A convergent scheme for the Bayesian filtering problem based on the Fokker--Planck equation and deep splitting
For researchers in Bayesian filtering and stochastic processes, this work provides a convergent and scalable numerical method for high-dimensional nonlinear filtering problems.
The authors introduce a numerical scheme for approximating the nonlinear filtering density, combining deep splitting for the Fokker-Planck equation with exact Bayesian updates, and prove convergence under a parabolic Hörmander condition. In a 10-dimensional numerical example, the method demonstrates robustness and mitigates the curse of dimensionality.
A numerical scheme for approximating the nonlinear filtering density is introduced and its convergence rate is established, theoretically under a parabolic Hörmander condition, and empirically in numerical examples. In a prediction step, between the noisy and partial measurements at discrete times, the scheme approximates the Fokker--Planck equation with a deep splitting scheme, followed by an exact update through Bayes' formula. This results in a classical prediction-update filtering algorithm that operates online for new observation sequences post-training. The algorithm employs a sampling-based Feynman--Kac approach, designed to mitigate the curse of dimensionality. As a corollary we obtain the convergence rate for the approximation of the Fokker--Planck equation alone, disconnected from the filtering problem. The convergence analysis is complemented by a nonlinear $10$-dimensional numerical example demonstrating the robustness of the method.