Fed-ensemble: Improving Generalization through Model Ensembling in Federated Learning
Fed-ensemble: Improving Generalization through Model Ensembling in Federated Learning
About this item
Full title
Author / Creator
Publisher
Ithaca: Cornell University Library, arXiv.org
Journal title
Language
English
Formats
Publication information
Publisher
Ithaca: Cornell University Library, arXiv.org
Subjects
More information
Scope and Contents
Contents
In this paper we propose Fed-ensemble: a simple approach that bringsmodel ensembling to federated learning (FL). Instead of aggregating localmodels to update a single global model, Fed-ensemble uses random permutations to update a group of K models and then obtains predictions through model averaging. Fed-ensemble can be readily utilized within est...
Alternative Titles
Full title
Fed-ensemble: Improving Generalization through Model Ensembling in Federated Learning
Authors, Artists and Contributors
Author / Creator
Identifiers
Primary Identifiers
Record Identifier
TN_cdi_proquest_journals_2554512334
Permalink
https://devfeature-collection.sl.nsw.gov.au/record/TN_cdi_proquest_journals_2554512334
Other Identifiers
E-ISSN
2331-8422
DOI
10.48550/arxiv.2107.10663
