BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Bayesian Generative Adversarial Networks  - Andrew Wilson\, Cornel
 l University
DTSTART:20171212T110000Z
DTEND:20171212T120000Z
UID:TALK95533@talks.cam.ac.uk
CONTACT:Microsoft Research Cambridge Talks Admins
DESCRIPTION:Through an adversarial game\, generative adversarial networks 
 (GANs) can implicitly learn rich distributions over images\, audio\, and d
 ata which are hard to model with an explicit likelihood. I will present a 
 practical Bayesian formulation for unsupervised and semi-supervised learni
 ng with GANs. Within this framework\, we use a stochastic gradient Hamilto
 nian Monte Carlo for marginalizing parameters. The resulting approach can 
 automatically discover complementary and interpretable generative hypothes
 es for collections of images. Moreover\, by exploring an expressive poster
 ior over these hypotheses\, we show that it is possible to achieve state-o
 f-the-art quantitative results on image classification benchmarks\, even w
 ith less than 1% of the labelled training data. 
LOCATION:Auditorium \, Microsoft Research Ltd\, 21 Station Road\, Cambridg
 e\, CB1 2FB
END:VEVENT
END:VCALENDAR