Supervised learning In supervised learning we are given an input domain and an output codomain with some observations of a function . The goal is to approximate the best we can given these observations.