Information for prospective students

I am interested in improving our ability to solve inference problems. Examples of inference problems are: Given what I see through this telescope, what is the mass of this black hole? What are the underlying patterns that can be found in this stack of images? What processes might have generated this data?

Specifically my research has focussed on developing flexible latent representations for use in probabilistic modelling, and the inference algorithms that go with them. Much of my attention has been on developing novel sampling (Monte Carlo) algorithms to make using complex models practical. For an introduction to both sampling and my teaching style you could see my tutorial lectures or the introduction to my thesis. I am also interested in model representations designed to make inference easier, and application to scientific domains. See my publications for more information.

Application procedure

Please note: I am unable to support internships.

Information about applying and studying for a PhD here is available from the School of Informatics Postgraduate Prospectus, including information on how to make a formal application.

You cannot send an application to me directly, but can contact me to ask any informal questions you may have. Let me know specifically why you are interested in working with me and what you are interested in doing (mass emails will be ignored). Please also include a CV/resume with your email.

The timescale for applications is given on the application page. After certain dates it is hard to get funding, especially for those from outside the EU.

Personal advice

Typically UK PhDs are short, 3 years. This can be an advantage: you get out faster if you want to leave academia, and if you stay as an academic you can afford to spend more time as a postdoc. It can be a rush however. Personally I had 4 years of funding for my PhD and, having come from a different field (Physics), felt that I needed it. This is why ANC generally encourages potential PhD students to first study for an MSc degree. Fortunately Edinburgh Informatics offers an excellent MSc programme (see below).

New from 2014: we also have a Data Science 4-year PhD programme, which starts with a training year with taught courses. Although this particular programme is new, it leans on our long-established MSc programme, and experience from a previous doctoral training programme. This programme has been working really well, and I highly recommend it.

MSc Students

Please investigate the MSc programmes in the School of Informatics, and in particular the Machine Learning specialism. Candidates with a good degree in a numerate discipline should apply here.