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/

Zendy Plus

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

Zendy Tools

Zendy Open

Federated learning (FL) enables distributed training while preserving dataprivacy, but stragglers-slow or incapable clients-can significantly slow downthe total training time and degrade performance. To mitigate the impact ofstragglers, system heterogeneity, including heterogeneous computing and networkbandwidth, has been addressed. While previous studies have addressed systemheterogeneity by splitting models into submodels, they offer limitedflexibility in model architecture design, without considering potentialinconsistencies arising from training multiple submodel architectures. Wepropose nested federated learning (NeFL), a generalized framework thatefficiently divides deep neural networks into submodels using both depthwiseand widthwise scaling. To address the inconsistency arising from trainingmultiple submodel architectures, NeFL decouples a subset of parameters fromthose being trained for each submodel. An averaging method is proposed tohandle these decoupled parameters during aggregation. NeFL enablesresource-constrained devices to effectively participate in the FL pipeline,facilitating larger datasets for model training. Experiments demonstrate thatNeFL achieves performance gain, especially for the worst-case submodel comparedto baseline approaches (7.63% improvement on CIFAR-100). Furthermore, NeFLaligns with recent advances in FL, such as leveraging pre-trained models andaccounting for statistical heterogeneity. Our code is available online.

NeFL: Nested Model Scaling for Federated Learning with System Heterogeneous Clients