Papers

8

Total Citations

97

H-Index

6

About

Younggyo Seo is a machine learning researcher whose work spans model-based reinforcement learning, reward learning, and vision-based robotic manipulation. His research addresses some of the most pressing challenges in making reinforcement learning more practical, data-efficient, and generalizable across real-world settings. Seo's most-cited work, "Context-aware Dynamics Model for Generalization in Model-Based Reinforcement Learning" (2020, 28 citations), tackles the difficult problem of training dynamics models that transfer across varying environments — a critical bottleneck in deploying RL agents at scale. His subsequent work on offline-to-online RL (2021, 18 citations) advanced the field's understanding of how to safely and effectively fine-tune agents pretrained on offline datasets, bridging a key gap between static training and live deployment. Beyond dynamics modeling, Seo has made notable contributions to feedback-efficient preference-based RL through semi-supervised reward learning (SURF, 2022) and explored language-conditioned reward modulation for sparse-reward tasks. More recently, his research has pushed into multimodal robot learning, combining vision and touch for dexterous manipulation and developing masked world models for visual control — work that reflects a growing ambition to build robots that perceive and act more like humans.

Research Focus

Key Achievements

6
H-Index
8
Papers
97
Total Citations
12
Avg Citations/Paper
🏆 Most Cited Paper
Context-aware Dynamics Model for Generalization in Model-Based\n Reinforcement Learning
28 citations · 2020
📈 Most Prolific Year: 2022 (2 Papers)
🤝 Key Collaborators: 18
🏛 Institutions: Korea Advanced Institute of Science and Technology

Top Papers

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8

Key Collaborators

Contact & Links

Available for collaboration
Content generated · 14 days ago