BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:MSR-Lecture: Gaussian Processes for Pattern Discovery\, Speaker: A
 ndrew Wilson - Andrew Wilson\, Cambridge Universtiy
DTSTART:20130410T090000Z
DTEND:20130410T100000Z
UID:TALK44535@talks.cam.ac.uk
CONTACT:32739
DESCRIPTION:Bayesian nonparametrics has the power to capture the infinite 
 complexity in the real world\, and could be used to explain how biological
  intelligence is capable of making extraordinary generalisations from a sm
 all number of examples. Gaussian processes are rich distributions over fun
 ctions\, which provide a Bayesian nonparametric approach to smoothing and 
 interpolation. When the large nonparametric support of a Gaussian process 
 is combined with the ability to automatically discover patterns in data an
 d extrapolate\, we are a step closer to developing truly intelligent agent
 s\, with applications in essentially any learning and prediction task.\n\n
 In this talk\, I derive simple closed form kernels that can be used with G
 aussian processes\, and other kernel machines\, to discover patterns and e
 nable extrapolation.  These kernels are derived by modelling a spectral de
 nsity -- the Fourier transform of a kernel -- with a Gaussian mixture. The
  proposed kernels support a broad class of stationary covariances\, but Ga
 ussian process inference remains simple and analytic.  I demonstrate the p
 roposed kernels by discovering patterns and performing long range extrapol
 ation on synthetic examples\, as well as atmospheric CO2 trends and airlin
 e passenger data.  I also show that we can reconstruct standard covariance
 s within the proposed framework.\n\nThis is joint work with Ryan P. Adams.
   A pre-print is available at http://arxiv.org/pdf/1302.4245v2\, and throu
 gh my website\, http://mlg.eng.cam.ac.uk/andrew\n
LOCATION:Small Lecture Theatre\, Microsoft Research Ltd\, 21 Station Road\
 , Cambridge\, CB1 2FB
END:VEVENT
END:VCALENDAR