Variational mean-field algorithm for efficient inference in large systems of stochastic differential equations

Article English OPEN
Vrettas, Michail D. ; Opper, Manfred ; Cornford, Dan (2015)

This work introduces a Gaussian variational mean-field approximation for inference in dynamical systems which can be modeled by ordinary stochastic differential equations. This new approach allows one to express the variational free energy as a functional of the marginal moments of the approximating Gaussian process. A restriction of the moment equations to piecewise polynomial functions, over time, dramatically reduces the complexity of approximate inference for stochastic differential equation models and makes it comparable to that of discrete time hidden Markov models. The algorithm is demonstrated on state and parameter estimation for nonlinear problems with up to 1000 dimensional state vectors and compares the results empirically with various well-known inference methodologies.
  • References (34)
    34 references, page 1 of 4

    Mid-points 1st interval Mid-points 2nd interval 1C. W. Gardiner, Handbook of Stochastic Methods: for physics, chemistry and the natural sciences, 3rd ed. (Springer Series in Synergetics, 2003).

    2A. Golightly and D. J. Wilkinson, Computational Statistics and Data analysis 52, 1674 (2007).

    3E. Kalnay, Atmospheric Modeling, Data Assimilation and Predictability (Cambridge University Press, 2003).

    4M. Opper and D. Saad, eds., Advanced Mean Field Methods: Theory and Practice (MIT Press, Cambridge, MA, 2001).

    5F. J. Alexander, G. Eyink, and J. Restrepo, Journal of Statistical Physics 119, 1331 (2005).

    6P. Fearnhead, O. Papaspiliopoulos, and G. O. Roberts, Journal of the Royal Statistical Society 70, 755 (2008).

    7A. S. Paul and E. A. Wan, in Acoustics, Speech and Signal Processing, 2008. ICASSP 2008. (2008) pp. 3621{3624.

    8Y. Tremolet, Quarterly Journal of the Royal Meteorological Society 132, 2483 (2006).

    9G. L. Eyink, J. M. Restrepo, and F. J. Alexander, Physica D: Nonlinear Phenomena 195, 347 (2004).

    10M. Gilorami and B. Calderhead, Journal of Royal Statistical Society - B 73, 123 (2011).

  • Metrics
    No metrics available
Share - Bookmark