• DocumentCode
    1203226
  • Title

    An expectation-driven response understanding paradigm

  • Author

    Shin, Dong-Guk

  • Author_Institution
    Dept. of Comput. Sci. & Eng., Connecticut Univ., Storrs, CT, USA
  • Volume
    6
  • Issue
    3
  • fYear
    1994
  • fDate
    6/1/1994 12:00:00 AM
  • Firstpage
    430
  • Lastpage
    443
  • Abstract
    This paper describes a model that can account for ad hoc user-responses to argument interrogative type of system-initiated questions. Successful implementation of the model can provide an alternative solution that is more effective than the menu-driven approach that has been proposed as a meager solution to enable the system to ask a question to the user. The proposed model assumes that when the system asks a question, it maintains an expectation of the potential answers. The system then uses the expectation as the focus to perform the most likely interpretation of the user´s response. Without using such a focus the interpretation process could be unbounded. The interpretation process is mapped into a heuristic search problem. The interpretation process results in identifying a particular expectation-response relationship type, which the system can use to tailor its response strategy with respect to the given user-response. A prototype has been constructed to demonstrate the soundness of the proposed model
  • Keywords
    human factors; interactive systems; natural language interfaces; user modelling; argument interrogative questions; expectation; expectation-driven response understanding; expectation-response relationship; heuristic search problem; menu-driven approach; natural language interface; response strategy; user modelling; user response interpretation; Cities and towns; Computer science; Database systems; Expert systems; Man machine systems; Motion pictures; Natural languages; Problem-solving; Prototypes; Search problems;
  • fLanguage
    English
  • Journal_Title
    Knowledge and Data Engineering, IEEE Transactions on
  • Publisher
    ieee
  • ISSN
    1041-4347
  • Type

    jour

  • DOI
    10.1109/69.334859
  • Filename
    334859