Understanding and Generalizing Monotonic Proximity Graphs for Approximate Nearest Neighbor Search

TL;DR

This paper provides a theoretical analysis of Monotonic Relative Neighborhood Graphs (MRNG) used in approximate nearest neighbor search, explaining their effectiveness and proposing ways to improve and generalize these graphs for large-scale applications.

Contribution

It offers the first theoretical insights into MRNGs, explores their structure, and suggests methods to enhance graph-based ANN search algorithms.

Findings

MRNGs tend to have good search performance due to their structure

Experiments guide how to approximate and generalize MRNGs for large-scale use

Conflicting nodes in MRNGs can be leveraged to improve ANN search

Abstract

Graph-based algorithms have shown great empirical potential for the approximate nearest neighbor (ANN) search problem. Currently, graph-based ANN search algorithms are designed mainly using heuristics, whereas theoretical analysis of such algorithms is quite lacking. In this paper, we study a fundamental model of proximity graphs used in graph-based ANN search, called Monotonic Relative Neighborhood Graph (MRNG), from a theoretical perspective. We use mathematical proofs to explain why proximity graphs that are built based on MRNG tend to have good searching performance. We also run experiments on MRNG and graphs generalizing MRNG to obtain a deeper understanding of the model. Our experiments give guidance on how to approximate and generalize MRNG to build proximity graphs on a large scale. In addition, we discover and study a hidden structure of MRNG called conflicting nodes, and we give theoretical evidence how conflicting nodes could be used to improve ANN search methods that are based on MRNG.