A research team led by Oak Ridge National Laboratory has developed a new method to uncover the atomic origins of anomalous material behavior. This approach uses Bayesian Deep Learning, a form of artificial intelligence that combines probability theory and neural networks, to analyze complex data sets with exceptional efficiency.
This technique reduces the time required for the experiment. It helps researchers explore sample regions that converge widely and rapidly on key features that exhibit interesting properties.
“This method allows us to study the properties of materials much more efficiently,” says Ganesh Narasimha of Ornl. “We usually need to scan large areas, then some small areas, and perform very time-consuming spectroscopy. Here, the AI algorithm needs to take control and do this process automatically and intelligently.”
This study investigated magnetic half-magnetic zinc zinc, known for its unique electronic behavior. However, this method can be generalized to a variety of materials. Using advanced scan tunnel microscopes, researchers have presented the connection between atomic structures and electronic properties. This streamlined approach simplifies the discovery process and advances the nation's capabilities related to artificial intelligence and quantum science.
The complete findings are available in NPJ calculation materials. – Scott Gibson
