Knowledge Representation by Means of Multilayer Perceptrons

摘要: Publisher Summary The correct specification of knowledge is supreme for the success any autonomous system. most common type intelligent system that can be defined by using symbolic processing (SP) tools an expert (ES), which a knowledge-based program provides quality solutions to problems in specific domain. manipulated incorporated into its main component, base, built one or more existing representation languages. Each those schemes has own particularities, and variety forms each take makes it necessary categorize requiring ES, so appropriate form(s) (KRF) used. Designing incorporates base with whose underlying theory embraces connectionist paradigm requires use at least various neural network (NN) models perform computation. This chapter reviews existence important differences between derived from paradigms, demonstrates useful results obtained through their integration. It outlines KRFs considered relevant this study. also advantages and/or disadvantages result process reproducing properties KRFs.