This $337,985 Project Grant awarded by the National Science Foundation (NSF) Mathematical and Physical Sciences (CFDA 47.049) program supports research to develop risk-sensitive statistical learning methods, algorithms, and theories. The research aims to advance decision-making processes in critical domains like medicine, finance, and robotics by incorporating risk assessments to improve outcomes and minimize harm. Key focus areas include quantile-constrained statistical learning, risk-sensitive deep reinforcement learning, and risk-sensitive bandit problems. The project will also create open-source software to make these advanced methods accessible to practitioners. Duke University, the primary awardee, will conduct this research over the 3-year project period from June 2024 to May 2027. No subawards are planned under this grant.
Generated 1/28/25, 9:02 AM