Model-Based Case Adaptation

Abstract

In this paper, we demonstrate an important role for model-based reasoning in case adaptation. Model-based reasoning can allow a case-based reasoner to apply cases to a wider range of problems than would otherwise be possible. We focus on case adaptation in brainstormer, a planner that uses abstract advice to help it plan in the domain of political and military policy as it relates to terrorism. We show that by equipping a case adapter with an explicit causal model of the planning process, cases presented as advice can be flexibly applied to difficulties that arise at a variety of different stages of planning. Publishing Information This report appeared in the Proceedings AAAI-92 Tenth National Conference on Artificial Intelligence, San Mateo, U. S. A. July 1992. 1 This research was conducted at the Institute for the Learning Sciences at Northwestern University, and was supported in part by the Air Force Office of Scientific Research (AFOSR). The Institute for the Learning Sciences wa...