English

Painless step size adaptation for SGD

Machine Learning 2024-10-28 v1 Neural and Evolutionary Computing

Abstract

Convergence and generalization are two crucial aspects of performance in neural networks. When analyzed separately, these properties may lead to contradictory results. Optimizing a convergence rate yields fast training, but does not guarantee the best generalization error. To avoid the conflict, recent studies suggest adopting a moderately large step size for optimizers, but the added value on the performance remains unclear. We propose the LIGHT function with the four configurations which regulate explicitly an improvement in convergence and generalization on testing. This contribution allows to: 1) improve both convergence and generalization of neural networks with no need to guarantee their stability; 2) build more reliable and explainable network architectures with no need for overparameterization. We refer to it as "painless" step size adaptation.

Keywords

Cite

@article{arxiv.2102.00853,
  title  = {Painless step size adaptation for SGD},
  author = {Ilona Kulikovskikh and Tarzan Legović},
  journal= {arXiv preprint arXiv:2102.00853},
  year   = {2024}
}

Comments

This work has been submitted to the IEEE for possible publication

R2 v1 2026-06-23T22:43:27.355Z