According to a recent study, machine learning could aid in the creation of new metal types with advantageous characteristics like resistance to rust and high temperatures. A variety of industries could benefit from this; for instance, spacecraft could be improved with metals that function well at lower temperatures, while boats and submarines could benefit from corrosion-resistant metals. Currently, attempts to produce new metals are mostly conducted in laboratories by scientists. Typically, they begin with one well-known element, such as iron, which is readily available and malleable, and then add one or two more to examine how it affects the base material. Trial & error is a hard process that invariably produces more failures than successful outcomes.

A little over 15 years ago, scientists at Kyoto University in Japan made a remarkable discovery. When they added just four proteins to a skin cell and waited about two weeks, some of the cells underwent an unexpected and astounding transformation: they became young again. They turned into stem cells almost identical to the kind found in a days-old embryo, just beginning life's journey. At least in a petri dish, researchers using the procedure can take withered skin cells from a 101-year-old and rewind them so they act as if they'd never aged at all. Now, after more than a decade of studying and tweaking so-called cellular reprogramming, a number of biotech companies and research labs say they have tantalizing hints that the process could be the gateway to an unprecedented new technology for age reversal.

The team managed to find these new metals through a combination of AI and lab experiments. First, they had to overcome a significant challenge: a lack of existing data they could use to train the machine-learning models. They trained the models on the data they had--several hundred data points describing the properties of existing metal alloys. The AI system used that data to make predictions for new metals that would exhibit low invar. The researchers then created those metals in a lab, measured the results, and fed those results back into the machine-learning model.