Mengzhen Liu
Papers
4
Total Citations
26
H-Index
3
About
Mengzhen Liu is a rising star in embodied AI and robot manipulation, whose work centers on bridging the gap between high-level reasoning and low-level physical action. Her major contributions include co-founding the **RoboMIND** benchmark—a massive dataset of 107k demonstration trajectories across 479 tasks and 96 object classes, collected via human teleoperation to standardize multi-embodiment robot learning. This resource has already garnered **14 citations** in its first year, signaling its foundational role in the field. Liu also developed **RoboMamba**, an efficient Vision-Language-Action (VLA) model that tackles a core challenge: enabling robots to both comprehend visual scenes and execute complex actions without excessive computational overhead. Her work directly addresses the insufficient reasoning ability of prior models, pushing toward more intelligent, real-world robotic systems. With over **26 combined citations** across her early publications, Liu’s research is shaping how robots learn from human demonstration and generalize across tasks. Her achievements mark her as a key contributor to the next generation of autonomous manipulation.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4