Bob jointed the Department of Statistics as a lecturer in July 2013, following a short stint as a post-doctoral researcher at the University of Birmingham, UK. He has a BSc(Hons) with first-class honours in Mathematical Sciences, obtained through the Open University while working in industry, an MSc with distinction in Natural Computation from the University of Birmingham for which he also received the 'best student' award for his year, and a PhD in Computer Science also from the University of Birmingham. His doctoral research concerned classification (discriminant analysis) and was focused on developing theory quantifying the effect of 'random projection' - a recent and very promising non-adaptive dimensionality reduction technique - on classifier performance.
Machine Learning and Statistical Pattern Recognition.
Dimensionality Reduction and Feature Selection.
Learning Theory and Mathematical Statistics, especially non-asymptotic (finite sample) theory.
Random Matrix Theory.
Applications of all of the above. In particular using theory to better understand existing techniques, and to develop efficient, effective, and principled methods for big data settings with performance guarantees.
Durrant, R.J. and Kaban, A. (2013) Sharp Generalization Error Bounds for Randomly-projected Classifiers Proc. 30th International Conference on Machine Learning; JMLR W&CP 28(3):693-701
Durrant, R.J., and Kaban, A. (2013) Random Projections as Regularizers: Learning a Linear Discriminant from Fewer Observations than Dimensions to appearProceedings 5th Asian Conference on Machine Learning
Ed. Vols. of Conference Proceedings
Durrant, R.J. and Kaban, A. (2012) Error Bounds for Kernel Fisher Linear Discriminant in Gaussian Hilbert Space Proc 15th International Conference on Artificial Intelligence and Statistics (AIStats 2012). JMLR W&CP 22: pp337-345.