Det-CGD: Compressed Gradient Descent with Matrix Stepsizes for Non-Convex Optimization

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Li, Hanmin, Karagulyan, Avetik, Richtárik, Peter
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929322446553088
author Li, Hanmin
Karagulyan, Avetik
Richtárik, Peter
author_facet Li, Hanmin
Karagulyan, Avetik
Richtárik, Peter
contents This paper introduces a new method for minimizing matrix-smooth non-convex objectives through the use of novel Compressed Gradient Descent (CGD) algorithms enhanced with a matrix-valued stepsize. The proposed algorithms are theoretically analyzed first in the single-node and subsequently in the distributed settings. Our theoretical results reveal that the matrix stepsize in CGD can capture the objective's structure and lead to faster convergence compared to a scalar stepsize. As a byproduct of our general results, we emphasize the importance of selecting the compression mechanism and the matrix stepsize in a layer-wise manner, taking advantage of model structure. Moreover, we provide theoretical guarantees for free compression, by designing specific layer-wise compressors for the non-convex matrix smooth objectives. Our findings are supported with empirical evidence.
format Preprint
id arxiv_https___arxiv_org_abs_2305_12568
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Det-CGD: Compressed Gradient Descent with Matrix Stepsizes for Non-Convex Optimization
Li, Hanmin
Karagulyan, Avetik
Richtárik, Peter
Optimization and Control
90C26
This paper introduces a new method for minimizing matrix-smooth non-convex objectives through the use of novel Compressed Gradient Descent (CGD) algorithms enhanced with a matrix-valued stepsize. The proposed algorithms are theoretically analyzed first in the single-node and subsequently in the distributed settings. Our theoretical results reveal that the matrix stepsize in CGD can capture the objective's structure and lead to faster convergence compared to a scalar stepsize. As a byproduct of our general results, we emphasize the importance of selecting the compression mechanism and the matrix stepsize in a layer-wise manner, taking advantage of model structure. Moreover, we provide theoretical guarantees for free compression, by designing specific layer-wise compressors for the non-convex matrix smooth objectives. Our findings are supported with empirical evidence.
title Det-CGD: Compressed Gradient Descent with Matrix Stepsizes for Non-Convex Optimization
topic Optimization and Control
90C26
url https://arxiv.org/abs/2305.12568