Adaptive sequential Monte Carlo by means of mixture of experts

Appropriately designing the proposal kernel of particle filters is an issue of significant importance, since a bad choice may lead to deterioration of the particle sample and, consequently, waste of computational power. In this paper we introduce a novel algorithm adaptively approximating the so-called optimal proposal kernel by a mixture of integrated curved exponential distributions with logistic weights. This family of distributions, referred to as mixtures of experts, is broad enough to be used in the presence of multi-modality or strongly skewed distributions. The mixtures are fitted, via online-EM methods, to the optimal kernel through minimisation of the Kullback-Leibler divergence between the auxiliary target and instrumental distributions of the particle filter. At each iteration of the particle filter, the algorithm is required to solve only a single optimisation problem for the whole particle sample, yielding an algorithm with only linear complexity. In addition, we illustrate in a simulation study how the method can be successfully applied to optimal filtering in nonlinear state-space models.

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

[2]  P. Robins,et al.  Realtime sequential inference of static parameters with expensive likelihood calculations , 2009 .

[3]  D. Rubin,et al.  ML ESTIMATION OF THE t DISTRIBUTION USING EM AND ITS EXTENSIONS, ECM AND ECME , 1999 .

[4]  R. Douc,et al.  Optimality of the auxiliary particle filter , 2009 .

[5]  Man-Suk Oh,et al.  Integration of Multimodal Functions by Monte Carlo Importance Sampling , 1993 .

[6]  G. McLachlan,et al.  The EM algorithm and extensions , 1996 .

[7]  Haikady N. Nagaraja,et al.  Inference in Hidden Markov Models , 2006, Technometrics.

[8]  P. Moral,et al.  On Adaptive Sequential Monte Carlo Methods , 2008 .

[9]  Timothy J. Robinson,et al.  Sequential Monte Carlo Methods in Practice , 2003 .

[10]  O. Cappé,et al.  On‐line expectation–maximization algorithm for latent data models , 2009 .

[11]  W. Rayens,et al.  Covariance pooling and stabilization for classification , 1991 .

[12]  Jean-Michel Marin,et al.  Adaptive importance sampling in general mixture classes , 2007, Stat. Comput..

[13]  G. McLachlan,et al.  The EM Algorithm and Extensions: Second Edition , 2008 .

[14]  Eric Moulines,et al.  Inference in hidden Markov models , 2010, Springer series in statistics.

[15]  Eric Moulines,et al.  Inference in Hidden Markov Models (Springer Series in Statistics) , 2005 .

[16]  M. Li,et al.  Particle Markov chain Monte Carlo methods , 2015 .

[17]  Rudolph van der Merwe,et al.  Sigma-point kalman filters for probabilistic inference in dynamic state-space models , 2004 .

[18]  C. Andrieu,et al.  On the ergodicity properties of some adaptive MCMC algorithms , 2006, math/0610317.

[19]  Eric Moulines,et al.  Adaptive methods for sequential importance sampling with application to state space models , 2008, 2008 16th European Signal Processing Conference.

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

[21]  Stephen S. Wilson,et al.  Random iterative models , 1996 .

[22]  Pierre Priouret,et al.  Adaptive Algorithms and Stochastic Approximations , 1990, Applications of Mathematics.

[23]  Robert A. Jacobs,et al.  Hierarchical Mixtures of Experts and the EM Algorithm , 1993, Neural Computation.

[24]  Ke Chen,et al.  Improved learning algorithms for mixture of experts in multiclass classification , 1999, Neural Networks.

[25]  E. Kuhn,et al.  Coupling a stochastic approximation version of EM with an MCMC procedure , 2004 .

[26]  Eric Moulines,et al.  Stability of Stochastic Approximation under Verifiable Conditions , 2005, Proceedings of the 44th IEEE Conference on Decision and Control.

[27]  J. Rosenthal,et al.  Coupling and Ergodicity of Adaptive Markov Chain Monte Carlo Algorithms , 2007, Journal of Applied Probability.

[28]  Geoffrey J. McLachlan,et al.  Robust mixture modelling using the t distribution , 2000, Stat. Comput..

[29]  Michael I. Jordan,et al.  Convergence results for the EM approach to mixtures of experts architectures , 1995, Neural Networks.

[30]  A. Doucet,et al.  Particle Markov chain Monte Carlo methods , 2010 .

[31]  Julien Cornebise,et al.  Adaptive Markov chain Monte Carlo forward projection for statistical analysis in epidemic modelling of human papillomavirus. , 2013, Statistics in medicine.

[32]  P. Moral,et al.  On adaptive resampling strategies for sequential Monte Carlo methods , 2012, 1203.0464.

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

[34]  H. Haario,et al.  An adaptive Metropolis algorithm , 2001 .

[35]  Gersende Fort,et al.  Convergence of the Monte Carlo expectation maximization for curved exponential families , 2003 .

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

[37]  Gareth O. Roberts,et al.  Examples of Adaptive MCMC , 2009 .

[38]  Dirk P. Kroese,et al.  Cross‐Entropy Method , 2011 .

[39]  Jun S. Liu,et al.  Monte Carlo strategies in scientific computing , 2001 .

[40]  Simon J. Godsill,et al.  On sequential Monte Carlo sampling methods for Bayesian filtering , 2000, Stat. Comput..

[41]  Branko Ristic,et al.  Beyond the Kalman Filter: Particle Filters for Tracking Applications , 2004 .

[42]  R. Quandt.  A New Approach to Estimating Switching Regressions , 1972 .

[43]  P. Moral,et al.  Sequential Monte Carlo samplers , 2002, cond-mat/0212648.