## Advances in Markov chain (2007)

### BibTeX

@MISC{Methods07advancesin,

author = {Monte Carlo Methods and Iain Murray},

title = {Advances in Markov chain},

year = {2007}

}

### OpenURL

### Abstract

I, Iain Murray, confirm that the work presented in this thesis is my own. Where information has been derived from other sources, I confirm that this has been indicated in the thesis. 3 Probability distributions over many variables occur frequently in Bayesian inference, statistical physics and simulation studies. Samples from distributions give insight into their typical behavior and can allow approximation of any quantity of interest, such as expectations or normalizing constants. Markov chain Monte Carlo (MCMC), introduced by Metropolis et al. (1953), allows sampling from distributions with intractable normalization, and remains one of most important tools for approximate computation with probability distributions. While not needed by MCMC, normalizers are key quantities: in Bayesian statistics marginal likelihoods are needed for model comparison; in statistical physics many physical quantities relate to the partition function. In this thesis we propose and investigate several new Monte Carlo algorithms, both for evaluating normalizing constants and for

### Citations

431 | A learning algorithm for Boltzmann machines
- Ackley, Hinton, et al.
- 1985
(Show Context)
Citation Context ...ts assume them.Approximation Schemes 128 The usual algorithm for learning BMs is a maximum likelihood version of the EM algorithm (assuming some of the variables are hidden sH and some observed sO) (=-=Ackley et al., 1985-=-). The gradient of the log probability is: ∂ log p(sO|W ) ∂Wij = Ec[sisj] − Eu[sisj] , (5.7) where Ec[·] denotes expectation under the “clamped” data distribution p(sH|sO, W ) and Eu[·] denotes expect... |

150 | Rao-bleickwellisation of Sampling Schemes - Casella, Robert - 1996 |

133 | Approximate Bayesian Inference with the Weighted Likelihood Bootstrap - Newton, Raftery - 1994 |

39 | Multicanonical ensemble: A new approach to simulate first-order phase transitions - Berg, Neuhaus - 1992 |

38 | Bayesian Training of Backpropagation Networks by the Hybrid Monte Carlo Method - Neal - 1992 |

37 | Monte Carlo Methods. Volume I: Basics - Kalos, Whitlock - 1986 |

32 | Exact sampling and approximate counting techniques - Huber - 1999 |

15 |
Erroneous results in “Marginal likelihood from the Gibbs output
- Neal
- 1999
(Show Context)
Citation Context ...he expense of an obvious increase in memory requirements and a harder-to-gauge cost for bringing a larger system to equilibrium. 2.5.3 Annealed importance sampling (AIS) Annealed importance sampling (=-=Neal, 1998-=-a, 2001) looks more like the original simulated annealing for optimization than the above ensemble methods. An initial point x0 is drawn from p0 a vague ‘high temperature’ base distribution. Further p... |

11 | A fast procedure for model search in multidimensional contingency tables - Havránek - 1987 |

3 | Probability Theory: The Logic of Science. Principles and Elementary Applications - Jaynes - 2003 |

2 | editors, Bayesian statistics 7 - Bayarri, Dawid, et al. - 2003 |

2 | Out of print. Available from http://cg.scs.carleton.ca/ ∼luc/rnbookindex.html. (page 22 - ISBN |

2 | Equation of state calculation by fast computing machines - Møller, Pettitt, et al. - 2004 |

2 |
pages 68 and 69
- Neal
- 1998
(Show Context)
Citation Context ...he expense of an obvious increase in memory requirements and a harder-to-gauge cost for bringing a larger system to equilibrium. 2.5.3 Annealed importance sampling (AIS) Annealed importance sampling (=-=Neal, 1998-=-a, 2001) looks more like the original simulated annealing for optimization than the above ensemble methods. An initial point x0 is drawn from p0 a vague ‘high temperature’ base distribution. Further p... |

2 | Facilitating the Gibbs sampler: the Gibbs stopper and the griddy-Gibbs sampler - Smith, West, et al. - 2007 |

2 | editors, Bayesian Statistics 8 - Bayarri, Dawid, et al. - 2007 |

1 | An essay towards solving a problem in the doctrine of chances. Philosophical Transations, 53:370–418, 1763. Communicated by Richard Price, in a letter to John Canton. Also available edited by G. A. Barnard in Biometrika 45(3/4):293–315 - Press, Diego - 1958 |

1 | Dependency networks for inference, collaborative filtering, and data visualization - Biometrika - 1970 |

1 | pages 119 and 143 - Newman, Ziff - 2005 |

1 | Carlo methods in statistical mechanics: foundations and new algorithms - Wang - 1995 |

1 | pages 36 and 37 - Letters - 1987 |