Normalized/Clipped SGD with Perturbation for Differentially Private Non-Convex Optimization.

Xiaodong Yang Huishuai Zhang Wei Chen Tie-Yan Liu

Published in: CoRR (2022)

Keyphrases

convex optimization
differentially private
differential privacy
privacy guarantees
low rank
interior point methods
total variation
primal dual
stochastic gradient descent
convex optimization problems
norm minimization
privacy preserving
machine learning
feature extraction
image restoration
operator splitting