TransERR: Translation-based Knowledge Graph Embedding via Efficient Relation Rotation

TL;DR

TransERR introduces an efficient translation-based knowledge graph embedding method using hypercomplex space and adaptive quaternion rotation, improving modeling of relation patterns and scalability on large datasets.

Contribution

It proposes a novel hypercomplex-valued embedding approach with adaptive relation rotation, enhancing translation freedom and pattern modeling in knowledge graphs.

Findings

Outperforms previous models on 10 benchmark datasets

Uses fewer parameters to encode large-scale datasets

Effectively models various relation patterns

Abstract

This paper presents a translation-based knowledge geraph embedding method via efficient relation rotation (TransERR), a straightforward yet effective alternative to traditional translation-based knowledge graph embedding models. Different from the previous translation-based models, TransERR encodes knowledge graphs in the hypercomplex-valued space, thus enabling it to possess a higher degree of translation freedom in mining latent information between the head and tail entities. To further minimize the translation distance, TransERR adaptively rotates the head entity and the tail entity with their corresponding unit quaternions, which are learnable in model training. We also provide mathematical proofs to demonstrate the ability of TransERR in modeling various relation patterns, including symmetry, antisymmetry, inversion, composition, and subrelation patterns. The experiments on 10 benchmark datasets validate the effectiveness and the generalization of TransERR. The results also indicate that TransERR can better encode large-scale datasets with fewer parameters than the previous translation-based models. Our code and datasets are available at~\url{https://github.com/dellixx/TransERR}.