Tech Breakthrough: AI Finds Over Two Million Crystals, 800 Years Faster Than Traditional Methods!

A British-American artificial intelligence research laboratory found a whopping 2.2 million crystal structures that could lead to leaps in progress in various fields like renewable energy and advanced computation.

“From microchips to batteries and photovoltaics, the discovery of inorganic crystals has been bottlenecked by expensive trial-and-error approaches,” as written in the paper. “Our work represents an order-of-magnitude expansion in stable materials known to humanity.”

Some 48,000 materials had already been identified – some known for millennia, like bronze and iron, some discovered more recently. The team behind Google DeepMind used the AI tool known as GNoME, which is more than 45 times larger than the number of such substances unearthed in the history of science. AI can shortcut years of experimental graft. Indeed, the number of substances found is the equivalent of almost 800 years of previous experimentally acquired knowledge. Machine learning was first used to generate candidate structures and then gauge their likely stability. The new compounds could be used to develop neuromorphic computing – with chips used to mirror the workings of the human brain – and invent versatile layered materials. From the 2 million crystal structures found, 381,000 of the most promising will be made available to fellow scientists to advance technologies such as solar cells and superconductors.