Data Assimilation with Sequential Gaussian Processes

We study a data assimilation problem using Gaussian processes (GPs) where the GPs act as latent variables for the observations. Inference is done using a convenient parameterisation and sequential learning for a faster algorithm. We are addressing the disadvantage of the GPs, namely the quadratic scaling of the parameters with data and eliminate the scaling by using a fixed number of parameters. The result is a sparse representation that allows us to treat problems with a large number of observations. We apply our method to the prediction of wind fields over the ocean surface from scatterometer data.

[1]  Dan Cornford,et al.  Online Approximations for Wind-Field Models , 2001, ICANN.

[2]  Michael E. Tipping.  The Relevance Vector Machine , 1999, NIPS.

[3]  David Saad,et al.  On-Line Learning in Neural Networks , 1999 .

[4]  Carl E. Rasmussen,et al.  In Advances in Neural Information Processing Systems , 2011 .

[5]  Dan Cornford,et al.  Bayesian inference for wind field retrieval , 2000, Neurocomputing.

[6]  Tom Minka,et al.  Expectation Propagation for approximate Bayesian inference , 2001, UAI.

[7]  David R. Wolf.  A Bayesian Reflection on Surfaces , 1999, Entropy.

[8]  Nadia Pinardi,et al.  Ocean forecasting : conceptual basis and applications , 2002 .

[9]  G. Wahba,et al.  Some results on Tchebycheffian spline functions , 1971 .

[10]  Andrew C. Lorenc,et al.  Analysis methods for numerical weather prediction , 1986 .

[11]  S. Haykin.  Kalman Filtering and Neural Networks , 2001 .

[12]  Vladimir N. Vapnik,et al.  The Nature of Statistical Learning Theory , 2000, Statistics for Engineering and Information Science.

[14]  D. Offiler.  The Calibration of ERS-1 Satellite Scatterometer Winds , 1994 .

[15]  Alexander J. Smola,et al.  Advances in Large Margin Classifiers , 2000 .

[16]  Christopher M. Bishop,et al.  Neural networks for pattern recognition , 1995 .

[17]  Lehel Csató,et al.  Sparse On-Line Gaussian Processes , 2002, Neural Computation.

[18]  Dan Cornford,et al.  Structured neural network modelling of multi-valued functions for wind vector retrieval from satellite scatterometer measurements , 2000, Neurocomputing.

[19]  Ad Stoffelen,et al.  Ambiguity removal and assimilation of scatterometer data , 1997 .

[20]  Philippe Courtier,et al.  Unified Notation for Data Assimilation : Operational, Sequential and Variational , 1997 .

[21]  R. Daley.  Atmospheric Data Analysis , 1991 .