Overcoming sampling limitations using machine-learned interatomic potentials: the case of water-in-salt electrolytes
Luca Brugnoli, Mathieu Salanne, A. Marco Saitta, Alessandra Serva, Arthur France-Lanord · 2026
Read Overcoming sampling limitations using machine-learned interatomic potentials: the case of water-in-salt electrolytes by Luca Brugnoli, Mathieu Salanne, A. Marco Saitta, Alessandra Serva, Arthur France-Lanord (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-22099.txt.
Description
Brugnoli and co-authors test machine-learned atomic potentials on a water-in-salt electrolyte. Such liquids resist ab initio modeling over long times. The team evaluates several MACE potentials on a 21 molal lithium salt solution. They try foundation models as shipped, fine-tuned versions, and models trained from scratch. Their simulations match measured structure factors closely. Fine-tuning beats training from scratch, above all on hard-to-sample setups like short lithium-lithium distances. They also warn that some dispersion corrections can hurt, depending on the reference functional.
Catalog record
| CREATOR | Luca Brugnoli, Mathieu Salanne, A. Marco Saitta, Alessandra Serva, Arthur France-Lanord |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 51557 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.22099