Classical-to-DFT multi-fidelity approach for developing machine learning potentials of aqueous systems
Authors: Siddique, Md Abu Bakar; Coronado, Kendra; Badhon, Mohammad M; Galib, Mirza
Source journal: The Journal of chemical physics (2026-11-21)
DOI: 10.1063/5.0341814
Abstract
Machine learning interatomic potentials (MLIPs) offer near-ab initio accuracy at a fraction of the cost of conventional electronic structure calculations, but their reliability depends critically on the quality and diversity of training data. Here, we present a classical-to-density functional theory (DFT) multi-fidelity workflow for efficient dataset generation, in which long classical molecular dynamics trajectories are supplemented with selectively DFT-optimized configurations to train neural network potentials. We demonstrate this approach on four systems: bulk water, water at the air-water interface, CaSO4 in bulk water, and CaSO4 at the interface. In all cases, the trained models faithfully reproduce reference DFT properties, validating the workflow across both simple and chemically complex environments. We further probe the extrapolation capabilities of the models, showing that bulk-trained networks can generalize to different temperatures, pressures, and interfacial geometries when datasets include targeted augmented configurations. This multi-fidelity strategy thus enables accurate and transferable MLIPs for complex solvated and interfacial systems, bridging the gap between computational efficiency and first-principles accuracy.
Related topics
Related open access articles
- DENSITY FUNCTIONAL THEORY ANALYSIS AND MOLECULAR DOCKING EVALUATION OF 8-CHLOROQUINOLINE 2-CARBALDEHYDE
- Thermochemical, DFT, Molecular Docking and Pharmacokinetic Studies of Methyl β-D-galactopyranoside Esters
- Mechanism and identify photolysis products of fluopyram under TiO 2 : Experiments, DFT and ab initio Molecular dynamics study
- Computational Study of the Effect of Physical Factors on the Halochromic Behavior of Azo Indicators Using Density Functional Theory
- Computed Local Chemical Reactivity of Melanoidins Red M1 and Red M2 Using Conceptual DFT
- Combining of DFT and QSAR Results to Predict the Antibacterial Activity of a Series of Azetidinones derived from Dapsone as Inhibitors of Bacillus Subtilis and Pseudomonas Aeruginosa
Abstract and bibliographic data from PubMed (U.S. National Library of Medicine). Copyright of the abstract belongs to the original publisher or authors.