Lecturer: Dr Christoph Schran
This module offers an overview of advanced techniques for atomistic simulations, with an emphasis on machine learning potentials and path integral techniques for the inclusion of nuclear quantum effects. It comprises twelve lectures and two practical sessions.
The aims of this module are:
- To provide an understanding of the limitations of standard atomistic simulation techniques and how to address these.
- To provide the conceptual foundation of the principles behind machine learning for the representation of potential energy surfaces and path integral techniques for the inclusion of nuclear quantum effects
- To provide training in performing and analysing of these two advanced simulation techniques for molecular and materials modelling.
Learning outcomes
Students attending this lecture will:
- gain a detailed understanding of two advanced simulation techniques, namely machine learning potentials for the representation of interactions and path integral techniques for the inclusion of the quantum nature of nuclei.
- be able to analyse and perform their own simulations using path integral techniques and machine learning potentials for molecular and materials modelling.
- be able to solve simple analytical problems (suitable for examination) involving atomic descriptors, regression models, path integrals and quantum mechanics.