Xiaoqian Liang

PhD student @Westlake_Uni

[Xiaoqian Liang @yubel426]

@BaadeAlan @drfeifei Great work! Is there a particular reason for training the prediction model and the representation model jointly, rather than first training the prediction model to convergence and then freezing it to train the representation model?

[Xiaoqian Liang @yubel426]

Also worth checking out:
https://t.co/GarUR8ppIu… by @CongrongX

https://t.co/GqxRZsuG73 by @zzigakovacic

https://t.co/utKEuTeY2m by @zgq1879

[Anpei Chen @AnpeiC]

1/5 Excited to share an early exploration.

Coding agents like Codex can already tackle much of real-to-sim—even for robot manipulation. Does that mean 3D vision is solved?

Not quite. 3D still matters.

https://t.co/1CPFCMFuhZ

[Weihan Luo @luo_weihan]

Lightweight harness + 3D tools helps coding agents do real-to-sim from video. Cool work by @AnpeiC @yubel426

[Xiaoqian Liang @yubel426]

@michaelgold @AnpeiC @Hyuan_Chen @binbinhuang00 You are too fast! It's public now.

[Anpei Chen @AnpeiC]

1/5 Excited to share an early exploration.

Coding agents like Codex can already tackle much of real-to-sim—even for robot manipulation. Does that mean 3D vision is solved?

Not quite. 3D still matters.

https://t.co/1CPFCMFuhZ