[HTML][HTML] Differentiability in unrolled training of neural physics simulators on transient dynamics
B List, LW Chen, K Bali, N Thuerey - Computer Methods in Applied …, 2025 - Elsevier
Unrolling training trajectories over time strongly influences the inference accuracy of neural
network-augmented physics simulators. We analyze these effects by studying three variants …
network-augmented physics simulators. We analyze these effects by studying three variants …
Predicting turbulent dynamics with the convolutional autoencoder echo state network
The dynamics of turbulent flows is chaotic and difficult to predict. This makes the design of
accurate reduced-order models challenging. The overarching objective of this paper is to …
accurate reduced-order models challenging. The overarching objective of this paper is to …
Prediction of chaotic dynamics and extreme events: A recurrence-free quantum reservoir computing approach
In chaotic dynamical systems, extreme events manifest in time series as unpredictable large-
amplitude peaks. Although deterministic, extreme events appear seemingly randomly, which …
amplitude peaks. Although deterministic, extreme events appear seemingly randomly, which …
Model scale versus domain knowledge in statistical forecasting of chaotic systems
W Gilpin - Physical Review Research, 2023 - APS
Chaos and unpredictability are traditionally synonymous, yet large-scale machine-learning
methods recently have demonstrated a surprising ability to forecast chaotic systems well …
methods recently have demonstrated a surprising ability to forecast chaotic systems well …
[HTML][HTML] Reconstruction, forecasting, and stability of chaotic dynamics from partial data
The forecasting and computation of the stability of chaotic systems from partial observations
are tasks for which traditional equation-based methods may not be suitable. In this …
are tasks for which traditional equation-based methods may not be suitable. In this …
How temporal unrolling supports neural physics simulators
B List, LW Chen, K Bali, N Thuerey - arXiv preprint arXiv:2402.12971, 2024 - arxiv.org
Unrolling training trajectories over time strongly influences the inference accuracy of neural
network-augmented physics simulators. We analyze these effects by studying three variants …
network-augmented physics simulators. We analyze these effects by studying three variants …
A 4D conservative chaotic system: Dynamics and realization
Z Yu, B Du, D Kong, Z Chai - Physica Scripta, 2024 - iopscience.iop.org
This paper proposes a novel four-dimensional conservative chaotic system (4D CCS) with a
simple algebraic representation, comprising only two quadratic nonlinear terms. The …
simple algebraic representation, comprising only two quadratic nonlinear terms. The …
Temperature modulation effects on chaos and heat transfer in Darcy–Bénard convection using a local thermal non-equilibrium model
This article investigates the effect of temperature modulation on convective heat transport in
a fluid-saturated porous layer under local thermal non-equilibrium (LTNE) conditions. The …
a fluid-saturated porous layer under local thermal non-equilibrium (LTNE) conditions. The …
Data-driven computation of adjoint sensitivities without adjoint solvers: An application to thermoacoustics
Adjoint methods have been the pillar of gradient-based optimization for decades. They
enable the accurate computation of a gradient (sensitivity) of a quantity of interest with …
enable the accurate computation of a gradient (sensitivity) of a quantity of interest with …
Neural networks for the prediction of chaos and turbulence
A Racca - 2023 - repository.cam.ac.uk
Chaos is a deterministic, yet unpredictable, phenomenon that appears in multiple
engineering applications. Predicting chaotic dynamics is challenging because infinitesimal …
engineering applications. Predicting chaotic dynamics is challenging because infinitesimal …