Utility-based Categorization

The ability to categorize and use concepts e ectively is a basic requirement of any intelligent actor. The utility-based approach to categorization is founded on the thesis that categorization is fundamentally in service of action, i.e., the choice of concepts made by an actor is critical to its choice of appropriate actions. This is in contrast to classical and similarity-based approaches which seek logical completeness in concept description with respect to sensory data rather than action-oriented e ectiveness. Utility-based categorization is normative and not descriptive. It prescribes how an intelligent agent ought to conceptualize to act e ectively. It provides ideals for categorization, speci es criteria for the design of e ective computational agents, and provides a model of ideal competence. A decision-theoretic framework for utilitybased categorization which involves reasoning about alternative categorization models of varying levels of abstraction is proposed. Categorization models that are too abstract may be blind to details that are critical for selecting the most appropriate action. On the other hand, categorization models that are too detailed may be too expensive to process or may contain information not critical for selecting the most appropriate action. Categorization models are, therefore, evaluated on the basis of the expected value of their recommended action, taking into account the associated resource consumption. A knowledge representation scheme, known as probabilistic conceptual networks, has been developed to support the dynamic construction of models at varying levels of abstraction. This knowledge representation scheme combines the formalisms of in uence diagram from decision analysis and inheritance/abstraction hierarchy from arti cial intelligence. An incremental approach to categorical reasoning which involves the dynamic construction and re nement of categorization models is also developed. Models may be improved by incrementally making the concepts under consideration in the model either more abstract or more detailed. The expected increase in value of the recommended actions due to model improvement can