EUVIMEDEUVIMED entdecken
Crossref

Statistical Model Checking of Ethical Autonomous System Under Uncertainty

Joanna Godawa, Krishnendu Ghosh

Proceedings of the AAAI Symposium Series · 06.11.2026 · 10 (1) · 915-922

DOI: 10.1609/aaaiss.v10i1.43507 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Proceedings of the AAAI Symposium Series, online erschienen am 06.11.2026. Verfasst von Joanna Godawa und 1 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Artificial intelligence-based systems are omnipresent in our daily lives. Thus, it is imperative automated systems perform decision-making ethical. The model abstraction for the evaluation of ethical decision-making of automated systems is limited because of limited data. Further, the model abstraction often do not account for environmental fluctuations during operation. In this work, we create a formalism based on internal discrete-time Markov chains to incorporate system uncertainty with integration of Deontic logic-based ethical rules. Statistical model checking is performed on a prototype to demonstrate computational feasibility and address epistemic uncertainty in the model. Experiments are conducted on a prototype of unmanned aerial vehicle on runaway using temporal logic queries and effectiveness of the results are discussed.

Zitieren

APA

Godawa, J., & Ghosh, K. (2026). Statistical Model Checking of Ethical Autonomous System Under Uncertainty. Proceedings of the AAAI Symposium Series, 10(1), 915-922. https://doi.org/10.1609/aaaiss.v10i1.43507

Vancouver

Godawa J, Ghosh K. Statistical Model Checking of Ethical Autonomous System Under Uncertainty. Proceedings of the AAAI Symposium Series. 2026;10(1):915-922. doi:10.1609/aaaiss.v10i1.43507

BibTeX
@article{Godawa2026,
  author = {Godawa, Joanna and Ghosh, Krishnendu},
  title = {Statistical Model Checking of Ethical Autonomous System Under Uncertainty},
  journal = {Proceedings of the AAAI Symposium Series},
  year = {2026},
  volume = {10},
  number = {1},
  pages = {915--922},
  doi = {10.1609/aaaiss.v10i1.43507}
}
Kostenlos starten →