Is the 1-norm the best convex sparse regularization?

Abstract : The 1-norm is a good convex regularization for the recovery of sparse vectors from under-determined linear measurements. No other convex regularization seems to surpass its sparse recovery performance. How can this be explained? To answer this question, we define several notions of " best " (convex) regulariza-tion in the context of general low-dimensional recovery and show that indeed the 1-norm is an optimal convex sparse regularization within this framework.
Complete list of metadatas

Cited literature [8 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-01819219
Contributor : Yann Traonmilin <>
Submitted on : Wednesday, June 20, 2018 - 10:26:56 AM
Last modification on : Friday, September 13, 2019 - 9:50:02 AM
Long-term archiving on : Monday, September 24, 2018 - 2:39:30 PM

Files

optimal_reg_abstract_hal.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01819219, version 1
  • ARXIV : 1806.08690

Citation

Yann Traonmilin, Samuel Vaiter, Rémi Gribonval. Is the 1-norm the best convex sparse regularization?. iTWIST'18 - international Traveling Workshop on Interactions between low-complexity data models and Sensing Techniques, Nov 2018, Marseille, France. pp.1-11. ⟨hal-01819219⟩

Share

Metrics

Record views

837

Files downloads

134