Zhixuan Liang

Physical robot learning can now autonomously discover and exploit informative experiments to continuously improve foundation models, achieving 77% success on RoboCasa365 and zero-shot transfer to real hardware.

Experimental results demonstrate the proposed $\mathcal{M}^2$Tok-based VLA achieves superior reconstruction fidelity but also significantly boosts the success rate of VLA models, and comprehensive ablation studies further confirm the effectiveness of the multi-head and multi-codebook mechanisms.