BEGIN:VCALENDAR VERSION:2.0 PRODID:IEEE vTools.Events//EN CALSCALE:GREGORIAN BEGIN:VTIMEZONE TZID:Europe/Skopje BEGIN:DAYLIGHT DTSTART:20240331T030000 TZOFFSETFROM:+0100 TZOFFSETTO:+0200 RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3 TZNAME:CEST END:DAYLIGHT BEGIN:STANDARD DTSTART:20241027T020000 TZOFFSETFROM:+0200 TZOFFSETTO:+0100 RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10 TZNAME:CET END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20250310T194042Z UID:78FCF94C-CF72-474D-A663-98E73CD0C172 DTSTART;TZID=Europe/Skopje:20240426T120000 DTEND;TZID=Europe/Skopje:20240426T130000 DESCRIPTION:The centralization of data and intelligence poses several risks \, including potential privacy breaches\, single points of failure\, and s tifling innovation. Decentralization\, on the other hand\, remains challen ging as edge devices often have to work with limited computing capabilitie s and battery energy budgets. In this talk\, I will present a line of rese arch conducted at the Faculty of Computer and Information Science\, Univer sity of Ljubljana\, Slovenia\, where we aim to reduce the computational an d energy burden of deep learning\, thus enabling AI on the edge. The basis of our approach is dynamically approximable computing\, which allows us t o adapt the amount of computation to the actual situation in which deep le arning occurs. The talk will demonstrate this philosophy applied to differ ent domains\, from physical activity recognition on smartphones to weed id entification on a camera-equipped unmanned aerial vehicle (UAV).\n\nSpeake r(s): Veljko Pejovic\, PhD\n\nRoom: Conference Room\, University Ss Cyril and Methodius\, Faculty of Electrical Engineering and Information Technolo gies\, Rugjer Boskovikj 18\, Skopje\, Macedonia\, Macedonia LOCATION:Room: Conference Room\, University Ss Cyril and Methodius\, Facult y of Electrical Engineering and Information Technologies\, Rugjer Boskovik j 18\, Skopje\, Macedonia\, Macedonia ORGANIZER:dushko.stavrov@feit.ukim.edu.mk SEQUENCE:30 SUMMARY:RoboMac 2024 Lecture - Decentralizing AI through Resource-Efficient Mobile Deep Learning URL;VALUE=URI:https://events.vtools.ieee.org/m/418356 X-ALT-DESC:Description: <br /><p style="text-align: justify\;">The centrali zation of data and intelligence poses several risks\, including potential privacy breaches\, single points of failure\, and stifling innovation. Dec entralization\, on the other hand\, remains challenging as edge devices of ten have to work with limited computing capabilities and battery energy bu dgets. In this talk\, I will present a line of research conducted at the F aculty of Computer and Information Science\, University of Ljubljana\, Slo venia\, where we aim to reduce the computational and energy burden of deep learning\, thus enabling AI on the edge. The basis of our approach is dyn amically approximable computing\, which allows us to adapt the amount of c omputation to the actual situation in which deep learning occurs. The talk will demonstrate this philosophy applied to different domains\, from phys ical activity recognition on smartphones to weed identification on a camer a-equipped unmanned aerial vehicle (UAV).&nbsp\;</p> END:VEVENT END:VCALENDAR