年 月 四川大学学报 自然科学版
2022 3 ( )
RENet tacticsandtechniuesclassificationsforcber
: q y
threatintelligencewithrelevanceenhancement
GE WenGHan1 WANGJunGFen 1 TANGBinGHui2 YUZhonGKun1 CHENBoGHan1 YUJian1
, g , , g , ,
(1.CollegeofComputerScience,SichuanUniversity,Chengdu610065,China;
2.SchoolofCyberScienceandEngineering,SichuanUniversity,Chengdu610065,China)
Abstract
:Tactics,TechniquesandProcedures(TTPs)analysisinCyberThreatIntelligence(CTI),isa
keytechniqueforcyberattacktraceabilitywhichprovidingaglobalviewofcyberattackeventsandreveaG
lingsystem weaknesses.ExistingTTPsclassificationschemesarepoorlyandunevenlyorientedtoabG
stractlanguageenvironments.Inthispaper,weproposeamultiGlabeldeeplearningmodelbasedonasG
sociationenhancement:RENet,whichclassifiestacticsandtechniquesbyusingamultiGlabelclassifier
thatcombinescontextualinformationandmultiplewordmeanings,andenhancestechniqueclassification
bytransferringtheclassificationresultsoftheoriginaltacticsthroughaconditionaltransfermatrixfrom
tacticstotechniques.Experimentsshowthat
基于关联增强的网络威胁情报技战术分类 来自淘豆网www.taodocs.com转载请标明出处.