AI agent helps identify material properties faster

A group headed by Dr. Phillip M. Maffettone (presently at National Synchrotron Light Source II in Upton, USA) and Professor Andrew Cooper from the Department of Chemistry and Materials Innovation Factory at the University of Liverpool signed up with forces with the Bochum-based group headed by Lars Banko and Professor Alfred Ludwig from the Chair of Materials Discovery and Interfaces and Yury Lysogorskiy from the Interdisciplinary Centre for Advanced Materials Simulation. The global group released their report in the journal Nature Computational Science from 19 April 2021.

Previously handbook, lengthy, error-prone

Efficient analysis of X-ray diffraction information (XRD) plays an important function in the discovery of brand-new products, for instance for the energy systems of the future. It is utilized to evaluate the crystal structures of brand-new products in order to discover, for which applications they may be appropriate. XRD measurements have actually currently been substantially sped up in the last few years through automation and offer big quantities of information when determining material libraries. “However, XRD analysis techniques are still largely manual, time-consuming, error-prone and not scalable,” states Alfred Ludwig. “In order to discover and optimise new materials faster in the future using autonomous high-throughput experiments, new methods are required.”

In their publication, the group demonstrates how expert system can be utilized to make XRD information analysis faster and more precise. The service is an AI agent called Crystallography Companion Agent (XCA), which teams up with the researchers. XCA can carry out self-governing stage recognitions from XRD information while it is determined. The agent appropriates for both natural and inorganic material systems. This is allowed by the massive simulation of physically proper X-ray diffraction information that is utilized to train the algorithm.

Expert conversation is simulated

What is more, a unique function of the agent that the group has actually adjusted for the present job is that it gets rid of the overconfidence of conventional neuronal networks: this is since such networks make a decision even if the information does not support a guaranteed conclusion. Whereas a researcher would interact their unpredictability and go over outcomes with other scientists. “This process of decision-making in the group is simulated by an ensemble of neural networks, similar to a vote among experts,” describes Lars Banko. In XCA, an ensemble of neural networks forms the specialist panel, so to speak, which sends a suggestion to the scientists. “This is accomplished without manual, human-labelled data and is robust to many sources of experimental complexity,” states Banko.

XCA can likewise be broadened to other types of characterisation such as spectroscopy. “By complementing recent advances in automation and autonomous experimentation, this development constitutes an important step in accelerating the discovery of new materials,” concludes Alfred Ludwig.


Disclaimer: We can make errors too. Have a good day.

Recommended For You

About the Author: livescience

Leave a Reply

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