Vietnamese to Chinese Machine Translation via Chinese Character as Pivot

Vietnamese to Chinese Machine Translation via Chinese Character as Pivot∗ Hai Zhao1,2† Tianjiao Yin3 Jingyi Zhang1,2 (1) MOE-Microsoft Key Laboratory of Intelligent Computing and Intelligent System (2) Department of Computer Science and Engineering, Shanghai Jiao Tong University #800 Dongchuan Road, Shanghai, China, 200240 (3) Facebook, Inc. 1601 Willow Rd. Menlo Park, CA 94025 USA zhaohai@cs.sjtu.edu.cn, ytj000@gmail.com zhangjingyizz@gmail.com Abstract Using Chinese characters as an intermedi- ate equivalent unit, we decompose machine translation into two stages, semantic transla- tion and grammar translation. This strategy is tentatively applied to machine translation between Vietnamese and Chinese. During the semantic translation, Vietnamese sylla- bles are one-by-one converted into the cor- responding Chinese characters. During the grammar translation, the sequences of Chi- nese characters in Vietnamese grammar or- der are modified and rearranged to form grammatical Chinese sentence. Compared to the existing single alignment model, the division of two-stage processing is more tar- geted for research and evaluation of machine translation. The proposed method is evalu- ated using the standard BLEU score and a new manual evaluation metric, understand- ing rate. Only based on a small number of dictionaries, the proposed method gives competitive and even better results com- pared to existing systems. 1 Introduction The statistical machine translation (SMT) has been well developed from a basis of data-drive idea s- ince the work of (Brown et al., 1993). How- ever, a large amount of parallel corpora are al- ways necessary to build a standard SMT system for a specific language pair, regardless of the pos- sible useful linkages between these two languages. There is existing work that considered using help- ful linguistic heuristics to enhance the curren- ∗This work was partially supported by the National Natu- ral Science Foundation of China (Grant No.60903119, Grant No.61170114, and Grant No.61272248), and the National Basic Research Program of China (Grant No.2009CB320901 and Grant No.2013CB329401). †corresponding author Figure 1: The phrase Chinese character culture sphere written in Chinese characters from different regions. t SMT (Chu et al., 2012), though their approach- es still follow the standard processing pipeline of SMT. For those resource-poor languages, a pivot language will be used as an expedience (Utiyama and Isahara, 2007; Wu and Wang, 2009). In this work, we focus on machine translation (MT) for language pairs with few parallel corpora but rich linguistic connections. A case study on Vietnamese and Chinese will be done. To exploit the shared linguistic characteristics between the language pair, the common written form, Chinese character, is adopted as a translation bridge. Be- ing the oldest continuously used writing system in the world, Chinese characters are logograms that are still used to write Chinese (汉字/漢字 in Chi- nese, hànzì in Chinese pinyin) and Japanese (kan- ji). Such characters were used but are currently less frequently used in Korean (hanja), and were also used in Vietnamese (chữ Hán). All the coun- tries that were historically under Chinese language and culture are unofficially referred to Chinese character cultural sphere or Sinosphere. These t- wo terms are often used interchangeably but …

Content truncated to 500 words