# Flow-based Bayesian filtering for high-dimensional nonlinear stochastic dynamical systems

Source: [Flow-based Bayesian filtering for high-dimensional nonlinear stochastic dynamical systems](<https://www.sota2.com/research/paper/68fa5d101e209131ce9d69f4>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-25T18:01:20.823Z

Authors: Xintong Wang, Xiaofei Guan, Ling Guo, Hao Wu

Published: 2025-02-22T14:04:23.000Z

## Abstract

Bayesian filtering for high-dimensional nonlinear stochastic dynamical systems is a fundamental yet challenging problem in many fields of science and engineering. Existing methods face significant obstacles: Gaussian-based filters struggle with non-Gaussian distributions, while sequential Monte Carlo methods are computationally intensive and prone to particle degeneracy in high dimensions. Although generative models in machine learning have made significant progress in modeling high-dimensional non-Gaussian distributions, their inefficiency in online updating limits their applicability to filtering problems. To address these challenges, we propose a flow-based Bayesian filter (FBF) that integrates normalizing flows to construct a novel latent linear state-space model with Gaussian filtering distributions. This framework facilitates efficient density estimation and sampling using invertible transformations provided by normalizing flows, and it enables the construction of filters in a data-driven manner, without requiring prior knowledge of system dynamics or observation models. Numerical experiments demonstrate the superior accuracy and efficiency of FBF.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Filtering on Burgers' equation nu=0.01](<https://www.sota2.com/research/sota/filtering-on-burgers-equation-nu-0-01.md>): Filtering; Burgers' equation nu=0.01; RMSE: 0.1152.

- [Stochastic Filtering on Burgers' equation (test)](<https://www.sota2.com/research/sota/stochastic-filtering-on-burgers-equation-test.md>): Stochastic Filtering; Burgers Equation (test); RMSE: 0.0752.

- [Filtering on Two-scale Lorenz model F=5](<https://www.sota2.com/research/sota/filtering-on-two-scale-lorenz-model-f-5.md>): Filtering; Two-scale Lorenz model F=5; RMSE: 0.5817.

- [Filtering on Two-scale Lorenz model F=8](<https://www.sota2.com/research/sota/filtering-on-two-scale-lorenz-model-f-8.md>): Filtering; Two-scale Lorenz model F=8; RMSE: 0.8301.

- [Filtering on Two-scale Lorenz model F=16](<https://www.sota2.com/research/sota/filtering-on-two-scale-lorenz-model-f-16.md>): Filtering; Two-scale Lorenz model F=16; RMSE: 3.7536.

- [Filtering on Single-scale Lorenz-96 (test)](<https://www.sota2.com/research/sota/filtering-on-single-scale-lorenz-96-test.md>): Filtering; Single-scale Lorenz-96 (test); RMSE: 0.2044.

- [Filtering on two-factor stochastic volatility model synthetic (test)](<https://www.sota2.com/research/sota/filtering-on-two-factor-stochastic-volatility-model-synthetic-test.md>): Filtering; two-factor stochastic volatility model synthetic (test); RMSE: 0.7481.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2502.16232>)

[PDF](<https://arxiv.org/pdf/2502.16232.pdf>)
