Falconn++: A Locality-sensitive Filtering Approach for Approximate Nearest Neighbor Search

TL;DR

Falconn++ introduces a new locality-sensitive filtering method for approximate nearest neighbor search that improves candidate quality and query efficiency on angular distance, outperforming existing hashing and graph-based methods.

Contribution

It proposes Falconn++, a novel filtering approach that reduces query time complexity and enhances recall-speed tradeoffs in approximate nearest neighbor search.

Findings

Achieves lower query time complexity than Falconn.

Outperforms Falconn in recall-speed tradeoffs on real-world datasets.

Is competitive with HNSW in high recall regimes.

Abstract

We present Falconn++, a novel locality-sensitive filtering approach for approximate nearest neighbor search on angular distance. Falconn++ can filter out potential far away points in any hash bucket \textit{before} querying, which results in higher quality candidates compared to other hashing-based solutions. Theoretically, Falconn++ asymptotically achieves lower query time complexity than Falconn, an optimal locality-sensitive hashing scheme on angular distance. Empirically, Falconn++ achieves higher recall-speed tradeoffs than Falconn on many real-world data sets. Falconn++ is also competitive with HNSW, an efficient representative of graph-based solutions on high search recall regimes.