Zhicheng An

About

Zhicheng An is a researcher advancing the frontiers of sample-efficient reinforcement learning, with a primary focus on model-based deep RL and intelligent exploration strategies. His most cited work, "Sample Efficient Reinforcement Learning via Model-Ensemble Exploration and Exploitation" (2021), tackles a critical bottleneck in RL: the need for vast amounts of interaction data. By introducing a model-ensemble approach that simultaneously explores and exploits learned dynamics, An’s method enables agents to learn accurate world models with far fewer environment steps—a breakthrough with direct implications for domains like robotics and game-playing (e.g., Go), where real-world data is expensive. This work has garnered over 15 citations, reflecting its growing influence in the RL community. An’s contributions are particularly notable for addressing the dual challenges of uncertainty estimation and efficient exploration, helping to bridge the gap between theoretical RL and practical deployment. His research stands as a key reference for students and engineers seeking to build agents that learn faster and more robustly, making him a rising voice in the push toward truly sample-efficient artificial intelligence.

Research Focus

Key Achievements

Top Papers

1. 1
   Sample Efficient Reinforcement Learning via Model-Ensemble Exploration and Exploitation

   15 citations · 2021
2. 2
   Sample Efficient Reinforcement Learning via Model-Ensemble Exploration and Exploitation

   2 citations · 2021

Key Collaborators

Contact & Links