Guided filtering and smoothing for infinite-dimensional diffusions

TL;DR

This paper introduces a novel approach for filtering and smoothing in infinite-dimensional diffusion processes using guided distributions, enhancing Monte Carlo methods and including parameter estimation, demonstrated through a stochastic Amari equation case study.

Contribution

The paper develops a new guided distribution framework for infinite-dimensional diffusions, improving filtering, smoothing, and parameter estimation techniques.

Findings

Effective proposal measure for Monte Carlo schemes

Enhanced filtering and smoothing accuracy

Successful application to stochastic Amari equation

Abstract

We consider the filtering and smoothing problems for an infinite-dimensional diffusion process X, observed through a finite-dimensional representation at discrete points in time. At the heart of our proposed methodology lies the construction of a path measure, termed the guided distribution of X, that is absolutely continuous with respect to the law of X, conditioned on the observations. We show that this distribution can be incorporated as a potent proposal measure for both sequential Monte Carlo as well as Markov Chain Monte Carlo schemes to tackle the filtering and smoothing problems respectively. In the offline setting, we extend our approach to incorporate parameter estimation of unknown model parameters. The proposed methodology is numerically illustrated in a case study for the stochastic Amari equation.