Understanding approximate and unrolled dictionary learning for pattern recovery - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year : 2022

Understanding approximate and unrolled dictionary learning for pattern recovery

Abstract

Dictionary learning consists of finding a sparse representation from noisy data and is a common way to encode data-driven prior knowledge on signals. Alternating minimization (AM) is standard for the underlying optimization, where gradient descent steps alternate with sparse coding procedures. The major drawback of this method is its prohibitive computational cost, making it unpractical on large real-world data sets. This work studies an approximate formulation of dictionary learning based on unrolling and compares it to alternating minimization to find the best trade-off between speed and precision. We analyze the asymptotic behavior and convergence rate of gradients estimates in both methods. We show that unrolling performs better on the support of the inner problem solution and during the first iterations. Finally, we apply unrolling on pattern learning in magnetoencephalography (MEG) with the help of a stochastic algorithm and compare the performance to a state-of-the-art method.
Fichier principal
Vignette du fichier
ICLR_2022___Autodiff_for_dictionary_learning.pdf (1.06 Mo) Télécharger le fichier
Origin Files produced by the author(s)

Dates and versions

hal-03258196 , version 1 (11-06-2021)
hal-03258196 , version 2 (11-06-2021)
hal-03258196 , version 3 (29-01-2022)
hal-03258196 , version 4 (08-02-2022)

Identifiers

  • HAL Id : hal-03258196 , version 4

Cite

Benoît Malézieux, Thomas Moreau, Matthieu Kowalski. Understanding approximate and unrolled dictionary learning for pattern recovery. International Conference on Learning Representations, Apr 2022, Online, France. ⟨hal-03258196v4⟩
255 View
477 Download

Share

Gmail Mastodon Facebook X LinkedIn More