ML SION2: Difference between revisions

From VASP Wiki
No edit summary
No edit summary
Line 1: Line 1:
{{TAGDEF|ML_SION2|[real]|1.5*{{TAG|ML_SION1}}}}
{{TAGDEF|ML_SION2|[real]|{{TAG|ML_SION1}}}}


Description: This tag specifies the width <math>\sigma_\text{atom}</math> of the Gaussian functions used for broadening the atomic distributions of the angular descriptor <math>\rho^{(3)}_i(r)</math> within the machine learning force field method (see [[Machine learning force field: Theory#Descriptors|this section]]).
Description: This tag specifies the width <math>\sigma_\text{atom}</math> of the Gaussian functions used for broadening the atomic distributions of the angular descriptor <math>\rho^{(3)}_i(r)</math> within the machine learning force field method (see [[Machine learning force field: Theory#Descriptors|this section]]).
----
----
The unit of {{TAG|ML_SION2}} is <math>\AA</math>.
The unit of {{TAG|ML_SION2}} is <math>\AA</math>.
{{BOX|tip|Our test calculations indicate that a ratio {{TAG|ML_SION2}} / {{TAG|ML_SION1}} {{=}} 1.5 results in an optimal training performance. Furthermore, a value of 0.5 was found to be a good default value for {{TAG|ML_SION2}}. Both findings together result in the default value for {{TAG|ML_SION1}}.|Background:}}
{{BOX|note|Our test calculations indicate that {{TAG|ML_SION1}} {{=}} {{TAG|ML_SION2}} results in an optimal training performance. Furthermore, a value of 0.5 was found to be a good default value for both. However, the best choice is system-dependent, careful testing may improve machine learning results.}}


== Related Tags and Sections ==
== Related Tags and Sections ==

Revision as of 20:26, 10 October 2021

ML_SION2 = [real]
Default: ML_SION2 = ML_SION1 

Description: This tag specifies the width of the Gaussian functions used for broadening the atomic distributions of the angular descriptor within the machine learning force field method (see this section).


The unit of ML_SION2 is .

Our test calculations indicate that ML_SION1 = ML_SION2 results in an optimal training performance. Furthermore, a value of 0.5 was found to be a good default value for both. However, the best choice is system-dependent, careful testing may improve machine learning results.

Related Tags and Sections

ML_LMLFF, ML_SION1, ML_RCUT1, ML_RCUT2, ML_MRB1, ML_MRB2

Examples that use this tag