Anthony G. Francis, Jr. and Ashwin Ram
The utility problem in learning systems occurs when knowledge learned in an attempt to improve a system’s performance degrades performance instead. We present a methodology for the analysis of the utility problem which uses computational models of problem solving systems to isolate the root causes of a utility problem, to detect the threshold conditions under which the problem will arise, and to design strategies to eliminate it. We present models of case-based reasoning and control-rule learning systems and compare them with respect to the swamping utility problem. Our analysis suggests that CBR systems are more resistant to the utility problem than CRL systems.