BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//talks.cam.ac.uk//v3//EN
BEGIN:VTIMEZONE
TZID:Europe/London
BEGIN:DAYLIGHT
TZOFFSETFROM:+0000
TZOFFSETTO:+0100
TZNAME:BST
DTSTART:19700329T010000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0100
TZOFFSETTO:+0000
TZNAME:GMT
DTSTART:19701025T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
CATEGORIES:Computer Laboratory Security Seminar
SUMMARY:Towards Meaningful Stochastic Defences in Machine 
 Learning - Ilia Shumailov\, University of Oxford
DTSTART;TZID=Europe/London:20221115T140000
DTEND;TZID=Europe/London:20221115T150000
UID:TALK184949AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/184949
DESCRIPTION:Machine learning (ML) has proven to be more fragil
 e than previously thought\, especially in adversar
 ial settings. A capable adversary can cause ML sys
 tems to break at training\, inference\, and deploy
 ment stages. In this talk\, I will cover the recen
 t work on attacking and defending machine learning
  pipelines using stochastic defences\; I will desc
 ribe how\, seemingly powerful defences fail to pro
 vide any security and end up being vulnerable to e
 ven standard attackers. I will then demonstrate a 
 number of possible randomness-based defences that 
 can provide theoretical and practical performance 
 improvements.\n\nBio: Ilia Shumailov holds a PhD i
 n Computer Science from University of Cambridge\, 
 specialising in Machine Learning and Computer Secu
 rity. During the PhD under the supervision of Prof
  Ross Anderson Ilia has worked on a number of proj
 ects spanning the fields of machine learning secur
 ity\, cybercrime analysis and signal processing. F
 ollowing the PhD\, Ilia joined Vector Institute in
  Canada as a Postdoctoral Fellow\, where he worked
  under the supervision of Prof Nicolas Papernot an
 d Prof Kassem Fawaz. Ilia is currently a Junior Re
 search Fellow at Christ Church\, University of Oxf
 ord. 
LOCATION:Webinar &amp\; FW11\, Computer Laboratory\, Willia
 m Gates Building.
CONTACT:Kieron Ivy Turk
END:VEVENT
END:VCALENDAR
