SuperpolynomialLowerBoundsforLearningOne-LayerNeuralNetworksusingGradientDescentSurbhiGoel1AravindGollakota1ZhihanJin2SushrutKarmalkar1AdamKlivans1AbstractOurResults.Inthispaperwegivethefirstsuperp...
SGDLearnsOne-LayerNetworksinWGANsQiLei1JasonD.Lee2AlexandrosG.Dimakis1ConstantinosDaskalakis3Abstractbutionwithlowsamplecomplexity.However,theseresultscannotbealgorithmicallyattainedviapracticalGAN...