SparseMult: A Tensor Decomposition model based on Sparse Relation Matrix

Abstract

Knowledge graphs (KGs) provide rich knowledge for lots of downstream tasks, such as recommendation system and question answering. However, KGs suffer from an incompleteness issue, namely, lots of relations between entities are missing. Link prediction, also known as knowledge graph completion (KGC), aims to predict missing relationships between entities. The models based on tensor decomposition, such as Rescal and DistMult, are one of the most effective methods to solve the link prediction task. However, previous Rescal method lack the ability to scale to large KGs due to the large amount of parameters. DistMult simplifies Rescal using diagonal matrices to represent relations, while it suffers from the limitation of dealing with antisymmetric relations. To address these problems, in this paper, we propose a novel tensor decomposition model based on sparse relation matrix, which is named as SparseMult. We conduct extensive experiments on link prediction task and experimental results show that our SparseMult model outperforms most of the state-of-the-art methods.