boltzhist_cur

Contents

boltzhist_cur#

autoplex.data.common.utils.boltzhist_cur(atoms, descriptor, isol_es, bolt_frac=0.1, bolt_max_num=3000, cur_num=100, kernel_exp=4, kT=0.3, energy_label='energy', P=None, random_seed=None)[source]#

Sample atoms from a list according to boltzmann energy weighting and CUR diversity.

Parameters:

  • atoms (list of ase.Atoms) – The atoms from which to select. If this is a list of lists, it is flattened.

  • descriptor (str) – The quippy SOAP descriptor string for CUR.

  • isol_es (dict) – Dictionary of isolated energy values for species.

  • bolt_frac (float) – The fraction to control the flat Boltzmann selection number.

  • bolt_max_num (int) – The maximum number of atoms to select by Boltzmann-weighted flat histogram.

  • cur_num (int) – The number of atoms to select by CUR.

  • kernel_exp (float) – The exponent for the dot-product SOAP kernel.

  • kT (float) – The product of the Boltzmann constant and the temperature, in eV.

  • energy_label (str) – The label for the energy property in the atoms.

  • P (list of float) – The pressures at which the atoms have been optimized, in GPa.

  • random_seed (int) – The seed for the random number generator.

Returns:

The selected atoms. These are copies of the atoms in the input list.

Return type:

list of ase.Atoms

Notes

This function selects the most diverse atoms based on the chosen algorithm. The algorithm uses a combination of CUR selection and Boltzmann weighting to select the atoms.

Adapted from: * Title: Research data supporting “De novo exploration and self-guided learning of potential-energy surfaces” * Script: select_enthalpy_flat_histogram.py * Author: Noam Bernstein, Gábor Csányi and Volker L. Deringer * Date 11/10/2019 * Availability: https://www.repository.cam.ac.uk/items/3aff252b-a583-4e7c-afc9-9dc1540cc37e * License: Attribution 4.0 International (CC BY 4.0) license.

Contents