深度学习技术在英译汉即时自动评分中的尝试

Deep Learning Technology in Instant Automated Scoring of Online English-Chinese Translation

本文探讨了将深度学习技术之一的卷积神经网络应用于在线英译汉即时自动评分系统的研制。针对系统的评分需要,本研究建设了在线翻译自动评分的网站以及移动平台,研制了训练评分模型所需要的语料库及标注规范。实验表明,将深度学习技术用于在线英译汉即时自动评分系统具有可行性。自动评分效果可以基本满足在线英译汉自测等低风险测试的需求。

This paper discusses the application of convolutional neural networks, one of the deep learning techniques, to the development of an instant automated scoring system of online English-Chinese translation. In response to the scoring needs of the system, a website and a mobile platform for automated scoring of online translation have been built, and the corpus and annotation specifications required for training the scoring model have been developed. The experiment shows the feasibility of applying deep learning technology to the instant automated scoring system of online English-Chinese translation. The automated scoring can basically meet the needs of low-risk tests such as online self-test of English-Chinese translation.