The article discusses the development of a large materials model for AI-driven materials discovery, focusing on the work of researchers from Tsinghua University. They developed the DeepH method, a deep-learning model based on Density Functional Theory (DFT), to predict material properties from their structures. This universal model can handle a wide variety of materials across the periodic table, achieving high accuracy in predicting properties. The research represents a significant step toward AI-enhanced materials science.
For more details, you can view the full article here: