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: An Introduction to Communication Efficient Edge Machine Learning

Abstract: In the near future, Internet-of-Things (IoT) is expected to connect billions of devices (e.g., smartphones and sensors), which generate massive real-time data at the network edge. Intelligence can be distilled from the data to support next-generation AI-powered applications, which is called edge machine learning. One challenge faced by edge learning is the communication bottleneck, which is caused by the transmission of high-dimensional data from many edge devices to edge servers for learning. Traditional wireless techniques focusing only on efficient radio access are ineffective in tackling the challenge. Solutions should be based on a new approach that seamlessly integrates communication and computation. This has led to the emergence of a new cross-disciplinary paradigm called communication efficient edge learning. The main theme in the area is to design new communication techniques and protocols for efficient implementation of different distributed learning frameworks (i.e., federated learning) in wireless networks. This article provides an overview of the emerging area by introducing new design principles, discussing promising research opportunities, and providing design examples based on recent work.
Comments: This is an English translation of the invited paper for Journal on Communications
Subjects: Information Theory (cs.IT); Signal Processing (eess.SP)
Cite as: arXiv:1912.01554 [cs.IT]
  (or arXiv:1912.01554v1 [cs.IT] for this version)

Submission history

From: Qiao Lan [view email]
[v1] Tue, 3 Dec 2019 17:55:26 GMT (2794kb)

Link back to: arXiv, form interface, contact.