AnInference-BasedPolicyGradientMethodforLearningOptionsMatthewJ.A.Smith1HerkeVanHoof2JoellePineau1Abstractatvariouslevelsofabstraction,itispossibletoinfer,learnandplanmuchmoreefficiently.Further,ab...
ASemismoothNewtonMethodforFast,GenericConvexProgrammingAlnurAli1EricWong1J.ZicoKolter2Abstracti.e.,thespaceofm×mpositivesemidefinitematricesSm+,yieldsasemidefiniteprogram;andtakingKasthesecond-Wei...
SARAH:ANovelMethodforMachineLearningProblemsUsingStochasticRecursiveGradientLamM.Nguyen1JieLiu1KatyaScheinberg12MartinTaka´cˇ1AbstractProblemsofthistypearisefrequentlyinsupervisedlearn-Inthispape...
Nystro¨mMethodwithKernelK-means++SamplesasLandmarksDinoOglic12ThomasGa¨rtner2AbstractoreigendecompositionwhichscaleasOn3.ToovercomethiscomputationalshortcomingandscalekernelMethodsWeinvestigate,t...
DissipativityTheoryforNesterov’sAcceleratedMethodBinHu1LaurentLessard1Abstracttories.OnceaLyapunovfunctionisfound,onecanrelatetherateofdecreaseofthisinternalenergytotherateofInthispaper,weadaptthe...