Recent advances in Monte Carlo based inference

Workshop theme

Monte Carlo methods, particular Markov Chain Monte Carlo (MCMC), are now the methods of choice for making inferences about complex stochastic systems. Whilst MCMC dates back over 50 years, and there has been extensive research in Monte Carlo methods over the past 20 years, there are still many challenges that face researchers today.

These include attempting to analyse the highly complicated stochastic models and large scientific data sets that are now commonplace; and trying to understand the theoretical properties of some of the novel ideas that are proposed.

Currently, Monte Carlo methods are used by researchers in numerous scientific fields, including statistics, physics, engineering, genetics, econometrics, bioinformatics, and machine learning. This interdisciplinary workshop will bring together researchers from a variety of such fields to discuss current and novel Monte Carlo methodology, and to cross-fertilise ideas across these different disciplines. The workshop will have a broad focus, covering both recent advances in more established methods such as MCMC and sequential Monte Carlo, together with more recent ideas and ideas that have had little exposure within the statistics community, such as Variational methods, Population Monte Carlo, Approximate Bayesian Computing, Quasi Monte Carlo, and Indirect Inference.

The application is closed

Timetable

Monday 30th October 2006
08:30 to 10:00	Registration	Room 1
10:00 to 10:05	Chris Wallace Welcome address	Room 1
10:05 to 11:00	Now you know your ABCs: examples and problems	Room 1
11:00 to 11:30	Coffee	No Room Required
11:30 to 12:30	Tony Pettitt Queensland University of Technology From doubly intractable distributions via auxiliary variables to likelihood free inference	Room 1
12:30 to 13:30	Lunch at Churchill College	No Room Required
14:00 to 15:15	Some developments of ABC	Room 1
15:15 to 15:45	Tea	No Room Required
15:45 to 17:00	Arnoldo Frigessi University of Oslo Estimating functions in indirect inference	Room 1
17:00 to 18:00	Welcome Wine Reception	No Room Required
18:15 to 18:45	Dinner at New Hall (Residents Only)	No Room Required

Tuesday 31st October 2006
09:00 to 10:00	Population-based MC for sampling trans-dimensional Bayesian regression models	Room 1
10:00 to 11:00	Sequentially interacting Markov Chain Monte Carlo	Room 1
11:00 to 11:30	Coffee	No Room Required
11:30 to 12:30	Olivier Cappe Adaptive population Monte Carlo	Room 1
12:30 to 13:30	Lunch at Churchill College	No Room Required
14:00 to 15:15	Deterministic alternatives to MCMC	Room 1
15:15 to 15:45	Tea	No Room Required
15:45 to 17:00	Poster session	Discussion Room
18:15 to 18:45	Dinner at New Hall (Residents Only)	No Room Required

Wednesday 1st November 2006
09:00 to 10:00	Yukito Iba The Institute of Statistical Mathematics (ISM) Applications of extended ensemble Monte Carlo	Room 1
10:00 to 11:00	David Leslie University of Bristol Sequential Monte Carlo for Generalized Linear Mixed Models	Room 1
11:00 to 11:30	Coffee	No Room Required
11:30 to 12:30	Toby Johnson University of Leeds A sequential importance sampler for reconstructing genetic pedigrees	Room 1
12:30 to 13:30	Lunch at Churchill College	No Room Required
14:00 to 15:15	Retrospective sampling	Room 1
15:15 to 15:45	Tea	No Room Required
15:45 to 17:00	The expected auxiliary variable method for Monte Carlo simulation	Room 1
18:15 to 18:45	Dinner at New Hall (Residents Only)	No Room Required

Thursday 2nd November 2006
09:00 to 10:00	Extensions of the CE method for statistical analysis	Room 1
10:00 to 11:00	Nested sampling	Room 1
11:00 to 11:30	Coffee	No Room Required
11:30 to 12:30	Sequential Monte Carlo methods: can we replace the resampling step?	Room 1
12:30 to 13:30	Lunch at Churchill College	No Room Required
14:00 to 15:15	Importance sampling for diffusion processes	Room 1
15:15 to 15:45	Tea	No Room Required
15:45 to 17:00	Bayesian inference for nonlinear multivariate diffusion processes	Room 1
18:00 to 19:30	CUP Bookshop Reception	No Room Required
20:00 to 18:00	Conference Dinner at Peterhouse (Dining Hall)	No Room Required

Friday 3rd November 2006
09:00 to 10:00	Configurationally-Biased MC and Virtual-move parallel tempering	Room 1
10:00 to 11:00	Peter Green University of Bristol Branching process Monte Carlo	Room 1
11:00 to 11:30	Coffee	No Room Required
12:30 to 13:30	Lunch at Churchill College	No Room Required
14:00 to 15:15	David Wales University of Cambridge Sampling the energy landscape: thermodynamics and rates	Room 1
15:15 to 15:45	Tea	No Room Required
15:45 to 17:00	Pierre L'Ecuyer Université de Montréal Randomized quasi-Monte Carlo for Markov Chains	Room 1
18:15 to 18:45	Dinner at New Hall (Residents Only)	No Room Required

Speakers

Olivier Cappe
Arnoldo Frigessi University of Oslo
Peter Green University of Bristol
Yukito Iba The Institute of Statistical Mathematics (ISM)
Toby Johnson University of Leeds
Pierre L'Ecuyer Université de Montréal
David Leslie University of Bristol
Tony Pettitt Queensland University of Technology
David Wales University of Cambridge
Chris Wallace

Recent advances in Monte Carlo based inference

Workshop theme

The application is closed

Timetable

Speakers

Supported By

Register for updates