Speaker: Erik G. Larsson, Linköping University, Sweden
Title: Topics in wireless federated learning: streaming data and over-the-air-aggregation
Date and Time: 3 July 2024, at 10:30, Lagrange Gris, Inria, Sophia Antipolis
Abstract: Federated learning (FL) has received significant attention in recent years for its advantages in efficient training of machine learning models across distributed clients without disclosing user-sensitive data. In federated edge learning (FEEL) systems, the time-varying nature of wireless channels introduces inevitable system dynamics in the communication process, thereby affecting training latency and energy consumption. In this talk, we will discuss two specific aspects of FEEL systems: first, algorithms for streaming-data scenarios where new training data samples are randomly generated over time at edge devices; second, aggregation via analog over-the-air combining in both single- and multiple-antenna systems.
Biography: Erik G. Larsson is Professor at Linköping University, Sweden, and a Fellow of the IEEE. He co-authored Fundamentals of Massive MIMO (Cambridge, 2016) and Space-Time Block Coding for Wireless Communications (Cambridge, 2003). He received, among others, the IEEE ComSoc Stephen O. Rice Prize in Communications Theory 2015, the IEEE ComSoc Leonard G. Abraham Prize 2017, the IEEE ComSoc Best Tutorial Paper Award 2018, the IEEE ComSoc Fred W. Ellersick Prize in 2019, and the IEEE SPS Donald G. Fink Overview Paper Award in 2023. His main interests are within wireless communications, distributed machine learning, and network science.
