Dipartimento d'Ingegneria


Written by  Friday, 21 March 2014 08:52

Recurrent neural network for approximate nonnegative matrix factorization

Giovanni Costantini, Renzo Perfetti, Massimiliano Todisco



A recurrent neural network solving the approximate nonnegative matrix factorization (NMF) problem is presented in this paper. The proposed network is based on the Lagrangian approach, and exploits a partial dual method in order to limit the number of dual variables. Sparsity constraints on basis or activation matrices are included by adding a weighted sum of constraint functions to the least squares reconstruction error. However, the corresponding Lagrange multipliers are computed by the network dynamics itself, avoiding empirical tuning or a validation process. It is proved that local solutions of the NMF optimization problem correspond to as many stable steady-state points of the network dynamics. The validity of the proposed approach is verified through several simulation examples concerning both synthetic and real-world datasets for feature extraction and clustering applications.

To be published in Neurocomputing (2014)

Read 39742 times Last modified on Friday, 21 March 2014 08:58
More in this category: « ARTICOLO LIBRO »

Leave a comment

Make sure you enter the (*) required information where indicated. HTML code is not allowed.


This site uses cookies, including third parties, for statistics and to help you navigate web pages. Further information available at the privacy policy page. Learn more

I understand