Symmetry-Adapted Machine Learning for Tensorial Properties of Atomistic Systems

Authors: Andrea Grisafi1, David M. Wilkins1*, Gabor Csányi2, Michele Ceriotti1

  1. Laboratory of Computational Science and Modeling, IMX, École Polytechnique Fédérale de Lausanne, 1015 Lausanne, Switzerland
  2. Engineering Laboratory, University of Cambridge, Trumpington Street, Cambridge CB21PZ, United Kingdom
  • Corresponding author email:

DOI10.24435/materialscloud:2018.0009/v1 (version v1, submitted on 19 May 2018)

How to cite this entry

Andrea Grisafi, David M. Wilkins, Gabor Csányi, Michele Ceriotti, Symmetry-Adapted Machine Learning for Tensorial Properties of Atomistic Systems, Materials Cloud Archive (2018), doi: 10.24435/materialscloud:2018.0009/v1.


Here we present 1,000 structures each of a water monomer, water dimer, Zundel cation and bulk water used to train tensorial machine-learning models in Phys. Rev. Lett. 120, 036002 (2018). The archive entry contains files in extended-XYZ format including the structures and several tensorial properties: for the monomer, dimer and Zundel cation, the dipole moment, polarizability and first hyperpolarizability are included, and for bulk water the dipole moment, polarizability and dielectric tensor are given.

Materials Cloud sections using this data

No Explore or Discover sections associated with this archive entry.


File name Size Description
MD5MD5: 2b977c3fc6b23f7ff894777c7bacc9cc
1.3 KiB Description of the content and format of the other files.
MD5MD5: efa6e057541497f86eeffa603cdaa539
583.0 KiB Structures and tensors for water monomers.
MD5MD5: 830dc887aa0562669d4eea80db4365f6
674.5 KiB Structures and tensors for water dimers.
MD5MD5: 1cee4b970d00be9893f06cbc0bda4adf
701.1 KiB Structures and tensors for Zundel cation.
MD5MD5: 7f79f3dbc9758210c8f2402e6742663a
3.3 MiB Structures and tensors for bulk water.


Files and data are licensed under the terms of the following license: Creative Commons Attribution 4.0 International.


water molecular bulk dipole moment polarizability hyperpolarizability dielectric tensor symmetry-adapted gaussian process regression machine learning

Version history

19 May 2018 [This version]