An Optimal and Scalable Matrix Mechanism for Noisy Marginals under Convex Loss Functions

Open in new window