Heecheol Kim
Papers
6
Total Citations
145
H-Index
5
About
Heecheol Kim is a robotics researcher specializing in deep imitation learning, dexterous robot manipulation, and human-guided robot learning. His work focuses on enabling robots to acquire complex manipulation skills from human demonstrations, with a particular emphasis on incorporating human perceptual mechanisms — especially gaze — to improve robot performance. Kim's most influential contributions include pioneering the use of human gaze signals to help robots filter task-irrelevant distractions during manipulation (28 citations) and developing gaze-based dual-resolution learning systems inspired by human foveal and peripheral vision for high-precision tasks like needle threading (29 citations). His goal-conditioned dual-action imitation learning framework (45 citations) tackles long-horizon dexterous manipulation of deformable objects such as banana peeling — a notoriously difficult challenge in robotics. He has also advanced master-to-robot policy transfer, enabling robot training without physical robots through force-feedback-aware imitation learning (26 citations). Across his career, Kim has accumulated over 145 citations, reflecting growing recognition of his interdisciplinary approach that bridges cognitive science and robot learning. His recent work on dual-arm manipulation datasets signals an expanding research agenda toward more generalizable, multi-task robotic systems capable of sophisticated real-world interactions.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 5
- 6