Puyuan Peng
Papers
1
Total Citations
5
H-Index
1
About
Puyuan Peng is a researcher advancing the intersection of speech processing, audio-visual learning, and robotics. His work focuses on enabling machines to understand complex human environments by integrating multimodal sensory data—particularly speech and visual cues—to drive autonomous action. In his most-cited paper, "Style-transfer based Speech and Audio-visual Scene understanding for Robot Action Sequence Acquisition from Videos" (2023), Peng introduces a novel framework that leverages style transfer techniques to parse audio-visual scenes from video demonstrations, allowing robots to learn action sequences without explicit programming. This contribution addresses a critical bottleneck in robot learning: the ability to generalize from noisy, real-world demonstrations. While his citation count is still building, the work signals a promising direction in embodied AI, where robots acquire skills through passive observation of human activities. Peng’s research is notable for its interdisciplinary approach, blending signal processing, computer vision, and reinforcement learning. As the field moves toward more intuitive human-robot interaction, his methods offer a pathway for robots to learn from natural, multimodal instruction—a key step toward autonomous agents that can operate in unstructured environments.
Research Focus
Key Achievements
Top Papers
- 1