摘要
近两年来,神经机器翻译(Neural Machine Translation,NMT)模型主导了机器翻译的研究,但是统计机器翻译(Statistical Machine Translation,SMT)在很多应用场合(尤其是专业领域)仍有较强的竞争力。如何利用深度学习技术提升现有统计机器翻译的水平成为研究者们关注的主要问题。由于语言模型是统计机器翻译中最核心的模块之一,本文主要从语言模型的角度入手,探索神经网络语言模型在统计机器翻译中的应用。本文分别探讨了基于词和基于短语的神经网络语言模型,在汉语到英语和汉语到日语的翻译实验表明神经网络语言模型能够显著改善统计机器翻译的译文质量。
Neural Machine Translation （NMT） dominates the research of machine translation in recent two years. However, Statistical Machine Translation （SMT） is very competitive in many scenarios such as some specific domains. It became a key issue how to apply deep learning technology to improve SMT performance. As language model is one of the most crucial modules in SMT, this paper investigated the usage of neural language model in statistical machine translation. We explored respectively the word-based and phrase- based neural language model, and evaluated the models on both Chinese-to-English and Chinese-to-Japanese translation tasks. The extensive experiments demonstrated that the neural language models can significantly improve the translation performance of statistical machine translation.
作者
张家俊
宗成庆
ZHANG JiaJun ZONG ChengQing(National Laboratory of Pattern Recognition, Institute of Automation Chinese Academy of Sciences University of Chinese Academy of Sciences, Beijing 100190, China CAS Center for Excellence in Brain Science and Intelligence Technology, Shanghai 200031, China)
出处
《情报工程》
2017年第3期21-28,共8页
Technology Intelligence Engineering
基金
国家自然科学基金“视听觉信息的认知计算”重点项目:面向汉语文本理解的语义计算方法(91520204)
国家自然科学基金面上项目:基于弱监督的神经网络翻译模型研究(61673380)的资助
