Abstract:We present a data-driven and interpretable approach for reducing the dimensionality of chaotic systems using spectral submanifolds (SSMs). Emanating from fixed points or periodic orbits, these SSMs are low-dimensional inertial manifolds containing the chaotic attractor of the underlying high-dimensional system. The reduced dynamics on the SSMs turn out to predict chaotic dynamics accurately over a few Lyapunov times and also reproduce long-term statistical features, such as the largest Lyapunov exponents and probability distributions, of the chaotic attractor. We illustrate this methodology on numerical data sets including a delay-embedded Lorenz attractor, a nine-dimensional Lorenz model, and a Duffing oscillator chain. We also demonstrate the predictive power of our approach by constructing an SSM-reduced model from unforced trajectories of a buckling beam, and then predicting its periodically forced chaotic response without using data from the forced beam.

What problem does this paper attempt to address?

### Problems the Paper Attempts to Solve This paper aims to address the problem of dynamical modeling and prediction in high-dimensional chaotic systems. Specifically, the authors propose a data-driven method based on Spectral Submanifolds (SSMs) to reduce the dimensionality of chaotic systems. This method can accurately predict the short-term behavior of chaotic systems and reproduce the long-term statistical characteristics of chaotic attractors, such as the largest Lyapunov exponent and probability distribution. ### Background and Motivation 1. **High-dimensional chaotic systems**: Many real-world systems, such as double pendulums and turbulence, are high-dimensional, but their chaotic dynamics usually occur on low-dimensional attractors. 2. **Need for dimensionality reduction**: To understand and predict the chaotic behavior of these systems, it is necessary to reduce the high-dimensional systems to low-dimensional inertial manifolds. 3. **Limitations of existing methods**: Traditional linear dimensionality reduction techniques (such as Principal Component Analysis PCA, Dynamic Mode Decomposition DMD) cannot capture nonlinear chaotic phenomena. Nonlinear dimensionality reduction techniques (such as Kernel PCA, Local Linear Embedding LLE) have improved but still have certain limitations. ### Method Overview 1. **Spectral Submanifolds (SSMs)**: SSMs are low-dimensional inertial manifolds originating from fixed points or periodic orbits, containing chaotic attractors in high-dimensional systems. 2. **Data-driven modeling**: Construct SSMs and extract reduced-dimensional dynamical models through observational data. 3. **Predictive capability**: The SSM reduced-dimensional model can accurately predict the short-term behavior of chaotic systems within several Lyapunov times and reproduce the long-term statistical characteristics of chaotic attractors. ### Application Examples 1. **Delay-embedded Lorenz and Rössler attractors**: Reconstruct SSMs containing chaotic attractors from a single scalar time series using delay embedding techniques. 2. **Nine-dimensional Lorenz model**: Handle higher-dimensional Lorenz models to verify the effectiveness of the SSM dimensionality reduction method. 3. **Periodically forced Duffing oscillator chain**: Study nonlinear dynamical systems under periodic forcing. 4. **Kuramoto-Sivashinsky equation**: Deal with low-dimensional attractors in high-dimensional systems. 5. **Periodically forced buckled beam**: Construct SSM reduced-dimensional models from unforced trajectories to predict chaotic responses under periodic forcing. ### Main Contributions 1. **Data-driven SSM modeling**: Propose a data-driven method for constructing and predicting SSMs in high-dimensional chaotic systems. 2. **Short-term and long-term prediction**: The SSM reduced-dimensional model can accurately predict the short-term behavior of chaotic systems and reproduce their long-term statistical characteristics. 3. **Wide applicability**: This method is applicable to various chaotic systems, including classical Lorenz and Rössler attractors, as well as more complex high-dimensional systems. ### Conclusion By introducing a data-driven dimensionality reduction method based on SSMs, this paper provides a new effective tool for modeling and predicting high-dimensional chaotic systems. The application of this method to multiple numerical datasets demonstrates that the SSM reduced-dimensional model has good short-term predictive capability and the ability to reproduce long-term statistical characteristics.

Data-Driven Modeling and Forecasting of Chaotic Dynamics on Inertial Manifolds Constructed as Spectral Submanifolds

Data-driven modeling and prediction of non-linearizable dynamics via spectral submanifolds

Nonlinear Model Reduction to Temporally Aperiodic Spectral Submanifolds

Data-driven prediction of large-scale spatiotemporal chaos with distributed low-dimensional models

Data-driven reduced-order modeling of spatiotemporal chaos with neural ordinary differential equations

Data-driven discovery of intrinsic dynamics

Nonlinear model identification and spectral submanifolds for multi-degree-of-freedom mechanical vibrations

Nonlinear Model Reduction to Random Spectral Submanifolds in Random Vibrations

Learning Chaotic Dynamics with Embedded Dissipativity

Nonlinear Model Reduction to Fractional and Mixed-Mode Spectral Submanifolds

Model selection of chaotic systems from data with hidden variables using sparse data assimilation

Data-driven modeling of the regular and chaotic dynamics of an inverted flag from experiments

Data-driven linearization of dynamical systems

Modeling Nonlinear Dynamics from Videos

Data-driven modelling of the regular and chaotic dynamics of an inverted flag from experiments

Model reduction for nonlinearizable dynamics via delay-embedded spectral submanifolds

Model reduction to spectral submanifolds in piecewise smooth dynamical systems

Natural Exponential and Three‐Dimensional Chaotic System

Optimal Reconstruction of Vector Fields from Data for Prediction and Uncertainty Quantification