Special Presentation by Dr. Mahdi Boloursaz Mashhadi (U. of Surrey, UK) Hosted by the Future Networks Artificial Intelligence & Machine Learning (AIML) Working Group Date/Time: Thursday, 20 February 2025 @ 12:00 UTC Topic: Goal-Oriented Generative Semantic Communications with Multimodal LLMs Abstract: The integration of Generative Artificial Intelligence (GenAI) models with wireless networks provides ample opportunities to develop innovative technologies with transformative potential. One such technologies is Generative Semantic Communications (Gen SemCom), which leverages the capabilities of state-of-the-art GenAI models to develop ultra-low bitrate semantic communication systems aiming to transmit only the semantic message of interest with high fidelity. GenAI models such as Diffusion, Flow-based, and GAN models, can learn the general distribution of natural signals through training and generate new samples at the inference time. This generative process can be guided or conditioned to synthesize outputs with a desired semantic content. In Gen SemCom, the semantics of interest are extracted at the transmitter, communicated over the channel, and then used at the receiver to guide a generative model to locally synthesize a semantically consistent signal. The emerging generative foundation AI models and Multi-modal Large Language Models (MLLMs) can be leveraged in the SemCom framework to convey the most important semantics of the source signal to the receiver through textual prompts in a super compact form. These models possess a vast general knowledge through intensive pre-training on huge amount of data. This alleviates the need for a shared knowledge base/graph between the semantic transmitter and receiver, obviating the need for corresponding knowledge sharing overheads imposed in current SemCom frameworks. Despite the above benefits, deployment of such large models in the SemCom framework is challenging due to their high computational complexity, energy consumption, and latency. This talk focuses on novel generative approaches to semantic communications, the fundamental bounds on Gen SemCom, and its emerging applications in wireless networks. It investigates the challenges and opportunities of deploying Gen SemCom at various layers in future wireless networks and provides the corresponding future research directions. Speaker: [] Dr. Mahdi Boloursaz Mashhadi (Senior Member, IEEE) is a Lecturer at the 5G/6G Innovation Centre (5G/6GIC) at the Institute for Communication Systems (ICS), University of Surrey (UoS), and a Surrey AI fellow. His research is focused at the intersection of AI/ML with wireless communication, learning and communication co-design, generative AI for telecommunications, and collaborative machine learning. He received B.S., M.S., and Ph.D. degrees in mobile telecommunications from the Sharif University of Technology (SUT), Tehran, Iran. He has more than 40 peer reviewed publications and patents in the areas of wireless communications, machine learning, and signal processing. He is a PI/Co-PI for various government and industry funded projects including the UKTIN/DSIT 12M£ national project TUDOR. He received the Best Paper Award from the IEEE EWDTS conference, and the Exemplary Reviewer Award from the IEEE ComSoc in 2021 and 2022. He served as a panel judge for the International Telecommunication Union (ITU) on the “AI/ML in 5G” challenge 2021- 2022. He is an editor for the Springer Nature Wireless Personal Communications Journal. Co-sponsored by: Future Networks Artificial Intelligence & Machine Learning (AIML) Working Group Virtual: https://events.vtools.ieee.org/m/463737