5. Multilayer perceptron network that uses Backpropagation algorithm for training is shown in Figure 1. Assume that the learning rate a = 0.2 and the threshold value = 0.5 and the activation function used is binary sigmoid function at both hidden and output layers. 0.2 0.4 -0.3 0.5 0.3 Z: X1 V1 -0.4 -0.2 0.2 0.5 22 X2 Figure 1. Multilayer perceptron Based on Figure 1, calculate the values for 21 in, Zz out and y. out in the feedforward phase. (6 marks) 6. Backpropagation algorithm is divided into THREE (3) phases. The third phase is the backpropagation of error. Calculate the error value for the output unit (8) when t = 0.5. (4 marks)