Coarsening of chiral domains in itinerant electron magnets: A machine learning force field approach