UnderstandingInvarianceviaFeedforwardInversionofDiscriminativelyTrainedClassifiersPiotrTeterwak1ChiyuanZhang2DilipKrishnan2MichaelC.Mozer23Abstract1.IntroductionAdiscriminativelytrainedneuralnetcla...
RegularizingtowardsCausalInvariance:LinearModelswithProxiesMichaelOberst1NikolajThams2JonasPeters2DavidSontag1AbstractβxAβyWeproposeamethodforlearninglinearmod-XαYelswhosepredictiveperformanceis...
HowrotationalInvarianceofcommonkernelspreventsgeneralizationinhighdimensionsKonstantinDonhauser1MingqiWu1FannyYang1Abstractconclusiveintheregimed/nβ→cwithβ∈(0,1).Inpar-ticular,eventhoughseveral...
RevisitingSpatialInvariancewithLow-RankLocalConnectivityGamaleldinF.Elsayed1PrajitRamachandran1JonathonShlens1SimonKornblith1AbstractH1:SpatialInvarianceisagoodinductivebias.H2:SpatialInvarianceiso...
FundamentalTradeoffsbetweenInvarianceandSensitivitytoAdversarialPerturbationsFlorianTramèr1JensBehrmann2NicholasCarlini3NicolasPapernot3Jörn-HenrikJacobsen4AbstractAdversarialexamplesarecommonlyf...
EliminatingtheInvarianceontheLossLandscapeofLinearAutoencodersRezaOftadeh1JiayiShen1ZhangyangWang1DylanShell1Abstractgenerallythecategorizationofextremepoints,hasflour-ished.Muchworkextendstheresul...
ConvexRepresentationLearningforGeneralizedInvarianceinSemi-Inner-ProductSpaceYingyiMa1VigneshGanapathiraman1YaoliangYu2XinhuaZhang1Abstractshipsbetweenclasseswhichholdforanyinput.Commonexamplesincl...
InvarianceofWeightDistributionsinRectifiedMLPsRussellTsuchida1FarbodRoosta-Khorasani23MarcusGallagher1Abstractplicationratherthananunderstandingofthecapabilitiesandtrainingofneuralnetworks.Recently...
AcceleratingNaturalGradientwithHigher-OrderInvarianceYangSong1JiamingSong1StefanoErmon1Abstractetal.,2013),andfirst-ordermethodswillhavetroubleinmakingprogress.Thecurvature,however,dependsonhowAnap...
WarpedConvolutions:EfficientInvariancetoSpatialTransformationsJoa˜oF.Henriques1AndreaVedaldi1Abstractacrossspace)andlesscomputation(duetotheirlimitedsupport)comparedtoafully-connectedlayer.Itsloca...