基于语义关联度的英语段落机器翻译方法  

English paragraph machine translation method based on semantic relevance

作　　者：梁亚敏[1] 李亚峰 LIANG Yamin;LI Yafeng(Xian Yang Normal University,Xianyang,Shanxi 712000,China)

机构地区：[1]咸阳师范学院,陕西咸阳712000

出　　处：《自动化与仪器仪表》2022年第8期208-212,共5页Automation & Instrumentation

基　　金：省级课题:需求分析理论下大学英语课程设置的教学改革与研究(SGH17H511);校级课题:生态翻译学理论在奥斯汀作品孙致礼译本中的应用与研究(XSYK18030);校级课题:咸阳师范学院2017“青蓝人才”培养项目(XSYQL201703)。

摘　　要：不同国家应用不同母语交际,造成了国际之间交流的语言壁垒,提出基于语义关联度的英语段落机器翻译方法研究。设计编码器处理英语段落信息,选取RelArtNet算法计算英语单词与句子语义关联度,构建英语段落翻译信息调序模型,通过联合训练方法训练神经机器翻译模型,将待翻译的英语段落输入至训练好的翻译模型中,实现了英语段落的机器翻译。实验数据显示:提出方法英语段落机器翻译时间低于最高限值,BLEU指数大于标准数值,证实提出方法应用性能较好。Different countries use different mother tongues to communicate,resulting in language barriers to international communication.a research on English paragraph machine translation method is put forward based on semantic relevance.The encoder is designed to process English paragraph information,the relartnet algorithm is selected to calculate the semantic correlation between English words and sentences,and the English paragraph translation information ordering model is constructed.The neural machine translation model is trained through the joint training method,and the English paragraphs to be translated are input into the trained translation model to realize the machine translation of English paragraphs.The experimental data show that the machine translation time of English paragraphs of the proposed method is lower than the maximum limit,and the Bleu index is greater than the standard value,which proves that the application performance of the proposed method is good.

关 键 词：语义关联度 英语 翻译 段落 机器 半监督学习 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]