My main research interests are in the fields of non-negative matrix factorisation (NMF) and neural networks. NMF is primarily a dimensionality reduction technique which is crucial in reducing the complexity of high-dimensional data. NMF has the additional benefit of tending to, naturally, lead to a sparse and parts based representation of the data, which may be significantly more interpretable than more holistic approaches such as principal component analysis. Neural networks in various forms have, in recent years, considerably improved upon the state of the art in various machine learning fields. These modern neural networks can produce excellent results but not the reasoning behind them. Improving the interpretabiliy of neural networks may lead us to produce better systems in the future and be able to trust the results provided.
MSc Computer Science with distinction, University of Southampton
MPhys Physics, First class, University of Manchester
Prospects in Datascience 2016
Advances in Datascience 2017