基于语义匹配的外语翻译机器人自动问答检索研究  

Research on automatic question answering retrieval of foreign language translation robot based on semantic matching

作　　者：李星[1] LI Xing(Xianyang Normal University,Xianyang,Shaanxi 712000,China)

机构地区：[1]咸阳师范学院,咸阳712000

出　　处：《自动化与仪器仪表》2022年第2期138-141,共4页Automation & Instrumentation

基　　金：陕西省教育厅科学研究计划项目:新中日关系与新“国标”下地方高校日语专业就业前景与对策研究(项目编号:19JK0921);咸阳师范学院“青年骨干教师”培养项目(项目编号:XSYGG201904)。

摘　　要：为提高外语翻译机器人自动问答的准确率,提出一种基于TF-IDF+语义匹配+深度学习的问答匹配方法。为提高问题检索的准确率,采用TF-IDF算法关键词匹配,以筛选出问题回复集;基于seq2seq模型进行语义相似度计算,以产生问题回复集,引入Dual-Encoder评分的方式筛选出最佳回复答案;构建检索回复的外语翻译机器人系统。通过搭建TensorFlow的测试环境进行测试结果表明,相较于其他匹配模型,构建的检索模型的匹配准确率更高,且系统性能更好,可实现外语翻译机器人的精准检索对话。To improve the translation accuracy of the foreign language translation robot, a foreign language translator system based on rule matching is proposed.first, Specific analysis of the TF-IDF algorithm, Realize the keyword matching, Then the sequence-to-sequence-based model, Performing the semantic similarity calculations, Then collecting the features of deep learning, Using it to score the output information, Thus to establish a retrieval model;Then a foreign language translation robot system is constructed based on this model, And the system to design and implementation of specific functions;Finally, the retrieval model and the translation robot system were tested, The test results showed that In contrast to the other matching models, The study-constructed search model had a higher matching accuracy, And a better system performance, It can realize the accurate retrieval dialogue of the foreign language translation robot.

关 键 词：规则匹配 外语翻译 机器人 TF-IDF算法 检索对话 

分 类 号：TP392[自动化与计算机技术—计算机应用技术]