PGTuner: An Efficient Framework for Automatic and Transferable Configuration Tuning of Proximity Graphs
Summary: PGTuner pre-trains a query-performance predictor to avoid costly proximity-graph rebuilds and uses deep reinforcement learning to recommend optimal construction/query parameters for target accuracy. It adds OOD detection and active learning for efficient transfer to new/dynamic datasets, yielding up to 14.7× speedups. (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. HAO DUAN (Shanghai Jiao Tong University)
- 2. YITONG SONG (Shanghai Jiao Tong University)
- 3. BIN YAO (Shanghai Jiao Tong University)
- 4. ANQI LIANG (Shanghai Jiao Tong University)
BibTeX Citation
@inproceedings{duan_sigmod26,
title = {{PGTuner: An Efficient Framework for Automatic and Transferable Configuration Tuning of Proximity Graphs}},
author = {DUAN, HAO and SONG, YITONG and YAO, BIN and LIANG, ANQI},
series = {{SIGMOD} '26},
booktitle = {Proceedings of the {ACM} {SIGMOD} International Conference on Management of Data},
publisher = {Association for Computing Machinery},
doi = {10.1145/3749179},
url = {https://dl.acm.org/doi/10.1145/3749179},
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 9 of 9 cited papers.
Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.
Previous
Page 1 / 1
Next