In this paper, we present a unified approach to handling uncertainty during plan inference in cooperative consultations. This approach assists with the following aspects of the plan inference process: inferring a user's intentions among a number of possibilities, deciding whether to admit an unlikely interpretation of the user's request or to actively acquire information from the user, determining whether a perceived ambiguity in a user's request is to be resolved by heuristics or by soliciting information from the user, and deciding whether a recognized intention is sufficiently detailed so that a plan may be proposed to satisfy it. We define an information-theoretic measure which allows us to determine the amount of information in an interpretation of a user's request, and show how this measure is combined with probabilities of interpretations to give preference to interpretations that are better defined. Our approach is implemented as part of a computerized consultant that operates as a travel agent.
CITATION STYLE
Raskutti, B., & Zukerman, I. (1996). A unified approach to handling uncertainty during cooperative consultations. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 1114, pp. 85–96). Springer Verlag. https://doi.org/10.1007/3-540-61532-6_8
Mendeley helps you to discover research relevant for your work.