2017 (modified: 08 Nov 2022)AISTATS 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...