基于英语翻译应用视角下的计算机智能校对系统开发研究
2020-05-11孙瑞



摘 要: 随着人工智能等计算机相关技术的发展,利用计算机进行英语翻译工作也逐渐成为研究方向之一,受限于算法及匹配度等因素影响,英语翻译仍存在瓶颈。因此,提出了基于英语翻译应用视角的计算机智能校对模型并进行开发研究,该模型能够在单词级别上直接整合附加注释——包括语言标记或自动生成的单词类。基于此模型的实验表明,基于英语翻译应用视角下的计算机智能校对模型可以在自动得分和更多语法连贯性方面带来更好的翻译表现,最后还从硬件、软件两方面探讨了计算机智能校对系统开发。
关键词: 英语翻译; 智能校对; 训练样本
中图分类号: TP311 文献标志码: A
Research on the Development of Computer Intelligent Proofreading System
Based on the Perspective of English Translation Application
SUN Rui
(Xian Innovation College, Yanan University, Xian 710100)
Abstract: With the development of computer-related technologies such as artificial intelligence, the use of computers for English translation has gradually become one of the research directions. Due to factors such as algorithm and matching degree, English translation still has bottlenecks. Therefore, this paper proposes a computer intelligent proofing model based on the perspective of English translation application and conducts research and development. This model can directly integrate additional annotations at the word level, including language markers or automatically generated word classes. Experiments based on this model show that the computer intelligent proofing model based on the perspective of English translation application can bring better translation performance in terms of automatic score and more grammatical coherence. Finally, the development of computer intelligent proofreading system is discussed from both hardware and software aspects.
Key words: English translation; Intelligent proofreading; Training samples
0 前言
随着人工智能等计算机相关技术的发展,利用计算机进行英语翻译工作也逐渐成为研究方向之一[1、2]。当前最先进的计算机机器翻译方法,即所谓基于短语的模型,但仅限于小文本块的映射,由于没有明确使用语言信息,可能是形态学、句法或语义。通过将其集成到预处理或后处理步骤中,已经证明这些附加信息是有价值的[3-5]。一般而言出于两类原因,需要将语言信息更紧密地整合到翻译模型中:以更一般表示形式操作的翻译模型,例如词条而不是表面形式的词,可以利用更丰富的统计数据并克服由有限的训练样本引起的数据稀疏性问题,翻译的许多方面可以在形态学、句法或语义层面上得到最好的解释[6]。将这些信息提供给翻译模型可以直接建模这些方面,例如:句子级别的重新排序主要由一般句法原则驱动,局部协议约束在形态学中出现等,然而这些翻译却只能针对短语有效[7]。……
