Reinforcement Learning the Chromatic Symmetric Function