×

Recommended by

Indexed by

Fixed node diffusion Monte Carlo energies for over one thousand small organic molecules

Bing Huang1*, Anatole von Lilienfeld2,3*, Jaron Krogel4*, Anouar Benali5*

1 Faculty of Physics, University of Vienna, 1090 Wien, Austria

2 Departments of Chemistry, Materials Science and Engineering, and Physics, University of Toronto, St. George Campus, Toronto, ON, Canada

3 Machine Learning Group, Technische Universit\"at Berlin and Institute for the Foundations of Learning and Data, 10587 Berlin, Germany

4 Materials Science and Technology Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831, United States

5 Computational Science Division, Argonne National Laboratory, Argonne, IL 60439, United States

* Corresponding authors emails: bing.huang@univie.ac.at, anatole.vonlilienfeld@utoronto.ca, krogeljt@ornl.gov, benali@anl.gov
DOI10.24435/materialscloud:9p-8a [version v2]

Publication date: Nov 15, 2023

How to cite this record

Bing Huang, Anatole von Lilienfeld, Jaron Krogel, Anouar Benali, Fixed node diffusion Monte Carlo energies for over one thousand small organic molecules, Materials Cloud Archive 2023.174 (2023), https://doi.org/10.24435/materialscloud:9p-8a

Description

In the past decade, quantum diffusion Monte Carlo (DMC) has been demonstrated to successfully predict the energetics and properties of a wide range of molecules and solids by numerically solving the electronic many-body Schrödinger equation. We show that when coupled with quantum machine learning (QML) based surrogate methods the computational burden can be alleviated such that QMC shows clear potential to undergird the formation of high quality descriptions across chemical space. We discuss three crucial approximations necessary to accomplish this: The fixed node approximation, universal and accurate references for chemical bond dissociation energies, and scalable minimal amons set based QML (AQML) models. Numerical evidence presented includes converged DMC results for over one thousand small organic molecules with up to 5 heavy atoms used as amons, and 50 medium sized organic molecules with 9 heavy atoms to validate the AQML predictions. Numerical evidence collected for 𝛥-AQML models suggests that already modestly sized QMC training data sets of amons suffice to predict total energies with near chemical accuracy throughout chemical space. In this archive, we present DMC energies for over one thousand small organic molecules with up to 5 heavy atoms, and 50 medium sized organic molecules with 9 heavy atoms, as well as energies computed at cheaper levels of theory such as HF, DFT, MP2 and CCSD(T).

Materials Cloud sections using this data

No Explore or Discover sections associated with this archive record.

Files

File name Size Description
amons-ni7.zip
MD5md5:d4f8b7cd9f71bf26158e5baa513767b4
845.9 KiB This archive includes the geometries and energies for atom-in-molecule based fragments (aka. amons) of the 50 test molecules (see targets.zip). Geometries were calculated at the level of B3LYP, while energies (in Hartree) were calculated at various levels of theory, including HF, DFT and post-HF (MP2 & CCSD(T)) levels of theory, all with cc-pVTZ basis.
targets.zip
MD5md5:605d8e3cf6edd4127146d9d7abf532e7
23.6 KiB This archive includes the B3LYP/cc-pVTZ geometries and energies (in Hartree, computed at QMC and various cheaper levels of theory shown above) of 50 test molecules randomly drawn from the QM9 dataset.
amons-ni5-qmc.zip
MD5md5:c48713f6d64de3cbd8ca565a295ab6be
878.9 KiB This archive includes the geometries and energies for amons of the QM9 dataset, made up of at most 5 heavy atoms. Geometries were calculated at the B3LYP/cc-pVTZ level, while energies (in kcal/mol) were calculated at the QMC level, as well as the other cheaper levels of theory.

License

Files and data are licensed under the terms of the following license: Creative Commons Attribution 4.0 International.
Metadata, except for email addresses, are licensed under the Creative Commons Attribution Share-Alike 4.0 International license.

Keywords

quantum machine learning Quantum Monte Carlo chemical space MARVEL

Version history:

2023.174 (version v2) [This version] Nov 15, 2023 DOI10.24435/materialscloud:9p-8a
2022.177 (version v1) Dec 19, 2022 DOI10.24435/materialscloud:p7-p8