Dohee Kim

About

Dohee Kim is a rising researcher in reinforcement learning (RL), with a focus on developing robust algorithms for physical systems operating under real-world constraints. Her work addresses critical challenges in robotics and control, specifically how to handle noisy state observations and unpredictable external disturbances. In her highly cited 2023 paper, "Distributional and hierarchical reinforcement learning for physical systems with noisy state observations and exogenous perturbations," Kim introduces a novel framework that combines distributional RL—which models the entire return distribution rather than just the expected value—with hierarchical learning to decompose complex tasks. This approach enables more resilient decision-making in environments where sensors are imperfect and dynamics are uncertain. Although early in her career, her work has already garnered 10 citations, signaling its relevance to the growing field of safe and practical RL. Kim’s contributions are particularly notable for bridging theoretical advances in distributional RL with the practical demands of physical systems, offering a pathway toward more reliable autonomous agents in manufacturing, robotics, and beyond. Her research promises to shape how RL is deployed in high-stakes, real-world settings.

Research Focus

Key Achievements

Top Papers

1. 1
   Distributional and hierarchical reinforcement learning for physical systems with noisy state observations and exogenous perturbations

   10 citations · 2023

Key Collaborators

Contact & Links