Published multiple papers, such as 'When Maximum Entropy Misleads Policy Optimization' and 'Improving Value Estimation Critically Enhances Vanilla Policy Gradient,' presented at conferences like ICML and ICRA.
Research Experience
Associate Professor at UCSD CSE, teaching courses such as CSE150 (Introduction to AI) and CSE257 (Search and Optimization). Supervises several PhD students and alumni, with research funded by awards like the Air Force Young Investigator Award, Amazon Research Award, and others.
Background
Research interests: Practical algorithms for NP-hard search and optimization problems that arise in the decision, control, and design aspects of computational systems. Focuses on topics that can benefit from the combinatorial perspectives of automated reasoning but are posed in numerical or statistical forms. The end goal is to build fundamentally reliable yet aggressively optimized forms of automation and autonomy.