A Gentle Push: Fine-Tuning Foundation Models for Complex Oxide Surface Reconstructions
- 1Institute of Materials Chemistry
- 2TU Wien
Abstract
Structure-prediction methods routinely sample highly distorted, far-from-equilibrium configurations. Conventional workflows based on machine learning interatomic potentials (MLIPs) therefore require the inclusion of unphysical local atomic environments in the training data to ensure stability during the search. We show that foundation MLIPs, fine-tuned on minimal datasets containing only physically reasonable structures, can serve as robust backbones for evolutionary structure prediction of complex surface reconstructions. We systematically benchmark fine-tuning strategies and identify that freezing all layers except the readouts as the most effective strategy in the low-data regime. Fine-tuning on as few as 23 structures is sufficient to perform evolutionary searches on reconstructed SrTiO 3 (110) surfaces. For BaTiO 3 (001), we further demonstrate that a lightweight active-learning approach, based on fine-tuning on as few as 20 structures, enables the identification of new low-energy (√5 × √5)R26.6° reconstructions.
Versions
Related
- First-principles Investigation of Reconstructions for the Polar CoO (111) Surface and the Impact on Nanoparticle Morphology
- Fine-tuning MLIP foundation models: strategies for accuracy and transferability
- Electronic tuning of tetranuclear Cu(i)–pyrazolate macrocycles: anti-Kasha emission through high-lying triplet states of a C2F5-substituted complex
- Fine-tuning universal machine learning potentials for transition state search in surface catalysis
- Enhanced Methane-to-Products Conversion in Solid Oxide Cells by Tuning Catalyst Surface Properties
BibTeX
@misc{wanzenbock2026gentle,
title = {{A Gentle Push: Fine-Tuning Foundation Models for Complex Oxide Surface Reconstructions}},
author = {Ralf Wanzenböck and Eva Doloszeski and Georg K. H. Madsen},
howpublished = {ChemRxiv},
year = {2026},
month = sep,
doi = {10.26434/chemrxiv.15001629/v3},
url = {https://chemrxiv.org/doi/full/10.26434/chemrxiv.15001629/v3},
note = {CC BY},
}
RIS
TY - UNPB
TI - A Gentle Push: Fine-Tuning Foundation Models for Complex Oxide Surface Reconstructions
AU - Ralf Wanzenböck
AU - Eva Doloszeski
AU - Georg K. H. Madsen
PY - 2026
DA - 2026/09/11
PB - ChemRxiv
DO - 10.26434/chemrxiv.15001629/v3
UR - https://chemrxiv.org/doi/full/10.26434/chemrxiv.15001629/v3
AB - Structure-prediction methods routinely sample highly distorted, far-from-equilibrium configurations. Conventional workflows based on machine learning interatomic potentials (MLIPs) therefore require the inclusion of unphysical local atomic environments in the training data to ensure stability during the search. We show that foundation MLIPs, fine-tuned on minimal datasets containing only physically reasonable structures, can serve as robust backbones for evolutionary structure prediction of complex surface reconstructions. We systematically benchmark fine-tuning strategies and identify that freezing all layers except the readouts as the most effective strategy in the low-data regime. Fine-tuning on as few as 23 structures is sufficient to perform evolutionary searches on reconstructed SrTiO 3 (110) surfaces. For BaTiO 3 (001), we further demonstrate that a lightweight active-learning approach, based on fine-tuning on as few as 20 structures, enables the identification of new low-energy (√5 × √5)R26.6° reconstructions.
ER -