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

Developing a non-asymptotic inference theory for scientific hypothesis testing with statistical deep modelling

Research Project

Project/Area Number 21K11780
Research Category

Grant-in-Aid for Scientific Research (C)

Allocation TypeMulti-year Fund
Section一般
Review Section Basic Section 60030:Statistical science-related
Research InstitutionThe University of Tokyo

Principal Investigator

Imaizumi Masaaki  東京大学, 大学院総合文化研究科, 准教授 (90814088)

Project Period (FY) 2021-04-01 – 2024-03-31
Project Status Completed (Fiscal Year 2023)
Budget Amount *help
¥3,900,000 (Direct Cost: ¥3,000,000、Indirect Cost: ¥900,000)
Fiscal Year 2023: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000)
Fiscal Year 2022: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000)
Fiscal Year 2021: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000)
Keywords高次元統計 / 深層学習 / ニューラルネットワーク / 比例的高次元 / 統計的推論 / 統計学 / 深層モデル / 過剰パラメータ / 汎化誤差 / ガウス近似
Outline of Research at the Start

近年強い注目を浴びるデータ解析法である深層学習は、物理法則や分子構造の探索や再現といったサイエンスの領域での使用も盛んである。しかし、仮説検定や信頼解析などの統計的推論法を深層学習に用いることは難しく、科学的仮説の厳密な検証法は十分には進展していない。これは、深層・大規模モデルでは適切な極限分布が使えないなど、複数の統計学的障害に由来する。本研究では、高次元ガウス近似などの非漸近な分布近似法を用いることでその障害を克服し、深層学習に適用可能な統計的推論の基盤を構築する。深層学習を用いた科学的仮説の検証・確定を可能にし、深層・大規模モデルによる統計的推論という新しい領域を開拓する。

Outline of Final Research Achievements

In this research project, we developed theories and methods for statistical estimation and inference in high-dimensional models to advance the understanding and utilization of contemporary, highly flexible statistical models. Specifically, we created theories for estimation and inference in high-dimensional settings where the data and the number of parameters diverge proportionally, without imposing sparsity on the parameters.
More precisely, our work includes: (i) evaluating the rate of estimation error in linear regression models where the noise is not independent, (ii) developing bias correction methods and statistical inference methods for high-dimensional single index models with nonlinear elements, and (iii) advancing Bayesian estimation methods using the spectral structure of high-dimensional data.

Academic Significance and Societal Importance of the Research Achievements

本研究は、現在発展の著しい深層学習や人工知能といった大規模統計モデルに対して、それらの性質や活用法を明らかにするという基礎研究の性質を持っている。これらの大規模統計モデルは高い性能を発揮してはいるが、内部がブラックボックスであったり計算コストが大きいなどの欠点も多く持っており、これらの原理を理解して効率的・安定的に運用する技術は社会に強く求められている。本研究が発展することで、大規模統計モデルの誤差を正しく補正して性能を上げたり、その予測の不安定性を評価した厳密な検定法を作ったり、学習のパラメータの選択を効率化して計算コストを下げるといった多くの実用的技術の開発が可能になる。

Report

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

    (125 results)

All 2024 2023 2022 2021 Other

All Int'l Joint Research (9 results) Journal Article (24 results) (of which Int'l Joint Research: 8 results,  Peer Reviewed: 18 results,  Open Access: 24 results) Presentation (85 results) (of which Int'l Joint Research: 27 results,  Invited: 48 results) Book (4 results) Remarks (3 results)

  • [Int'l Joint Research] チェンマイ大学(タイ)

    • Related Report
      2023 Annual Research Report
  • [Int'l Joint Research] ニューヨーク大学(米国)

    • Related Report
      2023 Annual Research Report
  • [Int'l Joint Research] エセック経営学校(シンガポール)

    • Related Report
      2023 Annual Research Report
  • [Int'l Joint Research] トゥウェンテ大学(オランダ)

    • Related Report
      2023 Annual Research Report
  • [Int'l Joint Research] マックスプランク研究所/CISPA(ドイツ)

    • Related Report
      2023 Annual Research Report
  • [Int'l Joint Research] Twenty University(オランダ)

    • Related Report
      2022 Research-status Report
  • [Int'l Joint Research] Max Planck Institute(ドイツ)

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

    • Related Report
      2021 Research-status Report
  • [Int'l Joint Research] University of Twente(オランダ)

    • Related Report
      2021 Research-status Report
  • [Journal Article] Minimax Analysis for Inverse Risk in Nonparametric Planer Invertible Regression2024

    • Author(s)
      Akifumi Okuno, Masaaki Imaizumi
    • Journal Title

      Electronic Journal of Statistics

      Volume: 18 Issue: 1 Pages: 355-394

    • DOI

      10.1214/23-ejs2202

    • Related Report
      2023 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] 論文解説:可逆関数推定の難しさ - 生成モデルを念頭に2024

    • Author(s)
      奥野彰文,今泉允聡
    • Journal Title

      Jxiv

      Volume: -

    • Related Report
      2023 Annual Research Report
    • Open Access
  • [Journal Article] Uniform confidence band for optimal transport map on one-dimensional data2024

    • Author(s)
      Ponnoprat Donlapark、Okano Ryo、Imaizumi Masaaki
    • Journal Title

      Electronic Journal of Statistics

      Volume: 18 Issue: 1 Pages: 515-552

    • DOI

      10.1214/23-ejs2211

    • Related Report
      2023 Annual Research Report
    • Peer Reviewed / Open Access / Int'l Joint Research
  • [Journal Article] Dimension-free bounds for sums of dependent matrices and operators with heavy-tailed distributions2024

    • Author(s)
      Shogo Nakakita,Pierre Alquier,Masaaki Imaizumi
    • Journal Title

      Electronic Journal of Statistics

      Volume: 18 Issue: 1

    • DOI

      10.1214/24-ejs2224

    • Related Report
      2023 Annual Research Report
    • Peer Reviewed / Open Access / Int'l Joint Research
  • [Journal Article] High-dimensional Contextual Bandit Problem without Sparsity2023

    • Author(s)
      Junpei Komiyama,Masaaki Imaizumi
    • Journal Title

      Advances in Neural Information Processing Systems

      Volume: -

    • Related Report
      2023 Annual Research Report
    • Peer Reviewed / Open Access / Int'l Joint Research
  • [Journal Article] Distribution-on-Distribution Regression with Wasserstein Metric: Multivariate Gaussian Case2023

    • Author(s)
      Ryo Okano,Masaaki Imaizumi
    • Journal Title

      arXiv

      Volume: -

    • Related Report
      2023 Annual Research Report
    • Open Access
  • [Journal Article] Sup-Norm Convergence of Deep Neural Network Estimator for Nonparametric Regression by Adversarial Training2023

    • Author(s)
      Masaaki Imaizumi
    • Journal Title

      arXiv

      Volume: -

    • Related Report
      2023 Annual Research Report
    • Open Access
  • [Journal Article] Statistical Inference in High-Dimensional Generalized Linear Models with Asymmetric Link Functions2023

    • Author(s)
      Kazuma Sawaya,Yoshimasa Uematsu,Masaaki Imaizumi
    • Journal Title

      arXiv

      Volume: -

    • Related Report
      2023 Annual Research Report
    • Open Access
  • [Journal Article] Bayesian Analysis for Over-parameterized Linear Model without Sparsity2023

    • Author(s)
      Tomoya Wakayama,Masaaki Imaizumi
    • Journal Title

      arXiv

      Volume: -

    • Related Report
      2023 Annual Research Report
    • Open Access
  • [Journal Article] Benign Overfitting of Non-Sparse High-Dimensional Linear Regression with Correlated Noise2023

    • Author(s)
      Toshiki Tsuda,Masaaki Imaizumi
    • Journal Title

      arXiv

      Volume: -

    • Related Report
      2023 Annual Research Report
    • Open Access
  • [Journal Article] Unified Perspective on Probability Divergence via the Density-Ratio Likelihood: Bridging KL-Divergence and Integral Probability Metrics2023

    • Author(s)
      Masahiro Kato,Masaaki Imaizumi,Kentaro Minami
    • Journal Title

      Artificial Intelligence and Statistics

      Volume: PMLR 206

    • Related Report
      2023 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] Unified Perspective on Probability Divergence via the Density-Ratio Likelihood: Bridging KL-Divergence and Integral Probability Metrics2023

    • Author(s)
      Masahiro Kato, Masaaki Imaizumi, Kentaro Minami
    • Journal Title

      Proceedings of Machine Learning Research

      Volume: 206

    • Related Report
      2022 Research-status Report
    • Peer Reviewed / Open Access
  • [Journal Article] On Generalization Bounds for Deep Networks Based on Loss Surface Implicit Regularization2023

    • Author(s)
      Masaaki Imaizumi, Johannes Schmidt-Hieber
    • Journal Title

      IEEE Transactions on Information Theory

      Volume: 69 Issue: 2 Pages: 1203-1223

    • DOI

      10.1109/tit.2022.3215088

    • Related Report
      2022 Research-status Report
    • Peer Reviewed / Open Access / Int'l Joint Research
  • [Journal Article] Instrumental variable regression via kernel maximum moment loss2023

    • Author(s)
      Zhang Rui、Imaizumi Masaaki、Sch?lkopf Bernhard、Muandet Krikamol
    • Journal Title

      Journal of Causal Inference

      Volume: 11 Issue: 1

    • DOI

      10.1515/jci-2022-0073

    • Related Report
      2022 Research-status Report
    • Peer Reviewed / Open Access / Int'l Joint Research
  • [Journal Article] In Estimating Functions with Singularity on Hypersurfaces and Advantages of Deep Neural Networks2022

    • Author(s)
      今泉 允聡
    • Journal Title

      Journal of the Japan Statistical Society, Japanese Issue

      Volume: 52 Issue: 1 Pages: 33-51

    • DOI

      10.11329/jjssj.52.33

    • ISSN
      0389-5602, 2189-1478
    • Year and Date
      2022-09-13
    • Related Report
      2022 Research-status Report
    • Peer Reviewed / Open Access
  • [Journal Article] Fast Convergence on Perfect Classification for Functional Data2022

    • Author(s)
      Tomoya Wakayama, Masaaki Imaizumi
    • Journal Title

      Statistica Sinica

      Volume: to appear

    • Related Report
      2022 Research-status Report
    • Peer Reviewed / Open Access
  • [Journal Article] Inference for Projection-Based Wasserstein Distances on Finite Spaces2022

    • Author(s)
      Ryo Okano, Masaaki Imaizumi
    • Journal Title

      Statistica Sinica

      Volume: to appear

    • Related Report
      2022 Research-status Report
    • Peer Reviewed / Open Access
  • [Journal Article] Advantage of Deep Neural Networks for Estimating Functions with Singularity on Hypersurfaces2022

    • Author(s)
      Masaaki Imaizumi, Kenji Fukumizu
    • Journal Title

      Journal of Machine Learning Research

      Volume: 23 Pages: 1-54

    • Related Report
      2022 Research-status Report
    • Peer Reviewed / Open Access
  • [Journal Article] Learning Causal Models from Conditional Moment Restrictions by Importance Weighting2022

    • Author(s)
      Masahiro Kato, Masaaki Imaizumi, Kenichiro McAlinn, Shota Yasui, Haruo Kakehi
    • Journal Title

      International Conference on Learning Representations

      Volume: to appear

    • Related Report
      2022 Research-status Report
    • Peer Reviewed / Open Access / Int'l Joint Research
  • [Journal Article] Learning Causal Relationships from Conditional Moment Restrictions by Importance Weighting2022

    • Author(s)
      M.Kato, M.Imaizumi, K.McAlinn, S.Yasui, H.Kakehi
    • Journal Title

      International Conference on Learning Representations

      Volume: -

    • Related Report
      2021 Research-status Report
    • Peer Reviewed / Open Access / Int'l Joint Research
  • [Journal Article] Hypothesis Test and Confidence Analysis with Wasserstein Distance on General Dimension2022

    • Author(s)
      M.Imaizumi, H.Ota, T.Hamaguchi
    • Journal Title

      Neural Computation

      Volume: -

    • Related Report
      2021 Research-status Report
    • Peer Reviewed / Open Access / Int'l Joint Research
  • [Journal Article] Advantage of Deep Neural Networks for Estimating Functions with Singularity on Hypersurface2022

    • Author(s)
      M.Imaizumi, K.Fukumizu
    • Journal Title

      Journal of Machine Learning Research

      Volume: 23 Pages: 1-54

    • Related Report
      2021 Research-status Report
    • Peer Reviewed / Open Access
  • [Journal Article] Improved Generalization Bounds of Group Invariant / Equivariant Deep Networks via Quotient Feature Spaces2021

    • Author(s)
      A.Sannai, M.Imaizumi, M.Kawano
    • Journal Title

      PMLR: Uncertainty on Artificial Intelligence

      Volume: 161 Pages: 771-780

    • Related Report
      2021 Research-status Report
    • Peer Reviewed / Open Access
  • [Journal Article] Frechet Kernels for Trajectory Data Analysis2021

    • Author(s)
      Koh Takeuchi, Masaaki Imaizumi, Shunsuke Kanda, Keisuke Fujii, Masakazu Ishihata, Takuya Maekawa, Ken Yoda & Yasuo Tabei
    • Journal Title

      In Proceedings of 29th ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems (ACM SIGSPATIAL 2021)

      Volume: - Pages: 221-224

    • DOI

      10.1145/3474717.3483949

    • Related Report
      2021 Research-status Report
    • Peer Reviewed / Open Access
  • [Presentation] インコンテキスト学習の統計学的解析2024

    • Author(s)
      今泉允聡
    • Organizer
      フォレストワークショップ2024
    • Related Report
      2023 Annual Research Report
  • [Presentation] Statistical Analysis on Overparameterized Models and In-Context Learning2024

    • Author(s)
      Masaaki Imaizumi
    • Organizer
      Workshop on Functional Inference and Machine Intelligence
    • Related Report
      2023 Annual Research Report
    • Int'l Joint Research / Invited
  • [Presentation] Statistics for Modern Data Science: Statistical Analysis on Overparameterized Models and In-Context Learning2024

    • Author(s)
      Masaaki Imaizumi
    • Organizer
      International Conference on Scientific Computing and Machine Learning 2024
    • Related Report
      2023 Annual Research Report
    • Int'l Joint Research / Invited
  • [Presentation] Deep Learning: Theory, Applications, and Implications2024

    • Author(s)
      Masaaki Imaizumi
    • Organizer
      Deep Learning: Theory, Applications, and Implications
    • Related Report
      2023 Annual Research Report
    • Int'l Joint Research / Invited
  • [Presentation] 深層学習と過剰パラメータの理論2024

    • Author(s)
      今泉允聡
    • Organizer
      計算物理春の学校2024
    • Related Report
      2023 Annual Research Report
    • Invited
  • [Presentation] ワッサースタイン空間における測地線主成分分析を用いた確率分布のクラスタリング2024

    • Author(s)
      岡野遼,今泉允聡
    • Organizer
      第18回日本統計学会春季集会
    • Related Report
      2023 Annual Research Report
  • [Presentation] Introduction to Theory of Deep Learning2024

    • Author(s)
      Masaaki Imaizumi
    • Organizer
      The Machine Learning Summer School in Okinawa 2024
    • Related Report
      2023 Annual Research Report
    • Int'l Joint Research / Invited
  • [Presentation] 確率過程サンプリングと動力学計算の融合による非線形力学系のランドスケープ解析2024

    • Author(s)
      仲田資季,今泉允聡
    • Organizer
      日本応用数理学会第20回研究部会連合発表会
    • Related Report
      2023 Annual Research Report
  • [Presentation] 汎用基盤技術研究グループパネルディスカッション2024

    • Author(s)
      今泉允聡
    • Organizer
      AIPシンポジウム 2023年度 成果報告会
    • Related Report
      2023 Annual Research Report
    • Int'l Joint Research / Invited
  • [Presentation] 深層学習に関する幾つかの理論研究2024

    • Author(s)
      今泉允聡
    • Organizer
      第4回TREFOIL研究会
    • Related Report
      2023 Annual Research Report
  • [Presentation] 現代的データ科学の数理:新しい高次元統計とインコンテキスト学習の統計理論2024

    • Author(s)
      今泉允聡
    • Organizer
      東京大学数理情報学談話会
    • Related Report
      2023 Annual Research Report
    • Invited
  • [Presentation] 現代的データ科学の数理:新しい高次元統計とインコンテキスト学習の統計理論2024

    • Author(s)
      今泉允聡
    • Organizer
      九州大学数学科統計科学セミナー
    • Related Report
      2023 Annual Research Report
    • Invited
  • [Presentation] 深層学習と過剰パラメータの理論2024

    • Author(s)
      今泉允聡
    • Organizer
      総合研究大学院大学葉山セミナー
    • Related Report
      2023 Annual Research Report
    • Invited
  • [Presentation] 深層学習と過剰パラメータの理論、人工知能の理解への試み2024

    • Author(s)
      今泉允聡
    • Organizer
      情報計測オンラインセミナー
    • Related Report
      2023 Annual Research Report
    • Invited
  • [Presentation] インコンテキスト学習の統計学的解析2024

    • Author(s)
      今泉允聡
    • Organizer
      統計科学・機械学習・情報数学の最前線
    • Related Report
      2023 Annual Research Report
    • Invited
  • [Presentation] Statistical Analysis on Generalization Ability of In-Context Learning2024

    • Author(s)
      Masaaki Imaizumi
    • Organizer
      The Mathematics of Data
    • Related Report
      2023 Annual Research Report
    • Int'l Joint Research / Invited
  • [Presentation] Statistical Analysis on Generalization Ability of In-Context Learning2024

    • Author(s)
      Masaaki Imaizumi
    • Organizer
      Institute of Statistical Mathematics Asia Pasific Rim Meeting 2024
    • Related Report
      2023 Annual Research Report
    • Int'l Joint Research / Invited
  • [Presentation] High-dimensional Contextual Bandit Problem without Sparsity2024

    • Author(s)
      Junpei Komiyama, Masaaki Imaizumi
    • Organizer
      Neural Information Processing Systems
    • Related Report
      2023 Annual Research Report
    • Int'l Joint Research
  • [Presentation] 深層学習と過剰パラメータの理論2023

    • Author(s)
      今泉允聡
    • Organizer
      固体地球科学データ同化/データ駆動型地球科学に関する研究会
    • Related Report
      2023 Annual Research Report
    • Invited
  • [Presentation] Non-sparse high-dimensional statistics and its applications2023

    • Author(s)
      Masaaki Imaizumi
    • Organizer
      International Symposium on Recent Advances in Theories and Methodologies for Large Complex Data
    • Related Report
      2023 Annual Research Report
    • Int'l Joint Research / Invited
  • [Presentation] 深層学習と過剰パラメータの理論2023

    • Author(s)
      今泉允聡
    • Organizer
      物理屋のための機械学習講義
    • Related Report
      2023 Annual Research Report
    • Invited
  • [Presentation] 深層学習・人工知能の原理に迫る理論の試み2023

    • Author(s)
      今泉允聡
    • Organizer
      日本地震学会春季大会
    • Related Report
      2023 Annual Research Report
    • Invited
  • [Presentation] 外れ値を含んだデータにおけるSGDの大域収束性について2023

    • Author(s)
      吉田直生,今泉允聡,仲北祥悟
    • Organizer
      IBIS2023
    • Related Report
      2023 Annual Research Report
  • [Presentation] 量子化回帰モデルの汎化誤差解析2023

    • Author(s)
      柏村周平,坂田綾香,今泉允聡
    • Organizer
      IBIS2023
    • Related Report
      2023 Annual Research Report
  • [Presentation] ワッサースタイン計量の下での多変量ガウス分布間の回帰モデル2023

    • Author(s)
      岡野遼,今泉允聡
    • Organizer
      IBIS2023
    • Related Report
      2023 Annual Research Report
  • [Presentation] 非スパースな高次元文脈付きバンディット問題2023

    • Author(s)
      小宮山純平,今泉允聡
    • Organizer
      IBIS2023
    • Related Report
      2023 Annual Research Report
  • [Presentation] ゼロから作る深層学習理論2023

    • Author(s)
      今泉允聡
    • Organizer
      IBIS2023チュートリアル
    • Related Report
      2023 Annual Research Report
    • Invited
  • [Presentation] Dimension-free concentration inequalities for sums of weakly dependent random matrices2023

    • Author(s)
      仲北祥悟,Pierre Alquier,今泉允聡
    • Organizer
      日本数学会秋季総合分科会
    • Related Report
      2023 Annual Research Report
  • [Presentation] 量子化回帰モデルの統計力学的解析2023

    • Author(s)
      柏村周平,坂田綾香,今泉允聡
    • Organizer
      日本物理学会第78回年次大会
    • Related Report
      2023 Annual Research Report
  • [Presentation] 高次元線形回帰モデルのもとでの条件付き平均処置効果の推定2023

    • Author(s)
      加藤真大,今泉允聡
    • Organizer
      2023年度統計関連学会連合大会
    • Related Report
      2023 Annual Research Report
  • [Presentation] A dimension-free concentration inequality for sums of random matrices under dependence2023

    • Author(s)
      仲北祥悟,Alquier Pierre,今泉允聡
    • Organizer
      2023年度統計関連学会連合大会
    • Related Report
      2023 Annual Research Report
  • [Presentation] 確率分布のクラスタリング2023

    • Author(s)
      岡野遼,今泉允聡
    • Organizer
      2023年度統計関連学会連合大会
    • Related Report
      2023 Annual Research Report
  • [Presentation] 最適腕識別:適応的実験計画による方策選択2023

    • Author(s)
      加藤真大,今泉允聡,石原卓弥,北川透
    • Organizer
      2023年度統計関連学会連合大会
    • Related Report
      2023 Annual Research Report
  • [Presentation] 非対称リンクをもつ高次元一般化線形モデルの統計的推論2023

    • Author(s)
      澤谷一磨,植松良公,今泉允聡
    • Organizer
      2024年度統計関連学会連合大会
    • Related Report
      2023 Annual Research Report
  • [Presentation] Augmented Projection Wasserstein 距離2023

    • Author(s)
      杉本実優,今泉允聡,岡野遼
    • Organizer
      2023年度統計関連学会連合大会
    • Related Report
      2023 Annual Research Report
  • [Presentation] Synthetic control methods through predictive synthesis2023

    • Author(s)
      M. Kato,A. Fukuda,K. Takanashi,K. McAlinn,A. Ohda,M. Imaizumi
    • Organizer
      Econometrics and Statistics
    • Related Report
      2023 Annual Research Report
    • Int'l Joint Research / Invited
  • [Presentation] Estimation of single index models in moderately high dimension2023

    • Author(s)
      Kazuma Sawaya,Yoshimasa Uematsu,Masaaki Imaizumi
    • Organizer
      Econometric and Statistics
    • Related Report
      2023 Annual Research Report
    • Invited
  • [Presentation] DEEP LEARNING: Theory, Algorithms, and Applications 20232023

    • Author(s)
      Masaaki Imaizumi
    • Organizer
      Sup-norm convergence of deep network estimator for nonparametric regression with corrected adversarial training
    • Related Report
      2023 Annual Research Report
    • Int'l Joint Research / Invited
  • [Presentation] On Generalization Bounds for Deep Networks based on Loss Surface Implicit Regularization2023

    • Author(s)
      Masaaki Imaizumi
    • Organizer
      Latin American Congress of Probability and Mathematical Statistics
    • Related Report
      2023 Annual Research Report
    • Int'l Joint Research / Invited
  • [Presentation] Sup-norm convergence of deep network estimator for nonparametric regression with corrected adversarial training2023

    • Author(s)
      Masaaki Imaizumi
    • Organizer
      Korian Statistical Society
    • Related Report
      2023 Annual Research Report
    • Int'l Joint Research / Invited
  • [Presentation] 深層学習の原理に迫る ──数学の挑戦──2023

    • Author(s)
      今泉允聡
    • Organizer
      高校生と大学生のための金曜特別講座
    • Related Report
      2023 Annual Research Report
    • Invited
  • [Presentation] 深層学習と過剰パラメータの理論2023

    • Author(s)
      今泉允聡
    • Organizer
      東京大学大学院 数理科学研究科/情報理工学系研究科 数値解析セミナー
    • Related Report
      2023 Annual Research Report
    • Invited
  • [Presentation] 深層学習と過剰パラメータの理論2023

    • Author(s)
      今泉允聡
    • Organizer
      創発的研究支援事業 融合の場
    • Related Report
      2023 Annual Research Report
    • Invited
  • [Presentation] 深層学習の原理に迫る -数学の挑戦-2023

    • Author(s)
      Masaaki Imaizumi
    • Organizer
      京都大学理学部第22回MACSコロキウム
    • Related Report
      2023 Annual Research Report
    • Invited
  • [Presentation] 高次元ガウス近似によるWasserstein距離推定の不確実性評価2022

    • Author(s)
      今泉允聡
    • Organizer
      Workshop OT 2023
    • Related Report
      2022 Research-status Report
    • Invited
  • [Presentation] High-Dimensional Estimators: Universality and Non-Linearity2022

    • Author(s)
      Masaaki Imaizumi
    • Organizer
      Workshop on Functional Inference and Machine Intelligence
    • Related Report
      2022 Research-status Report
    • Int'l Joint Research / Invited
  • [Presentation] フィルターが小さい深層CNNの最適近似レート2022

    • Author(s)
      佐藤佑真、今泉允聡
    • Organizer
      情報論的学習理論と機械学習研究会(IBISML)
    • Related Report
      2022 Research-status Report
  • [Presentation] 深層学習と過剰パラメータの理論2022

    • Author(s)
      今泉允聡
    • Organizer
      フォレストワークショップ2023
    • Related Report
      2022 Research-status Report
    • Invited
  • [Presentation] Why does SGD prefer flat minima?: Through the lens of dynamical systems2022

    • Author(s)
      Hikaru Ibayashi, Masaaki Imaizumi
    • Organizer
      AAAI When Machine Learning meets Dynamical Systems: Theory and Applications
    • Related Report
      2022 Research-status Report
    • Int'l Joint Research
  • [Presentation] 深層学習と過剰パラメータの理論2022

    • Author(s)
      今泉允聡
    • Organizer
      国立精神・神経医療研究センターの脳病態数理・データ科学セミナーシリーズ
    • Related Report
      2022 Research-status Report
    • Invited
  • [Presentation] Best Arm Identification with a Fixed Budget under a Small Gap2022

    • Author(s)
      M.Kato, K.Ariu, M.Imaizumi, M.Uehara, M.Nomura
    • Organizer
      2023 ASA Annual Meeting
    • Related Report
      2022 Research-status Report
    • Int'l Joint Research
  • [Presentation] High-dimensional asymptotics for single-index models via approximate message passing2022

    • Author(s)
      Y. Uematsu, K. Sawaya, M. Imaizumi
    • Organizer
      CMStatistics
    • Related Report
      2022 Research-status Report
    • Int'l Joint Research / Invited
  • [Presentation] Sup-norm convergence of deep network estimator for nonparametric regression with corrected adversarial training2022

    • Author(s)
      M.Imaizumi
    • Organizer
      CMStatistics
    • Related Report
      2022 Research-status Report
    • Int'l Joint Research / Invited
  • [Presentation] Theory of Deep Learning and Overparmeterization2022

    • Author(s)
      M.Imaizumi
    • Organizer
      Online Asian Machine Learning School, Asian Conference on Machine Learning
    • Related Report
      2022 Research-status Report
    • Int'l Joint Research / Invited
  • [Presentation] Hypothesis Test and Confidence Analysis with Wasserstein Distance on General Dimension2022

    • Author(s)
      M.Imaizumi
    • Organizer
      EcoSta
    • Related Report
      2022 Research-status Report
    • Int'l Joint Research / Invited
  • [Presentation] Benign overfitting in stochastic regression2022

    • Author(s)
      S.Nakakita, M.Imaizumi
    • Organizer
      EcoSta
    • Related Report
      2022 Research-status Report
    • Int'l Joint Research / Invited
  • [Presentation] Semiparametric Contextual Best Arm Identification with a Fixed Budget2022

    • Author(s)
      加藤真大, 今泉允聡, 石原卓弥, 北川透
    • Organizer
      IBIS2022
    • Related Report
      2022 Research-status Report
  • [Presentation] Dynamics of Deep Neural Network: A Functional and Diffusion Process Approach2022

    • Author(s)
      西澤渉, 今泉允聡
    • Organizer
      IBIS2022
    • Related Report
      2022 Research-status Report
  • [Presentation] 非スパースな高次元漸近論の理論と応用2022

    • Author(s)
      今泉允聡
    • Organizer
      大規模複雑データの理論と方法論~新たな発展と関連分野への応用~
    • Related Report
      2022 Research-status Report
  • [Presentation] メカニズムとの学際的統合による新しい分散学習理論基盤の構築2022

    • Author(s)
      今泉允聡
    • Organizer
      IPSJ連続セミナー2022「その先へ 情報技術が貢献できること」
    • Related Report
      2022 Research-status Report
    • Invited
  • [Presentation] High-Dimensional Asymptotics of Semiparametric Generalized Linear Models via Approximate Message Passing2022

    • Author(s)
      澤谷 一磨、植松 良公、今泉 允聡
    • Organizer
      統計関連学会連合大会
    • Related Report
      2022 Research-status Report
  • [Presentation] 操作変数による非スパース高次元な線形回帰モデルの良性過適合2022

    • Author(s)
      津田 俊樹、今泉 允聡
    • Organizer
      統計関連学会連合大会
    • Related Report
      2022 Research-status Report
  • [Presentation] 多変量ガウス分布間回帰2022

    • Author(s)
      岡野 遼、今泉 允聡
    • Organizer
      統計関連学会連合大会
    • Related Report
      2022 Research-status Report
  • [Presentation] 補正付敵対的訓練による深層ニューラルネットワーク推定量のSupノルム収束2022

    • Author(s)
      今泉 允聡
    • Organizer
      統計関連学会連合大会
    • Related Report
      2022 Research-status Report
  • [Presentation] Benign-Overfitting of Overparameterized Bayesian Linear Regression2022

    • Author(s)
      若山 智哉、今泉 允聡
    • Organizer
      統計関連学会連合大会
    • Related Report
      2022 Research-status Report
  • [Presentation] 線形時系列モデルにおける良性過適合2022

    • Author(s)
      仲北 祥悟、
    • Organizer
      統計関連学会連合大会
    • Related Report
      2022 Research-status Report
  • [Presentation] 深層学習の原理を明らかにする理論の試み2022

    • Author(s)
      今泉 允聡
    • Organizer
      電子情報通信学会Webinarチュートリアルシリーズ
    • Related Report
      2022 Research-status Report
    • Invited
  • [Presentation] 深層学習の原理に挑む理論の試み2022

    • Author(s)
      今泉 允聡
    • Organizer
      第46回IBISML研究会
    • Related Report
      2022 Research-status Report
  • [Presentation] 深層学習の原理を明らかにする理論の試み2022

    • Author(s)
      Masaaki Imaizumi
    • Organizer
      JFFoS
    • Related Report
      2022 Research-status Report
    • Invited
  • [Presentation] 深層学習の原理記述に向けた構造汎化理論スキームの開発2022

    • Author(s)
      今泉 允聡
    • Organizer
      JST創発的研究支援事業「融合の場」第1回公開シンポジウム
    • Related Report
      2022 Research-status Report
  • [Presentation] Benign Overfitting in Overparameterized Time Series Models2022

    • Author(s)
      Shogo Nakakita, Masaaki Imaizumi
    • Organizer
      Workshop on the Theory of Overparameterized Machine Learning
    • Related Report
      2022 Research-status Report
    • Int'l Joint Research
  • [Presentation] Benign Overfitting in Conditional Average Treatment Effect Prediction with Linear Regression2022

    • Author(s)
      Masahiro Kato, Masaaki Imaizumi
    • Organizer
      Workshop on the Theory of Overparameterized Machine Learning
    • Related Report
      2022 Research-status Report
    • Int'l Joint Research
  • [Presentation] Stability of Deep Network Estimator for Nonparametric Regression2022

    • Author(s)
      Masaaki Imaizumi
    • Organizer
      Workshop on Functional Inference and Machine Intelligence
    • Related Report
      2021 Research-status Report
    • Int'l Joint Research / Invited
  • [Presentation] Exponential escape efficiency of SGD from sharp minima2022

    • Author(s)
      H. Ibayashi, M. Imaizumi
    • Organizer
      Workshop on Functional Inference and Machine Intelligence
    • Related Report
      2021 Research-status Report
    • Int'l Joint Research
  • [Presentation] Minimum sharpness: Scale-invariant parameter-robus…2021

    • Author(s)
      H.Ibayashi, T.Hamaguchi, M.Imaizumi
    • Organizer
      ICML Workshop on Theoretic Foundation, Criticism, …
    • Related Report
      2021 Research-status Report
    • Int'l Joint Research
  • [Presentation] 深層学習の理論2021

    • Author(s)
      今泉允聡
    • Organizer
      松尾研セミナー
    • Related Report
      2021 Research-status Report
    • Invited
  • [Presentation] Implicit Regularization and Over-parameterization2021

    • Author(s)
      Masaaki Imaizumi
    • Organizer
      知の物理学センター
    • Related Report
      2021 Research-status Report
    • Invited
  • [Presentation] On Gaussian Approximation on M-Estimator2021

    • Author(s)
      Masaaki Imaizumi
    • Organizer
      International Chinese Statistical Association
    • Related Report
      2021 Research-status Report
    • Int'l Joint Research / Invited
  • [Presentation] 関数推定の理論に基づく深層学習の原理解析2021

    • Author(s)
      今泉允聡
    • Organizer
      微分方程式とデータサイエンス研究会
    • Related Report
      2021 Research-status Report
    • Invited
  • [Presentation] 深層学習の原理を明らかにする理論の試み2021

    • Author(s)
      今泉允聡
    • Organizer
      統計関連学会連合大会
    • Related Report
      2021 Research-status Report
    • Invited
  • [Presentation] 深層学習の原理を明らかにする理論の試み2021

    • Author(s)
      今泉允聡
    • Organizer
      中央大学理工学研究所特別講演会
    • Related Report
      2021 Research-status Report
    • Invited
  • [Presentation] Deep Neural Networks Learn Non-Smooth Functions Effectively2021

    • Author(s)
      今泉允聡
    • Organizer
      細谷賞セミナー
    • Related Report
      2021 Research-status Report
    • Invited
  • [Presentation] Gradient Descent Algorithm with Path Kernel-based …2021

    • Author(s)
      西澤 渉, 今泉允聡
    • Organizer
      IBIS2021
    • Related Report
      2021 Research-status Report
  • [Presentation] 深層学習の原理を明らかにする理論の試み2021

    • Author(s)
      今泉允聡
    • Organizer
      顕微鏡計測インフォマティックス部会
    • Related Report
      2021 Research-status Report
    • Invited
  • [Presentation] AIの原理を解き明かす新理論2021

    • Author(s)
      今泉允聡
    • Organizer
      JST理事長記者会見
    • Related Report
      2021 Research-status Report
  • [Book] 人工知能とどうつきあうか: 哲学から考える2023

    • Author(s)
      鈴木貴之, 柴田崇, 今泉允聡, 大塚淳, 中澤栄輔, 小野哲雄, 植原亮, 立花幸司, 上杉繁, 堀浩一, 関口海良
    • Total Pages
      256
    • Publisher
      勁草書房
    • Related Report
      2023 Annual Research Report
  • [Book] 応用基礎としてのデータサイエンス AI×データ活用の実践2023

    • Author(s)
      北川 源四郎、竹村 彰通、赤穂 昭太郎、今泉 允聡、内田 誠一、清 智也、高野 渉、辻 真吾、原 尚幸、久野 遼平、松原 仁、宮地 充子、森畑 明昌、宿久 洋
    • Total Pages
      384
    • Publisher
      講談社
    • ISBN
      4065307899
    • Related Report
      2022 Research-status Report
  • [Book] ニューラルネットワークとディープラーニング2022

    • Author(s)
      Charu C. Aggarwal、竹村 彰通、今泉 允聡、李 鍾賛、今井 貴史、今井 徹、紅林 亘、齋藤 邦彦、健山 智子、寺田 裕、西出 俊、西出 亮
    • Total Pages
      520
    • Publisher
      学術図書出版社
    • ISBN
      4780607140
    • Related Report
      2022 Research-status Report 2021 Research-status Report
  • [Book] 深層学習の原理に迫る2021

    • Author(s)
      今泉 允聡
    • Total Pages
      126
    • Publisher
      岩波書店
    • ISBN
      4000297031
    • Related Report
      2021 Research-status Report
  • [Remarks] 深層学習の原理を説明する新理論――ニューラルネットワークのエネルギー曲面上の滞留現象

    • URL

      https://www.u-tokyo.ac.jp/focus/ja/press/z0109_00065.html

    • Related Report
      2023 Annual Research Report 2022 Research-status Report
  • [Remarks] 深層学習が優位性を発揮する特異データ空間の存在を証明

    • URL

      https://research-er.jp/articles/view/111826

    • Related Report
      2022 Research-status Report
  • [Remarks] 深層学習によるデータ固有のフラクタル構造などへの適応を証明

    • URL

      https://www.u-tokyo.ac.jp/focus/ja/articles/z0508_00102.html

    • Related Report
      2021 Research-status Report

URL: 

Published: 2021-04-28   Modified: 2025-01-30  

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi