Range Retrieval with Graph-Based Indices

TL;DR

This paper introduces new graph-based range retrieval algorithms that efficiently find all points within a specified distance in high-dimensional datasets, significantly improving throughput over existing methods.

Contribution

The paper presents novel range retrieval algorithms for graph-based indices, along with a comprehensive dataset analysis and benchmarking for large-scale high-dimensional data.

Findings

Up to 100x faster query throughput compared to standard graph search.

Up to 10x improvement over previous range search modifications.

Effective performance on datasets with up to 1 billion points.

Abstract

Retrieving points based on proximity in a high-dimensional vector space is a crucial step in information retrieval applications. The approximate nearest neighbor search (ANNS) problem, which identifies the $k$ nearest neighbors for a query, has been extensively studied in recent years. However, comparatively little attention has been paid to the related problem of finding all points within a given distance of a query, the range retrieval problem, despite its applications in areas such as duplicate detection, plagiarism checking, and facial recognition. In this paper, we present new techniques for range retrieval on graph-based vector indices, which are known to achieve excellent performance on ANNS queries. Since a range query may have anywhere from no matching results to thousands of matching results in the database, we introduce a set of range retrieval algorithms based on modifications of the standard graph search that adapt to terminate quickly on queries in the former group, and to put more resources into finding results for the latter group. Due to the lack of existing benchmarks for range retrieval, we also undertake a comprehensive study of range characteristics of existing embedding datasets, and select a suitable range retrieval radius for eight existing datasets with up to 1 billion points in addition to one existing benchmark. We test our algorithms on these datasets, and find up to 100x improvement in query throughput over a standard graph search and the FAISS-IVF range search algorithm. We also find up to 10x improvement over a previously suggested modification of the standard beam search, and strong performance up to 1 billion data points.