Google has used DeepMind AI to analyze and predict thousands of new materials

Google-owned DeepMind has used artificial intelligence (AI) to predict the structures of more than 2 million new materials, a breakthrough that will drive real-world technological improvements. The news was released on November 30th

The relevant research results were published in "Nature" on Wednesday local time under the title "An autonomous laboratory for the accelerated synthesis of novel materials". Attached to this site is DOI:10.1038/s41586-023-06734-w.

DeepMind researchers noted in a paper that based on their hypothesis, most of the nearly 400,000 material designs could be rapidly produced under laboratory conditions. This research is expected to help produce better-performing batteries, solar panels and computer chips

DeepMind said that after using AI to predict the stability of these new materials, they will take the next step in their research The focus shifts to predicting how easy it will be to synthesize these materials in the laboratory

#What needs to be rewritten is: ▲ Source: Nature

Actually, The discovery and synthesis of new materials is actually a very expensive and time-consuming process. For example, the commercial application of lithium-ion batteries that we can see everywhere today took about 20 years, during which countless costs and efforts were consumed.

"DeepMind researcher Ekin Dogus Cubuk said that we hope to shorten this 10 to 20 years to a more manageable range through experiments, autonomous synthesis and huge improvements in machine learning models."

According to reports, DeepMind’s artificial intelligence is trained based on Materials Project data. Materials Project is an international research organization established at Lawrence Berkeley National Laboratory in 2011. It has researched about 50,000 known materials. The company said it will share the data with the research community to accelerate materials discovery. Further breakthroughs

“The industry is often somewhat risk-averse when it comes to cost increases, and new materials often take a while to become cost-effective.” Kristin Persson, director of Materials Project, said, “If If we can further shorten this time, that will be a real breakthrough."

The above is the detailed content of Google has used DeepMind AI to analyze and predict thousands of new materials. For more information, please follow other related articles on the PHP Chinese website!