Cormorant: Covariant Molecular Neural Networks

By Brandon M. Anderson, Truong Son Hy, Risi Kondor. 2019

498 citations recorded by the source corpus.

We propose Cormorant, a rotationally covariant neural network architecture for learning the behavior and properties of complex many-body physical systems. We apply these networks to molecular systems with two goals: learning atomic potential energy surfaces for use in Molecular Dynamics simulations, and learning ground state properties of molecules calculated by Density Functional Theory. Some of the key features of our network are that (a) each neuron explicitly corresponds to a subset of atoms; (b) the activation of each neuron is covariant to rotations, ensuring that overall the network is fully rotationally invariant. Furthermore, the non-linearity in our network is based upon tensor products and the Clebsch-Gordan decomposition, allowing the network to operate entirely in Fourier space. Cormorant significantly outperforms competing algorithms in learning molecular Potential Energy Surfaces from conformational geometries in the MD-17 dataset, and is competitive with other methods at learning geometric, energetic, electronic, and thermodynamic properties of molecules on the GDB-9 dataset.

Subfields: Computational Theory and Mathematics, Materials Chemistry, Molecular Biology, Neural Architecture Search & AutoML, Computational Chemistry & Drug Discovery

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Theory and Mathematics

  1. Tess E. Smidt
  2. Gisbert Schneider
  3. Michael Gastegger
  4. Francesca Grisoni
  5. K. Müller
  6. Erik J. Bekkers
  7. Kristof T. Schütt
  8. Mario Geiger

In Materials Chemistry

  1. Tess E. Smidt
  2. K. Müller
  3. Mario Geiger
  4. Michael Gastegger
  5. Max Welling
  6. Kristof T. Schütt
  7. Oliver T. Unke
  8. Yi Liao

In Molecular Biology

  1. Tess E. Smidt
  2. Max Welling
  3. K. Müller
  4. Michael Gastegger
  5. Stephan Günnemann
  6. Gisbert Schneider
  7. Francesca Grisoni
  8. Oliver T. Unke

In Neural Architecture Search & AutoML

  1. Maurice Weiler
  2. Max Welling
  3. Ron O. Dror
  4. Tess E. Smidt
  5. Stephan Eismann
  6. Raphael J. L. Townshend
  7. Mario Geiger
  8. Gabriele Cesa

In Computational Chemistry & Drug Discovery

  1. Tess E. Smidt
  2. K. Müller
  3. Mario Geiger
  4. Stephan Günnemann
  5. Johannes Klicpera
  6. Kristof T. Schütt
  7. Michael Gastegger
  8. Oliver T. Unke

Citations in the corpus