Shangtong Zhang

Papers

Total Citations

H-Index

About

Shangtong Zhang is a researcher advancing the frontiers of reinforcement learning, with a focus on continuous control and decision-making algorithms. His key contributions center on improving the efficiency and robustness of policy optimization in complex, high-dimensional action spaces. Zhang is best known for developing ACE (Actor Ensemble Algorithm for Continuous Control with Tree Search), a novel method that employs an ensemble of actors to more effectively search the global maxima of the critic function. This approach addresses a critical challenge in deterministic policy gradient methods—avoiding local optima—by leveraging multiple actors to explore the value landscape. The ACE algorithm, first presented in 2018 and refined in 2019, has garnered over 22 citations, reflecting its impact on the reinforcement learning community. By integrating tree search principles with actor ensembles, Zhang’s work bridges the gap between model-free and model-based methods, offering a scalable solution for tasks like robotic control and autonomous navigation. His research is widely recognized for its theoretical rigor and practical applicability, making him a notable figure in the ongoing evolution of deep reinforcement learning.