Scalable Hyperbolic Recommender Systems
arXiv:1902.08648
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.
11 pages, 8 figures, 2 tables
References in corpus (17)
- Emergence of scaling in random networks
- The structure and function of complex networks
- BPR: Bayesian Personalized Ranking from Implicit Feedback
- How to project a bipartite network?
- Hyperbolic Geometry of Complex Networks
- Sustaining the Internet with Hyperbolic Mapping
- Bipartite Graphs as Models of Complex Networks
- The Topology of Music Recommendation Networks
- Learning Continuous Hierarchies in the Lorentz Model of Hyperbolic Geometry
- Hyperbolic Attention Networks
- Hyperbolic Entailment Cones for Learning Hierarchical Embeddings
- Latent geometry of bipartite networks
- What is the dimension of citation space?
- Gradient descent in hyperbolic space
- Skip-gram word embeddings in hyperbolic space
- Product Characterisation towards Personalisation: Learning Attributes from Unstructured Data to Recommend Fashion Products
- WMRB: Learning to Rank in a Scalable Batch Training Approach
Cited by in corpus (8)
- Hyperbolic Deep Neural Networks: A Survey
- HCGR: Hyperbolic Contrastive Graph Representation Learning for Session-based Recommendation
- Multi-Manifold Learning for Large-scale Targeted Advertising System
- Hebbian Graph Embeddings
- HSR: Hyperbolic Social Recommender
- Learning Feature Interactions with Lorentzian Factorization Machine
- HyperText: Endowing FastText with Hyperbolic Geometry
- Hyperbolic Manifold Regression