Performance of Hyperbolic Geometry Models on Top-N Recommendation Tasks
This work addresses recommendation systems by proposing a minimalistic hyperbolic model, showing incremental improvements in efficiency and performance.
The authors tackled the collaborative filtering problem for top-N recommendation by introducing a simple hyperbolic geometry autoencoder with a single hidden layer, achieving competitive performance with state-of-the-art methods and outperforming Euclidean counterparts.
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.