Consider the following feedforward neural net with one hidden unit of type logistic, with g(z) = 1/(1+exp(-z)), and one linear output unit. What is the new value for output y of the network after one full iteration of the gradient descent. Let the learning rate be 0.5, and target t=5.

Use the square error E=1/2(t-y)^2

Assume: x = 3.0, w0 = 2.5, and w1 = 6.0