AI-Based Spectroscopy, ARTIST, Instantly Reveals Compounds’ Characteristics

A team of scientists from Aalto University, Finland and the Technical University of Denmark have advanced artificial intelligence (AI) to utterly speed up the development of innovative technologies from wearable electronic gadgets to flexible solar panels. Artificial Intelligence for Spectroscopy—abbreviated as ARTIST—rapidly decide the way a molecule will respond to light, which will better help in constructing the designer tools and products needed for upcoming technology.

Earlier, researchers study the sensitivity of molecular reactions through spectroscopy, a broadly used technique throughout the natural sciences and industry. Spectroscopy examines the internal characteristics of reactants or compounds by observing their response after a specific wavelength of light rays passed through them, which has initiated the development of new limitless technologies. Advancement in currently available analytical and computational spectroscopy can be, however, extremely expensive.

But with the help of AI-based spectroscopy, ARTIST, which offer a fundamental change in the existing approach to better decide the light spectrum range, where individual molecules respond.

Milica Todorovic, who is a postdoctoral researcher at Aalto University, said that finding and choosing appropriate molecules to be used in the device is an important part of the development, for which the scientists have to utilize previous knowledge along with some extent of chemical perception. Examining the individual spectra of molecules is then a hit-and-trial method, which is a tedious part and can consume even weeks or months, depending on the number of molecules sorted out. So, the AI developed by scientists will provide the characteristics of molecules immediately.

With its high precision and pace, ARTIST can possibly accelerate the growth of flexible electronics, comprising (LEDs) light-emitting diodes or flexible paper-like displays. ARTIST may also facilitate the production of improved batteries and catalysts.

The integrative team of experts has trained the AI by inserting a set of information regarding more than 132,000 organic molecules in just a few weeks