Papers
1
Total Citations
57
H-Index
1
About
Lisa Peng is a leading researcher in embodied AI and robot learning, with a focus on developing intelligent navigation systems that bridge perception and action. Her key contributions lie in leveraging structured representations—particularly 3D scene graphs and graph neural networks—to enable robots to learn effective, generalizable navigation policies. In her highly cited 2022 work, "Hierarchical Representations and Explicit Memory," Peng demonstrated how mid-level perceptual abstractions, such as depth estimates and semantic segmentation, can dramatically outperform raw sensor data like RGB images for policy learning. By integrating explicit memory and hierarchical scene understanding, her approach allows robots to navigate complex, unseen environments with greater efficiency and robustness. This paper has garnered 57 citations and is recognized as a foundational contribution to the field of learning-based navigation. Peng’s work is notable for its emphasis on interpretable, structured representations that reduce the need for massive training data while improving transfer across domains. Her research continues to shape how autonomous systems perceive and act in the physical world, making her a rising voice in the intersection of computer vision, robotics, and reinforcement learning.
Research Focus
Key Achievements
Top Papers
- 1