Kevin Huang

prof_pic.jpg

Hi! I am a fourth-year PhD student at the University of Washington, working with Byron Boots and Abhishek Gupta, researching robotics and AI. Broadly, I am interested in developing end-to-end intelligent, embodied systems that continuously learn from their environment, as well as robust control and planning in uncertain and dynamic environments. My current research interests include policy improvement, world models, and leveraging large scale data in simulation and closing the sim2real gap.

Previously, I completed my undergrad at Caltech and am fortunate to have worked with Yisong Yue and Anima Anandkumar at Caltech, and Pablo Moscato at the University of Newcastle.

selected publications

  1. simdist.gif
    Simulation Distillation: Pretraining World Models in Simulation for Rapid Real-World Adaptation
    Jacob Levy*, Tyler Westenbroek*, Kevin Huang, and 6 more authors
    In , 2026
    arXiv Code Website
    3 Google Scholar citations
  2. mpail2-vis.gif
    Online World Modeling Enables Real-World Inverse Reinforcement Learning from Observation
    Tyler Han, Bat Nemekhbold, Siyang Shen, and 6 more authors
    In , 2026
    arXiv Code Website
    0 Google Scholar citations
  3. lampshade_stitching.gif
    Using Non-Expert Data to Robustify Imitation Learning via Offline Reinforcement Learning
    Kevin Huang*, Rosario Scalise*, Cleah Winston, and 6 more authors
    In IEEE International Conference on Robotics (ICRA), 2026
    arXiv Code Website
    2 Google Scholar citations
  4. sgft.gif
    Rapidly Adapting Policies to the Real World via Simulation-Guided Fine-Tuning
    Patrick Yin*, Tyler Westenbroek*, Simran Bagaria, and 4 more authors
    In International Conference on Learning Representations (ICLR), 2025
    arXiv Code Website
    26 Google Scholar citations
  5. sim2real.png
    Overcoming the Sim-to-Real Gap: Leveraging Simulation to Learn to Explore for Real-World RL
    Andrew Wagenmaker, Kevin Huang, Liyiming Ke, and 3 more authors
    In Advances in Neural Information Processing Systems, 2024
    arXiv
    49 Google Scholar citations
  6. datt.gif
    DATT: Deep adaptive trajectory tracking for quadrotor control
    Kevin Huang, Rwik Rana, Alex Spitzer, and 2 more authors
    In Conference on Robot Learning, 2023
    arXiv Code Website
    66 Google Scholar citations
  7. dppo.png
    Deep Model Predictive Optimization
    Jake Sacks, Rwik Rana, Kevin Huang, and 3 more authors
    In IEEE International Conference on Robotics and Automation (ICRA), 2023
    arXiv
    25 Google Scholar citations
  8. grad.png
    Automatic Gradient Descent: Deep Learning without Hyperparameters
    Jermey Bernstein*, Chris Mingard*, Kevin Huang, and 2 more authors
    arXiv preprint arXiv:2304.05187, 2023
    arXiv
    23 Google Scholar citations
  9. cem-gd.png
    CEM-GD: Cross Entropy Method with Gradient Descent for Model Based Reinforcement Learning
    Kevin Huang, Sahin Lale, Rosalia Ugo, and 2 more authors
    arXiv preprint arXiv:2112.07746, 2021
    arXiv Blog
    10 Google Scholar citations