Index transformation techniques for facilitating creative use of multiple cases
Katia P. Sycara, D. Navinchandra · International Joint Conference on Artificial Intelligence · 1991
Using cases to find innovative solutions to problems is mainly the result of two processes: (1) cross-contextual rem in dings, and (2) composition of multiple cases or case parts. Although the ability to use cases taken from across contextual boundaries is desirable, there is a tension between representing and accessing cases across contexts and in using parts of multiple cases to synthesize a solution. One way of alleviating this difficulty is through index transformation. In this paper, we represent two index transformation techniques that facilitate both cross-contextual remindings and the access of multiple appropriate case parts. The mechanisms are general and principled (based on a qualitative calculus). They are also behavior-preserving, a needed requirement for case synthesis in many domains of interest. The transformation techniques have been implemented in CADET, a case-based problem solver mat operates in the domain of mechanical design.