Why is this innovation so noteworthy to NNSA’s mission, as well as other applications? A chef inventing a new dish might test hundreds, even thousands, of ingredient combinations to get a recipe just right. Materials science—vital to keeping the U.S. nuclear stockpile safe, secure and effective—traditionally has worked the same way. Through trial and error, scientists make educated guesses about how a “recipe” for a new material will turn out, based on the known properties its ingredients. LANL researchers have described how they can supercharge the trial-and-error process by teaching a computer to learn and adapt as it virtually combines ingredients. Using adaptive design strategy, based in information science and fed by data from past experiments, the researchers can accelerate the discovery of new materials with desired properties.
“What we’ve done is show that, starting with a relatively small data set of well-controlled experiments, it is possible to iteratively guide subsequent experiments toward finding the material with the desired target,” said Turab Lookman, a physicist and materials scientist in the Physics of Condensed Matter and Complex Systems group at Los Alamos National Laboratory.
“With increasing chemical complexity, the combination possibilities become too large for trial-and-error approaches to be practical,” Lookman said. “The goal is to cut in half the time and cost of bringing materials to market.”
The White House’s Materials Genome Initiative spurred interest in accelerated materials discovery, and the LANL study is one of the first to demonstrate how machine learning can actually lead to the discovery of new materials.
Learn more in a LANL news release, and read the full article, “Accelerated search for materials with targeted properties by adaptive design.”