BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Variational inference for some models with Polya-Gamma latent vari
 ables and Gaussian process priors - Manfred Opper\, TU Berlin
DTSTART:20171213T133000Z
DTEND:20171213T143000Z
UID:TALK96934@talks.cam.ac.uk
CONTACT:Dr R.E. Turner
DESCRIPTION:Polson et al [1] have shown that the logistic sigmoidal functi
 on can be represented as a mixture of Gaussians with the Polya-Gamma (PG) 
 density as the mixture distribution. This PG augmentation has attracted co
 nsiderable interest in the machine learning community. I will discuss a si
 mple variational inference approximation for such models with Gaussian (pr
 ocess) priors and discuss applications to classification\, Poisson process
 es and continuous time Ising models. \n\n[1] N G Polson\, J G Scott and J 
 Windle: Bayesian inference for logistic models using Polya-Gamma latent va
 riables\; J. Am Stat. Ass. (2015)
LOCATION:Engineering Department\, CBL Room BE-438.
END:VEVENT
END:VCALENDAR