面向中文抽象语义表示的复句研究综述

(1.南京师范大学国际文化教育学院,2.南京师范大学文学院,3.南京师范大学计算机科学与技术学院, 江苏 南京 210097; 4.福建省信息处理与智能控制重点实验室(闽江学院),福建 福州 350121)

中文抽象语义表示; 复句; 篇章关系

A Survey on the Study of Compound Sentences with Chinese Abstract Meaning Representation
WEI Tingxin1,2,QU Weiguang2,3,4*,SONG Li2,DAI Rubing2

(1.International College for Chinese Studies,2.School of Chinese Language and Literature,3.School of Computer Science and Technology,Nanjing Normal University,Nanjing 210097,China; 4.Fujian Provincial Key Laboratory of Information Processing and Intellige

Chinese abstract meaning representation(CAMR); compound sentence; discourse relation

DOI: 10.6043/j.issn.0438-0479.201805011

备注

抽象语义表示(AMR)是一种新型的句子语义表示方式.中文AMR在英文AMR的基础上,针对汉语特点,增加了复句逻辑语义关系的表示.中文AMR以句子为基本标注单位,以层次结构树形式表示各分句间的逻辑关系.由于允许论元共享,因此在树结构基础上形成图结构,从而对复句的语义表示更加完整全面.为了进一步研究中文AMR,对目前复句关系研究现状、复句及篇章关系资源的建设进行了综述,指出目前研究存在的问题,并提出将来工作研究的方向.

Abstract meaning representation(AMR)is a novel framework of representing sentential meaning.Due to linguistic characteristics of the Chinese language,Chinese AMR(CAMR)annotates the semantic compound-sentence meaning,which is ignored in English AMR annotation.Sentence is the elementary unit of CAMR and hierarchical tree structure is used to represent the logical relations of all minimal sentences in compound sentences.As arguments can be shared,the graph based on the tree structure expresses the semantic meaning of compound sentences more comprehensively.This paper introduces the current resource construction and methodology of compound sentence relations and discourse relations,as well as points out the key problems lying in present studies.Then future work is discussed.