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

Study on fast and accurate classifier learning method from unlabeled big data

Research Project

Project/Area Number 20K21815
Research Category

Grant-in-Aid for Challenging Research (Exploratory)

Allocation TypeMulti-year Fund
Review Section Medium-sized Section 61:Human informatics and related fields
Research InstitutionOsaka University

Principal Investigator

Washio Takashi  大阪大学, 産業科学研究所, 教授 (00192815)

Project Period (FY) 2020-07-30 – 2024-03-31
Project Status Completed (Fiscal Year 2023)
Budget Amount *help
¥6,370,000 (Direct Cost: ¥4,900,000、Indirect Cost: ¥1,470,000)
Fiscal Year 2021: ¥2,990,000 (Direct Cost: ¥2,300,000、Indirect Cost: ¥690,000)
Fiscal Year 2020: ¥3,380,000 (Direct Cost: ¥2,600,000、Indirect Cost: ¥780,000)
Keywords弱教師有り学習 / 分類器学習 / 機械学習 / UUC / 教師ラベル無しデータ / 分類器 / 回帰式 / クラス事前確率 / 非結合回帰 / 教師無し学習 / UUC学習 / クラス事前分布推定 / ガウス過程回帰 / 弱学習 / 教師無し分類器学習 / ラベル無しデータ / ビッグデータ
Outline of Research at the Start

研究は、(1)事例データの分布密度差推定の原理の構築、(2)理論的な性質や性能保証に関する解析、(3)理論的性質の人工検証データを用いた確認、(4)実用性に関するフィジビリティスタディとして心機能健診データから個人の心不全発症リスク分類、(5)同じく微小生体の形状観測情報から種類識別する分類器構築、の5つの項目からなる。(1)(2)(3)は鷲尾と国際共同研究者のK.M.Ting教授とで取り組み、(4)は国立循環器病研究センターの医療チーム、(5)は大阪大学産業科学研究所の谷口教授の研究室と共同で取り組む。これによって、目指すUUC手法の原理的基礎の確立と、その実用性に関する見通しを得る。

Outline of Final Research Achievements

With the widespread adoption of AI technology, there is an increasing demand for classifier learning from unlabeled big data due to constraints and costs associated with data collection. In response to this issue, the UUC method, which learns classifiers from two unlabeled datasets with different proportions of positive and negative examples, has been proposed. However, existing methods require vast computational resources for large-scale data and suffer from bias error in classification.
In this study, we propose a versatile UUC method which requires low computational cost only, and is free from bias error. We applied this method to the classification of various datasets, including real data, and verified that unsupervised learning without teaching labels is possible with almost the same accuracy as supervised learning. This establishes a UUC method that far exceeds the application range limitations of the existing UUC methods.

Academic Significance and Societal Importance of the Research Achievements

IoT社会の深化とAI技術の普及に伴い、ビッグデータからの分類器学習ニーズが増しているが、多くの場合にデータ収集の制約やコストから教師ラベルが得られないことが問題となっている。これに対し近年、正負例割合の異なる2つのラベル無し事例集合から分類器を学習するUUC手法が提案されている。しかし、これらはカーネル法を用いており、訓練データ数NについてO(N3)の学習計算量を要し、またN→∞でも分類に偏り誤差を生じる場合がある。従って、複雑な事例分布を持つビッグデータに適用可能な高速高精度なUUC手法の開発が強く待たれていた。本研究成果は、この社会的要請に応えるものである。

Report

(5 results)
  • 2023 Annual Research Report   Final Research Report ( PDF )
  • 2022 Research-status Report
  • 2021 Research-status Report
  • 2020 Research-status Report
  • Research Products

    (26 results)

All 2023 2022 2021 2020 Other

All Int'l Joint Research (6 results) Journal Article (7 results) (of which Int'l Joint Research: 5 results,  Peer Reviewed: 7 results,  Open Access: 7 results) Presentation (13 results) (of which Int'l Joint Research: 3 results,  Invited: 2 results)

  • [Int'l Joint Research] Nanjing University(中国)

    • Related Report
      2022 Research-status Report
  • [Int'l Joint Research] Federation University in Australia(オーストラリア)

    • Related Report
      2022 Research-status Report
  • [Int'l Joint Research] Nanjing University(中国)

    • Related Report
      2021 Research-status Report
  • [Int'l Joint Research] Federation University in Australia(オーストラリア)

    • Related Report
      2021 Research-status Report
  • [Int'l Joint Research] Federation University in Australia(オーストラリア)

    • Related Report
      2020 Research-status Report
  • [Int'l Joint Research] Nanjing University(中国)

    • Related Report
      2020 Research-status Report
  • [Journal Article] Isolation Kernel Estimators2022

    • Author(s)
      Kai Ming Ting, Takashi Washio, Jonathan Wells, Hang Zhang, Ye Zhu
    • Journal Title

      Knowledge and Information Systems (KAIS Journal)

      Volume: 65 Issue: 2 Pages: 759-787

    • DOI

      10.1007/s10115-022-01765-7

    • Related Report
      2022 Research-status Report
    • Peer Reviewed / Open Access / Int'l Joint Research
  • [Journal Article] Bayesian optimization-driven parallel-screening of multiple parameters for the flow synthesis of biaryl compounds2022

    • Author(s)
      Kondo Masaru、Wathsala H. D. P.、Salem Mohamed S. H.、Ishikawa Kazunori、Hara Satoshi、Takaai Takayuki、Washio Takashi、Sasai Hiroaki、Takizawa Shinobu
    • Journal Title

      Communications Chemistry

      Volume: 5 Issue: 1 Pages: 148-148

    • DOI

      10.1038/s42004-022-00764-7

    • Related Report
      2022 Research-status Report
    • Peer Reviewed / Open Access
  • [Journal Article] Isolation Kernel: The X Factor in Efficient and Effective Large Scale Online Kernel Learning2021

    • Author(s)
      Kai Ming Ting, Jonathan R. Wells, and Takashi Washio
    • Journal Title

      Data Mining and Knowledge Discovery

      Volume: 35 Issue: 6 Pages: 2282-2312

    • DOI

      10.1007/s10618-021-00785-1

    • Related Report
      2021 Research-status Report
    • Peer Reviewed / Open Access / Int'l Joint Research
  • [Journal Article] Isolation Kernel Density Estimation2021

    • Author(s)
      Kai Ming Ting, Takashi Washio, Jonathan Wells, and Hang Zhang
    • Journal Title

      IEEE ICDM 2021: IEEE ICDM 2021 21st IEEE International Conference on Data Mining

      Volume: 1 Pages: 619-628

    • DOI

      10.1109/icdm51629.2021.00073

    • Related Report
      2021 Research-status Report
    • Peer Reviewed / Open Access / Int'l Joint Research
  • [Journal Article] Isolation Distributional Kernel: A New Tool for Point & Group Anomaly Detection2021

    • Author(s)
      Kai Ming Ting, Takashi Washio, Bi-Cun Xu, Zhi-Hua Zhou
    • Journal Title

      IEEE Transactions on Knowledge and Data Engineering

      Volume: 1 Pages: 1-1

    • DOI

      10.1109/tkde.2021.3120277

    • Related Report
      2021 Research-status Report
    • Peer Reviewed / Open Access / Int'l Joint Research
  • [Journal Article] Classification from Positive and Unlabeled Data Based on Likelihood Invariance for Measurement2021

    • Author(s)
      Takeshi Yoshida, Takashi Washio, Takahito Ohshiro, Masateru Taniguchi
    • Journal Title

      Intelligent Data Analysis

      Volume: 25 Issue: 1 Pages: 57-79

    • DOI

      10.3233/ida-194980

    • Related Report
      2020 Research-status Report
    • Peer Reviewed / Open Access
  • [Journal Article] Isolation Distributional Kernel: A new tool for kernel based anomaly detection2020

    • Author(s)
      Kai Ming Ting, Takashi Washio, Bi-Cun Xu, Zhi-Hua Zhou
    • Journal Title

      KDD2020: Knowledge Discovery and Data Mining, 2020

      Volume: 1 Pages: 233-233

    • DOI

      10.1145/3394486.3403062

    • Related Report
      2020 Research-status Report
    • Peer Reviewed / Open Access / Int'l Joint Research
  • [Presentation] 対象事前知識に基づく回帰モデリングと不確定性推定法2023

    • Author(s)
      東 大介, 鷲尾 隆
    • Organizer
      2023年度人工知能学会全国大会(第37回)
    • Related Report
      2023 Annual Research Report
  • [Presentation] 機械学習ベイズ最適化を活用するケチミンの電解合成反応条件最適化2022

    • Author(s)
      Khalid Md Imrul,近藤健,杉嵜晃将,H.D.P. Wathsala,石川一宣,原聡,鷹合孝之,鷲尾隆,笹井宏明,滝澤忍
    • Organizer
      日本プロセス化学会2018サマーシンポジウム
    • Related Report
      2022 Research-status Report
  • [Presentation] ベイズ最適化を用いたイオン源制御手法の開発2022

    • Author(s)
      森田 泰之,福田 光宏,依田 哲彦,神田 浩樹,畑中 吉治,斎藤 高嶺,田村 仁志,安田 祐介,鷲尾 隆,中島 悠太,岩崎 昌子
    • Organizer
      第19回日本加速器学会年会
    • Related Report
      2022 Research-status Report
  • [Presentation] Bayesian optimization-assisted multi-parameter screening for laboratory- and industrial-scale syntheses2022

    • Author(s)
      H.D.P. Wathsala, M. Kondo, M.S. H. Salem, K. Ishikawa, S. Hara, T. Takaai, T. Miyazaki, D. Yamashita, T. Washio, H. Sasai and S. Takizawa
    • Organizer
      2022年度有機合成化学北陸セミナー
    • Related Report
      2022 Research-status Report
  • [Presentation] Measurement Informatics and Its Application in Science2022

    • Author(s)
      Takashi Washio
    • Organizer
      SciX2022: SciX (The Great SCIentific eXchange) Conference 2022 (The Federation of Analytical Chemistry and Spectroscopy Societies (FACSS))
    • Related Report
      2022 Research-status Report
    • Int'l Joint Research / Invited
  • [Presentation] 革新的先端計測の方程式:計測 + AI = 計測インフォマティクス2022

    • Author(s)
      鷲尾 隆
    • Organizer
      人工知能学会「シンポジウム BigDataDX 2022」
    • Related Report
      2022 Research-status Report
    • Invited
  • [Presentation] Unsupervised Noise Reduction for Nanochannel Measurement Using Noise2Noise Deep Learning2021

    • Author(s)
      Takayuki Takaai and Makusu Tsutsui
    • Organizer
      PAKDD 2021 Workshops MLMEIN
    • Related Report
      2021 Research-status Report
    • Int'l Joint Research
  • [Presentation] Class Prior Probability Estimation Using Density Ratio from Unlabeled and Contaminated Positive Datasets2021

    • Author(s)
      Takeshi Yoshida and Eitaro Shinya
    • Organizer
      PAKDD 2021 Workshops MLMEIN
    • Related Report
      2021 Research-status Report
    • Int'l Joint Research
  • [Presentation] 一対比較データによる非結合ガウス過程回帰手法の提案2021

    • Author(s)
      山川 将輝,鷲尾 隆
    • Organizer
      2021年人工知能学会全国大会
    • Related Report
      2021 Research-status Report
  • [Presentation] クラス事前確率を用いたラベル無しデータからの分類器学習の性能解析2021

    • Author(s)
      松本 瑞季,鷲尾 隆
    • Organizer
      2021年人工知能学会全国大会
    • Related Report
      2021 Research-status Report
  • [Presentation] Noise2Noise 深層学習を用いた教師無しのナノチャンネル計測ノイズ低減2020

    • Author(s)
      鷹合孝之, 筒井真楠, 鷲尾隆
    • Organizer
      人工知能学会第4回計測インフォマティクス研究会
    • Related Report
      2020 Research-status Report
  • [Presentation] ラベルなし事例集合と負事例混入正事例集合からの密度比を用いたクラス事前確率推定2020

    • Author(s)
      吉田剛, 新家英太郎, 鷲尾隆
    • Organizer
      人工知能学会第4回計測インフォマティクス研究会
    • Related Report
      2020 Research-status Report
  • [Presentation] アンサンブル最近傍距離を用いたラベル無しデータからの分類器学習2020

    • Author(s)
      松本 瑞季, 鷲尾 隆
    • Organizer
      第34回人工知能学会全国大会(2020)
    • Related Report
      2020 Research-status Report

URL: 

Published: 2020-08-03   Modified: 2025-01-30  

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi