Albert Musaelian, Simon Batzner, Anders Johansson, Lixin Sun, Cameron J. Owen...
TL;DR本文介绍了一种基于equivariant deep learning的新算法，Allegro，该算法使用了学习到的等变性表示的张量积序列，实现了对于分子和材料的能量和原子力的高精度预测，并能进行大规模的并行计算和分子动力学仿真。
Abstract
A simultaneously accurate and computationally efficient parametrization of the energy and atomic forces of molecules and materials is a long-standing goal in the natural sciences. In pursuit of this goal, neural message passing has lead to a paradigm shift by describing many-body correlations