Learning Continuous Hierarchies in the Lorentz Model of Hyperbolic Geometry

Abstract

We are concerned with the discovery of hierarchical relationships from large-scale unstructured similarity scores. For this purpose, we study different models of hyperbolic space and find that learning embeddings in the Lorentz model is substantially more efficient than in the Poincaré-ball model. We show that the proposed approach allows us to learn high-quality embeddings of large taxonomies which yield improvements over Poincaré embeddings, especially in low dimensions. Lastly, we apply our model to discover hierarchies in two real-world datasets: we show that an embedding in hyperbolic space can reveal important aspects of a company's organizational structure as well as reveal historical relationships between language families.

arXiv commentAccepted at ICML'18

Cited by in corpus (10)

• Discriminative Topic Mining via Category-Name Guided Text Embedding
• Skip-gram word embeddings in hyperbolic space
• Riemannian adaptive stochastic gradient algorithms on matrix manifolds
• McTorch, a manifold optimization library for deep learning
• Scalable Hyperbolic Recommender Systems
• Hyperbolic Interaction Model For Hierarchical Multi-Label Classification
• Hyperbolic Deep Learning for Foundation Models: A Survey
• A unified Fourier slice method to derive ridgelet transform for a variety of depth-2 neural networks
• Directed Graph Embeddings in Pseudo-Riemannian Manifolds
• Exploring Semantic Capacity of Terms