Bayesian differential programming for robust systems identification under uncertainty
From MaRDI portal
Abstract: This paper presents a machine learning framework for Bayesian systems identification from noisy, sparse and irregular observations of nonlinear dynamical systems. The proposed method takes advantage of recent developments in differentiable programming to propagate gradient information through ordinary differential equation solvers and perform Bayesian inference with respect to unknown model parameters using Hamiltonian Monte Carlo. This allows us to efficiently infer posterior distributions over plausible models with quantified uncertainty, while the use of sparsity-promoting priors enables the discovery of interpretable and parsimonious representations for the underlying latent dynamics. A series of numerical studies is presented to demonstrate the effectiveness of the proposed methods including nonlinear oscillators, predator-prey systems, chaotic dynamics and systems biology. Taken all together, our findings put forth a novel, flexible and robust workflow for data-driven model discovery under uncertainty.
Recommendations
- Bayesian Distributionally Robust Optimization
- Bayesian solution uncertainty quantification for differential equations
- Bayesian augmented Lagrangian algorithm for system identification
- Efficient Bayesian estimation and uncertainty quantification in ordinary differential equation models
- Bayesian system ID: optimal management of parameter, model, and measurement uncertainty
- Computational system identification of continuous-time nonlinear systems using approximate Bayesian computation
- Bayesian computation and adaptive control for dynamic linear system
- A Bayesian approach to identification of hybrid systems
- Bayesian identification of stochastic linear systems with observations at multiple scales
- Dynamic systems identification with Gaussian processes
Cites work
- A First Course in the Numerical Analysis of Differential Equations
- Adaptive Rejection Metropolis Sampling within Gibbs Sampling
- Adversarial uncertainty quantification in physics-informed neural networks
- Bayesian data analysis.
- Compressed sensing
- Conditional deep surrogate models for stochastic, high-dimensional, and multi-fidelity systems
- Control of systems integrating logic, dynamics, and constraints
- Crisis of the chaotic attractor of a climate model: a transfer operator approach
- Data driven governing equations approximation using deep neural networks
- Data-driven closures for stochastic dynamical systems
- Data-driven discovery of coordinates and governing equations
- Discovering governing equations from data by sparse identification of nonlinear dynamical systems
- Hidden fluid mechanics: Learning velocity and pressure fields from flow visualizations
- Hidden physics models: machine learning of nonlinear partial differential equations
- scientific article; zbMATH DE number 3167340 (Why is no real title available?)
- scientific article; zbMATH DE number 3667770 (Why is no real title available?)
- scientific article; zbMATH DE number 45867 (Why is no real title available?)
- Inferring solutions of differential equations using noisy multi-fidelity data
- Lagrangian coherent structures from approximate velocity data
- Large-scale machine learning with stochastic gradient descent
- Machine learning in cardiovascular flows modeling: predicting arterial blood pressure from non-invasive 4D flow MRI data using physics-informed neural networks
- Markov Chain Monte Carlo Convergence Diagnostics: A Comparative Review
- MCMC using Hamiltonian dynamics
- Non-intrusive reduced order modeling of unsteady flows using artificial neural networks with application to a combustion problem
- Non-Linear State Space Modelling of Fisheries Biomass Dynamics by Using Metropolis-Hastings within-Gibbs Sampling
- Nonparametric inference of interaction laws in systems of agents from trajectory data
- Numerical Gaussian processes for time-dependent and nonlinear partial differential equations
- Physics-constrained deep learning for high-dimensional surrogate modeling and uncertainty quantification without labeled data
- Physics-informed generative adversarial networks for stochastic differential equations
- Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
- Robust Bayesian analysis: sensitivity to the prior
- Solving high-dimensional partial differential equations using deep learning
- The geometric foundations of Hamiltonian Monte Carlo
- The no-U-turn sampler: adaptively setting path lengths in Hamiltonian Monte Carlo
- Understanding and Mitigating Gradient Flow Pathologies in Physics-Informed Neural Networks
Cited in
(20)- Learning stable and predictive structures in kinetic systems
- Bayesian augmented Lagrangian algorithm for system identification
- Bayesian frequentist bounds for machine learning and system identification
- A Bayesian approach for data-driven dynamic equation discovery
- Sparse Bayesian deep learning for dynamic system identification
- On the eigenvector bias of Fourier feature networks: from regression to solving multi-scale PDEs with physics-informed neural networks
- Automated learning of interpretable models with quantified uncertainty
- Uncertainty quantification in scientific machine learning: methods, metrics, and comparisons
- Physics-integrated neural differentiable (PiNDiff) model for composites manufacturing
- ASYMMETRICAL CONDITIONAL BAYES PARAMETER IDENTIFICATION FOR CONTROL ENGINEERING
- Pseudo-Hamiltonian system identification
- A Review of Data‐Driven Discovery for Dynamic Systems
- Operator inference with roll outs for learning reduced models from scarce and low-quality data
- Scalable Bayesian optimization with randomized prior networks
- NeuralUQ: A Comprehensive Library for Uncertainty Quantification in Neural Differential Equations and Operators
- Memory-based parameterization with differentiable solver: application to Lorenz '96
- Bayesian identification of nonseparable Hamiltonians with multiplicative noise using deep learning and reduced-order modeling
- A Bayesian approach for spatio-temporal data-driven dynamic equation discovery
- Neural closure models for dynamical systems
- Ensemble-SINDy: robust sparse model discovery in the low-data, high-noise limit, with active learning and control
This page was built for publication: Bayesian differential programming for robust systems identification under uncertainty
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5161165)
