The main aim of my research project is to further the understanding of uncertainty quantification in probabilistic machine learning systems. Specifically, the project is focused on better understanding the validity, and extending the scope, of the practice of decomposing uncertainty into epistemic (lack of information) and aleatoric (inherent variability) components. Such a decomposition is useful in many applied scenarios including active learning, but the principle of the decomposition has recently been called into question. My main research interest is the probabilistic behaviour of machine learning systems, in particular uncertainty quantification. I’m also interested in the legal and policy perspectives on AI, specifically I am interested in structural solutions to the serious ethical questions arising from the bias and fairness (or lack thereof) of AI systems.
I am conducting research here at the STAI CDT because of the aforementioned ethical shortcomings of current AI systems. The CDT is a strong and friendly community of research students and staff who are all motivated by ensuring AI is a genuine force for good – it is a pleasure and privilege to work with these like minded individuals on a daily basis.
–
Masters Qualification: MSc Scientific Computing and Data Analysis, University of Durham
Undergraduate Qualification: BSc Mathematics, University of Durham
Work Experience: Management consultant, Arup
