Ultrahyperbolic Knowledge Graph Embeddings
arXiv:2206.00449 · doi:10.1145/3534678.3539333
Abstract
Recent knowledge graph (KG) embeddings have been advanced by hyperbolic geometry due to its superior capability for representing hierarchies. The topological structures of real-world KGs, however, are rather heterogeneous, i.e., a KG is composed of multiple distinct hierarchies and non-hierarchical graph structures. Therefore, a homogeneous (either Euclidean or hyperbolic) geometry is not sufficient for fairly representing such heterogeneous structures. To capture the topological heterogeneity of KGs, we present an ultrahyperbolic KG embedding (UltraE) in an ultrahyperbolic (or pseudo-Riemannian) manifold that seamlessly interleaves hyperbolic and spherical manifolds. In particular, we model each relation as a pseudo-orthogonal transformation that preserves the pseudo-Riemannian bilinear form. The pseudo-orthogonal transformation is decomposed into various operators (i.e., circular rotations, reflections and hyperbolic rotations), allowing for simultaneously modeling heterogeneous structures as well as complex relational patterns. Experimental results on three standard KGs show that UltraE outperforms previous Euclidean- and hyperbolic-based approaches.
arXiv commentAccepted to SIGKDD'22
References in corpus (11)
- Complex Embeddings for Simple Link Prediction
- Unifying Knowledge Graph Learning and Recommendation: Towards a Better Understanding of User Preferences
- SimplE Embedding for Link Prediction in Knowledge Graphs
- Hyperbolic Graph Convolutional Neural Networks
- Analogical Inference for Multi-Relational Embeddings
- Embedding Graphs in Lorentzian Spacetime
- Irreflexive and Hierarchical Relations as Translations
- Fully Hyperbolic Neural Networks
- On Procrustes Analysis in Hyperbolic Space
- Ultrahyperbolic Representation Learning
- Directed Graph Embeddings in Pseudo-Riemannian Manifolds