This training session provides an overview on MedeA's new machine-learned potential capabilities.

An introduction to the new Machine-Learned Potential Generator (MLPG) of MedeA is provided which allows the user to create their own machine-learned potential (MLP). These can then be used with MedeA LAMMPS to perform MD simulations with ab initio accuracy.

We'll discuss guidelines for constructing a training set, performing DFT calculations for all structures included in the training set, and efficient storage of relevant data for use with the MLPG.

Attendees will learn to perform LAMMPS calculations with MLPs and create and apply their own MLPs.