An Iterative Discrete Least Square Estimator with Dynamic Parameterization via Deep-Unfolding

We propose a new dynamic parameterization approach via deep unfolding as an extension of the recently-introduced iterative discrete least square (IDLS) scheme, shown to elegantly generalize the conventional linear minimum mean squared error (LMMSE) method to enable the solution of inversion problems...

Full description

Saved in:
Bibliographic Details
Published inConference record - Asilomar Conference on Signals, Systems, & Computers pp. 32 - 36
Main Authors Ando, Kengo, Iimori, Hiroki, Rou, Hyeon Seok, Freitas de Abreu, Giuseppe Thadeu, G, David Gonzalez, Gonsa, Osvaldo
Format Conference Proceeding
LanguageEnglish
Published IEEE 31.10.2022
Subjects
Computer simulation
Deep learning
deep unfolding
Discrete inversion
Heuristic algorithms
least square
Linear systems
multidimensional linear systems
Neural networks
Symbols
Online AccessGet full text

Cover

More Information
Summary:We propose a new dynamic parameterization approach via deep unfolding as an extension of the recently-introduced iterative discrete least square (IDLS) scheme, shown to elegantly generalize the conventional linear minimum mean squared error (LMMSE) method to enable the solution of inversion problems in complex multidimensional linear systems subject to discrete inputs. Configuring a layer-wise structure analogous to a deep neural network, the new approach enables an efficient optimization of the iterative IDLS algorithm, by finding optimal hyper-parameters for the related optimization problem through backpropagation and stochastic gradient descent techniques. The effectiveness of the proposed approach is confirmed via computer simulations.
ISSN:2576-2303
DOI:10.1109/IEEECONF56349.2022.10051860