JADE: Corpus for Japanese Definition Modelling

Proceedings of the 13th Conference on Language Resources and Evaluation (LREC 2022), pages 6884–6888 Marseille, 20-25 June 2022 © European Language Resources Association (ELRA), licensed under CC-BY-NC-4.0 6884 JADE: Corpus for Japanese Definition Modelling Han Huang†, Tomoyuki Kajiwara‡, Yuki Arase⋆ †amanotaiga3665@gmail.com ‡Graduate School of Science and Engineering, Ehime University kajiwara@cs.ehime-u.ac.jp ⋆Graduate School of Information Science and Technology, Osaka University arase@ist.osaka-u.ac.jp Abstract This study investigated and released the JADE, a corpus for Japanese definition modelling, which is a technique that automatically generates definitions of a given target word and phrase. It is a crucial technique for practical applications that assist language learning and education, as well as for those supporting reading documents in unfamiliar domains. Although corpora for development of definition modelling techniques have been actively created, their languages are mostly limited to English. In this study, a corpus for Japanese, named JADE, was created following the previous study that mines an online encyclopedia. The JADE provides about 630k sets of targets, their definitions, and usage examples as contexts for about 41k unique targets, which is sufficiently large to train neural models. The targets are both words and phrases, and the coverage of domains and topics is diverse. The performance of a pre-trained sequence-to-sequence model and the state-of-the-art definition modelling method was also benchmarked on JADE for future development of the technique in Japanese. The JADE corpus has been released and available online. Keywords: definition generation, language learning and education, computer-assisted language learning (CALL) 1. Introduction Definition modelling aims to generate a definition of a word and phrase in context. It is a crucial technology for language learning/education support and reading assistance, such that users can refer to meaning of un- familiar expressions on the fly. For developing sophis- ticated definition modelling techniques, a corpus that provides definitions of words and phrases is crucial. Since the meanings of words and phrases are context dependent, such corpus should provide context-aware definitions, i.e. a set of word/phrase, its definition, and context in which the word/phrase is used. Addition- ally, to help language learning and reading, the corpus should be updated with neologisms and trendy entities. To meet these requirements, previous studies proposed to automatically construct a corpus from dictionar- ies (Noraset et al., 2017; Gadetsky et al., 2018) and online encyclopedias (Ni and Wang, 2017; Ishiwatari et al., 2019). These corpora have significantly con- tributed to the development of definition modelling techniques (Mickus et al., 2019; Washio et al., 2019; Li et al., 2020, to name a few). However, despite its im- portance in critical applications, such a corpus is scarce in languages other than English. This study created a corpus for definition modelling in Japanese, JADE (corpus for JApanese DEfinition mod- elling) using Wikipedia1 and Wikidata2 following Ishi- †This study was conducted when the first author was an graduate student at Osaka University. 1https://en.wikipedia.org/wiki/Main_ Page 2https://www.wikidata.org/wiki/ Wikidata:Main_Page watari et al. (2019). Wikipedia has two advantages with regard to creating a corpus for definition mod- elling. The first is the coverage as Wikipedia …

Content truncated to 500 words