Machine Learning

The University has developed leading-edge machine learning methods, implemented in widely-used open-source computational packages, to build probabilistic predictive models from large-scale datasets in business, engineering, biology and, particularly, health, in partnership with the Manchester-based Health e-Research Centre.

Machine Learning and Robotics

The Machine Learning and Robotics group host a diverse range of research machine learning and cognitive robotics. Working on the full spectrum from foundations to applications, the MLR group prides itself on breadth as well as depth.

Our research spans a wide spectrum of ML theory and applications, and Cognitive Robotics science and engineering.

Fundamentals of statistical machine learning

We have developed modern dimensionality reduction methods, preserving local and global structure relationships. We have contributed to the foundations of information theoretic variable selection, and to state of the art Deep Learning methodologies, applied in image and video understanding.

Medical decision making

We shape the way that machine learning research is conducted to ensure high quality, robustness, reproducibility and fairness in decision making. As a consequence, our work has wide-ranging translational impact from the NHS to the pharmaceutical industry, and from finance to engineering. We aim to deploy state of the art safe predictive models to understand and improve patient treatment in hospitals.

The Cognitive Robotics Lab

We specialise in machine learning models of language learning and cognitive development in humanoid robots. This for example uses the approach of Developmental Robotics, to simulated the gradual acquisition of sensorimotor and cognitive capabilities in robot, taking direct inspiration from developmental psychology theories and mechanisms. The models are used for studies and applications in human-robot interaction, eg: for robot companions for older people, for collaborative robots and for robot tutors for children.