Prune, Don't Rebuild: Efficiently Tuning $α$-Reachable Graphs for Nearest Neighbor Search

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Zhang, Tian, Padaki, Ashwin, Liang, Jiaming, Ives, Zack, Waingarten, Erik
Format: Preprint
Publié: 2026
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866918327848271872
author Zhang, Tian
Padaki, Ashwin
Liang, Jiaming
Ives, Zack
Waingarten, Erik
author_facet Zhang, Tian
Padaki, Ashwin
Liang, Jiaming
Ives, Zack
Waingarten, Erik
contents Vector similarity search is an essential primitive in modern AI and ML applications. Most vector databases adopt graph-based approximate nearest neighbor (ANN) search algorithms, such as DiskANN (Subramanya et al., 2019), which have demonstrated state-of-the-art empirical performance. DiskANN's graph construction is governed by a reachability parameter $α$, which gives a trade-off between construction time, query time, and accuracy. However, adaptively tuning this trade-off typically requires rebuilding the index for different $α$ values, which is prohibitive at scale. In this work, we propose RP-Tuning, an efficient post-hoc routine, based on DiskANN's pruning step, to adjust the $α$ parameter without reconstructing the full index. Within the $α$-reachability framework of prior theoretical works (Indyk and Xu, 2023; Gollapudi et al., 2025), we prove that pruning an initially $α$-reachable graph with RP-Tuning preserves worst-case reachability guarantees in general metrics and improved guarantees in Euclidean metrics. Empirically, we show that RP-Tuning accelerates DiskANN tuning on four public datasets by up to $43\times$ with negligible overhead.
format Preprint
id arxiv_https___arxiv_org_abs_2602_08097
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Prune, Don't Rebuild: Efficiently Tuning $α$-Reachable Graphs for Nearest Neighbor Search
Zhang, Tian
Padaki, Ashwin
Liang, Jiaming
Ives, Zack
Waingarten, Erik
Data Structures and Algorithms
Information Retrieval
Vector similarity search is an essential primitive in modern AI and ML applications. Most vector databases adopt graph-based approximate nearest neighbor (ANN) search algorithms, such as DiskANN (Subramanya et al., 2019), which have demonstrated state-of-the-art empirical performance. DiskANN's graph construction is governed by a reachability parameter $α$, which gives a trade-off between construction time, query time, and accuracy. However, adaptively tuning this trade-off typically requires rebuilding the index for different $α$ values, which is prohibitive at scale. In this work, we propose RP-Tuning, an efficient post-hoc routine, based on DiskANN's pruning step, to adjust the $α$ parameter without reconstructing the full index. Within the $α$-reachability framework of prior theoretical works (Indyk and Xu, 2023; Gollapudi et al., 2025), we prove that pruning an initially $α$-reachable graph with RP-Tuning preserves worst-case reachability guarantees in general metrics and improved guarantees in Euclidean metrics. Empirically, we show that RP-Tuning accelerates DiskANN tuning on four public datasets by up to $43\times$ with negligible overhead.
title Prune, Don't Rebuild: Efficiently Tuning $α$-Reachable Graphs for Nearest Neighbor Search
topic Data Structures and Algorithms
Information Retrieval
url https://arxiv.org/abs/2602.08097