I will present our work on memorization and optimization with minimum over-parameterization at Deep-Math, in San Diego!