Abstract

This paper describes the integration of abstraction and explanation-based learning (EBL) in the context of the PRODIGY system. PRODIGY'S abstraction module creates a hierarchy of abstract problem spaces, so problem solving can proceed in a more directed fashion. The EBL module acquires search control knowledge by analyzing problemsolving traces. When the two modules are integrated, they tend to complement each other's capabilities, resulting in performance improvements that neither system can achieve independently. We present empirical results showing the effect of combining the two modules and describe the factors that influence the overall performance of the integrated system.

Date

July 14, 1991

Authors

Craig A Knoblock, Steven Minton, Oren Etzioni

Book

Proceedings of the ninth National conference on Artificial intelligence-Volume 2

Pages

541-546