Accelerating Molecular Dynamics Simulations with Foundation Neural Network Models using Multiple Time-Step and Distillation
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866910013752082432 |
|---|---|
| author | Cattin, Côme Plé, Thomas Adjoua, Olivier Gouraud, Nicolaï Lagardère, Louis Piquemal, Jean-Philip |
| author_facet | Cattin, Côme Plé, Thomas Adjoua, Olivier Gouraud, Nicolaï Lagardère, Louis Piquemal, Jean-Philip |
| contents | We present a distilled multi-time-step (DMTS) strategy to accelerate molecular dynamics simulations using foundation neural network models. DMTS uses a dual-level neural network where the target accurate potential is coupled to a simpler but faster model obtained via a distillation process. The 3.5 Å-cutoff distilled model is sufficient to capture the fast-varying forces, i.e., mainly bonded interactions, from the accurate potential allowing its use in a reversible reference system propagator algorithms (RESPA)-like formalism. The approach conserves accuracy, preserving both static and dynamical properties, while enabling to evaluate the costly model only every 3 to 6 fs depending on the system. Consequently, large simulation speedups over standard 1 fs integration are observed: nearly 4-fold in homogeneous systems and 3-fold in large solvated proteins through leveraging active learning for enhanced stability. Such a strategy is applicable to any neural network potential and reduces their performance gap with classical force fields. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_06562 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Accelerating Molecular Dynamics Simulations with Foundation Neural Network Models using Multiple Time-Step and Distillation Cattin, Côme Plé, Thomas Adjoua, Olivier Gouraud, Nicolaï Lagardère, Louis Piquemal, Jean-Philip Chemical Physics We present a distilled multi-time-step (DMTS) strategy to accelerate molecular dynamics simulations using foundation neural network models. DMTS uses a dual-level neural network where the target accurate potential is coupled to a simpler but faster model obtained via a distillation process. The 3.5 Å-cutoff distilled model is sufficient to capture the fast-varying forces, i.e., mainly bonded interactions, from the accurate potential allowing its use in a reversible reference system propagator algorithms (RESPA)-like formalism. The approach conserves accuracy, preserving both static and dynamical properties, while enabling to evaluate the costly model only every 3 to 6 fs depending on the system. Consequently, large simulation speedups over standard 1 fs integration are observed: nearly 4-fold in homogeneous systems and 3-fold in large solvated proteins through leveraging active learning for enhanced stability. Such a strategy is applicable to any neural network potential and reduces their performance gap with classical force fields. |
| title | Accelerating Molecular Dynamics Simulations with Foundation Neural Network Models using Multiple Time-Step and Distillation |
| topic | Chemical Physics |
| url | https://arxiv.org/abs/2510.06562 |