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

Blockchain promises to enhance distributed machine learning (ML) approachessuch as federated learning (FL) by providing further decentralization,security, immutability, and trust, which are key properties for enablingcollaborative intelligence in next-generation applications. Nonetheless, theintrinsic decentralized operation of peer-to-peer (P2P) blockchain nodes leadsto an uncharted setting for FL, whereby the concepts of FL round and globalmodel become meaningless, as devices' synchronization is lost without thefigure of a central orchestrating server. In this paper, we study the practicalimplications of outsourcing the orchestration of FL to a democratic networksuch as in a blockchain. In particular, we focus on the effects that modelstaleness and inconsistencies, endorsed by blockchains' modus operandi, have onthe training procedure held by FL devices asynchronously. Using simulation, weevaluate the blockchained FL operation on the well-known CIFAR-10 dataset andfocus on the accuracy and timeliness of the solutions. Our results show thehigh impact of model inconsistencies on the accuracy of the models (up to a~35% decrease in prediction accuracy), which underscores the importance ofproperly designing blockchain systems based on the characteristics of theunderlying FL application.

The Implications of Decentralization in Blockchained Federated Learning: Evaluating the Impact of Model Staleness and Inconsistencies