Predicting Stellar Masses of the First Galaxies Using Graph Neural Networks

Abstract Theoretical models of galaxy formation and evolution are primarily investigated through cosmological simulations and semi-analytical models. The former method consumes O ( 10 6 ) core-hours explicitly modeling the dynamics of the galaxies, whereas the latter method only requires O ( 10 3 ) core-hours foregoing directly simulating internal structure for computational efficiency. In this work, we present a proof-of-concept machine learning regression model, using a graph neural network architecture, to predict the stellar mass of high-redshift galaxies solely from their dark matter merger trees, trained from a radiation hydrodynamics cosmological simulation of the first galaxies.