Projects per year
In this work, we study L∗-based learning of deterministic Markov decision processes, a class of Markov decision processes where an observation following an action uniquely determines a successor state. For this purpose, we first assume an ideal setting with a teacher who provides perfect information to the student. Then, we relax this assumption and present a novel learning algorithm that collects information by sampling execution traces of the system via testing.
Experiments performed on an implementation of our sampling-based algorithm suggest that our method achieves better accuracy than state-of-the-art passive learning techniques using the same amount of test obser vations. In contrast to existing learning algorithms which assume a predefined number of states, our algorithm learns the complete model structure including the state space.
- Active automata learning
- Markov decision processes
- Model inference
ASJC Scopus subject areas
- Theoretical Computer Science
Fields of Expertise
- Information, Communication & Computing
FingerprintDive into the research topics of 'L*-Based Learning of Markov Decision Processes (Extended Version)'. Together they form a unique fingerprint.
- 1 Active
Boano, C. A., Kubin, G., Bloem, R., Horn, M., Pernkopf, F., Zakany, N., Mangard, S., Witrisal, K., Römer, K. U., Aichernig, B., Bösch, W., Baunach, M. C., Tappler, M., Malenko, M., Weiser, S., Eichlseder, M., Leitinger, E., Grosinger, J., Großwindhager, B., Ebrahimi, M., Alothman Alterkawi, A. B., Knoll, C., Teschl, R., Saukh, O., Rath, M., Steinberger, M., Steinbauer-Wagner, G. & Tranninger, M.
1/01/16 → 31/03/22
Project: Research project