BEGIN:VCALENDAR VERSION:2.0 PRODID:IEEE vTools.Events//EN CALSCALE:GREGORIAN BEGIN:VTIMEZONE TZID:Europe/Rome BEGIN:DAYLIGHT DTSTART:20260329T030000 TZOFFSETFROM:+0100 TZOFFSETTO:+0200 RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3 TZNAME:CEST END:DAYLIGHT BEGIN:STANDARD DTSTART:20261025T020000 TZOFFSETFROM:+0200 TZOFFSETTO:+0100 RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10 TZNAME:CET END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20260929T071426Z UID:66437161-3462-4634-BFAB-BB04C2C9D812 DTSTART;TZID=Europe/Rome:20261012T140000 DTEND;TZID=Europe/Rome:20261012T150000 DESCRIPTION:Large Language Models (LLMs) are poised to fundamentally reshap e the design and operation of future wireless networks. Rather than relyin g solely on narrowly specialized optimization pipelines\, emerging wireles s systems may increasingly incorporate foundation models capable of reason ing\, adaptation\, and decision-making across diverse network tasks. This talk explores how large language models may drive a shift from conventiona l wireless optimization to foundation-model-driven intelligence in future 6G networks. The talk presents three complementary advances in LLM-driven wireless intelligence. First\, multi-agent LLM frameworks for radio resour ce management are introduced for solving constrained non-convex optimizati on problems such as beamforming\, power control\, and user association thr ough iterative prompting\, evaluator-guided refinement\, and uncertainty-a ware reasoning. Second\, the talk explores LLM-based wireless time-series forecasting\, where specialized agents model trends and fluctuations in co rrelated network signals to enable predictive and adaptive network intelli gence. Third\, intelligent autonomous vehicular systems are considered thr ough the joint optimization of vehicle-to-infrastructure (V2I) communicati ons and autonomous driving policies. Across these applications\, LLM-drive n frameworks demonstrate adaptable and scalable wireless intelligence with reduced computational complexity and minimal task-specific retraining. Th e talk concludes by discussing the broader vision of unified foundation-mo del-driven architectures capable of enabling end-to-end intelligence for f uture wireless systems and 6G networks beyond traditional communication pa radigms.\n\nCo-sponsored by: Tor Vergata University of Rome\n\nSpeaker(s): Hina Tabassum\n\nRoom: Classroom C12\, Bldg: Macroarea di Ingegneria\, Vi a del Politecnico\, Roma\, Lazio\, Italy\, 00133\, Virtual: https://events .vtools.ieee.org/m/578948 LOCATION:Room: Classroom C12\, Bldg: Macroarea di Ingegneria\, Via del Poli tecnico\, Roma\, Lazio\, Italy\, 00133\, Virtual: https://events.vtools.ie ee.org/m/578948 ORGANIZER:daniele.tarchi@unifi.it SEQUENCE:25 SUMMARY:Multi-Agent Large Language Models for Wireless Optimization and For ecasting URL;VALUE=URI:https://events.vtools.ieee.org/m/578948 X-ALT-DESC:Description: <br /><p>Large Language Models (LLMs) are poised to fundamentally reshape the design and operation of future wireless network s. Rather than relying solely on narrowly specialized optimization pipelin es\, emerging wireless systems may increasingly incorporate foundation mod els capable of reasoning\, adaptation\, and decision-making across diverse network tasks. This talk explores how&nbsp\;large language models may dri ve a shift from conventional wireless optimization to foundation-model-dri ven intelligence in future 6G networks. The talk presents three complement ary advances in LLM-driven wireless intelligence. First\, multi-agent LLM frameworks for radio resource management are introduced for solving constr ained non-convex optimization problems such as beamforming\, power control \, and user association through iterative prompting\, evaluator-guided ref inement\, and uncertainty-aware reasoning. Second\, the talk explores LLM- based wireless time-series forecasting\, where specialized agents model tr ends and fluctuations in correlated network signals to enable predictive a nd adaptive network intelligence. Third\, intelligent autonomous vehicular systems are considered through the joint optimization of vehicle-to-infra structure (V2I) communications and autonomous driving policies. Across the se applications\, LLM-driven frameworks demonstrate adaptable and scalable wireless intelligence with reduced computational complexity and minimal t ask-specific retraining. The talk concludes by discussing the broader visi on of unified foundation-model-driven architectures capable of enabling en d-to-end intelligence for future wireless systems and 6G networks beyond t raditional communication paradigms.</p> END:VEVENT END:VCALENDAR