Graph-Based Vector Search: An Experimental Evaluation of the State-of-the-Art
Summary: Large-scale survey and evaluation of in-memory graph-based vector search, comparing 12 methods on up to 1B vectors. Five paradigms—seed, incremental insertion, neighborhood propagation, diversification, divide-and-conquer—frame the space; incremental insertion and diversification perform best, base-graph choice hurts scalability, and data-adaptive seeding/diversification is a key future direction. (summarized by gpt-5-nano on Feb 09 2026)
Incoming Non-self Citations Over Time
Authors
- 1. Ilias Azizi (Mohammed V University; University of Paris)
- 2. Karima Echihabi (Mohammed V University)
- 3. Themis Palpanas (University of Paris)
BibTeX Citation
@inproceedings{azizi_sigmod25,
title = {{Graph-Based Vector Search: An Experimental Evaluation of the State-of-the-Art}},
author = {Azizi, Ilias and Echihabi, Karima and Palpanas, Themis},
series = {{SIGMOD} '25},
booktitle = {Proceedings of the {ACM} {SIGMOD} International Conference on Management of Data},
publisher = {Association for Computing Machinery},
doi = {10.1145/3709693},
url = {https://dl.acm.org/doi/10.1145/3709693},
year = {2025}
}
Incoming Citations (Sorted by Pagerank)
Showing 24 of 24 citing papers.
Previous
Page 1 / 1
Next
Outgoing Citations (Sorted by Pagerank)
Showing 21 of 21 cited papers.
Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.
Previous
Page 1 / 1
Next