Yijia Zhao

Ph.D

Statistics

University of California

Graduate Student Researcher

- Causal Bandits: Learning Good Interventions on Unknown Graphs with Limited Budget - Created a novel causal upper confidence bound algorithm by seamlessly integrating estimates derived from observational data through backdoor adjustment on Gaussian Bayesian networks - Proved a theoretical guarantee for the algorithm, demonstrating a remarkable cumulative regret bound - Conducted comparative analyses against established methods, leveraging parallel computing in R and managing batch jobs on the Hoffman2 computer cluster for efficient execution