English

On Inexact Solution of Auxiliary Problems in Tensor Methods for Convex Optimization

Optimization and Control 2021-06-07 v3

Abstract

In this paper we study the auxiliary problems that appear in pp-order tensor methods for unconstrained minimization of convex functions with ν\nu-H\"{o}lder continuous ppth derivatives. This type of auxiliary problems corresponds to the minimization of a (p+ν)(p+\nu)-order regularization of the ppth order Taylor approximation of the objective. For the case p=3p=3, we consider the use of Gradient Methods with Bregman distance. When the regularization parameter is sufficiently large, we prove that the referred methods take at most O(log(ϵ1))\mathcal{O}(\log(\epsilon^{-1})) iterations to find either a suitable approximate stationary point of the tensor model or an ϵ\epsilon-approximate stationary point of the original objective function.

Keywords

Cite

@article{arxiv.1907.13023,
  title  = {On Inexact Solution of Auxiliary Problems in Tensor Methods for Convex Optimization},
  author = {Geovani Nunes Grapiglia and Yurii Nesterov},
  journal= {arXiv preprint arXiv:1907.13023},
  year   = {2021}
}
R2 v1 2026-06-23T10:35:00.522Z