Abstract:This work presents a hybrid modeling approach to data-driven learning and representation of unknown physical processes and closure parameterizations. These hybrid models are suitable for situations where the mechanistic description of dynamics of some variables is unknown, but reasonably accurate observational data can be obtained for the evolution of the state of the system. In this work, we propose machine learning to account for missing physics and then data assimilation to correct the prediction. In particular, we devise an effective methodology based on a recurrent neural network to model the unknown dynamics. A long short-term memory (LSTM) based correction term is added to the predictive model in order to take into account hidden physics. Since LSTM introduces a black-box approach for the unknown part of the model, we investigate whether the proposed hybrid neural-physical model can be further corrected through a sequential data assimilation step. We apply this framework to the weakly nonlinear Lorenz model that displays quasiperiodic oscillations, the highly nonlinear Lorenz model, and two-scale Lorenz model. The hybrid neural-physics model yields accurate results for the weakly nonlinear Lorenz model with the predicted state close to the true Lorenz model trajectory. For the highly nonlinear Lorenz model and the two-scale Lorenz model, the hybrid neural-physics model deviates from the true state due to the accumulation of prediction error from one time step to the next time step. The ensemble Kalman filter approach takes into account the prediction error and updates the diverged prediction using available observations in order to provide a more accurate state estimate. The successful synergistic integration of neural network and data assimilation for low-dimensional system shows the potential benefits of the proposed hybrid-neural physics model for complex systems.

Hybrid Neural Differential Equations to Model Unknown Mechanisms and States in Biology

Hybrid modeling and prediction of dynamical systems

Robust parameter estimation and identifiability analysis with Hybrid Neural Ordinary Differential Equations in Computational Biology

A nonintrusive hybrid neural-physics modeling of incomplete dynamical systems: Lorenz equations

Generative Modeling of Neural Dynamics via Latent Stochastic Differential Equations

Cubature Kalman Filter Based Training of Hybrid Differential Equation Recurrent Neural Network Physiological Dynamic Models

Identifying Latent Dynamic Components in Biological Systems

Data-Driven Modeling of Partially Observed Biological Systems

Learning Governing Equations of Unobserved States in Dynamical Systems

A Data-Driven Hybrid Automaton Framework to Modeling Complex Dynamical Systems

Incomplete to complete multiphysics forecasting -- a hybrid approach for learning unknown phenomena

A Hybrid Neural Network - First Principles Approach For Modelling Of Cell Metabolism

Hybrid Modelling of the General Middle-Sized Genetic Regulatory Networks

Hybrid$^2$ Neural ODE Causal Modeling and an Application to Glycemic Response

Hybrid Modelling of Biological Systems: Current Progress and Future Prospects

DiffHybrid-UQ: Uncertainty Quantification for Differentiable Hybrid Neural Modeling

Latent State Space Extension for interpretable hybrid mechanistic models

A Data-Driven Framework for Learning Hybrid Dynamical Systems.

Learning Interactions in Reaction Diffusion Equations by Neural Networks

Collocation based training of neural ordinary differential equations

A Probabilistic State Space Model for Joint Inference from Differential Equations and Data