Speaker: Nathan Cummings, UKAEA

Abstract: Modern developments in artificial intelligence and machine learning (AI/ML), cloud- computing and collaborative communities present powerful opportunities for growth in the field of nuclear fusion. Fusion is no stranger to collaboration, as demonstrated by the scale and ambition of the ITER project, and this shared vision and drive for clean and sustainable energy production can be further leveraged by sharing our data in a Findable, Accessible, Interoperable and Reusable (FAIR) [1] way.

Through the development and adoption of community-driven standards for describing our data, backed-up by thoughtfully curated provenance descriptions and metadata, we can deliver interoperability between our myriad datasets and enable a new generation of fusion researchers. For example, we can apply modern AI/ML methods, whose efficacy is so dependent on the volume, velocity, and quality of input data, to expand on our understanding of magnetic confinement fusion (MCF) and tokamak operation.

We will look at some of the considerations around making fusion data ‘more FAIR’, how to store and share the data, how to update it, and crucially, what modern technology stacks will enable scalable, performant access to data, discussing the UK’s MAST tokamak as an demonstration. In addition, we will see examples of the kind of work that can be facilitated by such an approach.

We will also look at the solutions developed within other scientific fields that produce large volumes of data. The Pangeo project addresses these issues for climate modelling and earth-sciences, so we will have a brief introduction to their approach and discuss how we can learn and borrow from them to make fusion data more open and FAIR.

[1] Wilkinson, M., Dumontier, M., Aalbersberg, I. et al. The FAIR Guiding Principles for scientific data management and stewardship. Sci Data 3, 160018 (2016).

Bio: Nathan is a Data Engineer at the UK Atomic Energy Authority with a background in Physics. He specialises in data quality, advocating for FAIR (Findable, Accessible, Interoperable and Reusable) and open data policies. He was involved in the FAIR4fusion project focussing on data provenance, as well being involved with FAIR data pipelines for COVID-19 modelling. His current focus is on developing an open and FAIR fusion database for both experimental and simulation data for as many fusion labs as possible, to be leveraged by ML engineers and researchers everywhere to push the boundaries of fusion energy research.