发表机构
École Polytechnique Fédérale de Lausanne; Fritz Haber Institute of the Max Planck Society（洛桑联邦理工学院; 马克斯·普朗克学会弗里茨·哈伯研究所）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本文针对原子级机器学习中体序势的局限性，提出一种融合局域性、平滑性与对称性先验的架构，无需假设体序级数低阶截断，可实现高效计算与良好性能。
AI 中文摘要
机器学习原子间势已成为原子模拟不可或缺的工具，最流行的模型严重依赖 locality（局域性）、smoothness（平滑性）和 symmetry（对称性）等物理先验，这些先验被认为可提升训练模型的精度与迁移性。但近期证据表明，在数据充足的场景下，从数据中学习对称性的无约束模型可实现极具竞争力的精度与计算效率，且不牺牲稳定性与泛化能力。许多等变对称架构历史上依赖原子间势可由收敛团簇展开（即基于原子对、三体、四体及更高阶原子组的分解）良好近似的假设。本文描述了一种融合局域性、平滑性与对称性先验的架构，其得益于计算高效的张量积公式，无需假设体序级数的低阶截断。
英文摘要
Machine-learning interatomic potentials have become indispensable tools in atomistic simulations. The most popular models rely heavily on physical priors, such as locality, smoothness and symmetry, that are assumed to improve the accuracy and transferability of the trained models. However, recent evidence suggests that, in the data-rich regime, unconstrained models that learn symmetry from the data can achieve very competitive accuracy and computational efficiency, without sacrificing stability and generalization power. Many equivariant symmetric architectures historically rely on the assumption that the interatomic potential can be approximated well by a convergentcluster expansion, i.e., a decomposition in terms associated to atomic pairs, triplets, quadruplets, and higher-order tuples. We describe an architecture that incorporates locality, smoothness, and symmetry priors, but that -- thanks to a computationally efficient tensor product formulation -- does not assume low-order truncation of the body-ordered series.
