Gaussian Processes for Molecular Simulation

Gaussian processes can model complex experimental data and interatomic interactions while enabling rigorous uncertainty quantification and propagation within a Bayesian framework. We apply state-of-the-art GP methods, including spectral and non-stationary kernel designs [1], to develop next-generation molecular simulation tools such as uncertainty-aware machine learning potentials .

[1] Sullivan, H. W., Shanks, B. L., Cervenka, M. & Hoepfner, M. P. Physics-Informed Gaussian Process Inference of Liquid Structure from Scattering Data. https://doi.org/10.48550/arXiv.2507.07948 (2025).