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:Machine Learning @ CUED
SUMMARY:Inference and Learning in the Anglican Probabilist
 ic Programming System - Jan-Willem van de Meent (O
 xford)
DTSTART;TZID=Europe/London:20160405T110000
DTEND;TZID=Europe/London:20160405T120000
UID:TALK65534AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/65534
DESCRIPTION:Probabilistic programming systems aim to accelerat
 e iterative development of machine learning approa
 ches by introducing an abstraction boundary: model
 s are defined using a domain-specific language\, a
 nd a back end implements generic inference methods
  for such programs. The aim of this research endea
 vor is to do for the domains of data science and a
 rtificial intelligence what compiler technologies 
 have done for software development: enable practit
 ioners to reason about their models at a higher le
 vel of abstraction.\n\nIn this talk I will discuss
  inference strategies employed in Anglican\, a pro
 babilistic programing system closely integrated wi
 th the language Clojure. Anglican has pioneered in
 ference techniques based on sequential Monte Carlo
  that apply to programs written in general-purpose
  languages that support recursion\, higher-order f
 unctions\, and black box deterministic primitives.
  In addition to strategies for posterior inference
 \, I will discuss extensions to policy search and 
 marginal MAP estimation problems.\n\nBIO \n\nJan-W
 illem is post-doc in Machine Learning at the Depar
 tment of Engineering Science at Oxford. He works p
 rimarily on the Anglican probabilistic programming
  system\, which he co-created with Frank Wood and 
 David Tolpin. His broader research agenda is to un
 derstand how programs may be used to define struct
 ured and composable models for machine learning an
 d artificial intelligence. To facilitate this agen
 da\, he also works on inference techniques for pro
 babilistic programs.
LOCATION:Engineering Department\, CBL Room BE-438
CONTACT:Zoubin Ghahramani
END:VEVENT
END:VCALENDAR