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:Modern Deep Learning through Bayesian Eyes - Yarin
Gal\, University of Cambridge
DTSTART;TZID=Europe/London:20151126T130000
DTEND;TZID=Europe/London:20151126T140000
UID:TALK62720AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/62720
DESCRIPTION:Bayesian models are rooted in Bayesian statistics\
, and easily benefit from the vast literature in t
he field. In contrast\, deep learning lacks a soli
d mathematical grounding. Instead\, empirical deve
lopments in deep learning are often justified by m
etaphors\, evading the unexplained principles at p
lay. These two fields are perceived as fairly anti
podal to each other in their respective communitie
s. It is perhaps astonishing then that most modern
deep learning models can be cast as performing ap
proximate inference in a Bayesian setting. The imp
lications of this statement are profound: we can u
se the rich Bayesian statistics literature with de
ep learning models\, explain away many of the curi
osities with these\, combine results from deep lea
rning into Bayesian modelling\, and much more. \nI
n this talk I will explore the new theory linking
Bayesian modelling and deep learning. The practica
l impact of the framework will be demonstrated wit
h a range of real-world applications: from uncerta
inty modelling in deep learning\, through training
on small datasets\, to new state-of-the-art resul
ts in image processing. I will finish by surveying
open problems to research\, problems which stand
at the forefront of a new and exciting field combi
ning modern deep learning and Bayesian techniques.
\n
LOCATION:Auditorium\, Microsoft Research Ltd\, 21 Station R
oad\, Cambridge\, CB1 2FB
CONTACT:Microsoft Research Cambridge Talks Admins
END:VEVENT
END:VCALENDAR