Stochastic Subsampling for Factorizing Huge Matrices - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Journal Articles IEEE Transactions on Signal Processing Year : 2018

Stochastic Subsampling for Factorizing Huge Matrices

Abstract

We present a matrix-factorization algorithm that scales to input matrices with both huge number of rows and columns. Learned factors may be sparse or dense and/or non-negative, which makes our algorithm suitable for dictionary learning, sparse component analysis, and non-negative matrix factorization. Our algorithm streams matrix columns while subsampling them to iteratively learn the matrix factors. At each iteration, the row dimension of a new sample is reduced by subsampling, resulting in lower time complexity compared to a simple streaming algorithm. Our method comes with convergence guarantees to reach a stationary point of the matrix-factorization problem. We demonstrate its efficiency on massive functional Magnetic Resonance Imaging data (2 TB), and on patches extracted from hyperspectral images (103 GB). For both problems, which involve different penalties on rows and columns, we obtain significant speed-ups compared to state-of-the-art algorithms.
Fichier principal
Vignette du fichier
modl_tsp.pdf (2.29 Mo) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-01431618 , version 1 (11-01-2017)
hal-01431618 , version 2 (26-07-2017)
hal-01431618 , version 3 (30-10-2017)

Identifiers

Cite

Arthur Mensch, Julien Mairal, Bertrand Thirion, Gaël Varoquaux. Stochastic Subsampling for Factorizing Huge Matrices. IEEE Transactions on Signal Processing, 2018, 66 (1), pp.113-128. ⟨10.1109/TSP.2017.2752697⟩. ⟨hal-01431618v3⟩
2671 View
1434 Download

Altmetric

Share

Gmail Facebook X LinkedIn More