Title:
Radio Frequency Machine Learning – Building Latent Representations from Signals
Abstract:
Radio frequency (RF) machine learning (ML) continues to expand its use in communication applications. This has largely followed the same trend as other domains, with initial deployment of supervised machine learning techniques taking over the role of traditional methods, followed by expansion into novel uses and generative AI. This talk walks through the emergence of deep learning approaches to the RF domain, helping to shed some light on methods that are now well established in the field. The tutorial will largely focus on the physical layer (wireless signals), looking at how the various types of RF ML can be applied to communications waveforms. Emphasis will be given on what it means to learn in this context as well as representation learning, which is framed as the core component to supervised, unsupervised, and generative AI. Throughout, application and domain perspectives will be shared.
Speaker:
Scott Kuzdeba
Chief Scientist and Director of the AI-enabled RF & Cyber Group, BAE Systems FAST Labs™
Bio:
Dr. Scott Kuzdeba is a Chief Scientist and Director of the AI-enabled RF & Cyber group within FAST Labs™, the research & development branch of BAE Systems, Inc. He is heavily involved in the technical oversight and development of Artificial Intelligence (AI) and Machine Learning (ML) efforts, with heavy emphasis in the radio frequency (RF) domain. His expertise includes determining how to combine domain phenomenology with machine learning, with particular focus on developing machine learning solutions to RF applications, including communications systems, radars, EW, and SIGINT. Dr. Kuzdeba recently published a book on this subject – Radio Frequency Machine Learning: A Deep Learning Perspective (Artech House, 2025). He has vast experience working R&D programs for the government S&T community and helping to bridge the gap to programs of record. Dr. Kuzdeba holds a PhD in computational neuroscience from Boston University, where he studied how human intelligence and communication systems are formed.
Host: Professor Alex Wyglinski