Machine Learning for Nuclear Fusion

Hello! I’m Emily, a third year PhD student at UCL’s Centre for Data Intensive Science and the Culham Centre for Fusion Energy. I’m interested in helping make nuclear fusion a viable energy source in the effort to combat climate change.

My research focus is applying data driven tools to nuclear fusion applications. Currently I am investigating physics-informed neural networks for their ability to accelerate plasma simulations on the JET reactor. My other research interests include deep learning surrogate models, cybersecurity and plasma disruption prediction.

I’m also a fellow of the Software Sustainabilty Institute where I advocate for better software practices in academia and teach introductory programming skills with the Software Carpentries.

Find an overview of my research and technical experience below or on my CV. Find me on twitter where I mainly talk about science outreach and unorthodox uses of ML. Checkout github for my personal programming projects.

Publications

  • E. Lewis, S. Pamela, Y. Andreopoulos, G. Szepesi, “Neural surrogates for fast Grad-Shafranov equilibria reconstruction”, 48th EPS Conference on Plasma Physics (2022) (Poster presentation)
  • E. Lewis, T. Mlinarevic, A. Wilkinson, “Machine Learning for Static Malware Analysis”, NCC Technical Report (2021) (Technical note)
  • E. Lewis, G. Cartland-Glover, S. Rolfo, C. Moulinec, D. Emerson, B. Merk “Modelling the draining of a molten chloride salt reactor”, 18th International Topical Meeting on Nuclear Reactor Thermal Hydraulics, Portland, Oregon (2019)

Selected Technical Experience

  • Research software engineer at the Science and Technology Facilities Council (STFC). Developed a React front-end tool for the Diamond Light Source data storage service. Deployed the threat intelligence platform MISP, for a proof-of-concept security operations centre on the STFC Cloud.
  • Data science and cybersecurity placement at NCC Group. Built an ensemble deep learning model to classify between malware and benign software.
  • Teaching assistant for UCL’s software carpentry workshop. Demonstrated the Python and Git sessions.

This site was built using Hyde, a Jekyll theme by @mdo.