Injecting Knowledge Base Information into End-to-End Joint Entity and Relation Extraction and Coreference Resolution

TL;DR

This paper presents a joint information extraction model that incorporates knowledge base information via unsupervised entity linking, improving performance on entity, relation, and coreference tasks across datasets.

Contribution

It introduces a method to inject KB information into a joint IE model using entity linking representations from Wikipedia and Wikidata, with attention-based candidate selection.

Findings

Up to 5% F1-score improvement on IE tasks

Attention scheme outperforms prior-based weighting

Knowledge base representations enhance extraction accuracy

Abstract

We consider a joint information extraction (IE) model, solving named entity recognition, coreference resolution and relation extraction jointly over the whole document. In particular, we study how to inject information from a knowledge base (KB) in such IE model, based on unsupervised entity linking. The used KB entity representations are learned from either (i) hyperlinked text documents (Wikipedia), or (ii) a knowledge graph (Wikidata), and appear complementary in raising IE performance. Representations of corresponding entity linking (EL) candidates are added to text span representations of the input document, and we experiment with (i) taking a weighted average of the EL candidate representations based on their prior (in Wikipedia), and (ii) using an attention scheme over the EL candidate list. Results demonstrate an increase of up to 5% F1-score for the evaluated IE tasks on two datasets. Despite a strong performance of the prior-based model, our quantitative and qualitative analysis reveals the advantage of using the attention-based approach.