We have developed a methodology for predicting the properties of elastically deformed silicon in a wide range of strains. Screening the large 6D strain space with a tailored neural network as a surrogate model, we have predicted a possibility of the direct bandgap silicon crystal as well as the crystal with the metallic properties. The proposed method paves a new way to fast research in the field of elastic engineering for a wide range of materials. The corresponding article is published in the Proceedings of the National Academy of Sciences.
Take a look at the press releases and articles in news outlets:
- AI Predicts Effect Of Strain On Semiconductors
- Neural network strains to confer novel properties on materials
- Scientists develop AI to predict and engineer material properties
- Using Artificial Intelligence to Engineer Materials’ Properties
- Using AI to facilitate deep elastic strain engineering
- AI that predicts and engineers material properties
- Scientists develop AI to predict and engineer material properties
- NTU, MIT and Russian scientists develop AI to predict and engineer material properties
- Using artificial intelligence to engineer materials’ properties
