AI has swept the field of materials science, and DeepMind GNoME has successfully predicted 2.2 milli

Mondo Technology Updated on 2024-01-19

On November 30, Google's DeepMind once again shook the technology world, successfully ** 2.2 million crystal structures through the new AI tool GNOME, subverting the field of materials science. This breakthrough result was published in Nature and attracted widespread attention. This is not only a revolution in materials science, but also the potential of AI in chemistry, providing scientists with a powerful tool to accelerate the discovery and synthesis of new materials.

Behind the AlphaFold series leading the revolution in the field of biology, DeepMind once again demonstrated its strong scientific and technological strength. The advent of GNOME (Graph Neural Network for Organic Materials Exploration) has injected new vitality into materials science. The tool leverages advanced graph neural network (GNN) models to successfully identify 2.2 million crystal structures, of which 380,000 are considered potentially transformative。This achievement not only makes human civilization take an important step forward in material exploration, but also makes people look forward to the application of AI in other scientific fields.

GNOME works based on two key workflows: structural piping and component piping. Structural pipelines create candidates with similar crystal structures to known crystals, while compositional pipelines take a more random approach. The outputs of these two workflows are evaluated by density functional theory calculations, and the results are added to the GNOME database to inform the next round of active learning Xi. GNOME uses a graph neural network model to process the total energy of the crystal and process the connections between elements in the form of a graph, which greatly improves the discovery efficiency of new crystal materials.

Of the 2.2 million crystal structures, 380,000 are considered to be the most stable, offering great potential for materials for transformative technologies in the future. These materials can not only be used in the research and development of superconductors and electric vehicle batteries, but also are expected to be used in supercomputing power supply and other fields. In addition, scientists around the world have successfully synthesized 41 new materials with the assistance of GNOME, demonstrating the potential role of AI in material synthesis.

Gnome not only provides the best structure of the new material, but also publishes a database of newly discovered crystals to the research community, providing scientists with a new toolbox. Through this"AI formulations"Scientists can quickly develop new materials to support applications in a variety of fields. In the research at the Lawrence National Laboratory in the United States, more than 41 new materials have been successfully synthesized under the guidance of GNOME, opening up new possibilities for AI-driven material synthesis.

In the past, scientists searching for new materials often had to tweak known crystals or experiment with new combinations of elements, which was an expensive and time-consuming trial-and-error process. However, with the advent of GNOME, the cost of material discovery has been greatly reduced, and the discovery rate has increased significantly. The advent of GNOME has not only accelerated the process of material discovery, but also provided scientists with a new perspective on how the power of AI may be able to unlock the mysteries of the natural world more quickly and advance human civilization.

The success of Deepmind's GNOME tool**2.2 million crystal structures has revolutionized the field of materials science. This is not only a huge breakthrough in materials science, but also the pinnacle of artificial intelligence in the field of chemistry.

In the future, we have reason to expect more applications of AI in scientific research, accelerate human cognition of the laws of nature, promote the development of science and technology, and lead us into a new era of data-driven. Deepmind's GNOME will become a powerful assistant for human beings to explore the unknown and unveil the mystery of science for us.

I'm a tech creator

Related Pages