AI Paper Atlas

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

Data-Free Client Contribution Estimation via Logit Maximization for Federated Learning

Asim Ukaye, Nurbek Tastan, Mubarak Abdu-Aguye, Karthik Nandakumar

ECCV 2026 · Federated learning and privacy › Personalized and heterogeneous FL · approach: Federated learning with pretrained models

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 CELM class-wise contribution weights instead of uniform FedAvgFedAvg · CIFAR-10 PLSaccuracy: 56.82 -> 59.11Improved
use CELM weights instead of FedAvgFedAvg · FashionMNIST PLSaccuracy: 80.62 -> 83.70Improved
use CELM weights instead of FedAvgFedAvg · CIFAR-10 Dirichlet 0.10accuracy: 68.98 -> 69.21No clear effect
use CELM instead of CFFL on step label skewCFFL · FashionMNIST SLSaccuracy: 88.13 vs 87.00Hurt
use CELM on natural clinical partitionFedAvg · FedISICbalanced accuracy: 61.25 -> 70.18Improved
use class-wise normalized evidence instead of similarity-to-average (CGSV)CGSV · Maverick splitrare-class accuracy: 0.00 -> 61.21Improved
use CELM under Maverick splitFedAvg · CIFAR-10 Maverickrare-class accuracy: 39.95 -> 61.21Improved
combine CELM server aggregation with FedRS restricted softmaxFedRS · CIFAR-10 PLSaccuracy: 62.90 -> 64.05Improved
combine CELM aggregation with FedRS on Maverick splitCELM · CIFAR-10 Maverickaccuracy: 68.60 -> 63.74Hurt
use longer warm-up horizon (5% -> 30%) (ablation)CELM · FashionMNIST Dirichlet 0.05accuracy: 83.64 -> 82.70Hurt
use longer warm-up horizon (5% -> 30%) (ablation)CELM · FashionMNIST PLSaccuracy: 83.70 -> 84.14Improved
use CELM for free-rider plus Maverick detectionCFFL/CGSV · FashionMNIST FRMAUROC: 0.79 -> 1.00Improved

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