Advances in artificial intelligence (AI) research have led to a wide range of developments in this industry over the last ten years. Autonomously driven cars are one example, but everyday applications like search engines and spam filters illustrate the versatility of methods from the field of AI.
Infrared spectroscopy is a very valuable experimental method to gain insights into molecules’ world. Infrared spectra provide information on the composition and properties of substances and materials. In many cases, these spectra are complex. A detailed analysis makes computer-aided simulations indispensable. Quantum chemical calculations in principle enable precise prediction of infrared spectra and their applicability in practice is made difficult by high computational effort associated with them. For this reason, reliable infrared spectra can be calculated for relatively small chemical systems.
An international group of researchers, led by Philipp Marquetand from the Faculty of Chemistry at the University of Vienna, has found a way to accelerate these simulations using AI. For this, artificial neural networks are used, mathematical models of the human brain. These can learn the complex quantum mechanical relationships that are necessary for the modeling of infrared spectra by using a few examples. The scientists can carry out simulations within a few minutes that would otherwise take thousands of years, even with modern supercomputers, without sacrificing reliability.
"We can now finally simulate chemical problems that could not be overcome with the simulation techniques used up to now," says Michael Gastegger, the first author of the study.
Based on the study’s results, the researchers are confident that their method of spectra prediction will be widely used in the analysis of experimental infrared spectra in the future.
A paper on this research was published in Chemical Science.