Artificial Intelligence Designed Stronger Metallic Glass From Steel!

In a study published in Science Advances, a group of researchers explained that they were able to quickly design and manufacture long-lasting metallic glass using artificial intelligence. If you mix two or more metals, you will create a new material different from the properties of these two metals. It’s an alloy. In ordinary alloys, atoms are arranged in regular order. When we obtain a different sequence from this regular sequence of atoms, we obtain a new and strong alloy called metallic glass.

The atoms in these materials are not regular as in ordinary alloys but are irregular like glass. Metallic glass is both lighter and stronger than the best steel. They are also more resistant to wear. Over the past fifty years, much research has been done to develop new types of metallic glass. Thousands of different material combinations have been tried. Very few of the metallic lenses developed to date have been useful.

Professor of the research team. Dr. Chris Wolverton says 10 to 20 years have passed between the discovery of a material and its commercial use. The developed artificial intelligence application helped to discover three new metallic glass species 200 times faster than in the past.

SEE ALSO: What is Enigma Machine? How Does Enigma Work?

Researchers say that the most important difference in their work is the combination of artificial intelligence, which is called machine learning, and experimental methods that can produce and analyze hundreds of material samples very quickly. In machine learning, computer algorithms learn new things from a huge database. Algorithms establish connections between data and make inferences.

The researchers have built a large database of the results of experiments on metallic glass in the last 50 years. They then test the predictions made through machine learning very quickly and add the results to the database.

It has been seen that the accuracy of estimates made after a few repetitions has increased greatly. Initially, one of every 300-400 estimates of artificial intelligence about which material combinations could produce metallic glass gave a positive result, but one of every two or three estimates started to give a positive result.