On the determination of Lagrange multipliers for a weighted Lasso problem using geometric and convex analysis techniques

DOI10.1007/S00245-023-10096-0arXiv2301.09083MaRDI QIDQ6189680FDOQ6189680

Authors: Gianluca Giacchi, Bastien Milani, B. Franceschiello 

Publication date: 8 February 2024

Published in: Applied Mathematics and Optimization (Search for Journal in Brave)

Abstract: Compressed Sensing (CS) comprises a wide range of theoretical and applied techniques to recover signals given a partial knowledge of their coefficients. It finds its applications in several fields, such as mathematics, physics, engineering, and many medical sciences, to name a few. Driven by our interest in the mathematics behind Magnetic Resonance Imaging (MRI) and Compressed Sensing (CS), we use convex analysis techniques to determine analytically the optimal tuning parameters of the space-variant LASSO with voxel-wise weighting, under assumptions on the fidelity term, either on the sign of its gradient or orthogonality-like conditions on its matrix. Finally, we conclude conjecturing what the explicit form of optimal parameters should be in the most general setting (hypotheses-free) of the space-variant LASSO.

Full work available at URL: https://arxiv.org/abs/2301.09083

zbMATH Keywords

convex optimizationcompressed sensingLASSOLagrange dualityMRItuning parameters

Mathematics Subject Classification ID

Numerical solutions to overdetermined systems, pseudoinverses (65F20) Ill-posedness and regularization problems in numerical linear algebra (65F22) Biomedical imaging and signal processing (92C55) Mathematical programming (90Cxx)

Cites Work

• Title not available (Why is that?)
• Title not available (Why is that?)
• Generalized Cross-Validation as a Method for Choosing a Good Ridge Parameter
• Stable recovery of sparse overcomplete representations in the presence of noise
• Robust uncertainty principles: exact signal reconstruction from highly incomplete frequency information
• Compressed sensing
• The Generalized Lasso With Non-Linear Observations
• A mathematical introduction to compressive sensing
• Stein Unbiased GrAdient estimator of the Risk (SUGAR) for Multiple Parameter Selection
• The restricted isometry property and its implications for compressed sensing
• Tikhonov regularization and the L-curve for large discrete ill-posed problems
• Atomic decomposition by basis pursuit
• A continuous exact \(\ell_0\) penalty (CEL0) for least squares regularized problem
• Sharp MSE bounds for proximal denoising
• A Fast Adaptive Parameter Estimation for Total Variation Image Restoration
• Linear Inversion of Band-Limited Reflection Seismograms
• Residual whiteness principle for parameter-free image restoration
• Solving inverse problems using data-driven models
• Convex optimization in sums of Banach spaces
• A Flexible Space-Variant Anisotropic Regularization for Image Restoration with Automated Parameter Selection
• A unifying representer theorem for inverse problems and machine learning
• Weighted LASSO for Sparse Recovery With Statistical Prior Support Information
• Compressive imaging: structure, sampling, learning. With contributions by Vegard Antun
• Generating structured nonsmooth priors and associated primal-dual methods
• Sensitivity of \(\ell_1\) minimization to parameter choice

This page was built for publication: On the determination of Lagrange multipliers for a weighted Lasso problem using geometric and convex analysis techniques

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6189680)