Improving efficiency of federated learning using access network infrastructure

The first aspect of this project aims to address the so-called digital divide, the challenges for broadband access in remote areas. The second is data privacy in artificial intelligence solutions. With the advent of 5G, we can now use fixed wireless access (FWA) to enable broadband access in remote areas. It consists of installing a few radio equipment while deploying a core network in central locations around a community with connections back to the Internet. In order to support the FWA operation, this research project proposes improvements in the scalability and convergence of the so-called federated learning algorithms (FL). FL enables multiple local actors to build a common, robust ML model without sharing the data, consequently addressing critical issues related to data privacy. We expect the proposed solution to reduce the convergence time and improve the accuracy of FL, making it suitable for use in practice.

Faculty Supervisor:

Tristan Glatard

Student:

Partner:

Chalmers University of Technology

Discipline:

Computer science

Sector:

Education

University:

Concordia University

Program: