SparseandImperceptibleAdversarialAttackviaaHomotopyAlgorithmMingkangZhu1TianlongChen1ZhangyangWang1AbstractSparseadversarialAttackscanfooldeepneural(a)otterhound(b)stingray(c)0=1803(d)0=12264networ...
Progressive-ScaleBoundaryBlackboxAttackviaProjectiveGradientEstimationJiaweiZhang1LinyiLi2HuichenLi2XiaoluZhang3ShuangYang4BoLi2Abstract1.IntroductionBoundarybasedblackboxAttackhasbeenrec-Blackboxa...
PopSkipJump:Decision-BasedAttackforProbabilisticClassifiersCarl-JohannSimon-Gabriel1NomanAhmedSheikh1AndreasKrause1Abstractnoisyorprobabilisticclassificationoutputs–aquitenaturalandcommonsettingin...
DefendingAgainstSaddlePointAttackinByzantine-RobustDistributedLearningDongYin1YudongChen2KannanRamchandran1PeterBartlett13Abstractarestoredandprocessedlocallyinendusers’cellphonesandpersonalcomput...
Adversarialcamerastickers:Aphysicalcamera-basedAttackondeeplearningsystemsJunchengB.Li12FrankR.Schmidt1J.ZicoKolter12AbstractFigure1.Illustrationofourapproach:(left)ouradversarialstickeraffixedtoac...
AdversarialAttackonGraphStructuredDataHanjunDai1HuiLi2TianTian3XinHuang2LinWang2JunZhu3LeSong12Abstractyieldgoodpredictions,butcouldalsoputthemodelinarisk.Acriminalcouldtrytodisguisehimselfbyconnec...