Performance of Hyperbolic Geometry Models on Top-N Recommendation Tasks

TL;DR

This paper presents a hyperbolic geometry-based autoencoder for collaborative filtering that outperforms Euclidean models and is competitive with state-of-the-art methods, while also analyzing the impact of space curvature.

Contribution

It introduces a simple hyperbolic autoencoder for recommendation tasks and proposes a data-driven approach to optimize space curvature.

Findings

Hyperbolic autoencoder outperforms Euclidean models.

Minimalistic single hidden layer achieves competitive results.

Optimal space curvature significantly affects model performance.

Abstract

We introduce a simple autoencoder based on hyperbolic geometry for solving standard collaborative filtering problem. In contrast to many modern deep learning techniques, we build our solution using only a single hidden layer. Remarkably, even with such a minimalistic approach, we not only outperform the Euclidean counterpart but also achieve a competitive performance with respect to the current state-of-the-art. We additionally explore the effects of space curvature on the quality of hyperbolic models and propose an efficient data-driven method for estimating its optimal value.