机器学习简介课件ppt

Machine Learning：An Overview石立臣1Machine Learning：An Overview石OutlineWhat is machine learning(ML)Types of machine learningWork flowPopular modelsApplicationsFutures2OutlineWhat is machine learninWhat is machine learningTraining set(labels known)Test set(labels unknown)f()=“apple”f()=“tomato”f()=“cow”3What is machine learningTrainiWhat is machine learningDefinitionMachine learning refers to a system capable of the autonomous acquisition and integration of knowledgeMachine learning is programming computers to optimize a performance criterion using example data or past experienceComputerDataAlgorithmProgramKnowledgeKnowledge(new)4What is machine learningDefiniWhat is machine learningEvery machine learning algorithm has three componentsRepresentationModel(rules,statistics,instance;logic,KNN,SVM,DNN,)EvaluationPerformance(accuracy,mse,energy,entropy,)OptimizationParameters Combinatorial optimizationConvex optimizationConstrained optimization5What is machine learningEvery Types of machine learningSupervised learningTraining data includes desired outputsUnsupervised learningTraining data does not include desired outputsSemi-supervised learningTraining data includes a few desired outputsReinforcement learningRewards from sequence of actions6Types of machine learningSuperTypes of machine learningSupervised learningClassification:discrete outputRegression:continuous outputBias-variance7Types of machine learningSuperTraining and Validation DataFull Data SetTraining DataValidation DataIdea:train eachmodel on the“training data”and then testeach modelsaccuracy onthe validation data8Training and Validation DataFuUnderfitting&OverfittingPredictiveErrorModel ComplexityError on Training DataError on Test DataIdeal Rangefor Model ComplexityOverfittingUnderfitting9Underfitting&OverfittingPredTypes of machine learningUnsupervised learningClusteringDimensionality reductionFactor analysis10Types of machine learningUnsupTypes of machine learningSemi-supervised learningClustering or classification11Types of machine learningSemi-Types of machine learningReinforcement learningRobot&control12Types of machine learningReinfWork flowPredictionTraining LabelsTrainingTrainingImage FeaturesImage FeaturesTestingTest ImageLearned modelLearned modelSlide credit:D.Hoiem and L.Lazebnik13Work flowPredictionTraining LaWork flowFeatures14Work flowFeatures14Work flowModelsLogic,RulesStatistical,Black box modelStatic,dynamic modelOnline learningEnsemble learning15Work flowModels15Work flowArchitectureModelFeatureHardware16Work flowArchitectureModelFeatPopular modelsLinear model:logistic regression,linear discriminant analysis,linear regression(with basis function)17Popular modelsLinear model:loPopular modelsNearest neighborFeature&distance18Popular modelsNearest neighborPopular modelsSupport vector machine19Popular modelsSupport vector mPopular modelsArtificial neural network20Popular modelsArtificial neuraPopular modelsDecision tree21Popular modelsDecision tree21Popular modelsCollaborative filtering22Popular modelsCollaborative fiPopular modelsHierarchical clusteringK-meansSpectral clusteringManifold learning23Popular modelsHierarchical cluPopular modelsHidden markov modelConditional random fields24Popular modelsHidden markov moApplications25Applications25Applications26Applications26Applications27Applications27Applications28Applications28Applications29Applications29Applications30Applications30Applications31Applications31Applications32Applications32ApplicationsAttention 33ApplicationsAttention 33ApplicationsImage classification34ApplicationsImage classificatiApplications35Applications35ApplicationsBrain machine interface36ApplicationsBrain machine inteApplications37Applications37Applications38Applications38Applications39Applications39Applications40Applications40ApplicationsIndirect illuminationRegression41ApplicationsIndirect illuminatApplicationsIndirect illuminationkd-tree42ApplicationsIndirect illuminatApplicationsThe core is the data set!Othersfeaturesmodel&optimization43ApplicationsThe core is the daFuturesDecisionControlKnowledgePrediction44FuturesDecision44