DeepCybo
“The core achievement is a pipeline that converts synchronized ego-exo human video into controller-aligned 60-DoF humanoid joint action chunks — not just visual features or hand trajectories, but the same joint-space commands the robot executes at deployment.”
Source→“Hardware developer of the PrimeU humanoid platform. Acknowledged for 'substantial support in the development and integration of the humanoid platform' (Acknowledgments).”
Source→“The core achievement is a pipeline that converts synchronized ego-exo human video into controller-aligned 60-DoF humanoid joint action chunks — not just visual features or hand trajectories, but the same joint-space commands the robot executes at deployment.”
Source→“Xiaopeng Lin — HKUST (Guangzhou) / DeepCybo. Co-first author. Corresponding work on the human-to-humanoid pipeline.”
Source→“Kai Chen — DeepCybo / ZGCA / ZGCI. Corresponding author. Notable as the senior figure across multiple institutions, suggesting a coordinated effort bridging academia and industry for humanoid data scaling.”
Source→AI-extracted from podcast / newsletter / paper summaries. May contain errors.