On coupling particle filter trajectories
From MaRDI portal
Abstract: Particle filters are a powerful and flexible tool for performing inference on state-space models. They involve a collection of samples evolving over time through a combination of sampling and re-sampling steps. The re-sampling step is necessary to ensure that weight degeneracy is avoided. In several situations of statistical interest, it is important to be able to compare the estimates produced by two different particle filters; consequently, being able to efficiently couple two particle filter trajectories is often of paramount importance. In this text, we propose several ways to do so. In particular, we leverage ideas from the optimal transportation literature. In general, though, computing the optimal transport map is extremely computationally expensive; to deal with this, we introduce computationally tractable approximations to optimal transport couplings. We demonstrate that our resulting algorithms for coupling two particle filter trajectories often perform orders of magnitude more efficiently than more standard approaches.
Recommendations
- Copula particle filters
- Smoothing with couplings of conditional particle filters
- Particle Filters for Partially Observed Diffusions
- Particle filters
- Particle Filtering
- Coupled conditional backward sampling particle filter
- Particle filters for continuous-time jump models in tracking applications
- Central limit theorems for coupled particle filters
- Particle filters for continuous likelihood evaluation and maximisation
- Particle Filters for Multiscale Diffusions
Cites work
- scientific article; zbMATH DE number 5919872 (Why is no real title available?)
- scientific article; zbMATH DE number 1240224 (Why is no real title available?)
- scientific article; zbMATH DE number 3231692 (Why is no real title available?)
- scientific article; zbMATH DE number 3023295 (Why is no real title available?)
- A Second Course in Probability
- A guided sequential Monte Carlo method for the assimilation of data into stochastic dynamical systems
- A nonparametric ensemble transform method for Bayesian inference
- A sparse multiscale algorithm for dense optimal transport
- An Algorithm for Finding Best Matches in Logarithmic Expected Time
- Analyses of Infectious Disease Data from Household Outbreaks by Markov Chain Monte Carlo Methods
- Analytical and numerical studies of noise-induced synchronization of chaotic systems
- Bayesian forecasting and dynamic models.
- Convergence properties of pseudo-marginal Markov chain Monte Carlo algorithms
- Data-driven optimal transport
- Efficient implementation of Markov chain Monte Carlo when using an unbiased likelihood estimator
- Establishing some order amongst exact approximations of MCMCs
- Inference in hidden Markov models.
- Iterated filtering
- Monte Carlo Inference for State–Space Models of Wild Animal Populations
- Multilevel Monte Carlo Path Simulation
- Multilevel Particle Filters
- Multilevel Sequential Monte Carlo Samplers for Normalizing Constants
- Multilevel ensemble Kalman filtering
- Multilevel ensemble transform particle filtering
- Multilevel particle filters: normalizing constant estimation
- Multilevel sequential Monte Carlo samplers
- Noisy Monte Carlo: convergence of Markov chains with approximate transition kernels
- On a Least Squares Adjustment of a Sampled Frequency Table When the Expected Marginal Totals are Known
- On particle Gibbs sampling
- On particle methods for parameter estimation in state-space models
- On the efficiency of pseudo-marginal random walk Metropolis algorithms
- On the translocation of masses
- Particle Markov chain Monte Carlo for efficient numerical simulation
- Particle approximations of the score and observed information matrix in state space models with application to parameter estimation
- Practical Markov Chain Monte Carlo
- Regularized discrete optimal transport
- Scikit-learn: machine learning in Python
- Sequential Imputations and Bayesian Missing Data Problems
- Sequential Monte Carlo Methods for State and Parameter Estimation in Abruptly Changing Environments
- Smoothing with couplings of conditional particle filters
- Stability of noisy Metropolis-Hastings
- Stochastic chemical kinetics. Theory and (mostly) systems biological applications
- The Sinkhorn–Knopp Algorithm: Convergence and Applications
- The pseudo-marginal approach for efficient Monte Carlo computations
- The rate of convergence of Sinkhorn balancing
- Time series analysis by state space methods.
- Turbulent models of oil flow in a circular pipe with sudden enlargement
Cited in
(16)- Inference via low-dimensional couplings
- Multilevel particle filters: normalizing constant estimation
- A Wasserstein coupled particle filter for multilevel estimation
- On backward smoothing algorithms
- Advanced Multilevel Monte Carlo Methods
- Feedback Particle Filter on Riemannian Manifolds and Matrix Lie Groups
- Bayesian static parameter estimation for partially observed diffusions via multilevel Monte Carlo
- Multilevel Monte Carlo for smoothing via transport methods
- A seamless multilevel ensemble transform particle filter
- Smoothing with couplings of conditional particle filters
- Score-Based Parameter Estimation for a Class of Continuous-Time State Space Models
- Multilevel particle filters for Lévy-driven stochastic differential equations
- Central limit theorems for coupled particle filters
- Interacting and annealing particle filters: mathematics and a recipe for applications
- Multilevel Particle Filters
- Path storage in the particle filter
This page was built for publication: On coupling particle filter trajectories
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1702025)
