Introduction

We aim to integrate perspectives from cognitive neuroscience, robotics, and machine learning to achieve (1) a constructive understanding of the computational mechanisms underlying human cognitive functions, and (2) the realization of intelligent robots capable of cooperation with others based on this understanding. Furthermore, we also aim (3) to understand the computational mechanisms underlying psychiatric disorders such as autism spectrum disorder and schizophrenia.

■Cognitive Robotics
・Research Objective: Proposal of a real-world robot control framework based on contrastive deep active inference with demonstration data
- Title: Real-World Robot Control Based on Contrastive Deep Active Inference with Demonstrations, Fujii Kentaro, Isomura Takuya, Shingo Murata (IEEE Access) 12, 172343–172357, Oct. 2024
・Research Objective: Proposal of an active vision framework for real robots based on the free energy principle
- Title: Active Vision for Physical Robots Using the Free Energy Principle, Haddon-Hill Gabriel W., Murata Shingo (International Conference on Artificial Neural Networks (ICANN) 2024) 270-284, Sep. 2024
・Research Objective: Proposal of a deep active inference framework for real robots capable of selecting exploratory or goal-directed behavior
- Title: Selection of Exploratory or Goal-Directed Behavior by a Physical Robot Implementing Deep Active Inference, Igari Ko, Haddon-Hill Gabriel W., Fujii Kentaro, Shingo Murata (The 5th International Workshop on Active Inference (IWAI 2024)) 165-178, Sep. 2024
・Research Objective: Proposal of a hierarchical latent dynamics model with multiple timescales for achieving long-horizon tasks
- Title: Hierarchical Latent Dynamics Model with Multiple Timescales for Learning Long-Horizon Tasks, Fujii Kentaro, Murata Shingo (2023 IEEE International Conference on Development and Learning (ICDL)) 479-485, Nov. 2023

■Robot Learning
・Research Objective: Proposal of a real-world robot control and data augmentation framework based on world model learning from play data
- Title: Real-World Robot Control and Data Augmentation by World-Model Learning from Play, Nomura Yuta, Murata Shingo (2023 IEEE International Conference on Development and Learning (ICDL)) 133-138, Nov. 2023
・Research Objective: Proposal of a self-supervised learning framework from play data for flexible object manipulation in real-world environments
- Title: Goal-Conditioned Flexible Object Manipulation by Self-Supervised Learning from Play, Ishii Keigo, Hiramatsu Shun, Nomura Yuta, Murata Shingo (2023 IEEE International Conference on Development and Learning (ICDL)) 150-155, Nov. 2023
・Research Objective: Proposal of a deep predictive learning framework for collaborative robots capable of task goal inference and dynamic optimization
- Title: Deep Predictive Network for Inference and Dynamic Optimization of Task Goals during Human-Robot Collaboration, Hiramatsu Shun, Murata Shingo (2023 International Joint Conference on Neural Networks (IJCNN)) 1-6, Jun. 2023
・Research Objective: Proposal of a deep predictive learning model capable of proposing subgoals for long-horizon object manipulation
- Title: Structured Motion Generation with Predictive Learning: Proposing Subgoal for Long-Horizon Manipulation, Saito Namiko, Moura João, Ogata Tetsuya, Aoyama Marina Y., Murata Shingo, Sugano Shigeki, Vijayakumar Sethu (2023 IEEE International Conference on Robotics and Automation (ICRA)) 9566-9572, May 2023
・Research Objective: Proposal of a variational recurrent neural network using reservoir computing for analyzing physical human-robot interactions in latent state space
- Title: Latent Representation in Human–Robot Interaction With Explicit Consideration of Periodic Dynamics, Kobayashi Taisuke, Murata Shingo, Inamura Tetsunari (IEEE Transactions on Human-Machine Systems) 52(5) 928-940, Oct. 2022
・Research Objective: Proposal of a tool-use model considering the relationship among tools, objects, actions, and effects using multimodal deep neural networks
- Title: Tool-Use Model to Reproduce the Goal Situations Considering Relationship Among Tools, Objects, Actions and Effects Using Multimodal Deep Neural Networks, Saito Namiko, Ogata Tetsuya, Mori Hiroki, Murata Shingo, Sugano Shigeki (Frontiers in Robotics and AI) 8, Sep. 2021

■Computational Psychiatry
・Research Objective: Proposal of a deep generative model for extracting shared and private latent representations from multimodal data
- Title: Toward Understanding Psychiatric and Cognitive Characteristics: A Deep Generative Model for Extracting Shared and Private Representations and Its Evaluation with Synthetic Multimodal Data, Kusumoto Kaito, Murata Shingo (2023 IEEE International Conference on Development and Learning (ICDL)) 455-460, Nov. 2023
・Research Objective: Proposal of a neural network model based on the predictive processing framework to understand altered facial expression recognition in autism spectrum disorders
- Title: Neural network modeling of altered facial expression recognition in autism spectrum disorders based on predictive processing framework, Takahashi Yuta, Murata Shingo, Idei Hayato, Tomita Hiroaki, Yamashita Yuichi (Scientific Reports) 11(1), Jul. 2021
・Research Objective: Proposal of a neurorobotics model capable of explaining the relationship among developmental functional disconnection, abnormal sensory precision, and abnormal sensory reactivity in neurodevelopmental disorders
- Title: Paradoxical sensory reactivity induced by functional disconnection in a robot model of neurodevelopmental disorder, Idei Hayato, Murata Shingo, Yamashita Yuichi, Ogata Tetsuya (Neural Networks) 138 150-940, Jun. 2021

Areas of Research

・Intelligent Informatics
・Intelligent Robotics

Social Contributions

・Enhancing industrial and daily life support through robotics
・Realizing natural human-robot collaboration
・Developing understanding and support technologies for psychiatric and developmental disorders