Limitations of AI-based material prediction: Crystallographic disorder represents a stumbling block

Summary

A study led by the University of Bayreuth highlights that computer simulations and artificial intelligence frequently make significant errors in predicting the properties of new high-performance materials due to crystallographic disorder. The researchers published their findings in Advanced Materials, offering tools to mitigate these prediction challenges.