Irreflexive and Hierarchical Relations as Translations

TL;DR

This paper introduces a translation-based embedding method tailored for irreflexive and hierarchical relations in knowledge bases, achieving state-of-the-art results with a simple, efficient model.

Contribution

It presents a novel translation-based approach specifically designed for irreflexive and hierarchical relations, outperforming previous models on standard benchmarks.

Findings

Achieves state-of-the-art performance on WordNet and Freebase datasets.

Uses fewer parameters than previous approaches.

Effectively models hierarchical and irreflexive relations.

Abstract

We consider the problem of embedding entities and relations of knowledge bases in low-dimensional vector spaces. Unlike most existing approaches, which are primarily efficient for modeling equivalence relations, our approach is designed to explicitly model irreflexive relations, such as hierarchies, by interpreting them as translations operating on the low-dimensional embeddings of the entities. Preliminary experiments show that, despite its simplicity and a smaller number of parameters than previous approaches, our approach achieves state-of-the-art performance according to standard evaluation protocols on data from WordNet and Freebase.