Heeyeon Kwon
Papers
1
Total Citations
5
H-Index
1
About
Heeyeon Kwon is a researcher advancing the frontiers of reinforcement learning and human-robot interaction, with a focus on enabling robots to learn complex, multi-step visual tasks more efficiently. In her seminal work, ““Good Robot!”: Efficient reinforcement learning for multi-step visual tasks via reward shaping” (2019), Kwon introduced a novel reward-shaping framework that leverages human-like feedback to accelerate learning in visually guided robotic systems. This approach, which has garnered 5 citations, addresses a critical bottleneck in robotics: the time and data required for agents to master sequential actions in dynamic environments. By integrating reward shaping with deep reinforcement learning, her research bridges the gap between abstract human instruction and concrete robotic behavior, making autonomous systems more adaptable and intuitive. Kwon’s contributions are particularly notable for their practical implications in assistive robotics and manufacturing, where efficient task learning is paramount. Her work stands as a key step toward robots that can learn from sparse, human-provided cues, reducing the need for extensive manual programming. As a rising voice in her field, Kwon continues to explore how cognitive principles can enhance machine learning, promising safer and more capable autonomous agents.
Research Focus
Key Achievements
Top Papers
- 1“Good Robot!”: Efficient reinforcement learning for multi-step visual tasks via reward shaping5 citations · 2019