| 摘要: |
| 为了提高基于短语的机器翻译系统的重排序能力,提出了一个基于源语言端的中心-修饰依存结构的重排序模型,并将该重排序模型以软约束的方式加入到机器翻译系统中.该排序模型提出了一种在机器翻译中应用句法树资源的方法,将句法树结构,通过将句法树映射成中心-修饰词的依存关系集合.该重排序模型在基于短语系统的默认参数设置下,显著地提升了系统的翻译质量.在系统原有的词汇化的重排序模型基础上,该重排序模型在翻译模型中融入了句法信息.实验结果显示,该模型可以明显地改善机器翻译系统的局部调序. |
| 关键词: 短语机器翻译 重排序模型 中心修饰依存关系 无词汇化 |
| DOI：10.3724/SP.J.1001.2012.04055 |
| 分类号: |
| 基金项目:国家自然科学基金(60603032); 国家高技术研究发展计划(863)(2006AA010108) |
|
| Head-Modifier Dependency Reordering Model for Phrased SMT |
|
LIU Shui1, LI Sheng1, ZHAO Tie-Jun1, LIU Peng-Yuan2
|
|
1.Department of Computer Science and Technology, Harbin Institute of Technology, Harbin 150001, China;2.Applied Linguistics Research Institute of Beijing Language and Culture University, Beijing 100083, China
|
| Abstract: |
| To enhance the reordering capacity of the phrase-based SMT (statistical machine translation), the study leverages the head-modifier dependency structure on the source to model the reordering. The model is added to baseline model in the form of soft-constraint way. The proposed model explores an approach to utilize the constituent based parse tree that the parse tree is mapped into sets of head-modifier relationships. Experimental results show that this model improves the local reordering significantly. |
| Key words: phrase-based SMT (statistical machine translation) reordering model head-modifier relationship non-lexicalized |
