BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Deep Gaussian Processes - Richard Turner\, Alex Navarro\, Thang Bu
 i
DTSTART:20150326T150000Z
DTEND:20150326T163000Z
UID:TALK58581@talks.cam.ac.uk
CONTACT:Rowan McAllister
DESCRIPTION:*Abstract*\n\nDeep Gaussian processes are a probabilistic appr
 oach to modelling data that employ a hierarchical structure with nonparame
 tric mappings between layers. This family of models is attractive as it ca
 n potentially discover layers of increasingly abstract representations of 
 data\, in much the same way as their deep parametric counterparts\, but it
  can also handle and propagate uncertainty in the hierarchy. Standard infe
 rence and learning schemes\, however\, are not analytically tractable for 
 this model class. In this talk\, we will review deep Gaussian processes an
 d their relationship to other models\, before describing a recently develo
 ped variational free energy treatment that sidesteps the analytical intrac
 tabilities and which provides approximate posteriors and a tractable lower
  bound to the marginal likelihood.\n\n*Recommended reading:*\n\nThe sugges
 t reading is Damianou and Lawrence 2013\, but for more detail on the flavo
 ur of variational inference used in deep GPs and related models\, see Tits
 ias and Lawrence 2010 and Damianou et al 2014.\n# Damianou\, Andreas\, and
  Neil Lawrence. "Deep Gaussian Processes.":http://jmlr.org/proceedings/pap
 ers/v31/damianou13a.pdf Proceedings of the Sixteenth International Confere
 nce on Artificial Intelligence and Statistics. 2013.\n# Titsias\, Michalis
  K.\, and Neil D. Lawrence. "Bayesian Gaussian Process Latent Variable Mod
 el.":http://jmlr.org/proceedings/papers/v9/titsias10a/titsias10a.pdf Inter
 national Conference on Artificial Intelligence and Statistics. 2010.\n# A.
  Damianou\, M. Titsias and N. Lawrence. "Variational Inference for Uncerta
 inty on the Inputs of Gaussian Process Models.":http://arxiv.org/abs/1409.
 2287 2014.
LOCATION:Engineering Department\, CBL Room 438
END:VEVENT
END:VCALENDAR