Decentralized Optimization on Compact Submanifolds by Quantized Riemannian Gradient Tracking

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Chen, Jun, Liu, Lina, Zhu, Tianyi, Liu, Yong, Dai, Guang, Jiang, Yunliang, Tsang, Ivor W.
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912419161309184
author Chen, Jun
Liu, Lina
Zhu, Tianyi
Liu, Yong
Dai, Guang
Jiang, Yunliang
Tsang, Ivor W.
author_facet Chen, Jun
Liu, Lina
Zhu, Tianyi
Liu, Yong
Dai, Guang
Jiang, Yunliang
Tsang, Ivor W.
contents This paper considers the problem of decentralized optimization on compact submanifolds, where a finite sum of smooth (possibly non-convex) local functions is minimized by $n$ agents forming an undirected and connected graph. However, the efficiency of distributed optimization is often hindered by communication bottlenecks. To mitigate this, we propose the Quantized Riemannian Gradient Tracking (Q-RGT) algorithm, where agents update their local variables using quantized gradients. The introduction of quantization noise allows our algorithm to bypass the constraints of the accurate Riemannian projection operator (such as retraction), further improving iterative efficiency. To the best of our knowledge, this is the first algorithm to achieve an $\mathcal{O}(1/K)$ convergence rate in the presence of quantization, matching the convergence rate of methods without quantization. Additionally, we explicitly derive lower bounds on decentralized consensus associated with a function of quantization levels. Numerical experiments demonstrate that Q-RGT performs comparably to non-quantized methods while reducing communication bottlenecks and computational overhead.
format Preprint
id arxiv_https___arxiv_org_abs_2506_07351
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Decentralized Optimization on Compact Submanifolds by Quantized Riemannian Gradient Tracking
Chen, Jun
Liu, Lina
Zhu, Tianyi
Liu, Yong
Dai, Guang
Jiang, Yunliang
Tsang, Ivor W.
Optimization and Control
Machine Learning
Systems and Control
This paper considers the problem of decentralized optimization on compact submanifolds, where a finite sum of smooth (possibly non-convex) local functions is minimized by $n$ agents forming an undirected and connected graph. However, the efficiency of distributed optimization is often hindered by communication bottlenecks. To mitigate this, we propose the Quantized Riemannian Gradient Tracking (Q-RGT) algorithm, where agents update their local variables using quantized gradients. The introduction of quantization noise allows our algorithm to bypass the constraints of the accurate Riemannian projection operator (such as retraction), further improving iterative efficiency. To the best of our knowledge, this is the first algorithm to achieve an $\mathcal{O}(1/K)$ convergence rate in the presence of quantization, matching the convergence rate of methods without quantization. Additionally, we explicitly derive lower bounds on decentralized consensus associated with a function of quantization levels. Numerical experiments demonstrate that Q-RGT performs comparably to non-quantized methods while reducing communication bottlenecks and computational overhead.
title Decentralized Optimization on Compact Submanifolds by Quantized Riemannian Gradient Tracking
topic Optimization and Control
Machine Learning
Systems and Control
url https://arxiv.org/abs/2506.07351