文献类型：会议论文

英文题名：Rule-Based Semantic Role Labeling of Bei-sentences

作者：Wang, Mengxiang[1]

第一作者：汪梦翔

通讯作者：Wang, MX[1]

机构：[1]Beijing Union Univ, Teachers Coll, Beijing, Peoples R China

第一机构：北京联合大学师范学院

通讯机构：[1]corresponding author), Beijing Union Univ, Teachers Coll, Beijing, Peoples R China.|[1141711]北京联合大学师范学院;[11417]北京联合大学;

会议论文集：25th International Conference on Asian Language Processing (IALP)

会议日期：DEC 11-13, 2021

会议地点：ELECTR NETWORK

语种：英文

外文关键词：Bei-sentence; Semantic Role Labeling; Rules

摘要：Case markers are commonly utilized for guiding semantic relations labeling during the semantic role labelling process. The determination of some semantic roles in Bei-sentences often relies on the use of the marker "Bei". However, Chinese expressions involve complicated semantic relations and the marker "Bei" is even immediately followed by another marker, which affect the accuracy of semantic role labeling. Therefore, by identifying the problems in Bei-sentences during the semantic role labeling process, the Bei-sentence libraries incorporating a great deal of various Bei-sentences are searched and then manually labeled to generate semantic role labeling rules for labeled Bei-sentences. Then aforesaid rules are then verified on testing data. The experimental results show that semantic role labeling rules proposed in this paper are useful for semantic role labeling of labeled Bei-sentences.