English

Convergence of linesearch-based generalized conditional gradient methods without smoothness assumptions

Optimization and Control 2025-05-05 v1

Abstract

The generalized conditional gradient method is a popular algorithm for solving composite problems whose objective function is the sum of a smooth function and a nonsmooth convex function. Many convergence analyses of the algorithm rely on smoothness assumptions, such as the Lipschitz continuity of the gradient of the smooth part. This paper provides convergence results of linesearch-based generalized conditional gradient methods without smoothness assumptions. In particular, we show that a parameter-free variant, which automatically adapts to the H\"older exponent, guarantees convergence even when the gradient of the smooth part of the objective is not H\"older continuous.

Keywords

Cite

@article{arxiv.2505.01092,
  title  = {Convergence of linesearch-based generalized conditional gradient methods without smoothness assumptions},
  author = {Shotaro Yagishita},
  journal= {arXiv preprint arXiv:2505.01092},
  year   = {2025}
}

Comments

arXiv admin note: text overlap with arXiv:2505.00381

R2 v1 2026-06-28T23:18:57.564Z