HP Labs Technical Reports
Intermediate Representations: Dynamically Optimizing the Depth of Knowledge Representation in Multilevel Systems
Abstract: The limitations of shallow knowledge representations have driven AI research to focus on deeper knowledge. These solve some problems, but come with a computational cost. This paper focuses on the computational and representational advantages that may exist in using representations intermediate between shallow and deep. The Roschian notion of basic level categories is used to develop the notion of cognitively most economic representation. For medical diagnostic systems that reason about time varying aspects of disease, it is proposed that qualitative disease histories are a good intermediate representation. Since no single representation will provide complete coverage, this paper considers how one could construct a reasoning system that uses multiple representations. Measures of intra and inter-representational adequacy are proposed to define the optimal level of such a knowledge base for a given problem and define the trade-offs that occur when using a particular representational level, and the conditions when a reasoner should switch representations.
Back to Index