This study presents an initial version of an e-learning system that assists learners of Japanese with their study of vocabulary. The system uses sentences from a corpus to generate context-based exercises. The sentences used in the context-based exercises are selected using a readability formula developed for this system. We used the system with two different types of corpora, a web corpus that we constructed for this system and a sample of the recently released Balanced Corpus of Contemporary Written Japanese (BCCWJ). We compared the two corpora and while the BCCWJ has better word coverage, our web corpus still covered a majority (96.1%) of the target vocabulary words even though it’s relatively small. Evaluation of this system showed that the readability formula performs well, especially when sentences contain the system’s target set of vocabulary words. A group of learners of Japanese were also asked to use the system and then fill out a survey. Results of the survey indicate that the learners thought the system was easy to use. Most of the learners also expressed a desire to use this type of system when studying vocabulary.

View full abstract

A systematic account of Japanese language modality forms as well as distant collocations between modal adverbs and clause-final modality forms is lacking in the field of natural language processing. The same stands for coverage of this kind of linguistic information in Japanese language education. In order to remedy this deficiency, in this paper we make extraction of Japanese adverbs and clause-final modality forms collocations possible using the corpus query system Sketch Engine and examine possibilities for its application in Japanese language learning, focusing on learner’s dictionaries. First, as a result of analyzing various Japanese language corpora, we create a long list of modality forms and their variations. Then, we examine how ChaSen morphologically analyzes the forms and retag a sample of the large-scale Japanese language web corpus, JpWaC, by grouping all morphemes that correspond to individual modality forms together under a new modality tag. Finally, we load the newly tagged corpus into the Sketch Engine (SkE), modify the gramrel file and as a result obtain Word Sketch results for collocations between suppositional adverbs and modality forms. The evaluation of the collocation results shows that the proposed method reaches accuracy of above 93%. The results can be utilized in the creation of Japanese learners’ dictionaries or other language material or directly in language teaching or learning.

View full abstract

Introduced in this paper is an English learner corpus built for the R & D of an e-Learning system. Analysis and application experiments of the corpus are also shown. The corpus consists of English sentences that were translated from Japanese by Japanese English learners. Each of them translated 300 Japanese sentences into English. Their English proficiencies were measured through TOEIC. Reference sentences, translated by bilinguals, were also collected for automatic evaluation of the translation quality. In the experiments, automatic scores such as BLEU, NIST, WER, PER, METEOR and GTM were used. According to the experimental results, GTM gives the highest correlation, 0.74 for an automatic score and TOEIC. By adding 4 parameters (sentence length, word length of the translation of the English learners, etc.) for the multiple linear regression analysis, the correlation improves to 0.76.

View full abstract

This paper proposed a mutual teaching model for assisting students in writing compositions, and implemented a writing aid system as a Web application based on the model where students, teachers and our system teach each other their knowledge of writing. We designed the system to use in first language writing courses in the university. The existing systems have two problems: (a) the limitation of assistance for structure or contents of composition, (b) few mechanisms that allow teachers to incorporate their educational objectives into systems. In our proposed model, a student annotates on his/her own composition and makes comments on other’s compositions. And teachers define “Composition Rules” for incorporating their educational objectives into systems. Using the rules and results of the annotation, our system provides various assistance for also structure or contents of composition. By the proposed model and a coventional model, we made two composition experiments whose results showed the effectiveness of the proposed model and “Composition Rules”.

View full abstract

The purpose of this study is to develop an issue-oriented automatic syllabus categorization system, in which natural language processing and machine learning based automatic categorization are combined. Recent explosion of scientific knowledge due to the rapid advancement of academia and society makes it difficult for learners and educators to recognize overall picture of syllabus. In addition, the growing number of interdisciplinary researches makes it harder for learners to find their proper subjects from the syllabi. In an attempt to present clear directions to their subjects, issue-oriented syllabus structure is expected to be more efficient in learning and education. However, it normally requires categorizing all the syllabi manually in advance, and it is generally a time consuming task. Thus, this emphasizes the importance of developing efficient methods for (semi-) automatic syllabus structuring in order to accelerate syllabus retrieval. In this paper, we introduce design and implementation of an issue-oriented automatic syllabus categorization system. And preliminary experiments using more than 850 engineering syllabi of University of Tokyo show that our proposed syllabus categorization system obtains sufficient accuracy.

View full abstract