Artificial Neural Networks Used To Predict New Stable Materials

Share

Lately, artificial neural networks, some algorithms inspired by connections in the brain, were trained to perform a variety of tasks, from detecting obstacles and pedestrians in self-driving cars to analyze medical images and beyond. More recently, scientists at the University of California San Diego are coaching artificial neural networks to predict new stable materials.

“Predicting the stability of materials is a central problem in materials science, physics, and chemistry. On the one hand, you have traditional chemical intuition such as Linus Pauling’s five rules that describe stability for crystals in terms of the radii and packing of ions. On the other, you have expensive quantum mechanical computations to calculate the energy gained from forming a crystal that has to be done on supercomputers. What we have done is to use artificial neural networks to bridge these two worlds,” explained the study’s leading author Shyue Ping Ong from the University of California San Diego.

Artificial neural networks are now coached to predict new stable materials

The researchers developed models that can identify stable materials in two categories of crystals known as garnets and perovskites by coaching artificial neural networks to predict crystal’s formation energy only by electronegativity and ionic radius of the constituent atoms as inputs.

“Garnets and perovskites are used in LED lights, rechargeable lithium-ion batteries, and solar cells. These neural networks have the potential to greatly accelerate the discovery of new materials for these and other important applications,” highlighted Weike Ye, the study’s co-author.

These new models are about ten times more precise than older machine learning models and are sufficiently fast to effectively analyze thousands of materials within a few hours on a regular computer.

The scientists made their result publicly available on Crystals.Ai where every person can use these new artificial neural networks to calculate the formation energy of any garnet or perovskite composition. However, the engineers hope to extend the AI to other groups of crystals, too.

mm

Jasmine holds a Master’s in Journalism from Ryerson University in Toronto and writes professionally in a broad variety of genres. She has worked as a senior manager in public relations and communications for major telecommunication companies, and is the former Deputy Director for Media Relations with the Modern Coalition. Jasmine writes primarily in our LGBTTQQIAAP and Science section.


Share

Recommended For You

Leave a Reply

Your email address will not be published. Required fields are marked *