返回
probabilistic inference using markov chain monte carlo methods - r. neal

probabilistic inference using markov chain monte carlo methods - r. neal

ID：14315922

大小：1017.62 KB

页数：144页

时间：2018-07-27

probabilistic inference using markov chain monte carlo methods - r. neal_第1页
probabilistic inference using markov chain monte carlo methods - r. neal_第2页
probabilistic inference using markov chain monte carlo methods - r. neal_第3页
probabilistic inference using markov chain monte carlo methods - r. neal_第4页
probabilistic inference using markov chain monte carlo methods - r. neal_第5页
资源描述：

《probabilistic inference using markov chain monte carlo methods - r. neal》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、ProbabilisticInferenceUsingMarkovChainMonteCarloMethodsRadfordM.NealTechnicalReportCRG-TR-93-1DepartmentofComputerScienceUniversityofTorontoE-mail:radford@cs.toronto.edu25September1993cCopyright1993byRadfordM.NealAbstractProbabilisticinferenceisanattractiveapproachtouncertainreas

2、oningandem-piricallearninginarti cialintelligence.Computationaldicultiesarise,however,becauseprobabilisticmodelswiththenecessaryrealismand exibilityleadtocom-plexdistributionsoverhigh-dimensionalspaces.Relatedproblemsinother eldshavebeentackledusingMonteCarlomethodsbasedonsampli

3、ngusingMarkovchains,providingaricharrayoftechniquesthatcanbeappliedtoproblemsinarti cialintelligence.TheMetropolisalgorithm"hasbeenusedtosolvedicultproblemsinstatisticalphysicsforoverfortyyears,and,inthelastfewyears,therelatedmethodofGibbssampling"hasbeenappliedtoproblemsofsta

4、tisticalinference.Concurrently,analternativemethodforsolvingproblemsinstatisticalphysicsbymeansofdynamicalsimulationhasbeendevelopedaswell,andhasrecentlybeenuni edwiththeMetropolisalgorithmtoproducethehybridMonteCarlo"method.Incomputerscience,Markovchainsamplingisthebasisofthehe

5、uristicoptimizationtechniqueofsimulatedannealing",andhasrecentlybeenusedinrandomizedalgorithmsforapproximatecountingoflargesets.Inthisreview,Ioutlinetheroleofprobabilisticinferenceinarti cialintelligence,presentthetheoryofMarkovchains,anddescribevariousMarkovchainMonteCarloalgor

6、ithms,alongwithanumberofsupportingtechniques.Itrytopresentacomprehensivepictureoftherangeofmethodsthathavebeendeveloped,includingtechniquesfromthevariedliteraturethathavenotyetseenwideapplicationinarti cialintelligence,butwhichappearrelevant.Asillustrativeexamples,Iusetheproblems

7、ofprobabilisticinferenceinexpertsystems,discoveryoflatentclassesfromdata,andBayesianlearningforneuralnetworks.AcknowledgementsIthankDavidMacKay,RichardMann,ChrisWilliams,andthemembersofmyPh.Dcommittee,Geo reyHinton,RudiMathon,DemetriTerzopoulos,andRobTibshirani,fortheirhelpfulcom

8、mentsonthisreview.Thisworkwassupportedby

当前文档最多预览五页，下载文档查看全文

此文档下载收益归作者所有

当前文档最多预览五页，下载文档查看全文
温馨提示：
1. 部分包含数学公式或PPT动画的文件，查看预览时可能会显示错乱或异常，文件下载后无此问题，请放心下载。
2. 本文档由用户上传，版权归属用户，天天文库负责整理代发布。如果您对本文档版权有争议请及时联系客服。
3. 下载前请仔细阅读文档内容，确认文档内容符合您的需求后进行下载，若出现内容与标题不符可向本站投诉处理。
4. 下载文档时可能由于网络波动等原因无法下载或下载错误，付费完成后未能成功下载的用户请联系客服处理。