Differentially Private Stochastic Optimization: New Results in Convex and Non-Convex Settings

TL;DR

This paper introduces new differentially private stochastic optimization algorithms for convex and non-convex problems, achieving near-optimal excess risk and efficient runtime, with significant improvements over prior methods in high-dimensional settings.

Contribution

It provides the first nearly dimension-independent rates for differentially private non-convex optimization with smooth losses and extends results to the    setting, improving efficiency and theoretical guarantees.

Findings

Optimal excess risk in  convex case achieved in near-linear time.

Nearly-optimal excess risk    ig(rac{ ext{log}d}{n ext{ extperthousand}}ig) for  convex losses.

First nearly dimension-independent rate for  smooth non-convex optimization.

Abstract

We study differentially private stochastic optimization in convex and non-convex settings. For the convex case, we focus on the family of non-smooth generalized linear losses (GLLs). Our algorithm for the $\ell_2$ setting achieves optimal excess population risk in near-linear time, while the best known differentially private algorithms for general convex losses run in super-linear time. Our algorithm for the $\ell_1$ setting has nearly-optimal excess population risk $\tilde{O}\big(\sqrt{\frac{\log{d}}{n\varepsilon}}\big)$, and circumvents the dimension dependent lower bound of \cite{Asi:2021} for general non-smooth convex losses. In the differentially private non-convex setting, we provide several new algorithms for approximating stationary points of the population risk. For the $\ell_1$-case with smooth losses and polyhedral constraint, we provide the first nearly dimension independent rate, $\tilde O\big(\frac{\log^{2/3}{d}}{{(n\varepsilon)^{1/3}}}\big)$ in linear time. For the constrained $\ell_2$-case with smooth losses, we obtain a linear-time algorithm with rate $\tilde O\big(\frac{1}{n^{1/3}}+\frac{d^{1/5}}{(n\varepsilon)^{2/5}}\big)$. Finally, for the $\ell_2$-case we provide the first method for {\em non-smooth weakly convex} stochastic optimization with rate $\tilde O\big(\frac{1}{n^{1/4}}+\frac{d^{1/6}}{(n\varepsilon)^{1/3}}\big)$ which matches the best existing non-private algorithm when $d= O(\sqrt{n})$. We also extend all our results above for the non-convex $\ell_2$ setting to the $\ell_p$ setting, where $1 < p \leq 2$, with only polylogarithmic (in the dimension) overhead in the rates.