CaltechAUTHORS
  A Caltech Library Service

Improved accuracy and transferability of molecular-orbital-based machine learning: Organics, transition-metal complexes, non-covalent interactions, and transition states

Husch, Tamara and Sun, Jiace and Cheng, Lixue and Lee, Sebastian J. R. and Miller, Thomas F., III (2021) Improved accuracy and transferability of molecular-orbital-based machine learning: Organics, transition-metal complexes, non-covalent interactions, and transition states. Journal of Chemical Physics, 154 (6). Art. No. 064108. ISSN 0021-9606. https://resolver.caltech.edu/CaltechAUTHORS:20201110-142606934

[img] PDF - Submitted Version
Creative Commons Attribution.

2169Kb
[img] PDF - Supplemental Material
See Usage Policy.

4Mb

Use this Persistent URL to link to this item: https://resolver.caltech.edu/CaltechAUTHORS:20201110-142606934

Abstract

Molecular-orbital-based machine learning (MOB-ML) provides a general framework for the prediction of accurate correlation energies at the cost of obtaining molecular orbitals. The application of Nesbet’s theorem makes it possible to recast a typical extrapolation task, training on correlation energies for small molecules and predicting correlation energies for large molecules, into an interpolation task based on the properties of orbital pairs. We demonstrate the importance of preserving physical constraints, including invariance conditions and size consistency, when generating the input for the machine learning model. Numerical improvements are demonstrated for different datasets covering total and relative energies for thermally accessible organic and transition-metal containing molecules, non-covalent interactions, and transition-state energies. MOB-ML requires training data from only 1% of the QM7b-T dataset (i.e., only 70 organic molecules with seven and fewer heavy atoms) to predict the total energy of the remaining 99% of this dataset with sub-kcal/mol accuracy. This MOB-ML model is significantly more accurate than other methods when transferred to a dataset comprising of 13 heavy atom molecules, exhibiting no loss of accuracy on a size intensive (i.e., per-electron) basis. It is shown that MOB-ML also works well for extrapolating to transition-state structures, predicting the barrier region for malonaldehyde intramolecular proton-transfer to within 0.35 kcal/mol when only trained on reactant/product-like structures. Finally, the use of the Gaussian process variance enables an active learning strategy for extending the MOB-ML model to new regions of chemical space with minimal effort. We demonstrate this active learning strategy by extending a QM7b-T model to describe non-covalent interactions in the protein backbone–backbone interaction dataset to an accuracy of 0.28 kcal/mol.


Item Type:Article
Related URLs:
URLURL TypeDescription
https://doi.org/10.1063/5.0032362DOIArticle
https://arxiv.org/abs/2010.03626arXivDiscussion Paper
ORCID:
AuthorORCID
Husch, Tamara0000-0002-2880-2481
Cheng, Lixue0000-0002-7329-0585
Lee, Sebastian J. R.0000-0001-7006-9378
Miller, Thomas F., III0000-0002-1882-5380
Additional Information:© 2021 Published under license by AIP Publishing. Submitted: 16 November 2020; Accepted: 17 January 2021; Published Online: 12 February 2021. This work was supported, in part, by the U.S. Army Research Laboratory (Grant No. W911NF-12-2-0023), the U.S. Department of Energy (Grant No. DE-SC0019390), the Caltech DeLogi Fund, and the Camille and Henry Dreyfus Foundation (Award No. ML-20-196). T.H. acknowledges funding through an Early Post-Doc Mobility Fellowship by the Swiss National Science Foundation (Award No. P2EZP2_184234). S.J.R.L. thanks the Molecular Software Sciences Institute (MolSSI) for a MolSSI investment fellowship. Computational resources were provided by the National Energy Research Scientific Computing Center (NERSC), a DOE Office of Science User Facility supported by the DOE Office of Science under Contract No. DE-AC02-05CH11231.
Funders:
Funding AgencyGrant Number
Army Research Office (ARO)W911NF-12-2-0023
Department of Energy (DOE)DE-SC0019390
Caltech De Logi FundUNSPECIFIED
Camille and Henry Dreyfus FoundationML-20-196
Swiss National Science Foundation (SNSF)P2EZP2_184234
Molecular Software Sciences InstituteUNSPECIFIED
Department of Energy (DOE)DE-AC02-05CH11231
Issue or Number:6
Record Number:CaltechAUTHORS:20201110-142606934
Persistent URL:https://resolver.caltech.edu/CaltechAUTHORS:20201110-142606934
Official Citation:Improved accuracy and transferability of molecular-orbital-based machine learning: Organics, transition-metal complexes, non-covalent interactions, and transition states. J. Chem. Phys. 154, 064108 (2021); doi: 10.1063/5.0032362
Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:106593
Collection:CaltechAUTHORS
Deposited By: George Porter
Deposited On:10 Nov 2020 23:08
Last Modified:12 Feb 2021 18:33

Repository Staff Only: item control page