研究人员开发了一种名为Shard的新方法，以增强密集检索系统的隐私性。密集检索系统常用于语义搜索和检索增强生成（RAG）。Shard通过引入一种保留检索的嵌入转换来解决向量存储容易受到揭示底层文本的攻击的漏洞。这种转换将嵌入分割成用于初始检索的公共前缀和在密钥下分片的私有残差，然后使用CKKS进行重新排序以取消密钥并保持精确的内积。该系统旨在防止对齐攻击和去匿名化，提供几何防御而非加密保证。 AI
影响 增强了语义搜索和RAG系统的隐私性，可能实现更安全的数据检索。
排序理由 该集群包含一篇详细介绍私有密集检索新方法的论文。
AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →
arXiv:2606.27976v1 Announce Type: cross Abstract: Dense embeddings underpin semantic search and RAG, yet a leaked vector store hands much of the underlying text back to whoever holds it. The attacks that make this possible (few-shot alignment, zero-shot inversion, unsupervised cr…
Dense embeddings underpin semantic search and RAG, yet a leaked vector store hands much of the underlying text back to whoever holds it. The attacks that make this possible (few-shot alignment, zero-shot inversion, unsupervised cross-space translation) share one weakness: the pro…
Dense embeddings underpin semantic search and retrieval-augmented generation, yet a leaked vector store hands much of the underlying text back. Modern inversion and alignment attacks share one weakness: the protected store is a single global geometry, and any single geometry can …