awesome_deep_learig_iterpretability
深度学习近年来关于模型解释性的相关论文。
按引用次数排序可见引用排序
159篇论文pdf(有2篇需要上scihub找)上传到腾讯微云。
不定期更新。
YearPublicatioPaperCitatiocode2020CVPRExplaiigKowledgeDistillatiobyQuatifyigtheKowledge32020CVPRHigh-frequecyCompoetHelpsExplaitheGeeralizatioofCovolutioalNeuralNetworks162020CVPRWScore-CAM:Score-WeightedVisualExplaatiosforCovolutioalNeuralNetworks7Pytorch2020ICLRKowledgecosistecybetweeeuraletworksadbeyod32020ICLRIterpretableComplex-ValuedNeuralNetworksforPrivacyProtectio22019AIExplaatioiartificialitelligece:Isightsfromthesocialscieces6622019NMIStopExplaiigBlackBoxMachieLearigModelsforHighStakesDecisiosadUseIterpretableModelsIstead3892019NeurIPSCayoutrustyourmodel'sucertaity?Evaluatigpredictiveucertaityuderdatasetshift136-2019NeurIPSThislookslikethat:deeplearigforiterpretableimagerecogitio80Pytorch2019NeurIPSAbechmarkforiterpretabilitymethodsideepeuraletworks282019NeurIPSFull-gradietrepresetatioforeuraletworkvisualizatio72019NeurIPSOthe(I)fidelityadSesitivityofExplaatios132019NeurIPSTowardsAutomaticCocept-basedExplaatios25Tesorflow2019NeurIPSCXPlai:Causalexplaatiosformodeliterpretatiouderucertaity122019CVPRIterpretigCNNsviaDecisioTrees852019CVPRFromRecogitiotoCogitio:VisualCommoseseReasoig97Pytorch2019CVPRAttetiobrachetwork:Learigofattetiomechaismforvisualexplaatio392019CVPRIterpretableadfie-graiedvisualexplaatiosforcovolutioaleuraletworks182019CVPRLearigtoExplaiwithComplemetalExamples122019CVPRRevealigSceesbyIvertigStructurefromMotioRecostructios20Tesorflow2019CVPRMultimodalExplaatiosbyPredictigCouterfactualityiVideos42019CVPRVisualizigtheResilieceofDeepCovolutioalNetworkIterpretatios12019ICCVU-CAM:VisualExplaatiousigUcertaitybasedClassActivatioMaps102019ICCVTowardsIterpretableFaceRecogitio72019ICCVTakigaHINT:LeveragigExplaatiostoMakeVisioadLaguageModelsMoreGrouded282019ICCVUderstadigDeepNetworksviaExtremalPerturbatiosadSmoothMasks17Pytorch2019ICCVExplaiigNeuralNetworksSematicallyadQuatitatively62019ICLRHierarchicaliterpretatiosforeuraletworkpredictios24Pytorch2019ICLRHowImportatIsaNeuro?322019ICLRVisualExplaatiobyIterpretatio:ImprovigVisualFeedbackCapabilitiesofDeepNeuralNetworks132018ICMLExtractigAutomatafromRecurretNeuralNetworksUsigQueriesadCouterexamples71Pytorch2019ICMLTowardsADeepadUifiedUderstadigofDeepNeuralModelsiNLP15Pytorch2019ICAISIterpretigblackboxpredictiosusigfisherkerels242019ACMFATExplaiigexplaatiosiAI1192019AAAIIterpretatioofeuraletworksisfragile130Tesorflow2019AAAIClassifier-agosticsaliecymapextractio82019AAAICaYouExplaiThat?LucidExplaatiosHelpHuma-AICollaborativeImageRetrieval12019AAAIWUsupervisedLearigofNeuralNetworkstoExplaiNeuralNetworks102019AAAIWNetworkTrasplatig42019CSURASurveyofMethodsforExplaiigBlackBoxModels6552019JVCIRIterpretablecovolutioaleuraletworksviafeedforwarddesig31Keras2019ExplaiAIThe(U)reliabilityofsaliecymethods1282019ACLAttetioisotExplaatio1572019EMNLPAttetioisototExplaatio572019arxivAttetioIterpretabilityAcrossNLPTasks162019arxivIterpretableCNNs22018ICLRTowardsbetteruderstadigofgradiet-basedattributiomethodsfordeepeuraletworks2452018ICLRLearighowtoexplaieuraletworks:PatterNetadPatterAttributio1432018ICLROtheimportaceofsigledirectiosforgeeralizatio134Pytorch2018ICLRDetectigstatisticaliteractiosfromeuraletworkweights56Pytorch2018ICLRIterpretablecoutigforvisualquestioaswerig29Pytorch2018CVPRIterpretableCovolutioalNeuralNetworks2502018CVPRTellmewheretolook:Guidedattetioifereceetwork134Chaier2018CVPRMultimodalExplaatios:JustifyigDecisiosadPoitigtotheEvidece126Caffe2018CVPRTrasparecybydesig:Closigthegapbetweeperformaceaditerpretabilityivisualreasoig79Pytorch2018CVPRNet2vec:Quatifyigadexplaiighowcoceptsareecodedbyfiltersideepeuraletworks602018CVPRWhathavewelearedfromdeeprepresetatiosforactiorecogitio?302018CVPRLearigtoActProperly:PredictigadExplaiigAffordacesfromImages242018CVPRTeachigCategoriestoHumaLearerswithVisualExplaatios20Pytorch2018CVPRWhatdodeepetworksliketosee?192018CVPRIterpretNeuralNetworksbyIdetifyigCriticalDataRoutigPaths13Tesorflow2018ECCVDeepclusterigforusupervisedlearigofvisualfeatures382Pytorch2018ECCVExplaiableeuralcomputatioviastackeuralmoduleetworks55Tesorflow2018ECCVGroudigvisualexplaatios442018ECCVTextualexplaatiosforself-drivigvehicles592018ECCVIterpretablebasisdecompositioforvisualexplaatio51Pytorch2018ECCVCovetsadimageetbeyodaccuracy:Uderstadigmistakesaducoverigbiases362018ECCVVqa-e:Explaiig,elaboratig,adehacigyouraswersforvisualquestios202018ECCVChooseYourNeuro:IcorporatigDomaiKowledgethroughNeuro-Importace16Pytorch2018ECCVDiversefeaturevisualizatiosrevealivariacesiearlylayersofdeepeuraletworks9Tesorflow2018ECCVExplaiGAN:ModelExplaatioviaDecisioBoudaryCrossigTrasformatios62018ICMLIterpretabilitybeyodfeatureattributio:Quatitativetestigwithcoceptactivatiovectors214Tesorflow2018ICMLLearigtoexplai:Aiformatio-theoreticperspectiveomodeliterpretatio1172018ACLDidtheModelUderstadtheQuestio?63Tesorflow2018FITEEVisualiterpretabilityfordeeplearig:asurvey2432018NeurIPSSaityChecksforSaliecyMaps2492018NeurIPSExplaatiosbasedothemissig:Towardscotrastiveexplaatioswithpertietegatives79Tesorflow2018NeurIPSTowardsrobustiterpretabilitywithself-explaiigeuraletworks145Pytorch2018NeurIPSAttacksmeetiterpretability:Attribute-steereddetectioofadversarialsamples552018NeurIPSDeepPINK:reproduciblefeatureselectioideepeuraletworks30Keras2018NeurIPSRepreseterpoitselectioforexplaiigdeepeuraletworks30Tesorflow2018NeurIPSWorkshopIterpretablecovolutioalfilterswithsicNet372018AAAIAchors:High-precisiomodel-agosticexplaatios3662018AAAIImprovigtheadversarialrobustessaditerpretabilityofdeepeuraletworksbyregularizigtheiriputgradiets178Tesorflow2018AAAIDeeplearigforcase-basedreasoigthroughprototypes:Aeuraletworkthatexplaisitspredictios102Tesorflow2018AAAIIterpretigCNNKowledgeviaaExplaatoryGraph79Matlab2018AAAIExamiigCNNRepresetatioswithrespecttoDatasetBias372018WACVGrad-cam++:Geeralizedgradiet-basedvisualexplaatiosfordeepcovolutioaletworks1742018IJCVTop-doweuralattetiobyexcitatiobackprop3292018TPAMIIterpretigdeepvisualrepresetatiosviaetworkdissectio872018DSPMethodsforiterpretigaduderstadigdeepeuraletworks7132018AccessPeekigisidetheblack-box:AsurveyoExplaiableArtificialItelligece(XAI)3902018JAIRLearigExplaatoryRulesfromNoisyData155Tesorflow2018MIPROExplaiableartificialitelligece:Asurvey1082018BMVCRise:Radomizediputsampligforexplaatioofblack-boxmodels852018arxivDistill-ad-Compare:AuditigBlack-BoxModelsUsigTrasparetModelDistillatio302018arxivMaipulatigadmeasurigmodeliterpretability1332018arxivHowcovolutioaleuraletworkseetheworld-Asurveyofcovolutioaleuraletworkvisualizatiomethods452018arxivRevisitigtheimportaceofidividualuitsicsviaablatio432018arxivComputatioallyEfficietMeasuresofIteralNeuroImportace12017ICMLUderstadigBlack-boxPredictiosviaIflueceFuctios767Pytorch2017ICMLAxiomaticattributiofordeepetworks755Keras2017ICMLLearigImportatFeaturesThroughPropagatigActivatioDiffereces6552017ICLRVisualizigdeepeuraletworkdecisios:Predictiodiffereceaalysis271Caffe2017ICLRExplorigLOTSiDeepNeuralNetworks272017NeurIPSAUifiedApproachtoIterpretigModelPredictios14112017NeurIPSRealtimeimagesaliecyforblackboxclassifiers161Pytorch2017NeurIPSSVCCA:SigularVectorCaoicalCorrelatioAalysisforDeepLearigDyamicsadIterpretability1602017CVPRMiigObjectPartsfromCNNsviaActiveQuestio-Aswerig202017CVPRNetworkdissectio:Quatifyigiterpretabilityofdeepvisualrepresetatios5402017CVPRImprovigIterpretabilityofDeepNeuralNetworkswithSematicIformatio562017CVPRMDNet:ASematicallyadVisuallyIterpretableMedicalImageDiagosisNetwork129Torch2017CVPRMakigtheViVQAmatter:ElevatigtheroleofimageuderstadigiVisualQuestioAswerig5822017CVPRKowigwhetolook:Adaptiveattetioviaavisualsetielforimagecaptioig620Torch2017CVPRWIterpretable3dhumaactioaalysiswithtemporalcovolutioaletworks1632017ICCVGrad-cam:Visualexplaatiosfromdeepetworksviagradiet-basedlocalizatio2444Pytorch2017ICCVIterpretableExplaatiosofBlackBoxesbyMeaigfulPerturbatio419Pytorch2017ICCVIterpretableLearigforSelf-DrivigCarsbyVisualizigCausalAttetio1142017ICCVUderstadigadcomparigdeepeuraletworksforageadgederclassificatio522017ICCVLearigtodisambiguatebyaskigdiscrimiativequestios122017IJCAIRightfortherightreasos:Traiigdifferetiablemodelsbycostraiigtheirexplaatios1492017IJCAIUderstadigadimprovigcovolutioaleuraletworksviacocateatedrectifiedliearuits276Caffe2017AAAIGrowigIterpretablePartGraphsoCovNetsviaMulti-ShotLearig37Matlab2017ACLVisualizigadUderstadigNeuralMachieTraslatio922017EMNLPAcausalframeworkforexplaiigthepredictiosofblack-boxsequece-to-sequecemodels922017CVPRWorkshopLookiguderthehood:Deepeuraletworkvisualizatiotoiterpretwhole-slideimageaalysisoutcomesforcolorectalpolyps212017surveyIterpretabilityofdeeplearigmodels:asurveyofresults992017arxivSmoothGrad:removigoisebyaddigoise3562017arxivIterpretable&explorableapproximatiosofblackboxmodels1152017arxivDistilligaeuraletworkitoasoftdecisiotree188Pytorch2017arxivTowardsiterpretabledeepeuraletworksbyleveragigadversarialexamples542017arxivExplaiableartificialitelligece:Uderstadig,visualizigaditerpretigdeeplearigmodels3832017arxivCotextualExplaatioNetworks35Pytorch2017arxivChallegesfortrasparecy832017ACMSOPPDeepxplore:Automatedwhiteboxtestigofdeeplearigsystems4312017CEURWWhatdoesexplaiableAIreallymea?Aewcoceptualizatioofperspectives1172017TVCGActiVis:VisualExploratioofIdustry-ScaleDeepNeuralNetworkModels1582016NeurIPSSythesizigthepreferrediputsforeurosieuraletworksviadeepgeeratoretworks321Caffe2016NeurIPSUderstadigtheeffectivereceptivefieldideepcovolutioaleuraletworks4362016CVPRIvertigVisualRepresetatioswithCovolutioalNetworks3362016CVPRVisualizigadUderstadigDeepTextureRepresetatios982016CVPRAalyzigClassifiers:FisherVectorsadDeepNeuralNetworks1102016ECCVGeeratigVisualExplaatios303Caffe2016ECCVDesigofkerelsicovolutioaleuraletworksforimageclassificatio142016ICMLUderstadigadimprovigcovolutioaleuraletworksviacocateatedrectifiedliearuits2762016ICMLVisualizigadcomparigAlexNetadVGGusigdecovolutioallayers412016EMNLPRatioalizigNeuralPredictios355Pytorch2016IJCVVisualizigdeepcovolutioaleuraletworksusigaturalpre-images281Matlab2016IJCVVisualizigObjectDetectioFeatures27Caffe2016KDDWhyshoulditrustyou?:Explaiigthepredictiosofayclassifier35112016TVCGVisualizigthehiddeactivityofartificialeuraletworks1702016TVCGTowardsbetteraalysisofdeepcovolutioaleuraletworks2412016NAACLVisualizigaduderstadigeuralmodelsilp364Torch2016arxivUderstadigeuraletworksthroughrepresetatioerasure)1982016arxivGrad-CAM:Whydidyousaythat?1302016arxivIvestigatigtheiflueceofoiseaddistractorsotheiterpretatioofeuraletworks412016arxivAttetiveExplaatios:JustifyigDecisiosadPoitigtotheEvidece542016arxivTheMythosofModelIterpretability13682016arxivMultifacetedfeaturevisualizatio:Ucoverigthedifferettypesoffeatureslearedbyeacheuroideepeuraletworks1612015ICLRStrivigforSimplicity:TheAllCovolutioalNet2268Pytorch2015CVPRUderstadigdeepimagerepresetatiosbyivertigthem1129Matlab2015ICCVUderstadigdeepfeatureswithcomputer-geeratedimagery109Caffe2015ICMLWorkshopUderstadigNeuralNetworksThroughDeepVisualizatio1216Tesorflow2015AASIterpretableclassifiersusigrulesadBayesiaaalysis:Buildigabetterstrokepredictiomodel3852014ECCVVisualizigadUderstadigCovolutioalNetworks9873Pytorch2014ICLRDeepIsideCovolutioalNetworks:VisualisigImageClassificatioModelsadSaliecyMaps2745Pytorch2013ICCVHoggles:Visualizigobjectdetectiofeatures301论文talk
评论