Inducing Interpretability in Knowledge Graph Embeddings

TL;DR

This paper introduces a method to enhance interpretability in knowledge graph embeddings by leveraging entity co-occurrence statistics, achieving better semantic clarity without sacrificing task performance.

Contribution

It proposes a novel approach to induce interpretability in KG embeddings, addressing a gap in existing vector space models.

Findings

Significantly improves interpretability of KG embeddings.

Maintains comparable performance in knowledge graph tasks.

Uses entity co-occurrence statistics effectively.

Abstract

We study the problem of inducing interpretability in KG embeddings. Specifically, we explore the Universal Schema (Riedel et al., 2013) and propose a method to induce interpretability. There have been many vector space models proposed for the problem, however, most of these methods don't address the interpretability (semantics) of individual dimensions. In this work, we study this problem and propose a method for inducing interpretability in KG embeddings using entity co-occurrence statistics. The proposed method significantly improves the interpretability, while maintaining comparable performance in other KG tasks.