7 papers
E2E: Efficient Filtered AKNN Search via Adaptive Termination
Wenxuan Xia, Mingyu Yang, Wentao Li +1
Approximate k-Nearest Neighbor (AKNN) search is widely used in vector databases. When vectors carry additional attributes (e.g., labels or numerical values), filtered AKNN search r…
CubeGraph: Efficient Retrieval-Augmented Generation for Spatial and Temporal Data
Mingyu Yang, Wentao Li, Wei Wang
Hybrid queries combining high-dimensional vector similarity search with spatio-temporal filters are increasingly critical for modern retrieval-augmented generation (RAG) systems. E…
Multiple Index Merge for Approximate Nearest Neighbor Search
Liuchang Jing, Mingyu Yang, Lei Li +2
Approximate nearest neighbor (AKNN) search in high-dimensional space is a foundational problem in vector databases with widespread applications. Among the numerous AKNN indexes…
Quantization Meets Projection: A Happy Marriage for Approximate k-Nearest Neighbor Search
Mingyu Yang, Liuchang Jing, Wentao Li +1
Approximate -nearest neighbor (AKNN) search is a fundamental problem with wide applications. To reduce memory and accelerate search, vector quantization is widely adopted. Howev…
Elastic Index Selection for Label-Hybrid AKNN Search
Mingyu Yang, Wenxuan Xia, Wentao Li +2
Real-world vector embeddings are usually associated with extra labels, such as attributes and keywords. Many applications require the nearest neighbor search that contains specific…
ESG: Elastic Graphs for Range-Filtering Approximate k-Nearest Neighbor Search
Mingyu Yang, Wentao Li, Zhitao Shen +2
Range-filtering approximate -nearest neighbor (RFAKNN) search takes as input a vector and a numeric value, returning points from a database of high-dimensional points. T…