BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Learning the structure of graphical models with latent variables -
  Zoubin Ghahramani (University of Cambridge)
DTSTART:20100315T160000Z
DTEND:20100315T170000Z
UID:TALK21624@talks.cam.ac.uk
CONTACT:Florian Markowetz
DESCRIPTION:I will describe our work on the problem of learning the struct
 ure of  \nprobabilistic graphical models from data with hidden or missing 
  \nvariables.  This general machine learning problem is applicable to  \ng
 ene regulatory network inference\, which I will touch upon briefly. In  \n
 particular I will review work in our group on (i) variational Bayesian  \n
 learning of graph structures\, (ii) inference of gene regulatory  \nnetwor
 ks from state-space models of  time series data\, (iii) how to  \ninfer th
 e number of latent variables\, and (iv) Bayesian inference in  \ndirected 
 mixed graphs.
LOCATION:Cancer Research UK Cambridge Research Institute\, Lecture Theatre
END:VEVENT
END:VCALENDAR