91¸£Àû

Thomas, who recently successfully passed his viva and now works as a Postdoctoral Researcher in Machine Learning Interatomic Potentials at the 91¸£Àû, led the workshop as part of ongoing efforts to support developments in computational materials science and machine learning-driven atomistic simulation.

The newly developed tool extends the existing Hyperactive Learning approach originally developed by Cas van der Oord to MACE models. The framework also enables expensive Density Functional Theory (DFT) calculations and model training to be moved outside of the core selection loop, helping researchers generate training data in larger and more efficient batches.

The project was supported through funding from the Physical Sciences Data Infrastructure (PSDI), with supervision provided by James Kermode.

The workshop also featured talks from fellow HetSys CDT researchers and students, including Fraser Birks, Mariia Radova, and George Simmons.

HetSys CDT is delighted to see alumni continuing to contribute to advances in computational modelling, machine learning, and interdisciplinary scientific research following completion of their PhD studies.

To find out more about former HetSys CDT students and where their careers have taken them, visit the HetSys Alumni page.