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 (ML12): Model order reduction method for Hamiltonian dynamics using deep learning

→ Europe/London
Guillaume Steimer
Description
Machine Learning, Uncertainty Quantification and Data Science 

 

Many mathematical models of physical processes pose significant challenges for numerical simulation due to their high dimensionality and complexity. This is particularly the case in nuclear fusion, where such models are very demanding in terms of computational time and resources. The issue becomes even more critical when repeated or real-time evaluations are required, as in optimization or control applications. To overcome these limitations, reduced-order models (ROMs) can be developed to provide fast approximations of the full systems, at the cost of some loss in accuracy. When dealing with Hamiltonian systems, it is essential to preserve their underlying geometric structure in the reduction process. In this work, we introduce the fundamentals of symplectic model reduction, based on the symplectic Galerkin projection, and extend this framework by replacing linear approximations with neural networks. The proposed approach is illustrated on a particle-in-cell (PIC) model of the Vlasov–Poisson equation.