Published: 2017, Last Modified: 12 May 2023AISTATS 2017Readers: Everyone
Abstract:We develop a unified and systematic framework for performing online nonnegative matrix factorization under a wide variety of important divergences. The online nature of our algorithms makes them pa...