Document-Level Event Extraction via Human-Like Reading Process

TL;DR

This paper introduces HRE, a human-inspired two-stage method for document-level event extraction that effectively handles scattered arguments and multiple simultaneous events, outperforming previous approaches.

Contribution

The paper presents a novel human-like reading process-based framework for DEE, decomposing it into rough and elaborate reading stages to address key challenges.

Findings

HRE outperforms prior methods in experimental evaluations.

Hierarchical argument extraction improves accuracy for scattered arguments.

Iterative multi-round reading detects multiple events effectively.

Abstract

Document-level Event Extraction (DEE) is particularly tricky due to the two challenges it poses: scattering-arguments and multi-events. The first challenge means that arguments of one event record could reside in different sentences in the document, while the second one reflects one document may simultaneously contain multiple such event records. Motivated by humans' reading cognitive to extract information of interests, in this paper, we propose a method called HRE (Human Reading inspired Extractor for Document Events), where DEE is decomposed into these two iterative stages, rough reading and elaborate reading. Specifically, the first stage browses the document to detect the occurrence of events, and the second stage serves to extract specific event arguments. For each concrete event role, elaborate reading hierarchically works from sentences to characters to locate arguments across sentences, thus the scattering-arguments problem is tackled. Meanwhile, rough reading is explored in a multi-round manner to discover undetected events, thus the multi-events problem is handled. Experiment results show the superiority of HRE over prior competitive methods.