Why Machines Learn: The Elegant Math Behind Modern AI

by Anil Ananthaswamy

and hence the network, into some stable low-energy state. This state was characterized by the strengths of the connections, or the weights, between the neurons. If you were to read off the outputs of the neurons in this stable state, they would be representative of some memory. Then, if you were to perturb the system by changing some inputs to the neurons, and hence their outputs, this would constitute a partial disruption of memory. If you read off the outputs of neurons now, they would represent distorted memory. But this perturbation would put the system into some high-energy state, and the network would dynamically find its way to the stable state. Because that low-energy, stable state represents the memory, then that memory could be retrieved. The dynamical process restores the memory.