摘要
语义分析是自然语言处理和人工智能的关键技术之一,在很多自然语言处理和AI系统中有广泛应用。本文选择语义分析的若干前沿问题,包括词汇融合、实体链、语义角色标注、事件分析、事件链、事件框架和事件回指等,从概念、任务、资源、意义和挑战等几个方面进行分析,最后展望语义分析的发展趋势。语义分析技术的进展一方面促进自然语言处理和AI的发展,另一方面倒逼语言研究贴近认知机理和实际需求。因此语义分析有望协同语言研究、认知模型和人工智能一起迎来一个新浪潮。
Semantic analysis is one of the most kernel techniques in natural language processing and artificial intelligence community,which can facilitate a wide range of applications in relevant fields.In this paper,we aim to conduct analysis on the frontier tasks of semantic analysis,including lexicon fusion,entity chain extraction,semantic role labeling,event analysis,event chain extraction,event framework and event anaphora.The aspects of analysis cover concept,task,resource,meaning and challenge.And the trend as well as prospect of semantic analysis tasks are presented.On the one hand,semantic analysis technology promotes the development of natural language processing and artificial intelligence community.On the other hand,it encourages the research on natural languages to get close to the cognitive mechanism,as well as the demands from real world.Therefore,semantic analysis is expected to usher in a new wave of collaborative research of languages,cognitive modeling and artificial intelligence.
作者
姬东鸿
Ji Donghong(School of Cyber Science and Engineering,Wuhan University,Wuhan 430072,Hubei,China)
出处
《长江学术》
CSSCI
2020年第2期99-114,共16页
Yangtze River Academic
基金
国家社会科学基金重大项目“基于本体演化和事件结构的语义网模型研究”(11&ZD189)
国家自然科学基金面上项目“汉语句法结构和事件结构的联合分析研究”(61772378)
教育部哲学社会科学研究重大项目“新时代国家语言文字事业的新使命与发展方略研究”(18JZD015)
国家重点研发计划“重大动物源性病原体传入风险评估和预警技术研究”(2017YFC1200500)。
关键词
语义分析
事件结构
关系模型
语言资源
深度学习
Semantic Analysis
Event Structure
Relational Model
Language Resources
Deep Learning