Coarse-grained graph architectures for all-atom force predictions

We introduce a machine-learning framework termed coarse-grained all-atom force field (CGAA-FF), which incorporates coarse-grained message passing within an all-atom force field using equivariant nature of graph models. The CGAA-FF model employs grain embedding to encode atomistic coordinates into nodes representing grains rather than individual atoms, enabling predictions of both grain-level energies and atom-level forces. Tested on organic electrolytes, CGAA-FF achieves root-mean-square errors of 4.96 meV atom-1 for energy and 0.201 eV A-1 for force predictions. CGAA-FF significantly reduces computational costs, achieving about 22- and 14-fold improvements in simulation speed and memory efficiency, respectively, compared to the all-atom potential (SevenNet-0). Since this CGAA framework can be integrated into any equivariant architecture, we believe this work opens the door to efficient all-atom simulations of soft-matter systems.