Structure-preserving neural networks

@article{Hernandez2020StructurepreservingNN,
  title={Structure-preserving neural networks},
  author={Quercus Hernandez and Alberto Badias and D. Gonz{\'a}lez and F. Chinesta and E. Cueto},
  journal={ArXiv},
  year={2020},
  volume={abs/2004.04653}
}
  • Quercus Hernandez, Alberto Badias, +2 authors E. Cueto
  • Published 2020
  • Computer Science, Physics, Mathematics
  • ArXiv
  • We develop a method to learn physical systems from data that employs feedforward neural networks and whose predictions comply with the first and second principles of thermodynamics. The method employs a minimum amount of data by enforcing the metriplectic structure of dissipative Hamiltonian systems in the form of the so-called General Equation for the Non-Equilibrium Reversible-Irreversible Coupling, GENERIC [M. Grmela and H.C Oettinger (1997). Dynamics and thermodynamics of complex fluids. I… CONTINUE READING
    4 Citations

    Figures from this paper.

    Deep learning of thermodynamics-aware reduced-order models from data
    • 1
    • Highly Influenced
    • PDF
    Nonseparable Symplectic Neural Networks
    Symplectic Neural Networks in Taylor Series Form for Hamiltonian Systems
    • 7
    • PDF
    Neural Vortex Method: from Finite Lagrangian Particles to Infinite Dimensional Eulerian Dynamics

    References

    SHOWING 1-10 OF 51 REFERENCES
    Thermodynamically consistent data-driven computational mechanics
    • 31
    • PDF
    On learning Hamiltonian systems from data.
    • 20
    • PDF
    Learning in Modal Space: Solving Time-Dependent Stochastic PDEs Using Physics-Informed Neural Networks
    • 22
    • PDF
    Thermodynamically consistent time‐stepping algorithms for non‐linear thermomechanical systems
    • 77
    Dynamics and thermodynamics of complex fluids. II. Illustrations of a general formalism
    • 548
    Constraint-aware neural networks for Riemann problems
    • 11
    • PDF
    Runge-Kutta neural network for identification of dynamical systems in high accuracy
    • 84
    • PDF