While artificial intelligence appears to hold promise in many applications, being able to imagine artificial proteins or create new molecules, a recent new study has invented 31 million materials that do not yet exist.
Before scientists are ready to bring a new material to market with a number of specific properties they need tests, data and a series of precise calculations. However, researchers at the Jacobs School of Engineering at the University of California San Diego have decided to entrust “superbrains” with this task.
So a new artificial intelligence algorithm called M3G Net has been able to predict the structure and dynamic properties of any material, existing or new, and has been used to construct a database of over 31 million new materials which have yet to be synthesized and whose properties are predicted by the algorithm itself.
It is a great tool because it will be able to facilitate (and greatly) the lives of scientists of materials, who can now invent and test new materials much faster. The news didn’t end there: artificial intelligence could also be used to run dynamic and complex simulations to further validate property predictions.
“For example, we are often interested in the rate at which lithium ions diffuse into a lithium ion battery electrode or electrolyte. The faster the spread, the faster you can charge or discharge a battery“, say the study authors.”We have demonstrated that M3GNet IAP can be used to predict the conductivity of a material’s lithium with good accuracy. We truly believe that the M3GNet architecture is a transformational tool that can greatly expand our ability to explore new material chemistries and structures.“