基于FrameNet和双层优化的一种译文质量量化评判模式

A Quantitative Model Based on FrameNet and Bilevel Optimization for Translation Quality Assessment

在英语翻译的译文质量评判过程中,对翻译译文的"差错"分类不易清晰界定,由此较难把握译文质量的罚分标准,这样导致对评判结果难以获得统一认可。本文基于FrameNet系统判定翻译中的错误种类,利用双层优化决策思想方法,提出了一种新的翻译质量的量化评判模式,从而使得翻译雇佣行为中的三方(翻译服务提供者、翻译雇佣者与翻译质量检查者)对评判结果有可能获得均较为满意的统一认可。

While assessing the quality of E-C translation samples, the classification in determining the type of errors are often not easily decided. Thus, the discrepancies and arbitrariness within current scoring standards has resulted in partial recognition. To solve the aforementioned problem, this paper argues a new quantitative model for translation quality assessment by utilizing Bilevel Optimization Methods to aid the feasibility of fulfilling the unanimous satisfaction of all parties(translate service providers, employers, and QA personnel) involved in the process of translation service. FrameNet, developed by UC Berkeley, is used as a determiner of error types.