ML CTIFOR: Difference between revisions

Latest revision as of 14:40, 18 December 2024

Default: ML_CTIFOR	= 0.002	if ML_CALGO = 0
	= 0.02	if ML_CALGO = 1

Description: This flag sets the threshold for the error estimation in the machine learning force field method.

The use of this tag in combination with the learning algorithms is described here: here. Generally, first principles calculations are only performed if the error estimate of one force exceeds the threshold.

The initial threshold is set to the value provided by the tag ML_CTIFOR (units of eV/Angstrom for ML_CALGO=0 and unitless for ML_CALGO=1).

For ML_CALGO=0, the threshold can be updated dynamically during ML. The details of the update are controlled by ML_ICRITERIA. Typically, after extensive training, attainable values for ML_CTIFOR are 0.02 around 300-500 K, and 0.06 around 1000-2000 K, so temperature but also system dependent. The initial default 0.002 is only sensible, if ML_CTIFOR is automatically updated (ML_ICRITERIA = 1 or 2). If ML_ICRITERIA = 0 is used, it is necessary to use significantly larger values around 0.02-0.06 for ML_CTIFOR.

For ML_CALGO=1, only a constant threshold during the calculation is available (ML_ICRITERIA=0).

The related tag ML_SCLC_CTIFOR determines how many local reference configurations are chosen from each first principles calculations.

@@ Line 1: / Line 1: @@
-{{TAGDEF|ML_FF_CTIFOR|[real]|<math> 10^{-16}</math>}}
+{{DISPLAYTITLE:ML_CTIFOR}}
+{{DEF|ML_CTIFOR|0.002|if {{TAG|ML_CALGO}} {{=}} 0|0.02|if {{TAG|ML_CALGO}} {{=}} 1}}
-Description: This flag sets the threshold for the Bayesian error estimation on the force in the machine learning force field method.
+Description: This flag sets the threshold for the error estimation in the machine learning force field method.
 ----
+The use of this tag in combination with the learning algorithms is described here: [[Machine learning force field calculations: Basics#Sampling of training data and local reference configurations|here]]. Generally, first principles calculations are only performed if the error estimate of one force exceeds the threshold.
-Within the learning step if a newly considered structure gives an error in the force larger than {{TAG|ML_FF_CTIFOR}}, the sampling is stopped and a new force field is generated. This flag is only used if {{TAG|ML_FF_IERR}}=2 or 3. The unit of {{TAG|ML_FF_CTIFOR}} is eV/Angstrom.
+The initial threshold is set to the value provided by the tag {{TAG|ML_CTIFOR}} (units of eV/Angstrom for {{TAG|ML_CALGO}}=0 and unitless for {{TAG|ML_CALGO}}=1).
-== Related Tags and Sections ==
+For {{TAG|ML_CALGO}}=0, the threshold can be updated dynamically during ML. The details of the update are controlled by {{TAG|ML_ICRITERIA}}. Typically, after extensive training, attainable values for ML_CTIFOR are 0.02 around 300-500 K, and 0.06 around 1000-2000 K, so temperature but also system dependent. The initial default 0.002 is only sensible, if {{TAG|ML_CTIFOR}}  is automatically updated ({{TAG|ML_ICRITERIA}} = 1 or 2). If  {{TAG|ML_ICRITERIA}} = 0 is used, it is necessary to use significantly larger values around 0.02-0.06 for {{TAG|ML_CTIFOR}}.
-{{TAG|ML_FF_LMLFF}}, {{TAG|ML_FF_IERR}}, {{TAG|ML_FF_CSF}}
-{{sc|ML_FF_C[[Category:VASP6]]
+For {{TAG|ML_CALGO}}=1, only a constant threshold during the calculation is available ({{TAG|ML_ICRITERIA}}=0).
-TIFOR|Examples|Examples that use this tag}}
+The related tag {{TAG|ML_SCLC_CTIFOR}} determines how many local reference configurations are chosen from each first principles calculations.
+== Related tags and articles ==
+{{TAG|ML_LMLFF}}, {{TAG|ML_ICRITERIA}}, {{TAG|ML_CALGO}}, {{TAG|ML_SCLC_CTIFOR}} , {{TAG|ML_MHIS}}, {{TAG|ML_CSIG}}, {{TAG|ML_CSLOPE}}, {{TAG|ML_CDOUB}}, {{TAG|ML_CX}}, {{TAG|ML_NMDINT}}, {{TAG|ML_MCONF_NEW}}
+{{sc|ML_CTIFOR|Examples|Examples that use this tag}}
 ----
-[[Category:INCAR]][[Category:Machine Learning]][[Category:Machine Learned Force Fields]][[Category: Alpha]]
+[[Category:INCAR tag]][[Category:Machine-learned force fields]]