KGNN: Distributed Framework for Graph Neural Knowledge Representation

TL;DR

KGNN is a distributed framework that enhances knowledge graph representation learning by capturing high-order structures and attributes, enabling scalable and effective link prediction and classification.

Contribution

It introduces a GNN-based encoder and knowledge-aware decoder to jointly utilize structure and attribute information, addressing scalability and inductive prediction limitations.

Findings

Outperforms existing methods on multiple datasets

Effectively captures high-order structural information

Scales well to large industrial graphs

Abstract

Knowledge representation learning has been commonly adopted to incorporate knowledge graph (KG) into various online services. Although existing knowledge representation learning methods have achieved considerable performance improvement, they ignore high-order structure and abundant attribute information, resulting unsatisfactory performance on semantics-rich KGs. Moreover, they fail to make prediction in an inductive manner and cannot scale to large industrial graphs. To address these issues, we develop a novel framework called KGNN to take full advantage of knowledge data for representation learning in the distributed learning system. KGNN is equipped with GNN based encoder and knowledge aware decoder, which aim to jointly explore high-order structure and attribute information together in a fine-grained fashion and preserve the relation patterns in KGs, respectively. Extensive experiments on three datasets for link prediction and triplet classification task demonstrate the effectiveness and scalability of KGNN framework.