Stochastic Physics-Informed Neural Ordinary Differential Equations
Preprint
Abstract
Stochastic differential equations (SDEs) are used to describe a wide variety of complex stochastic dynamical systems. Learning the hidden physics within SDEs is crucial for unraveling fundamental understanding of these systems' stochastic and nonlinear behavior. We propose a flexible and scalable framework for training artificial neural networks to learn constitutive equations that represent hidden physics within SDEs. The proposed stochastic physics-informed neural ordinary differential equation framework (SPINODE) propagates stochasticity through the known structure of the SDE (i.e., the known physics) to yield a set of deterministic ODEs that describe the time evolution of statistical moments of the stochastic states. SPINODE then uses ODE solvers to predict moment trajectories. SPINODE learns neural network representations of the hidden physics by matching the predicted moments to those estimated from data. Recent advances in automatic differentiation and mini-batch gradient descent with adjoint sensitivity are leveraged to establish the unknown parameters of the neural networks. We demonstrate SPINODE on three benchmark in-silico case studies and analyze the framework's numerical robustness and stability. SPINODE provides a promising new direction for systematically unraveling the hidden physics of multivariate stochastic dynamical systems with multiplicative noise.Keywords
All Related Versions
This publication has 52 references indexed in Scilit:
- B-PINNs: Bayesian physics-informed neural networks for forward and inverse PDE problems with noisy dataJournal of Computational Physics, 2021
- Quantifying total uncertainty in physics-informed neural networks for solving forward and inverse stochastic problemsJournal of Computational Physics, 2019
- Adversarial uncertainty quantification in physics-informed neural networksJournal of Computational Physics, 2019
- Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equationsJournal of Computational Physics, 2019
- A stochastic SIRS epidemic model with nonlinear incidence rateApplied Mathematics and Computation, 2017
- Coexistence and exclusion of stochastic competitive Lotka–Volterra modelsJournal of Differential Equations, 2017
- System Identification Methods for (Operational) Modal Analysis: Review and ComparisonArchives of Computational Methods in Engineering, 2012
- Stochastic Processes in Physics and Chemistry2007
- Influence of nonlinear incidence rates upon the behavior of SIRS epidemiological modelsJournal of Mathematical Biology, 1986
- A generalization of the Kermack-McKendrick deterministic epidemic modelMathematical Biosciences, 1978