TestingGroupFairnessviaOptimalTransportProjectionsNianSi1KarthyekMurthy2JoseBlanchet1VietAnhNguyen13AbstractAnaturalfirstexplanationforthereportedalgorithmicbi-asesisthatthedatausedtotrainthealgori...
JustTrainTwice:ImprovingGroupRobustnesswithoutTrainingGroupInformationEvanZheranLiu1BehzadHaghgoo1AnnieS.Chen1AditiRaghunathan1PangWeiKoh1ShioriSagawa1PercyLiang1ChelseaFinn1Abstractcanbeespecially...
GroupFisherPruningforPracticalNetworkCompressionLiyangLiu1ShilongZhang2ZhanghuiKuang3AojunZhou3Jing-HaoXue4XinjiangWang3YiminChen3WenmingYang1QingminLiao1WayneZhang235Abstract78OursOursOursAOFPNetw...
CommutativeLieGroupVAEforDisentanglementLearningXinqiZhu1ChangXu1DachengTao2Abstractwhereσdenotestherepresentationfunction,andg,grepre-sentthesametransformationactingonthedataspaceandWeviewdisenta...
ApproximateGroupFairnessforClusteringBoLi1LijunLi2AnkangSun3ChenhaoWang4YingfanWang5Abstractinunsupervisedlearning–specifically,centroidclustering.Acanonicalclusteringproblemisdescribedas:givenaWe...
OnSecond-OrderGroupInfluenceFunctionsforBlack-BoxPredictionsSamyadeepBasu1XuchenYou1SoheilFeizi1Abstracttotracebackmodelpredictionstoitstrainingdata.Morespecifically,onecanaskwhichtrainingsampleswe...
LorentzGroupEquivariantNeuralNetworkforParticlePhysicsAlexanderBogatskiy1BrandonAnderson23JanT.Offermann1MarwahRoussi1DavidW.Miller14RisiKondor256Abstractproblemsmayrequireorbenefitfromlatentspacer...
LearnableGroupTransformForTime-SeriesRomainCosentino1BehnaamAazhang1AbstractAcommonapproachtoperforminginferenceontime-seriesconsistsofbuildingaDeepNeuralNetwork(DNN)thatWeproposeanovelapproachtofi...
AttentiveGroupEquivariantConvolutionalNetworksDavidW.Romero1ErikJ.Bekkers2JakubM.Tomczak1MarkHoogendoorn1AbstractFigure1.Meaningfulrelationshipsamongobjectsymmetries.Thougheveryfigureiscomposedbyth...
oi-VAE:OutputInterpretableVAEsforNonlinearGroupFactorAnalysisSamuelK.Ainsworth1NicholasJ.Foti1AdrianK.C.Lee2EmilyB.Fox1Abstractofthedata,and(ii)themappingcanbeusedtodescribethecorrelationstructureo...
MultilabelClassificationwithGroupTestingandCodesShashankaUbaru1AryaMazumdar2Abstractarevectorsindicatingthecorrespondinglabels(classes)thedatainstancesbelongto.ThevectoryihasaoneinthejthInrecentyea...
CombinedGroupandExclusiveSparsityforDeepNeuralNetworksJaehongYoon1SungJuHwang12Abstractnumberofparameters,deepnetworksrequirelargeamountofmemoryandcomputationpowertotrain.Further,largeThenumberofpa...