Physics-incorporated convolutional recurrent neural networks for source identification and forecasting of dynamical systems. (December 2021)

• Record Type: Journal Article
• Authors: Saha, Priyabrata; Dash, Saurabh; Mukhopadhyay, Saibal
• Is Part Of: Neural networks. Volume 144(2021)
• Page Start: 359
• Page End: 371
• Publication Date: 2021-12
• Subjects: Dynamical systems -- Partial differential equation -- Recurrent neural networks -- Physics-incorporated neural networks; Neural computers -- Periodicals; Neural networks (Computer science) -- Periodicals; Neural networks (Neurobiology) -- Periodicals; Nervous System -- Periodicals; Ordinateurs neuronaux -- Périodiques; Réseaux neuronaux (Informatique) -- Périodiques; Réseaux neuronaux (Neurobiologie) -- Périodiques; Neural computers; Neural networks (Computer science); Neural networks (Neurobiology); Periodicals; 006.32
• DOI: 10.1016/j.neunet.2021.08.033
• Languages: English
• ISSNs: 0893-6080

Abstract

Abstract: Spatio-temporal dynamics of physical processes are generally modeled using partial differential equations (PDEs). Though the core dynamics follows some principles of physics, real-world physical processes are often driven by unknown external sources. In such cases, developing a purely analytical model becomes very difficult and data-driven modeling can be of assistance. In this paper, we present a hybrid framework combining physics-based numerical models with deep learning for source identification and forecasting of spatio-temporal dynamical systems with unobservable time-varying external sources. We formulate our model PhICNet as a convolutional recurrent neural network (RNN) which is end-to-end trainable for spatio-temporal evolution prediction of dynamical systems and learns the source behavior as an internal state of the RNN. Experimental results show that the proposed model can forecast the dynamics for a relatively long time and identify the sources as well.