The company says its artificial intelligence model has already been used by researchers to create 736 new materials in the lab.

DeepMind claims to have made a new discovery that could lead to the creation of new materials for future technologies, Silicon Republic reports.

The company said that one of its artificial intelligence models has discovered 2.2 million new crystals, 380,000 of which could be used to create new materials that could drive various forms of technology.

“Our research boosted the discovery rate of materials stability prediction from around 50pc to 80pc – based on an external benchmark set by previous state-of-the-art models,” DeepMind said in a blog post.

According to DeepMind, these materials could be used to develop devices such as superconductors, supercomputers and next-generation batteries for electric vehicles. The company said it has provided researchers with the predictions of stable materials made by the artificial intelligence model.

“We also managed to scale up the efficiency of our model by improving the discovery rate from [less than] 10pc to [more than] 80pc – such efficiency increases could have significant impact on how much compute is required per discovery”, the company continued in the blog post.

The AI model, known as Graph Networks for Materials Exploration (Gnome), is denoted as a “graph neural network model.” It employs two distinct pipelines for the exploration of stable, low-energy materials.

According to DeepMind, one of these pipelines generates potential material candidates with structures resembling known crystals. The second pipeline takes a more experimental approach, employing a “randomized approach” based on chemical formulas.

The company states that the outputs from both pipelines undergo evaluation through “established density functional theory calculations” before being incorporated into the Gnome database.

DeepMind asserts that external researchers have successfully synthesized 736 novel materials in laboratories, leveraging predictions from the AI model. However, the company emphasizes that the realization of new technologies stemming from these materials hinges on our capability to effectively manufacture them.

“Our research – and that of collaborators at the Berkeley Lab, Google Research and teams around the world – shows the potential to use AI to guide materials discovery, experimentation and synthesis,” DeepMind said.

In the preceding year, DeepMind declared a significant scientific advancement as its AlphaFold model successfully predicted the structure of nearly every known protein in the scientific domain, totaling more than 200 million.