Answer all the following questions. [1+4+2+3=10 Marks ] A. Consider a fully connected autoencoder (each hidden node is connected to all inputs and all outputs) with two dimensional real valued input and one hidden layer with sigmoid activation function. Linear activation function is used at the output layer. Sum of squared differences L=21i(actualoutputtitargeti)2 is the loss function where i is the number of output nodes. At iteration t, the weights are shown in the following autoencoder architecture along with the input vector (x1=1,x2=0). All bias values are zero. Assume learning rate =0.2 and regularization constant =0.8. i) What will be value of the loss function at iteration t ? ii) What will be the weights w1 and w3 in iteration t+1 assuming backpropagation with ordinary gradient descent is used? iii) What will be the weights w1 and w3 in iteration t+1 assuming backpropagation with L2 regularization is used? B. Calculate the KL divergence KLD(p(x)q(x)) of two discrete distributions p(x) and q(x) in bits. Associated histograms are given by