一种人机互动的多策略机器翻译系统IHSMTS的设计与实现原理  

The Design and Im plementation Principle of an Interactive Hybrid Strategies Machine Translation System I H S M T S

作　　者：黄河燕[1] 陈肇雄[1] 宋继平[1] 

机构地区：[1]中国科学院计算机语言信息工程研究所中心,北京100083

出　　处：《中文信息学报》1999年第5期43-50,共8页Journal of Chinese Information Processing

基　　金：国家自然科学基金

摘　　要：现有单一策略的机器翻译系统很难有效地解决机器翻译所面临的所有问题。本文，提出一种基于人机交互互动的多策略机器翻译系统设计方法，该方法把基于多知识一体化描述的规则推理、基于经验记忆的类比启发式搜索推理和基于统计知识的概率方法及适当程度的人机交互有机地结合起来，利用现有基于规则的智能机器翻译系统自动产生具有各种特征知识的特征事例模式库，从而既可以通过与以往翻译实例的类比启发式搜索有效地利用以往系统成功的句子分析经验解决相似句子的分析，同时对特征事例模式库中没有相似实例的句子，又可以利用原有基于规则的方法和统计概率方法进行翻译转换处理，并在系统本身的知识不足以解决所遇到的多义区分问题时适时由人介入，从而可以大提高系统的翻译速度和翻译准确率，增强系统的实用性。It is hard for the existed single strategy machine translation approaches to solve all the prob lems in machine translation . In this paper ,we propose an interactive hybrid strategies machine translationapproaches ,which isincorporated with rule based parsing with multiple knowledge description ,analogy in ference with heuristic search based on case translation memory ,statistic knowledge and some extend hu man computerinteraction . In which ,the featured case model base including variousfeaturesis set up auto matically based on the existed intelligent machine translation system . It is not only possible for it to usecase based heuristic analogy reasoning to produce translation resultfor same or similar sentence with a for mer translated sentence ,but also can use rule based parsing to translate the sentences which are notsimilar to any modelin case base . In addition ,when the knowledge is not sufficient to disambiguate the ambi guities in the current translated sentence ,it can use human interacive mode to guide the translation pro cess ,so as to greatly enhance the translation speed and accuracy ,and made the translation system morepracticable .

关 键 词：智能型机器翻译 人机互动 基于事例的机器翻译 类比推理 事例特征 

分 类 号：H085[语言文字—语言学]