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

Federated Learning (FL) trains a machine learning model on distributedclients without exposing individual data. Unlike centralized training that isusually based on carefully-organized data, FL deals with on-device data thatare often unfiltered and imbalanced. As a result, conventional FL trainingprotocol that treats all data equally leads to a waste of local computationalresources and slows down the global learning process. To this end, we proposeFedBalancer, a systematic FL framework that actively selects clients' trainingsamples. Our sample selection strategy prioritizes more "informative" datawhile respecting privacy and computational capabilities of clients. To betterutilize the sample selection to speed up global training, we further introducean adaptive deadline control scheme that predicts the optimal deadline for eachround with varying client training data. Compared with existing FL algorithmswith deadline configuration methods, our evaluation on five datasets from threedifferent domains shows that FedBalancer improves the time-to-accuracyperformance by 1.20~4.48x while improving the model accuracy by 1.1~5.0%. Wealso show that FedBalancer is readily applicable to other FL approaches bydemonstrating that FedBalancer improves the convergence speed and accuracy whenoperating jointly with three different FL algorithms.

FedBalancer: Data and Pace Control for Efficient Federated Learning on Heterogeneous Clients