Researchers from the Southern Federal University (SFedU) have created a technology that allows artificial intelligence to analyze X-ray absorption spectroscopy data in a short time — the PyFitit framework. Now, work that usually took six months can be done in a couple of minutes — incredible productivity.
The arrangement of atoms, their chemical state, and other important characteristics of materials are areas where X-ray absorption spectroscopy is indispensable. Thanks to the use of the PyFitit framework, a neural network trained on specially collected databases is now able to analyze spectra with high accuracy and reliability.
Using machine learning based on the PyFitIt framework (which has been developed in our Institute for many years), as well as comprehensive databases (so far only for chromium and vanadium, but even this is significant - for example, chromium compounds are the basis of the Phillips catalyst, with which half of all polyethylene in the world is produced), we have achieved high-quality and reproducible results.
SFedU scientists noted that the accuracy of the analysis is also increased by the procedure of joint analysis of neighboring absorption edges for different elements ("two in one"), which was proposed for the first time. The use of the PyFitit framework has made data analysis faster and more accurate.
Read more on the topic:
Russian scientists have found out how to protect space electronics from radiation
PNIPU students have created prostheses that can be printed on a 3D printer