Professor Christian Robert from the Department of Applied Mathematics at the Université Paris-Dauphine will be giving a public lecture and a technical lecture as part of AMSI-SSAI-QUT lecture series.
Professor Christian P. Robert’s research areas cover Bayesian statistics, with a focus on decision theory and model selection, numerical probability, with works cantering on the application of Markov chain theory to simulation, and computational statistics, developing and evaluating new methodologies for the analysis of statistical models. He has written or co-written eight books on Bayesian statistics and computational methods, as well as over 150 research papers in these areas and their applications.
When: 21st August 2012
Where: QUT Garden’s Point Gibson Room (Z block level 10, room 1064)
Public Lecture: Simulation as a universal tool for statistics
Statistics is a child of Mathematics in that it relies on mathematical models and uses methods validated by mathematical theorems. It is also an offspring of Computer Science in that it cannot produce answers on realistic problems without the help of advanced computational devices. Computer simulation, that is the computer reproduction of randomness, is presumably the most ubiquitous of these tools, with the additional appeal of being validated by probability theory.
In this talk, I will introduce the basics of computer-based simulation like uniform simulation and accept-reject algorithms, then present some more advanced methods like simulated annealing, Markov chain Monte Carlo methods, and likelihood-free methods, with illustrations ranging from Sudokus to cosmological background noise (CMB) to ancestral trees of Pygmy tribes.
This is a public lecture, suitable for those without any technical background.
Technical Talk: “ABC methods for Bayesian model choice”
Approximate Bayesian computation (ABC) have become a essential tool for the analysis of complex stochastic models. Having implemented ABC-based model choice in a wide range of phylogenetic models in the DIY-ABC software (Cornuet et al., 2008), we first present theoretical background as to why a generic use of ABC for model choice is ungrounded, since it depends on an unknown amount of information loss induced by the use of a summary statistic (Robert et al., 2011). We then present necessary and sufficient conditions on the summary statistics for ABC based model choice procedure to be consistent, a solution that avoids the call to additional empirical verifications of the performances of the ABC procedure as those available in DIYABC and advocated in Ratman et al. (2011).
Open forum 3:00-5:00pm
More information can be found at www.amsi.org.au/robert.php