Differentially Private Coordinate Descent for Composite Empirical Risk Minimization

Open in new window