LearnLib

LearnLib: a framework for extrapolating behavioral models. In this paper, we present the LearnLib, a library of tools for automata learning, which is explicitly designed for the systematic experimental analysis of the profile of available learning algorithms and corresponding optimizations. Its modular structure allows users to configure their own tailored learning scenarios, which exploit specific properties of their envisioned applications. As has been shown earlier, exploiting application-specific structural features enables optimizations that may lead to performance gains of several orders of magnitude, a necessary precondition to make automata learning applicable to realistic scenarios.