ML ICRITERIA: Difference between revisions

From VASP Wiki
No edit summary
mNo edit summary
(18 intermediate revisions by 2 users not shown)
Line 1: Line 1:
{{TAGDEF|ML_FF_LCRITERIA|[logical]|.TRUE.}}
{{TAGDEF|ML_FF_LCRITERIA|[logical]|.TRUE.}}


Description: Decides whether the threshold in the learning decision step for the Bayesian error estimation is renewed or not in the machine learning force field methods
Description: Decides whether the threshold ({{TAG|ML_FF_CTIFOR}}) is updated in the machine learning force field methods. {{TAG|ML_FF_CTIFOR}} determines whether a first principles calculations is performed.
----
----


This flag is only used if Bayesian error estimation is switched on ({{TAG|ML_FF_IERR}}=2 or 3).
This flag is only used if Bayesian error estimation is switched on ({{TAG|ML_FF_IERR}}=2 or 3, 3 is the default). Generally it is recommended to automatically update the criteria {{TAG|ML_FF_CTIFOR}} during machine learning. Details on how and when the update is performed are controlled by {{TAG|ML_FF_CSLOPE}}, {{TAG|ML_FF_CSIG}} and {{TAG|ML_FF_MHIS}}.
 
{{TAG|ML_FF_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  {{TAG|ML_FF_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  {{TAG|ML_FF_CSIG}}, for the update to take place. Furthermore, the slope of the stored data must be below the threshold  {{TAG|ML_FF_CSLOPE}} (we recommend to set only  {{TAG|ML_FF_CSIG}}).
 
If the previous conditions are met, the criteria {{TAG|ML_FF_CTIFOR}} is updated. To avoid too abrupt changes the average Bayesian error is mixed with the current value of  {{TAG|ML_FF_CTIFOR}}. The mixing ratio can be determined by the tag {{TAG|ML_FF_XMIX}}.


== Related Tags and Sections ==
== Related Tags and Sections ==
{{TAG|ML_FF_LMLFF}}, {{TAG|ML_FF_IERR}, {{TAG|ML_FF_ISAMPLE}}
{{TAG|ML_FF_LMLFF}}, {{TAG|ML_FF_IERR}}, {{TAG|ML_FF_CTIFOR}},  {{TAG|ML_FF_CSLOPE}}, {{TAG|ML_FF_CSIG}}, {{TAG|ML_FF_MHIS}}, {{TAG|ML_FF_XMIX}}


{{sc|ML_FF_LCRITERIA|Examples|Examples that use this tag}}
{{sc|ML_FF_LCRITERIA|Examples|Examples that use this tag}}
----
----


[[Category:INCAR]][[Category:Machine Learning]][[Category:Machine Learned Force Fields]][[Categor:VASP6]]
[[Category:INCAR]][[Category:Machine Learning]][[Category:Machine Learned Force Fields]][[Category: Alpha]]

Revision as of 10:51, 16 April 2021

ML_FF_LCRITERIA = [logical]
Default: ML_FF_LCRITERIA = .TRUE. 

Description: Decides whether the threshold (ML_FF_CTIFOR) is updated in the machine learning force field methods. ML_FF_CTIFOR determines whether a first principles calculations is performed.


This flag is only used if Bayesian error estimation is switched on (ML_FF_IERR=2 or 3, 3 is the default). Generally it is recommended to automatically update the criteria ML_FF_CTIFOR during machine learning. Details on how and when the update is performed are controlled by ML_FF_CSLOPE, ML_FF_CSIG and ML_FF_MHIS.

ML_FF_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_FF_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_FF_CSIG, for the update to take place. Furthermore, the slope of the stored data must be below the threshold ML_FF_CSLOPE (we recommend to set only ML_FF_CSIG).

If the previous conditions are met, the criteria ML_FF_CTIFOR is updated. To avoid too abrupt changes the average Bayesian error is mixed with the current value of ML_FF_CTIFOR. The mixing ratio can be determined by the tag ML_FF_XMIX.

Related Tags and Sections

ML_FF_LMLFF, ML_FF_IERR, ML_FF_CTIFOR, ML_FF_CSLOPE, ML_FF_CSIG, ML_FF_MHIS, ML_FF_XMIX

Examples that use this tag