2004 | OriginalPaper | Chapter
Case Based Reasoning
An Evolutionary Computation Perspective
Author : Vivek Balaraman
Published in: Frontiers of Evolutionary Computation
Publisher: Springer US
Included in: Professional Book Archive
Activate our intelligent search to find suitable subject content or patents.
Select sections of text to find matching patents with Artificial Intelligence. powered by
Select sections of text to find additional relevant content using AI-assisted search. powered by
Case-Based Reasoning (CBR) is the model of human problem solving using prior experiences (calledcases). A case memory is a learning environment where new cases are being injected, existing cases getting purged and yet others adapted to fit situations in response to environmental pressures. As experience accumulates, case memories ideally progress incrementally towards expertise. At present however there is the lack of a framework to understand the nature of this progression and how various factors influence it. In this essay, using the analogy with natural selection, we cast case memories as evolutionary systems to see whether the perspective affords us any insights. We examine case memory processes in the light of their evolutionary role and enquire into the nature of search in evolutionary case memories (ECM). We show that while there are several unresolved questions, the perspective affords us interesting speculations and observations. As an extended application of the ECM perspective we examine whether the evolution of cases can lead to abstract knowledge levels. There is evidence that experiential knowledge may not suffice to achieve higher levels of task expertise but may require abstract knowledge structures such as schema. However, little is currently known about how schema come into existence. We explore the intriguing possibility that schema may evolve from cases as an evolutionary operation. The essay also raises a number of research problems that can be attempted by both CBR and Evolutionary System communities.