BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Extracting spatiotemporal patterns from data with dynamics-adapted
  kernels - Giannakis\, D (New York University)
DTSTART:20140318T113000Z
DTEND:20140318T121000Z
UID:TALK51488@talks.cam.ac.uk
CONTACT:Mustapha Amrani
DESCRIPTION:Kernel methods provide an attractive way of extracting feature
 s from data by biasing the geometry of the data in a controlled manner. In
  this talk\, we discuss a family of kernels for dynamical systems featurin
 g an explicit dependence on the dynamical vector field operating in the ph
 ase-space manifold\, estimated empirically through finite differences of t
 ime-ordered data samples. In a suitable asymptotic limit\, the associated 
 diffusion operator generates diffusions along the integral curves of the d
 ynamical vector field. We present applications to toy dynamical systems an
 d data generated by comprehensive climate models.\n
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR