• Search Research Projects
  • Search Researchers
  • How to Use
  1. Back to previous page

An Application of Deep Learning to generate Simplified Japanese by using "Surfece Charancteristics" of text.

Research Project

Project/Area Number 19K12247
Research Category

Grant-in-Aid for Scientific Research (C)

Allocation TypeMulti-year Fund
Section一般
Review Section Basic Section 62030:Learning support system-related
Research InstitutionKobe University

Principal Investigator

Hajime Murao  神戸大学, 国際文化学研究科, 教授 (70273761)

Project Period (FY) 2019-04-01 – 2023-03-31
Project Status Completed (Fiscal Year 2022)
Budget Amount *help
¥4,160,000 (Direct Cost: ¥3,200,000、Indirect Cost: ¥960,000)
Fiscal Year 2021: ¥1,040,000 (Direct Cost: ¥800,000、Indirect Cost: ¥240,000)
Fiscal Year 2020: ¥1,690,000 (Direct Cost: ¥1,300,000、Indirect Cost: ¥390,000)
Fiscal Year 2019: ¥1,430,000 (Direct Cost: ¥1,100,000、Indirect Cost: ¥330,000)
Keywordsやさしい日本語 / 機械学習 / 機械翻訳 / 学習支援 / 深層学習 / BERT / Transformer / GAN / 文章の表面的特徴 / 自然言語処理 / 日本語教育
Outline of Research at the Start

本研究では,表面的な特徴に基づいて「難解な日本語」の文章を「やさしい日本語」の文章に自動変換するシステムの試作および評価を行う。このために深層学習の一種であるVRAEを用いる。研究は,まず,「やさしい日本語」の「表面的な特徴」を明らかにし,その「表面的な特徴」を「内容」と分離して抽出する手法を構築,さらに抽出した「表面的な特徴」を適用して,難解な文章を「やさしい日本語」に変換する手法を構築する。

Outline of Final Research Achievements

We tried to construct a system to translate ordinary Japanese texts into simplified ones by using the "surface characteristics" of texts, which are not semantic features but lexical features such as how separators like commas, periods, and spaces were used. As a result, we achieved the following: 1. we clarified the difference in the surface characteristics between regular and simplified texts. 2. We constructed a system to evaluate the difficulty of Japanese texts, showing a pretty good result of over 90% accuracy. 3. We employed T5 to convert standard Japanese texts into simplified ones.

Academic Significance and Societal Importance of the Research Achievements

本研究により,日本語テキストの「やさしさ」を自動的に判定することが可能となり,また,通常文をやさしい日本語に変換する可能性が示された。これにより,やさしい日本語に関する特別な知識がなくとも,子どもや日本に滞在する外国人に対して情報提供を行うことができる。また,従来より研究されてきた,意味・内容に基づく文章変換と組み合わせることにより,さらに精度を高めることが可能となり,より広範に適用できる可能性がある。

Report

(5 results)
  • 2022 Annual Research Report   Final Research Report ( PDF )
  • 2021 Research-status Report
  • 2020 Research-status Report
  • 2019 Research-status Report
  • Research Products

    (9 results)

All 2023 2022 2021 2020 2019

All Journal Article (6 results) (of which Peer Reviewed: 6 results,  Open Access: 2 results) Presentation (3 results) (of which Int'l Joint Research: 2 results)

  • [Journal Article] A PROPOSAL TO CREATE A PSEUDO-PARALLEL TEXT CORPUS FOR SIMPLIFYING JAPANESE USING DTW2023

    • Author(s)
      Eri Maekawa, Hajime Murao
    • Journal Title

      INTED2023 Proceedings (The Proc. of the 17th Int. Technology, Education and Development Conf.)

      Volume: 1 Pages: 6542-6550

    • DOI

      10.21125/inted.2023.1745

    • Related Report
      2022 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] A study on analyzing differences between native Japanese speakers and non-native speakers based on facial muscle EMG signals2022

    • Author(s)
      Jiawen Xu, Hajime Murao
    • Journal Title

      The Proceedings of the 16th International Conference on Innovative Computing, Information and Control (ICICIC2022)

      Volume: -

    • Related Report
      2022 Annual Research Report
    • Peer Reviewed
  • [Journal Article] Interpreting BERT Attention Trained for Japanese Difficulty Classification from the Viewpoint of Grammatical Features2022

    • Author(s)
      Eri Maekawa and Hajime Murao
    • Journal Title

      ICIC Express Letters, Part B: Applications

      Volume: 13 Issue: 07 Pages: 697

    • DOI

      10.24507/icicelb.13.07.697

    • ISSN
      2185-2766
    • Related Report
      2022 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] The Comparison of Word Embeddings and Feature Vectors in Text Classification by Difficulty Level2021

    • Author(s)
      Eri Maekawa, Hajime Murao
    • Journal Title

      Proceedings of the 15th International Conference on Innovative Computing, Information and Control (ICICIC2021)

      Volume: Online

    • Related Report
      2021 Research-status Report
    • Peer Reviewed
  • [Journal Article] A Study on Finding Differences in Movement of Expert and Novice Darts Players by Using a Kinect-Like 3D Image Sensor2019

    • Author(s)
      Hajime Murao
    • Journal Title

      Proc. of the 14th International Conf. on Innovative Computing, Information and Control

      Volume: ICICIC2019-065 Pages: 1-6

    • NAID

      40022249144

    • Related Report
      2019 Research-status Report
    • Peer Reviewed
  • [Journal Article] Estimating Desk Work Status from Video Stream Using a Deep Neural Network2019

    • Author(s)
      Megumi Kawata, Hajime Murao
    • Journal Title

      Proc. of the 14th International Conf. on Innovative Computing, Information and Control

      Volume: ICICIC2019-159 Pages: 1-4

    • NAID

      40022434305

    • Related Report
      2019 Research-status Report
    • Peer Reviewed
  • [Presentation] 日本語の難易度に関する特徴分析2021

    • Author(s)
      前川 絵吏, 村尾 元
    • Organizer
      言語処理学会 第27回年次大会
    • Related Report
      2020 Research-status Report
  • [Presentation] Analysis of the Behavior of Foreign Tourists Using Mobile Translation devices2020

    • Author(s)
      Eri Maekawa, Hajime Murao
    • Organizer
      The SICE Annual Conference 2020
    • Related Report
      2020 Research-status Report
    • Int'l Joint Research
  • [Presentation] Study on the Effect of Appearance of Personified Agents in Persuation2020

    • Author(s)
      Megumi Kawata, Hajime Murao
    • Organizer
      The SICE Annual Conference 2020
    • Related Report
      2020 Research-status Report
    • Int'l Joint Research

URL: 

Published: 2019-04-18   Modified: 2024-01-30  

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi