Neural Dynamical Systems

Viraj Mehta, Ian Char, Willie Neiswanger, Youngseog Chung, Andrew Oakleigh Nelson, Mark D. Boyer, Egemen Kolemen, and Jeff Schneider

Workshop Paper, ICLR '20 Workshop on Integration of Deep Neural Models and Differential Equations, April, 2020

Abstract

We introduce Neural Dynamical Systems (NDS), a method of learning dynamical models which incorporates prior knowledge in the form of systems of ordinary differential equations. NDS uses neural models to estimate free parameters of the system, predicts residual terms, and numerically integrates over time to predict future states. It also natively handles irregularly sampled data and implicitly learns values of interpretable system parameters. We find that NDS learns dynamics with higher accuracy and fewer samples than a variety of deep learning methods that do not incorporate the prior knowledge. We demonstrate these advantages first on synthetic dynamical systems and then on real data captured from deuterium shots from a nuclear fusion reactor.

BibTeX

@workshop{Mehta-2020-126321,
author = {Viraj Mehta and Ian Char and Willie Neiswanger and Youngseog Chung and Andrew Oakleigh Nelson and Mark D. Boyer and Egemen Kolemen and Jeff Schneider},
title = {Neural Dynamical Systems},
booktitle = {Proceedings of ICLR '20 Workshop on Integration of Deep Neural Models and Differential Equations},
year = {2020},
month = {April},
}

Copyright notice: This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each author's copyright. These works may not be reposted without the explicit permission of the copyright holder.