a) Given the neural network below, calculate and show the weight changes that would be made by one step of BACKPROPAGATION for the training instance (X1,X2)=(0,05,0.10) and (Y1, Y2)=(0.01,0.99). Assume that the hidden (H1 & H2) and output (Y1 & Y2) units use sigmoid functions, the network is being trained to minimize squared error, and the learning rate is 1. Edges from constant b1 and b2 on the side indicate bias parameters.

b) Consider an input image that has been converted into a matrix of size 28 X 28 and apply a kernel/filter of size 7 X 7 with a stride of 1. What will be the size of the convoluted matrix? Given below is an input matrix of shape 7 X 7. What will be the output on applying a max pooling of size 3 X 3 with a stride of 2

W1 W5 X1 H1 Y1 W6 W2 W3 W7 X2 H2 Y2 W4 W8 b1 b2 X1=0.05 b1=0.35 Y1=0.01 W1=0.15 W2=0.20 W3=0.25 W4=0.30 X2=0.10 b2=0.60 Y2=0.99 W5=0.40 W6=0.45 W7=0.50 W8=0.55 1 2 4 1 4 0 1 O 0 1 1 1 5 5 1 4 4 4 1 4 1 4 1 5 6 5 0 1 5 1 0 6 1 1 1 1 8 2 3 1 8 5 8 1 0 9 1 2 3 1 4 4