Abstract:This paper revisits the convergence of Stochastic Mirror Descent (SMD) in the contemporary nonconvex optimization setting. Existing results for batch-free nonconvex SMD restrict the choice of the distance generating function (DGF) to be differentiable with Lipschitz continuous gradients, thereby excluding important setups such as Shannon entropy. In this work, we present a new convergence analysis of nonconvex SMD supporting general DGF, that overcomes the above limitations and relies solely on the standard assumptions. Moreover, our convergence is established with respect to the Bregman Forward-Backward envelope, which is a stronger measure than the commonly used squared norm of gradient mapping. We further extend our results to guarantee high probability convergence under sub-Gaussian noise and global convergence under the generalized Bregman Proximal Polyak-{\L}ojasiewicz condition. Additionally, we illustrate the advantages of our improved SMD theory in various nonconvex machine learning tasks by harnessing nonsmooth DGFs. Notably, in the context of nonconvex differentially private (DP) learning, our theory yields a simple algorithm with a (nearly) dimension-independent utility bound. For the problem of training linear neural networks, we develop provably convergent stochastic algorithms.

A Mirror Inertial Forward–Reflected–Backward Splitting: Convergence Analysis Beyond Convexity and Lipschitz Smoothness

Three-operator reflected forward-backward splitting algorithm with double inertial effects

Taming Nonconvex Stochastic Mirror Descent with General Bregman Divergence

A Forward-Backward Bregman Splitting Scheme for Regularized Distributed Optimization Problems

New inertial forward-backward type for variational inequalities with Quasi-monotonicity

Convergence Rates of Inertial Splitting Schemes for Nonconvex Composite Optimization

Forward–Reflected–Backward Splitting Algorithms with Momentum: Weak, Linear and Strong Convergence Results

Strong convergence of Bregman projection algorithms for solving split feasibility problems

On inertial non-lipschitz stepsize algorithms for split feasibility problems

Forward-Backward algorithms for weakly convex problems

The splitting algorithms by Ryu, by Malitsky-Tam, and by Campoy applied to normal cones of linear subspaces converge strongly to the projection onto the intersection

First Order Methods beyond Convexity and Lipschitz Gradient Continuity with Applications to Quadratic Inverse Problems

Accelerated forward-backward and Douglas-Rachford splitting dynamics

Two-step inertial Bregman alternating structure-adapted proximal gradient descent algorithm for nonconvex and nonsmooth problems

Convergence of a class of algorithms with the level-set subdifferential error bound

Accelerated forward–backward algorithms for structured monotone inclusions

The splitting algorithms by Ryu and by Malitsky-Tam applied to normal cones of linear subspaces converge strongly to the projection onto the intersection

A modified inertial proximal minimization algorithm for structured nonconvex and nonsmooth problem

A primal-dual backward reflected forward splitting algorithm for structured monotone inclusions

Convergence Rate Bounds for the Mirror Descent Method: IQCs, Popov Criterion and Bregman Divergence

Parallel inertial forward–backward splitting methods for solving variational inequality problems with variational inclusion constraints