Living Crazy

(42) 'pattern of activation levels across a very large population of neurons. And the brain performs computations on those representations by effecting various complex vector-to-vectortransformations from one neural population to another...Mathematically, the process is an instance of multiplying a vector by a matrix and pushing the result through a nonlinear filter...Such neural networks have been shown to be 'universal approximators' 41 On the c. They can also learn to 'approximate any desired function, from repeated presentation of its instances, by means of various automatic learning procedures that adjust the synaptic weights in response to various pressures induced by the specific input-output examples presented to the network.'