摘要 自设计之初,机器翻译面临的一个主要挑战就是对语境中意义的把握。目前采用的交互式机器翻译引擎,试图尽可能地使机器吸收人类的智慧与认知能力,并取得了一定的成果。文章由交互式机器翻译的技术环境入手,根据译者处理语境时付出的认知努力,提出包括本地语境、全局语境、语境效果在内的基于交互式机器翻译环境的语境架构。该等级框架从译者认知的角度出发,能够有效地收集、管理、分析译者反馈的数据,评估翻译系统对译者的依赖程度,从而更有效地在引擎自动切分、人机互动翻译等方面融合语境因素,以不断提高交互式机器翻译引擎的水平。 Since its inception,one of the biggest challenges for machine translation is meaning in context.Nowadays,the fields of artificial intelligence(AI)and human-computer interaction(HCI)are influencing each other like never before.Recent breakthroughs in the translation are made possible by a healthy AI-HCI collaboration.This article proposed a hierarchical structure of context for interactive machine translation environment tools,including local context,global context and contextual effects,based on translators'cognitive efforts when interacting with machines.This framework helps software developers,project managers and linguists who work with the interactive machine translation system better incorporate the contextual factors when collecting,managing and analyzing data from human feedback,which leads to relevant strategic plans for automatic segmentation as well as effective estimation for the degree of human involvement.
机构地区 渥太华大学翻译与口译学院 中山大学南方学院
出处 《北京科技大学学报:社会科学版》 2021年第2期138-146,共9页 Journal of University of Science and Technology Beijing(Social Sciences Edition)
关键词 交互式机器翻译环境 神经机器翻译 语境架构 interactive machine translation environment tools neural machine translation hierarchical structure of context
分类号 H085 [语言文字—语言学]
