This paper provides an Angluin-style learning algorithm for a class of register automata supporting the notion of fresh data values. More specifically, we introduce session automata which are well suited for modeling protocols in which sessions using fresh values are of major interest, like in security protocols or ad-hoc networks. We show that session automata (i) have an expressiveness partly extending, partly reducing that of register automata, (ii) admit a symbolic regular representation, and (iii) have a decidable equivalence problem. We use these results to establish a learning algorithm that can infer inherently non-deterministic session automata. Moreover, in the case of deterministic automata, it has a better complexity wrt. membership queries than existing learning algorithms for register automata. We strengthen the importance of our automaton model by its characterization in monadic second-order logic.
