Explanation-Based Learning: A Survey of Programs and Perspectives

"Explanation-Based learning" (EBl) is a technique by which an intelligent system can learn by observing examples. EBl systems are characterized by the ability to create justified generalizations from single training instances. They are also distinguished by their reliance on background knowledge of the domain under study. Although EBl is usually viewed as a method for performing generalization, it can be viewed in other ways as well. In particular, EBl can be seen as a method that performs four different learning tasks: generalization, chunking, operationalization and analogy. This paper provides a general introduction to the field of explanation-based learning. It places considerable emphasis on showing how EBl combines the four learning tasks mentioned above. The paper begins by presenting an intuitive example of the EBl technique. It subsequently places EBl in its historical context and describes the relation between EBl and other areas of machine learning. The major part of this paper is a survey of selected EBl programs. The programs have been chosen to show how EBl manifests each of the four learning tasks. Attempts to formalize the EBl technique are also briefly discussed. The paper concludes by discussing the limitations of EBl and the major open questions in the field