Jul 2026
DynStabNet: A Deep Learning Framework for Fast Dynamical Stability Prediction of Crystal Structures.
DynStabNet is designed as a fast surrogate model to accelerate the prescreening of candidate structures prior to subsequent MLIP- or DFT-based phonon validation, and achieves an accuracy of 97% while reducing the evaluation time per structure from several hours to approximately 1 ms.
Haichao Li, Zian Chen, Tao He et al.
· Journal of Physical Chemistr... · 0 citations