SPaDe-CSP first predicts most probable space groups and crystal densities using machine learning and then employs an efficient neural network potential for structure refinement. Prediction of crystal ...

“Crystal Math” uses equations—and minimal resources—to rapidly predict the 3D structures of molecular crystals, which could speed up R&D for drugs and electronic devices Researchers at New York ...

The new method can determine crystal structures underlying experimental data thus far difficult to analyze. A joint research team led by Yuuki Kubo and Shiji Tsuneyuki of the University of Tokyo has ...