Skip to main content Skip to navigation


Term 1      
date speakers title/paper comments
Fri, Oct 8, 2010 Krys M. Girolami and B. Calderhead
"Riemann manifold Langevin and Hamiltonian Monte Carlo methods"
link to the paper [pdf]
Fri, Oct 15, 2010 Krys Girolami & Calderhead continued  
Fri, Oct 22, 2010 Murray Introduction to particle filters notes [pdf] (by Murray Pollock)
Fri, Nov 6, 2010 Daniel K. Kalogeropoulos, G. Roberts, P Delaportas
"Inference for Stochastic Volatility Models using Time Change Transformations"
Ann. Stat. 2010, 38(2), 784-807
link to the paper [pdf]
Fri, Nov 13, 2010
Krys O. Stramer, G. Roberts "On inference for partially observed nonlinear diffusion models using the Metropolis-Hastings algorithm".Biometrika 88 (2001), no. 3, 603–621 link to thearticle in journal
Fri, Nov 20, 2010
Krys K. Kalogeropoulos, G. Roberts, P Delaportas
"Inference for Stochastic Volatility Models using Time Change Transformations"
Ann. Stat. 2010, 38(2), 784-807
link to the paper [pdf]
Fri, Dec 3, 2010 Christiane  Golightly, A; Wilkinson, D. J.
"Bayesian inference for nonlinear multivariate diffusion models observed with error."
Comput. Statist. Data Anal. 52 (2008), no. 3, 1674–1693.
link to the paper [pdf]
room change:seminar in the Statistics Common Room 
Term 2      
Fri, Jan 14, 2011 Alex Roberts, G. O.; Gelman, A.; Gilks, W. R.
"Weak convergence and optimal scaling of random walk Metropolis algorithms"
Ann. Appl. Probab. 7 (1997), no. 1, 110–120.
link to the paper [pdf]
Fri, Jan 21, 2011 Kasia Exact AlgorithmsAlexandros Beskos, Omiros Papaspiliopoulos, Gareth O. Roberts
"A factorisation of diffusion measure and finite sample path constructions" 881061678851u258/Alexandros Beskos, Omiros Papaspiliopoulos, Gareth O. Roberts, Paul Fearnhead
"Exact and computationally efficient likelihood-based estimation for discretely observed diffusion processes"
http://onlinelibrary.wiley. com/doi/10.1111/j.1467-9868. 2006.00552.x/full
room change: A1.01 
Fri, Jan 28, 2011 Dalia Inverting 2-D photos into densities in 3-D

The pursuit of the intrinsic material density in $\mathbb[R}^3$, given the available two dimensional image of the system, is commonly encountered in different areas of applied physics and is an integral step in astronomical modelling. The problem will be briefly discussed within the paradigm of integral equations. In the presence of density heterogeneities, this inverse problem is, however, ill-posed, unless extra information - in terms of measurements in addition to a single image - are made available. Such measurements sometimes constitute imaging the system at different viewing angles or when multiple viewing is impossible (as in astronomy), ancillary data are invoked. Modelling of such data (Paneratos 2009, Chakrabarty 2008,2010), will be discussed. Discussion of a new non-invasive methodology that attempts the estimation of the multi-modal 3-D density of material samples imaged by electron scattering techniques, without resorting to multiple viewing, will be included.
Fri, Feb 4, 2011 Adam Johansen Sequential Monte Carlo without State Space ModelsThe task of sampling from from a sequence of distributions arises naturally in many context within statistics and related fields. In some settings in which a single distribution is of interest it can also be useful to construct an artificial sequence which forms a bridge from a simple distribution to a more complicated distribution of interest. Sequential Monte Carlo provides a convenient framework in which importance sampling and resampling can be used to provide properly weighted samples from each distribution within a sequence. Some illustrative examples will be discussed if time permits.The ideas presented will focus on:
"Sequential Monte Carlo Samplers", P. Del Moral, A. Doucet & A. Jasra, J. Royal Statist. Soc. /B, vol. 68, no. 3, pp. 411-436, 2006. Pdf
< 7Earnaud/delmoral_doucet_ jasra_ sequentialmontecarlosamplersJR SSB.pdf>"Sequential Monte Carlo for Bayesian Computation" with discussion, P. Del Moral, A. Doucet & A. Jasra, Bayesian Statistics 8, Oxford University Press, 2006. Pdf draft version
< 7Earnaud/delmoral_doucet_ jasra_ sequentialmontecarloforbayesia ncomputation_ valenciaalmostfinal.pdf>
Fri, Feb 18, 2011 Chris Jewell MCMC in Practice: Bayesian computation for the real world  
Fri, Feb 25, 2011 Krys Markov chain CLTs and asymptotic confidence intervals for MCMC estimation.

I will show two approaches to the Markov chains CLT proof, one by the Poisson equation and martingale approximation and one by regeneration. Only the the most regular cases will be presented. Then I will briefly discuss CLT based asymptotic confidence intervals in MCMC estimation.

related papers are:
- Roberts, Gareth O.; Rosenthal, Jeffrey S. General state space Markov chains and MCMC algorithms. Probab. Surv. 1 (2004), 20-71
- Hobert, James P.; Jones, Galin L.; Presnell, Brett; Rosenthal, Jeffrey S. On the applicability of regenerative simulation in Markov chain Monte Carlo. Biometrika 89 (2002), no. 4, 731-743.
- Jones, Galin L.; Haran, Murali; Caffo, Brian S.; Neath, Ronald Fixed-width output analysis for Markov chain Monte Carlo. J. Amer. Statist. Assoc. 101 (2006), no. 476, 1537-1547.
- Häggström, Olle; Rosenthal, Jeffrey S. On variance conditions for Markov chain CLTs. Electron. Comm. Probab. 12 (2007), 454-464
- Bednorz, Witold; Łatuszyński, Krzysztof; Latała, Rafał A regeneration proof of the central limit theorem for uniformly ergodic Markovchains. Electron. Commun. Probab. 13 (2008), 85-98.
room change: B3.03 (Maths)
Fri, March 4, 2011 Kasia Roberts, G. & Tweedie, R.
Exponential convergence of Langevin distributions and their discrete approximations
Bernoulli, 1996, 2, 341-363 
Fri, March 11, 2011   seminar cancelled-> see instead AMSTAT seminar poster session starting from 10am in "the street" - atrium where the main entrance to the maths part of the building is.  
Fri, March 18, 2011 Alex Joulin, A. & Ollivier, Y.
Curvature, concentration, and error estimates for Markov chain Monte Carlo
The Annals of Probability, 2010, 38, 2418-2442 
Fri, March 25, 2011 Giorgos Sermaidis (Lancaster) MCMC for exact inference for diffusions   
Fri, April 1, 2011 Peter Windridge Mixing times, cutoff and critical slowdown of certain MCMC algorithms for the Ising model.The focus will be Lubetzky/Sly's recent result 
Fri, April 8, 2011 Peter Windridge will continue  
Term 3      
Fri, May 6, 2011      
Fri, May 13, 2011 Sebastian Hairer, M.; Stuart, A.; Voss, J. & Wiberg, P.
Analysis of SPDEs arising in path sampling. Part I: The Gaussian case
Communications in Mathematical Sciences, 2005, 3, 587-603, M.; Stuart, A. & Voss, J.
Analysis of SPDEs arising in path sampling part II: The nonlinear case
The Annals of Applied Probability, 2007, 17, 1657-1706 
2 hours session 
Fri, May 27, 2011 Alberto Kantas, N.; Doucet, A.; Singh, S. & Maciejowski, J.
An overview of sequential Monte Carlo methods for parameter estimation in general state-space models 
Fri, June 10, 2011      
Fri, June 17, 2011 Sebastian Non-asymptotic mixing of the MALA algorithm 
Martin Hairer, N. Bou-Rabee and E. Vanden-Eijnden. 
???, ???, 2011 Alberto continues on
Kantas, N.; Doucet, A.; Singh, S. & Maciejowski, J.
An overview of sequential Monte Carlo methods for parameter estimation in general state-space models