Explanation-Based Learning: A Problem Solving Perspective

By Steven Minton, Jaime G. Carbonell, Craig A. Knoblock, D. Kuokka, Oren Etzioni, Yolanda Gil. 1989

Abstract This article outlines explanation-based learning (EBL) and its role in improving problem solving performance through experience. Unlike inductive systems, which learn by abstracting common properties from multiple examples, EBL systems explain why a particular example is an instance of a concept. The explanations are then converted into operational recognition rules. In essence, the EBL approach is analytical and knowledge-intensive, whereas inductive methods are empirical and knowledge-poor. This article focuses on extensions of the basic EBL method and their integration with the prodigy problem solving system. prodigy 's EBL method is specifically designed to acquire search control rules that are effective in reducing total search time for complex task domains. Domain-specific search control rules are learned from successful problem solving decisions, costly failures, and unforeseen goal interactions. The ability to specify multiple learning strategies in a declarative manner enables EBL to serve as a general technique for performance improvement. prodigy 's EBL method is analyzed, illustrated with several examples and performance results, and compared with other methods for integrating EBL and problem solving.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Approximate Policy Iteration with a Policy Language Bias (2003, unknown): “[ 23 , 27, 10, 15, 1], and, more closely related, to learn stand-alone control policies [17 , 19, 28].”
Learning Relational Navigation Policies (2006, methodology): “Therefore, it is not surprising that EBL has been used as generalization algorithm within the Prodigy system [21] to learn general control rules from specific examples of problem solving episodes.”
Relational Reinforcement Learning An Overview (2004, methodology): “Indeed, EBL is at the heart of the generalization algorithms used in systems such as Prodigy and SOAR (Minton et al., 19899 Laird et al., 1986) to learn general control rules from speciic examples of problem-solving episodes.”
Approximate Policy Iteration with a Policy Language Bias: Solving Relational Markov Decision Processes (2011, background): “Two primary approaches are to learn domain-speciﬁc control rules for guiding search-based planners e.g., Minton, Carbonell, Knoblock, Kuokka, Etzioni, and Gil (1989), Veloso, Carbonell, Perez, Borrajo, Fink, and Blythe (1995), Estlin and Mooney (1996), Huang, Selman, and Kautz (2000), Ambite,…”
Training Deep Reactive Policies for Probabilistic Planning Problems (2018, background): “While these approaches have shown promise, they are not guaranteed to reduce planning times and can even result in net slow down of a planner (Minton et al. 1989).”
Solving Relational and First-Order Logical Markov Decision Processes: A Survey (2012, background): “The inspiration for this work is the similarity between Bellman backups and explanation-based generalization ( EBG) in explanation-based learning ( EBL) (see Minton et al, 1989).”