A Variant of Anderson Mixing with Minimal Memory Size
Abstract
Anderson mixing (AM) is a useful method that can accelerate fixed-point iterations by exploring the information from historical iterations. Despite its numerical success in various applications, the memory requirement in AM remains a bottleneck when solving large-scale optimization problems in a resource-limited machine. To address this problem, we propose a novel variant of AM method, called Min-AM, by storing only one vector pair, that is the minimal memory size requirement in AM. Our method forms a symmetric approximation to the inverse Hessian matrix and is proved to be equivalent to the full-memory Type-I AM for solving strongly convex quadratic optimization. Moreover, for general nonlinear optimization problems, we establish the convergence properties of Min-AM under reasonable assumptions and show that the mixing parameters can be adaptively chosen by estimating the eigenvalues of the Hessian. Finally, we extend Min-AM to solve stochastic programming problems. Experimental results on logistic regression and network training problems validate the effectiveness of the proposed Min-AM.
Cite
Text
Wei et al. "A Variant of Anderson Mixing with Minimal Memory Size." Neural Information Processing Systems, 2022.Markdown
[Wei et al. "A Variant of Anderson Mixing with Minimal Memory Size." Neural Information Processing Systems, 2022.](https://mlanthology.org/neurips/2022/wei2022neurips-variant/)BibTeX
@inproceedings{wei2022neurips-variant,
title = {{A Variant of Anderson Mixing with Minimal Memory Size}},
author = {Wei, Fuchao and Bao, Chenglong and Liu, Yang and Yang, Guangwen},
booktitle = {Neural Information Processing Systems},
year = {2022},
url = {https://mlanthology.org/neurips/2022/wei2022neurips-variant/}
}