AI Paper Atlas

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

Toward Understanding Generalization and Stability Gaps Between Centralized and Decentralized Federated Learning

Yan Sun, Li Shen, Dacheng Tao

TPAMI 2025 · Federated learning and privacy › Communication-efficient FL · approach: Client participation analysis · 7 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
use centralized FedAvg with 2log m active clients instead of ring-topology decentralized FedAvgFedAvg · CIFAR-10 Dir-0.1, m=100, E=5test accuracy (%): 59.20 -> 77.19Improved
use centralized FedAvg with 2log m clients instead of ring topologyFedAvg · CIFAR-10 Dir-0.1, m=500, E=20test accuracy (%): 45.20 -> 71.49Improved
use centralized FedAvg-6 instead of grid D-FedAvg at larger scaleFedAvg · CIFAR-10 Dir-0.1, m=500, E=5test accuracy (%): 58.20 -> 67.10Improved
use D-FedAvg on exp topology instead of FedAvg-2log mD-FedAvg · CIFAR-10 Dir-0.1, m=100, E=20test accuracy (%): 76.12 -> 76.09Mixed
DFL with ring topology at 1x bandwidth vs CFL FedAvg-3D-FedAvg · CIFAR-10 Dir-0.1, m=100, E=5test accuracy (%): 59.20 -> 67.71Hurt
increase local epochs from 5 to 20 for D-FedAvg-ringD-FedAvg-ring · CIFAR-10 Dir-0.1, m=100test accuracy (%): 59.20 -> 46.65Hurt
increase total clients m from 300 to 500 with fixed S=100 on full topologyD-FedAvg-full · CIFAR-10 fixed S=100test accuracy (%): 67.86 -> 73.12Improved
increase total clients m from 300 to 500 with fixed S=100 on exp topologyD-FedAvg-exp · CIFAR-10 fixed S=100test accuracy (%): 64.64 -> 67.28Improved
increase total clients m from 300 to 500 with fixed S=100 on ring topologyD-FedAvg-ring · CIFAR-10 fixed S=100test accuracy (%): 43.68 -> 41.19Hurt
increase total clients m from 300 to 500 with fixed S=100 on grid topologyD-FedAvg-grid · CIFAR-10 fixed S=100test accuracy (%): 57.28 -> 58.20Hurt
increase m from 30 to 50 with fixed S=1000 on ring topologyD-FedAvg-ring · CIFAR-10 fixed S=1000test accuracy (%): 63.50 -> 58.30Hurt
increase m from 30 to 50 with fixed S=1000 on full topologyD-FedAvg-full · CIFAR-10 fixed S=1000test accuracy (%): 71.36 -> 75.69Improved

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