# Efficient Large-scale Approximate Nearest Neighbor Search on the GPU

## TL;DR

This paper introduces a GPU-optimized approximate nearest neighbor search method using a novel product quantization tree, significantly improving speed and efficiency for high-dimensional data in real-world applications.

## Contribution

The paper proposes a new two-level product and vector quantization tree with a parallel re-ranking method, enabling efficient GPU implementation for large-scale high-dimensional ANN search.

## Key findings

- Outperforms recent state-of-the-art methods on standard datasets
- Demonstrates GPU performance surpassing CPU in high-dimensional ANN tasks
- Enables real-time applications like loop-closing in videos

## Abstract

We present a new approach for efficient approximate nearest neighbor (ANN) search in high dimensional spaces, extending the idea of Product Quantization. We propose a two-level product and vector quantization tree that reduces the number of vector comparisons required during tree traversal. Our approach also includes a novel highly parallelizable re-ranking method for candidate vectors by efficiently reusing already computed intermediate values. Due to its small memory footprint during traversal, the method lends itself to an efficient, parallel GPU implementation. This Product Quantization Tree (PQT) approach significantly outperforms recent state of the art methods for high dimensional nearest neighbor queries on standard reference datasets. Ours is the first work that demonstrates GPU performance superior to CPU performance on high dimensional, large scale ANN problems in time-critical real-world applications, like loop-closing in videos.