ABSTRACT
The history of neural nets is traced, beginning with the associationist ideas of Aristotle and the physiological speculations of Hartley and Hebb. Lashley’s empirical work led him to conclude that memory traces are not localized. However, it was not until the development of the hologram by Gabor and its application to brain function by van Heerden that a possible theoretical basis for distributed information storage was proposed. Willshaw and his colleagues demonstrated that non-linear associative models were simpler, more efficient models, which retained the property of tolerance of degraded input. Since then there have been improvements in the realism of these models and attempts to integrate findings from the structure and function of the nervous system, notably the cerebellum and hippocampus.
