15.00 - 16.00 Simon Tavaré (Cambridge)
 Approximate Bayesian Computation for modeling DNA replication dynamics

Department of Applied Mathematics and Theoretical Physics, University of Cambridge
Approximate Bayesian Computation (ABC) provides a useful approach to inference in complex scientific problems where likelihoods are hard or impossible to compute. From its origins in the mid-90s, ABC has become an established part of a Bayesian’s toolkit. I will give a brief overview of the principles behind ABC, and discuss some of its pros and cons. To illustrate the ideas, I will describe some recent work in my lab that studies the dynamics of DNA replication in yeast, using genome-wide microarray and next-generation sequencing data that provide the highest fidelity map of DNA replication to date.

16.00 - 16.30 Tea, Coffee & Biscuits (Main Atrium)

16.30 - 17.30 Simon French (Warwick)
Aggregating expert judgement in major societal decisions

Society and its systems are becoming more and more complex and that means that risk analyses are having to rely more and more on expert judgement rather than data, simply for the reason that there are insufficient data available. But experts seldom agree, and even when they do there there is no guarantee that they are right. How should their judgements be used in large scale risk analyses. Moreover, how might they be re-used in subsequent analyses on similar but not identical issues?
In this talk we will survey the theory of aggregating expert judgement which has been developed over the past 30-40 years, noting that there are a variety of contexts in which their aggregated judgement may be needed and that different methods may be need for different contexts. We shall also note the lack of a theory of 'Meta-Analysis of Expert Judgement Studies' and that one is greatly needed.

 17.30 - 18.00 Wine & Cheese (Main Atrium)