Unveiling stellar physics, from birth to explosion, using AI

Project: FDCRGP

Project Details

Grant Program

Faculty Development Competitive Research Grants Program 2022-2024

Project Description

Modern astronomy produces a lot of scientific data which are not yet fully analyzed. Using AI algorithms is a promising way to move forward. In this project, we will use AI algorithms to gain insights into three stages of stellar lifetime: (a) star formation, (b) evolution of stellar clusters, and (c) core-collapse supernova (CCSN) explosion. More specifically, we will work on (a) identification of filamentary structures in giant molecular clouds, (b) finding stars in open clusters and their tidal tails, and (c) on learning about the central regions of CCSNe from the emitted gravitational wave (GW) signals.

We will approach each of the sub-projects in three steps: (1) produce databases, (2) find optimal AI algorithms, and (3) infer the physical parameters of each astrophysical system using these algorithms.
1.Construct databases:
a.For the “star formation” part, we construct a database of magnetic field and filament structures from observations of interstellar dust emission and from numerical simulations.
b.For the “star cluster evolution” part, we generate mock astro-photometric database observations from N-body simulations of star cluster evolution. Then we integrate our database with background Galactic field stars from the mock Gaia DR2/DR3 catalogs.
c.For the “CCSN” part, we perform CCSN simulations and produce GWs for a wide range of stellar progenitors with different rotational configurations and different parameters of nuclear matter.
2.Analyse different machine learning and statistical techniques and determine the most optimal AI algorithm for each sub-project.
3.Using the selected AI algorithm, develop and implement strategies for obtaining insight and constraints on the physical parameters of the astrophysical systems.

The proposed sub-projects will be conducted under leadership of the PI E. Abdikamalov, co-PI Dana Alina, and senior postdoc Bekdaulet Shukirgaliyev. The PI is an expert of CCSNe and GWs, and co-PI Alina is an expert on the dynamics of interstellar medium, its interaction with magnetic field and star formation. Dr. Shukirgaliyev is an expert on the dynamic evolution of stellar clusters. In addition to leading the three sub-projects, these scientists will work together on finding the optimal AI algorithms for each of the projects and on extracting the insights into astrophysical science using machine learning. They will supervise graduate students participating in the project, which will help students not only acquire training in AI and computational modeling, but also develop into world-class researchers.
Effective start/end date1/1/2212/31/24