Machine Learning Approaches to Bayesian Inference for Stochastic systems

Statisticians are investigating increasingly complex datasets and models in diverse fields such as astrophysics, infectious disease dynamics, genetics and molecular biology. The Bayesian approach to statistics has been very successful in providing thorough scientific analyses of small systems of this kind, but struggles to scale up to larger problems. On the other hand, machine learning methods, such as neural networks, scale up well to large datasets, but are better suited to producing predictions than answering scientific questions. This project will use machine learning predictions within Bayesian methods to harness the advantages of both methods and provide a new generation of scientific tools.

Successful applicants will also be given the opportunity to complete teaching and demonstrating duties within the school amounting to up to £1500 per annum.