Corpus ID: 212644628

Lagrangian Neural Networks

@article{Cranmer2020LagrangianNN,
  title={Lagrangian Neural Networks},
  author={Miles D. Cranmer and Sam Greydanus and Stephan Hoyer and Peter W. Battaglia and David N. Spergel and Shirley Ho},
  journal={ArXiv},
  year={2020},
  volume={abs/2003.04630}
}
  • Miles D. Cranmer, Sam Greydanus, +3 authors Shirley Ho
  • Published 2020
  • Computer Science, Mathematics, Physics
  • ArXiv
  • Accurate models of the world are built upon notions of its underlying symmetries. In physics, these symmetries correspond to conservation laws, such as for energy and momentum. Yet even though neural network models see increasing use in the physical sciences, they struggle to learn these symmetries. In this paper, we propose Lagrangian Neural Networks (LNNs), which can parameterize arbitrary Lagrangians using neural networks. In contrast to models that learn Hamiltonians, LNNs do not require… CONTINUE READING

    Paper Mentions

    Citations

    Publications citing this paper.
    SHOWING 1-10 OF 14 CITATIONS

    Discovering Symbolic Models from Deep Learning with Inductive Biases

    VIEW 1 EXCERPT
    CITES METHODS

    Forward Prediction for Physical Reasoning

    VIEW 2 EXCERPTS
    CITES BACKGROUND

    Learning Dynamics Models with Stable Invariant Sets

    VIEW 1 EXCERPT
    CITES BACKGROUND

    Learning Physical Constraints with Neural Projections

    VIEW 3 EXCERPTS
    CITES BACKGROUND

    Lipschitz Recurrent Neural Networks

    VIEW 1 EXCERPT
    CITES BACKGROUND

    Meta-Learning Symmetries by Reparameterization

    VIEW 1 EXCERPT
    CITES BACKGROUND

    References

    Publications referenced by this paper.
    SHOWING 1-10 OF 21 REFERENCES

    Hamiltonian Neural Networks

    VIEW 5 EXCERPTS

    Distilling Free-Form Natural Laws from Experimental Data

    VIEW 1 EXCERPT
    HIGHLY INFLUENTIAL

    JAX: composable transformations of Python+NumPy programs

    • James Bradbury, Roy Frostig, +4 authors Skye Wanderman-Milne
    • 2018
    VIEW 4 EXCERPTS
    HIGHLY INFLUENTIAL

    JAX: composable transformations of Python+NumPy programs, 2018

    • James Bradbury, Roy Frostig, +4 authors Skye Wanderman-Milne
    • URL http://github.com/google/jax
    • 2018
    VIEW 4 EXCERPTS
    HIGHLY INFLUENTIAL

    Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification

    VIEW 1 EXCERPT

    Discovering Physical Equations with Graph Networks

    • Miles Cranmer, Alvaro Sanchez-Gonzalez, +4 authors Shirley Ho
    • 2020
    VIEW 2 EXCERPTS

    Discovering Physical Equations with Graph Networks. forthcoming

    • Miles Cranmer, Alvaro Sanchez-Gonzalez, +4 authors Shirley Ho
    • 2020
    VIEW 1 EXCERPT