Google DeepMind’s machine-learning tool, GN o ME, predicts and synthesizes millions of new crystal structures, expanding the possibilities for scientific research.
Crystals have long fascinated scientists and captivated the imagination with their unique properties and intricate structures. Now, Google DeepMind, an artificial intelligence company, has developed a groundbreaking machine-learning tool called GN o ME (Graph Networks for Materials Exploration) that has the ability to predict and synthesize millions of new crystal structures. This remarkable achievement has the potential to revolutionize scientific research and open doors to new discoveries in various fields.
The Vast World of Crystals:
Crystals, with their repeating atomic structures, form a vast family of compounds, each with its own distinct properties. In a recent paper published in Nature, DeepMind revealed that there are approximately 48,000 known crystal structures. However, using GN o ME, the AI system generated an astonishing 2.2 million new crystal structures, previously unknown to science.
Validation through Synthesis:
To validate the accuracy of GN o ME’s predictions, DeepMind collaborated with researchers at the University of California, Berkeley, on a second paper, also published in Nature. They selected 58 of the predicted compounds and successfully synthesized 41 of them in just over two weeks. Since then, other research groups have produced over 700 additional crystals from the predictions made by DeepMind.
Expanding the Possibilities:
DeepMind’s work has not only added to the vast library of crystal structures but has also increased the number of candidate materials for research purposes exponentially. By making a subset of the 381,000 most stable structures available to the public, DeepMind has provided a valuable resource for laboratories interested in exploring these new materials. Among the structures are thousands of crystals with the potential for superconductivity and hundreds of potential conductors of lithium ions, which could have applications in batteries.
The Beginning of Exploration:
While DeepMind’s achievement is impressive, it is just the beginning of a new era of exploration. Aron Walsh, a materials scientist at Imperial College London, acknowledges the significance of the work but emphasizes that there is much more to discover. The AI system focused on crystals that form under specific conditions, and there is still an entire universe of materials to explore, from amorphous solids to gases, gels, and liquids.
The Promise of AI in Crystal Research:
The potential applications of DeepMind’s AI tool go beyond the discovery of new crystals. It may also shed light on the underlying rules that govern crystal formation, providing valuable insights for scientists. The AI’s ability to predict crystals made from six elements, previously believed to be rare, has already challenged existing knowledge. Additionally, the predictions made by AI could save researchers significant time and effort by eliminating the need to synthesize each new material manually.
Conclusion:
DeepMind’s AI breakthrough in generating 2.2 million new crystal structures marks a significant milestone in scientific exploration. While the true value and applications of these crystals are yet to be determined, the techniques used by GN o ME have the potential to transform the field of materials research. As scientists delve deeper into the possibilities presented by AI, they may uncover a wealth of knowledge about crystal formation and discover materials with extraordinary properties that could shape the future of technology and innovation.
Leave a Reply