Machine learning based techniques to study chemical reactions
Identyfikator grantu: PT01110
Kierownik projektu: Marta Łabuda
Realizatorzy:
- Bartosz Majewski
- Aleksandra Rzepka
Politechnika Gdańska
Wydział Fizyki Technicznej i Matematyki Stosowanej
Gdańsk
Data otwarcia: 2024-01-16
Streszczenie projektu
The main focus of our work is devoted to the application and development of quantum chemistry and quantum dynamics methods as well as, recently, machine learning based techniques in order to study chemical reactions. Our current topics focus on: 1) studying the mechanisms of interactions induced by collisions of atoms, highly charged ions, electrons, and protons with molecules of biological meaning in the context of radiation damage and astrobiology, 2) performing calculations to observe, analyze and understand the mechanisms of decay of molecules in the collision processes, 3) developing numerical protocols based on a data-driven approach and machine learning methods (supervised and unsupervised learning) to perform simulations of the chemical reactions and to check their efficiency with respect to the statistical and structural methods 4) preparation and creation of the datasets and database of the possible identified fragments with the idea to build up a new tool for the investigation of chemical reaction paths. We plan to apply and develop several predictive models in order to automatize the quantum chemical calculations on structure and dynamics of the chemical processes.