Federal Grant Award Summary

Northwestern University's Sponsored Research Division received a $244,168 CAREER (Faculty Early Career Development) Project Grant award from the National Science Foundation (NSF) Division of Mathematical Sciences under the Mathematical and Physical Sciences program (CFDA 47.049). The five-year award, beginning June 1, 2026 and concluding May 31, 2031, supports fundamental research and educational initiatives aimed at developing compositional learning models that bridge the gap between interpretable statistical methods and modern predictive systems. The primary research deliverable involves creating a unified framework for learning compositional architectures—such as deep neural networks—with rigorous statistical foundations to reliably uncover low-dimensional structure in high-dimensional complex data across applications in biology, econometrics, and aerodynamics.

Beyond core research activities, the award integrates educational and dissemination components, including mentoring of undergraduate and graduate students, development of new courses in statistics and machine learning at Northwestern University, and release of open-source software tools and methodologies. The project emphasizes advancing multi-index modeling through joint optimization of linear and nonlinear transformations while establishing theoretical understanding of when compositional models discover true underlying structure rather than overfitting patterns. These deliverables reflect the NSF Mathematical and Physical Sciences program's emphasis on combining intellectual merit with broader impacts through workforce development and knowledge dissemination across the scientific community.