English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Personalized federated learning is aimed at allowing numerous clients totrain personalized models while participating in collaborative training in acommunication-efficient manner without exchanging private data. However, manypersonalized federated learning algorithms assume that clients have the sameneural network architecture, and those for heterogeneous models remainunderstudied. In this study, we propose a novel personalized federated learningmethod called federated classifier averaging (FedClassAvg). Deep neuralnetworks for supervised learning tasks consist of feature extractor andclassifier layers. FedClassAvg aggregates classifier weights as an agreement ondecision boundaries on feature spaces so that clients with not independentlyand identically distributed (non-iid) data can learn about scarce labels. Inaddition, local feature representation learning is applied to stabilize thedecision boundaries and improve the local feature extraction capabilities forclients. While the existing methods require the collection of auxiliary data ormodel weights to generate a counterpart, FedClassAvg only requires clients tocommunicate with a couple of fully connected layers, which is highlycommunication-efficient. Moreover, FedClassAvg does not require extraoptimization problems such as knowledge transfer, which requires intensivecomputation overhead. We evaluated FedClassAvg through extensive experimentsand demonstrated it outperforms the current state-of-the-art algorithms onheterogeneous personalized federated learning tasks.

FedClassAvg: Local Representation Learning for Personalized Federated Learning on Heterogeneous Neural Networks