We demonstrate the potential for machine learning systems to predict three-dimensional (3D)-relevant NMR properties beyond traditional ¹H- and ¹³C-based data, with comparable accuracy to density functional theory (DFT) (but orders of magnitude faster). Predictions of DFT-calculated ¹⁵N chemical shifts for 3D molecular structures can be achieved using a machine learning system—IMPRESSION (Intelligent Machine PREdiction of Shift and Scalar information Of Nuclei), with an accuracy of 6.12-ppm mean absolute error (∼1% of the δ¹⁵N chemical shift range) and an error of less than 20 ppm for 95% of the chemical shifts. It provides less accurate raw predictions of experimental chemical shifts, due to the limited size and chemical space diversity of the training dataset used in its creation, coupled with the limitations of the underlying DFT methodology in reproducing experiment.

中文翻译：

---

通过机器学习预测 15N 化学位移

我们展示了机器学习系统预测三维 (3D) 相关 NMR 属性的潜力，超越传统的基于¹ H 和¹³ C 的数据，其精度与密度泛函理论 (DFT) 相当（但速度快几个数量级）。可以使用机器学习系统 IMPRESSION（智能机器预测核的位移和标量信息）来预测 3D 分子结构的 DFT 计算的^{15 N 化学位移，准确度为 6.12 ppm 平均绝对误差（~1%} δ ¹⁵ _N 化学位移范围）和 95% 的化学位移误差小于 20 ppm。由于在其创建过程中使用的训练数据集的大小和化学空间多样性有限，再加上基础 DFT 方法在重现实验中的局限性，它对实验化学位移的原始预测不太准确。