PhysiCalWorld: In-Context Physics Calibration for World Models
Calibrating world models to the physics of a new scene through in-context interaction, without fine-tuning or physics labels.
Email: wentaohe [at] umich [dot] edu
Master's studentUniversity of Michigan School of Information
I am a master's student at the University of Michigan School of Information, working with Professor Nima Fazeli and Professor JJ (Jeong Joon) Park.
Previously, I studied Electronic and Computer Engineering at Shanghai Jiao Tong University, where I worked with Professor Junguo Lu in the Machine Vision and Autonomous System Laboratory (MVASL).
Calibrating world models to the physics of a new scene through in-context interaction, without fine-tuning or physics labels.
A unified framework for motion retargeting and whole-body imitation learning across different humanoid robots.

An LLM-guided framework connecting reward design, reinforcement learning, and sim-to-real evaluation for bipedal robot locomotion.
