De-Sequentialized Monte Carlo: a parallel-in-time particle smoother

02/04/2022
โˆ™
by   Adrien Corenflos, et al.
โˆ™
1
โˆ™

Particle smoothers are SMC (Sequential Monte Carlo) algorithms designed to approximate the joint distribution of the states given observations from a state-space model. We propose dSMC (de-Sequentialized Monte Carlo), a new particle smoother that is able to process T observations in ๐’ช(log T) time on parallel architecture. This compares favourably with standard particle smoothers, the complexity of which is linear in T. We derive โ„’_p convergence results for dSMC, with an explicit upper bound, polynomial in T. We then discuss how to reduce the variance of the smoothing estimates computed by dSMC by (i) designing good proposal distributions for sampling the particles at the initialization of the algorithm, as well as by (ii) using lazy resampling to increase the number of particles used in dSMC. Finally, we design a particle Gibbs sampler based on dSMC, which is able to perform parameter inference in a state-space model at a ๐’ช(log(T)) cost on parallel hardware.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
โˆ™ 06/18/2018

Stability of Conditional Sequential Monte Carlo

The particle Gibbs (PG) sampler is a Markov Chain Monte Carlo (MCMC) alg...
research
โˆ™ 01/03/2019

Learning Nonlinear State Space Models with Hamiltonian Sequential Monte Carlo Sampler

State space models (SSM) have been widely applied for the analysis and v...
research
โˆ™ 01/06/2023

The one step fixed-lag particle smoother as a strategy to improve the prediction step of particle filtering

Sequential Monte Carlo methods have been a major breakthrough in the fie...
research
โˆ™ 03/27/2023

Parameter estimation for many-particle models from aggregate observations: A Wasserstein distance based sequential Monte Carlo sampler

In this work we study systems consisting of a group of moving particles....
research
โˆ™ 04/21/2019

Particle filter efficiency under limited communication

Sequential Monte Carlo (SMC) methods are typically not straightforward t...
research
โˆ™ 03/12/2012

Decentralized, Adaptive, Look-Ahead Particle Filtering

The decentralized particle filter (DPF) was proposed recently to increas...
research
โˆ™ 07/15/2020

The Monte Carlo Transformer: a stochastic self-attention model for sequence prediction

This paper introduces the Sequential Monte Carlo Transformer, an origina...

Please sign up or login with your details

Forgot password? Click here to reset