用无需训练的量化方法，让企业私有检索更安全高效
TurboVec: A Case Study in Cost-Efficient Private Retrieval for Enterprise RAG via Codebook-Oblivious Quantization
- 采用无码本的标量量化，避免语料统计泄露
- 4比特下召回率比传统方法高8.5-8.9个百分点
- 适合多租户场景下的隐私敏感向量检索
检索增强生成（RAG）系统日益支撑企业级大模型应用，但向量检索层带来两个未被充分关注的问题：(1) 训练码本量化器在构建索引时可能暴露语料统计信息，导致多租户部署中的数据泄露；(2) 事后租户隔离过滤会降低选择性查询的召回率。我们提出TurboVec，一个基于TurboQuant的开源向量索引——一种无需语料依赖训练的码本无关标量量化器。在DBpedia OpenAI嵌入基准（d=1536，10万至99.9万向量）上，TurboQuant 4比特在相同内存预算下，相比训练过的FAISS乘积量化，在所有规模上的Recall@5提升8.5-8.9个百分点。相较于HNSW（R@5=0.991）和IVF-PQ（R@5=0.840），TurboQuant在无需训练的前提下实现更高召回率，且内存消耗仅为HNSW的1/4至1/8。在Snowpark容器服务中部署，10万向量下中位查询延迟为11毫秒，远低于仓库暴力扫描的707毫秒。内核级白名单过滤在10至1000个租户负载下维持0.86-0.93的Recall@10，显著优于事后过滤基线（0.09-0.19）。码本无关设计将成员推理准确率降至接近随机水平（50.0%），而PQ码本为57.3%。局限性包括仅单数据集评估、未压缩HNSW对比以及隐私评估仅限合成数据。
原文摘要 · Abstract (English)
Retrieval-Augmented Generation (RAG) systems increasingly power enterprise LLM applications, yet the vector retrieval layer introduces two underexplored challenges: (1) trained codebook quantizers may expose corpus statistics during index construction, creating a leakage channel in multi-tenant deployments, and (2) post-hoc filtering for tenant isolation degrades recall on selective queries. We study TurboVec, an open-source vector index built on TurboQuant - a codebook-oblivious scalar quantizer requiring no corpus-dependent training. On the DBpedia OpenAI embeddings benchmark (d=1536, 100K-999K vectors), TurboQuant 4-bit outperforms trained FAISS Product Quantization at the same memory budget by 8.5-8.9 percentage points in Recall@5 across all scales. Compared to HNSW (R@5=0.991) and IVF-PQ (R@5=0.840), TurboQuant occupies a distinct design point: higher recall than IVF-PQ without training, at 4-8x less memory than HNSW. Deployed on Snowpark Container Services, TurboVec achieves 11ms median query latency at 100K vectors versus 707ms for warehouse brute-force scan. Kernel-level allowlist filtering maintains 0.86-0.93 Recall@10 across 10-1000 tenant workloads versus 0.09-0.19 for post-filter baselines. Codebook-oblivious design reduces membership inference accuracy to near-random (50.0%) versus 57.3% for PQ codebooks. Limitations include single dataset evaluation, uncompressed HNSW comparison, and privacy evaluation on synthetic data only.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。