Sangwoo Shin

SangwooShin.jpeg

Hello! I am a second-year Ph.D. student in Computer Sciences at the University of Wisconsin–Madison, advised by Professor Josiah Hanna. Previously, I received my M.S. and B.S. degrees from Sungkyunkwan University, where I was a member of the CSI-Agent Group advised by Professor Honguk Woo.

My research focuses on robot learning that works reliably in the real world. I work on reinforcement learning and sim-to-real transfer for legged and humanoid robots, and more recently on learning-based manipulation for automating routine tasks in science labs.

I also have a strong interest in pure mathematics, particularly Lie theory and smooth manifolds, which underlie much of the geometry of robot dynamics.

selected publications

  1. Preprint
    Informationally Decoupled Trajectory Design for Sim-to-Real System Identification
    Sangwoo Shin, Ashvin Anilkumar, Ryan Gao, and 1 more author
    arXiv preprint arXiv:2610.10905, 2026
  2. CoRL
    RoboFQE: Rollout-Free Evaluation of Visual Robot Policies from Offline Trajectories
    Yunfu Deng, Sangwoo Shin, Brahma Pavse, and 1 more author
    In Conference on Robot Learning (CoRL), 2026
  3. IROS
    Articulated-Body Dynamics Network: Dynamics-Grounded Prior for Robot Learning
    Sangwoo Shin, Kunzhao Ren, Xiaobin Xiong, and 1 more author
    In IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2026