Andrea Pferscher

Our paper “Attribution-based Explanations for Markov Decision Processes” has been accepted at the 35^th International Joint Conference on Artificial Intelligence (IJCAI) in Bremen, Germany.

In this paper, co-authored by Paul Kobialka, Francesco Leofante, Erika Ábráham, Einar Broch Johnsen, and S. Lizeth Tapia Tarifa, we propose a technique for attribution-based explanations for Markov decision processes. This enables the attribution of importance in a sequential decision-making setting.

Our article “A Self-Adaptive Digital Twin Architecture for Dynamic Resource Management” has been accepted at the 21st International Conference on Software Engineering for Adaptive and Self-Managing Systems.

In this paper, co-authored by Riccardo Sieve, Paul Kobialka, Nelly Bencomo, S. Lizeth Tapia Tarifa, and Einar Broch Johnsen, we present a software architecture for self-adaptive digital twins using semantic reflection, lifecycle management, and optimization. We showcase our architecture for the bed-bay allocation problem in a hospital.

Our article “Automata Learning versus Process Mining: The Case for User Journeys” has been published in IEEE Transactions on Software Engineering (10.1109/TSE.2026.3679253).

In this article, co-authored by Paul Kobialka, Bernhard K. Aichernig, Einar Broch Johnsen, and S. Lizeth Tapia Tarifa, we compare automata learning with process mining techniques based on the use case of learning user journeys. Our results show that both techniques have advantages and disadvantages based on the underlying event log. Hence, we propose a technique that automatically switches between automata learning and process mining depending on the quality of the underlying event log.