Particle Smoothing in Continuous Time: A Fast Approach via Density Estimation

We consider the particle smoothing problem for state-space models where the transition density is not available in closed form, in particular for continuous-time, nonlinear models expressed via stochastic differential equations (SDEs). Conventional forward-backward and two-filter smoothers for the particle filter require a closed-form transition density, with the linear-Gaussian Euler-Maruyama discretization usually applied to the SDEs to achieve this. We develop a pair of variants using kernel density approximations to relieve the dependence, and in doing so enable use of faster and more accurate discretization schemes such as Runge-Kutta. In addition, the new methods admit arbitrary proposal distributions, providing an avenue to deal with degeneracy issues. Experimental results on a functional mag- netic resonance imaging (fMRI) deconvolution task demonstrate comparable accuracy and significantly improved runtime over conventional techniques.

[1]  R. E. Kalman,et al.  A New Approach to Linear Filtering and Prediction Problems , 2002 .

[2]  E. Fehlberg,et al.  Low-order classical Runge-Kutta formulas with stepsize control and their application to some heat transfer problems , 1969 .

[3]  P. Kloeden,et al.  Numerical Solution of Stochastic Differential Equations , 1992 .

[4]  J. Dormand,et al.  A family of embedded Runge-Kutta formulae , 1980 .

[5]  C. W. Gardiner,et al.  Handbook of stochastic methods - for physics, chemistry and the natural sciences, Second Edition , 1986, Springer series in synergetics.

[6]  B. Silverman.  Density estimation for statistics and data analysis , 1986 .

[7]  Numerical integration of stochastic differential equations , 1988 .

[8]  D. Tank,et al.  Brain magnetic resonance imaging with contrast dependent on blood oxygenation. , 1990, Proceedings of the National Academy of Sciences of the United States of America.

[9]  N. Gordon,et al.  Novel approach to nonlinear/non-Gaussian Bayesian state estimation , 1993 .

[10]  E. Hairer,et al.  Solving ordinary differential equations I (2nd revised. ed.): nonstiff problems , 1993 .

[11]  Jun S. Liu,et al.  Blind Deconvolution via Sequential Imputations , 1995 .

[12]  G. Kitagawa.  Monte Carlo Filter and Smoother for Non-Gaussian Nonlinear State Space Models , 1996 .

[13]  Karl J. Friston,et al.  Human Brain Function , 1997 .

[14]  R. Buxton,et al.  Dynamics of blood flow and oxygenation changes during brain activation: The balloon model , 1998, Magnetic resonance in medicine.

[15]  Michael Isard,et al.  A Smoothing Filter for CONDENSATION , 1998, ECCV.

[16]  B. Rosen,et al.  Evidence of a Cerebrovascular Postarteriole Windkessel with Delayed Compliance , 1999, Journal of cerebral blood flow and metabolism : official journal of the International Society of Cerebral Blood Flow and Metabolism.

[17]  M. Pitt,et al.  Filtering via Simulation: Auxiliary Particle Filters , 1999 .

[18]  Michael I. Jordan,et al.  Learning with Mixtures of Trees , 2001, J. Mach. Learn. Res..

[19]  Nando de Freitas,et al.  The Unscented Particle Filter , 2000, NIPS.

[20]  Andrew W. Moore,et al.  'N-Body' Problems in Statistical Learning , 2000, NIPS.

[21]  Karl J. Friston,et al.  Nonlinear Responses in fMRI: The Balloon Model, Volterra Kernels, and Other Hemodynamics , 2000, NeuroImage.

[22]  Neil J. Gordon,et al.  Editors: Sequential Monte Carlo Methods in Practice , 2001 .

[23]  Nando de Freitas,et al.  Sequential Monte Carlo Methods in Practice , 2001, Statistics for Engineering and Information Science.

[24]  Karl J. Friston,et al.  Modeling regional and psychophysiologic interactions in fMRI: the importance of hemodynamic deconvolution , 2003, NeuroImage.

[25]  Richard N. Henson,et al.  Introduction to Functional Magnetic Resonance Imaging: Principles and Techniques , 2002 .

[26]  Karl J. Friston,et al.  Dynamic causal modelling , 2003, NeuroImage.

[27]  Joshua Wilkie.  Numerical methods for stochastic differential equations. , 2004, Physical review. E, Statistical, nonlinear, and soft matter physics.

[28]  Nando de Freitas,et al.  Fast particle smoothing: if I had a million particles , 2006, ICML.

[29]  Darren J. Wilkinson,et al.  Bayesian sequential inference for nonlinear multivariate diffusions , 2006, Stat. Comput..

[30]  P. Fearnhead,et al.  Exact and computationally efficient likelihood‐based estimation for discretely observed diffusion processes (with discussion) , 2006 .

[31]  P. Fearnhead,et al.  Particle filters for partially observed diffusions , 2007, 0710.4245.

[32]  Amos J. Storkey,et al.  Continuous Time Particle Filtering for fMRI , 2007, NIPS.

[33]  Dan Cornford,et al.  Variational Inference for Diffusion Processes , 2007, NIPS.

[34]  A. Doucet,et al.  Smoothing algorithms for state–space models , 2010 .

[35]  Karl J. Friston,et al.  Dynamic causal modeling , 2010, Scholarpedia.

[36]  P. Fearnhead,et al.  A sequential smoothing algorithm with linear computational cost. , 2010 .