Unsupervised Skill Discovery with Bottleneck Option Learning

UnsupervisedSkillDiscoverywithBottleneckOptionLearningJaekyeomKim1SeohongPark1GunheeKim1Abstractlearnedskillscanencouragetheexplorationforencounter-ingrewards,notonlybyprovidingusefulprimitivesfort...

Skill Discovery for Exploration and Planning using Deep Skill Graphs

SkillDiscoveryforExplorationandPlanningusingDeepSkillGraphsAkhilBagaria1JasonSenthil1GeorgeKonidaris1AbstractWeintroduceanewskill-Discoveryalgorithmthatbuildsadiscretegraphrepresentationoflargecon-...

Quantifying Availability and Discovery in Recommender Systems via Stochastic Reachability

QuantifyingAvailabilityandDiscoveryinRecommenderSystemsviaStochasticReachabilityMihaelaCurmei1SarahDean1BenjaminRecht1Abstractwhichhavebeenimplicatedinunintendedconsequenceslikepolarizationorradica...

PAPRIKA Private Online False Discovery Rate Control

PAPRIKA:PrivateOnlineFalseDiscoveryRateControlWanrongZhang1GautamKamath2RachelCummings3Abstractoflarge-scaledatasetsandeaseofdataanalysis,whilebene-ficialtosociety,hascreatedaseverecrisisofreproduc...

Unsupervised Discovery of Interpretable Directions in the GAN Latent Space

UnsupervisedDiscoveryofInterpretableDirectionsintheGANLatentSpaceAndreyVoynov1ArtemBabenko12Abstract-Thickness+ThelatentspacesofGANmodelsoftenhave-Hair+semanticallymeaningfuldirections.Mov-inginthe...

Quantile Causal Discovery

DistinguishingCausefromEffectUsingQuantiles:BivariateQuantileCausalDiscoveryNatasaTagasovska12Vale´rieChavez-Demoulin1ThibaultVatter3Abstract(a)Additive(b)Multiplicative(c)Location−ScaleCausalinf...

Option Discovery in the Absence of Rewards with Manifold Analysis

OptionDiscoveryintheAbsenceofRewardswithManifoldAnalysisAmitayBar1RonenTalmon1RonMeir1Abstractthegraphedgesrepresentthestatesconnectivity.Suchanapproachledtotheintroductionofproto-valuefunctionsOpt...

Online Dense Subgraph Discovery via Blurred-Graph Feedback

OnlineDenseSubgraphDiscoveryviaBlurred-GraphFeedbackYukoKuroki12AtsushiMiyauchi12JunyaHonda12MasashiSugiyama21Abstractsity),whichisdefinedashalftheaveragedegreeofthesub-graphinducedbythesubset.Unli...

Explore, Discover and Learn Unsupervised Discovery of State-Covering Skills

Explore,DiscoverandLearn:UnsupervisedDiscoveryofState-CoveringSkillsV´ıctorCampos1AlexTrott2CaimingXiong2RichardSocher2XavierGiro-i-Nieto3JordiTorres1AbstractFigure1.Skillslearnedonamazewithbottl...

Efficient Intervention Design for Causal Discovery with Latents

EfficientInterventionDesignforCausalDiscoverywithLatentsRaghavendraAddanki1ShivaPrasadKasiviswanathan2AndrewMcGregor1CameronMusco1Abstractmorethanthreedecades,forthesereasonsithasreceivedincreasing...

Causal Structure Discovery from Distributions Arising from Mixtures of DAGs

CausalStructureDiscoveryfromDistributionsArisingfromMixturesofDAGsBasilSaeed1SnigdhaPanigrahi2CarolineUhler13AbstractseparationstatementsintheDAG(knownasd-separation)throughtheMarkovproperty(Laurit...

Unsupervised Deep Learning by Neighbourhood Discovery

UnsupervisedDeepLearningbyNeighbourhoodDiscoveryJiaboHuang1QiDong1ShaogangGong1XiatianZhu2Abstractturerepresentations,whichhasrecentlydrawnincreasingattention(Wuetal.,2018;Caronetal.,2018).Deepconv...

The Kernel Interaction Trick Fast Bayesian Discovery of Pairwise Interactions in High Dimensions

TheKernelInteractionTrick:FastBayesianDiscoveryofPairwiseInteractionsinHighDimensionsRajAgrawal1JonathanH.Huggins2BrianL.Trippe1TamaraBroderick1Abstractsearchersseektocharacterizehowindividual-leve...

EDDI Efficient Dynamic Discovery of High-Value Information with Partial VAE

Causal Discovery and Forecasting in Nonstationary Environments with State-Space Models

CausalDiscoveryandForecastinginNonstationaryEnvironmentswithState-SpaceModelsBiweiHuang1KunZhang1MingmingGong12ClarkGlymour1Abstractnonstationarytimeseries,andconcernedwithbothfindingcausalrelation...

Minimal I-MAP MCMC for Scalable Structure Discovery in Causal DAG Models

MinimalI-MAPMCMCforScalableStructureDiscoveryinCausalDAGModelsRajAgrawal123TamaraBroderick12CarolineUhler23Abstractamongrandomvariables.LearningBNsamongobservedvariablesfromdatahasprovenusefulindec...

A Laplacian Framework for Option Discovery in Reinforcement Learning

ALaplacianFrameworkforOptionDiscoveryinReinforcementLearningMarlosC.Machado1MarcG.Bellemare2MichaelBowling1Abstracttheoptimalpolicyforthatrewardfunction.Inthispaperweintroduceanalgorithmforoptiondi...

Uncertainty Assessment and False Discovery Rate Control in High-Dimensional Granger Causal Inference

UncertaintyAssessmentandFalseDiscoveryRateControlinHigh-DimensionalGrangerCausalInferenceAdityaChaudhry1PanXu2QuanquanGu2AbstractInstatistics,causalityisoftenestablishedbymeansofacon-trolled,random...

Automatic Discovery of the Statistical Types of Variables in a Dataset

AutomaticDiscoveryoftheStatisticalTypesofVariablesinaDatasetIsabelValera1ZoubinGhahramani12Abstractplore,findpatternsormakepredictionsonthedata.Asanexample,apredictiontaskissolveddifferentlydependi...