An efficient indexing approach for multi-tenant vector databases that handles low-selectivity filters effectively. Curator addresses the challenge of maintaining high performance when serving multiple tenants with filtered vector search queries.
Curator presents efficient indexing techniques for multi-tenant vector databases, specifically addressing the challenge of low-selectivity filters—queries where filter conditions match only a small fraction of the dataset.
In multi-tenant vector database deployments:
Curator proposes indexing strategies that:
Efficient Filter Integration: Novel techniques for incorporating filter predicates into graph-based ANN indexes
Adaptive Routing: Graph traversal strategies that quickly navigate to relevant filtered regions
Space-Time Tradeoffs: Methods to balance index size, query latency, and filter selectivity
The paper demonstrates:
Published as arXiv preprint arXiv:2401.07119 (2024) by Jin, Yicheng, et al. The work addresses an increasingly important problem as vector databases move to production multi-tenant deployments.
Loading more......