Customizing an English-Korean Machine Translation System for Patent Translation

Customizing an English-Korean Machine Translation System for Patent Translation* Sung-Kwon Choi, Young-Gil Kim Natural Language Processing Team, Electronics and Telecommunications Research Institute, 161 Gajeong-dong, Youseong-gu, Daejon, Korea, 305-350 {choisk, kimyk}@etri.re.kr Abstract. This paper addresses a method for customizing an English-to-Korean machine translation system from general domain to patent domain. The customizing method consists of following steps: 1) linguistically studying about characteristics of patent documents, 2) extracting unknown words from large patent documents and constructing large bilingual terminology, 3) extracting and constructing the patent-specific translation patterns 4) customizing the translation engine modules of the existing general MT system according to linguistic study about characteristics of patent documents, and 5) evaluating the accuracy of translation modules and the translation quality. This research was performed under the auspices of the MIC (Ministry of Information and Communication) of Korean government during 2005-2006. The translation accuracy of the customized English-Korean patent translation system is 82.43% on the average in 5 patent fields (machinery, electronics, chemistry, medicine and computer) according to the evaluation of 7 professional human translators. In 2006, the patent MT system started an on-line patent MT service in IPAC (International Patent Assistance Center) under MOCIE (Ministry of Commerce, Industry and Energy) in Korea. In 2007, KIPO (Korean Intellectual Property Office) tries to launch an English-Korean patent MT service. Keywords: Machine Translation, Customization, Patent Machine Translation 1. Introduction An English-Korean machine translation system has been developed in earnest in Korea since 1996. We have applyed it to different areas such as web translation (Choi, 1999) and broadcasting subtitle translation (Choi, 2001). Recently, the natural language processing(NLP) of intellectual property documents is attracting many researchers and NLP-related companies, because NLP techniques associated with specificity of patent domain have promise for improving the translation qualtiy. It is well known that a sentence style and a dominant translation for a word vary with domains. Therefore, if the domain to be translated is fixed to patents, a adaptation of bilingual dictionary to the patent domain and a customization of natural language analyzer to the linguistic specificity of patent style would be one of effective ways to improve the translation quality of MT system. There have been studies concerned specifically with patent MT using these domain- specific advantages (Shinmori et al., 2003; Hong et al., 2005; Kaji, 2005; Shimihata, 2005). * This work was supported by the IT R&D program of MIC/IITA, Domain Customization Machine Translation Technology Development for Korean, Chinese, and English. Copyright 2007 by Sung-Kwon Choi and Young-Gil Kim 105 Though intensive research has been made on patent MT for the domain-specific advantages, there still remain many issues to be tackled. We focus on the several issues that have continusely been problems in existing English-to-Korean MT systems: (1) new terminology construction, (2) patent-specific probabilities of POS tagger, (3) long and complex sentence analysis, and (4) target word selection. This paper addresses the customization of an English-Korean MT system for patent translation. The English-Korean patent MT system described in this paper is based on an English-Korean MT system …

Content truncated to 500 words