an introduction to conditional random fields

an introduction to conditional random fields

ID:15173386

大小:674.86 KB

页数:90页

时间:2018-08-01

an introduction to conditional random fields_第页
预览图正在加载中,预计需要20秒,请耐心等待
资源描述:

《an introduction to conditional random fields》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、AnIntroductiontoConditionalRandomFieldsCharlesSuttonUniversityofEdinburghcsutton@inf.ed.ac.ukAndrewMcCallumUniversityofMassachusettsAmherstmccallum@cs.umass.edu17November2010AbstractOftenwewishtopredictalargenumberofvariablesthatdependoneachotheraswellasonotherobservedvariables.Structuredpredic

2、-tionmethodsareessentiallyacombinationofclassi cationandgraph-icalmodeling,combiningtheabilityofgraphicalmodelstocompactlymodelmultivariatedatawiththeabilityofclassi cationmethodstoperformpredictionusinglargesetsofinputfeatures.Thistutorialde-scribesconditionalrandom elds,apopularprobabilisticm

3、ethodforstructuredprediction.CRFshaveseenwideapplicationinnaturallan-guageprocessing,computervision,andbioinformatics.WedescribemethodsforinferenceandparameterestimationforCRFs,includingarXiv:1011.4088v1[stat.ML]17Nov2010practicalissuesforimplementinglargescaleCRFs.Wedonotassumepreviousknowledg

4、eofgraphicalmodeling,sothistutorialisintendedtobeusefultopractitionersinawidevarietyof elds.Contents1Introduction12Modeling52.1GraphicalModeling62.2GenerativeversusDiscriminativeModels102.3Linear-chainCRFs182.4GeneralCRFs212.5ApplicationsofCRFs232.6FeatureEngineering242.7NotesonTerminology263In

5、ference273.1Linear-ChainCRFs283.2InferenceinGraphicalModels323.3ImplementationConcerns404ParameterEstimation43i4.1MaximumLikelihood444.2StochasticGradientMethods524.3Parallelism544.4ApproximateTraining544.5ImplementationConcerns615RelatedWorkandFutureDirections635.1RelatedWork635.2FrontierAreas

6、701IntroductionFundamentaltomanyapplicationsistheabilitytopredictmultiplevariablesthatdependoneachother.Suchapplicationsareasdiverseasclassifyingregionsofanimage[60],estimatingthescoreinagameofGo[111],segmentinggenesinastrandofDNA[5],andextractingsyntaxfromnatural-languagetext[123].Insuchapplic

7、ations,wewishtopredictavectory=fy0;y1;:::;yTgofrandomvariablesgivenanobservedfeaturevectorx.Arelativelysimpleexamplefromnatural-languageprocessingispart-of-speechtagging,inwhicheachvariableysisthepart-of-speechtagofthewordatpositi

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。