a- Calculate the outputs of the ANN shown below if the inputs: x * 1 = 1 , x ^ 2 = 1 , * 3 = 27 Show your work. The circles are nodes, and the number inside each node is the value of the bias of input 1. The weight on each connection is given next to the line. The activation functions are sigmoid functions defined by f(x) = 1 / (1 + exp(- v)) .
b- Part (a) was the forward pass. The desired outputs are A = 1 , B = 2 and C = 3 Using the back propagation algorithm, how would all the weights and biases be adjusted in the backwards pass? The learning rate is 0.2.
c- Calculate the Cost function