Project Grant 2312204

The National Science Foundation (NSF) awarded a $444,679 Project Grant to Stanford University under the Computer and Information Science and Engineering (CFDA 47.070) program. The grant supports a collaborative research project focused on developing "statistical and algorithmic foundations for distributionally robust policy learning in unknown environments, under a possibly misspecified generative model."

The key objectives are to study the fundamental learning limits for distributionally robust policy learning, characterize how distributional uncertainty impacts learning difficulty, and develop computationally efficient and statistically optimal estimation schemes. The research aims to provide a framework for capturing model-agnostic environment changes and translate efficiency gains from learning distributionally robust policies. The award period runs from October 1, 2023 to September 30, 2027.