English to Chinese Translation: How Chinese Character Matters

English to Chinese Translation: How Chinese Character Matters? Rui Wang1,2, Hai Zhao1,2 ∗ † and Bao-Liang Lu1,2 1Center for Brain-Like Computing and Machine Intelligence, Department of Computer Science and Engineering, Shanghai Jiao Tong University, Shanghai, 200240, China 2Key Laboratory of Shanghai Education Commission for Intelligent Interaction and Cognitive Engineering, Shanghai Jiao Tong University, Shanghai, 200240, China wangrui.nlp@gmail.com and {zhaohai, blu}@cs.sjtu.edu.cn Abstract Word segmentation is helpful in Chinese nat- ural language processing in many aspects. However it is showed that different word seg- mentation strategies do not affect the per- formance of Statistical Machine Translation (SMT) from English to Chinese significant- ly. In addition, it will cause some confu- sions in the evaluation of English to Chinese SMT. So we make an empirical attempt to translation English to Chinese in the charac- ter level, in both the alignment model and lan- guage model. A series of empirical compari- son experiments have been conducted to show how different factors affect the performance of character-level English to Chinese SMT. We also apply the recent popular continuous s- pace language model into English to Chinese SMT. The best performance is obtained with the BLEU score 41.56, which improve base- line system (40.31) by around 1.2 BLEU s- core. ∗Correspondence author. †Thank all the reviewers for valuable comments and sug- gestions on our paper. This work was partially supported by the National Natural Science Foundation of China (No. 61170114, and No. 61272248), the National Basic Re- search Program of China (No. 2013CB329401), the Science and Technology Commission of Shanghai Municipality (No. 13511500200), the European Union Seventh Framework Pro- gram (No. 247619), the Cai Yuanpei Program (CSC fund 201304490199 and 201304490171), and the art and science interdiscipline funds of Shanghai Jiao Tong University, No. 14X190040031, and the Key Project of National Society Sci- ence Foundation of China, No. 15-ZDA041. 1 Introduction Word segmentation is necessary in most Chinese language processing doubtlessly, because there are no natural spaces between characters in Chinese tex- t (Xi et al., 2012). It is defined in this paper as character-based segmentation if Chinese sentence is segmented into characters, otherwise as word seg- mentation. In Statistical Machine Translation (SMT) in which Chinese is target language, few work have shown that better word segmentation will lead to better result in SMT (Zhao et al., 2013; Chang et al., 2008; Zhang et al., 2008). Recently Xi et al. (2012) demonstrate that Chinese character alignment can improve both of alignment quality and translation performance, which also motivates us the hypothe- sis whether word segmentation is not even necessary for SMT where Chinese as target language. From the view of evaluation, the difference be- tween the word-based segmentation methods will al- so makes the evaluation of SMT where Chinese as target language confusing. The automatic evalua- tion methods (such as BLEU and NIST BLEU s- core) in SMT are mostly based on n-gram preci- sion. If the segmentation of test sets are differen- t, the elements of the n-gram of test sets will al- so be …

Content truncated to 500 words