Accelerating Moment Tensor Potentials through Post-Training Pruning
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2025
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866911226983874560 |
|---|---|
| author | Meng, Zijian Zongo, Karim Thoms, Matthew Grant, Ryan Eric Béland, Laurent Karim |
| author_facet | Meng, Zijian Zongo, Karim Thoms, Matthew Grant, Ryan Eric Béland, Laurent Karim |
| contents | Moment Tensor Potentials (MTPs) are machine-learning interatomic potentials whose basis functions are typically selected using a level-based scheme that is data-agnostic. We introduce a post-training, cost-aware pruning strategy that removes expensive basis functions with minimal loss of accuracy. Applied to nickel and silicon-oxygen systems, it yields models up to seven times faster than standard MTPs. The method requires no new data and remains fully compatible with current MTP implementations. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_19737 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Accelerating Moment Tensor Potentials through Post-Training Pruning Meng, Zijian Zongo, Karim Thoms, Matthew Grant, Ryan Eric Béland, Laurent Karim Materials Science Moment Tensor Potentials (MTPs) are machine-learning interatomic potentials whose basis functions are typically selected using a level-based scheme that is data-agnostic. We introduce a post-training, cost-aware pruning strategy that removes expensive basis functions with minimal loss of accuracy. Applied to nickel and silicon-oxygen systems, it yields models up to seven times faster than standard MTPs. The method requires no new data and remains fully compatible with current MTP implementations. |
| title | Accelerating Moment Tensor Potentials through Post-Training Pruning |
| topic | Materials Science |
| url | https://arxiv.org/abs/2510.19737 |