In this paper, we give an overview on some algorithms for learning automata. Starting with Biermann's and Angluin's algorithms, we describe some of the extensions catering for specialized or richer classes of automata. Furthermore, we survey their recent application to verification problems.