no code implementations • 29 Jun 2023 • Martin Tappler, Edi Muškardin, Bernhard K. Aichernig, Bettina Könighofer
We aim to provide insights into the decisions faced by the agent by learning an automaton model of environmental behavior under the control of an agent.
1 code implementation • 29 Jun 2023 • Edi Muškardin, Martin Tappler, Ingo Pill, Bernhard K. Aichernig, Thomas Pock
We examine the assumption that the hidden-state vectors of recurrent neural networks (RNNs) tend to form clusters of semantically similar vectors, which we dub the clustering hypothesis.
no code implementations • 4 Dec 2022 • Bettina Könighofer, Julian Rudolf, Alexander Palmisano, Martin Tappler, Roderick Bloem
The intuition behind online shielding is to compute at runtime the set of all states that could be reached in the near future.
1 code implementation • 4 Dec 2022 • Martin Tappler, Stefan Pranger, Bettina Könighofer, Edi Muškardin, Roderick Bloem, Kim Larsen
Iteratively, we use the collected data to learn new MDPs with higher accuracy, resulting in turn in shields able to prevent more safety violations.
no code implementations • 23 Jun 2022 • Edi Muskardin, Martin Tappler, Bernhard K. Aichernig, Ingo Pill
In practical applications, we can rarely assume full observability of a system's environment, despite such knowledge being important for determining a reactive control system's precise interaction with its environment.
no code implementations • 7 May 2022 • Martin Tappler, Filip Cano Córdoba, Bernhard K. Aichernig, Bettina Könighofer
We present a search-based testing framework that enables a wide range of novel analysis capabilities for evaluating the safety and performance of deep RL agents.
no code implementations • 17 Dec 2020 • Bettina Könighofer, Julian Rudolf, Alexander Palmisano, Martin Tappler, Roderick Bloem
The intuition behind online shielding is to compute during run-time the set of all states that could be reached in the near future.
Logic in Computer Science
no code implementations • 10 Jul 2019 • Bernhard K. Aichernig, Roderick Bloem, Masoud Ebrahimi, Martin Horn, Franz Pernkopf, Wolfgang Roth, Astrid Rupp, Martin Tappler, Markus Tranninger
Therefore, there is considerable interest in learning such hybrid behavior by means of machine learning which requires sufficient and representative training data covering the behavior of the physical system adequately.
no code implementations • 28 Jun 2019 • Martin Tappler, Bernhard K. Aichernig, Giovanni Bacci, Maria Eichlseder, Kim G. Larsen
In this work, we study L*-based learning of deterministic Markov decision processes, first assuming an ideal setting with perfect information.