基于双向LSTM模型的英语语法错误检测  

English grammar error detection based on bidirectional LSTM model

作　　者：马晓伟 MA Xiao-wei(Department of Basic Education,Yantai Automobile Engineering Professional College,Yantai 265500,Shandong Province,China)

机构地区：[1]烟台汽车工程职业学院基础教学部,山东烟台265500

出　　处：《信息技术》2022年第9期56-60,共5页Information Technology

基　　金：全国人文社会科学课题(2019年度)(19QRSK122)。

摘　　要：随着人工智能的兴起,利用计算机辅助英语学习成为了一种趋势。计算机能够提供可靠和有意义的反馈,帮助提高英语水平。因此提出一种基于深度学习的语法错误检测系统,基于双向长短期记忆人工神经网络框架,结合上下文语境对不同水平的英语任务进行研究;并且通过将口语内容转录为文本内容,对其进行语法检测。实验结果表明,提出的方法具有较好的性能,可满足非母语学习者的实际需求。With the rise of Artificial Intelligence,the use of computer-assisted English learning has become a trend.The computers can provide reliable and meaningful feedback to improve English.Therefore,this research proposes a grammatical error detection system based on deep learning.The system is based on a bidirectional Long Short-Term Memory framework and combines context to study English tasks at different levels.And by transcribing the spoken language content into text content,it has also been grammatically tested.Experiment results show that the proposed method has better performance,which could meet the actual needs of non-native language learners.

关 键 词：神经网络 深度学习 语法检测 长短期记忆人工神经网络 人工智能 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]