Controlled sequential Monte Carlo
From MaRDI portal
Publication:2215764
DOI10.1214/19-AOS1914zbMATH Open1455.62162arXiv1708.08396OpenAlexW3087215779MaRDI QIDQ2215764FDOQ2215764
Authors: Jeremy Heng, Adrian N. Bishop, George Deligiannidis, Arnaud Doucet 
Publication date: 14 December 2020
Published in: The Annals of Statistics (Search for Journal in Brave)
Abstract: Sequential Monte Carlo methods, also known as particle methods, are a popular set of techniques for approximating high-dimensional probability distributions and their normalizing constants. These methods have found numerous applications in statistics and related fields; e.g. for inference in non-linear non-Gaussian state space models, and in complex static models. Like many Monte Carlo sampling schemes, they rely on proposal distributions which crucially impact their performance. We introduce here a class of controlled sequential Monte Carlo algorithms, where the proposal distributions are determined by approximating the solution to an associated optimal control problem using an iterative scheme. This method builds upon a number of existing algorithms in econometrics, physics, and statistics for inference in state space models, and generalizes these methods so as to accommodate complex static models. We provide a theoretical analysis concerning the fluctuation and stability of this methodology that also provides insight into the properties of related algorithms. We demonstrate significant gains over state-of-the-art methods at a fixed computational complexity on a variety of applications.
Full work available at URL: https://arxiv.org/abs/1708.08396
Recommendations
state space modelsoptimal controlapproximate dynamic programmingreinforcement learningnormalizing constantsannealed importance sampling
Computational methods for problems pertaining to statistics (62-08) Asymptotic properties of parametric estimators (62F12) Markov processes: estimation; hidden Markov models (62M05) Time series, auto-correlation, regression, etc. in statistics (GARCH) (62M10) Monte Carlo methods (65C05) Sequential estimation (62L12) Dynamic programming (90C39)
Cites Work
- Sequential Monte Carlo Samplers
- Exponential convergence of Langevin distributions and their discrete approximations
- Log Gaussian Cox Processes
- Central limit theorem for sequential Monte Carlo methods and its application to Bayesian inference
- Lookahead strategies for sequential Monte Carlo
- Title not available (Why is that?)
- Filtering via Simulation: Auxiliary Particle Filters
- Particle Markov Chain Monte Carlo Methods
- Title not available (Why is that?)
- Monte Carlo strategies in scientific computing
- A sequential particle filter method for static models
- Sequential Monte Carlo Methods for Dynamic Systems
- Langevin diffusions and Metropolis-Hastings algorithms
- Boosting With theL2Loss
- Riemann manifold Langevin and Hamiltonian Monte Carlo methods. With discussion and authors' reply
- Simulating normalizing constants: From importance sampling to bridge sampling to path sampling
- Particle filters
- Title not available (Why is that?)
- Title not available (Why is that?)
- Smoothing algorithms for state-space models
- Computing a nearest symmetric positive semidefinite matrix
- Recursive Monte Carlo filters: algorithms and theoretical analysis
- Efficient high-dimensional importance sampling
- Iterated Random Functions
- Adaptive importance sampling for control and inference
- Inference for Lévy-driven stochastic volatility models via adaptive sequential Monte Carlo
- On the convergence of adaptive sequential Monte Carlo methods
- Sequential Monte Carlo on large binary sampling spaces
- Ergodicity for SDEs and approximations: locally Lipschitz vector fields and degenerate noise.
- Stochastic Integration with Jumps
- Central limit theorem for nonlinear filtering and interacting particle systems
- Two-filter formulae for discrete-time non-linear bayesian smoothing
- Scaling Limits for the Transient Phase of Local Metropolis–Hastings Algorithms
- Monte-Carlo methods and stochastic processes. From linear to non-linear
- Particle efficient importance sampling
- Path storage in the particle filter
- Particle Smoothing for Hidden Diffusion Processes: Adaptive Path Integral Smoother
- Negative association, ordering and convergence of resampling methods
Cited In (18)
- Solving high-dimensional Hamilton-Jacobi-Bellman PDEs using neural networks: perspectives from the theory of controlled diffusions and measures on path space
- Monte Carlo, Control Variates, and Stochastic Ordering
- Monte Carlo Methods for the Neutron Transport Equation
- An Invitation to Sequential Monte Carlo Samplers
- An optimal control approach to particle filtering
- Data assimilation: the Schrödinger perspective
- Continuous-discrete smoothing of diffusions
- Multi-index sequential Monte Carlo ratio estimators for Bayesian inverse problems
- A randomized multi-index sequential Monte Carlo method
- Elements of sequential Monte Carlo
- Sequential Monte Carlo for model predictive control
- Unbiased Markov chain Monte Carlo for intractable target distributions
- Controlling procedural modeling programs with stochastically-ordered sequential Monte Carlo
- Rejection Control and Sequential Importance Sampling
- Ensemble Kalman filter based sequential Monte Carlo sampler for sequential Bayesian inference
- Diffusion Schrödinger bridges for Bayesian computation
- A multilevel approach for stochastic nonlinear optimal control
- Sequential Monte Carlo methods for diffusion processes
This page was built for publication: Controlled sequential Monte Carlo
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2215764)
