BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Demystifying deep learning - Petar Veličković (University of Cam
 bridge)
DTSTART:20170517T130000Z
DTEND:20170517T140000Z
UID:TALK72571@talks.cam.ac.uk
CONTACT:Rengarajan
DESCRIPTION:Machine learning with deep neural networks (nowadays commonly 
 referred to as "deep learning") has brought about significant benefits acr
 oss a variety of domains (such as computer vision\, natural language proce
 ssing and game-playing) using essentially the same kind of learning algori
 thm. However\, in essence\, the bulk of what we know about deep learning r
 ight now has been established in the past century\, the innovations made b
 y large institutions are often hyperbolised by the popular media (for an e
 xcellent overview of related issues\, I highly recommend the following art
 icle: http://approximatelycorrect.com/2017/03/28/the-ai-misinformation-epi
 demic/)\, and the research landscape is difficult to navigate even for exp
 erts\, given the immense volume of papers consistently being released in t
 he area over the past few years. \n\nIn this talk\, I will channel the exp
 erience I've acquired in this area over the past few years to provide a co
 ncise\, formal and (hopefully!) exciting overview of where the field is ri
 ght now\, how it has gotten there\, and what are the likely next steps (wi
 th particular emphasis on the kinds of problems where this technique still
  notably struggles). This will be an applications-oriented talk\, accompan
 ied by a complementary overview of fundamental theory. Therefore\, the tal
 k is expected to be appropriate for people at all skill levels\, but it is
  expected that it will be mostly valuable for those that only vaguely inte
 racted with deep learning thus far. An entry-level grasp of linear algebra
  and probability will also be useful\, but not essential.
LOCATION:Seminar room 2\, Institute for Manufacturing\, Cambridge
END:VEVENT
END:VCALENDAR