Fetching the paper…

Gradient Perturbation is Underrated for Differentially Private Convex Optimization · Around