Cheongwoong Kang

Korea Advanced Institute of Science and Technology

Papers

Total Citations

H-Index

About

Cheongwoong Kang is a rising researcher advancing the frontiers of multi-agent systems and safe reinforcement learning. His work addresses critical scalability and robustness challenges in autonomous robotics. Kang’s most influential contribution, "Cooperative Multi-Robot Task Allocation with Reinforcement Learning" (2021, 40 citations), tackles the combinatorial explosion problem in assigning tasks to large robot teams. He demonstrated that traditional meta-heuristic methods falter as complexity grows, proposing a novel RL-based framework that dynamically optimizes task allocation, achieving superior scalability and efficiency—a pivotal advance for warehouse logistics and disaster response robotics. More recently, Kang has pioneered "Balanced Domain Randomization for Safe Reinforcement Learning" (2024), confronting RL agents’ tendency to overfit to training environments. By introducing a balanced randomization strategy, he enhances policy generalization and safety during sim-to-real transfer, a critical step for deploying RL in unpredictable real-world settings. His work bridges the gap between theoretical RL and practical deployment, earning recognition for its potential to make autonomous systems both more capable and reliable. Kang’s research is essential reading for anyone working in multi-robot coordination or robust policy learning.