Taming mmWave Complexity with Machine Learning Approaches


In the rapidly evolving landscape of 5G, 6G, and beyond, deep learning has emerged as a transformative force, enabling unprecedented advancements in wireless communication. Traditionally, wireless resource management and channel estimation have relied on model-based approaches, leveraging signal processing and numerical optimization techniques. However, these methods often face significant computational complexity and processing overhead, especially as the number of system parameters grows exponentially with the adoption of large antenna arrays in next-generation networks. In this talk, we explore the potential of data-driven, learning-based approaches to overcome these challenges in the context of mmWave systems. We present use cases based on mmWave massive MIMO base stations enhanced with Reflective Intelligent Surfaces (RIS), representing advanced state-of-the-art network deployment scenarios. We will delve into Machine Learning (ML) based complexity reduction approaches to beam pointing algorithms and codebook design, further showcasing the versatility of machine learning in optimizing network operations. A pivotal part of the discussion will address the robustness and reliability of these solutions. We highlight the role of explainable AI (XAI) in building trust and credibility in these advanced architectures, ensuring that they are not only powerful but also transparent and dependable.


Ahmed M. Eltawil is a Professor of Electrical and Computer Engineering at KAUST where he joined the Computer, Electrical and Mathematical Science and Engineering Division (CEMSE) in 2019. Prior to that, he was a Professor of Electrical Engineering and Computer Science at the University of California, Irvine (UCI) from 2005 to 2021. At KAUST, he established the Communication and Computing Systems Laboratory (CCSL) to conduct research on efficient architectures for computing and communications systems, with a particular focus on mobile wireless systems. His research interests encompass various application domains such as low-power mobile systems, machine learning platforms, sensor networks, body area networks, and critical infrastructure networks. Professor Eltawil earned his Doctorate degree from the University of California, Los Angeles in 2003. He serves as a distinguished lecturer for the IEEE COMSOC during the 2023/24 term. Additionally, he holds senior membership in both the IEEE and the National Academy of Inventors in the United States. His contributions have garnered numerous accolades and grants, including the US National Science Foundation CAREER award, which supported his research on low-power computing and communication systems. In 2021, he was honored as the "Innovator of the Year" by the Henry Samueli School of Engineering at the University of California, Irvine. His significant advancements in wireless communication technologies have been recognized by certificates of recognition presented by the United States Congress.