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:Microsoft Research Cambridge\, public talks
SUMMARY:Computational foundations of Bayesian inference an
d probabilistic programming - Daniel Roy\, Univers
ity of Cambridge
DTSTART;TZID=Europe/London:20140227T110000
DTEND;TZID=Europe/London:20140227T120000
UID:TALK51161AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/51161
DESCRIPTION:The complexity\, scale\, and variety of data sets
we now have access to have grown enormously\, and
present exciting opportunities for new application
s. Just as high-level programming languages and c
ompilers empowered experts to solve computational
problems more quickly\, and made it possible for n
on-experts to solve them at all\, a number of high
-level probabilistic programming languages with co
mputationally universal inference engines have bee
n developed with the potential to similarly transf
orm the practice of Bayesian statistics. These sy
stems provide formal languages for specifying prob
abilistic models compositionally\, and general alg
orithms for turning these specifications into effi
cient algorithms for inference.\n\nIn this talk\,
I will address three key questions at the theoreti
cal and algorithmic foundations of probabilistic p
rogramming---and probabilistic modelling more gene
rally---that can be answered using tools from prob
ability theory\, computability and complexity theo
ry\, and nonparametric Bayesian statistics. Which
Bayesian inference problems can be automated\, an
d which cannot? Can probabilistic programming lan
guages represent the stochastic processes at the c
ore of state-of-the-art nonparametric Bayesian mod
els? And if not\, can we construct useful approxi
mations? I’ll close by relating these questions t
o other challenges and opportunities ahead at the
intersections of computer science\, statistics\, a
nd probability.\n
LOCATION:Auditorium\, Microsoft Research Ltd\, 21 Station R
oad\, Cambridge\, CB1 2FB
CONTACT:Microsoft Research Cambridge Talks Admins
END:VEVENT
END:VCALENDAR