Event Argument Extraction with Enriched Prompts

TL;DR

This paper investigates how enriching prompts with different contextual information affects event argument extraction performance, achieving optimal results and demonstrating potential for further optimization.

Contribution

It introduces a comprehensive analysis of prompt-based EAE models with various information types and establishes performance benchmarks on RAMS dataset.

Findings

Enriching prompts improves EAE accuracy.

Optimal prompt configurations can be identified.

Models can be further optimized through training objectives.

Abstract

This work aims to delve deeper into prompt-based event argument extraction (EAE) models. We explore the impact of incorporating various types of information into the prompt on model performance, including trigger, other role arguments for the same event, and role arguments across multiple events within the same document. Further, we provide the best possible performance that the prompt-based EAE model can attain and demonstrate such models can be further optimized from the perspective of the training objective. Experiments are carried out on three small language models and two large language models in RAMS.