面向中文AMR标注体系的兼语语料库构建及兼语结构识别  

Construction of a concurrent corpus for a Chinese AMR annotation system and recognition of concurrent structures

作　　者：侯文惠 曲维光[1,2] 魏庭新 李斌[2] 顾彦慧[1] 周俊生[1] HOU Wenhui;QU Weiguang;WEI Tingxin;LI Bin;GU Yanhui;ZHOU Junsheng(School of Computer and Electronic Information,Nanjing Normal University,Nanjing 210023,China;School of Chinese Language and Literature,Nanjing Normal University,Nanjing 210097,China;International College for Chinese Studies,Nanjing Normal University,Nanjing 210097,China)

机构地区：[1]南京师范大学计算机与电子信息学院,南京210023 [2]南京师范大学文学院,南京210097 [3]南京师范大学国际文化教育学院,南京210097

出　　处：《清华大学学报（自然科学版）》2021年第9期920-926,共7页Journal of Tsinghua University(Science and Technology)

基　　金：国家自然科学基金面上项目(61772278);江苏省高校哲学社会科学基金一般项目(2019JSA0220);国家社会科学基金面上项目(18BYY127)。

摘　　要：兼语结构是汉语中常见的一种动词结构,由述宾短语与主谓短语共享兼语,结构复杂,给句法分析造成困难,因此兼语识别工作对于语义解析及下游任务都具有重要意义。但现存兼语语料库较少,面向中文抽象语义表示(AMR)标注体系的兼语语料库构建仍处于空白阶段。针对这一现状,该文总结出一套兼语语料库标注规范,构建了包含4760个兼语句的面向中文AMR标注体系的兼语语料库。基于构建的语料库,采用LA-BiLSTM-CRF模型识别兼语结构,达到了86.06%的F1,并分析了识别结果,提出了改进方向。Concurrent structures which are shared by the predicate-object phrase and the subject-predicate phrase in one sentence are common Chinese verb structures.However,their complexity makes such structures difficult to analyze.Therefore,recognition of concurrent structures is important for semantic analyses and downstream tasks.However,there are few existing concurrent corpora with no concurrent corpora for the Chinese AMR annotation system.This study summarizes a set of concurrent corpus annotation specifications and builds a concurrent corpus for Chinese AMR annotation systems which contains 4760 concurrent sentences.The LA-BiLSTM-CRF model is then used to recognize concurrent structures with an F1 score of 86.06%.The recognition results are analyzed to determine needed improvements.

关 键 词：抽象语义表示 兼语结构 识别 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]