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

Neural Networks Deep Learning Methods Utilizing Singular Regions

Research Project

Project/Area Number 16K00342
Research Category

Grant-in-Aid for Scientific Research (C)

Allocation TypeMulti-year Fund
Section一般
Research Field Soft computing
Research InstitutionChubu University

Principal Investigator

NAKANO Ryohei  中部大学, 工学部, 客員教授 (90324467)

Project Period (FY) 2016-04-01 – 2019-03-31
Project Status Completed (Fiscal Year 2018)
Budget Amount *help
¥2,990,000 (Direct Cost: ¥2,300,000、Indirect Cost: ¥690,000)
Fiscal Year 2018: ¥910,000 (Direct Cost: ¥700,000、Indirect Cost: ¥210,000)
Fiscal Year 2017: ¥910,000 (Direct Cost: ¥700,000、Indirect Cost: ¥210,000)
Fiscal Year 2016: ¥1,170,000 (Direct Cost: ¥900,000、Indirect Cost: ¥270,000)
Keywords多層パーセプトロン / 特異領域 / 特異階段追跡法 / 複素ニューラルネット / 深層学習 / モデル選択 / ニューラルネットワーク / 機械学習 / 特異モデル
Outline of Final Research Achievements

For neural networks learning, we have established SSF (singularity stairs following) principle, which enables us to find excellent solutions by utilizing singular regions. By applying SSF principle, we developed three very powerful learning methods: SSF1.4 for real-valued multilayer perceptron (MLP), RBF-SSF-pH for real-valued RBF network, and C-SSF1.3 for complex-valued MLP. In our experiments these methods showed quite excellent performances. Moreover, they were successfully employed for singular model selection using WAIC and WBIC, and for prediction of deterministic chaos behavior. Finally, we have also investigated and implemented mixture of non-linear regressions to find underlying functions for heterogeneous data.

Academic Significance and Societal Importance of the Research Achievements

学術的意義:深層学習の成功により、ニューラルネットの有用性が再認識されているが、本研究で技術確立した特異階段追跡(SSF)原理は、ニューラルネットの学習性能、特に解品質を著しく向上させるものである。深層学習と併用することにより、全体性能の一層の向上が期待される。また今後、SSF原理を他のニューラルネットモデルに適用する研究の展開も期待される。
社会的意義:近年の人工知能は多分野に導入されて大ブレークした。その中核には機械学習があり、特に、深層学習の成功が大きい。本研究は深層学習の性能に磨きをかけて有用性拡大に貢献すると期待される。

Report

(4 results)
  • 2018 Annual Research Report   Final Research Report ( PDF )
  • 2017 Research-status Report
  • 2016 Research-status Report
  • Research Products

    (16 results)

All 2019 2018 2017 2016 Other

All Journal Article (1 results) (of which Peer Reviewed: 1 results,  Open Access: 1 results) Presentation (13 results) (of which Int'l Joint Research: 7 results) Book (1 results) Remarks (1 results)

  • [Journal Article] How Learning Methods Influence the Performance of Complex-Valued Multilayer Perceptrons2017

    • Author(s)
      佐藤聖也, 中野良平
    • Journal Title

      電子情報通信学会論文誌D 情報・システム

      Volume: J100-D Issue: 6 Pages: 649-660

    • DOI

      10.14923/transinfj.2016JDP7109

    • ISSN
      1880-4535, 1881-0225
    • Year and Date
      2017-06-01
    • Related Report
      2017 Research-status Report 2016 Research-status Report
    • Peer Reviewed / Open Access
  • [Presentation] Faster RBF network learning utilizing singular regions2019

    • Author(s)
      Seiya Satoh, Ryohei Nakano
    • Organizer
      ICPRAM 2019 (8th International Conference on Pattern Recognition Applications and Methods)
    • Related Report
      2018 Annual Research Report
    • Int'l Joint Research
  • [Presentation] Mixture of multilayer perceptron regressions2019

    • Author(s)
      Ryohei Nakano
    • Organizer
      ICPRAM 2019 (8th International Conference on Pattern Recognition Applications and Methods)
    • Related Report
      2018 Annual Research Report
    • Int'l Joint Research
  • [Presentation] A new method for learning RBF networks by utilizing singular regions2018

    • Author(s)
      Seiya Satoh, Ryohei Nakano
    • Organizer
      ICAISC 2018 (17th International Conference on Artificial Intelligence and Soft Computing)
    • Related Report
      2018 Annual Research Report
    • Int'l Joint Research
  • [Presentation] Weak dependence on initialization in mixture of linear regressions2018

    • Author(s)
      Ryohei Nakano, Seiya Satoh
    • Organizer
      ICAIA 2018 (International Conference on Artificial Intelligence and Applications)
    • Related Report
      2017 Research-status Report
    • Int'l Joint Research
  • [Presentation] How new information criteria WAIC and WBIC worked for MLP model selection2017

    • Author(s)
      Seiya Satoh, Ryohei Nakano
    • Organizer
      ICPRAM 2017 (6th International Conference on Pattern Recognition Applications and Methods)
    • Place of Presentation
      Porto, Portugal
    • Year and Date
      2017-02-24
    • Related Report
      2016 Research-status Report
    • Int'l Joint Research
  • [Presentation] Performance of complex-valued multilayer perceptrons largely depends on learning methods2017

    • Author(s)
      Seiya Satoh, Ryohei Nakano
    • Organizer
      IJCCI 2017 (9th International Joint Conference on Computational Intelligence)
    • Related Report
      2017 Research-status Report
    • Int'l Joint Research
  • [Presentation] 特異領域を利用したRBFネットワーク探索法2017

    • Author(s)
      佐藤聖也, 中野良平
    • Organizer
      第15回情報学ワークショップ(WiNF 2017)
    • Related Report
      2017 Research-status Report
  • [Presentation] 特異階段追跡法を用いたカオス二重振り子の軌道予測2017

    • Author(s)
      小島久幸, 佐藤聖也, 中野良平
    • Organizer
      第15回情報学ワークショップ(WiNF 2017)
    • Related Report
      2017 Research-status Report
  • [Presentation] 特異領域を利用したRBFネットワーク新学習法2017

    • Author(s)
      佐藤聖也, 中野良平
    • Organizer
      電子情報通信学会技術研究報告 ニューロコンピューティング(NC)研究会
    • Related Report
      2017 Research-status Report
  • [Presentation] How complex-valued multilayer perceptron can predict the behavior of deterministic chaos2016

    • Author(s)
      Seiya Satoh, Ryohei Nakano
    • Organizer
      IJCNN 2016 (the International Joint Conference on Neural Networks)
    • Place of Presentation
      Vancouver, Canada
    • Year and Date
      2016-07-24
    • Related Report
      2016 Research-status Report
    • Int'l Joint Research
  • [Presentation] 複素特異階段追跡法の解品質と計算時間の実験評価2016

    • Author(s)
      佐藤聖也, 中野良平
    • Organizer
      計測自動制御学会(システム・情報部門)第9回コンピューテーショナル・インテリジェンス(CI)研究会
    • Place of Presentation
      千葉大学
    • Related Report
      2016 Research-status Report
  • [Presentation] 複素多層パーセプトロンの学習法と非線形性の関係に関する実験評価2016

    • Author(s)
      佐藤聖也, 中野良平
    • Organizer
      第14回情報学ワークショップ(WiNF 2016)
    • Place of Presentation
      愛知県立大学
    • Related Report
      2016 Research-status Report
  • [Presentation] 多層パーセプトロンを用いた二重振り子カオス軌道予測2016

    • Author(s)
      小島久幸, 佐藤聖也, 中野良平
    • Organizer
      第14回情報学ワークショップ(WiNF 2016)
    • Place of Presentation
      愛知県立大学
    • Related Report
      2016 Research-status Report
  • [Book] 人工知能学大事典2017

    • Author(s)
      人工知能学会
    • Total Pages
      1600
    • Publisher
      共立出版
    • ISBN
      9784320124202
    • Related Report
      2017 Research-status Report
  • [Remarks] ニューラル情報処理の研究

    • URL

      http://www.nipl.cs.chubu.ac.jp/~nakano/learning-j.html

    • Related Report
      2016 Research-status Report

URL: 

Published: 2016-04-21   Modified: 2020-03-30  

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi