An active learning approach to train machine-learning interatomic potentials (moment tensor potentials) for multicomponent alloys to ab initio data is presented. Employing this approach, the disordered body-centered cubic (bcc) ${\mathrm{TiZrHfTa}}_x$ system with varying Ta concentration is investigated via molecular dynamics simulations. Our results show a strong interplay between elastic properties and the structural $\omega$ phase stability, strongly affecting the mechanical properties. Based on these insights we systematically screen composition space for regimes where elastic constants show little or no temperature dependence (elinvar effect).

中文翻译：

---

来自ab initiotrained机器学习潜力的bcc多组分合金的结构稳定性、化学复杂性和弹性特性的有限温度相互作用

提出了一种主动学习方法，用于将多组分合金的机器学习原子间势（矩张量势）训练为ab initio数据。采用这种方法，无序体心立方 (bcc)${\mathrm{钛锆}}_X$通过分子动力学模拟研究了具有不同 Ta 浓度的系统。我们的结果表明弹性性能和结构之间有很强的相互作用$\omega$相稳定性，强烈影响机械性能。基于这些见解，我们系统地筛选弹性常数显示很少或没有温度依赖性（elinvar 效应）的区域的组成空间。