欢迎来到天天文库
浏览记录
ID:7296213
大小:299.97 KB
页数:5页
时间:2018-02-10
《bringing trust to collaborative content》由会员上传分享,免费在线阅读,更多相关内容在工程资料-天天文库。
1、ISSDMCollaborativeResearchProgramProposalProjectTitle:BringingTrusttoCollaborativeContentPrincipalInvestigator:LucadeAlfaroEmail:luca@soe.ucsc.eduPhone:(831)459-4982Web:http://www.soe.ucsc.edu/∼lucaComputerEngineeringDept.UniversityofCalifornia,SantaCruzOn-linecollaborationisfastbecomin
2、goneofthemosteffectivewaysinwhichinformationisbeingcreatedandshared.TheposterchildofthisphenomenonistheWikipedia,acollaborativeencyclopediawhichisbecomingoneofthemostcommonlyusedsourcesofinformationintheworld.Effectivecollaborationbenefitsfromtwomechanisms:incentivesystemsthatencourageco
3、nstructivebehavior,andtrustsystemsthathelpinformationusersjudgethereliabilityoftheinformationresultingfromthecollaborativeprocess.Specifically,wewilldevelopareputationsystemforauthors,andatrustsystemforcontent.Thereputationsystemwillbedrivenbycontentevolution,anditwillgivehigherreputatio
4、ntoauthorswhoprovidethemostlong-lived,andthususeful,contributions.Thetrustsystemwillcomputeavalueoftrustforthecontent,atwordgranularity:thetrustofawordwilldependonthereputationofitsauthor,aswellasthereputationoftheauthorswhorevisedthetextwherethewordappears.Contenttrustwillbedisplayedvi
5、aanintuitive,trust-basedcoloringofthetextbackground(seeFigure1).Wehavealreadyimplementedembryosofthesesystems.TheembryosworkonstaticdumpsofWikipediainformation,andarenotsuitedtodynamicinteractionwithvisitors:stilltheyprovideahintoftheirpotentialvalue.Thisprojectwilldevelopthekeyelements
6、thatarerequiredtoturntheseoff-lineembryosintoworkingon-linesystems,thatcanprovidereal-timetrustinformationtovisitors.Makingthesystemssuitabletoon-linedeploymentrequirestheinvestigation,andsolution,oftwomainresearchchallenges,whichformthetechnicalcoreoftheproject.Thefirstchallengeconsists
7、inmakingthereputationandtrustsystemsresistanttoattacks.Inanon-lineimplementation,somevisitorsmaywanttogamethesystemstogainreputationortocauseinformationoftheirchoicetobemarkedastrusted.Weproposetoinvestigatemethodsthatexploitthecontent-drivennatureofthereputationandtrustalgorit
此文档下载收益归作者所有