Federal Project Grant Award Summary

The National Science Foundation (NSF) Division of Mathematical Sciences awarded $180,000 to the University of Wisconsin - Madison under the Mathematical and Physical Sciences program (CFDA 47.049) on July 15, 2026, with a completion date of June 30, 2029. This project develops fast, flexible, and accurate covariance models for Gaussian processes—a foundational statistical approach used in prediction and uncertainty quantification across machine learning and artificial intelligence applications. The research addresses two critical challenges: enabling covariance functions to accommodate increasingly large and detailed modern datasets while maintaining computational feasibility for model fitting.

The project delivers research innovations through the integration of Fourier methods with covariance model design, leveraging classical algorithms including the Fast Fourier Transform (FFT) and nonuniform Fast Fourier Transform (NUFFT) to accelerate computational procedures. By enabling model validation in the Fourier domain, this approach facilitates the integration of new modeling degrees of freedom responsive to specific data features while abstracting technical complexities for practitioners. The award includes graduate and undergraduate student research training opportunities, positioning the work to advance both theoretical foundations and practical applications in machine learning and artificial intelligence across diverse data and application settings.