Dongjie Yu

Tsinghua University

Papers

2

Total Citations

16

H-Index

2

About

Dongjie Yu is a rising researcher at the forefront of safe reinforcement learning (RL), a critical field for deploying AI in real-world, safety-critical systems like robotics. His work directly tackles the challenge of training autonomous agents to not only achieve high performance but also to rigorously satisfy safety constraints during the learning process itself. Yu’s major contribution is the development of an **uncertainty-aware reachability certificate** for model-based RL. This innovative approach provides a formal, mathematical guarantee of safety by explicitly modeling and bounding the uncertainty in the agent’s learned dynamics model. By integrating this certificate into the RL loop, his method significantly reduces the number of safety violations that occur during training—a key bottleneck for real-world deployment. His most-cited paper (2023, 14 citations) has already established him as a promising voice in the community, offering a principled path toward trustworthy autonomy. For students and researchers, Yu’s work represents a vital bridge between theoretical safety guarantees and practical, high-performance control, making him a key figure to watch in the evolution of safe and reliable AI.

Research Focus

Key Achievements

2
H-Index
2
Papers
16
Total Citations
8
Avg Citations/Paper
🏆 Most Cited Paper
Safe Model-Based Reinforcement Learning With an Uncertainty-Aware Reachability Certificate
14 citations · 2023
📈 Most Prolific Year: 2023 (1 Papers)
🤝 Key Collaborators: 9
🏛 Institutions: Tsinghua University

Top Papers

  1. 1
  2. 2

Key Collaborators

Contact & Links

Available for collaboration
Content generated · 12 days ago