1. Home
  2. Research
  3. Centers & Labs
  4. TRIP Headquarters
  5. Data and Computational Sciences Integration Research Program (CoRe)

Data and Computational Sciences Integration Research Program Physical AI R&D Team

Team Director: Tatsuya Harada (D.Eng.)

Research Summary

Tatsuya Harada

The “Physical AI Research and Development Team” develops foundational technologies that enable AI to understand the real world through senses such as vision and touch, and to act flexibly in response to different situations. We will apply the results of projects such as those conducted under NEDO to a wide range of industrial sectors through joint research and demonstrations with companies, thereby accelerating the societal implementation of Physical AI.

Main Research Fields

  • Informatics

Related Research Fields

  • Engineering

Keywords

  • Physical AI
  • Multimodal Model
  • Machine Learning

Selected Publications

  • 1. Dexuan Zhang, Qianyu Zhou, Thomas Westfechtel, Yusuke Mukuta, Tatsuya Harada.
    "Bridging Theory and Practice in Source-Free Domain Adaptation via Adversarial Proxy Perturbation."
    The 19th European Conference on Computer Vision (ECCV), 2026.
  • 2. Ryo Umagami, Tomohiro Hashimoto, Xuangeng Chu, Ziteng Cui, Yusuke Mukuta, Jia Deng, Tatsuya Harada. TriSplat:
    "Adaptive Triplane for Sparse-View Large-Scale Scene Reconstruction."
    The 19th European Conference on Computer Vision (ECCV), 2026.
  • 3. Shuhong Liu, Xining Ge, Ziying Gu, Quanfeng Xu, Ziteng Cui, Lin Gu, Xuangeng Chu, Jun Liu, Dong Li, Tatsuya Harada.
    "Denoising the Deep Sky: Physics-Based CCD Noise Formation for Astronomical Imaging."
    The 19th European Conference on Computer Vision (ECCV), 2026.
  • 4. Kazuki Ota, Takayuki Osa, Motoki Omura, Tatsuya Harada.
    "Revisiting Regularized Policy Optimization for Stable and Efficient Reinforcement Learning in Two-Player Games."
    Forty-third International Conference on Machine Learning (ICML), 2026.
  • 5. Wenjun Huang, Ziteng Cui, Yinqiang Zheng, Yirui He, Tatsuya Harada, Mohsen Imani. Dr. RAW:
    "Towards General High-Level Vision from RAW with Efficient Task Conditioning."
    The Thirty-Ninth Annual Conference on Neural Information Processing Systems (NeurIPS), 2025.

Lab Members

Principal investigator

Tatsuya Harada
Team Director

Contact Information

Engineering Building #2
7-3-1, Hongo, Bunkyo-ku, Tokyo, 113-8656, Japan
Email: tatsuya.harada@riken.jp

Top