About Me

Hi! I am a Ph.D. candidate in Computer Science at Vanderbilt University, advised by Prof. Thomas Beckers, and I expect to graduate in December 2027.

I build generative world models for physical AI: models that learn how physical systems move from only a handful of sparse observations, respect the underlying physics, and know when they are uncertain. My work combines diffusion models, Gaussian processes, and conformal prediction with port-Hamiltonian structure, and puts the resulting models to work in forecasting, model predictive control, and robot planning.

Before Vanderbilt, I received my M.S. in Electrical Engineering from Washington University in St. Louis (2023), where I worked with Prof. Yiannis Kantaros on the robustness of trajectory predictors and on LLM-based robot planning. I received my B.Eng. in Information Engineering from Sun Yat-sen University (2021), where I was one of 18 students selected for its inaugural Information Engineering Experimental Program.

I am open to research internships in physical AI, generative modeling, and robotics. My resume is available here; feel free to reach out by email or LinkedIn.

Research Interests

  • Physical AI: world models of physical dynamics, physics-informed learning, learning-based control
  • Generative Models: diffusion models for trajectory forecasting from sparse observations
  • Uncertainty Quantification: conformal prediction, out-of-distribution detection, calibrated forecasts
  • Bayesian Methods: Gaussian processes, Bayesian learning of port-Hamiltonian systems
  • Robotics: safe planning and control, LLM task planning, reinforcement learning

Publications

Physical AI & Generative Models

PHDME
PHDMEGIF coming soon
Kaiyuan Tan, Kendra Givens, Peilun Li, Thomas Beckers
Preprint, under review at TMLR, 2026
CIPHER
CIPHERGIF coming soon
Peilun Li, Kaiyuan Tan, Daniel Moyer, Thomas Beckers
Preprint, under review at ICLR 2027
NAPI-MPC
NAPI-MPCGIF coming soon
Peilun Li, Kaiyuan Tan, Thomas Beckers
Learning for Dynamics and Control Conference (L4DC 2025)

Bayesian Methods & Uncertainty Quantification

PnP-PIML
PnP-PIMLGIF coming soon
Kaiyuan Tan, Peilun Li, Jun Wang, Thomas Beckers
IEEE International Conference on Robotics and Automation (ICRA 2025)
GP-dPHS
GP-dPHSGIF coming soon
Kaiyuan Tan, Peilun Li, Thomas Beckers
Learning for Dynamics and Control Conference (L4DC 2024)
PyGpPHs
PyGpPHsGIF coming soon
Peilun Li, Kaiyuan Tan, Thomas Beckers
IFAC Workshop on Lagrangian and Hamiltonian Methods for Nonlinear Control (LHMNC 2024)

Robotics: Planning, Learning & Safety

Conformal temporal logic planning
HERACLEsGIF coming soon
Jun Wang, Jiaming Tong, Kaiyuan Tan, Yevgeniy Vorobeychik, Yiannis Kantaros
ACM Transactions on Cyber-Physical Systems (TCPS 2025)
Mission-driven exploration
Mission-driven RLGIF coming soon
Jun Wang, Hosein Hasanbeig, Kaiyuan Tan, Zihe Sun, Yiannis Kantaros
Learning for Dynamics and Control Conference (L4DC 2025)
Targeted adversarial attacks
TA4TPGIF coming soon
Kaiyuan Tan, Jun Wang, Yiannis Kantaros
Learning for Dynamics and Control Conference (L4DC 2023)

News

  • 05/2026 Our preprint CIPHER on contrastive learning of latent port-Hamiltonian dynamics is on arXiv.
  • 01/2026 Our preprint PHDME on physics-informed diffusion models is on arXiv.
  • 09/2025 Conformal Temporal Logic Planning using LLMs was accepted to ACM TCPS.
  • 05/2025 I presented PnP-PIML at ICRA 2025 in Atlanta.
  • 02/2025 Two papers, NAPI-MPC and Mission-driven Exploration, were accepted to L4DC 2025.
  • 01/2025 PnP-PIML was accepted to ICRA 2025.
  • 07/2024 I gave two talks at the 16th World Congress on Computational Mechanics in Vancouver.
  • 07/2024 We presented our work on physics-constrained learning of PDE systems at L4DC 2024 in Oxford.
  • 06/2024 We presented PyGpPHs at the 8th IFAC LHMNC workshop in Besançon.
  • 12/2023 I started my Ph.D. with Prof. Beckers at Vanderbilt University.
  • 03/2023 Our paper on targeted adversarial attacks against trajectory predictors was accepted to L4DC 2023.

Teaching

Teaching Assistant, Vanderbilt University:

  • CS 4275/5275: Error Analysis in Safety-Critical Systems, Spring 2025
  • CS 3891: Autonomous Driving, Fall 2024
  • CS 4275/5275: Error Analysis in Safety-Critical Systems, Spring 2024

Honors & Awards

  • Selected as one of 18 students for the inaugural Information Engineering Experimental Program, Sun Yat-sen University, 2019
Visitor map