Robust and accurate analysis of pronunciation based on structural representation of speech and machine learning and its application to foreign language learning

2014 Fiscal Year Final Research Report

• Project/Area Number: 23300067
• Research Category: Grant-in-Aid for Scientific Research (B)
• Allocation Type: Single-year Grants
• Section: 一般
• Research Field: Perception information processing/Intelligent robotics
• Research Institution: The University of Tokyo
• Principal Investigator: MINEMATSU Nobuaki 東京大学, 工学(系)研究科(研究院), 教授 (90273333)
• Co-Investigator(Kenkyū-buntansha): 牧野 武彦 中央大学, 経済学部, 教授 (00269482)
• Co-Investigator(Renkei-kenkyūsha): 山内 豊 東京国際大学, 商学部, 教授 (30306245)
• Project Period (FY): 2011-04-01 – 2015-03-31
• Keywords: 音声の構造的表象 / 機械学習 / 外国語学習 / 発音教育 / 外国語訛り / 世界諸英語

Outline of Final Research Achievements

In this study, pronunciation analysis based on structural representation of speech was investigated and applied to foreign language learning, especially to pronunciation training. The following four results were obtained as contribution. 1) development of a technique to perform structural analysis of pronunciation even for a small number of utterances, 2) extension of a non-native speech corpus by labeling the utterances with IPA (International Phonetic Alphabets), 3) individual-based and precise clustering of native and non-native pronunciations of English, and 4) preliminary investigation of educationally-effective visualization of pronunciation diversity of English. In study 1), Chinese was adopted as target language and in studies 2), 3), and 4), English was used. The technical contributions of these studies are, however, language-independent and they can be applied to any language.

Free Research Field

音声工学