BEGIN:VCALENDAR VERSION:2.0 PRODID:IEEE vTools.Events//EN CALSCALE:GREGORIAN BEGIN:VTIMEZONE TZID:Europe/Dublin BEGIN:DAYLIGHT DTSTART:20260329T020000 TZOFFSETFROM:+0000 TZOFFSETTO:+0100 RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3 TZNAME:IST END:DAYLIGHT BEGIN:STANDARD DTSTART:20251026T010000 TZOFFSETFROM:+0100 TZOFFSETTO:+0000 RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10 TZNAME:GMT END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20260114T222119Z UID:77694EDA-82E3-4094-95FF-7431EF4FFE52 DTSTART;TZID=Europe/Dublin:20251118T150000 DTEND;TZID=Europe/Dublin:20251118T160000 DESCRIPTION:The IEEE Signal Processing Society UK & Ireland Chapter is plea sed to welcome Professor Urbashi Mitra to the University College Dublin fo r a Distinguished Lecturer seminar.\n\nTitle: Exploiting Statistical Hardn ess for Increased Privacy in Wireless Systems\n\nAbstract: Securing signal s from unintended eavesdroppers has become an increasingly important probl em with the emergence of the Internet-of-Things. Herein\, we examine learn ing problems in signal processing that are inherently hard without key sid e information. In particular\, we exploit necessary resolution limits for classical compressed sensing problems. To limit an eavesdropper's capabili ties\, we create an environment for the eavesdropper wherein the appropria te compressed sensing algorithm would probably fail. The intended receiver overcomes this ill-posed problem by leveraging secret side information sh ared between the intended transmitter and receiver. Two scenarios are cons idered: one for communication over a wireless channel where a novel block- sparsity based signaling strategy is employed and one for localization whe re novel structured noise is introduced to degrade the form of the eavesdr opper’s channel. In the latter scenario\, the transmitter designs a beam former that introduces spurious paths\, or spoofs the line-of-sight path\, in the channel without having access to the channel state information. In both private communication and private localization\, the amount of secre t information that must be shared is very modest. Theoretical guarantees c an be provided for both cases. Proposed algorithms are validated via numer ical results. Finally\, the idea of statistical hardness is applied to a d istributed learning (federated) problem and shown to be effective in provi ding model privacy against eavesdroppers.\n\nSpeaker(s): Urbashi Mitra \, \n\nRoom 204\, UCD Engineering & Materials Science Centre\, Dublin\, Dubli n\, Ireland\, 8Q4J+HJ LOCATION:Room 204\, UCD Engineering & Materials Science Centre\, Dublin\, D ublin\, Ireland\, 8Q4J+HJ ORGANIZER:g.katsaros@surrey.ac.uk SEQUENCE:11 SUMMARY:IEEE SPS UK & Ireland Distinguished Lecturer Seminar at UCD: Prof. Urbashi Mitra\, Exploiting Statistical Hardness for Increased Privacy in W ireless Systems URL;VALUE=URI:https://events.vtools.ieee.org/m/515303 X-ALT-DESC:Description: <br /><p class="MsoNormal"><strong><span style="col or: black\;">The IEEE Signal Processing Society UK &amp\; Ireland Chapter is pleased to welcome Professor Urbashi Mitra to the University College Du blin for a Distinguished Lecturer seminar.</span></strong></p>\n<p class=" MsoNormal"><strong><span style="color: black\;">Title: Exploiting Statisti cal Hardness for Increased Privacy in Wireless Systems</span></strong></p> \n<p class="MsoNormal" style="text-align: justify\;"><strong><span style=" color: black\;">Abstract:</span></strong><span style="color: black\;">&nbs p\;Securing signals from unintended eavesdroppers has become an increasing ly important problem with the emergence of the Internet-of-Things. Herein\ , we examine learning problems in signal processing that are inherently ha rd without key side information. In particular\, we exploit necessary reso lution limits for classical compressed sensing problems. To limit an eaves dropper's capabilities\, we create an environment for the eavesdropper whe rein the appropriate compressed sensing algorithm would probably fail. The intended receiver overcomes this ill-posed problem by leveraging secret s ide information shared between the intended transmitter and receiver. Two scenarios are considered: one for communication over a wireless channel wh ere a novel block-sparsity based signaling strategy is employed and one fo r localization where novel structured noise is introduced to degrade the f orm of the eavesdropper&rsquo\;s channel. In the latter scenario\, the tra nsmitter designs a beamformer that introduces spurious paths\, or spoofs t he line-of-sight path\, in the channel without having access to the channe l state information. In both private communication and private localizatio n\, the amount of secret information that must be shared is very modest. T heoretical guarantees can be provided for both cases. Proposed algorithms are validated via numerical results.&nbsp\;Finally\, the idea of statistic al hardness is applied to a distributed learning (federated) problem and s hown to be effective in providing&nbsp\;<em>model privacy</em>&nbsp\;again st eavesdroppers.</span></p>\n<p class="MsoNormal" style="text-align: just ify\;"><span style="color: black\;">&nbsp\;</span></p> END:VEVENT END:VCALENDAR