Analysis of the Impact of Negative Sampling on Link Prediction in Knowledge Graphs

TL;DR

This paper empirically investigates how different negative sampling strategies affect the quality of knowledge graph embeddings for link prediction, revealing dataset-dependent impacts of sampling methods.

Contribution

It provides a comprehensive comparison of negative sampling techniques, including novel embedding-based methods, across multiple embedding models and benchmark datasets.

Findings

Traditional corrupting positives works best on WN18

Embedding-based sampling improves results on FB15k

Sampling method impact varies significantly between datasets

Abstract

Knowledge graphs are large, useful, but incomplete knowledge repositories. They encode knowledge through entities and relations which define each other through the connective structure of the graph. This has inspired methods for the joint embedding of entities and relations in continuous low-dimensional vector spaces, that can be used to induce new edges in the graph, i.e., link prediction in knowledge graphs. Learning these representations relies on contrasting positive instances with negative ones. Knowledge graphs include only positive relation instances, leaving the door open for a variety of methods for selecting negative examples. In this paper we present an empirical study on the impact of negative sampling on the learned embeddings, assessed through the task of link prediction. We use state-of-the-art knowledge graph embeddings -- \rescal , TransE, DistMult and ComplEX -- and evaluate on benchmark datasets -- FB15k and WN18. We compare well known methods for negative sampling and additionally propose embedding based sampling methods. We note a marked difference in the impact of these sampling methods on the two datasets, with the "traditional" corrupting positives method leading to best results on WN18, while embedding based methods benefiting the task on FB15k.