ABSTRACT
In neural networks have been touted as the miracle solution, both to the problems that cognitive scientists face in trying to build models of cognition and to the practical problems that engineers face in trying to develop intelligent solutions for commerce and industry. However, neural networks have also been subject to much criticism. A substantial proportion of artificial intelligence research, and the application of the technology produced from this research, is based on the assumption that all the important aspects of human cognition, and all the useful tasks which humans require their machines to do, may be described or carried out by a computational model. The classical symbolic school states that the correct level at which to model the mind is that of the symbol, an entity in a computer program that is taken to refer to an entity in the real world.
