NSF-BSF: Choreographing Astrophysical Turbulence Using Machine Learning, Simulations, and Novel Analytic Modeling

Magnetohydrodynamic (MHD) turbulence, involving magnetic fields and fluid motion, is crucial to the movement of gas in galaxies and essential for understanding star and planet formation. However, the complexity of MHD turbulence has hindered the development of a comprehensive theory of star formation and stellar convection. This project leverages machine learning to enhance turbulence resolution in simulations.

By integrating machine learning with traditional methods, the project aims to uncover fundamental equations of turbulent systems and develop new star formation models. Beyond astrophysics, this research advances education and societal engagement via a novel collaboration with the Gibney Dance Company to use dance in communicating scientific concepts.

Additionally, the project enhances the Catalog for Astrophysical Turbulence Simulations (CATS) database, creating educational resources and tools for future research in MHD turbulence.

The primary goals are to: Develop machine learning tools for super-resolution in MHD turbulence and Rayleigh-Bénard convection simulations. Test and refine analytic models of turbulent star formation and apply machine learning to derive symbolic prescriptions for star formation processes. Additionally, the PIs will enhance the CATS database, creating Python notebooks and educational materials for classroom and broader use.

The improved CATS database will serve as a gold standard for training and testing machine learning models in MHD turbulence.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.