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

Download:

Current browse context:

cs.IT

Change to browse by:

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 > Information Theory

Title: Matching Pursuit Based Scheduling for Over-the-Air Federated Learning

Abstract: This paper develops a class of low-complexity device scheduling algorithms for over-the-air federated learning via the method of matching pursuit. The proposed scheme tracks closely the close-to-optimal performance achieved by difference-of-convex programming, and outperforms significantly the well-known benchmark algorithms based on convex relaxation. Compared to the state-of-the-art, the proposed scheme poses a drastically lower computational load on the system: For $K$ devices and $N$ antennas at the parameter server, the benchmark complexity scales with $\left(N^2+K\right)^3 + N^6$ while the complexity of the proposed scheme scales with $K^p N^q$ for some $0 < p,q \leq 2$. The efficiency of the proposed scheme is confirmed via numerical experiments on the CIFAR-10 dataset.
Comments: 47 Pages and 10 Figures
Subjects: Information Theory (cs.IT); Machine Learning (cs.LG)
Cite as: arXiv:2206.06679 [cs.IT]
  (or arXiv:2206.06679v2 [cs.IT] for this version)

Submission history

From: Ali Bereyhi [view email]
[v1] Tue, 14 Jun 2022 08:14:14 GMT (467kb)
[v2] Wed, 12 Oct 2022 16:58:05 GMT (822kb)

Link back to: arXiv, form interface, contact.