Institute for Advanced Study/Princeton University Early Universe/Cosmology Lunch Discussion

The so-called galaxy--halo connection is a vast topic concerning how galaxies populate their host dark matter halos.  I will present some works we have made under this grand theme, including (1) a unique method to estimate the host halo mass (and its probability distribution function) simply by counting the number of neighboring galaxies for a single galaxy, and (2) a graphic neural network model that can accurately predict the stellar mass and color of galaxies at redshift z=0-5 based purely on dark matter subhalo properties along the whole merger tree, and discuss the potential applications on improving our knowledge on galaxy-halo connection.