Machine Translation vs. Dictionary Term Translation - a Comparison for English-Japanese News Article Alignment

Machine Translation vs. Dictionary Term Translation - a Comparison for English-Japanese News Article Alignment Nigel Collier, Hideki Hirakawa and Akira Kumano Communication and Information Systems Laboratories Research and Development Center, Toshiba Corporation 1 Komukai Toshiba-cho, Kawasaki-shi, Kanagawa 210-8582, Japan {nigel, hirakawa, kmn}@eel, rdc. to shiba, co. j p Abstract Bilingual news article alignment methods based on multi-lingual information retrieval have been shown to be successful for the automatic production of so-called noisy-parallel corpora. In this paper we compare the use of machine translation (MT) to the commonly used dictionary term lookup (DTL) method for Reuter news article alignment in English and Japanese. The results show the trade-off be- tween improved lexical disambiguation provided by machine translation and extended synonym choice provided by dictionary term lookup and indicate that MT is superior to DTL only at medium and low recall levels. At high recall levels DTL has su- perior precision. 1 Introduction In this paper we compare the effectiveness of full ma- chine translation (MT) and simple dictionary term lookup (DTL) for the task of English-Japanese news article alignment using the vector space model from multi-lingual information retrieval. Matching texts depends essentially on lexical coincidence between the English text and the Japanese translation, and we see that the two methods show the trade-off be- tween reduced transfer ambiguity in MT and in- creased synonymy in DTL. Corpus-based approaches to natural language pro- cessing are now well established for tasks such as vo- cabulary and phrase acquisition, word sense disam- biguation and pattern learning. The continued prac- tical application of corpus-based methods is crit- ically dependent on the availability of corpus re- sources. In machine translation we are concerned with the provision of bilingual knowledge and we have found that the types of language domains which users are interested in such as news, current affairs and technology, are poorly represented in today's pub- lically available corpora. Our main area of interest is English-Japanese translation, but there are few clean parallel corpora available in large quantities. As a result we have looked at ways of automatically acquiring large amounts of parallel text for vocabu- lary acquisition. The World Wide Web and other Internet re- sources provide a potentially valuable source of par- allel texts. Newswire companies for example pub- lish news articles in various languages and various domains every day. We can expect a coincidence of content in these collections of text, but the de- gree of parallelism is likely to be less than is the case for texts such as the United Nations and par- liamentary proceedings. Nevertheless, we can expect a coincidence of vocabulary, in the case of names of people and places, organisations and events. This time-sensitive bilingual vocabulary is valuable for machine translation and makes a significant differ- ence to user satisfaction by improving the compre- hensibility of the output. Our goal is to automatically produce a parallel corpus of aligned articles from collections of English and Japanese news texts for bilingual vocabulary ac- quisition. The first stage in …

Content truncated to 500 words