(1) E is the L2 norm of the output of C-X_P. Treat X_i as a variable and apply the chain rule to arrive at a solution.

(2) Leave the solution in terms of the X_i's.

Question 7 (10 points) Let Xn = f(Wn, Xn-1) for n = 1, ..., P, for some function f(). Let us define the value of variable E as E = |IC - XPll2, (3 for some constant C. What is the value of the gradient DE ? Xo