Publication:

Delta Sum Learning: An Approach for Fast and Global Convergence in Gossip Learning

 
cris.virtual.department#PLACEHOLDER_PARENT_METADATA_VALUE#
cris.virtual.department#PLACEHOLDER_PARENT_METADATA_VALUE#
cris.virtual.department#PLACEHOLDER_PARENT_METADATA_VALUE#
cris.virtual.department#PLACEHOLDER_PARENT_METADATA_VALUE#
cris.virtual.orcid0000-0003-4824-1199
cris.virtual.orcid0000-0002-1332-2290
cris.virtual.orcid0000-0003-0575-5894
cris.virtual.orcid0000-0002-4093-7338
cris.virtualsource.department505a9fa2-2261-4859-8c77-73c2ba21244c
cris.virtualsource.departmented3e85a5-47e5-455c-97ad-13d052f80bf7
cris.virtualsource.departmentbe209fe9-cb8c-4c91-821b-9c93bd548ca7
cris.virtualsource.department9991ed04-e237-4d24-a487-66d0a142de3b
cris.virtualsource.orcid505a9fa2-2261-4859-8c77-73c2ba21244c
cris.virtualsource.orcided3e85a5-47e5-455c-97ad-13d052f80bf7
cris.virtualsource.orcidbe209fe9-cb8c-4c91-821b-9c93bd548ca7
cris.virtualsource.orcid9991ed04-e237-4d24-a487-66d0a142de3b
dc.contributor.authorGoethals, Tom
dc.contributor.authorSebrechts, Merlijn
dc.contributor.authorDe Schrijver, Stijn
dc.contributor.authorDe Turck, Filip
dc.contributor.authorVolckaert, Bruno
dc.date.accessioned2026-11-16T15:05:28Z
dc.date.available2026-11-16T15:05:28Z
dc.date.createdwos2026
dc.date.issued2026
dc.description.abstractFederated Learning is a popular approach for distributed learning due to its security and computational benefits. With the advent of powerful devices in the network edge, Gossip Learning further decentralizes Federated Learning by removing centralized integration and relying fully on peer to peer updates. However, the averaging methods generally used in both Federated and Gossip Learning are not ideal for model accuracy and global convergence. Additionally, there are few options to deploy Learning workloads in the edge as part of a larger application using a declarative approach such as Kubernetes manifests. This paper proposes Delta Sum Learning as a method to improve the basic aggregation operation in Gossip Learning, and implements it in a decentralized orchestration framework based on Open Application Model, which allows for dynamic node discovery and intent-driven deployment of multiworkload applications. Evaluation results show that Delta Sum performance is on par with alternative integration methods for 10 node topologies, but results in a 58% lower global accuracy drop when scaling to 50 nodes. Overall, it shows strong global convergence and a logarithmic loss of accuracy with increasing topology size compared to a linear loss for alternatives under limited connectivity.
dc.description.wosFundingTextResearch funded by Flanders Research Foundation Junior Postdoctoral Researcher grant number 1245725N, and the NATWORK Horizon Europe project.
dc.identifier.doi10.1109/noms69089.2026.11668145
dc.identifier.isbn979-8-3315-9268-4
dc.identifier.isbn979-8-3315-9269-1
dc.identifier.issn1542-1201
dc.identifier.urihttps://imec-publications.be/handle/20.500.12860/60820
dc.language.isoeng
dc.provenance.editstepusergreet.vanhoof@imec.be
dc.publisherIEEE
dc.relation.ispartofseriesIEEE IFIP Network Operations and Management Symposium
dc.source.beginpage1
dc.source.conferenceIEEE Network Operations and Management Symposium - NOMS
dc.source.conferencedate2026-05-18
dc.source.conferencelocationRome, Italy
dc.source.endpage9
dc.source.journalNOMS 2026-2026 IEEE NETWORK OPERATIONS AND MANAGEMENT SYMPOSIUM
dc.source.numberofpages9
dc.title

Delta Sum Learning: An Approach for Fast and Global Convergence in Gossip Learning

dc.typeProceedings paper
dspace.entity.typePublication
imec.internal.crawledAt2026-09-03
imec.internal.sourcecrawler
imec.internal.wosCreatedAt2026-11-11
Files

Original bundle

Name:
9150_acc.pdf
Size:
367.56 KB
Format:
Adobe Portable Document Format
Name:
9150.pdf
Size:
453.53 KB
Format:
Adobe Portable Document Format
Publication available in collections: