Abstract
In this paper, we aim to extend Nonnegative Matrix Factorization with Nesterov iterations (Ne-NMF)—well-suited to large-scale problems—to the situation when some entries are missing in the observed matrix. In particular, we investigate the Weighted and Expectation-Maximization strategies which both provide a way to process missing data. We derive their associated extensions named W-NeNMF and EM-W-NeNMF, respectively. The proposed approaches are then tested on simulated nonnegative low-rank matrix completion problems where the EM-W-NeNMF is shown to outperform state-of-the-art methods and the W-NeNMF technique.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
See for example the CPU-time consumptions of [27] at https://github.com/andrewssobral/lrslibrary.
- 2.
As explained above, the standard approach was shown to be slow to converge in [32], because many MUs were applied at each M-step. To prevent such an issue, at each M-stage, we here run one NMF update per matrix factor, i.e., one MU.
- 3.
References
Candès, E.J., Plan, Y.: Matrix completion with noise. Proc. IEEE 98(6), 925–936 (2010)
Candès, E.J., Recht, B.: Exact matrix completion via convex optimization. Found. Comput. Math. 9(6), 717–772 (2009)
Choo, J., Lee, C., Reddy, C.K., Park, H.: Weakly supervised nonnegative matrix factorization for user-driven clustering. Data Min. Knowl. Disc. 29(6), 1598–1621 (2015)
Chreiky, R., Delmaire, G., Puigt, M., Roussel, G., Courcot, D., Abche, A.: Split gradient method for informed non-negative matrix factorization. In: Vincent, E., Yeredor, A., Koldovský, Z., Tichavský, P. (eds.) Proceedings of LVA/ICA, pp. 376–383 (2015)
Dorffer, C., Puigt, M., Delmaire, G., Roussel, G.: Blind calibration of mobile sensors using informed nonnegative matrix factorization. In: Vincent, E., Yeredor, A., Koldovský, Z., Tichavský, P. (eds.) Proceedings of LVA/ICA, pp. 497–505 (2015)
Dorffer, C., Puigt, M., Delmaire, G., Roussel, G.: Blind mobile sensor calibration using an informed nonnegative matrix factorization with a relaxed rendezvous model. In: Proceedings of ICASSP, pp. 2941–2945 (2016)
Dorffer, C., Puigt, M., Delmaire, G., Roussel, G.: Nonlinear mobile sensor calibration using informed semi-nonnegative matrix factorization with a Vandermonde factor. In: Proceedings of SAM (2016)
Fazel, M.: Matrix rank minimization with applications. Ph.D. thesis, Stanford University (2002)
Guan, N., Tao, D., Luo, Z., Yuan, B.: NeNMF: an optimal gradient method for nonnegative matrix factorization. IEEE Trans. Signal Proc. 60(6), 2882–2898 (2012)
Guillamet, D., Vitrià, J., Schiele, B.: Introducing a weighted non-negative matrix factorization for image classification. Pattern Recogn. Lett. 24(14), 2447–2454 (2003)
Haldar, J.P., Hernando, D.: Rank-constrained solutions to linear matrix equations using powerfactorization. IEEE Signal Process. Lett. 16(7), 584–587 (2009)
Hamon, R., Emiya, V., Févotte, C.: Convex nonnegative matrix factorization with missing data. In: Procceedings of MLSP (2016)
Ho, N.D.: Nonnegative matrix factorizations algorithms and applications. Ph.D. thesis, Université Catholique de Louvain (2008)
Hoyer, P.O.: Non-negative matrix factorization with sparseness constraints. J. Mach. Learn. Res. 5, 1457–1469 (2004)
Kim, H., Park, H.: Nonnegative matrix factorization based on alternating nonnegativity constrained least squares and active set method. SIAM J. Matrix Anal. Appl. 30(2), 713–730 (2008)
Kim, Y.D., Choi, S.: Weighted nonnegative matrix factorization. In: Proceedings of ICASSP, pp. 1541–1544, April 2009
Kumar, R., Da Silva, C., Akalin, O., Aravkin, A.Y., Mansour, H., Recht, B., Herrmann, F.J.: Efficient matrix completion for seismic data reconstruction. Geophysics 80(5), V97–V114 (2015)
Lantéri, H., Theys, C., Richard, C., Févotte, C.: Split gradient method for nonnegative matrix factorization. In: Proceedings of EUSIPCO (2010)
Lee, D.D., Seung, H.S.: Algorithms for non-negative matrix factorization. In: NIPS, pp. 556–562 (2001)
Limem, A., Delmaire, G., Puigt, M., Roussel, G., Courcot, D.: Non-negative matrix factorization under equality constraints–a study of industrial source identification. Appl. Numer. Math. 85, 1–15 (2014)
Limem, A., Puigt, M., Delmaire, G., Roussel, G., Courcot, D.: Bound constrained weighted NMF for industrial source apportionment. In: Proceedings of MLSP (2014)
Lin, C.: Projected gradient methods for nonnegative matrix factorization. Neural Comput. 19(10), 2756–2779 (2007)
Liu, C., Yang, H.C., Fan, J., He, L.W., Wang, Y.M.: Distributed nonnegative matrix factorization for web-scale dyadic data analysis on MapReduce. In: Proceedings of WWW Conference, April 2010
Mairal, J., Bach, F., Ponce, J., Sapiro, G.: Online learning for matrix factorization and sparse coding. J. Mach. Learn. Res. 11, 19–60 (2010)
Nesterov, Y.: A method of solving a convex programming problem with convergence rate O(1/k2). Sov. Math. Doklady 27, 372–376 (1983)
Savvaki, S., Tsagkatakis, G., Panousopoulou, A., Tsakalides, P.: Application of matrix completion on water treatment data. In: Proceedings of CySWater, pp. 3:1–3:6 (2015)
Sobral, A., Bouwmans, T., Zahzah, E.: LRSLibrary: low-rank and sparse tools for background modeling and subtraction in videos. In: Robust Low-Rank and Sparse Matrix Decomposition: Applications in Image and Video Processing. CRC Press, Taylor and Francis Group
Srebro, N., Jaakkola, T.: Weighted low-rank approximations. In: Proceedings of ICML (2003)
Tepper, M., Sapiro, G.: Compressed nonnegative matrix factorization is fast and accurate. IEEE Trans. Signal Process. 64(9), 2269–2283 (2016)
Wang, Y.X., Zhang, Y.J.: Nonnegative matrix factorization: a comprehensive review. IEEE Trans. Knowl. Data Eng. 25(6), 1336–1353 (2013)
Wen, Z., Yin, W., Zhang, Y.: Solving a low-rank factorization model for matrix completion by a nonlinear successive over-relaxation algorithm. Math. Program. Comput. 4(4), 333–361 (2012)
Zhang, S., Wang, W., Ford, J., Makedon, F.: Learning from incomplete ratings using non-negative matrix factorization, chap. 58, pp. 549–553 (2006)
Zhou, G., Cichocki, A., Xie, S.: Fast nonnegative matrix/tensor factorization based on low-rank approximation. IEEE Trans. Signal Process. 60(6), 2928–2940 (2012)
Acknowledgments
This work was funded by the “OSCAR” project within the Région Hauts-de-France “Chercheurs Citoyens” Program.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
A Proof of Lemmas 1 and 2
A Proof of Lemmas 1 and 2
Following the structure of the proof in [9], the proof of Lemma 1 is shown by noticing that
Let us now focus on the proof of Lemma 2. The Lipschitz continuity of \(\nabla _F \mathcal {J}_W\) can be shown by extending the proof in [9] to the weighted situation considered in this paper. However, the key point lies in the estimation of the “best” Lipschitz constant. Indeed, let us first recall that Q is a Lipschitz constant of \(\nabla _F \mathcal {J}_W (G,.)\) if, for any matrices \(F_1\) and \(F_2\),
From Eq. (7), it is obvious that the larger is the majoring constant Q, the smaller is the associated gradient step size and thus the convergence speed.
Considering the gradient expression (14), we derive
The singular value decomposition of G yields
At this stage, we need to put W and G out of the norm to obtain a Lipschitz constant. This can be done by assuming that any column and any row of M contains at least one element—i.e., any column and row of W contains at least one 1—and thus noticing that
which provides, using another singular value decomposition of G:
i.e., the Lipschitz constant for W-NeNMF is L.
Rights and permissions
Copyright information
© 2017 Springer International Publishing AG
About this paper
Cite this paper
Dorffer, C., Puigt, M., Delmaire, G., Roussel, G. (2017). Fast Nonnegative Matrix Factorization and Completion Using Nesterov Iterations. In: Tichavský, P., Babaie-Zadeh, M., Michel, O., Thirion-Moreau, N. (eds) Latent Variable Analysis and Signal Separation. LVA/ICA 2017. Lecture Notes in Computer Science(), vol 10169. Springer, Cham. https://doi.org/10.1007/978-3-319-53547-0_3
Download citation
DOI: https://doi.org/10.1007/978-3-319-53547-0_3
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-53546-3
Online ISBN: 978-3-319-53547-0
eBook Packages: Computer ScienceComputer Science (R0)