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

Statistical mechanics and sparse modeling approach to large-scale inverse problems

Research Project

Project/Area Number 18K11463
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 InstitutionKyoto University (2019-2021)
Tokyo Institute of Technology (2018)

Principal Investigator

Obuchi Tomoyuki  京都大学, 情報学研究科, 准教授 (40588448)

Project Period (FY) 2018-04-01 – 2022-03-31
Project Status Completed (Fiscal Year 2021)
Budget Amount *help
¥4,290,000 (Direct Cost: ¥3,300,000、Indirect Cost: ¥990,000)
Fiscal Year 2021: ¥1,170,000 (Direct Cost: ¥900,000、Indirect Cost: ¥270,000)
Fiscal Year 2020: ¥1,040,000 (Direct Cost: ¥800,000、Indirect Cost: ¥240,000)
Fiscal Year 2019: ¥1,040,000 (Direct Cost: ¥800,000、Indirect Cost: ¥240,000)
Fiscal Year 2018: ¥1,040,000 (Direct Cost: ¥800,000、Indirect Cost: ¥240,000)
Keywords情報統計力学 / スパースモデリング / 統計的逆問題 / 機械学習 / 近似推論
Outline of Final Research Achievements

This project has been aimed at obtaining a deeper mathematical understanding and practical numerical solutions to variable selection and parameter estimation problems by extending the techniques of sparse modelling (SpM) and statistical mechanical informatics. As a result, a number of results (16 peer-reviewed papers, including several high impact journals, and three approximation algorithm packages published on Github) have been obtained. In particular, ingenious results include the development and related theoretical analysis of approximation algorithms that perform cross-validation methods with low computational complexity, theoretical analysis of inverse Ising problems related to structural learning, and an approximation algorithm for variable selection using the bootstrap method and its theoretical analysis.

Academic Significance and Societal Importance of the Research Achievements

近年、機械学習や人工知能といったキーワードで、ある種の数学的モデリング方法が注目を集めている。本研究では、そのモデリング方法の中でも、スパースモデリングという、様々な事物(変数)の中から重要なものを半自動的に取りだす技術に着目し、その技術を高めるべく研究を行った。多くの場合、計算量的問題から適切な変数の取り出しが難しくなるのだが、それを情報統計力学という物理分野にある手法で問題を近似することにより、計算量を削減するという方針で研究を行った。その結果、正確性を保ったまま以前よりも速やかに変数選択をすることが可能になった。機械学習を大規模に運用する上で重要な成果と成る。

Report

(5 results)
  • 2021 Annual Research Report   Final Research Report ( PDF )
  • 2020 Research-status Report
  • 2019 Research-status Report
  • 2018 Research-status Report
  • Research Products

    (39 results)

All 2022 2021 2020 2019 2018

All Journal Article (14 results) (of which Int'l Joint Research: 1 results,  Peer Reviewed: 14 results,  Open Access: 12 results) Presentation (25 results) (of which Int'l Joint Research: 7 results,  Invited: 4 results)

  • [Journal Article] Assessing transfer entropy from biochemical data2022

    • Author(s)
      Imaizumi Takuya、Umeki Nobuhisa、Yoshizawa Ryo、Obuchi Tomoyuki、Sako Yasushi、Kabashima Yoshiyuki
    • Journal Title

      Physical Review E

      Volume: 105 Issue: 3 Pages: 1-13

    • DOI

      10.1103/physreve.105.034403

    • Related Report
      2021 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] Perfect reconstruction of sparse signals with piecewise continuous nonconvex penalties and nonconvexity control2021

    • Author(s)
      Sakata Ayaka、Obuchi Tomoyuki
    • Journal Title

      Journal of Statistical Mechanics: Theory and Experiment

      Volume: 2021 Issue: 9 Pages: 093401-093401

    • DOI

      10.1088/1742-5468/ac1403

    • Related Report
      2021 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] Sharp Asymptotics of Matrix Sketching for a Rank-One Spiked Model2021

    • Author(s)
      Tagashira Fumito、Obuchi Tomoyuki、Tanaka Toshiyuki
    • Journal Title

      2021 IEEE International Symposium on Information Theory (ISIT)

      Volume: none Pages: 250-255

    • DOI

      10.1109/isit45174.2021.9517724

    • Related Report
      2021 Annual Research Report
    • Peer Reviewed
  • [Journal Article] Ising Model Selection Using $\ell_1$-Regularized Linear Regression: A Statistical Mechanics Analysis2021

    • Author(s)
      Xiangming Meng、Tomoyuki Obuchi、Yoshiyuki Kabashima
    • Journal Title

      Advances in Neural Information Processing Systems

      Volume: 34 Pages: 6290-6303

    • Related Report
      2021 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] Structure learning in inverse Ising problems using ? <sub>2</sub>-regularized linear estimator2021

    • Author(s)
      Meng Xiangming、Obuchi Tomoyuki、Kabashima Yoshiyuki
    • Journal Title

      Journal of Statistical Mechanics: Theory and Experiment

      Volume: 2021 Issue: 5 Pages: 053403-053403

    • DOI

      10.1088/1742-5468/abfa10

    • Related Report
      2021 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] Reconstructing Sparse Signals via Greedy Monte-Carlo Search2020

    • Author(s)
      Hayashi Kao、Obuchi Tomoyuki、Kabashima Yoshiyuki
    • Journal Title

      Journal of the Physical Society of Japan

      Volume: 89 Issue: 12 Pages: 124802-124802

    • DOI

      10.7566/jpsj.89.124802

    • NAID

      40022438412

    • Related Report
      2020 Research-status Report
    • Peer Reviewed
  • [Journal Article] Inferring Neuronal Couplings From Spiking Data Using a Systematic Procedure With a Statistical Criterion2020

    • Author(s)
      Terada Yu、Obuchi Tomoyuki、Isomura Takuya、Kabashima Yoshiyuki
    • Journal Title

      Neural Computation

      Volume: 32 Issue: 11 Pages: 2187-2211

    • DOI

      10.1162/neco_a_01324

    • Related Report
      2020 Research-status Report
    • Peer Reviewed / Open Access
  • [Journal Article] Learning performance in inverse Ising problems with sparse teacher couplings2020

    • Author(s)
      Abbara Alia、Kabashima Yoshiyuki、Obuchi Tomoyuki、Xu Yingying
    • Journal Title

      Journal of Statistical Mechanics: Theory and Experiment

      Volume: 2020 Issue: 7 Pages: 073402-073402

    • DOI

      10.1088/1742-5468/ab8c3a

    • Related Report
      2020 Research-status Report
    • Peer Reviewed / Open Access / Int'l Joint Research
  • [Journal Article] Empirical Bayes method for Boltzmann machines2019

    • Author(s)
      Muneki Yasuda and Tomoyuki Obuchi
    • Journal Title

      Journal of Physics A: Mathematical and Theoretical

      Volume: 53 Issue: 1 Pages: 014004-014004

    • DOI

      10.1088/1751-8121/ab57a7

    • Related Report
      2019 Research-status Report
    • Peer Reviewed / Open Access
  • [Journal Article] Cross validation in sparse linear regression with piecewise continuous nonconvex penalties and its acceleration2019

    • Author(s)
      Obuchi Tomoyuki、Sakata Ayaka
    • Journal Title

      Journal of Physics A: Mathematical and Theoretical

      Volume: 52 Issue: 41 Pages: 414003-414003

    • DOI

      10.1088/1751-8121/ab3e89

    • Related Report
      2019 Research-status Report
    • Peer Reviewed / Open Access
  • [Journal Article] Objective and efficient inference for couplings in neuronal networks2018

    • Author(s)
      Yu Terada, Tomoyuki Obuchi, Takuya Isomura, Yoshiyuki Kabashima
    • Journal Title

      Advances in Neural Information Processing Systems

      Volume: 31

    • Related Report
      2018 Research-status Report
    • Peer Reviewed / Open Access
  • [Journal Article] Mean-field theory of graph neural networks in graph partitioning2018

    • Author(s)
      Tatsuro Kawamoto, Masashi Tsubaki, Tomoyuki Obuchi
    • Journal Title

      Advances in Neural Information Processing Systems

      Volume: 31

    • Related Report
      2018 Research-status Report
    • Peer Reviewed / Open Access
  • [Journal Article] Statistical mechanical analysis of sparse linear regression as a variable selection problem2018

    • Author(s)
      Tomoyuki Obuchi, Yoshinori Nakanishi-Ohno, Masato Okada, Yoshiyuki Kabashima
    • Journal Title

      J. Stat. Mech.

      Volume: 103401 Pages: 103401-103401

    • Related Report
      2018 Research-status Report
    • Peer Reviewed / Open Access
  • [Journal Article] Accelerating Cross-Validation in Multinomial Logistic Regression with L1-Regularization2018

    • Author(s)
      Tomoyuki Obuchi, Yoshiyuki Kabashima
    • Journal Title

      Journal of Machine Learning Research

      Volume: 19

    • NAID

      120006629921

    • Related Report
      2018 Research-status Report
    • Peer Reviewed / Open Access
  • [Presentation] ニューラルネットにおける短期記憶の1ビット圧縮センシングによる想起2021

    • Author(s)
      杉崎嵐, 寺田裕, 小渕智之, 樺島祥介
    • Organizer
      日本物理学会 第76回年次大会
    • Related Report
      2020 Research-status Report
  • [Presentation] 高ランクテンソル分解のBPに基づく推定2021

    • Author(s)
      長澤莉希, 小渕智之, 吉野元
    • Organizer
      日本物理学会 第76回年次大会
    • Related Report
      2020 Research-status Report
  • [Presentation] SCAD正則化を用いた圧縮センシングに関する信号復元性能の分析2021

    • Author(s)
      石井智, 小渕智之, 樺島祥介
    • Organizer
      日本物理学会 第76回年次大会
    • Related Report
      2020 Research-status Report
  • [Presentation] リサンプリングの近似計算アルゴリズムの理論と実践2020

    • Author(s)
      小渕智之
    • Organizer
      スマートサンプリング講演会
    • Related Report
      2020 Research-status Report
  • [Presentation] 教師の結合が疎な場合における逆イジング問題のレプリカ解析と仮説検定2020

    • Author(s)
      Alia Abbara, 樺島祥介, 小渕智之, 許インイン
    • Organizer
      日本物理学会2020年秋季大会
    • Related Report
      2020 Research-status Report
  • [Presentation] 細胞内生化学反応に関する伝達情報量の解析2020

    • Author(s)
      今泉拓也, 小渕智之, 吉澤亮, 佐甲靖志, 樺島祥介
    • Organizer
      日本物理学会2020年秋季大会
    • Related Report
      2020 Research-status Report
  • [Presentation] 1+p体観測による多成分ベクトルの統計的推定2020

    • Author(s)
      長澤莉希, 小渕智之, 吉野元
    • Organizer
      日本物理学会2020年秋季大会
    • Related Report
      2020 Research-status Report
  • [Presentation] 逆イジング問題のレプリカ法による性能解析:教師の結合が疎な場合2020

    • Author(s)
      Alia Abbara, 樺島祥介, 小渕智之, 許インイン
    • Organizer
      第23回情報論的学習理論ワークショップ(IBIS2020)
    • Related Report
      2020 Research-status Report
  • [Presentation] 高次元確率モデルの扱いについて:統計物理の視点から2020

    • Author(s)
      小渕 智之
    • Organizer
      電子情報通信学会総合大会
    • Related Report
      2019 Research-status Report
    • Invited
  • [Presentation] Cross validation in sparse linear regression with piecewise continuous nonconvex penalties and its acceleration2019

    • Author(s)
      Obuchi Tomoyuki, Sakata Ayaka
    • Organizer
      2019 Workshop on Statistical Physics of Disordered Systems and Its Applications (SPDSA2019)
    • Related Report
      2019 Research-status Report
    • Int'l Joint Research / Invited
  • [Presentation] Cross validation in sparse linear regression with piecewise continuous nonconvex penalties and its acceleration2019

    • Author(s)
      Obuchi Tomoyuki, Sakata Ayaka
    • Organizer
      Statistical Physics and Neural Computation (SPNC-2019)
    • Related Report
      2019 Research-status Report
    • Int'l Joint Research
  • [Presentation] Statistical mechanical analysis of sparse linear regression as a variable selection problem2019

    • Author(s)
      Obuchi Tomoyuki
    • Organizer
      2019 International Workshop on Glass Physics in Beijing
    • Related Report
      2019 Research-status Report
    • Int'l Joint Research / Invited
  • [Presentation] ボルツマンマシンによる神経細胞集団の有効な非対称結合推定2019

    • Author(s)
      小渕 智之
    • Organizer
      情報数物研究会
    • Related Report
      2019 Research-status Report
    • Invited
  • [Presentation] レプリカ法を用いたボルツマンマシンに対する経験ベイズ推定2019

    • Author(s)
      安田宗樹, 小渕智之
    • Organizer
      日本物理学会2019年年次大会
    • Related Report
      2018 Research-status Report
  • [Presentation] Elastic-net正則化付き多項ロジスティック回帰における交差検証法の加速2019

    • Author(s)
      小渕智之, 樺島祥介
    • Organizer
      日本物理学会2019年年次大会
    • Related Report
      2018 Research-status Report
  • [Presentation] スパース線形回帰モデルにおけるレプリカ交換モンテカルロ法を用いた変数選択2019

    • Author(s)
      石井奨, 小渕智之, 樺島祥介
    • Organizer
      日本物理学会2019年年次大会
    • Related Report
      2018 Research-status Report
  • [Presentation] Mean-field theory of graph neural networks in graph partitioning2018

    • Author(s)
      Tatsuro Kawamoto, Masashi Tsubaki, Tomoyuki Obuchi
    • Organizer
      NeurIPS2018
    • Related Report
      2018 Research-status Report
    • Int'l Joint Research
  • [Presentation] Objective and efficient inference dor couplings in neuronal networks2018

    • Author(s)
      Yu Terada, Tomoyuki Obuchi, Takuya Isomura, Yoshiyuki Kabashima
    • Organizer
      NeurIPS2018
    • Related Report
      2018 Research-status Report
    • Int'l Joint Research
  • [Presentation] Statistical mechanical analysis of sparse linear regression as a variable selection problem2018

    • Author(s)
      Tomoyuki Obuchi, Y. Nakanishi-Ohno, M. Okada and Y. Kabashima
    • Organizer
      Disordered serendipity: a glassy path to discovery
    • Related Report
      2018 Research-status Report
    • Int'l Joint Research
  • [Presentation] Accelerating Cross-Validation in Multinomial Logistic Regression with L1-Regularization2018

    • Author(s)
      Tomoyuki Obuchi and Yoshiyuki Kabashima
    • Organizer
      Statistical physics and machine learning back together
    • Related Report
      2018 Research-status Report
    • Int'l Joint Research
  • [Presentation] 半解析的ブートストラップ法とその応用2018

    • Author(s)
      小渕智之
    • Organizer
      第21回情報論的学習理論ワークショップ(IBIS2018)
    • Related Report
      2018 Research-status Report
  • [Presentation] 統計力学的アプローチによるリサンプリング手法の軽量化2018

    • Author(s)
      小渕智之
    • Organizer
      第30回RAMPシンポジウム(RAMP2018)
    • Related Report
      2018 Research-status Report
  • [Presentation] 深層学習によるグラフのコミュニティ構造推定II2018

    • Author(s)
      川本達郎, 椿真史, 小渕智之
    • Organizer
      日本物理学会2018年秋季大会
    • Related Report
      2018 Research-status Report
  • [Presentation] 非凸正則化付き線形回帰における復号性能2018

    • Author(s)
      坂田綾香, 小渕智之
    • Organizer
      日本物理学会2018年秋季大会
    • Related Report
      2018 Research-status Report
  • [Presentation] LASSO+半解析的ブートストラップ法を用いた現実的な変数選択法2018

    • Author(s)
      小渕智之, 樺島祥介
    • Organizer
      日本物理学会2018年秋季大会
    • Related Report
      2018 Research-status Report

URL: 

Published: 2018-04-23   Modified: 2023-01-30  

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi