of adaptive MCMC methods
Summary Adaptive Markov
Chain Monte Carlo (MCMC) methods are currently a very active field of
research. MCMC methods are sampling methods, based on Markov Chains
which are ergodic with respect to the target probability measure. The
principle of adaptive methods is to optimize on the fly some design
parameters of the algorithm with respect to a given criterion
reflecting the sampler's performance (optimize the acceptance rate,
optimize an importance sampling function, etc...).
postdoctoral position is opened to work on the numerical analysis of
adaptive MCMC methods: convergence, numerical efficiency, development
and analysis of new algorithms. A particular emphasis will be
given to applications in statistics and molecular dynamics. (Detailed
Position funded by the French
National Research Agency (ANR) through the 2009-2012 project
Required diploma PhD thesis in
statistics orprobability, with a competitive track record.
skills experience in MCMC methods and their mathematical
Deadline for applications
: September 2010.
Applications must include :
a detailed CV with a description
of realized projects
and must be sent to Gersende FORT (firstname.lastname@example.org) in pdf
format; or by standard mail to : Gersende FORT (LTCI,
46 rue Barrault, 75 634 Paris Cedex 13, Paris, France).
a motivation letter
a summary of the thesis
2 or 3 recommendation letters
preferred starting dates and duration
Duration : 12 months.
: Paris. The position will benefit from an
interdisciplinary environment involving numerical analysts,
statisticians and probabilists, and of strong interactions between the
partners of the project ANR-08-BLAN-0218
Meeting of the project : December, 4 2008, at
Seminar : January, 29 ; March, 12 ; April, 30 ;
June, 4 ; June, 25 ; October, 22; November 26 : at IHP see the
webpage of the
Meeting of the project : September,
Seminar : January, 28; February, 18; March, 25;
April, 15; May, 27; Juin, 10; Ocotber 7; November 18 : at
see the webpage of the
Seminar : January, 6; February, 3; March, 3;
April, 7; May, 5; June, 9; September, 1st; October,
13; November, 15; December, 8 : at IHP, see the webpage of
Meeting of the project : October, 13.
: every month, at IHP. [see
Course : "Méthodes de
Monte Carlo par Chaînes de Markov adaptatives" organized by
ENPC &Univ. Paris-Est. Invited
: Y. ATCHADE (Univ. Michigan, USA), June 2010, Paris, France.
3rd Conference on numerical
methods in finance, organized by ENPC, April 15-17
2009, Paris, France.
Satellite meeting "Adaptive Monte
Carlo methods" of the MCMCski conference, organized by C.P.
(Univ. Dauphine), Jan 3-4 2011,
Special session "Monte
Carlo methods for Bayesian inverse problems " of the
ASMDA 2011 conference, organized by G. FORT (LTCI), June 2011, Roma,
meeting of the conference
"Méthodes particulaires pour les modèles de diffusion",
organized by C.P. ROBERT (Univ. Dauphine), July 2011, Barcelona,
intractability in Statistical Inference"
16-19 2012 Bristol, UK. With invited speakers C.P. ROBERT (Univ.
Dauphine) and N. CHOPIN (CREST).
Workshop "Advances in Markov
chain Monte Carlo", April 23-25 2012, Edinburgh, UK.
Organized by C.P. ROBERT (Univ. Dauphine).
Conference ISBA 2012,
June 2012, Kyoto, Japan. Invited sessions and Special topic sessions
organized by J. ROUSSEAU (Univ. Dauphine) and C.PO. ROBERT (Univ.
Y. Atchadé, G. Fort, E.
Moulines, P. Priouret. Adaptive MCMC
: theory and methods, submitted.
J.M. Cornuet, J.M. Marin, A. Mira, C.P.
Robert. Adaptive Multiple
Importance Sampling, ArXiv:0907.1254
R. Douc, C.P. Robert. A vanilla
Rao-Blackwellisation of Metropolis-Hastings algorithms, ArXiv:0904.2144
B. Jourdain, J. Lelong. Robust
adaptive Importance Sampling for Normal Random vectors, To
appear in Ann. Appl. Prob., [preprint]
B. Jourdain, T. Lelièvre, R. Roux. Existence, uniqueness and convergence of a
particle approximation for the Adaptive Biasing Force process; ArXiv:0903.4518