Xiangteng Zhang
Papers
1
Total Citations
19
H-Index
1
About
Xiangteng Zhang is a leading researcher in safe reinforcement learning (RL), with a primary focus on developing theoretically grounded methods that guarantee statewise safety—a critical requirement for deploying RL in real-world, safety-critical systems. His most notable contribution, the "Feasible Actor-Critic" (FAC) framework, introduced a constrained RL approach that ensures safety at every individual state, rather than merely on average across initial states. This work, published in 2021 and garnering 19 citations, directly addresses a fundamental limitation of prior safe RL methods, offering a more rigorous and practical solution for applications such as autonomous driving, robotics, and healthcare. Zhang's research bridges the gap between theoretical safety guarantees and real-world deployment, earning recognition for its innovative approach to constraint satisfaction. By advancing the feasibility and reliability of RL agents, his work has significant implications for the future of autonomous systems, where ensuring safety at every step is paramount. His contributions continue to influence the safe RL community, inspiring further research into statewise and per-step safety constraints.
Research Focus
Key Achievements
Top Papers
- 1