Epscor research fellows: NSF: enhancing astronomical multi-messenger inference of the neutron star equation of state through machine learning methods

The National Science Foundation Office of Integrative Activities awarded the University of Nevada, Las Vegas $300,000 on January 1, 2027, under the Integrative Activities program (CFDA 47.083) to develop machine-learning methods that infer neutron star equation-of-state physics from multi-messenger astronomical observations.

The project provides a research fellowship to an assistant professor and training support for a graduate student. The fellowship recipient and graduate student will develop a bi-directional emulator that connects parameters defining the MUSES (Modular Unified Solver of the Equation of State) for neutron star matter to the astrophysical observables that EOS produces. This emulator removes computational barriers to robust studies of extreme matter under neutron star conditions, enabling investigation of model uncertainties and correlations that would otherwise be computationally infeasible. The work leverages high-performance computational resources and machine-learning expert training from the NSF-Simons Artificial Intelligence Institute for the Sky (SKAI Institute).

The project is conducted in collaboration with the University of Illinois Urbana-Champaign and the SKAI Institute. Performance of work is conducted at Chicago, Illinois. Period of performance runs from January 1, 2027, through December 31, 2028. The funding is classified as a Project Grant, an assistance type under the Research Infrastructure Improvement (RII) EPSCOR Research Fellows program track.

Abstract

This research infrastructure improvement (RII) epscor research fellows project provides a fellowship to an assistant professor and training for a graduate student at the university of nevada, LAS vegas (UNLV). This work is conducted in collaboration with the university of illinois urbana-champaign and the NSF-simons artificial intelligence (AI) institute for the sky (skai institute). Through the fellowship, the PI will develop a machine-learning based method to explore the most physics-rich description of the most extreme form of matter in the universe, found inside neutron STARS. This will remove the computational roadblock currently making such studies infeasible, allow US to better understand some of the most violent astronomical processes in existence, and provide training and exploration in cutting-edge machine-learning methods for both the PI and the graduate student. In this project, the PI and a graduate student will develop a BI-directional emulator able to describe the connection between the parameters defining the muses (modular unified solver of the equation of state (EOS)) for neutron star matter and the astrophysical observables this EOS provides. This emulator will allow for robust studies of the muses EOS itself, how model uncertainties in both the muses EOS and astronomical observational constraints translate into each other and how model correlations emerge and evolve. Without this emulator, such studies will be computationally infeasible and will thus significantly limit our understanding of matter under its most extreme conditions. The project will make use of modern high-performance computational resources through the skai institute, provide training for the PI and graduate student from machine-learning experts at the skai institute, enable further collaboration between UNLV and skai institute scientists and align with UNLV?S goal to leverage advanced AI technologies to enhance education, research, and campus operations. This project is supported by the epscor research infrastructure improvement program: epscor research fellows (ERF). The ERF program supports early- and mid-career investigators in eligible jurisdictions to develop collaborations at the nation?S private, government or academic research institutions. 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.- subawards are not planned for this award.