[SOLVED] ECE 449 Assignment 8

1.  Briefly compare neurons used in a) perceptron-type networks and b) RBF-type networks.

[Hint: concentrate on what they compare, how tot is calculated, and how neuron output is determined]

2. Consider the modified Hebbian learning rule

wijnew = wijold (1−α)+hxio j

and assume the following values x_i = o_j = 1, learning rate  = 1, forgetting factor  = 0.1, and initial weight w_ij^old =0.

1. Compare standard ( = 0) and modified Hebbian learning by plotting weights over 30 subsequent learning steps (i.e. plot w_ij^new as a function of time) using the parameters provided.
2. Determine maximum value of weight w that can be obtained by modified Hebbian learning using the parameters provided [Hint: in the limit case, the values of w_ij^new and w_ij^old would be identical].