Given responses {yi} n i=1 and features {(xi , zi)} n i=1, formulate a level-(2, 2) fair optimization problem for the given decision rule (x0) = x0 + g(x0; h). Note that in this decision rule, and h are unknown parameters to be estimated and g is a known function. For this question, you do not need to estimate anything. You leave , h and g in your formulation. The loss function is L (y0, (x0)). The tolerance to approximate fairness is m,q for m {1, 2} and q {1, 2}