RECON: Relation Extraction using Knowledge Graph Context in a Graph Neural Network

TL;DR

RECON is a novel graph neural network-based method that enhances relation extraction by integrating comprehensive knowledge graph context, significantly outperforming existing methods on standard datasets.

Contribution

The paper introduces RECON, a new approach that combines entity attributes and factual triples in a GNN to improve relation extraction accuracy.

Findings

Achieves 87.23 F1 on Wikidata, surpassing previous best of 82.29.

Outperforms state-of-the-art on NYT Freebase with 87.5 P@10 and 74.1 P@30.

Effectively utilizes various KG representations to boost extraction performance.

Abstract

In this paper, we present a novel method named RECON, that automatically identifies relations in a sentence (sentential relation extraction) and aligns to a knowledge graph (KG). RECON uses a graph neural network to learn representations of both the sentence as well as facts stored in a KG, improving the overall extraction quality. These facts, including entity attributes (label, alias, description, instance-of) and factual triples, have not been collectively used in the state of the art methods. We evaluate the effect of various forms of representing the KG context on the performance of RECON. The empirical evaluation on two standard relation extraction datasets shows that RECON significantly outperforms all state of the art methods on NYT Freebase and Wikidata datasets. RECON reports 87.23 F1 score (Vs 82.29 baseline) on Wikidata dataset whereas on NYT Freebase, reported values are 87.5(P@10) and 74.1(P@30) compared to the previous baseline scores of 81.3(P@10) and 63.1(P@30).