Near-Optimal First Order Method for Saddle Point Problems with Higher Order Smoothness

MaRDI QIDQ6434194zbMATH OpenFDO

Authors Nuozhou Wang, Junyu Zhang, Shuzhong Zhang

Publication date 24 April 2023

Abstract: This paper studies the complexity of finding approximate stationary points for nonconvex-strongly-concave (NC-SC) smooth saddle point problem: $mi{n}_{x}ma{x}_{y}f(x,y)$. Under first-order smoothness conditions, and denote $kapp{a}_y$ the condition number for the inner strongly-concave problem, existing literature shows that the optimal complexity for first-order methods to obtain an $epsilon$-stationary point is $ildeTheta(sqrtkapp{a}_{y}epsilo{n}^{-2})$. However, when the objective function further satisfies certain high-order smoothness conditions, both the iteration complexity upper and lower bounds can be improved for first-order methods. In detail, given the additional assumption that the function $Phi(x):=ma{x}_{y}f(x,y)$ has Lipschitz continuous Hessian, we derive a first-order method with an improved $ildeO(sqrtkapp{a}_{y}epsilo{n}^{-7/4})$ complexity, by designing an inexact proximal point variant of the "Convex Until Proven Guilty" method that is compatible with the APPA acceleration technique for saddle point problems. In addition to an upper bound, we also construct an $Omega(sqrtkapp{a}_{y}epsilo{n}^{-12/7})$ lower bound for NC-SC problems with second-order smoothness. As a result, the complexity of our proposed algorithm is near-optimal and it almost matches the lower bound barring a negligible factor of $ildeO(epsilo{n}^{-1/28})$.

This page was built for publication: Near-Optimal First Order Method for Saddle Point Problems with Higher Order Smoothness

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6434194)