BEGIN:VCALENDAR
VERSION:2.0
PRODID:IEEE vTools.Events//EN
CALSCALE:GREGORIAN
BEGIN:VTIMEZONE
TZID:Asia/Shanghai
BEGIN:STANDARD
DTSTART:19910915T010000
TZOFFSETFROM:+0900
TZOFFSETTO:+0800
TZNAME:CST
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260617T095227Z
UID:97D080A5-381F-4114-95CE-5E34B3F3ABC6
DTSTART;TZID=Asia/Shanghai:20260624T150000
DTEND;TZID=Asia/Shanghai:20260624T170000
DESCRIPTION:this is a DLT academic events supported by IEEE comsoc and tong
 ji niveristy.\n\nAbstract: Large artificial intelligence model (LAIM) has 
 been tremendously developed in terms of capabilities and applications in r
 ecent years. Meanwhile\, 6G mobile technologies are in the process of fast
  development and standardization. 6G is widely believed to be AI-native. H
 owever\, the application of LAIM in wireless optimization is still very li
 mited\, and optimized deployment of LAIM in wireless networks is still lar
 gely unknown. We will first present integration of LAIM with traditional o
 ptimization algorithms to solve complex design problems in Integrated Sens
 ing and Communication (ISAC) networks. Then we will develop a framework of
  collaborative agent LLM in wireless networks. Our results show LAIM has g
 reat potentials for optimizing and design wireless networks. Then we will 
 discuss the optimized deployment of LAIM in wireless edge networks. In par
 ticular\, we will discuss a pruning-aware and quantization-based LAIM co-i
 nference scheme\, where a pre-trained LAIM is pruned or quantized and part
 itioned into on-device and on-server sub-models for deployment. We show th
 at the LAIM output distortion is upper bounded by its parameter distortion
 . Then a lower bound on parameter distortion via the rate-distortion theor
 y\, analytically capturing the relationship between pruning ratio and co-i
 nference performance is given. The analytic results enable the optimized d
 eployment of LAIM in wireless networks.\n\nCo-sponsored by: Prof. Chao Wan
 g\n\nSpeaker(s): Ming Xiao\n\nRoom: Room 703\, \, Bldg:  Zhixin Building\,
  Tongji University\,\, 4800 Cao'an Highway\, shanghai\, Shanghai\, China\,
  201804
LOCATION:Room: Room 703\, \, Bldg:  Zhixin Building\, Tongji University\,\,
  4800 Cao'an Highway\, shanghai\, Shanghai\, China\, 201804
ORGANIZER:lixinwan@sjtu.edu.cn
SEQUENCE:26
SUMMARY:Large AI Model Native Wireless Networks
URL;VALUE=URI:https://events.vtools.ieee.org/m/564090
X-ALT-DESC:Description: <br /><p>this is a DLT academic events supported by
  IEEE comsoc and tongji niveristy.&nbsp\;</p>\n<p><strong><span lang="EN-U
 S" style="font-size: 12.0pt\; font-family: 'Arial'\,sans-serif\; mso-farea
 st-font-family: '等线 Light'\; mso-fareast-theme-font: major-fareast\; m
 so-ansi-language: EN-US\; mso-fareast-language: ZH-CN\; mso-bidi-language:
  AR-SA\;">Abstract</span></strong><span lang="EN-US" style="font-size: 12.
 0pt\; font-family: 'Arial'\,sans-serif\; mso-fareast-font-family: '等线 
 Light'\; mso-fareast-theme-font: major-fareast\; mso-ansi-language: EN-US\
 ; mso-fareast-language: ZH-CN\; mso-bidi-language: AR-SA\;">: </span><span
  lang="EN-US" style="font-size: 12.0pt\; font-family: 'Arial'\,sans-serif\
 ; mso-fareast-font-family: 等线\; mso-fareast-theme-font: minor-fareast\
 ; color: black\; mso-ansi-language: EN-US\; mso-fareast-language: ZH-CN\; 
 mso-bidi-language: AR-SA\;">Large artificial intelligence model (LAIM) has
  been tremendously developed in terms of capabilities and applications in 
 recent years. Meanwhile\, 6G mobile technologies are in the process of fas
 t development and standardization. 6G is widely believed to be AI-native. 
 However\, the application of LAIM in wireless optimization is still very l
 imited\, and optimized deployment of LAIM in wireless networks is still la
 rgely unknown. We will first present integration of LAIM with traditional 
 optimization algorithms to solve complex design problems in Integrated Sen
 sing and Communication (ISAC) networks. &nbsp\;Then we will develop a fram
 ework of collaborative agent LLM in wireless networks. Our results show LA
 IM has great potentials for optimizing and design wireless networks. Then 
 we will discuss the optimized deployment of LAIM in wireless edge networks
 . In particular\, we will discuss a pruning-aware and quantization-based L
 AIM co-inference scheme\, where a pre-trained LAIM is pruned or quantized 
 and partitioned into on-device and on-server sub-models for deployment. We
  show that the LAIM output distortion is upper bounded by its parameter di
 stortion. Then a lower bound on parameter distortion via the rate-distorti
 on theory\, analytically capturing the relationship between pruning ratio 
 and co-inference performance is given. The analytic results enable the opt
 imized deployment of LAIM in wireless networks. </span></p>
END:VEVENT
END:VCALENDAR