4 papers
Attribution-based Explanations for Markov Decision Processes
Paul Kobialka, Andrea Pferscher, Francesco Leofante +3
Attribution techniques explain the outcome of an AI model by assigning a numerical score to its inputs. So far, these techniques have mainly focused on attributing importance to st…
A Hyperlogic for Strategies in Stochastic Games (Extended Version)
Lina Gerlach, Christof Löding, Erika Ãbrahám
We propose a probabilistic hyperlogic called HyperSt that can express hyperproperties of strategies in turn-based stochastic games. To the best of our knowledge, HyperSt is…
Efficient Probabilistic Model Checking for Relational Reachability (Extended Version)
Lina Gerlach, Tobias Winkler, Erika Ãbrahám +2
Markov decision processes model systems subject to nondeterministic and probabilistic uncertainty. A plethora of verification techniques addresses variations of reachability proper…
Counterfactual Strategies for Markov Decision Processes
Paul Kobialka, Lina Gerlach, Francesco Leofante +3
Counterfactuals are widely used in AI to explain how minimal changes to a model's input can lead to a different output. However, established methods for computing counterfactuals t…