This work presents a machine-learning approach that optimizes the learning of atomic interactions across diverse materials. By integrating heterogeneous data, it delivers accurate and transferable predictions beyond its training domain.
- Jaesun Kim
- Jinmu You
- Seungwu Han
