Abstract
Traditional communication systems, striving to transmit every bit accurately, are increasingly strained by the explosive growth of data-rich applications. Semantic communication emerges as a transformative paradigm, prioritizing the conveyance of meaning over mere bit-level fidelity. This chapter provides an overview of semantic communications, highlighting its core concepts, inherent challenges such as semantic noise and interference, and the synergistic potential of Generative Artificial Intelligence (GenAI) in overcoming these hurdles. We then present a detailed case study based on a novel multi-user semantic communication system, Semantic Feature Multiple Access (SFMA), specifically designed for GenAI-enhanced video transmission. This case study delves into the architecture of SFMA, the integration of GenAI for tasks like video frame interpolation, innovative user pairing strategies, and advanced power allocation algorithms. We also explore how traditional metrics like SINR need to be adapted for the semantic domain. Finally, the chapter discusses promising future research directions, including the development of more sophisticated GenAI models for semantic processing, scalable algorithms for large-scale deployments, and the extension of semantic communication principles to multi-modal and immersive experiences.