Toward a world model for corrosion science
Abstract Corrosion is a trajectory, not an event, yet current machine learning maps fixed inputs to outputs and cannot model reliably how a corroding interface responds to intervention. We argue that world models, deep learning frameworks that learn latent physical dynamics from observation sequences, can close this gap. A physically interpretable latent space anchored to corrosion mechanisms can enable prediction, counterfactual reasoning, and discovery. We chart a path toward realising it.
Authors
- Viacheslav Shkirskiy (ORCID: https://orcid.org/0000-0003-4289-9678)
- N. Birbilis (ORCID: https://orcid.org/0000-0002-7910-7470)
- Leonardo B. Coelho
Institutions
- Centre National de la Recherche Scientifique (FR)
- University of Mons (BE)
- Université Libre de Bruxelles (BE)
- Vrije Universiteit Brussel (BE)
- Deakin University (AU)
- Université Paris Cité (FR)
- Interfaces Traitements Organisation et Dynamique des Systèmes (FR)
- Instituto de Cardiologia de Santa Catarina (BR)
Publication Details
- Journal
- npj Materials Degradation
- Published
- 2026-06-13
- DOI
- https://doi.org/10.1038/s41529-026-00826-4
- Primary Topic
- Machine Learning in Materials Science
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- Agence Nationale de la Recherche