【书名】      Probabilistic Inference Using Markov Chain Monte Carlo Methods
【作者】      Radford M. Neal
【出版日期】25 September 1993
【文件格式】PDF【文件大小】2.18M
【页数】       144
【资料类别】论文
【资料方向】概率论 马氏过程 人工智能
【扫描版还是影印版】影印版
【是否缺页】完整
【关键词】Markov Chain , Monte Carlo , Artificial Intelligence ,Probability Inference
【目录】
Contents
1. Introduction 1
2. Probabilistic Inference for Arti cial Intelligence 4
3. Background on the Problem and its Solution 30
4. The Metropolis and Gibbs Sampling Algorithms 47
5. The Dynamical and Hybrid Monte Carlo Methods 70
6. Extensions and Re nements 87
7. Directions for Research 116
8. Annotated Bibliography 121
【原创书评】
作为理工科的学生，MCMC（马氏链蒙特卡洛）  方法在解决许多数值模拟和复杂数据处理问题上有许多独到的功效。本文的作者是多伦多大学计算机系的知名教授，他就依托自己的专业知识介绍了MCMC方法在人工智能方面的使用，并对优劣给出了评价，且给出了改进的意见。当然，这篇文章也是一篇非常好的入门读物。因为文中有三个章节用平实的语言对MCMC方法，统计物理，Gibbs抽样等抽象的方法和理论进行了介绍，非常适合非数学专业的学生阅读。文章也有自己的问题。比如说，对于若干理论，文中没有给出证明就直接使用。虽然这对于应用学科显得无足轻重，可是，却对文章的严谨性产生了质疑。不过这也给很多做概率理论的同学提供了机会，可以从中选取课题。   总之，作为对MCMC方法的介绍性文章以及MCMC方法的应用介绍性文章，本文是可以称的上优秀的！虽然年代有点早，但是，文中提出的很多问题到现在依然没有实质性解决！如果你做大规模数据模拟，如果你做马氏过程，做过你做蒙特卡洛方法，本文不可错过！
【内容简介】
Probabilistic inference is an attractive approach to uncertain reasoning and empirical learning in arti cial intelligence. Computational diculties arise, however, because probabilistic models with the necessary realism and exibility lead to complex distributions over high-dimensional spaces. Related problems in other elds have been tackled using Monte Carlo methods based on sampling using Markov chains, providing a rich array of techniques that can be applied to problems in arti cial intelligence. The \Metropolis algorithm" has been used to solve dicult problems in statistical physics for over forty years, and, in the last few years, the related method of \Gibbs sampling" has been applied to problems of statistical inference. Concurrently, an alternative method for solving problems in statistical physics by means of dynamical simulation has been developed as well, and has recently been uni ed with the Metropolis algorithm to produce the \hybrid Monte Carlo" method. In computer science, Markov chain sampling is the basis of the heuristic optimization technique of \simulated annealing", and has recently been used in randomized algorithms for approximate counting of large sets. In this review, I outline the role of probabilistic inference in arti cial intelligence, present the theory of Markov chains, and describe various Markov chain Monte Carlo algorithms, along with a number of supporting techniques. I try to present a comprehensive picture of the range of methods that have been developed, including techniques from the varied literature that have not yet seen wide application in arti cial intelligence, but which appear relevant. As illustrative examples, I use the problems of probabilistic inference in expert systems, discovery of latent classes from data, and Bayesian learning for neural networks.

[此贴子已经被作者于2008-3-11 23:34:59编辑过]