Published June 30, 2023 | Version v1
Dataset Open

Robustness of local predictions in atomistic machine learning models

  • 1. Laboratory of Computational Science and Modeling (COSMO), IMX, École Polytechnique Fédérale de Lausanne (EPFL), 1015 Lausanne, Switzerland
  • 2. Department of Chemistry, Inorganic Chemistry Laboratory, University of Oxford, Oxford OX1 3QR, United Kingdom

* Contact person

Description

Machine learning (ML) models for molecules and materials commonly rely on a decomposition of the global target quantity into local, atom-centered contributions. This approach is convenient from a computational perspective, enabling large-scale ML-driven simulations with a linear-scaling cost, and can also be used to deduce useful structure--property relations as they associate simple atomic motifs with complicated macroscopic properties. However, even though there exist practical justifications for these decompositions, only the global quantity is rigorously defined, and thus it is unclear to what extent the atomistic terms predicted by the model can be trusted. Here, we introduce a quantitative metric, which we call the local prediction rigidity (LPR), to assess how robust the locally decomposed predictions of ML models are. We investigate the dependence of LPR on the details of model training, e.g. composition of the dataset, for several different problems ranging from simple toy models to real chemical systems. We present strategies to systematically enhance the LPR, which can be used to improve the robustness, interpretability, and transferability of the resulting atomistic ML models. This repository contains datasets and Jupyter notebooks employed to corroborate the results of the above study.

Files

File preview

files_description.md

All files

Files (12.9 MiB)

Name Size
md5:e0decdaa149a560808418c18ffcfb3b0
286 Bytes Preview Download
md5:68664f36b80f27317378e76fa62a6406
12.9 MiB Download
md5:a0331a12aa07fc5269de7ccb738c8a13
1.8 KiB Preview Download

References

Preprint (Preprint where the data and workflow are discussed)
S. Chong, F. Grasselli, C. Ben Mahmoud, J. D. Morrow, V. L. Deringer, M. Ceriotti, arXiv:2306.15638, doi: 10.48550/arXiv.2306.15638