The reviewers all agree that this paper presents an interesting alternative output layer for classification and regression networks. The reviewers were especially satisfied with the additional results and content of the rebuttal, so please be sure to include this in the final draft.