Materials discovery often triggers new technological innovations and is therefore an exciting topic in materials research. In order to search a large chemistry space, however, lengthy trial-and-error testing is required. Recently, a Canadian research team devised a novel representation scheme for perovskite alloys. Combined with machine-learning methods, it enables efficient exploration of a large chemistry space with reasonable accuracy.

中文翻译：

---

大型化学领域的数据驱动材料发现

材料发现通常会引发新的技术创新，因此是材料研究中令人兴奋的主题。但是，为了搜索较大的化学空间，需要进行长时间的反复试验。最近，加拿大的一个研究小组设计了一种新颖的钙钛矿合金表征方案。结合机器学习方法，它可以合理的精度有效地探索大型化学空间。