Slides: AI for THz UM-MIMO: From Model-Driven Deep Learning to Foundation Models

(Slides only) NOTE: The recording of this webinar is not available. In these slides, the presenter show how artificial intelligence (AI) can tackle challenges posed by terahertz ultra-massive multiple-input multiple-output (THz UM-MIMO) systems. The presenter starts by outlining the characteristics of THz UM-MIMO systems, and identifying three primary challenges for transceiver design: ‘hard to compute’, ‘hard to model’, and ‘hard to measure’. They argue that AI can provide a promising solution to these challenges. The presenter then proposes two systematic research roadmaps for developing AI algorithms tailored for THz UM-MIMO systems. The first roadmap, called model-driven deep learning (DL), emphasizes the importance to leverage available domain knowledge and advocates for adopting AI only to enhance the bottleneck modules within an established signal processing or optimization framework. They discuss four essential steps to make it work, including algorithmic frameworks, basis algorithms, loss function design, and neural architecture design. Afterwards, the presenter discusses a forward-looking vision through the second roadmap, i.e., physical layer foundation models. This approach seeks to unify the design of different transceiver modules by focusing on their common foundation, i.e., the wireless channel. They propose to train a single, compact foundation model to estimate the score function of wireless channels, serving as a versatile prior for designing a wide variety of transceiver modules. The presenter will also introduce four essential steps, including general frameworks, conditioning, site-specific adaptation, and the joint design of foundation models and model-driven DL. To better illustrate the ideas, the presenter also discusses representative case studies on THz UM-MIMO channel estimation.