A Theory of Rerepresentation in Analogical Matching
Abstract
Psychologically, rerepresentation appears to be an important technique for achieving flexibility in analogical matching. This paper presents a concise theory of rerepresentation in analogical matching. It divides the problem into detecting opportunities for rerepresentation, generating rerepresentation suggestions based on libraries of general methods, and strategies for controlling the rerepresentation process. We show that the kinds of opportunities can be exhaustively derived from the principles of structure mapping, and the methods for detecting them derived from consideration of how the SME algorithm works. Four families of rerepresentation methods are proposed, as well as task-independent and task-dependent constraints on strategies. Implemented simulation examples are used for illustration.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 2003
- Accession Number
- ADA466013
Entities
People
- Dedre Gentner
- Jin Yan
- Ken Forbus
Organizations
- Northwestern University