The U.S. Department of Energy’s Argonne National Laboratory is working to use machine learning and artificial intelligence to build a better battery.

Should the DOE’s efforts prove fruitful, it could be a positive development for the medical device industry, where batteries have proven to be a technological stumbling block when it comes to device miniaturization.

Argonne researchers created a database of approximately 133,000 small organic molecules that could form the basis of battery electrolytes with a computationally intensive model called G4MP2, which represents 166 billion larger molecules that scientists wanted to probe for electrolyte candidates, according to a news release.

The researchers applied a machine-learning algorithm to relate the known structures from the small data set to more coarsely modeled structures from the larger set, using a less computationally taxing modeling framework based on density functional theory. It is less accurate than G4MP2, but density functional theory provides a good approximation, according to the DOE.

Argonne researchers say that using G4MP2 as a gold standard could result in those researchers training the density functional theory model to incorporate a correction factor, improving its accuracy while keeping computational costs down.

“When it comes to determining how these molecules work, there are big tradeoffs between accuracy and the time it takes to compute a result,” Argonne data science and learning division director Ian Foster said in the release. “We believe that machine learning represents a way to get a molecular picture that is nearly as precise at a fraction of the computational cost.”

“The machine learning algorithm gives us a way to look at the relationship between the atoms in a large molecule and their neighbors, to see how they bond and interact, and look for similarities between those molecules and others we know quite well,” added Argonne computational scientist Logan Ward. “This will help us to make predictions about the energies of these larger molecules or the differences between the low- and high-accuracy calculations.”

“This whole project is designed to give us the biggest picture possible of battery electrolyte candidates,” said Argonne chemist Rajeev Assary. “If we are going to use a molecule for energy storage applications, we need to know properties like its stability, and we can use this machine learning to predict properties of bigger molecules more accurately.”