Is There More Pattern in Knowledge Graph? Exploring Proximity Pattern for Knowledge Graph Embedding

TL;DR

This paper introduces the concept of proximity pattern in knowledge graphs, defining it based on shared queries, and proposes a novel CP-GNN model that combines proximity and relation patterns to improve knowledge graph embedding and completion.

Contribution

It defines proximity pattern in knowledge graphs, constructs a derived graph for it, and develops CP-GNN to integrate proximity with relation patterns, achieving state-of-the-art results.

Findings

CP-GNN outperforms existing models on FB15k-237 and WN18RR datasets.

Proximity pattern enhances modeling of complex multi-answer queries.

Integrating proximity with relation patterns improves knowledge graph embedding quality.

Abstract

Modeling of relation pattern is the core focus of previous Knowledge Graph Embedding works, which represents how one entity is related to another semantically by some explicit relation. However, there is a more natural and intuitive relevancy among entities being always ignored, which is that how one entity is close to another semantically, without the consideration of any explicit relation. We name such semantic phenomenon in knowledge graph as proximity pattern. In this work, we explore the problem of how to define and represent proximity pattern, and how it can be utilized to help knowledge graph embedding. Firstly, we define the proximity of any two entities according to their statistically shared queries, then we construct a derived graph structure and represent the proximity pattern from global view. Moreover, with the original knowledge graph, we design a Chained couPle-GNN (CP-GNN) architecture to deeply merge the two patterns (graphs) together, which can encode a more comprehensive knowledge embedding. Being evaluated on FB15k-237 and WN18RR datasets, CP-GNN achieves state-of-the-art results for Knowledge Graph Completion task, and can especially boost the modeling capacity for complex queries that contain multiple answer entities, proving the effectiveness of introduced proximity pattern.