A research team led by Associate Professor Hongliang Xin has not only identified new materials that can generate clean energy in fuel cells and wastewater treatment, but also can now explain how and why it works.

The team, including collaborators Gang Wu, at the University of Buffalo in New York and Yi Li, of Jiangsu University, published the work the journal Nature Communications. They used theory and artificial intelligence approaches to sort through more than 24,000 candidate materials and predict their effectiveness. The performance of some of those candidate materials were then verified in lab experiments.

"This work let us see inside the black box of deep neural networks to determine how and why these materials worked," Xin said. "This goes beyond what has so far been described in the literature and shows the potential of explainable AI in uncovering new physics."