Irreflexive and Hierarchical Relations as Translations
This addresses the limitation of existing methods that are inefficient for modeling non-equivalence relations in knowledge bases, though it appears incremental as it builds on translation-based embedding approaches.
The paper tackles the problem of embedding knowledge base relations by explicitly modeling irreflexive relations like hierarchies as translations in low-dimensional vector spaces, achieving state-of-the-art performance on WordNet and Freebase datasets.
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.