Computing Division Cross-Disciplinary Seminars (2026)

1 Jan 2026, 06:00 → 31 Dec 2026, 20:00 Europe/London

Culham/Remote

Aravinda Perera (Plasma Simulation)

The monthly Computing Division Cross-Disciplinary Seminars complement the Computing Division Technical Meetings.

They are intended to provide a spotlight to invite external speakers working on both new and well-established 

• computational methods,
• frameworks,
• software,
• practices,
• and other advancements

in non-fusion fields and sectors, or within groups working in fusion but are not currently collaborating with UKAEA, and could be of interest to the UKAEA computing community.

These seminars are normally recorded. Recordings of previous seminars can be found here:

Computing Division Cross-Disciplinary Seminar recordings

Recordings of past seminars

Tuesday 24 February

13:00 → 14:00 ML/AI, UQ and Data Science: GPGreen: Learning Linear Operators with Gaussian Processes

Conveners: Thomas Cowperthwaite (University of Cambridge), Henry Moss (Lancaster University)

34_GPGreen_Learning_Linear_Ope.pdf

13:00 GPGreen: Learning Linear Operators with Gaussian Processes

Thomas Cowperthwaite
Applied Mathematics PhD Student, DAMTP, University of Cambridge

Henry Moss
Lecturer, School of Mathematical Sciences, Lancaster University and
Early Career Research Fellow, University of Cambridge

Abstract

Operator learning has emerged as a promising data-driven approach to emulating solutions of partial differential equations (PDEs). Existing deep learning-based models lack principled uncertainty quantification, rely on access to large numbers of training examples, and remain largely uninterpretable. Here, we use Gaussian process regression to make uncertainty-aware estimates of PDE solutions. We show our method is competitively accurate compared to existing approaches, while additionally providing uncertainty quantification and improving sample efficiency. The framework exploits Kronecker structures and Fast Fourier Transforms to achieve resolution-invariant prediction cost scaling.

Speaker: Mr Thomas Cowperthwaite (University of Cambridge)

Paper: GPGreen: Learning Linear Operators with Gaussian Processes

Wednesday 11 March

14:15 → 15:15 ML/AI, UQ and Data Science: Uncertainty quantification, inverse problems and timescale acceleration in atomic simulation

Convener: Aravinda Perera (Plasma Simulation)

34_GPGreen_Learning_Linear_Ope.pdf

14:15 Uncertainty quantification, inverse problems and timescale acceleration in atomic simulation

Tom Swinburne is an Assistant Professor in Mechanical Engineering at the University of Michigan, Ann Arbor; a Visiting Researcher at CNRS Physique; and Associate Editor of Computational Materials Science

tomswinburne.github.io

Abstract

Recent years have seen an explosion in the use of data-driven tools in atomic simulation, following broader trends across computational science. Current efforts have focused on interpolating some approximate electronic structure method; I will discuss how these same tools can be used to quantify model form uncertainties[1], learn from experimental data sources[2], harness adjoint methods familiar to e.g. FEM[3] and provide new strategies to extend simulation timescales[4]. Applications specific to fusion materials science and transferability to general modelling objectives will be discussed.

[1] T.D. Swinburne and D. Perez, Mach. Learn.: Sci. Technol., (2025)
[2] T.D. Swinburne, C. Lapointe, C. Marinica, Nat. Comm. (2026) & NeurIPS AI4Mat (2025)
[3] I. Maliyov, P. Grigorev, T.D. Swinburne, NPJ Computational Materials (2025)
[4] T.D. Swinburne, Phys. Rev. Lett (2023)

Speaker: Dr Thomas Swinburne (University of Michigan)

Wednesday 25 March

15:00 → 16:00 Numerical Methods: Energy-preserving coupling of explicit particle-in-cell with Monte Carlo collisions

Convener: Aravinda Perera (Plasma Simulation)