Learning reversible symplectic dynamics

Riccardo Valperga; Kevin Webster; Dmitry Turaev; Victoria Klein; Jeroen S.W. Lamb

Learning reversible symplectic dynamics

Riccardo Valperga, Kevin Webster, Dmitry Turaev, Victoria Klein, Jeroen S.W. Lamb

Research output: Contribution to journal › Conference article › peer-review

7 Scopus citations

Abstract

Time-reversal symmetry arises naturally as a structural property in many dynamical systems of interest. While the importance of hard-wiring symmetry is increasingly recognized in machine learning, to date this has eluded time-reversibility. In this paper we propose a new neural network architecture for learning time-reversible dynamical systems from data. We focus in particular on an adaptation to symplectic systems, because of their importance in physics-informed learning.

Original language	English
Pages (from-to)	906-916
Number of pages	11
Journal	Proceedings of Machine Learning Research
Volume	168
State	Published - 1 Jan 2022
Externally published	Yes
Event	4th Annual Learning for Dynamics and Control Conference, L4DC 2022 - Stanford, United States Duration: 23 Jun 2022 → 24 Jun 2022

Keywords

Physics-informed machine learning
dynamical systems
symplectic neural networks
time-reversal symmetry

ASJC Scopus subject areas

Artificial Intelligence
Software
Control and Systems Engineering
Statistics and Probability

Cite this

@article{30f47cb4365842acb485313fc7e9add4,

title = "Learning reversible symplectic dynamics",

abstract = "Time-reversal symmetry arises naturally as a structural property in many dynamical systems of interest. While the importance of hard-wiring symmetry is increasingly recognized in machine learning, to date this has eluded time-reversibility. In this paper we propose a new neural network architecture for learning time-reversible dynamical systems from data. We focus in particular on an adaptation to symplectic systems, because of their importance in physics-informed learning.",

keywords = "Physics-informed machine learning, dynamical systems, symplectic neural networks, time-reversal symmetry",

author = "Riccardo Valperga and Kevin Webster and Dmitry Turaev and Victoria Klein and Lamb, {Jeroen S.W.}",

note = "Publisher Copyright: {\textcopyright} 2022 R. Valperga, K. Webster, D. Turaev, V. Klein & J.S.W. Lamb.; 4th Annual Learning for Dynamics and Control Conference, L4DC 2022 ; Conference date: 23-06-2022 Through 24-06-2022",

year = "2022",

month = jan,

day = "1",

language = "English",

volume = "168",

pages = "906--916",

journal = "Proceedings of Machine Learning Research",

issn = "2640-3498",

}

TY - JOUR

T1 - Learning reversible symplectic dynamics

AU - Valperga, Riccardo

AU - Webster, Kevin

AU - Turaev, Dmitry

AU - Klein, Victoria

AU - Lamb, Jeroen S.W.

PY - 2022/1/1

Y1 - 2022/1/1

N2 - Time-reversal symmetry arises naturally as a structural property in many dynamical systems of interest. While the importance of hard-wiring symmetry is increasingly recognized in machine learning, to date this has eluded time-reversibility. In this paper we propose a new neural network architecture for learning time-reversible dynamical systems from data. We focus in particular on an adaptation to symplectic systems, because of their importance in physics-informed learning.

AB - Time-reversal symmetry arises naturally as a structural property in many dynamical systems of interest. While the importance of hard-wiring symmetry is increasingly recognized in machine learning, to date this has eluded time-reversibility. In this paper we propose a new neural network architecture for learning time-reversible dynamical systems from data. We focus in particular on an adaptation to symplectic systems, because of their importance in physics-informed learning.

KW - Physics-informed machine learning

KW - dynamical systems

KW - symplectic neural networks

KW - time-reversal symmetry

UR - http://www.scopus.com/inward/record.url?scp=85140411677&partnerID=8YFLogxK

M3 - Conference article

AN - SCOPUS:85140411677

SN - 2640-3498

VL - 168

SP - 906

EP - 916

JO - Proceedings of Machine Learning Research

JF - Proceedings of Machine Learning Research

T2 - 4th Annual Learning for Dynamics and Control Conference, L4DC 2022

Y2 - 23 June 2022 through 24 June 2022

ER -

Learning reversible symplectic dynamics

Abstract

Keywords

ASJC Scopus subject areas

Other files and links

Fingerprint

Cite this