AI in CHEM Seminar Series: Albert Bartók (Warwick)

This talk is part of the “AI and Machine Learning in Chemical Research and Industry” Seminar Series organized by the Aalto University School of Chemical Engineering. It is open to all members of the public. Registered students in course CHEM-E4190 can also obtain 1cr by attending the seminars and completing the assignments.

Date and location

• Wednesday 7 May 2025 @ 13:15-14:00
• A304 Ke2 lecture hall in the main building of the School of Chemical Engineering, Kemistintie 1, 02150 Espoo.

Agenda

• 13:00-13:15. Setup and brief info for the registered students.
• 13:15-14:00. Seminar by Albert Bartók, lecture hall A304.
• 14:00-onwards. Coffee, netwoking and mingling in the lobby adjacent to the lecture hall.

Seminar info

Materials modelling across the scales

Albert P. Bartók

School of Engineering and Department of Physics, University of Warwick, UK

The past two decades have seen a transformative change in atomistic modelling with the development of machine-learned interatomic potentials, which allow quantum-accurate simulations at an affordable computational cost. While the formalism of these models has converged, there still remain open questions about the optimal way to generate training databases as well as about the reliability of potentials. In this talk, I will present our efforts to automatically generate atomic databases using a combination of active learning and advanced sampling methods and how the resulting potential results in exceptionally accurate potential energy surface for Mg at a pressure range of 0-600 GPa. I will also show how fast and accurate potentials can help us discover novel phenomena, illustrated by our observation on how helium affects dislocation mobility in tungsten. Finally, I will report how transfer learning may be used to fine-tune foundation models using a little amount of data, resulting in accurate, but application-specific potentials.

About the speaker

Albert Bartók-Pártay is an Associate Professor at the University of Warwick. He earned his Ph.D. degree in physics from the University of Cambridge in 2010, his research having been on developing interatomic potentials based on ab initio data using machine learning. He was a Junior Research Fellow at Magdalene College, Cambridge, and later a Leverhulme Early Career Fellow. Before taking up his current position, he was a Research Scientist at the Science and Technology Facilities Council. His research focuses on developing theoretical and computational tools to understand atomistic processes.