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) is a prospective distributed machine learningframework that can preserve data privacy. In particular, cross-silo FL cancomplete model training by making isolated data islands of differentorganizations collaborate with a parameter server (PS) via exchanging modelparameters for multiple communication rounds. In cross-silo FL, an incentivemechanism is indispensable for motivating data owners to contribute theirmodels to FL training. However, how to allocate the reward budget amongdifferent rounds is an essential but complicated problem largely overlooked byexisting works. The challenge of this problem lies in the opaque feedbackbetween reward budget allocation and model utility improvement of FL, makingthe optimal reward budget allocation complicated. To address this problem, wedesign an online reward budget allocation algorithm using Bayesian optimizationnamed BARA (\underline{B}udget \underline{A}llocation for \underline{R}everse\underline{A}uction). Specifically, BARA can model the complicated relationshipbetween reward budget allocation and final model accuracy in FL based onhistorical training records so that the reward budget allocated to eachcommunication round is dynamically optimized so as to maximize the final modelutility. We further incorporate the BARA algorithm into reverse auction-basedincentive mechanisms to illustrate its effectiveness. Extensive experiments areconducted on real datasets to demonstrate that BARA significantly outperformscompetitive baselines by improving model utility with the same amount of rewardbudget.

BARA: Efficient Incentive Mechanism with Online Reward Budget Allocation in Cross-Silo Federated Learning