Hongze Fu

Umich MS ECE robotic student.

[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 🧵

[Yuejiang Liu @liu_yuejiang]

Robotic foundation models need memory, but existing methods still fall short

We’ll present RoboMME, a challenge organized w/ @YinpeiD, @jed_yang, @chelseabfinn, @NimaFazeli7, @SLED_AI at #CVPR2026 — a preview of our ICML Oral (0.7%) 🔥

Join us tmr (June 4), Room 703, 3:45 PM

[Hongze Fu @HongzeF14218]

@ziqiao_ma Congratulations Martin! Following your work at @thinkymachines🥳

[Hongze Fu @HongzeF14218]

@DJiafei @UW @uwcse @fox_dieter17849 @RanjayKrishna @NUSComputing Congratulations, Jiafei!

[Yinpei Dai @YinpeiD]

🚀 RoboMME is accepted to ICML 2026 (Spotlight)! 🎉

Excited to push the frontier of memory-augmented robot manipulation 🤖🧠

We’re also hosting the challenge @ CVPR 2026 (FMEA workshop) - join us to build next-gen memory-based robotic policies! 💪🔥

🔗 https://t.co/79FpVd1ti3

[Yinpei Dai @YinpeiD]

🚀 The RoboMME Challenge @ CVPR 2026 is now LIVE!

Timeline:
• May 15 — Policy submission
• June 3 — Winner announcement

🏆 Top 3 teams will be awarded $500/300/200

Let’s push the frontier for memory-augmented robotic manipulation together 💪

🔗https://t.co/79FpVd1ti3

[Hongze Fu @HongzeF14218]

@hirasava_ui Thanks so much! MIKASA-Robo was a great inspiration when we were building RoboMME — really glad to see more work pushing on memory in robot learning. Excited to see where you take the VLA direction!

[Hongze Fu @HongzeF14218]

@JieWang_ZJUI Thanks for supporting our work!

[Hongze Fu @HongzeF14218]

@DJiafei Thanks @DJiafei for highlighting our work! We implemented SAM2Act+ for RoboMME. Code is coming soon! https://t.co/4Wl1zYuEBs