汉语词语离合现象识别研究  

Separated Word Recognition in Chinese

作　　者：周露[1] 曲维光[1,2,4] 魏庭新 周俊生[1] 李斌 顾彦慧[1] ZHOU LU;QU Weiguang;WEI Tingxin;ZHOU Junsheng;LI Bin;GU Yanhui(School of Computer and Electronic Information/School of Artificial Intelligence,Nanjing Normal University,Nanjing,Jiangsu 210023,China;School of Chinese Language and Literature,Nanjing Normal University,Nanjing,Jiangsu 210097,China;International College for Chinese Studies,Nanjing Normal University,Nanjing,Jiangsu 210097,China;Zhongbei College,Nanjing Nornal University,Danyang,Jiangsu 212334,China)

机构地区：[1]南京师范大学计算机与电子信息学院/人工智能学院,江苏南京210023 [2]南京师范大学文学院,江苏南京210097 [3]南京师范大学国际文化教育学院,江苏南京210097 [4]南京师范大学中北学院,江苏丹阳212334

出　　处：《中文信息学报》2023年第6期25-32,共8页Journal of Chinese Information Processing

基　　金：国家社会科学基金(21&ZD288)。

摘　　要：离合现象是指汉语中一种词语的前后语素之间可以插入其他成分,但分离后表达的意思仍然是一个整体的现象。该文采用字符级序列标注方法解决二字动词离合现象的自动识别问题,以避免自动分词及词性标注的错误传递;引入掩码机制,遮蔽句中离合词,以强化对中间插入成分的学习,并对前后语素采用不同的掩码以强调其出现顺序;设计双编码模型,对原始句子与掩码后的句子分别进行编码。实验结果表明,该文提出的BERT_MASK+2BiLSTMs+CRF模型比当前性能最优的离合词识别模型提高了2.85%的F1值。The separated words,as a unique grammatical phenomenon in Chinese,refers to the components of words can be divided by other grammatical constituents without changing the word meaning.This paper proposes a character-level sequence tagging method for automatic recognition of separated two-character verbs via the MASK mechanism.We mask the separated words,and assign different embeddings to the first and second components of a word to emphasize the morpheme sequence order.A double coding model is designed to encode the original sentence and the masked sentence,respectively.Experimental results show that the proposed BERT_MASK+2BiLSTMs+CRF model improves the F_1-value by 2.85%compared with the state-of-the-art model.

关 键 词：离合词 自动识别 掩码机制 神经网络 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]