4. Covariance estimation (10 points). Let (X, Y) be two r.v.s. To estimate their covari- ance, we sample a sequence of pairs (Xi, Yi), i = 1, . ..,n, independently and from the same distribution as (X, Y). Consider the covariance estimate n n n Covn = XiYi - 2=1 2=] Is Covn unbiased? If not, how do you modify it to make it unbiased