Yunuo Zhang

I am a Computer Science Ph.D. candidate at Vanderbilt University, working under the guidance of Dr. Abhishek Dubey. My research focuses on developing robust algorithms for sequential decision-making under uncertainty, with particular emphasis on belief representation methods for partially observable Markov decision processes (POMDPs) and adaptive planning in non-stationary environments.

My research interests span three key areas: Sequential Decision-Making Under Partial Observability, where I develop advanced algorithms for POMDPs including particle-based belief representation, annealed importance sampling, and Monte Carlo tree search with applications to real-time planning under uncertainty; Adaptive Planning in Non-Stationary Environments, focusing on online learning and planning methods that adapt to changing environmental dynamics, including in-context learning with temporal abstractions and policy-augmented search for evolving systems; and AI for Time-Critical Applications, deploying decision-theoretic algorithms to safety-critical domains including UAV search and rescue operations and smart grid vehicle-to-building systems.

I have published first-author papers at premier venues including NeurIPS, ICAPS, and AAMAS, advancing both theoretical foundations and practical applications of decision-making under uncertainty. Through industry collaborations, I have worked with DARPA's ANSR program on UAV search and rescue in urban environments, and partnered with Nissan Motor Corporation on vehicle-to-building energy optimization using real-world EV data.

Currently, I am particularly interested in using particle-based approaches to approximate high-dimensional and multi-modal posterior distributions, addressing fundamental challenges in belief representation for complex real-world systems where traditional parametric methods fail to capture the intricate uncertainty structures.

Prior to my doctoral studies, I earned my Bachelor's degree in Computer Science and Applied Mathematics from Vanderbilt University, where I developed a strong foundation in both theoretical computer science and mathematical optimization that underpins my current research.