Project Grant 2602036

The National Science Foundation Division of Electrical, Communications and Cyber Systems awarded the Regents of the University of Michigan $150,000 on September 1, 2026, under the Engineering program (CFDA 47.041) to develop mathematical foundations for reinforcement learning and generative artificial intelligence systems.

The project will use stochastic analysis, stochastic control, probability theory, and optimization to establish theory and algorithms for continuous-time reinforcement learning in continuous state-action spaces, including adaptive exploration and exploitation methods, policy optimization, and transfer learning for controlled diffusion processes. The work will also establish stochastic foundations for diffusion-based generative models that create data satisfying physical, economic, or scientific constraints. The research addresses AI systems operating in environments with uncertainty, sequential decision-making requirements, and evolving conditions, with applications in healthcare, transportation, energy systems, finance, and scientific computing.

The award funds research conducted at the University of Michigan in Ann Arbor, Michigan, with performance running through August 31, 2029. The project will train students and early-career researchers in mathematics and machine learning while advancing the mathematical principles underlying reliable, efficient, and controllable artificial intelligence systems aligned with national priorities in AI and translational science.