Event Extraction Based on Deep Learning: A Survey of Research Issue

Event extraction is a long-standing and challenging task in natural language processing and has achieved encouraging results. Given various research targets and concerns, it is difficult for readers to comprehensively understand the situations and trends of event extraction. Therefore, we review event extraction studies from the perspectives of research tasks, research issues and corresponding solving methods. Specifically, the event definition is discussed first, followed by an elaborate description and analysis for research tasks to clarify the targets of diverse research tasks. Meanwhile, the representative research achievements in various tasks are summarized. Then, the main aspects of research problems that existing event extraction achievements focus on addressing, why these problems exist, and why they need to be resolved, are analyzed. Subsequently, the technical line of each aspect is sorted out to investigate the development and advancement of each other. Finally, the future direction of event extraction is discussed. © 2024 Science Press. All rights reserved.