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

Cross-silo federated learning offers a promising solution to collaborativelytrain robust and generalized AI models without compromising the privacy oflocal datasets, e.g., healthcare, financial, as well as scientific projectsthat lack a centralized data facility. Nonetheless, because of the disparity ofcomputing resources among different clients (i.e., device heterogeneity),synchronous federated learning algorithms suffer from degraded efficiency whenwaiting for straggler clients. Similarly, asynchronous federated learningalgorithms experience degradation in the convergence rate and final modelaccuracy on non-identically and independently distributed (non-IID)heterogeneous datasets due to stale local models and client drift. To addressthese limitations in cross-silo federated learning with heterogeneous clientsand data, we propose FedCompass, an innovative semi-asynchronous federatedlearning algorithm with a computing power aware scheduler on the server side,which adaptively assigns varying amounts of training tasks to different clientsusing the knowledge of the computing power of individual clients. FedCompassensures that multiple locally trained models from clients are received almostsimultaneously as a group for aggregation, effectively reducing the stalenessof local models. At the same time, the overall training process remainsasynchronous, eliminating prolonged waiting periods from straggler clients.Using diverse non-IID heterogeneous distributed datasets, we demonstrate thatFedCompass achieves faster convergence and higher accuracy than otherasynchronous algorithms while remaining more efficient than synchronousalgorithms when performing federated learning on heterogeneous clients.

FedCompass: Efficient Cross-Silo Federated Learning on Heterogeneous Client Devices using a Computing Power Aware Scheduler