Synthesizability Prediction of Crystalline Structures with Structure-Aware Feature Learning and Uncertainty Quantification
SyntheFormer is a positive-unlabeled framework that learns synthesizability directly from crystal structure, combining Fourier-transformed crystal properties (FTCP) representation with structure-aware feature extraction, Random-Forest feature selection, and a compact deep MLP classifier, demonstrating that stability al...
Danial Ebrahimzadeh, S. Sharif, Y. Banad
· 0 citations