Dongqi Han

Okinawa Institute of Science and Technology Graduate University

Papers

Total Citations

H-Index

About

Dongqi Han is a researcher specializing in deep reinforcement learning (RL), with a particular focus on solving complex control tasks under partial observability. His major contribution lies in developing variational recurrent models that integrate representation learning with policy optimization, enabling RL agents to effectively extract task-relevant information from raw, incomplete observations. This work directly addresses a critical challenge in real-world applications, where agents often lack full access to the environment state. Han’s most-cited paper, “Variational Recurrent Models for Solving Partially Observable Control Tasks” (2019), has accumulated 20 citations, with a closely related follow-up (2020) garnering 15 citations, reflecting growing interest in his approach. By combining variational inference with recurrent neural networks, his models improve both state estimation and decision-making, leading to more robust and efficient policies. Han’s research is notable for bridging the gap between theoretical advances in RL and practical deployment in partially observable settings, making his work a valuable reference for students and researchers working on memory-based agents, POMDPs, and representation learning in reinforcement learning.