From kneecaps to spaceships: Explainable AI can help create superalloys faster than ever
The future of material design is here. A unique AI system can help scientists create perfect alloys faster than ever.
Can you guess the recipe for a cake just by looking at a bakery shelf filled with hundreds of finished desserts? That’s the kind of challenge materials scientists have faced for decades, trying to guess the elements that combine well and result in useful materials.
With over 100 elements on the periodic table, there are an endless number of ways in which these elements could potentially combine. Scientists can’t test each such combination, and therefore, it is very challenging to create materials and alloys with desirable properties.
A new study reveals an AI-based technique that can solve this problem. The AI can rapidly check countless possibilities, predict the most promising element combinations, and even suggest the ideal proportions of each element for creating alloys with superior properties.
Using this approach, the study authors produced multiple principal element alloys (MPEA) with exceptional mechanical properties. “This work demonstrates how data-driven frameworks and explainable AI can unlock new possibilities in materials design,” Sanket Deshmukh, one of the study’s authors and a professor at Virginia Tech, said.