BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Constructing expressive models using nonparametric Gaussian proces
 s convolutions. - Magnus Ross\, University of Manchester
DTSTART:20230525T140000Z
DTEND:20230525T150000Z
UID:TALK201241@talks.cam.ac.uk
CONTACT:Carl Henrik Ek
DESCRIPTION:Gaussian processes (GPs) are powerful models for probabilistic
  machine learning\, but good performance requires the selection of an appr
 opriate covariance function. Process convolutions are framework for the co
 nstruction of flexible covariances for both scalar and vector valued GPs\,
  but still require choosing a smoothing kernel\, which is non-trivial. Pre
 vious approaches have built covariance functions by using GP priors over t
 he smoothing kernel\, and by extension the covariance\, as a way to bypass
  the need to specify it in advance. However\, such models have been limite
 d in several ways: they are restricted to single dimensional inputs\, e.g.
  time\; they only allow modeling of single outputs and they do not scale t
 o large datasets since inference is not straightforward. In this talk I wi
 ll discuss the development of sampling and inference methods for nonparame
 tric process convolutions that resolve these limitations\, and allow covar
 iances to be inferred from data in a large variety of domains. Using these
  methods\, I'll discuss some extensions of process convolutions framework 
 to the case of a nonlinear operators via the Volterra series\, which resul
 t in non-Gaussian output processes\, and allow for the development of powe
 rful models for system identification. I will end by discussing the limita
 tions of the framework\, some ideas for future work. \n\nBio\n---\nMagnus 
 Ross is a third-year PhD student at the University of Manchester (and prev
 iously the University of Sheffield)\, working with Mauricio Álvarez and M
 ichael Smith. His research interests center around Gaussian processes\, an
 d in particular their use for physics informed machine learning.
LOCATION:Computer Lab\, SW00
END:VEVENT
END:VCALENDAR