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

Development of small data processing method combined with mathmatical model and machine learning approarch

Research Project

Project/Area Number 19K12139
Research Category

Grant-in-Aid for Scientific Research (C)

Allocation TypeMulti-year Fund
Section一般
Review Section Basic Section 61040:Soft computing-related
Research InstitutionUniversity of Miyazaki

Principal Investigator

Yamamori Kunihito  宮崎大学, 工学部, 教授 (50293395)

Project Period (FY) 2019-04-01 – 2023-03-31
Project Status Completed (Fiscal Year 2022)
Budget Amount *help
¥3,250,000 (Direct Cost: ¥2,500,000、Indirect Cost: ¥750,000)
Fiscal Year 2022: ¥650,000 (Direct Cost: ¥500,000、Indirect Cost: ¥150,000)
Fiscal Year 2021: ¥650,000 (Direct Cost: ¥500,000、Indirect Cost: ¥150,000)
Fiscal Year 2020: ¥1,170,000 (Direct Cost: ¥900,000、Indirect Cost: ¥270,000)
Fiscal Year 2019: ¥780,000 (Direct Cost: ¥600,000、Indirect Cost: ¥180,000)
Keywords機械学習 / スモールデータ / 数理モデル / 決定木 / ネットワークセキュリティ / 拡張重み更新型自己組織化マップ / 強化学習 / ヒューリスティック / 自己組織化マップ / 学習サンプル / 分散 / ソフトコンピューティング
Outline of Research at the Start

本研究では,多数の学習サンプルを得ることが困難な生命系現象を対象に,少数の未整理かつ分散の大きいデータから適切な学習サンプルを構築する手法,および少ない学習サンプルでも過学習を起こさず汎化能力の高い学習アルゴリズムを確立することを目的とする.具体的には,研究代表者が関与したたんぱく質発現量からの食品機能性(複数のがん抑制活性)推定を題材に,生命活動を反映した数理モデルにより生成した学習サンプルと,ベクトルの要素を波のサンプル点と見做した距離関数を導入した拡張重み更新型自己組織化マップを併用することで,高精度かつ汎化能力の高い学習モデルの確立を目指す.

Outline of Final Research Achievements

Recent machine learning algorithms require a large number of training samples. In this research, I try to combine the mathematical model that can approximate the phenomena and the machine learning approach to solve some problems that are hard to observe the phenomena or hard to reproduce the experimental results again. I picked up two problems; one was to estimate the physiological activities from the protein expression levels, and the other was to detect the intrusion into the computer systems. For the first one, I rewrote the Linux-based programs into an integrated program. For the second one, I showed Gradient Boosted Decision Tree algorithm was suitable and robust for the small number of training samples.

Academic Significance and Societal Importance of the Research Achievements

現代のAIでは,適切な結果を得るためには膨大な数の学習サンプルを必要とする.一方,観測が困難であったり再現が難しいなど,多数の学習サンプルを準備することが難しい課題も存在する.本研究では,学習データを補うため数理モデルを作成し,モデルに従って学習サンプルを生成することで精度よく学習が行うことが可能なアプローチを模索した.例題として,たんぱく質発現量から生理活性値を推定する問題,および学習データ数は豊富なものの信頼性に疑義があるコンピュータシステムへの侵入検知問題を取り上げた.前者については推定プログラムを作成し,後者についてはブースティングを併用した決定木アプローチが有効であることを示した.

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

    (12 results)

All 2023 2022 2021 2020 2019 Other

All Journal Article (1 results) (of which Int'l Joint Research: 1 results,  Peer Reviewed: 1 results) Presentation (10 results) (of which Int'l Joint Research: 7 results) Remarks (1 results)

  • [Journal Article] Affect of data unbalance in "Kyoto 2016 Dataset" for NIDS with machine learning2020

    • Author(s)
      Ryo SAITO, Masaru AIKAWA, Kentaro INOUE, Kunihito YAMAMORI
    • Journal Title

      Proc. International Symposium on Artificial Life and Robotics

      Volume: 25 Pages: 612-616

    • Related Report
      2019 Research-status Report
    • Peer Reviewed / Int'l Joint Research
  • [Presentation] Performance of Machine Learning base NIDS on Re-organized Kyoto 2016 Dataset2023

    • Author(s)
      Ryo Saito, Masaru Aikawa, Kunihito Yamamori
    • Organizer
      The 28th International Symposium on Artificial Life and Robotics
    • Related Report
      2022 Annual Research Report
    • Int'l Joint Research
  • [Presentation] Robustness comparison of machine learning algorithms for NIDS under the same environment2023

    • Author(s)
      Masaki Tagawa, Kunihito Yamamori, Masaru Aikawa
    • Organizer
      The 28th International Symposium on Artificial Life and Robotics
    • Related Report
      2022 Annual Research Report
    • Int'l Joint Research
  • [Presentation] 同一環境下における機械学習アルゴリズムの侵入検知精度比較評価2023

    • Author(s)
      田河雅輝,山森一人,齊藤燎
    • Organizer
      火の国情報シンポジウム2023
    • Related Report
      2022 Annual Research Report
  • [Presentation] Driving trajectory optimization by reinforcement learning for motorsports2022

    • Author(s)
      Akinobu Iwai, Masaru Aikawa, Kunihito Yamamori
    • Organizer
      The 27th International Symposium on Artificial Life and Robotics 2022
    • Related Report
      2021 Research-status Report
    • Int'l Joint Research
  • [Presentation] Heuristic base music arrangement suppressing on discord progression2022

    • Author(s)
      Kosuke Yoshida, Masaru Aikawa, Kunihito Yamamori
    • Organizer
      The 27th International Symposium on Artificial Life and Robotics 2022
    • Related Report
      2021 Research-status Report
    • Int'l Joint Research
  • [Presentation] Simultaneous estimation of multiple food functions of food components using a targeted proteomics approach2021

    • Author(s)
      Kiyoko Nagahama, Akira Ota, Katsuhisa Kurogi, Kunihito Yamamori, Yoichi Sakakibara
    • Organizer
      10th Asia-oceania Human Proteome Organization Congress
    • Related Report
      2021 Research-status Report
    • Int'l Joint Research
  • [Presentation] ターゲットプロテオミクスによる複数の食品機能性の同時推定2021

    • Author(s)
      永濱清子, 太田輝, 黒木勝久, 山森一人, 水光正仁, 榊原陽一
    • Organizer
      第27回 日本生物工学会九州支部 大分大会
    • Related Report
      2021 Research-status Report
  • [Presentation] Network Design for Session Type NIDS2021

    • Author(s)
      Ryo Saito, Kunihito Yamamori, Masaru Aikawa, Kentaro Inoue,
    • Organizer
      The 26th International Symposium on Artificial Life and Robotics 2021
    • Related Report
      2020 Research-status Report
    • Int'l Joint Research
  • [Presentation] Tuning Support Tool for WAF Mod Security by Log Analysis with Machine Learning2021

    • Author(s)
      Chihiro Kudo, Kunihito Yamamori, Masaru Aikawa, Kentaro Inoue, Ryo Saito
    • Organizer
      The 26th International Symposium on Artificial Life and Robotics 2021
    • Related Report
      2020 Research-status Report
    • Int'l Joint Research
  • [Presentation] 「Kyoto 2016 Dataset」における冗長性と同一特徴量異ラベルデータに関する報告2019

    • Author(s)
      齋藤燎,相川勝,井上健太郎,山森一人
    • Organizer
      2019年度(第72回)電気・情報関係学会九州支部連合大会
    • Related Report
      2019 Research-status Report
  • [Remarks] 宮崎大学研究者データベース

    • URL

      https://srhumdb.miyazaki-u.ac.jp/html/247_ja.html?k=%E3%83%A4%E3%83%9E%E3%83%A2%E3%83%AA

    • Related Report
      2022 Annual Research Report

URL: 

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

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi