Advanced Monte Carlo Methods for Complex Inference Problems

In recent years there has been an explosion of complex data-sets in areas as diverse as Bioinformatics, Ecology, Epidemiology, Finance and Population genetics. In a wide variety of these applications, the stochastic models devised to realistically represent the data generating processes are very high dimensional and the only computationally feasible and accurate way to perform statistical inference is with Monte Carlo.

The focus of this programme is on recent innovations in the field of Monte Carlo methods for inference in complex and intractable statistical problems. This programme will bring together researchers from a broad base, for the first time since 2009, to promote discussion and development of this important and rapidly advancing cross-disciplinary area. It will leverage on the two very successful past programmes which were the INI Programme on Stochastic Computation in the Biological Sciences (23 October - 15 December 2006) and the SAMSI programme on Sequential Monte Carlo (SMC) Methods (September 2008 to August 2009), by taking up the following research threads that have genuinely enthused the wider community of research and applied statisticians over the past couple of years: Approximate Bayesian Computation; SMC and Markov Chain Monte Carlo and their integration; and recent theoretical advancements underpinning these areas.

This programme will also hold a workshop in the first week covering the two major themes of this proposal to launch the 4-week programme. The workshop will serve as a catalyst for the remaining 3 weeks of intensive research and aims to cover the following specific areas:

ABC: new applications, methodology and theory SMC/MCMC for high dimensional computation The workshop will also have an introductory element to it, aimed at acquainting postgraduate students and postdoctoral researchers with the subject area. Details to follow soon.