Jaeyong Sung
Papers
8
Total Citations
1,103
H-Index
7
About
Jaeyong Sung is a leading researcher in robotics and artificial intelligence, with a focus on enabling robots to perceive, understand, and interact with unstructured human environments. His key research areas include human activity detection, natural language grounding for robot instruction, and manipulation of novel objects. Sung's pioneering work on unstructured human activity detection from RGBD images (528 citations) established a low-cost, reliable framework for recognizing complex human behaviors in home settings, a critical step toward personal assistive robotics. He further advanced human-robot interaction with the "Tell me Dave" project (194 citations), which developed context-sensitive methods for grounding natural language commands into actionable manipulation instructions, allowing robots to adapt tasks like "boil water" based on environmental context. His Robobarista system introduced deep multimodal embedding techniques to enable robots to manipulate novel objects by transferring manipulation trajectories from crowd-sourced data, bridging vision, language, and motion. Sung's contributions have been widely cited and recognized for their practical impact, with his work appearing in top venues such as AAAI and RSS. His research continues to push the boundaries of how robots can learn from and assist humans in dynamic, real-world settings.
Research Focus
Key Achievements
Top Papers
- 1Unstructured human activity detection from RGBD images528 citations · 2012
- 2Human Activity Detection from RGBD Images273 citations · 2011
- 3
- 4
- 5
- 6
- 7Learning to represent haptic feedback for partially-observable tasks7 citations · 2017
- 8