Wentao He

MS @UMich · Robot learning: world models, generative models & embodied AI · Prev. SJTU · Looking for Research Engineer roles

[Wentao He @wthe2004]

Excited to share our new work on calibrating world models to instance-specific dynamics! 🌍 Very happy to have co-led this with @yluo_y 🎉 Huge thanks to Joyce Chai, @NimaFazeli7, @jjpark3D and all our amazing coauthors. Details in the thread 👇

[Yunhao Luo @yluo_y]

🌏 World models need to capture a scene’s hidden physics properties, e.g., mass or friction, to generate predictions consistent with physical reality.

🚀 Excited to introduce PhysiCalWorld: In-Context Physics Calibration for World Models! We calibrate world models through an interaction clip, grounding future predictions in scene-specific dynamics.

🏷️ Supervision: trained without physics labels
⚡ Efficiency: minimal additional inference cost
⚛️ Generalization: extrapolation beyond the physics ranges seen during training

https://t.co/k9PRgg7qmR

More below 🧵