Program
Machine Learning for Material Science (a biased introduction)
- achieving the accuracy of quantum mechanics, without the electrons
- kernel ridge regression and Gaussian processes
- artificial neural networks with predefined descriptors
- message passing neural networks with learnable descriptors
- more data or more physics ?
- challenges: uncertainty prediction, active learning, data efficiency, delta-learning,...
full schedule
17/03/25 2PM rm 003, 18/03/25 2PM rm 003, 19/03/25 2PM rm 003, 20/03/25 2PM rm 003, 25/03/25 4PM rm 131, 27/03/25 4PM rm 131, 01/04/25 4PM rm 131, 03/04/25 4PM rm 131, 08/04/25 4PM rm 131, 10/04/25 4PM rm 131, 17/04/25 4PM rm 131