About me

Hi! I’m Zechen, a PhD student in Computer Science at Duke University advised by Ronald Parr. Previously, I was a research fellow in Computer Science at Brown University, working under the guidance of Amy Greenwald. Before that, I did my Master’s degree in Computational Science at EPFL & USI in Switzerland, where I conducted research under the guidance of Volkan Cevher. Prior to graduate school, I worked as a quantitative trader.

News

• Nov 17th, 2025: I will be presenting my NeurIPS 2025 Spotlight paper on Thursday, Dec 4th at Poster #513 during Poster Session 3 (11:00 AM – 2:00 PM, Exhibit Hall C/D/E) in San Diego. If you’re attending NeurIPS, feel free to stop by and chat! Full session details on my NeurIPS paper page.

  Many thanks to the NeurIPS Foundation for the NeurIPS 2025 Scholar Award, which is supporting my travel to the conference.

• Sept 18th, 2025: My first-author paper, A Unifying View of Linear Function Approximation in Off-Policy RL Through Matrix Splitting and Preconditioning, was accepted to NeurIPS 2025 as a Spotlight presentation (top 3% of submissions).

Research Interests

Keywords: Theoretical foundations of Reinforcement Learning (RL), deep RL, Continual learning, foundation model, Large Language Model(LLM)

I study the theoretical foundations of reinforcement learning (RL) and sequential decision-making, aiming to develop a fundamental understanding of their mechanisms, bridge theory and practice, and design efficient algorithms that solve industrial problems. I am also interested in applying these insights to emerging domains such as the foundation model, with a focus on topics including reasoning, alignment, efficiency, and optimization.