Equivariant Generative Diffusion Learns and Generalizes the Structural Ensemble of Amorphous Oxides
My rating
This paper presents an SE(3)-equivariant denoising diffusion model for learning the configurational distribution of amorphous oxides. The model serves as a structure database, enabling generation of structural ensembles and generalization beyond the small subsets sampled by conventional DFT and machine-learned-potential simulations; the authors also report data-efficient learning.
Tags: