ADAPT: Lightweight, Long-Range Machine Learning Force Fields Without Graphs
This work introduces the ADAPT, an MLFF that replaces graph representations with a direct coordinates-in-space formulation and explicitly considers all pairwise atomic interactions, achieving a roughly 22% reduction in force and a roughly 40 percent reduction in energy prediction error relative to a state-of-the-art GN...
Evan Dramko, Yihuang Xiong, Yizhi Zhu et al.
· Physical Review Research · 4 citations