MURATA Shingo

写真a

Affiliation

Faculty of Science and Technology, Department of Electronics and Electrical Engineering ( Yagami )

Position

Associate Professor

Related Websites
Scopus Paper Info  
Paper Count: 53 Citation Count: 442 h-index: 11

Click to view the Scopus page. The data was downloaded from Scopus API in May 17, 2026, via http://api.elsevier.com and http://www.scopus.com .

Page Top ▲

Career 【 Display / hide 】

  • 2016.04
    -
    2018.03

    Waseda University, Department of Modern Mechanical Engineering, School of Creative Science and Engineering, Research Associate

  • 2018.04
    -
    2020.03

    National Institute of Informatics, Principles of Informatics Research Division, Assistant Professor

  • 2018.04
    -
    2020.03

    The Graduate University for Advanced Studies, Department of Informatics, School of Multidisciplinary Science, Assistant Professor

  • 2020.04
    -
    2024.03

    Keio University, Department of Electronics and Electrical Engineering, Faculty of Science and Technology, Assistant Professor

  • 2024.04
    -
    Present

    Keio University, Department of Electronics and Electrical Engineering, Faculty of Science and Technology, Associate Professor

Page Top ▲

Academic Background 【 Display / hide 】

  • 2007.04
    -
    2011.03

    Waseda University, School of Creative Science and Engineering, Department of Modern Mechanical Engineering

    University, Graduated

  • 2011.04
    -
    2013.03

    Waseda University, Graduate School of Creative Science and Engineering, Department of Modern Mechanical Engineering

    Graduate School, Completed, Master's course

  • 2013.04
    -
    2016.03

    Waseda University, Graduate School of Creative Science and Engineering, Department of Modern Mechanical Engineering

    Graduate School, Completed, Doctoral course

Page Top ▲

Academic Degrees 【 Display / hide 】

  • Doctor of Engineering, Waseda University, Coursework, 2016.03

Page Top ▲
 

Research Areas 【 Display / hide 】

  • Cognitive Robotics

  • Robot Learning

  • Computational Psychiatry

  • Informatics / Intelligent informatics

  • Informatics / Intelligent robotics

Page Top ▲

Research Keywords 【 Display / hide 】

  • Deep Learning

  • Cognitive Robotics

  • ロボット学習

  • Computational Psychiatry

  • Neural Network

display all >>

Page Top ▲
 

Books 【 Display / hide 】

  • Cognitive Robotics

    Angelo Cangelosi, Minoru Asada, The MIT Press, 2022.04

    Scope: Machine Learning for Cognitive Robotics

Page Top ▲

Papers 【 Display / hide 】

  • Real-World Robot Control by Deep Active Inference With a Temporally Hierarchical World Model

    Fujii K., Murata S.

    IEEE Robotics and Automation Letters 11 ( 1 ) 890 - 897 2026

    　View Summary

    Robots in uncertain realworld environments must perform both goaldirected and exploratory actions. However, most deep learning-based control methods neglect exploration and struggle under uncertainty. To address this, we adopt deep active inference, a framework that accounts for human goal-directed and exploratory actions. Yet, conventional deep active inference approaches face challenges due to limited environmental representation capacity and high computational cost in action selection. We propose a novel deep active inference framework that consists of a world model, an action model, and an abstract world model. The world model encodes environmental dynamics into hidden state representations at slow and fast timescales. The action model compresses action sequences into abstract actions using vector quantization, and the abstract world model predicts future slow states conditioned on the abstract action, enabling low-cost action selection. We evaluate the framework on object-manipulation tasks with a real-world robot. Results show that it achieves high success rates across diverse manipulation tasks and switches between goal-directed and exploratory actions in uncertain settings, while making action selection computationally tractable. These findings highlight the importance of modeling multiple timescale dynamics and abstracting actions and state transitions.

  • System 0/1/2/3: Quad-Process Theory for Multitimescale Embodied Collective Cognitive Systems.

    Tadahiro Taniguchi, Yasushi Hirai, Masahiro Suzuki, Shingo Murata, Takato Horii, Kazutoshi Tanaka

    Artificial life 31 ( 4 ) 465 - 496 2025.12

    　View Summary

    This article introduces the System 0/1/2/3 framework as an extension of dual-process theory, employing a quad-process model of cognition. Expanding upon System 1 (fast, intuitive thinking) and System 2 (slow, deliberative thinking), we incorporate System 0, which represents precognitive embodied processes, and System 3, which encompasses collective intelligence and symbol emergence. We contextualize this model within Bergson's philosophy by adopting multiscale time theory to unify the diverse temporal dynamics of cognition. System 0 emphasizes morphological computation and passive dynamics, illustrating how physical embodiment enables adaptive behavior without explicit neural processing. Systems 1 and 2 are explained from a constructive perspective, incorporating neurodynamical and artificial intelligence (AI) viewpoints. In System 3, we introduce collective predictive coding to explain how societal-level adaptation and symbol emergence operate over extended timescales. This comprehensive framework ranges from rapid embodied reactions to slow-evolving collective intelligence, offering a unified perspective on cognition across multiple timescales, levels of abstraction, and forms of human intelligence. The System 0/1/2/3 model provides a novel theoretical foundation for understanding the interplay between adaptive and cognitive processes, thereby opening new avenues for research in cognitive science, AI, robotics, and collective intelligence.

  • Variational Adaptive Noise and Dropout towards Stable Recurrent Neural Networks

    Taisuke Kobayashi, Shingo Murata

    2025 IEEE International Conference on Development and Learning (ICDL) (IEEE)     1 - 6 2025.09

    　View Summary

    This paper proposes a novel stable learning theory for recurrent neural networks (RNNs), so-called variational adaptive noise and dropout (VAND). As stabilizing factors for RNNs, noise and dropout on the internal state of RNNs have been separately confirmed in previous studies. We reinterpret the optimization problem of RNNs as variational inference, showing that noise and dropout can be derived simultaneously by transforming the explicit regularization term arising in the optimization problem into implicit regularization. Their scale and ratio can also be adjusted appropriately to optimize the main objective of RNNs, respectively. In an imitation learning scenario with a mobile manipulator, only VAND is able to imitate sequential and periodic behaviors as instructed.

  • Active Inference with Dynamic Planning and Information Gain in Continuous Space by Inferring Low-Dimensional Latent States

    Matsumoto T., Fujii K., Murata S., Tani J.

    Entropy 27 ( 8 )  2025.08

    　View Summary

    Active inference offers a unified framework in which agents can exhibit both goal-directed and epistemic behaviors. However, implementing policy search in high-dimensional continuous action spaces presents challenges in terms of scalability and stability. Our previously proposed model, T-GLean, addressed this issue by enabling efficient goal-directed planning through low-dimensional latent space search, further reduced by conditioning on prior habituated behavior. However, the lack of an epistemic term in minimizing expected free energy limited the agent’s ability to engage in information-seeking behavior that can be critical for attaining preferred outcomes. In this study, we present EFE-GLean, an extended version of T-GLean that overcomes this limitation by integrating epistemic value into the planning process. EFE-GLean generates goal-directed policies by inferring low-dimensional future posterior trajectories while maximizing expected information gain. Simulation experiments using an extended T-maze task—implemented in both discrete and continuous domains—demonstrate that the agent can successfully achieve its goals by exploiting hidden environmental information. Furthermore, we show that the agent is capable of adapting to abrupt environmental changes by dynamically revising plans through simultaneous minimization of past variational free energy and future expected free energy. Finally, analytical evaluations detail the underlying mechanisms and computational properties of the model.

  • Selection of Exploratory or Goal-Directed Behavior by a Physical Robot Implementing Deep Active Inference

    Ko Igari, Kentaro Fujii, Gabriel W. Haddon-Hill, Shingo Murata

    Communications in Computer and Information Science (Springer Nature Switzerland)  2193 CCIS   165 - 178 2024.12

    Last author, Corresponding author, Accepted,  ISSN  1865-0929

    　View Summary

    Intelligent robots are being developed with the expectation that they will perform various tasks in diverse environments. Such robots need to autonomously engage in both exploratory behavior to reduce environmental uncertainty and goal-directed behavior to achieve their preferred observations (or goals). In this study, we focus on active inference, which provides a unified scheme for these distinct behavioral modes. Policy selection in active inference is based on minimizing expected free energy (EFE), which consists of one term representing epistemic value and another representing extrinsic value. Specifically, we investigate the influence of preference precision, which controls the balance between these two terms, on policy selection by a physical robot receiving high-dimensional and uncertain observations. We developed a deep active inference framework comprising a world model and a policy suggester. The world model predicts future hidden states and observations based on candidate policies from the policy suggester. The EFE for each policy is approximated using the predicted future hidden states and observations as well as the preferred observation. We implemented our proposed framework in a robot, requiring it to select a policy that minimizes EFE and then generate actions accordingly. The experimental results showed that the robot implementing the proposed framework selected exploratory or goal-directed behavior depending on the level of preference precision. These findings suggest that adjusting preference precision plays a crucial role in the autonomous selection of exploratory or goal-directed behavior in real-world situations with potential uncertainty.

display all >>

Page Top ▲

Papers, etc., Registered in KOARA 【 Display / hide 】

Page Top ▲

Reviews, Commentaries, etc. 【 Display / hide 】

Page Top ▲

Presentations 【 Display / hide 】

  • Deep Active Inference with Reconstructive and Contrastive Learning

    Kentaro Fujii, Takuya Isomura, Shingo Murata

    The 5th International Workshop on Active Inference, 

    2024.09
    , 

    Poster presentation

  • Real-World Robot Control Based on Contrastive Active Inference with Learning from Demonstration

    Kentaro Fujii, Takuya Isomura, Shingo Murata

    The 4th International Workshop on Active Inference, 

    2023.09
    , 

    Poster presentation

  • A Deep Generative Model for Extracting Shared and Private Latent Representations from Multimodal Data

    Kaito Kusumoto, Shingo Murata

    International Symposium on Predictive Brain and Cognitive Feelings,, 

    2023.07
    , 

    Poster presentation

  • Multiple Timescale Recurrent State-Space Model for Learning Long-Horizon Tasks

    Kentaro Fujii, Shingo Murata

    International Symposium on Predictive Brain and Cognitive Feelings, 

    2023.07
    , 

    Poster presentation

  • Action Modification Based on Real-time Amortized Inference of Others’ Intentions Using Backward RNN

    Yukiko Orui, Shingo Murata

    The 54th ISCIE International Symposium on Stochastic Systems Theory and Its Applications (SSS '22), 

    2022.10
    , 

    Oral presentation (general)

display all >>

Page Top ▲

Research Projects of Competitive Funds, etc. 【 Display / hide 】

  • Computational Nosology of Psychiatric Disorders: Towards a Next-Generation Disease Concepts via Integration of Dimensional and Categorical Perspectives

    2025.04
    -
    2030.03

    Japan Society for the Promotion of Science, Grants-in-Aid for Scientific Research, Grant-in-Aid for Scientific Research (A), No Setting

    　View Summary

    現行の精神障害診断カテゴリーは生物学的・臨床的妥当性が不十分であり、新たな疾病概念の確立が求められている。そこで、本研究は機械学習（ML）・人工知能（AI）技術を基盤とする計算論的精神医学に基づき、精神障害の背後にある連続的・離散的潜在状態を捉え、それを次元的・カテゴリー的臨床判断に統合する新たな疾病概念（「計算論的診断学」）を構築する。具体的には、階層的・動的、かつ連続的・離散的性質を併せ持つ潜在状態抽出を可能にする計算モデルの開発と、抽出した潜在状態を次元的・カテゴリー的性質を併せ持つ臨床判断に融合させる計算論的手法を確立し、WEB実験により疾病概念の妥当性と臨床的有用性を検証する。

  • Robotic Interaction Framework Based on Free-Energy Principle

    2024.04
    -
    2027.03

    日本学術振興会, Grants-in-Aid for Scientific Research, Grant-in-Aid for Scientific Research (B), Principal investigator

    　View Summary

    人とロボットの長期にわたる持続的なインタラクションの実現が期待される．本研究は，その実現のための一つのアプローチとして，人同士のインタラクションを支える脳内情報処理の「計算原理」を理解し，その原理に基づいてロボットの知的情報処理機構を開発する．具体的には，理論神経科学の分野で提唱されている自由エネルギー原理（Free-Energy Principle: FEP）に注目する．理論的枠組みであるFEPに対して，深層学習分野で培われてきた実世界応用可能なアルゴリズムを適用することでロボットの知的情報処理機構を構築し，人―ロボット間インタラクション実験を通じて，開発技術を検証する．
    本研究課題では，人同士のインタラクションを支える脳内情報処理の「計算原理」を理解し，その原理に基づいてロボットの知的情報処理機構を開発することを目的とする．具体的には，理論神経科学の分野で提唱されている自由エネルギー原理（Free-Energy Principle: FEP）に注目する．理論的枠組みであるFEPに対して，深層学習分野で培われてきた実世界応用可能なアルゴリズムを適用することでロボットの知的情報処理機構を構築し，人―ロボット間インタラクション実験を通じて，開発技術を検証する．
    令和6年度は，FEPを実世界環境でのインタラクションにスケールさせるためのフレームワークの構築を行なった．具体的には，高次元の観測情報時系列から潜在的なダイナミクスを抽出するための世界モデル，観測情報に応じたロボットの行動系列候補を生成可能な方策モデルを実装し，数値実験及びロボット実験による評価を行なった．世界モデルについては，時系列データの処理機構としてRNNを用いたモデルとTransformerを用いたモデルを実装した．また，方策モデルについては，RNNを用いたモデルと拡散モデルを用いたモデルを実装した．
    数値実験及び実ロボット実験によって，世界モデルの自己回帰的な観測予測能力，方策モデルの観測に応じた行動系列候補の生成能力を確認した．これらの成果は，人とインタラクションしながらロボットが期待自由エネルギーを最小化する行動を選択するための基盤となるものであり，本研究の目的に向けた重要な進展である．
    現在までに，自由エネルギー原理（FEP）に基づくロボットの知的情報処理機構の構築に向け，世界モデルおよび方策モデルの実装と評価を進めた．これらの成果の一部は，査読付き国際会議において発表を行い，対外的な評価も得ている．加えて，人とロボットの物理的なインタラクションを実現するために，ロボットのインピーダンス制御の環境整備も進めており，実験系のさらなる拡充に取り組んでいる．これらの取り組みにより，実環境下での人ーロボットインタラクションの実現に向けた基盤を着実に構築しつつある．
    今後の研究では令和6年度に構築した計算フレームワークを利用して，人とロボット間のインタラクション実験を実施していく．具体的には，人の動作や力の変化に応じてロボットが適応的に行動を選択する状況を設定し，期待自由エネルギーの最小化に基づく行動選択メカニズムの有効性を検証する．また，インピーダンス制御による柔軟な力のやり取りを活用し，物理的インタラクションにおける双方向の適応過程を明らかにすることを目指す．これにより，FEPに基づくロボットの意思決定機構が実環境下で人との協調的行動にどのように寄与するかを体系的に評価する．

  • 脳の計算原理とプレイデータに基づく実世界ロボット学習

    2022.10
    -
    2026.03

    科学技術振興機構, 戦略的創造研究推進事業（さきがけ）, Principal investigator

  • Understanding hierarchical statistical learning of sequential perception with dynamic multiple timescale models

    2021.10
    -
    2025.03

    日本学術振興会, Grants-in-Aid for Scientific Research, Fund for the Promotion of Joint International Research (Fostering Joint International Research (B)), Coinvestigator(s)

  • Computational nosology based on artificial intelligence and dimensional approach for psychiatric disorders

    2020.04
    -
    2025.03

    National Center of Neurology and Psychiatry, Grants-in-Aid for Scientific Research Grant-in-Aid for Scientific Research (A), Yamashita Yuichi, Grant-in-Aid for Scientific Research (A), Coinvestigator(s)

    　View Summary

    We applied computational psychiatry methodologies -including machine learning and artificial intelligence techniques- to large-scale datasets encompassing psychiatric symptoms, neurophysiology, and cognitive behaviors. Using transdiagnostic and dimensional approaches, our research aimed to identify novel end-phenotypes beyond existing psychiatric diagnostic categories. Specifically, employing large-scale web-based experimental frameworks, I explored symptom-based subtypes (symptomatological phenotypes) and computational phenotypes that transcend traditional disorder classifications. Additionally, we utilized transdiagnostic neuroimaging databases to identify neurophysiological psychiatric subtypes (biotypes). These efforts collectively sought to establish new conceptual frameworks for understanding psychiatric disorders.

display all >>

Page Top ▲

Awards 【 Display / hide 】

  • Advanced Robotics Best Survey Paper Award

    Tadahiro Taniguchi, Shingo Murata, Masahiro Suzuki, Dimitri Ognibene, Pablo Lanillos, Emre Ugur, Lorenzo Jamone, Tomoaki Nakamura, Alejandra Ciria, Bruno Lara, Giovanni Pezzulo, 2024.09, The Robotics Society of Japan

  • JSAI Annual Conference Award

    2020.07, The Japanese Society for Artificial Intelligence (JSAI 2020)

    Type of Award： Award from Japanese society, conference, symposium, etc.

  • Best Paper Award

    Tatsuro Yamada, Shingo Murata, Hiroaki Arie, Tetsuya Ogata, 2016.09, The 25th International Conference on Artificial Neural Networks (ICANN 2016)

    Type of Award： Award from international society, conference, symposium, etc.

  • Best Paper Award for Young Researcher of IPSJ National Convention

    2014.03, Information Processing Society of Japan (IPSJ 2014)

    Type of Award： Award from Japanese society, conference, symposium, etc.

Page Top ▲
 

Courses Taught 【 Display / hide 】

  • INTRODUCTION TO MACHINE LEARNING

    2026

  • DOCTORAL RESEARCH ON ENGINEERING AND DESIGN

    2026

  • LABORATORY IN SCIENCE

    2026

  • GRADUATE RESEARCH ON ENGINEERING AND DESIGN 2

    2026

  • INDEPENDENT STUDY ON INTEGRATED DESIGN ENGINEERING

    2026

display all >>

Page Top ▲
 

Memberships in Academic Societies 【 Display / hide 】

  • IEEE, 

    2016
    -
    Present

  • The Japanese Society for Artificial Intelligence, 

    2014
    -
    Present

  • The Robotics Society of Japan, 

    2012
    -
    Present
Page Top ▲

Committee Experiences 【 Display / hide 】

  • 2024
    -
    Present

    Section Editor, Special Section on "Cognitive Development and Symbol Emergence" in Advanced Robotics

  • 2024
    -
    Present

    Special Section on "Cognitive Development and Symbol Emergence" in Advanced Robotics

  • 2021.11
    -
    Present

    Associate Editor, IEEE Transactions on Cognitive and Developmental Systems

  • 2021

    Associate Editor, IEEE/RSJ IROS 2021

  • 2021

    Associate Editor, IEEE ICDL 2021

display all >>

Page Top ▲