Abstract:Data-driven discovery of governing equations in computational science has emerged as a new paradigm for obtaining accurate physical models and as a possible alternative to theoretical derivations. The recently developed physics-informed neural networks (PINNs) have also been employed to learn governing equations given data across diverse scientific disciplines, e.g., in biology and fluid dynamics. Despite the effectiveness of PINNs for discovering governing equations, the physical models encoded in PINNs may be misspecified in complex systems as some of the physical processes may not be fully understood, leading to the poor accuracy of PINN predictions. In this work, we present a general approach to correct the misspecified physical models in PINNs for discovering governing equations, given some sparse and/or noisy data. Specifically, we first encode the assumed physical models, which may be misspecified in PINNs, and then employ other deep neural networks (DNNs) to model the discrepancy between the imperfect models and the observational data. Due to the expressivity of DNNs, the proposed method is capable of reducing the computational errors caused by the model misspecification and thus enables the applications of PINNs in complex systems where the physical processes are not exactly known. Furthermore, we utilize the Bayesian physics-informed neural networks (B-PINNs) and/or ensemble PINNs to quantify uncertainties arising from noisy and/or gappy data in the discovered governing equations. A series of numerical examples including reaction-diffusion systems and non-Newtonian channel and cavity flows demonstrate that the added DNNs are capable of correcting the model misspecification in PINNs and thus reduce the discrepancy between the physical models encoded in PINNs and the observational data. In addition, the B-PINNs and ensemble PINNs can provide reasonable uncertainty bounds in the discovered physical models, which makes the predictions more reliable. We also demonstrate that we can seamlessly combine the present approach with the symbolic regression to obtain the explicit governing equations upon the training of PINNs. We envision that the proposed approach will extend the applications of PINNs for discovering governing equations in problems where the physico-chemical or biological processes are not well understood.

Process‐Informed Neural Networks: A Hybrid Modelling Approach to Improve Predictive Performance and Inference of Neural Networks in Ecology and Beyond

Physics-informed Neural Networks to Solve Lumped Kinetic Model for Chromatography Process.

A Spatial-Temporally Adaptive PINN Framework for 3D Bi-Ventricular Electrophysiological Simulations and Parameter Inference

Physics-Informed Neural Networks for Dynamic Process Operations with Limited Physical Knowledge and Data

Physics informed integral neural network for dynamic modelling of solvent-based post-combustion CO 2 capture process

Can deep neural networks learn process model structure? An assessment framework and analysis

Physics‐informed neural networks for biopharmaceutical cultivation processes: Consideration of varying process parameter settings

Adapting Physics-Informed Neural Networks for Bifurcation Detection in Ecological Migration Models

Correcting model misspecification in physics-informed neural networks (PINNs)

Domain-decoupled Physics-informed Neural Networks with Closed-form Gradients for Fast Model Learning of Dynamical Systems

Enhancing process-based hydrological models with embedded neural networks: A hybrid approach

Biologically-informed neural networks guide mechanistic modeling from sparse experimental data

Scientific Machine Learning Through Physics–Informed Neural Networks: Where we are and What's Next

Modeling of additive manufacturing processes with time‐dependent material properties using physics‐informed neural networks

Inverse Physics-Informed Neural Networks for transport models in porous materials

Machine Learning Through Physics–Informed Neural Networks: Progress and Challenges

A metalearning approach for Physics-Informed Neural Networks (PINNs): Application to parameterized PDEs

Enhanced physics‐informed neural networks for efficient modelling of hydrodynamics in river networks

Bayesian Physics Informed Neural Networks for Data Assimilation and Spatio-Temporal Modelling of Wildfires

An Expert's Guide to Training Physics-informed Neural Networks