On 15 May 2026, a one-day Artificial Intelligence in Molecular and Materials Modeling school took place at the Bogoliubov Laboratory of Theoretical Physics in a hybrid format. Seventy people from more than five countries attended; 20 of them participated in person.
For the third time, the event was organized and conducted by employees of the Small-Particle System Sector at the Laboratory of Theoretical Physics Miroslav Iliaš and Dipayan Sen. The school first took place at BLTP in August 2025. Unlike the two previous online events, this school was held in a hybrid format.
The programme of the event focused on modelling molecules and materials on personal computers using free access software and simple but illustrative examples not requiring singnificant computing power. The participants were engaged in programming using the Python Atomic Simulation Environment (ASE) library, which allows writing, running, and refining scripts for modelling molecules.
A special feature of the third school was the introduction of a new element, machine-learning interatomic potentials (MLIPs). When integrated with ASE, these neural network-based models can predict the properties of molecular systems and materials with accuracy approaching that of first principle calculations.
Artificial intelligence assistants were added to the learning process, such as Google AI, Microsoft Copilot, and DeepSeek. With their help, special Python modelling programs were generated and tested.
In the future, the school is planned to take place four times a year in a hybrid format. The organizers are sure the event will help participants develop their research and master the modelling of molecules and materials using both traditional computational methods and modern approaches based on artificial intelligence.
