Skip to Main content Skip to Navigation
Conference papers

Short and squeezed: accelerating the computation of antisparse representations with safe squeezing

Clément Elvira Cédric Herzet 1
1 SIMSMART - SIMulation pARTiculaire de Modèles Stochastiques
IRMAR - Institut de Recherche Mathématique de Rennes, Inria Rennes – Bretagne Atlantique
Abstract : Antisparse coding aims at spreading the information uniformly over representation coefficients and can be expressed as the solution of an ∞-norm regularized problem. In this paper, we propose a new methodology, coined "safe squeezing", accelerating the computation of antisparse representations. The idea consists in identifying saturated entries of the solution via simple tests and compact their contribution to achieve some form of dimensionality reduction. Numerical experiments show that the proposed approach leads to significant computational gain.
Complete list of metadatas

https://hal.inria.fr/hal-03070344
Contributor : Cédric Herzet <>
Submitted on : Tuesday, December 15, 2020 - 6:07:31 PM
Last modification on : Friday, January 8, 2021 - 3:40:07 AM

File

IcasspAntisparse.pdf
Files produced by the author(s)

Identifiers

Citation

Clément Elvira, Cédric Herzet. Short and squeezed: accelerating the computation of antisparse representations with safe squeezing. ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), May 2020, Barcelona, Spain. pp.5615-5619, ⟨10.1109/ICASSP40776.2020.9053156⟩. ⟨hal-03070344⟩

Share

Metrics

Record views

13

Files downloads

90