Tomsk State University Develops AI Model to Identify Energy Materials
The model could significantly reduce the time needed to study promising compounds.

Tomsk State University is developing a new artificial intelligence model to help identify energy materials. The idea is to assess the properties of compounds computationally and focus experimental research on the most promising candidates.
According to the researchers, developing new materials using traditional approaches can take several years. Before finding a suitable compound, scientists have to synthesize and study numerous substances. The new technology is expected to significantly reduce the time required.
“When searching for new materials, researchers usually start with known compounds and modify their structures—for example, by replacing or adding specific groups of atoms. AI can be trained on computational and experimental results to identify relationships between a substance’s structure and its properties. Given a particular structure, the model could then estimate the desired characteristics and assess the compound’s potential. This is known as forward prediction,” says Leonid Kuznetsov, a graduate of the Faculty of Chemical Technology at the D. Mendeleev University of Chemical Technology of Russia and a junior researcher at the university’s Laboratory of Artificial Intelligence in Chemistry and Materials Science.








































