I am working with Prof. Lyndsay Fletcher on the implementation of machine learning algorithms in solar observations. My main focus is on flare spectropolarimetry and how machine learning techniques can aid the data analysis process for carrying out chromospheric magnetic field diagnostics in a flaring atmosphere.
Currently, I am developing a generative adversarial network (GAN) to correct for atmospheric seeing in ground-based flare observations. This is when we set up a generator to generate images corrected from seeing and a discriminator to decide whether or not these images are real data or generated. The goal of this is for the two to train each other into a state wherein the generator can produce images corrected for seeing that are perceptually convincing whilst still preserving spectral integrity.
I am also interested in a solar ImageNet development which would aid transfer learning and initialisation for a wide variety of solar machine learning problems in the future.
A C.V. can be found here.
School of Physics and Astronomy
University of Glasgow
Tel: +44 141 330 2960