Faster-than-fast NMF

Random projections have been recently implemented in Nonnegative Matrix Factorization (NMF) to speed-up the NMF computations, with a negligible loss of performance. In this paper, we investigate the effects of such projections when the NMF technique uses the fast Nesterov gradient descent (NeNMF). We experimentally show that structured random projections significantly speed-up NeNMF for very large data matrices.

The code was written in Matlab by Farouk YAHAYA and is maintained by Farouk YAHAYA, Matthieu PUIGT, Gilles DELMAIRE, and Gilles ROUSSEL (firstname.LASTNAME [at]

Link to the code:

You can use it for for research or educational purpose. In that case, please cite:

F. Yahaya, M. Puigt, G. Delmaire, and G. Roussel, « Faster-than-fast NMF using random projections and Nesterov iterations,«  in Proc. of iTWIST: international Traveling Workshop on Interactions between low-complexity data models and Sensing Techniques, Marseille, France, November 21-23, 2018.

For any suggestions or questions about this code, please contact: Farouk.Yahaya [at] and Matthieu.Puigt [at]


IN-Cal, ACIN-Cal, SpIN-Cal et SpAIN-Cal sont des méthodes d’étalonnage in situ de capteurs mobiles soumis à des dérives. Ces approches sont basées sur des formalismes

Aller au contenu principal