Abstract
Neural networks that are equivariant to rotations, translations, reflections,and permutations on n-dimensional geometric space have shown promise inphysical modeling for tasks such as accurately but inexpensively modelingcomplex potential energy surfaces to guiding the sampling of complex dynamicalsystems or forecasting their time evolution. Current state-of-the-art methodsemploy spherical harmonics to encode higher-order interactions among particles,which are computationally expensive. In this paper, we propose a simplealternative functional form that uses neurally parametrized linear combinationsof edge vectors to achieve equivariance while still universally approximatingnode environments. Incorporating this insight, we design spatial attentionkinetic networks with E(n)-equivariance, or SAKE, which are competitive inmany-body system modeling tasks while being significantly faster.