According to a new research paper, Google’s DeepMind has discovered hundreds of thousands of new hypothetical material designs. Their hope with this breakthrough is to help improve the production of materials such as computer chips, batteries, and solar panels.
Published in Nature, the discovery and synthesis of these new materials could take years to produce physical results. But their potential is quite clear. With machine learning and other advancements, these materials could see production much sooner than something such as lithium-ion batteries which took decades.
Ekin Dogus Cubuk, a research scientist at DeepMind said, “We’re hoping that big improvements in experimentation, autonomous synthesis, and machine learning models will significantly shorten that 10 to 20-year timeline to something that’s much more manageable.”
The AI used by the DeepMind team was trained on data from the Materials Project. That’s an international research group founded at the Lawrence Berkeley National Laboratory in 2011. Currently, it’s researching around 50,000 known materials.
This research hopes that the shared information from this breakthrough could help other researchers in the quest for material discovery. “Industry tends to be a little risk-averse when it comes to cost increases, and new materials typically take a bit of time before they become cost-effective,” said Kristin Persson, director of the Materials Project.
Kristin went on to say, “If we can shrink that even a bit more, it would be considered a real breakthrough.” Though not exactly sci-fi sci-fi-inspired breakthrough, the foundation of this research could in theory produce materials that could revalorize architecture, space travel, and more.
Since this paper, DeepMind said that it will now turn its focus to predicting how easily the materials discovered can be synthesized in a lab setting. If successfully done, synthesizing new materials could help reduce costs associated with this kind of R&D and lay the groundwork for even greater possibilities of discoveries in the future.