Yihe Tang

Robot Learning @USC | Prev @Stanford @UCSD

[Yue Wang @yuewang314]

Introducing Ψ₀ (https://t.co/qqH1PiIJS8) — an open foundation model for universal humanoid loco-manipulation.

🏆 Outperforms GR00T N1.6 by 40%+ overall success rate
📉 Uses only ~10% of the pre-training data
📦 Fully open-source: model, data, code, and deployment pipeline

1/10

[Chengshu Li @ChengshuEricLi]

We are excited to release MoMaGen, a data generation method for multi-step bimanual mobile manipulation.
MoMaGen turns 1 human-teleoped robot trajectory into 1000s of generated trajectories automatically.🚀

Website: https://t.co/DYKvqY4bII
arXiv: https://t.co/lDffi0FXHl

[Yihe Tang @Yihe_yihe]

@YunfanJiang @wenlong_huang Thank you Yunfan!

[Yihe Tang @Yihe_yihe]

@neil_nie_ @wenlong_huang Many thanks Neil! Can't wait to see your next work!

[Yihe Tang @Yihe_yihe]

@yuewang314 @CSatUSC Thank you Yue!

[Yihe Tang @Yihe_yihe]

Excited to share our work Unsupervised Affordance Distillation (UAD) on visual affordance learning! We'll be presenting the work at ICRA 2025 on May 20th, TuDT1 Session, Room 302. Feel free to stop by and chat!

[Wenlong Huang @wenlong_huang]

How to scale visual affordance learning that is fine-grained, task-conditioned, works in-the-wild, in dynamic envs?

Introducing Unsupervised Affordance Distillation (UAD): distills affordances from off-the-shelf foundation models, *all without manual labels*.

Very excited this is nominated as Best Paper Finalist at #ICRA2025!

https://t.co/OOEi6lX2vM

🧵👇