A Web-based English Abstract Writing Tool Using a Tagged E-J Parallel Corpus

Masumi Narita (Software Research Center, Ricoh Co., Ltd., 1-1-17 Koishikawa, Bunkyo-ku, Tokyo, Japan)

Kazuya Kurokawa (Computational Linguistics Laboratory, Department of Information and Computer Sciences, Toyohashi University of Technology, 1-1 Hibarigaoka, Tempaku-cho, Toyohashi, Japan)

Takehito Utsuro (Computational Linguistics Laboratory, Department of Information and Computer Sciences, Toyohashi University of Technology, 1-1 Hibarigaoka, Tempaku-cho, Toyohashi, Japan)

WO24: Applications Based On Written LRs

In this paper, we present a Web-based English abstract writing tool, the "BEAR (Building English Abstracts by Ricoh)." This English writing tool is aimed at helping Japanese software engineers improve the organization of their writing by enabling them to select a rhetorical template of the target abstract and to build up component sentences while having access to good-quality sample sentences. To provide this kind of language assistance, we constructed an E-J parallel corpus of 539 sample abstracts as the core language resource. After analyzing the rhetorical structure of these sample abstracts, we tagged the corpus with textual and linguistic information. The "BEAR" is not designed for beginners but for intermediate to advanced EFL learners who very often need to write a research paper or a technical report in English. Software development has not yet been completed, but we have already gathered some user feedback at preliminary user trials. We show that the "BEAR" has been positively evaluated by our users and thus our tagged E-J parallel corpus of sample abstracts can support our users in the difficult task of working with a foreign language. We also discuss the outlook for further development of the "BEAR." 

BEAR (Building English Abstracts by Ricoh)

137.pdf