BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//University of Liverpool Computer Science Seminar System//v2//EN
BEGIN:VEVENT
DTSTAMP:20260920T141617Z
UID:Seminar-dept-286@lxserverM.csc.liv.ac.uk
ORGANIZER:CN=Lutz Oettershagen:MAILTO:Lutz.Oettershagen@liverpool.ac.uk
DTSTART:20120515T110000
DTEND:20120515T120000
SUMMARY:School Seminar Series
DESCRIPTION:Prof. Neil Lawrence: Latent Force Models: Bridging the Divide between Mechanistic and Data Modelling Paradigms\n\nPhysics based approaches to data modeling involve\n\nconstructing an accurate mechanistic model of data,\n\noften based on differential equations. Machine\n\nlearning and statistical approaches are typically data\n\ndriven---perhaps through regularized function\n\napproximation.   \n\n \n\nThese two approaches to data modeling are often seen\n\nas polar opposites, but in reality they are two\n\ndifferent ends to a spectrum of approaches we might\n\ntake.            \n\n \n\nIn this talk we introduce latent force models. Latent\n\nforce models are a new approach to data \n\nrepresentation that model data through unknown\n\nforcing functions that drive differential equation\n\nmodels.  By treating the unknown forcing functions\n\nwith Gaussian process priors we can create\n\nprobabilistic models that exhibit particular physical\n\ncharacteristics of interest, for example, in\n\ndynamical systems resonance and inertia. This allows\n\nus to perform a synthesis of the data driven and\n\nphysical modeling paradigms.\n\nhttps://www.csc.liv.ac.uk/research/seminars/abstract.php?id=286
LOCATION:Ashton Lecture Theatre
END:VEVENT
END:VCALENDAR