Publications

Peer-Reviewed Papers

2024

  1. Open X-Embodiment: Robotic Learning Datasets and RT-X Models
    Open X-Embodiment Collaboration (Google DeepMind and Open X-Embodiment Collaboration)
    ICRA 2024 Best Robotic Manipulation Paper - Finalist
    [Project Page] [Paper] [Code]

    Large-scale multi-robot dataset and transformer-based models for robotic learning across different embodiments.

2022

  1. Development of Amphibious Humanoid Platform for Sensor-based Behavior Acquisition of Whole-body Manipulation Tasks
    T. Makabe, J. Oh, et al.
    IEEE-RAS International Conference on Humanoid Robots (Humanoids), 2022

    Development of an amphibious humanoid robot platform capable of both terrestrial and aquatic manipulation tasks.

Conference Presentations

  1. Dynamic parameter error identification of throwing robot and optimization of throwing based on sensitivity analysis
    J. Oh, Advisor: Prof. [Name]
    Robomech 2022, Sapporo
    Bachelor Graduation Research Presentation

    Analysis of throwing robot dynamics with parameter identification and trajectory optimization based on sensitivity analysis.

Current Research

  1. 3D Affordance Learning from Human Video Data for Robotic Manipulation (2024-Present)
    J. Oh, Advisor: Prof. Kei Okada
    PhD Research Project

    Extracting 3D affordances from large-scale human video data using ego-centric structure-from-motion and 3D reconstruction techniques, with application to robotic object manipulation learning.

Thesis

  • Master’s Thesis: “3D Affordance Learning from Human Video Data for Robotic Manipulation”
    The University of Tokyo, 2025
    Advisor: Prof. Kei Okada

  • Bachelor’s Thesis: “Dynamic parameter error identification of throwing robot and optimization of throwing based on sensitivity analysis”
    Tokyo Institute of Technology, 2022
    Advisor: Prof. [Name]