BEGIN:VCALENDAR VERSION:2.0 PRODID:IEEE vTools.Events//EN CALSCALE:GREGORIAN BEGIN:VEVENT DTSTAMP:20240320T141637Z UID:C41A9337-6B67-48CC-8FA6-CFB53950B0E8 DTSTART;TZID=Etc/UTC:20240418T120000 DTEND;TZID=Etc/UTC:20240418T130000 DESCRIPTION:Special Presentation by Prof. Petar Popovski (Aalborg Universit y\, Denmark)\n\nHosted by the Future Networks Artificial Intelligence & Ma chine Learning (AIML) Working Group\n\nDate/Time: Thursday\, April 18th\, 2024 @ 12:00 UTC\n\nTopic:\n\nLearning and Intelligence over Weak Communic ation Links\n\nAbstract:\n\nBesides the fascinating questions on how to tr ain increasingly capable Machine Learning (ML) models and explain their be havior\, there is a suite of highly relevant challenges that emerge when M L models become elements of distributed connected systems and networks. A popular instance of this set of problems is federated learning. The first part of the talk will present a federated learning setup over LEO satellit e constellation. It will be seen that the predictability of satellite move ment can be used to speed up the training process. The second part will de al with a model for supervised learning in which Alice has access to abund ant data features but does not have the labels\, while Bob is able to prov ide a correct label for any data point. Alice is connected to Bob through a low-rate communication link and the talk will present strategies that co mbine active learning and data compression that enable Alice to get the la bels. Finally\, the third part of the talk discusses generative network la yer of communication protocols. This is implemented in an intermediate net work node that contains a Generative AI module. When the link to the sourc e is weak\, instead of waiting for packets to be routed\, the node can gen erate the packets that need to be sent to the destination. Generative netw ork layer is an early step towards the potential changes in communication protocols based on increasingly capable AI.\n\nCo-sponsored by: IEEE Futur e Networks\n\nSpeaker(s): Prof. Petar Popovski\n\nVirtual: https://events. vtools.ieee.org/m/413460 LOCATION:Virtual: https://events.vtools.ieee.org/m/413460 ORGANIZER:c.polk@comsoc.org SEQUENCE:11 SUMMARY:Learning and Intelligence over Weak Communication Links URL;VALUE=URI:https://events.vtools.ieee.org/m/413460 X-ALT-DESC:Description: <br /><p class="MsoNormal" style="margin-top: 12.0p t\;">Special Presentation by<strong> Prof. Petar Popovski (Aalborg Univers ity\, Denmark)</strong></p>\n<p class="MsoNormal" style="margin-top: 12.0p t\;">Hosted by the Future Networks<strong> Artificial Intelligence &amp\; Machine Learning (AIML) Working Group</strong></p>\n<p class="MsoNormal" s tyle="margin-top: 12.0pt\;"><strong><span style="font-size: 14.0pt\; font- family: Copperplate\; mso-fareast-font-family: PMingLiU\; mso-fareast-them e-font: minor-fareast\; mso-bidi-font-family: Arial\; mso-bidi-theme-font: minor-bidi\; mso-ansi-language: EN-US\; mso-fareast-language: ZH-TW\; mso -bidi-language: AR-SA\;">Date/Time</span><span style="font-size: 12.0pt\; font-family: 'Calibri'\,sans-serif\; mso-ascii-theme-font: minor-latin\; m so-fareast-font-family: PMingLiU\; mso-fareast-theme-font: minor-fareast\; mso-hansi-theme-font: minor-latin\; mso-bidi-font-family: Arial\; mso-bid i-theme-font: minor-bidi\; mso-ansi-language: EN-US\; mso-fareast-language : ZH-TW\; mso-bidi-language: AR-SA\;">: Thursday\, April 18<sup>th</sup>\, 2024 @ 12:00 UTC</span></strong></p>\n<p><img src="https://events.vtools. ieee.org/vtools_ui/media/display/db6c7d08-8291-49f7-ad2e-f0826406ec70" wid th="1035" height="271"></p>\n<p class="MsoNormal" style="margin-top: 12.0p t\;"><strong><u><span style="font-size: 16.0pt\; font-family: Copperplate\ ;">Topic</span></u></strong><strong><span style="font-size: 16.0pt\; font- family: Copperplate\;">:</span></strong></p>\n<p class="MsoNormal"><span s tyle="background-color: rgb(255\, 255\, 255)\; color: rgb(0\, 0\, 0)\;"><s trong><span style="font-size: 16pt\; background: rgb(255\, 255\, 255)\;">L earning and Intelligence over Weak Communication Links</span></strong></sp an></p>\n<p class="MsoNormal" style="margin-top: 12.0pt\;"><strong><u><spa n style="font-size: 16.0pt\; font-family: Copperplate\;">Abstract</span></ u></strong><strong><span style="font-size: 16.0pt\; font-family: Copperpla te\;">:</span></strong></p>\n<p class="MsoNormal" style="margin-top: 6.0pt \;">Besides the fascinating questions on how to train increasingly capable Machine Learning (ML) models and explain their behavior\, there is a suit e of highly relevant challenges that emerge when ML models become elements of distributed connected systems and networks. A popular instance of this set of problems is federated learning. The first part of the talk will pr esent a federated learning setup over LEO satellite constellation. It will be seen that the predictability of satellite movement can be used to spee d up the training process. The second part will deal with a model for supe rvised learning in which Alice has access to abundant data features but do es not have the labels\, while Bob is able to provide a correct label for any data point. Alice is connected to Bob through a low-rate communication link and the talk will present strategies that combine active learning an d data compression that enable Alice to get the labels. Finally\, the thir d part of the talk discusses generative network layer of communication pro tocols. This is implemented in an intermediate network node that contains a Generative AI module. When the link to the source is weak\, instead of w aiting for packets to be routed\, the node can generate the packets that n eed to be sent to the destination. Generative network layer is an early st ep towards the potential changes in communication protocols based on incre asingly capable AI.</p> END:VEVENT END:VCALENDAR