RotatE: Knowledge Graph Embedding by Relational Rotation in Complex Space

TL;DR

RotatE introduces a novel knowledge graph embedding method that models relations as rotations in complex space, effectively capturing diverse relation patterns and outperforming existing models in link prediction tasks.

Contribution

The paper proposes RotatE, a new embedding model that encodes relations as rotations in complex space, enabling it to model various relation patterns and improve link prediction accuracy.

Findings

RotatE outperforms state-of-the-art models on benchmark datasets.

It effectively models symmetry, inversion, and composition relations.

The model is scalable and training is enhanced by a self-adversarial negative sampling technique.

Abstract

We study the problem of learning representations of entities and relations in knowledge graphs for predicting missing links. The success of such a task heavily relies on the ability of modeling and inferring the patterns of (or between) the relations. In this paper, we present a new approach for knowledge graph embedding called RotatE, which is able to model and infer various relation patterns including: symmetry/antisymmetry, inversion, and composition. Specifically, the RotatE model defines each relation as a rotation from the source entity to the target entity in the complex vector space. In addition, we propose a novel self-adversarial negative sampling technique for efficiently and effectively training the RotatE model. Experimental results on multiple benchmark knowledge graphs show that the proposed RotatE model is not only scalable, but also able to infer and model various relation patterns and significantly outperform existing state-of-the-art models for link prediction.