A parallel distributed computational model for reasoning and learning is discussed based on a belief network paradigm. Issues like reasoning and learning for the proposed model are discussed. Comparisons between our m e t h o d and other methods are also given.
Wilson X. Wen