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

Karakida Ryo  唐木田 亮

ORCIDConnect your ORCID iD *help
Researcher Number 30803902
Other IDs
Affiliation (Current) 2025: 国立研究開発法人産業技術総合研究所, 情報・人間工学領域, 主任研究員
Affiliation (based on the past Project Information) *help 2021 – 2024: 国立研究開発法人産業技術総合研究所, 情報・人間工学領域, 主任研究員
2017 – 2020: 国立研究開発法人産業技術総合研究所, 情報・人間工学領域, 研究員
Review Section/Research Field
Principal Investigator
Basic Section 61040:Soft computing-related / Soft computing
Except Principal Investigator
Sections That Are Subject to Joint Review: Basic Section61020:Human interface and interaction-related , Basic Section62040:Entertainment and game informatics-related / Basic Section 62040:Entertainment and game informatics-related / Basic Section 61020:Human interface and interaction-related / Transformative Research Areas, Section (II)
Keywords
Principal Investigator
深層学習 / 機械学習 / ニューラルネットワーク / 統計力学的解析 / ランダム行列 / 統計力学 / 数理工学 / 統計物理 / 最適化 / 力学系 … More / 継続学習 / レプリカ法 / ソフトコンピューティング / 人工知能 / 情報幾何 … More
Except Principal Investigator
継続学習 / 転移学習 / 筋電 / 機械学習 / 深層学習 Less
  • Research Projects

    (5 results)
  • Research Products

    (64 results)
  • Co-Researchers

    (4 People)
  •  Development of an sEMG-based Human-Computer Interface Utilizing Deep Transfer Learning and Continual Learning

    • Principal Investigator
      叶賀 卓
    • Project Period (FY)
      2023 – 2026
    • Research Category
      Grant-in-Aid for Scientific Research (B)
    • Review Section
      Basic Section 61020:Human interface and interaction-related
      Basic Section 62040:Entertainment and game informatics-related
      Sections That Are Subject to Joint Review: Basic Section61020:Human interface and interaction-related , Basic Section62040:Entertainment and game informatics-related
    • Research Institution
      National Institute of Advanced Industrial Science and Technology
  •  線形性と非線形性の協同による可解なランダム神経回路の深化Principal Investigator

    • Principal Investigator
      唐木田 亮
    • Project Period (FY)
      2023 – 2027
    • Research Category
      Grant-in-Aid for Early-Career Scientists
    • Review Section
      Basic Section 61040:Soft computing-related
    • Research Institution
      National Institute of Advanced Industrial Science and Technology
  •  Mathematics and application of deep learning

    • Principal Investigator
      田中 章詞
    • Project Period (FY)
      2022 – 2026
    • Research Category
      Grant-in-Aid for Transformative Research Areas (A)
    • Review Section
      Transformative Research Areas, Section (II)
    • Research Institution
      Institute of Physical and Chemical Research
  •  Mathematical Foundations of Random Deep Neural Networks and their applications to machine-learning problemsPrincipal Investigator

    • Principal Investigator
      Karakida Ryo
    • Project Period (FY)
      2019 – 2022
    • Research Category
      Grant-in-Aid for Early-Career Scientists
    • Review Section
      Basic Section 61040:Soft computing-related
    • Research Institution
      National Institute of Advanced Industrial Science and Technology
  •  Theoretical construction and control of deep learning based on the geometry of hierarchical modelsPrincipal Investigator

    • Principal Investigator
      Karakida Ryo
    • Project Period (FY)
      2017 – 2018
    • Research Category
      Grant-in-Aid for Research Activity Start-up
    • Research Field
      Soft computing
    • Research Institution
      National Institute of Advanced Industrial Science and Technology

All 2024 2023 2022 2021 2020 2019 2018 2017 Other

All Journal Article Presentation Book Other

  • [Book] 数理科学 深層神経回路網の幾何~ 統計神経力学とのつながり ~2020

    • Author(s)
      唐木田 亮
    • Total Pages
      7
    • Publisher
      サイエンス社
    • Data Source
      KAKENHI-PROJECT-19K20366
  • [Book] 数理科学 (深層学習の数理)2018

    • Author(s)
      唐木田亮, 麻生英樹
    • Total Pages
      8
    • Publisher
      サイエンス社
    • Data Source
      KAKENHI-PROJECT-17H07390
  • [Journal Article] Attention in a family of Boltzmann machines emerging from modern Hopfield networks2023

    • Author(s)
      Ota, Toshihiro and Karakida, Ryo
    • Journal Title

      Neural Computation

      Volume: in press

    • Peer Reviewed
    • Data Source
      KAKENHI-PLANNED-22H05116
  • [Journal Article] Understanding Gradient Regularization in Deep Learning: Efficient Finite-Difference Computation and Implicit Bias2023

    • Author(s)
      Ryo Karakida, Tomoumi Takase, Tomohiro Hayase & Kazuki Osawa
    • Journal Title

      Proceedings of ICLR (PMLR)

      Volume: 202 Pages: 1-19

    • Peer Reviewed / Open Access
    • Data Source
      KAKENHI-PROJECT-23K16965
  • [Journal Article] Attention in a Family of Boltzmann Machines Emerging From Modern Hopfield Networks2023

    • Author(s)
      Ota Toshihiro、Karakida Ryo
    • Journal Title

      Neural Computation

      Volume: 35 Issue: 8 Pages: 1463-1480

    • DOI

      10.1162/neco_a_01597

    • Peer Reviewed
    • Data Source
      KAKENHI-PROJECT-23K16965
  • [Journal Article] Attention in a family of Boltzmann machines emerging from modern Hopfield networks2023

    • Author(s)
      Toshihiro Ota, Ryo Karakida
    • Journal Title

      Neural Computation

      Volume: -

    • Data Source
      KAKENHI-PROJECT-19K20366
  • [Journal Article] Learning Curves for Continual Learning in Neural Networks: Self-Knowledge Transfer and Forgetting2022

    • Author(s)
      Ryo Karakida, Shotaro Akaho
    • Journal Title

      International Conference on Learning Representations

      Volume: - Pages: 1-27

    • Peer Reviewed / Open Access
    • Data Source
      KAKENHI-PROJECT-19K20366
  • [Journal Article] Learning Curves for Continual Learning in Neural Networks: Self-Knowledge Transfer and Forgetting2022

    • Author(s)
      Karakida, Ryo and Shotaro, Akaho
    • Journal Title

      International Conference on Learning Representations (ICLR)

      Volume: なし Pages: 1-27

    • Peer Reviewed / Open Access
    • Data Source
      KAKENHI-PLANNED-22H05116
  • [Journal Article] Understanding approximate Fisher information for fast convergence of natural gradient descent in wide neural networks*2021

    • Author(s)
      Karakida Ryo, Osawa Kazuki
    • Journal Title

      Journal of Statistical Mechanics: Theory and Experiment

      Volume: 2021 Issue: 12 Pages: 124010-124010

    • DOI

      10.1088/1742-5468/ac3ae3

    • Peer Reviewed / Open Access
    • Data Source
      KAKENHI-PROJECT-19K20366
  • [Journal Article] The Spectrum of Fisher Information of Deep Networks Achieving Dynamical Isometry2021

    • Author(s)
      Tomohiro Hayase, Ryo Karakida
    • Journal Title

      Proceedings of AISTATS

      Volume: -

    • Peer Reviewed / Open Access
    • Data Source
      KAKENHI-PROJECT-19K20366
  • [Journal Article] Self-paced data augmentation for training neural networks2021

    • Author(s)
      Tomoumi Takase, Ryo Karakida, Hideki Asoh
    • Journal Title

      Neurocomputing

      Volume: 442 Pages: 296-306

    • DOI

      10.1016/j.neucom.2021.02.080

    • Peer Reviewed / Open Access
    • Data Source
      KAKENHI-PROJECT-20K19888, KAKENHI-PROJECT-19K20366
  • [Journal Article] Understanding Approximate Fisher Information for Fast Convergence of Natural Gradient Descent in Wide Neural Networks2020

    • Author(s)
      Ryo Karakida, Kazuki Osawa
    • Journal Title

      Proceedings of Conference on Neural Information Processing Systems (NeurIPS)

      Volume: 33

    • Peer Reviewed / Open Access
    • Data Source
      KAKENHI-PROJECT-19K20366
  • [Journal Article] Collective dynamics of repeated inference in variational autoencoder rapidly find cluster structure2020

    • Author(s)
      Nagano Yoshihiro、Karakida Ryo、Okada Masato
    • Journal Title

      Scientific Reports

      Volume: 10 Issue: 1

    • DOI

      10.1038/s41598-020-72593-4

    • Peer Reviewed / Open Access
    • Data Source
      KAKENHI-PROJECT-19K20366, KAKENHI-PROJECT-18H04106
  • [Journal Article] Fisher Information and Natural Gradient Learning in Random Deep Networks2019

    • Author(s)
      Shun-ichi Amari, Ryo Karakida, Masafumi Oizumi
    • Journal Title

      Proceedings of Conference on Artificial Intelligence and Statistics

      Volume: -

    • Peer Reviewed / Open Access
    • Data Source
      KAKENHI-PROJECT-19K20366
  • [Journal Article] The Normalization Method for Alleviating Pathological Sharpness in Wide Neural Networks2019

    • Author(s)
      Ryo Karakida, Shotaro Akaho, Shun-ichi Amari
    • Journal Title

      Proceedings of Conference on Neural Information Processing Systems

      Volume: -

    • Peer Reviewed / Open Access
    • Data Source
      KAKENHI-PROJECT-19K20366
  • [Journal Article] Statistical neurodynamics of deep networks: geometry of signal spaces2019

    • Author(s)
      Shun-ichi Amari, Ryo Karakida, Masafumi Oizumi
    • Journal Title

      NOLTA

      Volume: 10 Issue: 4 Pages: 322-336

    • DOI

      10.1587/nolta.10.322

    • NAID

      130007722599

    • ISSN
      2185-4106
    • Language
      English
    • Peer Reviewed
    • Data Source
      KAKENHI-PROJECT-19K20366
  • [Journal Article] Statistical mechanical analysis of learning dynamics of two-layer perceptron with multiple output units2019

    • Author(s)
      Yuki Yoshida, Ryo Karakida, Masato Okada, Shun-ichi Amari
    • Journal Title

      Journal of Physics A: Mathematical and Theoretical

      Volume: 52 Pages: 1-17

    • Peer Reviewed / Open Access
    • Data Source
      KAKENHI-PROJECT-17H07390
  • [Journal Article] Fisher Information and Natural Gradient Learning in Random Deep Networks2019

    • Author(s)
      Shun-ichi Amari, Ryo Karakida, Masafumi Oizumi
    • Journal Title

      Proceedings of Machine Learning Research (AISTATS)

      Volume: 89 Pages: 694-702

    • Peer Reviewed / Open Access
    • Data Source
      KAKENHI-PROJECT-17H07390
  • [Journal Article] Universal Statistics of Fisher Information in Deep Neural Networks: Mean Field Approach2019

    • Author(s)
      Ryo Karakida, Shotaro Akaho, Shun-ichi Amari
    • Journal Title

      Proceedings of Conference on Artificial Intelligence and Statistics

      Volume: -

    • Peer Reviewed / Open Access
    • Data Source
      KAKENHI-PROJECT-19K20366
  • [Journal Article] Universal Statistics of Fisher Information in Deep Neural Networks: Mean Field Approach2019

    • Author(s)
      Ryo Karakida, Shotaro Akaho, Shun-ichi Amari
    • Journal Title

      Proceedings of Machine Learning Research (AISTATS)

      Volume: 89 Pages: 1032-1041

    • Peer Reviewed / Open Access
    • Data Source
      KAKENHI-PROJECT-17H07390
  • [Presentation] 特徴学習領域における最終層のスケール解析2024

    • Author(s)
      唐木田亮
    • Organizer
      日本物理学会春季大会
    • Data Source
      KAKENHI-PROJECT-23K16965
  • [Presentation] 対角線形ネットにおける勾配正則化の陰的バイアス2023

    • Author(s)
      唐木田亮, 高瀬朝海, 早瀬友裕, 大沢和樹
    • Organizer
      日本物理学会 2023年春季大会
    • Data Source
      KAKENHI-PLANNED-22H05116
  • [Presentation] 対角線形ネットにおける勾配正則化の陰的バイアス2023

    • Author(s)
      唐木田亮, 高瀬朝海, 早瀬友裕, 大沢和樹
    • Organizer
      日本物理学会2023年春季大会
    • Data Source
      KAKENHI-PROJECT-19K20366
  • [Presentation] Self-knowledge forgetting in continual learning: Insight from a solvable overparameterized model2023

    • Author(s)
      唐木田亮
    • Organizer
      人工知能学会
    • Data Source
      KAKENHI-PROJECT-23K16965
  • [Presentation] Understanding deep-learning algorithms through learning regimes2023

    • Author(s)
      Ryo Karakida
    • Organizer
      International conference on MACHINE LEARNING PHYSICS
    • Int'l Joint Research
    • Data Source
      KAKENHI-PROJECT-23K16965
  • [Presentation] 深層学習アルゴリズムを操る2023

    • Author(s)
      唐木田亮
    • Organizer
      MCMEセミナー
    • Invited
    • Data Source
      KAKENHI-PROJECT-23K16965
  • [Presentation] Understanding Implicit Bias of Learning Dynamics in Overparameterized Regimes2023

    • Author(s)
      Ryo Karakida
    • Organizer
      9IDMRCS
    • Invited / Int'l Joint Research
    • Data Source
      KAKENHI-PROJECT-23K16965
  • [Presentation] 効率的な勾配正則化アルゴリズムとその陰的バイアスの解析2022

    • Author(s)
      唐木田亮, 高瀬朝海, 早瀬友裕, 大沢和樹
    • Organizer
      IIBIS2022
    • Data Source
      KAKENHI-PLANNED-22H05116
  • [Presentation] 効率的な勾配正則化アルゴリズムとその陰的バイアスの解析2022

    • Author(s)
      唐木田亮, 高瀬朝海, 早瀬友裕, 大沢和樹
    • Organizer
      IBIS2022
    • Data Source
      KAKENHI-PROJECT-19K20366
  • [Presentation] Neural tangent kernel regimeにおける継続学習の学習曲線2022

    • Author(s)
      唐木田亮
    • Organizer
      日本応用数理学会2022年度年会
    • Invited
    • Data Source
      KAKENHI-PROJECT-19K20366
  • [Presentation] Neural tangent kernel regimeにおける継続学習の学習曲線2022

    • Author(s)
      唐木田亮
    • Organizer
      日本応用数理学会2022年度年会
    • Invited
    • Data Source
      KAKENHI-PLANNED-22H05116
  • [Presentation] カーネル法の統計力学的解析とそれによる継続学習の評価2022

    • Author(s)
      唐木田亮
    • Organizer
      統計物理と統計科学のセミナー
    • Invited
    • Data Source
      KAKENHI-PLANNED-22H05116
  • [Presentation] Learning Curves for Continual Learning in Neural Networks: Self-Knowledge Transfer and Forgetting2022

    • Author(s)
      Karakida, Ryo and Akaho, Shotaro
    • Organizer
      International Conference on Learning Representations (ICLR)
    • Int'l Joint Research
    • Data Source
      KAKENHI-PLANNED-22H05116
  • [Presentation] カーネル法の統計力学的解析とそれによる継続学習の評価2022

    • Author(s)
      唐木田亮
    • Organizer
      統計物理と統計科学のセミナー
    • Invited
    • Data Source
      KAKENHI-PROJECT-19K20366
  • [Presentation] 継続学習における自己知識転移と忘却2022

    • Author(s)
      唐木田亮
    • Organizer
      第51回統計的機械学習セミナー(オンライン)
    • Invited
    • Data Source
      KAKENHI-PLANNED-22H05116
  • [Presentation] 継続学習における自己知識転移と忘却2022

    • Author(s)
      唐木田亮
    • Organizer
      第51回統計的機械学習セミナ-
    • Invited
    • Data Source
      KAKENHI-PROJECT-19K20366
  • [Presentation] Understanding Approximate Fisher Information for Fast Convergence of Natural Gradient Descent in Wide Neural Networks2021

    • Author(s)
      Ryo Karakida
    • Organizer
      Math Machine Learning Seminar MPI MIS + UCLA
    • Invited
    • Data Source
      KAKENHI-PROJECT-19K20366
  • [Presentation] Improving the trainability of deep neural networks: A perspective from the infinite width limit2021

    • Author(s)
      Ryo Karakida
    • Organizer
      4th international conference on econometrics and statistics
    • Invited / Int'l Joint Research
    • Data Source
      KAKENHI-PROJECT-19K20366
  • [Presentation] 幅無限大深層モデルにおける近似自然勾配法の収束解析2021

    • Author(s)
      唐木田 亮
    • Organizer
      日本物理学会 第76回年次大会
    • Data Source
      KAKENHI-PROJECT-19K20366
  • [Presentation] 継続学習における転移と忘却: NTK regimeのレプリカ解析2021

    • Author(s)
      唐木田 亮, 赤穗 昭太郎
    • Organizer
      IBIS2021
    • Data Source
      KAKENHI-PROJECT-19K20366
  • [Presentation] The Spectrum of Fisher Information of Deep Networks Achieving Dynamical Isometry2021

    • Author(s)
      Tomohiro Hayase, Ryo Karakida
    • Organizer
      AISTATS
    • Int'l Joint Research
    • Data Source
      KAKENHI-PROJECT-19K20366
  • [Presentation] 深層学習の数理: 統計力学的アプローチ2020

    • Author(s)
      唐木田 亮
    • Organizer
      ディープラーニングと物理学2020 オンライン
    • Invited
    • Data Source
      KAKENHI-PROJECT-19K20366
  • [Presentation] 深層学習の数理: ランダム行列と統計力学的視点2020

    • Author(s)
      唐木田 亮
    • Organizer
      Random Matrices, Free Probability, and Machine Learning ワークショップ
    • Invited
    • Data Source
      KAKENHI-PROJECT-19K20366
  • [Presentation] Understanding Approximate Fisher Information for Fast Convergence of Natural Gradient Descent in Wide Neural Networks2020

    • Author(s)
      Ryo Karakida
    • Organizer
      Conference on Neural Information Processing Systems (NeurIPS)
    • Int'l Joint Research
    • Data Source
      KAKENHI-PROJECT-19K20366
  • [Presentation] 深層モデルにおいて高速に収束する近似自然勾配法の理論解析2020

    • Author(s)
      唐木田 亮, 大沢 和樹
    • Organizer
      IBIS2020
    • Data Source
      KAKENHI-PROJECT-19K20366
  • [Presentation] ランダムなBackpropagation学習における巨視的ダイナミクスの生成汎関数法的解析2020

    • Author(s)
      唐木田亮
    • Organizer
      日本物理学会 第75回年次大会
    • Data Source
      KAKENHI-PROJECT-19K20366
  • [Presentation] The Normalization Method for Alleviating Pathological Sharpness in Wide Neural Networks2019

    • Author(s)
      Ryo Karakida, Shotaro Akaho, Shun-ichi Amari
    • Organizer
      Conference on Neural Information Processing Systems
    • Data Source
      KAKENHI-PROJECT-19K20366
  • [Presentation] Fisher Information of Deep Neural Networks With Random Weights2019

    • Author(s)
      Ryo Karakida
    • Organizer
      The 11th ICSA international conference
    • Invited / Int'l Joint Research
    • Data Source
      KAKENHI-PROJECT-19K20366
  • [Presentation] ランダム深層ニューラルネットの摂動に対する応答の普遍性2019

    • Author(s)
      唐木田亮, 赤穂昭太郎, 甘利俊一
    • Organizer
      日本物理学会第74回年次大会
    • Data Source
      KAKENHI-PROJECT-17H07390
  • [Presentation] Universal Statistics of Fisher Information in Deep Neural Networks: Mean Field Approach2019

    • Author(s)
      Ryo Karakida, Shotaro Akaho, Shun-ichi Amari
    • Organizer
      Conference on Artificial Intelligence and Statistics
    • Data Source
      KAKENHI-PROJECT-19K20366
  • [Presentation] 層ニューラルネットワークにおけるFisher情報行列の普遍性2018

    • Author(s)
      唐木田亮
    • Organizer
      第30回RAMPシンポジウム
    • Invited
    • Data Source
      KAKENHI-PROJECT-17H07390
  • [Presentation] ランダム深層ニューラルネットワークにおけるFisher情報行列の巨視的理論2018

    • Author(s)
      唐木田亮、赤穂昭太郎、甘利俊一
    • Organizer
      日本物理学会年次大会
    • Data Source
      KAKENHI-PROJECT-17H07390
  • [Presentation] 機械学習から見たニューラルネットワークの数理2018

    • Author(s)
      唐木田亮
    • Organizer
      東京理科大学 脳学際研究部門 第2回公開シンポジウム "脳のサイエンス"
    • Invited
    • Data Source
      KAKENHI-PROJECT-17H07390
  • [Presentation] 深層ニューラルネットワークにおけるFisher情報行列の普遍性2018

    • Author(s)
      唐木田亮, 赤穂昭太郎, 甘利俊一
    • Organizer
      IBIS2018
    • Data Source
      KAKENHI-PROJECT-17H07390
  • [Presentation] 深層ニューラルネットワークの数理: 平均場理論の視点2018

    • Author(s)
      唐木田亮
    • Organizer
      産総研AIセミナー
    • Data Source
      KAKENHI-PROJECT-17H07390
  • [Presentation] Theoretical analysis of RBMs with Gaussian visible units - Dynamical analysis and Riemannian optimization -2018

    • Author(s)
      Ryo Karakida
    • Organizer
      Americal Institute of Mathematics (AIM) workshop "Boltzmann machines"
    • Invited / Int'l Joint Research
    • Data Source
      KAKENHI-PROJECT-17H07390
  • [Presentation] エントロピー正則化付きWasserstein距離の情報幾何2017

    • Author(s)
      唐木田亮
    • Organizer
      第64回幾何学シンポジウム
    • Invited
    • Data Source
      KAKENHI-PROJECT-17H07390
  • [Presentation] Information Geometry of Wasserstein Divergence2017

    • Author(s)
      Ryo Karakida、Shun-ichi Amari
    • Organizer
      Geometric Science of Information
    • Int'l Joint Research
    • Data Source
      KAKENHI-PROJECT-17H07390
  • []

  • []

  • []

  • []

  • []

  • []

  • 1.  田中 章詞 (20791924)
    # of Collaborated Projects: 1 results
    # of Collaborated Products: 0 results
  • 2.  瀧 雅人 (70548221)
    # of Collaborated Projects: 1 results
    # of Collaborated Products: 0 results
  • 3.  叶賀 卓 (40803903)
    # of Collaborated Projects: 1 results
    # of Collaborated Products: 0 results
  • 4.  高瀬 朝海
    # of Collaborated Projects: 0 results
    # of Collaborated Products: 1 results

URL: 

Are you sure that you want to link your ORCID iD to your KAKEN Researcher profile?
* This action can be performed only by the researcher himself/herself who is listed on the KAKEN Researcher’s page. Are you sure that this KAKEN Researcher’s page is your page?

この研究者とORCID iDの連携を行いますか?
※ この処理は、研究者本人だけが実行できます。

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi