5.From Frequency to Meaning- Vector Space Models of Semantics

5.From Frequency to Meaning- Vector Space Models of Semantics

ID:40703773

大小:298.44 KB

页数:48页

时间:2019-08-06

5.From Frequency to Meaning- Vector Space Models of Semantics_第1页
5.From Frequency to Meaning- Vector Space Models of Semantics_第2页
5.From Frequency to Meaning- Vector Space Models of Semantics_第3页
5.From Frequency to Meaning- Vector Space Models of Semantics_第4页
5.From Frequency to Meaning- Vector Space Models of Semantics_第5页
资源描述:

《5.From Frequency to Meaning- Vector Space Models of Semantics》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、JournalofArti cialIntelligenceResearch37(2010)141-188Submitted10/09;published02/10FromFrequencytoMeaning:VectorSpaceModelsofSemanticsPeterD.Turneypeter.turney@nrc-cnrc.gc.caNationalResearchCouncilCanadaOttawa,Ontario,Canada,K1A0R6PatrickPantelme@patrickpantel.comYahoo!LabsSunnyval

2、e,CA,94089,USAAbstractComputersunderstandverylittleofthemeaningofhumanlanguage.Thisprofoundlylimitsourabilitytogiveinstructionstocomputers,theabilityofcomputerstoexplaintheiractionstous,andtheabilityofcomputerstoanalyseandprocesstext.Vectorspacemodels(VSMs)ofsemanticsarebeginningt

3、oaddresstheselimits.ThispapersurveystheuseofVSMsforsemanticprocessingoftext.WeorganizetheliteratureonVSMsaccordingtothestructureofthematrixinaVSM.TherearecurrentlythreebroadclassesofVSMs,basedonterm{document,word{context,andpair{patternmatrices,yieldingthreeclassesofapplications.W

4、esurveyabroadrangeofapplicationsinthesethreecategoriesandwetakeadetailedlookataspeci copensourceprojectineachcategory.OurgoalinthissurveyistoshowthebreadthofapplicationsofVSMsforsemantics,toprovideanewperspectiveonVSMsforthosewhoarealreadyfamiliarwiththearea,andtoprovidepointersin

5、totheliteratureforthosewhoarelessfamiliarwiththe eld.1.IntroductionOneofthebiggestobstaclestomakingfulluseofthepowerofcomputersisthattheycurrentlyunderstandverylittleofthemeaningofhumanlanguage.Recentprogressinsearchenginetechnologyisonlyscratchingthesurfaceofhumanlanguage,andyett

6、heimpactonsocietyandtheeconomyisalreadyimmense.Thishintsatthetransformativeimpactthatdeepersemantictechnologieswillhave.Vectorspacemodels(VSMs),surveyedinthispaper,arelikelytobeapartofthesenewsemantictechnologies.Inthispaper,weusethetermsemanticsinageneralsense,asthemeaningofaword

7、,aphrase,asentence,oranytextinhumanlanguage,andthestudyofsuchmeaning.Wearenotconcernedwithnarrowersensesofsemantics,suchasthesemanticweborapproachestosemanticsbasedonformallogic.WepresentasurveyofVSMsandtheirrelationwiththedistributionalhypothesisasanapproachtorepresentingsomeaspe

8、ctsofnaturallanguagesemantics.The

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。