AdapskIII, the satellite meeting to MCMC'ski 3

The Canyons, Park City, Utah.              3-4 January, 2011


Workshop on Advances in Monte Carlo organised by


and sponsored by
      


[ Motivations | Invited Speakers | Registration | News | Practicals | Acknowledgements | MCMC'ski ]


Motivation and format

Following an enthustiastic reponse to the earlier editions of the workshop in 2005 and 2008, this workshop is intended to provide an updated snapshot of the methodological and theoretical advances in Monte Carlo methods with an emphasis on adaptive Monte Carlo methods in the broad sense (adaptive MCMC, adaptive population Monte Carlo, and various breeds of adaptive importance sampling amongst others), that is algorithms that attempt to automatically optimise their performance to a given task.

The workshop will consist of 4 half-day sessions on 3rd and 4th January. The whole workshop will be held at The Canyons. There will be breaks on both afternoons in order to allow both informal discussions and relaxation (skiing!).

 Please note that registration is mandatory to attend the workshop.



Invited speakers

The main speakers have confirmed their participation

Yves Atchadé
Radu Craiu

Nando de Freitas
Galin Jones
Faming Liang

Eric Moulines
Jeff Rosenthal
Scott Schmidler
Matti Vihola
Dawn Woodard

in addition the following people have accepted to act as discussants,

Luke Bornn
David Dunson
Heikki Haario
Pierre Jacob
Krzysztof Latuszynski
Jean-Michel Marin
Antonietta Mira
Feng Liang

Programme

Location: The Canyons

Monday 3rd January, 8:30-12:15am: Session 1
  


8:30 Jeff Rosenthal (University of Toronto) Ergodicity of Adaptive MCMC Algorithms  - Slides
Discussant: Antonietta Mira (University of Lugano, Switzerland and University of Insubria, Italy) - Slides
9:30 Eric Moulines (Telecom-Paritech, Paris) Some results on adaptive MCMC Algorithms
Discussant: Krzysztof Latuszynski (University of Warwick) - Slides
10:30-11am: Coffee break
11:00 Radu Craiu (University of Toronto) Recent Advances in Regional Adaptation for MCMC - Slides
Discussant: Jean-Michel Marin (Université Montpeliers 2) - Slides


 

Monday 3rd January, apres-ski (4:30-7pm): Session 2
 

 

4:30-5pm: Coffee break
5:00 Nando de Freitas (University of British Columbia, Vancouver), Bayesian optimization for adaptive MCMC - Slides
Discussant: David Dunson (Duke University)
6:00 Faming Liang (Texas A&M University) An adaptive Monte Carlo Metropolis-Hastings algorithm for Bayesian inference of spatial autologistic models
Discussant: Pierre Jacob (Universite Paris-Dauphine) - Slides





Tuesday 4th January, 8:30-12am: Session 3
  


8:30 Yves Atchadé (University of Michigan) Asymptotic variance estimation for Adaptive Markov Chain Monte Carlo.
Discussant: Luke Bornn (UBC)
9:30 Galin Jones (University of Minnesota) Using regeneration to construct componentwise adaptive MCMC samplers.
Discussant: Heikki Haario (Lappeenrannan University of Technology) - Slides
10:30-11am: Coffee break
11:00 Matti Vihola (University of Jyväskylä) On the stability and convergence of adaptive MCMC - Slides
Discussant: Christian Robert (Université Paris-Dauphine)

 

Tuesday 4th January, apres-ski(4:30-7pm):  Session 4

 

4:30-5pm: Coffee break
5:00 Dawn Woodard (Cornell University) Convergence rate of Markov chain methods for Genomic Motif Discovery
Discussant: Bala Rajaratnam (Stanford) - Slides
6:00 Scott Schmidler (Duke University) Exploration vs. Exploitation in Adaptive MCMC - Slides
Discussant: Feng Liang (University of Illinois at Urbana-Champaign) - Slides

 



Registration

Registration to attend the workshop is mandatory and is available as an option on the main MCMSki meeting registration form. The registration fees of $50 cover the coffee breaks.

Practical details

Directions can be found on the main MCMSki website.

Lodging information can also be found on the main MCMSki website.

[ Motivations  | Registration | News | Practicals | MCMC'ski ]

Last update 13 Dec. 2010