BEGIN:VCALENDAR VERSION:2.0 PRODID:IEEE vTools.Events//EN CALSCALE:GREGORIAN BEGIN:VTIMEZONE TZID:Europe/London BEGIN:DAYLIGHT DTSTART:20250330T020000 TZOFFSETFROM:+0000 TZOFFSETTO:+0100 RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3 TZNAME:BST END:DAYLIGHT BEGIN:STANDARD DTSTART:20241027T010000 TZOFFSETFROM:+0100 TZOFFSETTO:+0000 RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10 TZNAME:GMT END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20250321T231740Z UID:E5C48A8D-7FDE-4E16-806C-2D8F2E0E684A DTSTART;TZID=Europe/London:20250321T110000 DTEND;TZID=Europe/London:20250321T115900 DESCRIPTION:Special Talk: Fast Adaptation for Deep Learning based Wireless Communications\n\nSpeaker: Professor Geoffrey Li\, Chair in Wireless Syste ms\, Imperial College London\n\nAbstract: The integration of wireless comm unications with artificial intelligence (AI)\, especially deep learning (D L)\, is recognized as one of the six usage scenarios in next-generation ce llular systems. In this talk\, we first provide a brief overview on DL for wireless communications in the past around 10 years and present several c ritical challenges hinder the widespread applications of DL techniques in wireless communications. Since the existing DL-based wireless communicatio ns struggle to address the dynamic environments\, we then discuss fast ada ptation for DL-based wireless communications by using the few-shot learnin g (FSL) techniques. After identifying the difference between fast adaptati on in wireless communications and traditional AI tasks\, we outline two de sign requirements for applying FSL techniques to wireless communications a nd provide a comprehensive discussion on FSL techniques in wireless commun ications that satisfy these two design requirements. In particular\, we em phasize the crucial role of domain knowledge in achieving fast adaptation. At the end of this talk\, we highlight several open issues for future res earch.\n\nBiography: Geoffrey Ye Li\, FREng (Fellow of Royal Academy of En gineering) and Fellow of IEEE\, is a Chair Professor at Imperial College L ondon\, UK. Before joining Imperial in 2020\, he was a Professor at Georgi a Institute of Technology\, USA\, for 20 years and a Principal Technical S taff Member with AT&T Labs – Research (previous Bell Labs)\, USA\, for f ive years. He is the first to introduce deep learning to communications in 2016\, which has become a popular research area now. He made fundamental contributions to orthogonal frequency division multiplexing (OFDM) for wir eless communications\, which made him win 2024 IEEE Eric E. Sumner Technic al-Field Award. He also won several awards from IEEE Signal Processing\, V ehicular Technology\, and Communications Societies\, including 2019 IEEE C omSoc Edwin Howard Armstrong Achievement Award.\n\nCo-sponsored by: CH0803 1\n\nSpeaker(s): \, Geoffrey Li\n\nRoom: Room A_2A.014\, Bldg: Nancy Rothw ell\, Oxford Road\, Manchester\, England\, United Kingdom\, M13 9PL\, Virt ual: https://events.vtools.ieee.org/m/472836 LOCATION:Room: Room A_2A.014\, Bldg: Nancy Rothwell\, Oxford Road\, Manches ter\, England\, United Kingdom\, M13 9PL\, Virtual: https://events.vtools. ieee.org/m/472836 ORGANIZER:hujun.yin@manchester.ac.uk SEQUENCE:17 SUMMARY:Seminar: Fast Adaptation for Deep Learning based Wireless Communica tions URL;VALUE=URI:https://events.vtools.ieee.org/m/472836 X-ALT-DESC:Description: <br /><p><strong><span data-olk-copy-source="Messag eBody">Special Talk</span></strong><span data-olk-copy-source="MessageBody ">: Fast Adaptation for Deep Learning based Wireless Communications</span> </p>\n<p>&nbsp\;</p>\n<p><strong><span data-olk-copy-source="MessageBody"> Speaker</span></strong>: Professor Geoffrey Li\, Chair in Wireless Systems \, Imperial College London</p>\n<p>&nbsp\;</p>\n<p><strong>Abstract</stron g>:&nbsp\;The integration of wireless communications with artificial intel ligence (AI)\, especially deep learning (DL)\, is recognized as one of the six usage scenarios in next-generation cellular systems. In this talk\, w e first provide a brief overview on DL for wireless communications in the past around 10 years and present several critical challenges hinder the wi despread applications of DL techniques in wireless communications. Since t he existing DL-based wireless communications struggle to address the dynam ic environments\, we then discuss fast adaptation for DL-based wireless co mmunications by using the few-shot learning (FSL) techniques. After identi fying the difference between fast adaptation in wireless communications an d traditional AI tasks\, we outline two design requirements for applying F SL techniques to wireless communications and provide a comprehensive discu ssion on FSL techniques in wireless communications that satisfy these two design requirements. In particular\, we emphasize the crucial role of doma in knowledge in achieving fast adaptation. At the end of this talk\, we hi ghlight several open issues for future research.</p>\n<p>&nbsp\;</p>\n<p>< strong>Biography</strong>: Geoffrey Ye Li\, FREng (Fellow of Royal Academy of Engineering) and Fellow of IEEE\, is a Chair Professor at Imperial Col lege London\, UK.&nbsp\; Before joining Imperial in 2020\, he was a Profes sor at Georgia Institute of Technology\, USA\, for 20 years and a Principa l Technical Staff Member with AT&amp\;T Labs &ndash\; Research (previous B ell Labs)\, USA\, for five years. He is the first to introduce deep learni ng to communications in 2016\, which has become a popular research area no w. He made fundamental contributions to orthogonal frequency division mult iplexing (OFDM) for wireless communications\, which made him win <em>2024 IEEE Eric E. Sumner Technical-Field Award</em>. He also won several awards from IEEE Signal Processing\, Vehicular Technology\, and Communications S ocieties\, including&nbsp\;<em>2019 IEEE ComSoc</em>&nbsp\;<em>Edwin Howar d Armstrong Achievement Award.</em></p> END:VEVENT END:VCALENDAR