Mirror Descent Under Generalized Smoothness
Abstract
Smoothness is crucial for attaining fast rates in first-order optimization. However, many optimization problems in modern machine learning involve non-smooth objectives. Recent studies relax the smoothness assumption by allowing the Lipschitz constant of the gradient to grow with respect to the gradient norm, which accommodates a broad range of objectives in practice. Despite this progress, existing generalizations of smoothness are restricted to Euclidean geometry with -norm and only have theoretical guarantees for optimization in the Euclidean space. In this paper, we address this limitation by introducing a new -smoothness concept that measures the norm of Hessians in terms of a general norm and its dual, and establish convergence for mirror-descent-type algorithms, matching the rates under the classic smoothness. Notably, we propose a generalized self-bounding property that facilitates bounding the gradients via controlling suboptimality gaps, serving as a principal component for convergence analysis. Beyond deterministic optimization, we establish sharp convergence for stochastic mirror descent, matching state-of-the-art under classic smoothness. Our theory also extends to non-convex and composite optimization, which may shed light on practical usages of mirror descent, including pre-training and post-training of LLMs.
Cite
@article{arxiv.2502.00753,
title = {Mirror Descent Under Generalized Smoothness},
author = {Dingzhi Yu and Wei Jiang and Hongyi Tao and Yuanyu Wan and Lijun Zhang},
journal= {arXiv preprint arXiv:2502.00753},
year = {2026}
}
Comments
Major Update: new empirical results; weaker assumptions for AMD; new theory for SMD; new theory for non-convex CMD