Variational Markov chain Monte Carlo for Bayesian smoothing of non-linear diffusions
- 204 Downloads
In this paper we develop set of novel Markov chain Monte Carlo algorithms for Bayesian smoothing of partially observed non-linear diffusion processes. The sampling algorithms developed herein use a deterministic approximation to the posterior distribution over paths as the proposal distribution for a mixture of an independence and a random walk sampler. The approximating distribution is sampled by simulating an optimized time-dependent linear diffusion process derived from the recently developed variational Gaussian process approximation method. The novel diffusion bridge proposal derived from the variational approximation allows the use of a flexible blocking strategy that further improves mixing, and thus the efficiency, of the sampling algorithms. The algorithms are tested on two diffusion processes: one with double-well potential drift and another with SINE drift. The new algorithm’s accuracy and efficiency is compared with state-of-the-art hybrid Monte Carlo based path sampling. It is shown that in practical, finite sample applications the algorithm is accurate except in the presence of large observation errors and low observation densities, which lead to a multi-modal structure in the posterior distribution over paths. More importantly, the variational approximation assisted sampling algorithm outperforms hybrid Monte Carlo in terms of computational efficiency, except when the diffusion process is densely observed with small errors in which case both algorithms are equally efficient.
KeywordsStochastic dynamic systems Data assimilation Bridge sampling
Unable to display preview. Download preview PDF.
- Archambeau C, Cornford D, Opper M, Shawe-Tayler J (2007) Gaussian Process approximations of stochastic differential equations. J Mach Learn Res Workshop and Conference Proceedings 1: 1–16Google Scholar
- Archambeau C, Opper M, Shen Y, Cornford D, Shawe-Tayler J (2008) Variational inference for diffusion processes. In: Platt C, Koller D, Singer Y, Roweis S (eds) Neural information processing systems (NIPS), vol 20. The MIT Press, Cambridge, pp 17–24Google Scholar
- de Freitas N, H⌽jen-S⌽rensen P, Jordan M, Russell S (2001) Variational MCMC. In: Proceedings of the 17th annual conference on uncertainty in artificial intelligence. Morgan Kaufmann Publishers Inc., San Francisco, CA, pp 120–127Google Scholar
- Honerkamp J (1994) Stochastic dynamical systems. VCH, WeinheimGoogle Scholar
- Kalnay E (2003) Atmospheric modelling, data assimilation and predictability. Cambridge University Press, CambridgeGoogle Scholar
- Papaspiliopolous O, Roberts GO, Skold M (2003) Non-centered parameterisations for hierarchical models and data augmentation. In: Bayesian Statistics 7, pp 307–326Google Scholar
- Wan E, van der Merwe R (2001) The unscented Kalman filter. In: Haykin S (ed) Kalman filtering and neural networks. Wiley, New York, pp 207–219Google Scholar
- Wilkinson DJ, Golightly A (2010) Markov Chain Monte Carlo algorithms for SDE parameter estimation. In: Learning and inference in computational systems biology, pp 253–276Google Scholar