Machine learning and non-thermal external fields: A dual-track paradigm for engineering high-performance of Mg-based hydrogen storage materials

Confronted with the inherent thermodynamic and kinetic challenges of Mg-based hydrogen storage materials (HSMs), research strategies have evolved from traditional alloying, nano-structuring, and catalytic modification toward a new dual-track paradigm that integrates machine learning (ML)-based rational design with non-thermal external field (NTEF)-assisted precision regulation. ML significantly accelerates the design and screening of novel HSMs through efficient performance prediction, key descriptor extraction, and microscopic mechanism elucidation, while NTEF provides a powerful experimental means for non-equilibrium synthesis, microstructure optimization, and de-/hydrogenation behavior control by leveraging its unique high energy efficiency and tunability. These two approaches complement each other and jointly advance Mg-based HSMs toward higher performance, lower cost, and superior cycling stability, thereby establishing a solid scientific and technological foundation for safe, efficient hydrogen storage and transportation as well as large-scale hydrogen-thermal coupled application.