liu.seSearch for publications in DiVA
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • oxford
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Decentralized Learning over Wireless Networks with Broadcast-Based Subgraph Sampling
Linköping University, Department of Electrical Engineering, Communication Systems. Linköping University, Faculty of Science & Engineering.ORCID iD: 0000-0002-6097-7935
Linköping University, Department of Electrical Engineering, Communication Systems. Linköping University, Faculty of Science & Engineering.ORCID iD: 0000-0001-5621-2860
Linköping University, Department of Electrical Engineering, Communication Systems. Linköping University, Faculty of Science & Engineering.ORCID iD: 0000-0002-7599-4367
2024 (English)In: ICC 2024 - IEEE INTERNATIONAL CONFERENCE ON COMMUNICATIONS, IEEE , 2024, p. 932-937Conference paper, Published paper (Refereed)
Abstract [en]

This work focuses on the communication perspective of decentralized learning over wireless networks, using consensus-based decentralized stochastic gradient descent (D-SGD). Considering the actual communication cost or delay caused by innetwork information exchange in every iteration, our goal is to achieve fast convergence of the algorithm measured by improvement per transmission slot. We propose BASS, an efficient communication framework for D-SGD over wireless networks with broadcast-based subgraph sampling. More explicitly, in every iteration, we activate multiple subsets of non-interfering nodes to broadcast model updates to their neighbors. These subsets are activated randomly over time with some probabilities under a given communication cost (e.g., number of transmission slots per iteration). During the consensus update step, only bi-directional links are effectively considered to preserve the communication symmetry. As compared to existing link-based scheduling methods, the broadcasting nature of wireless channels provides inherent advantages in speeding up convergence of decentralized learning by creating more communicated links under the same number of transmission slots.

Place, publisher, year, edition, pages
IEEE , 2024. p. 932-937
Series
IEEE International Conference on Communications, ISSN 1550-3607, E-ISSN 1938-1883
Keywords [en]
Decentralized machine learning; consensus optimization; wireless networks; partial communication; broadcast scheduling
National Category
Communication Systems
Identifiers
URN: urn:nbn:se:liu:diva-212055DOI: 10.1109/ICC51166.2024.10622904ISI: 001300022501009ISBN: 9781728190549 (electronic)ISBN: 9781728190556 (print)OAI: oai:DiVA.org:liu-212055DiVA, id: diva2:1942537
Conference
59th Annual IEEE International Conference on Communications (IEEE ICC), Denver, CO, jun 09-13, 2024
Available from: 2025-03-05 Created: 2025-03-05 Last updated: 2025-03-05

Open Access in DiVA

No full text in DiVA

Other links

Publisher's full text

Search in DiVA

By author/editor
Pérez Herrera, DanielChen, ZhengLarsson, Erik G
By organisation
Communication SystemsFaculty of Science & Engineering
Communication Systems

Search outside of DiVA

GoogleGoogle Scholar

doi
isbn
urn-nbn

Altmetric score

doi
isbn
urn-nbn
Total: 55 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • oxford
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf