The large fields of convex optimization and active learning have been developed fairly independent of each other, from the design of algorithms to the techniques of proof. Given the growing literature in both these subjects, we believe that understanding the connections between them is