在最新发表于Nature Computational Science的研究中，清华大学王笑楠团队提出了催化剂设计领域的全新基础模型SurFF。 在无需依赖昂贵实验或耗时DFT计算的前提下，这一模型可实现对金属间化合物表面暴露与形貌的高效、高精度预测，计算速度相比传统方法提升10⁵倍 ...

(a) The distribution of magnetic moments for the transition metal halides. A histogram of DFT magnetic moment values (right axis, blue) is displayed alongside a histogram of ML predictions (left axis, ...

Using state-of-the-art density functional theory (DFT) simulations, they showed that when a single layer of the semiconductor ...