AI Paper Atlas

Research map / Federated learning and privacy / Personalized and heterogeneous FL

Factor-Assisted Federated Learning for Personalized Optimization With Heterogeneous Data

Feifei Wang, Huiyun Tang, Yang Li

TPAMI 2026 · Federated learning and privacy › Personalized and heterogeneous FL · approach: Sparse and collaboration-graph personalization · 2 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 dynamic decomposition instead of static decomposition (ablation)FedFac · CIFAR10 pi=1accuracy: 0.7047 -> 0.7153Improved
use dynamic decomposition instead of static (ablation)FedFac · CIFAR100 S=30accuracy: 0.6107 -> 0.6275Improved
use dynamic decomposition instead of static (ablation)FedFac · FEMNISTaccuracy: 0.8139 -> 0.8301Improved
use dynamic decomposition instead of static (ablation)FedFac · CIFAR100 S=5accuracy: 0.7138 -> 0.7136No clear effect
use dynamic decomposition instead of static (ablation)FedFac · CIFAR10 pi=0.1accuracy: 0.8307 -> 0.8307No clear effect
factor-analysis split of shared and personalized neurons instead of FedAvgFedFac · CIFAR100 S=5accuracy: 0.0710 -> 0.7136Improved
factor-analysis split instead of FedAvgFedFac · CIFAR10 pi=0.1accuracy: 0.8033 -> 0.8307Improved
factor-analysis split instead of FedEMFedFac · CIFAR100 S=30accuracy: 0.6183 -> 0.6107Hurt
use Ensemble instead of LocalTrain for new clients (ablation)FedFac · CIFAR10 new clients pi=0.5accuracy: 0.5750 -> 0.7154Improved
use Ensemble instead of LocalTrain for new clients (ablation)FedFac · CIFAR10 new clients pi=0.1accuracy: 0.6113 -> 0.6410Improved
random decomposition instead of factor-analysis decomposition (ablation)FedFac · SimulationaccuracyHurt
decompose deeper layers (largest Lambda) (ablation)FedFac · CIFAR100 VGG-16accuracyImproved

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