摘要
考虑到传统系统在检测英语机器翻译微小误差时的功能和性能无法满足设计要求,提出了基于CNN(Convolutional Neural Network)的英语机器翻译微小误差检测系统设计。基于英语机器翻译微小误差检测系统总体架构,利用卷积神经网络设计了英语机器翻译微小误差解码器隐层单元,结合英语机器翻译微小误差检测器的设计,完成系统的硬件设计;在系统的原理设计中,根据提取出的英语机器翻译微小误差特征,设计了英语机器翻译微小误差检测算法,实现了英语机器翻译的微小误差检测。系统测试结果表明,系统的主页功能和微小误差检测功能满足设计要求,还可以通过提高英语机器翻译微小误差检测覆盖率和缩短检测耗时,提高系统的性能。
Considering that the function and performance of the traditional system can’t meet the design requirements when detecting small errors in English machine translation,a design of small errors detection system for English machine translation based on CNN is proposed.Based on the overall architecture of English machine translation micro-error detection system,the hidden layer unit of English machine translation micro-error decoder is designed by using convolutional neural network,and the hardware design of the system is completed by combining with the design of English machine translation micro-error detector.In the principle design of the system,according to the extracted small error characteristics of English machine translation,the small error detection algorithm of English machine translation is designed to realize the small error detection of English machine translation.The system test results show that the homepage function and micro-error detection function of the system in this paper meet the design requirements,and the system performance can be improved by improving the coverage of micro-error detection in English machine translation and shortening the detection time.
作者
赵崇俊
ZHAO Chongjun(Xianyang Normal University,Xianyang Shanxi 712000,China)
出处
《自动化与仪器仪表》
2022年第4期210-213,共4页
Automation & Instrumentation
基金
陕西省教育科学“十三五”规划2020年度课题:大学英语教学《指南》视域下教师隐性课程研究(SGH20Y1243)
陕西省体育局科研常规课题:奥运会全运会视域下城市竞技赛事语言环境建设研究(2021249)
陕西省哲学社会科学重大理论与现实问题研究重点项目:“一带一路”背景下师范类高校大学英语精品教材资源建设研究(2021HZ-842)。
关键词
卷积神经网络
机器翻译
解码器
检测系统
特征提取
微小误差
convolutional neural network
machine translation
decoder
detection system
feature extraction
minor error
