We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.LG

Change to browse by:

cs

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: Communication-Efficient ADMM-based Federated Learning

Abstract: Federated learning has shown its advances over the last few years but is facing many challenges, such as how algorithms save communication resources, how they reduce computational costs, and whether they converge. To address these issues, this paper proposes exact and inexact ADMM-based federated learning. They are not only communication-efficient but also converge linearly under very mild conditions, such as convexity-free and irrelevance to data distributions. Moreover, the inexact version has low computational complexity, thereby alleviating the computational burdens significantly.
Subjects: Machine Learning (cs.LG)
Cite as: arXiv:2110.15318 [cs.LG]
  (or arXiv:2110.15318v3 [cs.LG] for this version)

Submission history

From: Shenglong Zhou [view email]
[v1] Thu, 28 Oct 2021 17:32:46 GMT (665kb,D)
[v2] Thu, 13 Jan 2022 20:58:06 GMT (665kb,D)
[v3] Sat, 29 Jan 2022 12:23:27 GMT (665kb,D)

Link back to: arXiv, form interface, contact.