AI Paper Atlas

Research map / Federated learning and privacy / Communication-efficient FL

Decentralized Federated Learning by Partial Message Exchange

Shan Sha, Shenglong Zhou, Xin Wang, Lingchen Kong, Geoffrey Ye Li

TPAMI 2026 · Federated learning and privacy › Communication-efficient FL · approach: Local drift correction · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
increase transmission ratio s from 0.1 to 0.6 (ablation)PaME · Linear regression, m=1000MSE: 0.462 -> 0.015Improved
increase transmission ratio s from 0.1 to 0.6 (ablation)PaME · Linear regression, m=2000MSE: 0.136 -> 0.011Improved
increase transmission ratio s from 0.1 to 0.6 (ablation)PaME · Linear regression, m=1000iterations to MSE 0.5: 96 -> 39Improved
increase participation rate v from 0.1 to 0.6 (ablation)PaME · Linear regression, m=1000MSE: 0.136 -> 0.011Improved
increase participation rate v from 0.1 to 0.6 (ablation)PaME · Linear regression, m=2000iterations to MSE 0.5: 45 -> 28Improved
increase number of nodes m from 1000 to 2000 at s=0.1 (ablation)PaME · Linear regressionMSE: 0.462 -> 0.136Improved
raise transmission rate s/n above 0.2 (ablation)PaME · Linear regression, m in {32,64,128}final objectiveNo clear effect
raise participation rate beyond 0.2 (ablation)PaME · Linear regression, m in {32,64,128}convergence speedNo clear effect
shorten communication period (homogeneous kappa) (ablation)PaME · Linear regressionfinal objectiveNo clear effect
use heterogeneous instead of homogeneous communication periods (ablation)PaME · Linear regressionconvergence speedHurt
increase number of nodes m (ablation)PaME · Linear regressionconvergence speedImproved
use partial message exchange (PaME) versus D-PSGD, DFedSAM, BEER, ANQ-NIDSPaME · Fashion-MNIST, CIFAR-10, Tiny-ImageNetaccuracyImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy