The paper entitled “Inductive Synthesis of Finite-State Controllers for POMDPs” by Roman Andriushchenko, Milan Ceska, Sebastian Junges, Joost-Pieter Katoen has been accepted for the 38th Conference on Uncertainty in Artificial Intelligence (UAI) in Eindhoven, The Netherlands. The paper was one of the 230 accepted papers out of 718 submissions.
The paper proposes an inductive synthesis-based approach for the synthesis of finite memory controllers for indefinite-horizon reachability of POMDPs. Experiments indicate that this technique is (i) competitive to state-of-the-art belief-based approaches for indefinite-horizon properties, (ii) yields smaller FSCs than existing methods for several models, and (iii) naturally treats multi-objective specifications.
The reviews of the paper and the final paper are publicly available here: