Brain, neural networks, and computation
The method by which brain produces mind has for centuries been discussed in terms of the most complex engineering and science metaphors of the day. Descartes described mind in terms of interacting vortices. Psychologists have metaphorized memory in terms of paths or traces worn in a landscape, a geological record of our experiences. To McCulloch and Pitts (1943) and von Neumann (1958), the appropriate metaphor was the digital computer, then in its infancy. The field of “neural networks” is the study of the computational properties and behavior of networks of “neuronlike” elements. It lies somewhere between a model of neurobiology and a metaphor for how the brain computes. It is inspired by two goals: to understand how neurobiology works, and to understand how to solve problems which neurobiology solves rapidly and effortlessly and which are very hard on present digital machines.