BIGDML: Towards Exact Machine Learning Force Fields for Materials
Not on the map. Why?
Abstract (arXiv)
Machine-learning force fields (MLFF) should be accurate, computationally and data efficient, and applicable to molecules, materials, and interfaces thereof. Currently, MLFFs often introduce tradeoffs that restrict their practical applicability to small subsets of chemical space or require exhaustive datasets for training. Here, we introduce the Bravais-Inspired Gradient-Domain Machine Learning (BIGDML) approach and demonstrate its ability to construct reliable force fields using a training set with just 10-200 geometries for materials including pristine and defect-containing 2D and 3D semiconductors and metals, as well as chemisorbed and physisorbed atomic and molecular adsorbates on surfaces. The BIGDML model employs the full relevant symmetry group for a given material, does not assume artificial atom types or localization of atomic interactions and exhibits high data efficiency and state-of-the-art energy accuracies (errors substantially below 1 meV per atom) for an extended set of materials. Extensive path-integral molecular dynamics carried out with BIGDML models demonstrate the counterintuitive localization of benzene--graphene dynamics induced by nuclear quantum effects and allow to rationalize the Arrhenius behavior of hydrogen diffusion coefficient in a Pd crystal for a wide range of temperatures.
Key facts. Cited by 5 papers in the map (data through 7 Oct 2026). Builds on 9 of its 144 references that are in the map. First posted on arXiv 8 Jun 2021.
Builds on
9 of 144 references.
References · 9
- 2018Tensor field networks: Rotation- and translation-equivariant neural networks for 3D point cloudscited by 735
- 2021Equivariant message passing for the prediction of tensorial properties and molecular spectracited by 438
- 2020Machine Learning Force Fieldscited by 204
- 2019Machine learning for molecular simulationcited by 109
- 2021SpookyNet: Learning Force Fields with Electronic Degrees of Freedom and Nonlocal Effectscited by 103
- 2020Deep Learning in Protein Structural Modeling and Designcited by 25
- 2015On Invariance and Selectivity in Representation Learningcited by 18
- 2019Machine learning for protein folding and dynamicscited by 15
- 2020Autonomous robotic nanofabrication with reinforcement learningcited by 8
Built on by
Citing papers per year: 2022: 3; 2023: 1; 2024: 1.
5 papers.
- 2024Euclidean Fast Attention -- Machine Learning Global Atomic Representations at Linear Costcited by 12
- 2022A Transferable Recommender Approach for Selecting the Best Density Functional Approximations in Chemical Discoverycited by 3
- 2022Reconstructing Kernel-based Machine Learning Force Fields with Super-linear Convergencecited by 2
- 2022Active Learning Exploration of Transition Metal Complexes to Discover Method-Insensitive and Synthetically Accessible Chromophorescited by 1
- 2023From Peptides to Nanostructures: A Euclidean Transformer for Fast and Stable Machine Learned Force Fieldscited by 1
- 2024Euclidean Fast Attention -- Machine Learning Global Atomic Representations at Linear Cost11 Dec 2024
- 2023From Peptides to Nanostructures: A Euclidean Transformer for Fast and Stable Machine Learned Force Fields21 Sep 2023
- 2022Reconstructing Kernel-based Machine Learning Force Fields with Super-linear Convergence24 Dec 2022
- 2022Active Learning Exploration of Transition Metal Complexes to Discover Method-Insensitive and Synthetically Accessible Chromophores10 Aug 2022
- 2022A Transferable Recommender Approach for Selecting the Best Density Functional Approximations in Chemical Discovery21 Jul 2022