ASPEC: Asian Scientific Paper Excerpt Corpus

ASPEC: Asian Scientific Paper Excerpt Corpus Toshiaki Nakazawa∗, Manabu Yaguchi∗, Kiyotaka Uchimoto∗∗, Masao Utiyama∗∗, Eiichiro Sumita∗∗ Sadao Kurohashi†, Hitoshi Isahara†† ∗Japan Science and Technology Agency 5-3, Yonbancho, Chiyoda-ku, Tokyo, 102-8666, Japan nakazawa@pa.jst.jp, yaguchi@jst.go.jp ∗∗National Institute of Information and Communications Technology 3-5 Hikaridai, Seika-cho, Soraku-gun, Kyoto, 619-0289, Japan {uchimoto, mutiyama, eiichiro.sumita}@nict.go.jp †Kyoto University Yoshida-honmachi, Sakyo-ku, Kyoto, 606-8501, Japan kuro@i.kyoto-u.ac.jp ††Toyohashi University of Technology 1-1 Hibarigaoka, Tempaku, Toyohashi, Aichi, 441-8580, Japan isahara@tut.jp Abstract In this paper, we describe the details of the ASPEC (Asian Scientific Paper Excerpt Corpus), which is the first large-size parallel corpus of scientific paper domain. ASPEC was constructed in the Japanese-Chinese machine translation project conducted between 2006 and 2010 using the Special Coordination Funds for Promoting Science and Technology. It consists of a Japanese-English scientific paper abstract corpus of approximately 3 million parallel sentences (ASPEC-JE) and a Chinese-Japanese scientific paper excerpt corpus of approximately 0.68 million parallel sentences (ASPEC-JC). ASPEC is used as the official dataset for the machine translation evaluation workshop WAT (Workshop on Asian Translation). Keywords: parallel corpus, scientific paper, asian languages 1. Introduction The parallel corpus is no doubt an essential resource for almost all the current machine translation systems. There are many parallel corpora available online1 for various lan- guage pairs and domains. However, there is no available parallel corpus which includes a variety of scientific paper domains so far. There are plenty of useful scientific and technical docu- ments which are written in languages other than English, and are referenced domestically. Accessing these domestic documents in other countries is very important in order to know what has been accomplished and what is needed next in the science and technology fields. However, we need to surmount the language barrier to directly access these valu- able documents. One obvious way to achieve this is using machine translation systems to translate foreign documents into the users’ language. ASPEC (Asian Scientific Paper Excerpt Corpus) is the first large-size parallel corpus of scientific paper domain in the world2. This corpus is constructed by the Japan Science and Technology Agency (JST) in collaboration with the Na- tional Institute of Information and Communications Tech- nology (NICT). It consists of a Japanese-English scientific paper abstract corpus of approximately 3 million parallel sentences (ASPEC-JE) and a Chinese-Japanese scientific paper excerpt corpus of approximately 0.68 million paral- 1http://www.statmt.org/moses/?n=Moses.LinksToCorpora 2It is released in January, 2014 lel sentences (ASPEC-JC). ASPEC is free and available on- line3 for non-commercial purposes. There are some related work which try to automatically ex- tract scientific paper domain parallel corpus using scientific paper Web portals such as MEDLINE (Jimeno Yepes et al., 2013) and Elsevier (Morin and Prochasson, 2011). How- ever the corpus size is not large enough to achieve the sat- isfactory machine translation quality of the paper abstracts. In addition, the domain of the corpora are limited. They do not contain a variety of scientific paper domains. In this paper, we describe the details of the ASPEC and briefly introduce the Workshop on Asian Translation (WAT) as an …

Content truncated to 500 words