Effective and General Distance Computation for Approximate Nearest Neighbor Search

TL;DR

This paper introduces a novel data-driven distance computation method for approximate nearest neighbor search that outperforms existing approaches in speed and accuracy by leveraging data distribution and orthogonal projection.

Contribution

The paper proposes a new distance computation technique that improves effectiveness and generality over ADSampling by using data distribution and decoupled correction.

Findings

Achieves 1.6 to 2.1 times speedup over ADSampling

Provides higher accuracy in real-world datasets

Demonstrates superior performance through extensive experiments

Abstract

Approximate K Nearest Neighbor (AKNN) search in high-dimensional spaces is a critical yet challenging problem. In AKNN search, distance computation is the core task that dominates the runtime. Existing approaches typically use approximate distances to improve computational efficiency, often at the cost of reduced search accuracy. To address this issue, the state-of-the-art method, ADSampling, employs random projections to estimate approximate distances and introduces an additional distance correction process to mitigate accuracy loss. However, ADSampling has limitations in both effectiveness and generality, primarily due to its reliance on random projections for distance approximation and correction. To address the effectiveness limitations of ADSampling, we leverage data distribution to improve distance computation via orthogonal projection. Furthermore, to overcome the generality limitations of ADSampling, we adopt a data-driven approach to distance correction, decoupling the correction process from the distance approximation process. Extensive experiments demonstrate the superiority and effectiveness of our method. In particular, compared to ADSampling, our method achieves a speedup of 1.6 to 2.1 times on real-world datasets while providing higher accuracy.