Automatic learning of efficient behaviour

Many of the artificial neural network models so far proposed `learn´ nonlinear functional mappings from training examples. For example, the multilayer perceptron of D.E. Rumelhart and J.L. McClelland (1984) and the CMAC of J.A. Albus (1981) are both devices of this type. Neural networks are not the only function approximation methods available, and there is interest in other methods; a number of types of function learning module have been reviewed by S. Omohundro (1987). The paper describes how to use such function learning modules as components of larger learning systems that learn efficient strategies for performing multistage tasks, a method by which machines could acquire skill through experience