The growing demand for running machine learning (ML) services over wireless networks has stimulated the development of new communication paradigms capable of efficiently supporting distributed learning at the network edge. In wireless environments, ML services face fundamental challenges related to computation, bandwidth efficiency, scalability, privacy, and security. A prominent approach to addressing these challenges is over-the-air computation (OAC), where wireless devices transmit values using analog amplitude modulation such that a desired function—such as gradient aggregation in federated learning—is computed directly through the wireless channel at a common receiver. By exploiting the natural superposition of electromagnetic waves, OAC can dramatically reduce communication energy consumption, significantly improve spectral efficiency, and provide inherent privacy benefits.

Despite these advantages, OAC is almost exclusively confined to analog communication systems. In contrast, digital communication offers important benefits, including error correction, synchronization, channel state information acquisition, and seamless integration with existing wireless standards. In this talk, I present ChannelComp, a fundamentally new computing paradigm that enables over-the-air computation using arbitrary digital modulation schemes. I show how digital modulation not only enables reliable computation over wireless channels but also supports a broader class of computable functions than traditional OAC. Furthermore, I formulate a feasibility optimization problem that identifies the optimal digital modulation for over-the-air function computation. Simulation results demonstrate that ChannelComp significantly outperforms conventional OAC in terms of robustness and accuracy.

About Carlo Fischione

Dr. Carlo Fischione is full Professor at KTH Royal Institute of Technology, Electrical Engineering and Computer Science, Division of Network and Systems Engineering (NSE), Stockholm, Sweden. Prof. Fischione is Fellow of IEEE (the Institute of Electrical and Electronic Engineers), AAIA (Asian-Pacific Artificial Intelligence Association), DF (KTH Digital Futures research center), and DASP (the Italian academy of history Deputazione Abruzzese di Storia Patria) and is Distinguished Lecturer of the IEEE Communication Society. He is Director of the undergraduate education at NSE, Chair of the IEEE Machine Learning for Communications Emerging Technologies Initiative, founding General Chair and Steering Commitee Chair of the IEEE International Conference on Machine Learning for Communications and Networking – IEEE ICMLCN, and was the funding Director of the KTH-Ericsson Data Science Micro Degree Program directed to Ericsson globally. Prof. Fischione is the IEEE ComSoc delegate to the IEEE AI Alliance. He received the Ph.D. degree in Electrical and Information Engineering (3/3 years) in May 2005 and the Laurea degree in Electronic Engineering (Laurea, Summa cum Laude, 5/5 years) in April 2001, both from University of L’Aquila, Italy. He received the Starting Grant of the Swedish Research Council in 2008. Prof. Fischione has held research positions at Massachusetts Institute of Technology, Cambridge, MA (2015, Visiting Professor); Harvard University, Cambridge, MA (2015, Associate); and University of California at Berkeley, CA (2004-2005, Visiting Scholar, and 2007-2008, Research Associate). He is Professor at the Doctoral School of University of L’Aquila, Italy, Department of Mathematics, Information Engineering, and Computer Science. His research interests include applied optimization, wireless Internet of Things, and machine learning. He received a number of awards, such as the “IEEE Communication Society S. O. Rice” award for the best IEEE Transactions on Communications paper of 2018, the best paper award of IEEE Transactions on Industrial Informatics (2007). He is Editor of IEEE Transactions on Communications (Machine Learning for Communications area) and IEEE Transactions on Machine Learning for Communication and Networking, and has served as Associated Editor of IFAC Automatica (2014-2019). He is co-founder of the company ELK.Audio.

Este evento se impartirá en inglés