Papers
Topics
Authors
Recent
Search
2000 character limit reached

Pseudo-Hamiltonian system identification

Published 9 May 2023 in eess.SY, cs.LG, cs.NA, cs.SY, math.DS, and math.NA | (2305.06920v2)

Abstract: Identifying the underlying dynamics of physical systems can be challenging when only provided with observational data. In this work, we consider systems that can be modelled as first-order ordinary differential equations. By assuming a certain pseudo-Hamiltonian formulation, we are able to learn the analytic terms of internal dynamics even if the model is trained on data where the system is affected by unknown damping and external disturbances. In cases where it is difficult to find analytic terms for the disturbances, a hybrid model that uses a neural network to learn these can still accurately identify the dynamics of the system as if under ideal conditions. This makes the models applicable in some situations where other system identification models fail. Furthermore, we propose to use a fourth-order symmetric integration scheme in the loss function and avoid actual integration in the training, and demonstrate on varied examples how this leads to increased performance on noisy data.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (62)
  1. Noether networks: meta-learning useful conserved quantities. Advances in Neural Information Processing Systems, 34:16384–16397, 2021.
  2. V. I. Arnold. Mathematical methods of classical mechanics, volume 60 of Graduate Texts in Mathematics. Springer-Verlag, New York, second edition, 1989. Translated from the Russian by K. Vogtmann and A. Weinstein.
  3. Identification of port-Hamiltonian systems from frequency response data. Systems Control Lett., 143:104741, 10, 2020.
  4. Discovering governing equations from data by sparse identification of nonlinear dynamical systems. Proceedings of the national academy of sciences, 113(15):3932–3937, 2016.
  5. J. R. Cash and A. Singhal. High order methods for the numerical solution of two-point boundary value problems. BIT, 22(2):184–199, 1982.
  6. J. R. Cash and A. Singhal. Mono-implicit Runge-Kutta formulae for the numerical integration of stiff differential systems. IMA J. Numer. Anal., 2(2):211–227, 1982.
  7. Neural ordinary differential equations. In Advances in Neural Information Processing Systems, volume 31, pages 6572–6583, 2018.
  8. Symplectic recurrent neural networks. In International Conference on Learning Representations, 2020.
  9. A non-intrusive method to inferring linear port-Hamiltonian realizations using time-domain data. Electron. Trans. Numer. Anal., 56:102–116, 2022.
  10. Lagrangian neural networks. ICLR 2020 Workshop on Integration of Deep Neural Models and Differential Equations, 2020.
  11. Discovering symbolic models from deep learning with inductive biases. Advances in Neural Information Processing Systems, 33:17429–17442, 2020.
  12. G. Cybenko. Approximation by superpositions of a sigmoidal function. Mathematics of control, signals and systems, 2(4):303–314, 1989.
  13. M. David and F. Méhats. Symplectic learning for Hamiltonian neural networks. arXiv preprint arXiv:2106.11753, 2021.
  14. Variational integrator graph networks for learning energy-conserving dynamical systems. Physical Review E, 104(3):035310, 2021.
  15. Port-Hamiltonian neural networks for learning explicit time-dependent dynamical systems. Phys. Rev. E, 104:034312, 9 2021.
  16. Sparse symplectically integrated neural networks. Advances in Neural Information Processing Systems, 33:6074–6085, 2020.
  17. T. Duong and N. Atanasov. Hamiltonian-based neural ODE networks on the SE(3) manifold for dynamics learning and control. In Robotics: Science and Systems (RSS), 2021.
  18. T. Duong and N. Atanasov. Adaptive control of SE(3) Hamiltonian dynamics with learned disturbance features. IEEE Control Syst. Lett., 6:2773–2778, 2022.
  19. S. Eidnes and K. O. Lye. Pseudo-Hamiltonian neural networks for learning partial differential equations. arXiv preprint, arXiv:2304.14374, 2023.
  20. Pseudo-Hamiltonian neural networks with state-dependent external forces. Phys. D, 446:Paper No. 133673, 13, 2023.
  21. Ensemble-sindy: Robust sparse model discovery in the low-data, high-noise limit, with active learning and control. Proceedings of the Royal Society A, 478(2260):20210904, 2022.
  22. Simplifying Hamiltonian and Lagrangian neural networks via explicit constraints. Advances in neural information processing systems, 33:13880–13889, 2020.
  23. Hamiltonian neural networks. In H. Wallach, H. Larochelle, A. Beygelzimer, F. d'Alché-Buc, E. Fox, and R. Garnett, editors, Advances in Neural Information Processing Systems, volume 32. Curran Associates, Inc., 2019.
  24. Geometric numerical integration, volume 31 of Springer Series in Computational Mathematics. Springer-Verlag, Berlin, second edition, 2006. Structure-preserving algorithms for ordinary differential equations.
  25. M. Hénon and C. Heiles. The applicability of the third integral of motion: Some numerical experiments. Astronom. J., 69:73–79, 1964.
  26. K. Hornik. Approximation capabilities of multilayer feedforward networks. Neural networks, 4(2):251–257, 1991.
  27. Learning Poisson systems and trajectories of autonomous systems via Poisson neural networks. IEEE Transactions on Neural Networks and Learning Systems, 2022.
  28. SympNets: Intrinsic structure-preserving symplectic networks for identifying Hamiltonian systems. Neural Networks, 132:166–179, 2020.
  29. Sindy-pi: a robust algorithm for parallel implicit sparse identification of nonlinear dynamics. Proceedings of the Royal Society A, 476(2242):20200279, 2020.
  30. Sparse identification of nonlinear dynamics for model predictive control in the low-data limit. Proceedings of the Royal Society A, 474(2219):20180335, 2018.
  31. Physics-informed machine learning. Nature Reviews Physics, 3(6):422–440, 2021.
  32. Data-driven approximation of the Koopman generator: Model reduction, system identification, and control. Phys. D, 406:132416, 15, 2020.
  33. J. R. Koza. Genetic programming as a means for programming computers by natural selection. Statistics and computing, 4(2):87–112, 1994.
  34. Structure-preserving sparse identification of nonlinear dynamics for data-driven modeling. Proceedings of Mathematical and Scientific Machine Learning, 190:65–80, 2022.
  35. Z. Liu and M. Tegmark. Machine learning conservation laws from trajectories. Phys. Rev. Lett., 126(18):Paper No. 180604, 6, 2021.
  36. L. Ljung. System identification: theory for the user. Prentice Hall Information and System Sciences Series. Prentice Hall, Inc., Englewood Cliffs, NJ, 1987.
  37. PDE-Net 2.0: learning PDEs from data with a numeric-symbolic hybrid deep network. J. Comput. Phys., 399:108925, 17, 2019.
  38. Introduction to mechanics and symmetry, volume 17 of Texts in Applied Mathematics. Springer-Verlag, New York, second edition, 1999. A basic exposition of classical mechanical systems.
  39. Deep energy-based modeling of discrete-time physics. In H. Larochelle, M. Ranzato, R. Hadsell, M. F. Balcan, and H. Lin, editors, Advances in Neural Information Processing Systems, volume 33, pages 13100–13111. Curran Associates, Inc., 2020.
  40. T. Matsubara and T. Yaguchi. Finde: Neural differential equations for finding and preserving invariant quantities. arXiv preprint arXiv:2210.00272, 2022.
  41. Geometric integration using discrete gradients. R. Soc. Lond. Philos. Trans. Ser. A Math. Phys. Eng. Sci., 357(1754):1021–1045, 1999.
  42. Port-Hamiltonian dynamic mode decomposition. arXiv preprint arXiv:2204.13474, 2022.
  43. H. Noren. Learning Hamiltonian systems with mono-implicit Runge–Kutta methods. arXiv preprint, arXiv:2303.03769, 2023.
  44. C. Offen and S. Ober-Blöbaum. Symplectic integration of learned Hamiltonian systems. Chaos, 32(1):Paper No. 013122, 13, 2022.
  45. Universal differential equations for scientific machine learning. arXiv preprint arXiv:2001.04385, 2020.
  46. Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. J. Comput. Phys., 378:686–707, 2019.
  47. Data-driven discovery of partial differential equations. Science advances, 3(4):e1602614, 2017.
  48. H. Schaeffer. Learning partial differential equations via data discovery and sparse optimization. Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, 473(2197):20160446, 2017.
  49. M. Schmidt and H. Lipson. Distilling free-form natural laws from experimental data. science, 324(5923):81–85, 2009.
  50. M. Tao. Explicit symplectic approximation of nonseparable hamiltonians: Algorithm and long time performance. Physical Review E, 94(4):043303, 2016.
  51. S.-M. Udrescu and M. Tegmark. Ai feynman: A physics-inspired method for symbolic regression. Science Advances, 6(16):eaay2631, 2020.
  52. W. M. G. van Bokhoven. Efficient higher order implicit one-step methods for integration of stiff differential equations. BIT, 20(1):34–43, 1980.
  53. A. Van Der Schaft. Port-Hamiltonian systems: an introductory survey. In Proceedings of the international congress of mathematicians, volume 3, pages 1339–1365. Citeseer, 2006.
  54. A. Van Der Schaft and D. Jeltsema. Port-Hamiltonian systems theory: An introductory overview. Foundations and Trends in Systems and Control, 1(2-3):173–378, 2014.
  55. A. van der Schaft and B. Maschke. The Hamiltonian formulation of energy conserving physical systems with external ports. AEU. Archiv für Elektronik und Übertragungstechnik, 49(5-6):362–371, 1995.
  56. Integrating scientific knowledge with machine learning for engineering and environmental systems. ACM Comput. Surv., 55(4), nov 2022.
  57. H. Yoshida. Construction of higher order symplectic integrators. Phys. Lett. A, 150(5-7):262–268, 1990.
  58. GFINNs: GENERIC formalism informed neural networks for deterministic and stochastic dynamical systems. Philos. Trans. Roy. Soc. A, 380(2229):Paper No. 20210207, 21, 2022.
  59. Dissipative SymODEN: Encoding Hamiltonian dynamics with dissipation and control into deep learning. In ICLR 2020 Workshop on Integration of Deep Neural Models and Differential Equations, 2020.
  60. Symplectic ODE-net: Learning Hamiltonian dynamics with control. In International Conference on Learning Representations, 2020.
  61. Deep Hamiltonian networks based on symplectic integrators. arXiv preprint, arXiv:2004.13830, 2020.
  62. On numerical integration in neural ordinary differential equations. In International Conference on Machine Learning, pages 27527–27547. PMLR, 2022.
Citations (3)
View on Semantic Scholar

Summary

No one has generated a summary of this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Sign Up to Summarize

Paper to Video (Beta)

Whiteboard

No one has generated a whiteboard explanation for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Sign Up to Generate

Open Problems

We haven't generated a list of open problems mentioned in this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Generate Now

Continue Learning

We haven't generated follow-up questions for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Generate Now

Collections

Sign up for free to add this paper to one or more collections.

User Circle Single Streamline Icon: https://streamlinehq.com Sign Up