English

Non-Uniform Smoothness for Gradient Descent

Optimization and Control 2023-11-16 v1 Machine Learning

Abstract

The analysis of gradient descent-type methods typically relies on the Lipschitz continuity of the objective gradient. This generally requires an expensive hyperparameter tuning process to appropriately calibrate a stepsize for a given problem. In this work we introduce a local first-order smoothness oracle (LFSO) which generalizes the Lipschitz continuous gradients smoothness condition and is applicable to any twice-differentiable function. We show that this oracle can encode all relevant problem information for tuning stepsizes for a suitably modified gradient descent method and give global and local convergence results. We also show that LFSOs in this modified first-order method can yield global linear convergence rates for non-strongly convex problems with extremely flat minima, and thus improve over the lower bound on rates achievable by general (accelerated) first-order methods.

Keywords

Cite

@article{arxiv.2311.08615,
  title  = {Non-Uniform Smoothness for Gradient Descent},
  author = {Albert S. Berahas and Lindon Roberts and Fred Roosta},
  journal= {arXiv preprint arXiv:2311.08615},
  year   = {2023}
}
R2 v1 2026-06-28T13:21:32.186Z