4. Backpropagation. (15 points) Consider the model with D-dimensional inputs x=(x1,,xD) and M-dimensional outputs y=(y1,,yM), given by ym==1LBmz,z=j=1DxjAj,m=1,,M with parameters Aj and Bm. (a) (5 points) You are given training samples xi=(xi1,,xid) and yi=(yi1,,yiM). Write yim in terms of the inputs xij. (b) (5 points) Draw the computation graph between the data, parameters and loss function using the loss function, J=i=1Nm=1M(yimyim)2. (c) (5 points) Show how to compute J/Aj from J/zi