Shengbo Eben Li

Tsinghua University

Papers

16

Total Citations

295

H-Index

9

About

Shengbo Eben Li is a prominent researcher at the intersection of reinforcement learning, optimal control, and safety-critical autonomous systems. His work spans model-free and model-based reinforcement learning, safe RL, and intelligent control for autonomous driving and robotics, establishing him as a leading voice in bridging theoretical RL advances with real-world engineering demands. Li's most impactful contributions include GOPS, a general optimal control problem solver that addresses the heavy computational burdens of traditional model predictive control while unlocking RL's potential for industrial applications (62 citations). His foundational work on safe reinforcement learning has been particularly influential — developing neural barrier certificates for model-free safety guarantees, uncertainty-aware reachability certificates for model-based approaches, and statewise safety frameworks through the Feasible Actor-Critic algorithm. These contributions collectively address one of the field's most pressing challenges: ensuring reliable, constraint-satisfying behavior in safety-critical real-world deployments. His research on distributional reinforcement learning, notably the Distributional Soft Actor-Critic with three refinements (53 citations), advances value estimation accuracy in complex decision-making tasks. With over 275 citations across his most prominent works, Li's research consistently shapes how the control and autonomous systems communities approach the safe, efficient deployment of learned controllers.

Research Focus

Key Achievements

9
H-Index
16
Papers
295
Total Citations
18
Avg Citations/Paper
🏆 Most Cited Paper
GOPS: A general optimal control problem solver for autonomous driving and industrial control applications
62 citations · 2023
📈 Most Prolific Year: 2023 (5 Papers)
🤝 Key Collaborators: 60
🏛 Institutions: Tsinghua University

Top Papers

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10

Key Collaborators

Contact & Links

Available for collaboration
Content generated · 15 days ago