BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Gamma Chains priors for audio processing - Onur Dikmen\, Bogazici 
 University
DTSTART:20071126T113000Z
DTEND:20071126T123000Z
UID:TALK9283@talks.cam.ac.uk
CONTACT:Taylan Cemgil
DESCRIPTION:Useful models for nonstationary audio sources can be obtained 
 by modelling the dependencies in the time-frequency representations of sig
 nals. Gamma Markov chains enable us to introduce such dependencies. We use
 d IGMCs to model the sources in denoising and single channel source separa
 tion applications. We will discuss the pros and cons of various inference 
 schemata such as Gibbs sampling\, variational Bayes and sequential Monte C
 arlo. 
LOCATION:Engineering Department\, Signal Processing Lab meeting room
END:VEVENT
END:VCALENDAR