Predicting Massive Death: A Deep Learning Study on the Stellar Structure at Core Collapse

Abstract

Supernovae take millions of years to evolve, less than a second to collapse, and a couple months to radiate their light after they explode. In my thesis, I evolve 1,000-2,000 progenitor stars with a range of initial masses up until their point of core collapse using the Modules for Experimental Stellar Astrophysics (MESA) software. I then take the density versus mass profiles and compare the results with those of another paper by Sukhbold, who used different software for stellar evolution. Finally, we use the two datasets to train a deep learning program to reproduce these profiles using just the initial mass.