|
Dongsu Lee Hello! I'm Dongsu. I will start as a PhD student in the 2025 Fall, United States. Previously, I worked at Brain and Machine Intelligence Lab (BMIL) (Prof. Minhae Kwon) and Post-genome Informatics Lab (Prof. Sangsoo Kim) at the Soongsil University in Seoul, South Korea. I am interested in reinforcement learning, representation learning, and robotics. Recently, I explore the way to enhance the scalability of RL for robotics and multi-agent system. |
|
|
Selected Publications |
| Offline Reinforcement Learning on Foresighted Decision-making for Autonomous Driving System |
| Episodic Future Thinking Mechanism for Multi-agent Reinforcement Learning |
| AD4RL: Autonomous Driving Benchmarks for Offline Reinforcement Learning with Value-based Dataset |
| Instant Inverse Modeling of Stochastic Driving Behavior with Deep Reinforcement Learning |
|
Stability Analysis in Mixed-autonomous Traffic with Deep Reinforcement Learning
Dongsu Lee,
Minhae Kwon
IEEE T-VT 2023; NeurIPS DeepRL Workshop 2022 paper | slides | invited talk (korean) | media |
| ADAS-RL: Safety Learning Approach for Stable Autonomous Driving |
| Patents |
| Method for Combating Stop-and-Go Wave Problem Using Deep Reinforcement Learning based Autonomous Vehicles, Recording medium and device for performing this the method |
| Honors and Awards |
Scholarships |
|
AI Intensive Program at Carnegie Mellon University (Aug 2024 - Feb 2025) IITP & Sogang University (USD 41K) |
|
Future Industrial Talent Scholarship (Aug 2023 - Feb 2025) CMK Hyundai Group (Full tuition + KRW 3,600K / year) |
Awards |
|
Global Excellence Scholarship (Prize: KRW 3,000K), CMK Hyundai Group (Mar, Dec, Dec 2024) |
|
ICT Challenge IITP President’s Award (Prize: KRW 5,000K), IITP (Sep 2023) |
| Services |
Reviews |
|
Journals: IEEE T-VT (2023), IEEE IoTJ (2023) Conferences: ICRA (2025), ICML (2022) Workshops: ICML Theory of Mind Workshop (2023; Program committee), NeurIPS ML4AD (2022) |