• Media type: Report; E-Book
  • Title: The myth of domain-independent persistence
  • Contributor: Weber, Jay C. [Author]
  • Published: Scientific publications of the Saarland University (UdS), 1990
  • Language: English
  • DOI: https://doi.org/10.22028/D291-25018
  • ISSN: 0946-0071
  • Keywords: Künstliche Intelligenz
  • Origination:
  • Footnote: Diese Datenquelle enthält auch Bestandsnachweise, die nicht zu einem Volltext führen.
  • Description: The frame problem can be reduced to the problem of inferring the non-existence of causes for change. This paper concerns how these non-existence inferences are made, and shows how many popular approaches lack generality because they rely on a domain-independent assumption of occurrence omniscience. Also, this paper shows how to represent and use appropriate domain-dependent knowledge in three successively more expressive versions, where the causal theories are deductive, non-monotonic, and statistical.
  • Access State: Open Access