BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:A medley of geometry\, optimal transport\, and machine learning - 
 Varun Jog (Cambridge)
DTSTART:20211110T140000Z
DTEND:20211110T150000Z
UID:TALK164626@talks.cam.ac.uk
CONTACT:Hamza Fawzi
DESCRIPTION:Modern machine learning algorithms are surprisingly fragile to
  adversarial perturbations of data. In this talk\, we present some theoret
 ical contributions towards understanding fundamental bounds on the perform
 ance of machine learning algorithms in the presence of adversaries. We sha
 ll discuss how optimal transport emerges as a natural mathematical tool to
  characterize "robust risk"\, a notion of risk in the adversarial machine 
 learning literature analogous to Bayes risk in hypothesis testing. We shal
 l also show how\, in addition to tools from optimal transport\, we may use
  reverse-isoperimetric inequalities from geometry to provide theoretical b
 ounds on the sample size of estimating robust risk.\n\n*Join Zoom Meeting*
 \nhttps://maths-cam-ac-uk.zoom.us/j/93933342683?pwd=dUlHZGJYVEhiWXNIaGhsRD
 VkbmpPZz09\n\nMeeting ID: 939 3334 2683\nPasscode: U22EKPmS
LOCATION:Virtual (Zoom details under abstract)
END:VEVENT
END:VCALENDAR