Chinese Syntactic Reordering for Adequate Generation of Korean Verbal Phrases in Chinese-to-Korean SMT

Proceedings of the Fourth Workshop on Statistical Machine Translation , pages 190–196, Athens, Greece, 30 March – 31 March 2009. c⃝2009 Association for Computational Linguistics Chinese Syntactic Reordering for Adequate Generation of Korean Verbal Phrases in Chinese-to-Korean SMT Jin-Ji Li, Jungi Kim, Dong-Il Kim*, and Jong-Hyeok Lee Department of Computer Science and Engineering, Electrical and Computer Engineering Division, Pohang University of Science and Technology (POSTECH), San 31 Hyoja Dong, Pohang, 790-784, R. of Korea E-mail: {ljj, yangpa, jhlee}@postech.ac.kr *Language Engineering Institute, Department of Computer, Electron and Telecommunication Engineering, Yanbian University of Science and Technology (YUST), Yanji, Jilin, 133-000, P.R. of China E-mail: {dongil}@ybust.edu.cn Abstract Chinese and Korean belong to different lan- guage families in terms of word-order and morphological typology. Chinese is an SVO and morphologically poor language while Ko- rean is an SOV and morphologically rich one. In Chinese-to-Korean SMT systems, systemat- ic differences between the verbal systems of the two languages make the generation of Ko- rean verbal phrases difficult. To resolve the difficulties, we address two issues in this paper. The first issue is that the verb position is dif- ferent from the viewpoint of word-order ty- pology. The second is the difficulty of com- plex morphology generation of Korean verbs from the viewpoint of morphological typology. We propose a Chinese syntactic reordering that is better at generating Korean verbal phrases in Chinese-to-Korean SMT. Specifi- cally, we consider reordering rules targeting Chinese verb phrases (VPs), preposition phrases (PPs), and modality-bearing words that are closely related to Korean verbal phras- es. We verify our system with two corpora of different domains. Our proposed approach significantly improves the performance of our system over a baseline phrased-based SMT system. The relative improvements in the two corpora are +9.32% and +5.43%, respectively. 1 Introduction Recently, there has been a lot of research on en- coding syntactic information into statistical ma- chine translation (SMT) systems in various forms and in different stages of translation processes. During preprocessing source language sen- tences undergo reordering and morpho-syntactic reconstruction phases to generate more target language-like sentences. Also, fixing erroneous words, generating complex morphology, and re- ranking translation results in post-processing phases may utilize syntactic information of both source and target languages. A syntax-based SMT system encodes the syntactic information in its translation model of the decoding step. A number of researchers have proposed syn- tactic reordering as a preprocessing step (Xia and McCord, 2004; Collins et al., 2005; Wang et al., 2007). In these syntactic reordering approaches, source sentences are first parsed and a series of reordering rules are applied to the parsed trees to reorder the source sentences into target language- like word orders. Such an approach is an effec- tive method for a phrase-based SMT system that employs a relatively simple distortion model in the decoding phase. This paper concentrates upon reordering source sentences in the preprocessing step of a Chinese-to-Korean phrase-based SMT system using syntactic information. Chinese-to-Korean SMT has more difficulties than the language pairs studied in previous research (French- English, German-English, …

Content truncated to 500 words