Compositional Automata Learning of Synchronous Systems

Abstract Automata learning is a technique to infer an automaton model of black-box system via queries the system. In recent years it has found widespread use both in industry and academia, as enables formal verification when no available or too complex create one manually. this paper we consider problem individual components synchronous system, assuming can only query whole We introduce compositional approach which several learners cooperate, each aiming learn components. Our experiments show that, many cases, our requires significantly fewer than widely-used non-compositional algorithm such $$\mathtt {L^*}$$ L ∗ .