The Dialogue Systems group is part of the Machine Intelligence Lab at the University of Cambridge. Our work focuses on the use of statistical approaches to Spoken Dialogue Systems including understanding, belief tracking, policy optimisation and generation. The principal goal is to design systems that can be trained on real dialogue data, explicitly modelling the uncertainty present in human-machine interaction and providing a natural user experience over a wide range of topic domains. The techniques we employ include traditional Bayesian methods and modern deep learning approaches. Wherever possible, we make our work freely available through open source and open data repositories.