Project Grant 2544082

Federal Grant Award Summary

The University of Connecticut received a $394,370 Project Grant from the National Science Foundation (NSF) Division of Computing and Communication Foundations under the Computer and Information Science and Engineering (CISE) program (CFDA 47.070). The award, effective July 1, 2026 through June 30, 2031, supports a CAREER research project focused on developing machine learning methods that function effectively with minimal labeled training data—a critical bottleneck in scientific research applications. The research deliverables include algorithmic advancements for learning families of evolution equations from limited dynamic trajectories and computer vision algorithms that leverage discovered evolution equations, enabling predictive modeling in scientific domains including materials science, biology, chemistry, and engineering.

The project addresses the challenge of applying machine learning to complex scientific systems where obtaining large quantities of labeled data is prohibitively expensive, time-consuming, or physically impossible. By adopting a dynamical systems approach that leverages temporal information about system behavior, the research aims to enhance model extrapolation to new experimental contexts and improve model interpretability—ultimately reducing scientific research costs while accelerating innovation in critical areas such as battery development, manufacturing, and materials failure prediction. The award represents NSF's investment in foundational computing research that bridges machine learning methodology with practical scientific discovery needs.