Project: Augmenting DFT modelling of vibrational spectra through Machine Learning and Deep Neural networks using small data sets.

Thijs van Wijk, MSc. Physics’23
[Materiomics AAP-UHasselt, 2023-2027]

Quantum chemical modelling, specifically density functional theory (DFT) plays an important role in modern materials design, as it provides direct access to atomic scale properties not available experimentally. However, as the system size and complexity grows so does the computational cost of the DFT calculations. The aim of this project is to extend the range of problem sizes that can be investigated at DFT accuracy through the use of machine learning and deep neural networks. For this, methods will be developed focusing on the efficient learning from small data sets. This AI-enhanced DFT approach will be applied on the modelling of vibrational spectra for the fingerprinting of defects in diamond and the characterisation of weakly bonded crystals.

Acknowledgement financial/compute support

Permanent link to this article: