Dohyeok Lee

I'm a first year Ph.D. student at Cognitive Machine Learning Laboratory, advised by Prof. Jungwoo Lee, in Department of Electrical and Computer Engineering at Seoul National University (SNU). Previously, I received my M.S. in ECE from SNU in 2024 and B.S. in EE from KAIST in 2020.

Research Keywords: Robot Learning, Robotics, Learning from Demonstration (LfD), Reinforcement Learning (RL).

Research Statement: My research focuses on advancing robotic manipulation through machine learning approaches. Drawing from my hands-on experience building real robot systems across personal projects and start-up companies, I recognized the limitations of classical robotics theory in real-world applications. This motivated my current research in robot learning, where I combine Learning from Demonstration (LfD) and Reinforcement Learning (RL) to develop foundation models for robotic manipulation.

Email  /  CV  /  Google Scholar  /  Github  /  LinkedIn

Research

My research lies at the intersection of robotics, machine learning, and reinforcement learning. I focus on solving real-world bottleneck problems in robotics through machine learning approaches, while grounding solutions in rigorous mathematical theory. My work spans both theoretical foundations and practical implementations. Some papers and projects are highlighted.

-Developing generalizable policy architectures for Learning from Demonstration
-Designing observation representing systems for physical understanding
-Advancing robotic manipulation through embodied datasets and reinforcement learning

We developed SPQR, a novel independence regularization method for ensemble Q-learning based on random matrix theory and spiked random model. Our approach demonstrates significant improvements in reducing overestimation bias while achieving better computational efficiency in ensemble Q-learning.

We created ARTE, an interactive drawing system enabling collaboration between humans and robots. The system features a CLIP-based reward mechanism for real-time drawing state assessment and a reinforcement learning policy trained on diverse datasets using brush stroke simulation.

Implemented a robust control system for the Furuta pendulum combining swing-up and balancing tasks using DDPG and PPO algorithms, with sim2real transfer and in-the-wild training.

Work Experience

Developing autonomous delivery robot
-Developing: driving controller module, sensor noise filtering system
-Managing: development of mobile robot hardware platform, sensor system, sensor fusion system, planning module, segmentation module

Developing autonomous filming robot
-Mobile robot platform: hardware(frame,battery system), BLDC motor control system, UART communication system
-ROS system for motor control, IMU, LiDAR, and depth camera data processing

Project Experience

*DC: Domestic Conference, *O: Open Source Contribution, *P: Personal Project, *W: Work Experience, *R: Research Project

Developing AR system integrating robot Spot with Unity, ROS

We propose isometric regularization for RL to minimize distorsion of latent space embedding

Implement nonlinear control (robust, adaptive, sliding mode) algorithms on two-arm manipulator simulator

Developing autonomous mobile robot with YOLO, Tmap API, GPS and compass sensor, etc.

Developing 3D box fitting algorithm for given point cloud data, collabolation with Hubo lab

Implement IMPALA(Scalable Distributed Deep-RL with Importance Weighted Actor-Learner Architectures) in distributed machine system with ray, redis, UDP

Developing wheel-based humanoid for moving and picking object

We propose separated batch ensemble DQN for diversification of ensemble using separated batch for Bellman Q-target

Implement genetic algorithm for earth surface decomposition with arbitrary basis function

Implement EKF(Extended Kalman Filter) for sensor fusion of GPS and IMU data with Kitti dataset

Implement RRT(Rapid Random Tree) algorithms

Creating A.I media artwork with A.I based emotion recognition and autonomous vending machine system

Developing simulation environment for surveillance/reconnaissance and reinforcement learning algorithms for surveillance agent

Developing hand-shape manipulator and glove-shape interface for teleoperation

Developing mobile robot for maker-based localization and mapping

Miscellanea

template adapted from here.