Using Machine Learning to find Heavily Obscured AGN with X-rays and Infrared Data

The cosmic X-ray background (CXB) is thought to be mainly produced by obscured and unobscured active galactic nuclei (AGN). Compton-thick (CT-) AGNs (with absorbing column density NH >10²⁴ cm^-2) are responsible for ~30% of the CXB at its peak and are expected to be numerous in population (30-50% as predicted by population synthesis models). However, as of today, CT-AGNs have never been detected in large numbers, with their observed fraction in the local universe being only ~5-10%. I present our newly created machine learning algorithm with the capability to discover heavily obscured sources. Using X-ray and infrared data, we can accurately predict the column density of AGN, thus giving us a better understanding of the intrinsic fraction of CT-AGN and the physics of the AGN obscuration process.