Machine learning (ML) has become a widely used technique in materials science study. Recently, a research team led by Prof. Feng Pan from Peking University, Shenzhen Graduate School, has trained an ML model for the relationship between atomic structures and band gaps of crystals. After analyzing structures whose band gap deviates much from the prediction made by the model, some structures with unusual structure units or other abnormities are found.
Source: EurekaAlert, https://www.eurekalert.org