Bayesian Case - Based Reasoning with Neural

| Given a problem, a case-based reasoning (CBR) system will search its case memory and use the stored cases to nd the solution, possibly modifying retrieved cases to adapt to the required input speciications. In this paper we introduce a neural network architecture for eecient case-based reasoning. We show how a rigorous Bayesian probability propagation algorithm can be implemented as a feedforward neural network and adapted for CBR. In our approach the eecient indexing problem of CBR is naturally implemented by the parallel architecture, and heuristic matching is replaced by a probability metric. This allows our CBR to perform theoretically sound Bayesian reasoning. We also show how the probability propagation actually ooers a solution to the adaptation problem in a very natural way.