DOE OSTI · 2998246
Teacher-student training improves the accuracy and efficiency of machine learning interatomic potentials
Abstract
Machine learning interatomic potentials (MLIPs) are revolutionizing the field of molecular dynamics (MD) simulations. Recent MLIPs have tended towards more complex architectures trained on larger datasets. The resulting increase in computational and memory costs may prohibit the application of these MLIPs to perform large-scale MD simulations. Herein, we present a teacher-student training framework in which the latent knowledge from the teacher (atomic energies) is used to augment the students' training. We show that the light-weight student MLIPs have faster MD speeds at a fraction of the memory footprint compared to the teacher models. Remarkably, the student models can even surpass the accuracy of the teachers, even though both are trained on the same quantum chemistry dataset. Our work highlights a practical method for MLIPs to reduce the resources required for large-scale MD simulations.
Explore related subjects
Keep this discovery
Explore connections, maps & timelines
Matin, Sakib [Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)] (ORCID:0000000151662838), Allen, Alice Elisabeth Anastasia [Los Alamos National Laboratory (LANL), Los Alamos, NM (United States); Max Planck Institute for Polymer Research, Mainz (Germany)] (ORCID:0000000287278333), Shinkle, Emily Suzanne [Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)] (ORCID:0000000322550540), Pachalieva, Aleksandra Atanasova [Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)] (ORCID:0000000312460410), Craven, Galen Thomas [Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)] (ORCID:0000000151172345), Nebgen, Benjamin Tyler [Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)] (ORCID:0000000153103263), Smith, Justin S. [NVIDIA Corporation, Santa Clara, CA (United States)], Messerly, Richard Alma [Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)] (ORCID:0000000313214246), Li, Ying Wai [Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)] (ORCID:0000000301248262), Tretiak, Sergei [Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)] (ORCID:0000000155473647), Barros, Kipton Marcos [Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)] (ORCID:0000000213335972), Lubbers, Nicholas Edward [Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)] (ORCID:0000000290019973). 2025-08-07. Teacher-student training improves the accuracy and efficiency of machine learning interatomic potentials. https://doi.org/10.1039/d5dd00085h
Cite the original work for its findings. Save a collection to share your selection of sources.