BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Defending Against Adversarial Attacks - Ross Clarke (University of
  Cambridge)
DTSTART:20181107T140000Z
DTEND:20181107T153000Z
UID:TALK114559@talks.cam.ac.uk
CONTACT:75379
DESCRIPTION:Adversarial examples are inputs which have been maliciously pe
 rturbed to induce inappropriate responses from a machine learning system\,
  but which are generally indistinguishable from innocent inputs by humans.
  They thus represent a substantial threat to the reliability and practicab
 ility of ML applications\, as systems vulnerable to manipulation in this w
 ay cannot be trusted with important decisions. Despite this\, surprisingly
  little is understood about the mechanisms by which adversarial examples a
 rise\, and how we might construct systems which are resilient to attack by
  these samples. We chart the evolution of the literature on adversarial at
 tacks by considering some initially proposed explanations for how they ari
 se. We discuss some defence mechanisms such as adversarial training and th
 e less obvious approach of network distillation. We then briefly summarise
  the current state of the field.
LOCATION:Engineering Department\, CBL Room 438
END:VEVENT
END:VCALENDAR