Details of Grant 

EPSRC Reference:	GR/R61857/01
Title:	Sequential, sparse Gaussian processes for data assimilation
Principal Investigator:	Cornford, Dr D
Other Investigators:	Opper, Professor M
Researcher Co-Investigators:	Professor P Challenor Professor M Niranjan Professor C Williams
Project Partners:
Department:	Sch of Engineering and Applied Science
Organisation:	Aston University
Scheme:	Fast Stream
Starts:	01 May 2002	Ends:	30 April 2003	Value (£):	45,604
EPSRC Research Topic Classifications:	Artificial Intelligence Information & Knowledge Mgmt
EPSRC Industrial Sector Classifications:	Environment
Related Grants:
Panel History:
Summary on Grant Application Form
In current work, to be presented at the ICANN conference in August 2001 the investigators demonstrate the application of a sequential, sparse Gaussian process learning algorithm to the assimilation of satellite scatterometer observations for wind field retrieval. This initial investigation has revealed two major obstacles to more widespread application of these sequential, sparse methods in data assimilation. First, the method must be adapted to use the more generally available forward models (which map the state vector to the observations) rather than the direct inverse model currently used. This will require the evaluation of a non-trivial integral which will be optimised in the project. Secondly, the estimation of the approximation to the full posterior distribution will be improved using expectation propagation based ideas which allow for data re-use, and thus an iterative improvement by recycling the observations. This will reduce the importance of the order of insertion of observations and be of particular benefit where the posterior distribution has multiple modes, such as in wind field retrieval. These improvements to the sequential, sparse Gaussian process algorithm will demonstrate the applicability of this framework to general data assimilation problems. By providing carefully written software we will make these tools available to a wide community. The research will also improve the retrieval of wind fields from scatterometer observations, itself important for numerical weather prediction.
Key Findings
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk
Potential use in non-academic contexts
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk
Impacts
Description This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk Summary Date Materialised
Sectors submitted by the Researcher
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk
Project URL:
Further Information:
Organisation Website:	http://www.aston.ac.uk