Materials scientists have historically used trial-and-error and intuition to discover new materials, but as chemical complexities increase, this becomes harder and harder — but machine learning could be the solution.

Researchers at Los Alamos National Laboratory attempted to combine machine learning with targeted experiments using “informatics-based adaptive design strategy,” and were able to massively increase the materials discovery process, according to a statement from the laboratory.

Using a relatively small data set of well-controlled experiments, scientists showed it was possible to find the material with the desired target, the scientists said.

Los Alamos’ supercomputers used a machine-learning algorithm that basically digitizes the trial-and-error process, vastly accelerating it from how long it would take if humans were doing it. The end result is that it takes half hte time to bring new materials to market.

“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,” Turab Lookman, a physicist and materials scientist in the Physics of Condensed Matter and Complex Systems group at Los Alamos National Laboratory, said in the statement. “Finding new materials has traditionally been guided by intuition and trial and error,” said Lookman.”But with increasing chemical complexity, the combination possibilities become too large for trial-and-error approaches to be practical.”

He added: “The goal is to cut in half the time and cost of bringing materials to market. What we have demonstrated is a data-driven framework built on the foundations of machine learning and design that can lead to discovering new materials with targeted properties much faster than before.”