Going beyond density functional theory accuracy: Leveraging experimental data to refine pre-trained machine learning interatomic potentials
Authors:
Shriya Gumber,
Lorena Alzate-Vargas,
Benjamin T. Nebgen,
Arjen van Veelen,
Smit Kadvani,
Tammie Gibson,
Richard Messerly
Abstract:
Machine learning interatomic potentials (MLIPs) are inherently limited by the accuracy of the training data, usually consisting of energies and forces obtained from quantum mechanical calculations, such as density functional theory (DFT). Since DFT itself is based on several approximations, MLIPs may inherit systematic errors that lead to discrepancies with experimental data. In this paper, we use…
▽ More
Machine learning interatomic potentials (MLIPs) are inherently limited by the accuracy of the training data, usually consisting of energies and forces obtained from quantum mechanical calculations, such as density functional theory (DFT). Since DFT itself is based on several approximations, MLIPs may inherit systematic errors that lead to discrepancies with experimental data. In this paper, we use a trajectory re-weighting technique to refine DFT pre-trained MLIPs to match the target experimental Extended X-ray Absorption Fine Structure (EXAFS) spectra. EXAFS spectra are sensitive to the local structural environment around an absorbing atom. Thus, refining an MLIP to improve agreement with experimental EXAFS spectra also improves the MLIP prediction of other structural properties that are not directly involved in the refinement process. We combine this re-weighting technique with transfer learning and a minimal number of training epochs to avoid overfitting to the limited experimental data. The refinement approach demonstrates significant improvement for two MLIPs reported in previous work, one for an established nuclear fuel: uranium dioxide (UO$_2$) and second one for a nuclear fuel candidate: uranium mononitride (UN). We validate the effectiveness of our approach by comparing the results obtained from the original (unrefined) DFT-based MLIP and the EXAFS-refined MLIP across various properties, such as lattice parameters, bulk modulus, heat capacity, point defect energies, elastic constants, phonon dispersion spectra, and diffusion coefficients. An accurate MLIP for nuclear fuels is extremely beneficial as it enables reliable atomistic simulation, which greatly reduces the need for large number of expensive and inherently dangerous experimental nuclear integral tests, traditionally required for the qualification of efficient and resilient fuel candidates.
△ Less
Submitted 11 June, 2025;
originally announced June 2025.
Toward machine learning interatomic potentials for modeling uranium mononitride
Authors:
Lorena Alzate-Vargas,
Kashi N. Subedi,
Nicholas Lubbers,
Michael W. D Cooper,
Roxanne M. Tutchton,
Tammie Gibson,
Richard A. Messerly
Abstract:
Uranium mononitride (UN) is a promising accident-tolerant fuel because of its high fissile density and high thermal conductivity. In this study, we developed the first machine learning interatomic potentials for reliable atomic-scale modeling of UN at finite temperatures. We constructed a training set using density functional theory (DFT) calculations that was enriched through an active learning p…
▽ More
Uranium mononitride (UN) is a promising accident-tolerant fuel because of its high fissile density and high thermal conductivity. In this study, we developed the first machine learning interatomic potentials for reliable atomic-scale modeling of UN at finite temperatures. We constructed a training set using density functional theory (DFT) calculations that was enriched through an active learning procedure, and two neural network potentials were generated. Both potentials successfully reproduce key thermophysical properties of interest, such as temperature-dependent lattice parameter, specific heat capacity, and bulk modulus. We also evaluated the energy of stoichiometric defect reactions and defect migration barriers and found close agreement with DFT predictions, demonstrating that our potentials can be used for modeling defects in UN. Additional tests provide evidence that our potentials are reliable for simulating diffusion, noble gas impurities, and radiation damage.
△ Less
Submitted 5 June, 2025; v1 submitted 21 November, 2024;
originally announced November 2024.
