Managing Expectations and Imbalanced Training Data in Reactive Force Field Development: An Application to Water Adsorption on Alumina
- 19 April 2024
- journal article
- research article
- Published by American Chemical Society (ACS) in Journal of Chemical Theory and Computation
- Vol. 20 (9) , 3779-3797
Abstract
ReaxFF is a computationally efficient model for reactive molecular dynamics simulations that has been applied to a wide variety of chemical systems. When ReaxFF parameters are not yet available for a chemistry of interest, they must be (re)optimized, for which one defines a set of training data that the new ReaxFF parameters should reproduce. ReaxFF training sets typically contain diverse properties with different units, some of which are more abundant (by orders of magnitude) than others. To find the best parameters, one conventionally minimizes a weighted sum of squared errors over all of the data in the training set. One of the challenges in such numerical optimizations is to assign weights so that the optimized parameters represent a good compromise among all the requirements defined in the training set. This work introduces a new loss function, called Balanced Loss, and a workflow that replaces weight assignment with a more manageable procedure. The training data are divided into categories with corresponding "tolerances", i.e., acceptable root-mean-square errors for the categories, which define the expectations for the optimized ReaxFF parameters. Through the Log-Sum-Exp form of Balanced Loss, the parameter optimization is also a validation of one's expectations, providing meaningful feedback that can be used to reconfigure the tolerances if needed. The new methodology is demonstrated with a nontrivial parametrization of ReaxFF for water adsorption on alumina. This results in a new force field that reproduces both the rare and frequent properties of a validation set not used for training. We also demonstrate the robustness of the new force field with a molecular dynamics simulation of water desorption from a γ-Al2O3 slab model.Keywords
All Related Versions
Funding Information
- IFP Energies Nouvelles
- Bijzonder Onderzoeksfonds UGent (BOF/24J/2023/121)
This publication has 130 references indexed in Scilit:
- A Reactive Molecular Dynamics Study of Chlorinated Organic Compounds. Part I: Force Field DevelopmentChemphyschem, 2023
- Neural network and ReaxFF comparison for Au propertiesInternational Journal of Quantum Chemistry, 2016
- Efficient global optimization of reactive force‐field parametersJournal of Computational Chemistry, 2015
- MOF‐FF – A flexible first‐principles derived force field for metal‐organic frameworksPhysica Status Solidi (b), 2013
- Parameterization of a reactive force field using a Monte Carlo algorithmJournal of Computational Chemistry, 2013
- Ab‐initiosimulations of materials using VASP: Density‐functional theory and beyondJournal of Computational Chemistry, 2008
- γ‐Alumina as a Support for Catalysts: A Review of Fundamental AspectsEuropean Journal of Inorganic Chemistry, 2005
- Alumina2002
- Computer Simulation of Molecular Dynamics: Methodology, Applications, and Perspectives in ChemistryAngewandte Chemie International Edition, 1990
- Dehydration of Alcohols on Aluminum OxideAngewandte Chemie International Edition, 1968