Papers
1
Total Citations
24
H-Index
1
About
Renxing Li is a rising figure in reinforcement learning (RL), whose work tackles two of the field’s most persistent challenges: sample efficiency and learning stability in continuous action spaces. Li’s most cited paper, “Relative Entropy Regularized Sample-Efficient Reinforcement Learning With Continuous Actions” (2023, 24 citations), introduces continuous dynamic policy programming (CDPP). This novel approach extends relative entropy regularization—a concept from discrete RL—to continuous domains, enabling agents to learn more robustly from fewer interactions. By balancing exploration and exploitation through a principled information-theoretic constraint, CDPP mitigates the instability and data hunger that plague many deep RL algorithms. This contribution is particularly significant for real-world applications like robotics and autonomous control, where data collection is costly. Li’s work has already garnered attention for its elegant theoretical grounding and practical promise, positioning them as a key innovator in advancing RL toward real-world deployment.
Research Focus
Key Achievements
Top Papers
- 1