Machine learning is concerned with the design of computer programs that can improve their performance by learning from experience.  This often takes the form of algorithms for supervised statistical pattern recognition, regression or unsupervised clustering, and so is closely related to the field of data mining.  Unlike data mining, however the focus is on optimising performance (often with theoretical bounds on generalisation), rather than exploration of the data.  Machine learning also has strong links with statistics, with many approaches based on Bayesian principles, or results from computational learning theory.  Unlike statistics, however, there is often greater interest in situations where the underlying assumptions of statistics are violated, for instance learning from non i.i.d. (indentically and independently distributied) data, learning under covariate shift and semi-supervised or transductive learning. 

Our main areas of research at present include:

• Kernel learning methods
• Model selection
• Robust performance evaluation
• Bayesian modelling
• Approximate inference for Gaussian process classifiers
• Semi-supervised learning

 Applications include:

• Analysis and classification of gene microarray data
• Detecting natively disordered regions in proteins
• Modelling the growth of microbial populations
• Predicting gene regulation from sequence data
• Predictive Modelling of Bone Ageing
• Statistical downscaling in climatology
• Time Series Data Mining Electricity Usage Patterns
• Machine learning in immunology, to predict peptide binding affinity to MHC complex molecules