Siliang Zeng
Papers
1
Total Citations
4
H-Index
1
About
Siliang Zeng is an emerging researcher specializing in reinforcement learning theory, inverse reinforcement learning, and optimization algorithms for sequential decision-making. His most notable work, "Maximum-Likelihood Inverse Reinforcement Learning with Finite-Time Guarantees" (2022), addresses a fundamental challenge in the field: recovering reward functions from expert demonstrations with rigorous theoretical backing. Prior to Zeng's contribution, many inverse reinforcement learning algorithms relied on nested optimization structures that lacked formal convergence guarantees, making their theoretical properties poorly understood. By developing a maximum-likelihood framework equipped with finite-time convergence guarantees, Zeng provided the research community with both a principled methodology and the mathematical rigor needed to understand algorithmic behavior in practice. This work bridges the gap between empirical performance and theoretical soundness — a critical advancement for deploying IRL in real-world applications such as robotics, autonomous driving, and human-AI interaction. Though early in his career with growing citation counts, Zeng's focus on provably efficient algorithms positions him as a promising contributor to the theoretical foundations of modern machine learning and artificial intelligence research.
Research Focus
Key Achievements
Top Papers
- 1