or
Looking to list your PhD opportunities? Log in here.
We seek a student to develop deep learning models of magnetic field changes and investigate processes in Earth’s magnetosphere. There is currently an exciting growth of machine learning methods in space physics; this research will inform future research directions and space weather prediction techniques.
Earth’s magnetosphere is the region of near-Earth space where processes are dominated by Earth’s magnetic field. This highly dynamic region responds to the solar wind, a constant buffeting flow of plasma from the Sun. Large numbers of ground and space observations mean that space physics is an area ripe for deep learning, as we need to (1) understand all the complex physics (2) predict space weather conditions to protect ground and space infrastructure and (3) understand and adapt deep learning techniques to datasets with some extreme properties.
Graph neural networks (GNNs) perform deep learning on data represented as networks or graphs (i.e. a series of nodes and edges) rather than a 3D grid. Considering space physics data as coming from a network more naturally represents how we sample the rapidly changing magnetosphere.
In this project you would train graph neural networks across ground magnetometer chains. These are magnetic field observations across a wide area of the Earth, sampling a large portion of the magnetosphere. We can use these to map global magnetospheric processes. Graph neural networks can be used to predict or classify nodes (i.e. each magnetometer station), edges (i.e. spatial or magnetic relation between stations) or the entire graph (i.e. a snapshot of the magnetic field configuration). There are many physics questions that could be explored with this model, or you could alternatively focus your attention on adapting the theory behind deep learning techniques to space weather applications. There are several research directions that could significantly improve our understanding and so the project focus will be determined by student interests.
Dr Sarah Bentley and Prof Clare Watt use a variety of methods to represent and understand space physics processes. The Solar and Space Physics group is in Maths and Physics and is growing rapidly. E-mail: [Email Address Removed] and [Email Address Removed]).
In addition to regular supervisor meetings, students are encouraged to participate in group activities, including our seminar series with national and international experts. We offer Masters-level modules to provide any necessary subject-specific knowledge or skills, and you are encouraged to attend national summer schools and apply to competitive international summer schools. We will help you choose conferences and workshops to discuss your results with the scientific community.
The Principal Supervisor for this project is Dr Sarah Bentley.
Please note eligibility requirement:
This project would particularly suit either a mathematics/computer science student with an interest in space physics, or a student with a physics background and an interest in deep learning. A numerate background and some physics knowledge are essential. Prior knowledge of programming or deep learning is desirable.
For further details of how to apply, entry requirements and the application form, see
https://www.northumbria.ac.uk/research/postgraduate-research-degrees/how-to-apply/
Please note: Applications that do not include a research proposal of approximately 1,000 words (not a copy of the advert), or that do not include the advert reference (e.g. RDF22/…) will not be considered.
Deadline for applications: 18 February 2022
Start Date: 1 October 2022
Northumbria University takes pride in, and values, the quality and diversity of our staff and students. We welcome applications from all members of the community.
Based on your current searches we recommend the following search filters.
Check out our other PhDs in Newcastle, United Kingdom
Start a New search with our database of over 4,000 PhDs
Based on your current search criteria we thought you might be interested in these.
Analysis for Power System Security by Graphical Deep Learning on Complex Networks
Xi’an Jiaotong-Liverpool University
Deep Learning Based Machine Vision for Space-Informed Applications
Durham University
Bayesian Deep Learning for cosmology with Euclid
Royal Holloway, University of London