知识图谱的构建与应用-朱小燕

知识图谱的构建与应用-朱小燕

ID:46825712

大小:3.26 MB

页数:32页

时间:2019-11-28

知识图谱的构建与应用-朱小燕_第1页
知识图谱的构建与应用-朱小燕_第2页
知识图谱的构建与应用-朱小燕_第3页
知识图谱的构建与应用-朱小燕_第4页
知识图谱的构建与应用-朱小燕_第5页
资源描述:

《知识图谱的构建与应用-朱小燕》由会员上传分享,免费在线阅读,更多相关内容在行业资料-天天文库

1、知识图谱的构建与应用清华大学计算机系朱小燕zxy-dcs@tsinghua.edu.cn,@朱小燕THU2014.4.18Contents1•Information&Knowledge2•Challenges3•Ideas&AchievementsInformation&KnowledgeWhatdoweusetheinternetfor?INFORMATIONSERVICEAccesstheInternetBuildknowledgeforgettingnews,graphforURL,etc.generatinganswersNewsMicr

2、oblogPictureVideoEncyclopediaForumInformationServices•BaiduBoxComputing•GoogleKnowledgeGraphKnowledge•Whatisknowledge?–Entity+Rule–Triple(predicate,semanticnetwork,rule,framework,……)AbstractedInformationDrawback:LackofabilitytoCodedInformationcommunicatewithmachineDIRECTLY!知识

3、的作用:帮助信息计算、理解、评价ChallengesScientificPerspective•OurGoal–InformationMeasurable–KnowledgeComputableMeasureMiningAcquisitionStructurizationInformation/KnowledgeApplicationPerspectiveInformationAcquisition•HowtomeasuretheinformationKnowledgeBase•Howtobuildknowledgebase–Tobuildfro

4、moriginalmaterials–Toextendandupdate–Tomergedifferencedatabases–Toverifytheknowledge•Howtouseknowledgebases–Foranswergeneration–Foranswerre-ranking–Forinference–Forrecommendation–Forverticalsearch–……IdeasandAchievementsKnowledgeBaseConstruction•KnowledgeBaseConstruction–Gener

5、alKnowledgeBaseConstruction–InformationOrganizationforMulti-sourceUGCs–DomainspecificKnowledgeBase•KeyPoints–Organizationwithsemanticinformation•TopicHierarchystructure–Practicability•Timeliness,IterativelyUpdate•Customization,QueryOrientedInformationOrganizationforMulti-sour

6、ceUGCsviaTopicHierarchyConstructionPowerOfStatistics:FindaTopic•Multi-sourceUGCs–Hugevolume&diversifiedtopics–ComplementaryQualityofContents:DescribeaTopicCanweorganizeMulti-sourceUGCsbyaunifiedstructure,•TopicHierarchywhichcaneffectivelyguideuserstotheirrequiredknowledge?–ro

7、otnode:usertopicpollpolicyIslam–non-rootnode:subtopics…–leafnode:linktoUGCs…mediccare…InformationOrganizationforMulti-sourceUGCsviaTopicHierarchyConstruction(cont.)•TopicExtractionBarackObamaPolicy–CandidateselectionviaMulti-sourcep=0.009basedVoterp=0.015p=0.013–Sub-topicrela

8、tionviaMulti-sourceevidencep=0.006learningandcombinationTaxDebate•To

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。