AI Paper Atlas

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

Intra-Class Unbiased Prototype Aggregation and Classifier Collaboration for Personalized Federated Learning

Hao Zheng, Shiyu Song, Zhigang Hu, Meiguang Zheng, Liu Yang, Aikun Xu, Rongchang Zhao, Ruizhi Pu, Ruiyi Fang, Boyu Wang

AAAI 2026 · Federated learning and privacy › Personalized and heterogeneous FL · approach: Non-IID federated learning · 3 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
intra-class prototype clustering and classifier collaboration instead of FedCCFAFedIC · CIFAR100 pathologicalaccuracy: 67.93 -> 75.50Improved
FedIC vs FedCCFAFedIC · CIFAR100 practicalaccuracy: 56.72 -> 65.89Improved
FedIC vs FedCCFAFedIC · FMNIST pathologicalaccuracy: 99.20 -> 99.42No clear effect
FedIC vs best baseline with ResNet18 backboneFedIC · CIFAR10 practical (ResNet18)accuracy: 92.89 -> 92.83Hurt
deeper backbone (ResNet10) vs shallow CNNFedIC · CIFAR100 practicalaccuracy: 65.89 -> 72.54Improved
increase clients from 40 to 60FedIC · CIFAR100 practicalaccuracy: 65.89 -> 65.36No clear effect
lower Dirichlet beta from 0.5 to 0.01FedIC · CIFAR100 practicalaccuracy: 71.43 -> 59.82Hurt
lower Dirichlet beta from 0.5 to 0.01baselines · CIFAR100 practicalaccuracy: 62.85 -> 48.03 (FedCCFA)Hurt
add intra-class prototype clustering to global prototype aggregation (ablation)FedIC · CIFAR100accuracyImproved
add classifier collaboration alone (ablation)FedIC · CIFAR100accuracyImproved
combine clustering and classifier collaboration (ablation)FedIC · CIFAR100accuracyImproved
use more clusters than optimal (ablation)FedIC · FMNIST, CIFAR10, CIFAR100accuracyHurt

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