期刊文献+

类人机译系统原理 被引量:10

The Principle of a Human-Like Machine Translation System
在线阅读 下载PDF
导出
摘要 本文从人作为一个翻译系统的特点出发,讨论了类人机译系统主要特点:程序系统独立于语言知识,有很强的通用性、学习能力和重组能力,统一了一些重要方法——语法分析、语义文法、模式匹配和格式文法,翻译过程以语法、语义和环境(条件)统一的最小翻译单位的“最小语言模式”的对应为主、分析为辅,引入与程序块相对应的公用函数及各语言专用函数,立足于不久将来的功能更强、价格更廉的高档微机上,等等.并讨论了这些特点的实现.最后,还讨论了与本系统相适应的语言知识的描述语言. In this paper, we first discuss the main characteristics of a humanlike M. T. System from viewpoint of that human as a M. T. System: program system independent of language knowledge, generality, ability of learning and reorganization; some important methods such as syntax analysis,semantic grammar, match pattern and case grammar are unified; translation procedure is based on minimum language pattern (i. e. minimum translating elements combined with syntax, semantics,pattern and case) assisted by analysis, common functions and special functions of different languages which correspond to subroutines are introduced; the support of system is based on futural workstations, etc. Then, we discuss implementation of those characteristics. At last, the descriptive language of language knowledge in this system is introduced.
出处 《计算机研究与发展》 EI CSCD 北大核心 1989年第2期1-8,共8页 Journal of Computer Research and Development
  • 相关文献

参考文献1

  • 1马希文,1988年

同被引文献69

引证文献10

二级引证文献22

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部