Active learning of industrial software with data

L. Sanchez, Jan Friso Groote, Ramon Schiffelers

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review


Active automata learning allows to learn software in the form of an automaton representing its behavior. The algorithm SL∗, as implemented in RALib, is one of few algorithms today that allows learning automata with data parameters. In this paper we investigate the suitability of SL∗ to learn software in an industrial environment.
For this purpose we learned a number of industrial systems, with and without data. Our conclusion is that SL∗ appears to be very suitable for learning systems of limited size with data parameters in an industrial environment. However, as it stands, SL∗ is not scalable enough to deal with more complex systems. Moreover, having more data theories available will increase practical usability.
Original languageEnglish
Title of host publicationPreproceedings of Fundamentals of Software Engineering (FSEN) 2019
EditorsHossein Hojjat, Mieke Massink
Place of PublicationTehran
PublisherInstitute for Studies in Theoretical Physics and Mathematics (IPM), School of Mathematics
Publication statusPublished - Feb 2019
Event8th IPM International Conference on Fundamentals of Software Engineering - Tehran, Iran, Islamic Republic of
Duration: 1 May 20193 May 2019


Conference8th IPM International Conference on Fundamentals of Software Engineering
Country/TerritoryIran, Islamic Republic of


Dive into the research topics of 'Active learning of industrial software with data'. Together they form a unique fingerprint.

Cite this