Abstract

In order to solve problems more effectively with accumulating experience, a system must be able to extract, analyze, represent and exploit search control knowledge. While previous research has demonstrated that explanation-based learning is a viable method for acquiring search control knowledge, in practice explanation-based techniques may generate complex expressions that are computationally expensive to use. Better results may be obtained by explicitly reformulating the learned knowledge to maximize its effectiveness. This paper reports on the PRODIGY learning apprentice, an instructable, general-purpose problem solver that combines compression analysis with explanation-based learning, in order to formulate useful search control rules that satisfy the dual goals of generality and simplicity.
This research was sponsored in part by the Dftfense Advanced Research Projects Agency (DOD), ARPA Order …

Date

January 1, 1970

Authors

Steven Minton, Jaime G Carbonell, Craig A Knoblock, Daniel Kuokka, Henrik Nordin

Journal

Proc. Workshop on Knowledge Compilation

Pages

77-87