Simon Shaw Department of Mathematical Sciences University of Durham, Science Laboratories, South Road, Durham, DH1 3LE UNITED KINGDOM Tel: +44-(0)191-374-7781 (direct) Fax: +44-(0)191-374-7388 E-mail: S.C.Shaw@durham.ac.uk

Biography:

I'm originally from Nettleham, Lincoln, arriving as an undergraduate at Durham in 1993.  In 1996 I obtained my BSc Mathematics with First Class Honours and in 2000 obtained my PhD Statistics, having been supervised by Michael Goldstein . I am currently a Senior Research Associate in the third, and final, year of an EPSRC funded project entitled `High Reliability Testing for Complex Software Systems using generalised Bayesian graphical modelling and program comprehension analysis'.  This work is a collaboration between members of the Statistics department and members of the Computer Science department.

Research interests:

Principally, I'm interested in Bayes (linear) methods and Bayesian methods for software testing. In Bayes linear methods, expectation rather than probability is taken as primitive for expressing our beliefs about random quantities of interest. The Bayes linear methodology offers a framework for the adjustment of beliefs given only a partial prior specification. Within this realm, my interests lie in the analysis of  collections of (second-order) exchangeable sequences, where the sequences may be infinite or finite; uses of (generalised) conditional independence and graphical models. The Bayes linear methods homepage may be of interest. Software testing is ideally suited to a Bayesian formulation for we may consider it to be the process of reducing uncertainty about software reliability through careful choice of test-suite. The process is modelled using a Bayesian graphical model. I am interested in the development of graphical models that combine differing levels of prior specification, for example fully specified data inputs with only a Bayes linear specification made over the parameters. I'm also interested in program comprehension, the means of understanding the structure and inter-relationships of a piece of software. The homepage of the Durham software testing group is worth a visit.

Descriptions (and downloadable documentation) of my research work may be found be visiting the following three sections:

• generalised Bayesian graphical modelling (development of graphical models containing both full Bayes and Bayes linear prior specifications)
• program comprehension methodology
• multivariate partially exchangeable systems (PhD work, including downloadable thesis)