Fully Hyperbolic Rotation for Knowledge Graph Embedding

TL;DR

This paper introduces a fully hyperbolic knowledge graph embedding model using Lorentz rotations directly in hyperbolic space, achieving state-of-the-art results on challenging datasets with fewer parameters.

Contribution

The paper proposes a novel model that operates entirely within hyperbolic space using Lorentz rotations, unlike previous models that relied on feature mappings.

Findings

Achieves competitive results on standard benchmarks.

Sets new state-of-the-art on CoDEx-s and CoDEx-m datasets.

Uses fewer parameters than existing models.

Abstract

Hyperbolic rotation is commonly used to effectively model knowledge graphs and their inherent hierarchies. However, existing hyperbolic rotation models rely on logarithmic and exponential mappings for feature transformation. These models only project data features into hyperbolic space for rotation, limiting their ability to fully exploit the hyperbolic space. To address this problem, we propose a novel fully hyperbolic model designed for knowledge graph embedding. Instead of feature mappings, we define the model directly in hyperbolic space with the Lorentz model. Our model considers each relation in knowledge graphs as a Lorentz rotation from the head entity to the tail entity. We adopt the Lorentzian version distance as the scoring function for measuring the plausibility of triplets. Extensive results on standard knowledge graph completion benchmarks demonstrated that our model achieves competitive results with fewer parameters. In addition, our model get the state-of-the-art performance on datasets of CoDEx-s and CoDEx-m, which are more diverse and challenging than before. Our code is available at https://github.com/llqy123/FHRE.