Scalable Hyperbolic Recommender Systems

TL;DR

This paper introduces a scalable hyperbolic recommender system that leverages hyperbolic geometry to outperform Euclidean models on complex network datasets, enabling large-scale recommendations.

Contribution

It presents a novel hyperbolic model and the Einstein midpoint technique, demonstrating their effectiveness for large-scale, asymmetric recommendation tasks.

Findings

Hyperbolic models outperform Euclidean models on complex network datasets.

The proposed approach scales to millions of users and hundreds of thousands of items.

Use of Einstein midpoint enables asymmetric recommendations in hyperbolic space.

Abstract

We present a large scale hyperbolic recommender system. We discuss why hyperbolic geometry is a more suitable underlying geometry for many recommendation systems and cover the fundamental milestones and insights that we have gained from its development. In doing so, we demonstrate the viability of hyperbolic geometry for recommender systems, showing that they significantly outperform Euclidean models on datasets with the properties of complex networks. Key to the success of our approach are the novel choice of underlying hyperbolic model and the use of the Einstein midpoint to define an asymmetric recommender system in hyperbolic space. These choices allow us to scale to millions of users and hundreds of thousands of items.