PhD project: Generative AI for lattice field theory

Project description

This project is being offered by Dr Gurtej Kanwar with a start date of September 2024.

Lattice simulations allow us to non-perturbatively understand strongly interacting field theories, importantly including the strong nuclear force, as described by Quantum Chromodynamics (QCD). Unfortunately, state-of-the-art lattice field theory calculations are often limited by the large computational cost required to perform Monte Carlo sampling of field fluctuations in vacuum or thermal states. A breakthrough in this sampling step would have profound consequences on our understanding of QCD and other strongly interacting field theories.

Inspired by the parallels between generative AI tasks and Monte Carlo sampling for lattice simulations, this project aims to develop specialised generative models for lattice field theory. The idea is to challenge these models to independently generate snapshots of field configurations that reproduce the rigorous physical distribution prescribed by the theory under study. Early work over the past several years has shown that such techniques can indeed be applied without introducing bias into the sampling process and can accelerate Monte Carlo sampling by orders of magnitude. As a PhD student on this project, you will work at the frontier between physics and computational research, with the particular balance subject to your interests.

Project supervisors

The project supervisors welcome informal enquiries about this project.

Find out more about this research area

The links below summarise our research in the area(s) relevant to this project:

What next?

More PhD projects