HyperKG: Hyperbolic Knowledge Graph Embeddings for Knowledge Base Completion
arXiv:1908.04895
Abstract
Learning embeddings of entities and relations existing in knowledge bases allows the discovery of hidden patterns in data. In this work, we examine the geometrical space's contribution to the task of knowledge base completion. We focus on the family of translational models, whose performance has been lagging, and propose a model, dubbed HyperKG, which exploits the hyperbolic space in order to better reflect the topological properties of knowledge bases. We investigate the type of regularities that our model can capture and we show that it is a prominent candidate for effectively representing a subset of Datalog rules. We empirically show, using a variety of link prediction datasets, that hyperbolic space allows to narrow down significantly the performance gap between translational and bilinear models.
arXiv comment10 pages, 2 figures
References in corpus (15)
- Emergence of scaling in random networks
- Power laws, Pareto distributions and Zipf's law
- A Review of Relational Machine Learning for Knowledge Graphs
- Hyperbolic Geometry of Complex Networks
- Powerlaw: a Python package for analysis of heavy-tailed distributions
- A Novel Embedding Model for Knowledge Base Completion Based on Convolutional Neural Network
- RotatE: Knowledge Graph Embedding by Relational Rotation in Complex Space
- Sustaining the Internet with Hyperbolic Mapping
- Analogical Inference for Multi-Relational Embeddings
- Poincaré Embeddings for Learning Hierarchical Representations
- Network Mapping by Replaying Hyperbolic Growth
- Poincaré GloVe: Hyperbolic Word Embeddings
- Network Geometry Inference using Common Neighbors
- On Multi-Relational Link Prediction with Bilinear Models
- Multi-relational Poincaré Graph Embeddings
Cited by in corpus (5)
- Knowledge Graph Embeddings and Explainable AI
- Hyperbolic Deep Neural Networks: A Survey
- DyERNIE: Dynamic Evolution of Riemannian Manifold Embeddings for Temporal Knowledge Graph Completion
- A Statistical Relational Approach to Learning Distance-based GCNs
- HyperText: Endowing FastText with Hyperbolic Geometry