Foundation-Model Supervision of Distilled Potentials for Million-Atom Molecular Dynamics

Abstract

Machine-learned interatomic potentials distilled from foundation models enable fast large-scale molecular dynamics (MD), but they can fail locally in chemical environments that their training data did not sample. System-averaged error metrics and small random audits often miss such sparse or isolated failures. Here we use an exact equivariant foundation model, MACE-MPA-0, to check every atom of million-atom production simulations. Reorganizing two-layer MACE inference into a receiver-chunked three-pass schedule that keeps no per-edge intermediates resident reduces the marginal memory per atom 340-fold, evaluates up to $2.86\times10^{7}$ atoms on a single H200 GPU, and supervises $10^{6}$-atom trajectories in LAMMPS at 1.3% of the run time. Distilled MACE students run four to nine times slower than NEP89 with 0.95-1.28 times the teacher's force error against density functional theory (DFT). In molten SiO$_2$, whole-system checks uncover a persistent tail of under-coordinated atoms hidden by the system-averaged error and an isolated, unphysical O-O contact that a 1% random audit misses 94% of the time. On-the-fly retraining on teacher-labelled 13-{\AA} clusters, with the force loss restricted to their inner 6 {\AA}, removes three quarters of the failure tail and the kilobar-level pressure biases in SiO$_2$ and compressed water; the repair converges to the teacher, not to DFT.