Adaptive methods for sequential importance sampling with application to state space models
By Julien Cornebise, Julien Cornebise, Eric Moulines, Jimmy Olsson. 2008
91 citations recorded by the source corpus.
In this paper we discuss new adaptive proposal strategies for sequential Monte Carlo algorithms—also known as particle filters—relying on criteria evaluating the quality of the proposed particles. The choice of the proposal distribution is a major concern and can dramatically influence the quality of the estimates. Thus, we show how the long-used coefficient of variation (suggested by Kong et al. in J. Am. Stat. Assoc. 89(278–288):590–599, 1994) of the weights can be used for estimating the chi-square distance between the target and instrumental distributions of the auxiliary particle filter. As a by-product of this analysis we obtain an auxiliary adjustment multiplier weight type for which this chi-square distance is minimal. Moreover, we establish an empirical estimate of linear complexity of the Kullback-Leibler divergence between the involved distributions. Guided by these results, we discuss adaptive designing of the particle filter proposal distribution and illustrate the methods on a numerical example.
Subfields: Artificial Intelligence, Statistics and Probability, Probabilistic & Bayesian Methods
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Brooks Paige
- Frank D. Wood
- Paul Fearnhead
- Jimmy Olsson
- Julien Cornebise
- Eric Moulines
- Gareth William Peters
- Geoffrey R. Hosack
In Statistics and Probability
- Brooks Paige
- Frank D. Wood
- Paul Fearnhead
- Jimmy Olsson
- Julien Cornebise
- Eric Moulines
- Gareth William Peters
- Geoffrey R. Hosack
In Probabilistic & Bayesian Methods
- Brooks Paige
- Frank D. Wood
- Paul Fearnhead
- Benjamin M. Taylor
- Gareth William Peters
- Eric Moulines
- Jimmy Olsson
- Julien Cornebise
Citations in the corpus
- Adaptive sequential Monte Carlo by means of mixture of experts (2011, methodology): “Relying on the results of Cornebise et al. (2008), we have built new algorithms approximating the so-called optimal proposal kernel at a given time step of an auxiliary particle filter by means of minimisation of the KLD between the auxiliary target and instrumental distributions of the particle Fig.”
- An Adaptive Sequential Monte Carlo Sampler (2010, background): “Sections 2.1 and 2.2 provide a brief description of two such Monte Carlo approaches.”
- On sequential Monte Carlo, partial rejection control and approximate Bayesian computation (2008, unknown): “It is well known that the performance of SMC methods are strongly dependent on the mutation kernel (Cornebise et al. 2008).”
- Inference Networks for Sequential Monte Carlo in Graphical Models (2016, methodology): “, 2008) and sequential Monte Carlo (Cornebise et al., 2008; 2014; Gu et al., 2015), a parametric family q(x|λ) is proposed, with λ is a free parameter, and the adaptive algorithms aim to minimize either the reverse KullbackLeibler (KL) divergence or Chi-squared distance between the approximating fam”
- Ecological non-linear state space model selection via adaptive particle Markov chain Monte Carlo (AdPMCMC) (2010, methodology): “Finally, we point out that adaption of the mutation kernel in the SMC component of the algorithm can be performed arbitrarily without any constraint on the adaption rate and one possible approach to consider involves the methodology of [Cornebise et al., 2008].”