Automating Discovery of Physics-Informed Neural State Space Models via Learning and Evolution

Elliott Skomski, Ján Drgoňa, Aaron Tuor
Proceedings of the 3rd Conference on Learning for Dynamics and Control, PMLR 144:980-991, 2021.

Abstract

Recent works exploring deep learning application to dynamical systems modeling have demonstrated that embedding physical priors into neural networks can yield more effective, physically-realistic, and data-efficient models. However, in the absence of complete prior knowledge of a dynamical system’s physical characteristics, determining the optimal structure and optimization strategy for these models can be difficult. In this work, we explore methods for discovering neural state space dynamics models for system identification. Starting with a design space of block-oriented state space models and structured linear maps with strong physical priors, we encode these components into a model genome alongside network structure, penalty constraints, and optimization hyperparameters. Demonstrating the overall utility of the design space, we employ an asynchronous genetic search algorithm that alternates between model selection and optimization and obtains accurate physically consistent models of three physical systems: an aerodynamics body, a continuous stirred tank reactor, and a two tank interacting system.

Cite this Paper

BibTeX
@InProceedings{pmlr-v144-skomski21a, title = {Automating Discovery of Physics-Informed Neural State Space Models via Learning and Evolution}, author = {Skomski, Elliott and Drgo\v{n}a, J\'an and Tuor, Aaron}, booktitle = {Proceedings of the 3rd Conference on Learning for Dynamics and Control}, pages = {980--991}, year = {2021}, editor = {Jadbabaie, Ali and Lygeros, John and Pappas, George J. and A. Parrilo, Pablo and Recht, Benjamin and Tomlin, Claire J. and Zeilinger, Melanie N.}, volume = {144}, series = {Proceedings of Machine Learning Research}, month = {07 -- 08 June}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v144/skomski21a/skomski21a.pdf}, url = {https://proceedings.mlr.press/v144/skomski21a.html}, abstract = {Recent works exploring deep learning application to dynamical systems modeling have demonstrated that embedding physical priors into neural networks can yield more effective, physically-realistic, and data-efficient models. However, in the absence of complete prior knowledge of a dynamical system’s physical characteristics, determining the optimal structure and optimization strategy for these models can be difficult. In this work, we explore methods for discovering neural state space dynamics models for system identification. Starting with a design space of block-oriented state space models and structured linear maps with strong physical priors, we encode these components into a model genome alongside network structure, penalty constraints, and optimization hyperparameters. Demonstrating the overall utility of the design space, we employ an asynchronous genetic search algorithm that alternates between model selection and optimization and obtains accurate physically consistent models of three physical systems: an aerodynamics body, a continuous stirred tank reactor, and a two tank interacting system.} }
Endnote
%0 Conference Paper %T Automating Discovery of Physics-Informed Neural State Space Models via Learning and Evolution %A Elliott Skomski %A Ján Drgoňa %A Aaron Tuor %B Proceedings of the 3rd Conference on Learning for Dynamics and Control %C Proceedings of Machine Learning Research %D 2021 %E Ali Jadbabaie %E John Lygeros %E George J. Pappas %E Pablo A. Parrilo %E Benjamin Recht %E Claire J. Tomlin %E Melanie N. Zeilinger %F pmlr-v144-skomski21a %I PMLR %P 980--991 %U https://proceedings.mlr.press/v144/skomski21a.html %V 144 %X Recent works exploring deep learning application to dynamical systems modeling have demonstrated that embedding physical priors into neural networks can yield more effective, physically-realistic, and data-efficient models. However, in the absence of complete prior knowledge of a dynamical system’s physical characteristics, determining the optimal structure and optimization strategy for these models can be difficult. In this work, we explore methods for discovering neural state space dynamics models for system identification. Starting with a design space of block-oriented state space models and structured linear maps with strong physical priors, we encode these components into a model genome alongside network structure, penalty constraints, and optimization hyperparameters. Demonstrating the overall utility of the design space, we employ an asynchronous genetic search algorithm that alternates between model selection and optimization and obtains accurate physically consistent models of three physical systems: an aerodynamics body, a continuous stirred tank reactor, and a two tank interacting system.
APA
Skomski, E., Drgoňa, J. & Tuor, A.. (2021). Automating Discovery of Physics-Informed Neural State Space Models via Learning and Evolution. Proceedings of the 3rd Conference on Learning for Dynamics and Control, in Proceedings of Machine Learning Research 144:980-991 Available from https://proceedings.mlr.press/v144/skomski21a.html.

Related Material