Once upon a time, explorers traveled the world in search of new discoveries. Currently, various types of spacecraft and robots are being sent into space to discover new information and gravitational objects.

Now, artificial intelligence is adding a new dimension to exploring the world. Chemists are trying to discover new substances using artificial intelligence technology developed by Google's DeepMind AI. In general, we know about 48,000 crystals or crystal structures.

This information helps to understand the properties of different objects. Now DeepMind’s AI networks have created the potential to create millions of possible objects. The artificial intelligence network Google DeepMind could revolutionize the creation of new objects.

This kind of artificial intelligence technology can be a new means of innovating a variety of important technologies, including high-performance batteries, solar panels and computer processors.

In this regard, DeepMind researcher Ekin Dougus Cubuk said: "Everyone has the opportunity to further develop their own technology through artificial intelligence at any time." The AI ​​model used in new material design is called Graph Networks for Materials Exploration.

In a nutshell, this is what is called the genome. These genomes can be used to predict inorganic crystal structures. We know that the six-dimensional symmetry of snowflakes is due to the formation of ice crystals. Six-dimensional symmetry means that an object looks the same when rotated six times, or 60 degrees and 120 degrees.

Organic crystals contain carbon-hydrogen bonds. To date, we know about 48,000 possible inorganic crystals. New AI models have now discovered more than 2 million new crystals. The new structure is a bit weak, but the goal is a stable structure.

To date, the lab has produced 700 crystals. GNOME is a graph neural network. This artificial intelligence network can examine various aspects of relationships between objects. This also allows the network to collect information about atoms and their chemical bonds.

To do this, the researchers used various data and databases of known inorganic crystals to train the artificial intelligence network. Based on this training, the model can create new possible objects and crystals by changing different elements or creating symmetries of known crystals. It also provides insight into the stability of newly formed crystals.