摘要
事件关系是一种客观存在于事件之间的逻辑关系,事件关系检测是一项面向文本信息流进行事件关系判定的自然语言处理技术。事件关系检测的核心任务是以事件为基本语义单元,通过分析事件的篇章结构特征及语义特征,借助语义关系识别和推理,对事件关系进行自动分析与理解。事件关系检测技术在自动文摘,自动问答,信息检索等领域有着广泛的应用。首先介绍事件关系检测的任务定义、语言学资源和评测方法;然后,回顾国内外现有的主要研究方法;最后,给出这一研究的关键问题及技术难点。
Event relation is a kind of logic relation,which exists between events objectively. And event relation detection,as one of natural language processing technologies,faces information stream of texts detecting unknown event relations. The key to event relation detection is to recognize the event relation automatically by analyzing the corresponding discourse structure and semantic features of events,with the techniques of semantic relation recognition and inference,which treats event as the basic semantic unit. The notion of event relation detection has been applied to many areas such as automatic summarization,automatic question answering and information retrieval. This paper mainly introduces background,definition,linguistics resource,evaluation and methods of event relation extraction,and explores its future development trend through analyzing current research. Finally,the paper correspondingly points out the critical issues and the technical difficulties of the research.
出处
《智能计算机与应用》
2014年第2期5-9,15,共6页
Intelligent Computer and Applications
基金
国家自然科学基金(61003152
61272259
60970056)
江苏省高校自然科学基金重大项目(11KIJ520003)
苏州市自然科学基金(SYG201030
SH201212)
关键词
事件抽取
事件关系
事件推理
机器学习
Event Extraction
Event Relation
Event Inference
Machine Learning
