ML ICRITERIA: Difference between revisions

From VASP Wiki
No edit summary
No edit summary
Line 6: Line 6:
* {{TAG|ML_ICRITERIA}} = 0: No update of the threshold {{TAG|ML_CTIFOR}} is done.
* {{TAG|ML_ICRITERIA}} = 0: No update of the threshold {{TAG|ML_CTIFOR}} is done.
* {{TAG|ML_ICRITERIA}} = 1: Update of criteria using average of the Bayesian errors of the forces from history (see description of method below).
* {{TAG|ML_ICRITERIA}} = 1: Update of criteria using average of the Bayesian errors of the forces from history (see description of method below).
* {{TAG|ML_ICRITERIA}} = 2: Update of criteria using gliding average of Bayesian errors (experimental and not well tested).  
* {{TAG|ML_ICRITERIA}} = 2: Update of criteria using gliding average of Bayesian errors ('''experimental and not well tested''').  


Generally it is recommended to automatically update the threshold {{TAG|ML_CTIFOR}} during machine learning. Details on how and when the update is performed are controlled by {{TAG|ML_CSLOPE}}, {{TAG|ML_CSIG}} and {{TAG|ML_MHIS}}.
Generally it is recommended to automatically update the threshold {{TAG|ML_CTIFOR}} during machine learning. Details on how and when the update is performed are controlled by {{TAG|ML_CSLOPE}}, {{TAG|ML_CSIG}} and {{TAG|ML_MHIS}}.

Revision as of 08:44, 23 August 2021

ML_ICRITERIA = [integer]
Default: ML_ICRITERIA = 1 

Description: Decides whether ([[ML_ICRITERIA>0]]) or how the Bayesian error threshold (ML_CTIFOR) is updated in the machine learning force field methods. ML_CTIFOR determines whether a first principles calculations is performed.


The following options are possible for ML_ICRITERIA:

  • ML_ICRITERIA = 0: No update of the threshold ML_CTIFOR is done.
  • ML_ICRITERIA = 1: Update of criteria using average of the Bayesian errors of the forces from history (see description of method below).
  • ML_ICRITERIA = 2: Update of criteria using gliding average of Bayesian errors (experimental and not well tested).

Generally it is recommended to automatically update the threshold ML_CTIFOR during machine learning. Details on how and when the update is performed are controlled by ML_CSLOPE, ML_CSIG and ML_MHIS.

Description of ML_ICRITERIA=1:

ML_CTIFOR is generally set to the average of the Bayesian errors of the forces stored in a history. The number of entries in the history are controlled by ML_MHIS. To avoid that noisy data or an abrupt jump of the Bayesian error causes issues, the standard error of the history must be below the threshold ML_CSIG, for the update to take place. Furthermore, the slope of the stored data must be below the threshold ML_CSLOPE (we recommend to set only ML_CSIG).

If the previous conditions are met, the threshold ML_CTIFOR is updated. To avoid too abrupt changes the average Bayesian error can be mixed with the current value of ML_CTIFOR. The mixing ratio can be determined by the tag ML_XMIX (default is no mixing).

Related Tags and Sections

ML_LMLFF, ML_CTIFOR, ML_CSLOPE, ML_CSIG, ML_MHIS, ML_XMIX

Examples that use this tag