• DocumentCode
    2458516
  • Title

    A topos over an ontology as a model of experts intuition

  • Author

    Voinov, Alexander

  • Author_Institution
    Inst. for High-Performance Comput. & Databases, St. Petersburg, Russia
  • fYear
    2002
  • fDate
    2002
  • Firstpage
    26
  • Lastpage
    31
  • Abstract
    A formal representation of an expert\´s world model is described, which covers not only the domain ontology, but also the relationships of similarity between its entities. Similarities are formally treated in the framework of pretopological, topological and metric spaces, where an actually achievable grade of metricity of similarity relationships is determined experimentally. An automated inference under uncertainty in the proposed model is based on the internal logic of a (pre)sheaf where the base category is the formal domain ontology (a field of knowledge). The ordinary skeleton of entities and relationships is embedded into a (pre)topological structure, formed by the subjectively significant, for the expert, "regions of evidence", where the decision process is "intuitively obvious" and formalized with the lattice of conceptual similarities of the corresponding local neighbourhoods in his/her world model.
  • Keywords
    category theory; inference mechanisms; knowledge acquisition; knowledge representation; uncertainty handling; automated inference; category theory; experiment; formal domain ontology; knowledge acquisition; metric spaces; model of experts intuition; pretopological spaces; reasoning; topological spaces; topological structure; uncertainty; world model representation; Databases; Dictionaries; Extraterrestrial measurements; Large-scale systems; Lattices; Logic; Ontologies; Psychology; Skeleton; Uncertainty;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Artificial Intelligence Systems, 2002. (ICAIS 2002). 2002 IEEE International Conference on
  • Print_ISBN
    0-7695-1733-1
  • Type

    conf

  • DOI
    10.1109/ICAIS.2002.1048045
  • Filename
    1048045