The University of Florida was awarded a $125,701 Project Grant from the National Science Foundation Division of Mathematical Sciences under the Mathematical and Physical Sciences federal grant program (CFDA 47.049). The grant will support the development of a mathematical foundation for novel artificial intelligence learning algorithms with applications to biology and engineering. Specifically, researchers will establish a theoretical framework for the minimax optimization of machine learning problems in the weak sense, inspired by Friedrichs' theory for partial differential equation systems. This includes approximation theory, convergence analysis for deep neural networks, and generalization analysis of deep learning in the weak form. The results are intended to advance deep learning techniques with more reliable predictions and risk-informed computation. The theoretical work will also lay the groundwork for new theories in standard strong-form deep learning. The period of performance is from August 1, 2022 to July 31, 2025.
Mod # | Description | Reason For Modification | Federal Obligation (Click to sort descending) | Date (Click to sort ascending) |
|---|---|---|---|---|
| Not listed | $125.7k | 7/27/22 |