Unified and Efficient Approach for Multi-Vector Similarity Search
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915912506933248 |
|---|---|
| author | Yang, Binhan Zeng, Yuxiang Zhang, Hengxin Zheng, Zhuanglin Chi, Yunzhen Tong, Yongxin Xu, Ke |
| author_facet | Yang, Binhan Zeng, Yuxiang Zhang, Hengxin Zheng, Zhuanglin Chi, Yunzhen Tong, Yongxin Xu, Ke |
| contents | Multi-Vector Similarity Search is essential for fine-grained semantic retrieval in many real-world applications, offering richer representations than traditional single-vector paradigms. Due to the lack of native multi-vector index, existing methods rely on a filter-and-refine framework built upon single-vector indexes. By treating token vectors within each multi-vector object in isolation and ignoring their correlations, these methods face an inherent dilemma: aggressive filtering sacrifices recall, while conservative filtering incurs prohibitive computational cost during refinement. To address this limitation, we propose MV-HNSW, the first native hierarchical graph index designed for multi-vector data. MV-HNSW introduces a novel edge-weight function that satisfies essential properties (symmetry, cardinality robustness, and query consistency) for graph-based indexing, an accelerated multi-vector similarity computation algorithm, and an augmented search strategy that dynamically discovers topologically disconnected yet relevant candidates. Extensive experiments on seven real-world datasets show that MV-HNSW achieves state-of-the-art search performance, maintaining over 90% recall while reducing search latency by up to 14.0$\times$ compared to existing methods. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_02815 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Unified and Efficient Approach for Multi-Vector Similarity Search Yang, Binhan Zeng, Yuxiang Zhang, Hengxin Zheng, Zhuanglin Chi, Yunzhen Tong, Yongxin Xu, Ke Databases Multi-Vector Similarity Search is essential for fine-grained semantic retrieval in many real-world applications, offering richer representations than traditional single-vector paradigms. Due to the lack of native multi-vector index, existing methods rely on a filter-and-refine framework built upon single-vector indexes. By treating token vectors within each multi-vector object in isolation and ignoring their correlations, these methods face an inherent dilemma: aggressive filtering sacrifices recall, while conservative filtering incurs prohibitive computational cost during refinement. To address this limitation, we propose MV-HNSW, the first native hierarchical graph index designed for multi-vector data. MV-HNSW introduces a novel edge-weight function that satisfies essential properties (symmetry, cardinality robustness, and query consistency) for graph-based indexing, an accelerated multi-vector similarity computation algorithm, and an augmented search strategy that dynamically discovers topologically disconnected yet relevant candidates. Extensive experiments on seven real-world datasets show that MV-HNSW achieves state-of-the-art search performance, maintaining over 90% recall while reducing search latency by up to 14.0$\times$ compared to existing methods. |
| title | Unified and Efficient Approach for Multi-Vector Similarity Search |
| topic | Databases |
| url | https://arxiv.org/abs/2604.02815 |