A first-order primal-dual method with adaptivity to local smoothness
Optimization and Control
2021-10-29 v1 Machine Learning
Abstract
We consider the problem of finding a saddle point for the convex-concave objective , where is a convex function with locally Lipschitz gradient and is convex and possibly non-smooth. We propose an adaptive version of the Condat-V\~u algorithm, which alternates between primal gradient steps and dual proximal steps. The method achieves stepsize adaptivity through a simple rule involving and the norm of recently computed gradients of . Under standard assumptions, we prove an ergodic convergence rate. Furthermore, when is also locally strongly convex and has full row rank we show that our method converges with a linear rate. Numerical experiments are provided for illustrating the practical performance of the algorithm.
Cite
@article{arxiv.2110.15148,
title = {A first-order primal-dual method with adaptivity to local smoothness},
author = {Maria-Luiza Vladarean and Yura Malitsky and Volkan Cevher},
journal= {arXiv preprint arXiv:2110.15148},
year = {2021}
}