The Computing Division Technical Meetings are a platform for

  • presenting Pinboard papers under review for journal or conference publication,
  • inviting speakers who are current or prospective UKAEA collaborators at external organisations,
  • presenting work done in PhD projects funded by or co-supervised by UKAEA,
  • presenting work done during summer placements or other secondments to UKAEA.

 

If you would like to invite a speaker on a topic that would be of interest to one or more Units within the Computing Division, but is not currently collaborating on a UKAEA project, please consider nominating them for a Computing Division Cross-Disciplinary Seminar.

These meetings are normally recorded. Recordings of past meetings can be found here:

CD Technical Meetings Archive on UKAEA Sharepoint

CD Technical Meeting (ML15): Learning compact space representation of plasma states from the FairMAST dataset

→ Europe/London
Andrea Loreti
Description
Machine Learning, Uncertainty Quantification and Data Science 

 

The FairMAST dataset provides a unique machine-learning-ready collection of more than 11,000 MAST discharges, including magnetic, equilibrium, current, voltage and auxiliary diagnostic measurements. However, the data present significant challenges due to missing signals, heterogeneous diagnostic modalities and widely varying sampling frequencies. 

This work presents a Variational Autoencoder (VAE) framework designed to learn compact latent-space representations of multimodal plasma diagnostics while remaining robust to sparse and incomplete measurements. 

Using magnetic diagnostics from FairMAST, the model is applied to both equilibrium reconstruction and short-term forecasting tasks, demonstrating accurate recovery of EFIT-derived quantities, including poloidal flux maps and plasma boundary information.