面向研究问题的深度学习事件抽取综述  

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

作　　者：万齐智 万常选[1,2] 胡蓉[1,2] 刘德喜[1,2] 刘喜平[1,2] 廖国琼[1,3] WAN Qi-Zhi;WAN Chang-Xuan;HU Rong;LIU De-Xi;LIU Xi-Ping;LIAO Guo-Qiong(School of Computing and Artificial Intelligence,Jiangxi University of Finance and Economics,Nanchang 330032;Jiangxi Key Laboratory of Data and Knowledge Engineering,Jiangxi University of Finance and Economics,Nanchang 330013;Virtual Reality Modern Industrial Institute,Jiangxi University of Finance and Economics,Nanchang 330032)

机构地区：[1]江西财经大学计算机与人工智能学院,南昌330032 [2]江西财经大学数据与知识工程江西省高校重点实验室,南昌330013 [3]江西财经大学虚拟现实现代产业学院,南昌330032

出　　处：《自动化学报》2024年第11期2079-2101,共23页Acta Automatica Sinica

基　　金：国家自然科学基金(62272205,619721184,62272206,62076112);江西省教育厅科学技术项目(GJJ2400411);江西省自然科学基金(20242BAB25119,20232ACB202008);江西省主要学科学术和技术带头人培养计划领军人才项目(20213BCJL22041)资助。

摘　　要：事件抽取是一个历史悠久且极具挑战的研究任务,近年来取得了大量优异成果.由于事件抽取涉及的研究内容较多,它们的目标和重心各不相同,使得读者难以全面地了解事件抽取包含的研究任务、研究问题和未来热点趋势.为此,面向研究问题,对基于深度学习的事件抽取研究成果进行整理.首先,界定事件相关概念,论述事件抽取的研究任务,明确各研究任务的目标,再总结各任务上的代表性研究成果;接着,总结现有事件抽取成果主要致力于解决哪些方面研究问题,分析为什么会存在这些问题,分析为什么需要解决这些问题;然后,对各方面研究问题进行技术总结,分析各自研究方案和研究推进过程;最后,讨论事件抽取的发展趋势.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.

关 键 词：事件抽取 研究问题 研究进展及解决方案 深度学习 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术] TP18[自动化与计算机技术—计算机科学与技术]