BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Piecewise Deterministic Monte Carlo for latent variable models : a
  case study - Joris Bierkens (Delft University of Technology)
DTSTART:20241204T140000Z
DTEND:20241204T150000Z
UID:TALK224263@talks.cam.ac.uk
DESCRIPTION:Piecewise Deterministic Monte Carlo (PDMC) methods have been a
 round for some time now and our theoretical understanding has already grow
 n significantly in the last decade. However the application of these metho
 ds could do with a boost\; unfortunately this presents serious challenges 
 in real-world applications.\nOne of the attractive properties of PDMC is t
 he ability to use unbiased gradients of the target distribution. The appli
 cations of this property goes beyond subsampling of data. In this talk we 
 will discuss recent efforts to apply this properties in latent variable mo
 dels. As we will see\, this problem is not completely solved yet\, but its
  study leads to interesting insights.
LOCATION:Seminar Room 2\, Newton Institute
END:VEVENT
END:VCALENDAR