Machine learning based prediction of the electronic structure of quasi-one-dimensional materials under strain

Materials scientists from UCLA and Michigan Technological University have trained artificial intelligence (AI) to predict the electronic properties of nanoscale strands of building blocks of materials after they have been twisted, stretched, or compressed. Described as ``quasi-one-dimensional,'' this class of materials include nanotubes, nanoribbons, nanowires, and similar configurations of materials that are long and thin. The researchers use specialized quantum mechanical simulation techniques to produce the data that the AI models are trained on. Remarkably, the AI models, based on artificial neural networks, are able to discern hidden patterns in the simulations data and to predict accurately the complex electronic structure of the strained quasi-one-dimensional materials, based on very limited training. The researchers anticipate that such AI-assisted simulation capabilities of nanomaterials may lead to the discovery of new materials for energy storage technologies, quantum computing hardware, advanced sensors, and a range of other electronic devices.