Advanced Deep Learning with Keras
Recentdevelopmentsindeeplearning,includingGANs,VariationalAutoencoders,andDeepReinforcementLearning,arecreatingimpressiveAIresultsinournewsheadlines-suchasAlphaGoZerobeatingworldchesschampions,andgenerativeAIthatcancreateartpaintingsthatsellforover$400kbecausetheyaresohuman-like.AdvancedDeepLearningwithKerasisacomprehensiveguidetotheadvanceddeeplearningtechniquesavailabletoday,soyoucancreateyourowncutting-edgeAI.UsingKerasasanopen-sourcedeeplearninglibrary,you'llfindhands-onprojectsthroughoutthatshowyouhowtocreatemoreeffectiveAIwiththelatesttechniques.ThejourneybeginswithanoverviewofMLPs,CNNs,andRNNs,whicharethebuildingblocksforthemoreadvancedtechniquesinthebook.You’lllearnhowtoimplementdeeplearningmodelswithKerasandTensorflow,andmoveforwardstoadvancedtechniques,asyouexploredeepneuralnetworkarchitectures,includingResNetandDenseNet,andhowtocreateAutoencoders.YouthenlearnallaboutGenerativeAdversarialNetworks(GANs),andhowtheycanopennewlevelsofAIperformance.VariationalAutoEncoders(VAEs)areimplemented,andyou’llseehowGANsandVAEshavethegenerativepowertosynthesizedatathatcanbeextremelyconvincingtohumans-amajorstrideforwardformodernAI.Tocompletethissetofadvancedtechniques,you'lllearnhowtoimplementDeepReinforcementLearning(DRL)suchasDeepQ-LearningandPolicyGradientMethods,whicharecriticaltomanymodernresultsinAI.
·7萬字