve relevant documents from a remote server storing public documents D, without revealing q or the documents. Fully homomorphic encryption (FHE)-based private information retrieval (PIR) reduces communication but increases server computation. Existing solutions minimize overhead but may lead to (1) incorrect results, (2) query leakage, (3) excessive homomorphic operations, and (4) high client-side storage. We propose a two-server approach: Server 1 manages dataset attributes, while Server 2 stores data using binary Knapsack algorithm to minimize storage waste. This optimizes the above limitations while ensuring query privacy.