Koopman learning with episodic memory
- 1 January 2025
- journal article
- research article
- Published by AIP Publishing in Chaos: An Interdisciplinary Journal of Nonlinear Science
- Vol. 35 (1)
Abstract
Koopman operator theory has found significant success in learning models of complex, real-world dynamical systems, enabling prediction and control. The greater interpretability and lower computational costs of these models, compared to traditional machine learning methodologies, make Koopman learning an especially appealing approach. Despite this, little work has been performed on endowing Koopman learning with the ability to leverage its own failures. To address this, we equip Koopman methods—developed for predicting non-autonomous time series—with an episodic memory mechanism, enabling global recall of (or attention to) periods in time where similar dynamics previously occurred. We find that a basic implementation of Koopman learning with episodic memory leads to significant improvements in prediction on synthetic and real-world data. Our framework has considerable potential for expansion, allowing for future advances, and opens exciting new directions for Koopman learning.Keywords
All Related Versions
Funding Information
- Defense Sciences Office, DARPA (W31P4Q-21-C-0007)
- Air Force Office of Scientific Research (FA9550-22-1-0531)
This publication has 50 references indexed in Scilit:
- Transformers for modeling physical systemsNeural Networks, 2022
- Geometric considerations of a good dictionary for Koopman analysis of dynamical systems: Cardinality, “primary eigenfunction,” and efficient representationCommunications in Nonlinear Science and Numerical Simulation, 2021
- Spectrum of the Koopman Operator, Spectral Expansions in Functional Spaces, and State-Space GeometryJournal of Nonlinear Science, 2019
- Data-driven spectral decomposition and forecasting of ergodic dynamical systemsApplied and Computational Harmonic Analysis, 2019
- Delay-Coordinate Maps and the Spectra of Koopman OperatorsJournal of Statistical Physics, 2019
- Linear predictors for nonlinear dynamical systems: Koopman operator meets model predictive controlAutomatica, 2018
- Koopman Mode Decomposition for Periodic/Quasi-periodic Time Dependence**The funding provided by UTRC is greatly appreciated.IFAC-PapersOnLine, 2016
- A Data–Driven Approximation of the Koopman Operator: Extending Dynamic Mode DecompositionJournal of Nonlinear Science, 2015
- Spectral Properties of Dynamical Systems, Model Reduction and DecompositionsNonlinear Dynamics, 2005
- Nonlinear prediction of chaotic time seriesPhysica D: Nonlinear Phenomena, 1989