Plug-And-Play Learned Gaussian-mixture Approximate Message Passing

  title={Plug-And-Play Learned Gaussian-mixture Approximate Message Passing},
  author={Osman Musa and Peter Jung and G. Caire},
  journal={ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},
  • O. MusaP. JungG. Caire
  • Published 18 November 2020
  • Computer Science
  • ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
Deep unfolding showed to be a very successful approach for accelerating and tuning classical signal processing algorithms. In this paper, we propose learned Gaussian-mixture AMP (L-GM-AMP) - a plug-and-play compressed sensing (CS) recovery algorithm suitable for any i.i.d. source prior. Our algorithm builds upon Borgerding’s learned AMP (LAMP), yet significantly improves it by adopting a universal denoising function within the algorithm. The robust and flexible denoiser is a byproduct of… 
2 Citations

Figures from this paper

Deep Unfolding of Iteratively Reweighted ADMM for Wireless RF Sensing

This work proposes a non-convex approach based on the iteratively reweighted nuclear and ℓ1-norm (a double-reweighted approach) to obtain a higher accuracy compared to the conventional nuclear norm and ™1- norm minimization for defect detection.

Algorithm Unfolding for Block-sparse and MMV Problems with Reduced Training Overhead

This paper considers algorithm unfolding for the Multiple Measurement Vector (MMV) problem in the case where only few training samples are available and evaluates the unfolded algorithms in numerical experiments and discusses connections to other sparse recovering algorithms.



AMP-Inspired Deep Networks for Sparse Linear Inverse Problems

This paper proposes two novel neural-network architectures that decouple prediction errors across layers in the same way that the approximate message passing (AMP) algorithms decouple them across iterations: through Onsager correction.

Asymptotic Analysis of Complex LASSO via Complex Approximate Message Passing (CAMP)

The approximate message passing (AMP) algorithm is extended to solve the complex-valued LASSO problem and obtained the complex approximate message passed algorithm (CAMP), and the state evolution framework recently introduced for the analysis of AMP is generalized to the complex setting.

Optimally Tuned Iterative Reconstruction Algorithms for Compressed Sensing

  • A. MalekiD. Donoho
  • Computer Science
    IEEE Journal of Selected Topics in Signal Processing
  • 2010
It is shown that the phase transition is a well-defined quantity with the suite of random underdetermined linear systems chosen, and the optimally tuned algorithms dominate such proposals.

The Dynamics of Message Passing on Dense Graphs, with Applications to Compressed Sensing

This paper proves that indeed it holds asymptotically in the large system limit for sensing matrices with independent and identically distributed Gaussian entries, and provides rigorous foundation to state evolution.

Message passing algorithms for compressed sensing: II. analysis and validation

The state evolution formalism for analyzing these algorithms, and some of the conclusions that can be drawn from this formalism are described, and a few representative results are presented.

Message passing algorithms for compressed sensing: I. motivation and construction

The present paper outlines the derivation of AMP from standard sum-product belief propagation, and its extension in several directions, and discusses relations with formal calculations based on statistical mechanics methods.

Learning Fast Approximations of Sparse Coding

Two versions of a very fast algorithm that produces approximate estimates of the sparse code that can be used to compute good visual features, or to initialize exact iterative algorithms are proposed.

A Review of Sparse Recovery Algorithms

A comprehensive study and a state-of-the-art review of compressive sensing theory algorithms used in imaging, radar, speech recognition, and data acquisition and some open research challenges are presented.

Graphical Models Concepts in Compressed Sensing

Surprisingly, the analysis of fast approximate message passing algorithms allows to prove exact high-dimensional limit results for the LASSO risk.

Algorithm Unrolling: Interpretable, Efficient Deep Learning for Signal and Image Processing

The increasing popularity of unrolled deep networks is due, in part, to their potential in developing efficient, high-performance (yet interpretable) network architectures from reasonably sized training sets.