[Dataset submission | request] aleatoric_epistemic_error_AIC2023

jvita commented 11 months ago

Name

Josh Vita

Email

vita1@llnl.gov

Dataset name

aleatoric_epistemic_error_AIC2023

Authors

Sugata Goswami, Silvan Käser, Raymond J. Bemish, Markus Meuwly

Publication link

https://www.sciencedirect.com/science/article/pii/S2949747723000337

Data link

https://github.com/MMunibas/noise

Additional links

Dataset description

Datasets for H2CO and HONO with, and without, added noise for testing the effects of noise on quality of fit

File details

(From GitHub repository for data) The data are saved in compressed numpy binary (.npz) files with seven numpy arrays. These are described below along with the file names.

File, “ch2o_3601-RKHS.npz” contains the clean RKHS energies
File, “ch2o_3601-RKHS-gaussian-noise-SD-0-0000001.npz” contains the perturbed energies with amplitude of 10-7 eV (2.31 X 10-6 kcal/mol)
File, “ch2o_3601-RKHS-gaussian-noise-SD-0-000001.npz” contains the perturbed energies with amplitude of 10-6 eV (2.31 X 10-5 kcal/mol)
File, “ch2o_3601-RKHS-gaussian-noise-SD-0-00001.npz” contains the perturbed energies with amplitude of 10-5 eV (2.31 X 10-4 kcal/mol)
File, “ch2o_3601-RKHS-force-energy.npz” contains clean RKHS energies and forces.
File, “ch2o_3601-RKHS-gaussian-noise-force-SD-10-7.npz” contains clean RKHS energies and noisy forces with Gaussian noise with amplitude of 10-7 eV/angstrom (2.31 X 10-6 kcal/mol/angstrom)
File, “ch2o_3601-RKHS-gaussian-noise-force-SD-10-6.npz” contains clean RKHS energies and noisy forces with Gaussian noise with amplitude of 10-6 eV/angstrom (2.31 X 10-5 kcal/mol/angstrom)
File, “ch2o_3601-RKHS-gaussian-noise-force-SD-10-5.npz” contains clean RKHS energies and noisy forces with Gaussian noise with amplitude of 10-5 eV/angstrom (2.31 X 10-4 kcal/mol/angstrom)

Method

CCSD(T)

Method (other)

No response

Software

None

Software (other)

No response

Software version(s)

No response

Additional details

(From GitHub repository for data) In addition to the energies (eV, stored as/in ‘E’) and forces (eV/angstrom, stored as/in ‘F’), the files also contain information regarding dipole moment (stored as/in ‘D’) of the molecule, number of atoms (stored as/in ‘N’) present in the molecule, charge (stored as/in ‘Q’) of the molecule, coordinates of the atoms (stored as/in ‘R’), thus the geometry of the molecule and atomic numbers of the atoms (stored as/in ‘Z’).

Property types

Atomic forces, Potential energy

Other/additional property

No response

Property details

No response

Elements

No response

Number of Configurations

No response

Naming convention

No response

Configuration sets

No response

Configuration labels

No response

Distribution license

No response

Permissions

[ ] I confirm that I have the necessary permissions to submit this dataset

gpwolfe commented 11 months ago

Have reached out to see if GH repo is still private (possibly accidentally)

gpwolfe commented 3 months ago

Repo has since been made public.

gpwolfe commented 1 month ago

The GitHub link contains the CH2O data only. The HONO data used in the publication is available at https://zenodo.org/records/4585449 -- will be ingested as part of a separate dataset (VibML)

gpwolfe commented 1 month ago

Now available at colabfit.org pending export of dataset files

colabfit / data-lake