Requests for technical support from the VASP group should be posted in the VASP-forum.

Category:Machine Learning

From Vaspwiki
Revision as of 12:31, 21 August 2019 by Schlipf (talk | contribs)
Jump to navigationJump to search

This section describes how to use machine learning to generate force fields. First, one should become familiar with the theoretical background On-the-fly machine learning force field generation using Bayesian linear regression. Secondly, one can focus on how to set up and run the machine learning calculations within VASP Machine learning force field calculations: Basics. Finally, one gains some hands-on experience with the following tutorial: Liquid Si - MLFF.

Input

  • Besides the usual input files (INCAR, POSCAR, etc.) the machine learning force field method requires the following input files:
    • ML_ABCAR Ab initio data used to create the training data.
    • ML_FFCAR File containing force field parameters.

Output

  • The machine learning force field method generates the following output files:
    • ML_LOGFILE Main output file for the machine learning force field method.
    • ML_ABNCAR New abinitio data (used as ABCAR in the next run).
    • ML_REGCAR Output file summarizing regression results.
    • ML_HISCAR Output file summarizing the histogram data.
    • ML_FFNCAR File containing new force field parameters (used as FFCAR in the next run).
    • ML_EATOM Output file containing local atomic energies
    • ML_HEAT Output file including local heat flux


All INCAR tags belonging to the machine learning force field method start with the prefix ML_FF_. Input tags that are related to the many-body term end with _MB.

Theoretical Background

How to

Tutorial

  • Basic tutorial on how to perform on-the-fly learning and how to control the accuracy of the force field: Liquid Si - MLFF.