In collaboration with Francesca Menescardi, details to be  agreed with interested students

The course is a practical introduction to the Local Atomic Tensors Trainable Expansion and other functionalities of the in-house developed PANNA (Properties from Artificial Neural Network Architectures)  for the construction of Machine Learned Interatomic Potentials based on ab-initio data.

An overview of the many alternative approaches that have been developed in recent years and the directions we want to move will be given.