TRIM: Accelerating High-Dimensional Vector Similarity Search with Enhanced Triangle-Inequality-Based Pruning
Summary: TRIM enhances triangle‑inequality lower‑bound pruning for high‑dimensional vector search via optimized landmark selection and tunable bound relaxation to combat distance concentration. Pluggable into graph/quantization/disk indexes (HNSW, IVFPQ, DiskANN), yields massive reductions in distance/I/O (up to 99% pruning, 90–200% speedups, up to 58% I/O cut) while maintaining accuracy. (summarized by gpt-5-mini on Feb 11 2026)
Incoming Non-self Citations Over Time
No non-self incoming citations found for this paper in this database.
Authors
- 1. Yitong Song (Hong Kong Baptist University; Shanghai Jiao Tong University)
- 2. Pengcheng Zhang (Shanghai Jiao Tong University)
- 3. Chao Gao (Zilliz)
- 4. Bin Yao (Shanghai Jiao Tong University)
- 5. Kai Wang (Shanghai Jiao Tong University)
- 6. Zongyuan Wu (Alibaba)
- 7. Lin Qu (Taobao)
BibTeX Citation
@inproceedings{song_sigmod26,
title = {{TRIM: Accelerating High-Dimensional Vector Similarity Search with Enhanced Triangle-Inequality-Based Pruning}},
author = {Song, Yitong and Zhang, Pengcheng and Gao, Chao and Yao, Bin and Wang, Kai and Wu, Zongyuan and Qu, Lin},
series = {{SIGMOD} '26},
booktitle = {Proceedings of the {ACM} {SIGMOD} International Conference on Management of Data},
publisher = {Association for Computing Machinery},
doi = {10.1145/3769838},
url = {https://dl.acm.org/doi/10.1145/3769838},
year = {2026}
}
Incoming Citations (Sorted by Pagerank)
Showing 1 of 1 citing papers.
| Rank | Citing Paper | Year | Venue | Pagerank |
|---|---|---|---|---|
| 10,201 | Balancing Global and Local: Representative Sampling for Large-Scale Vector Data | 2026 | SIGMOD | 5.093636e-05 |
Previous
Page 1 / 1
Next
Outgoing Citations (Sorted by Pagerank)
Showing 22 of 22 cited papers.
Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.
Previous
Page 1 / 1
Next