Главная

Yuantao Gu

  1. Performance of Orthogonal Matching Pursuit for Multiple Measurement Vectors.

    Пт, 09/30/2011 - 15:01 — SomNambul
    Authors: Yuantao Gu, Jie Ding, Laming Chen
    Subjects: Information Theory
    Abstract

    In this paper, we consider orthogonal matching pursuit (OMP) algorithm for
    multiple measurement vectors (MMV) problem. The robustness of OMPMMV is studied
    under general perturbations---when the measurement vectors as well as the
    sensing matrix are incorporated with additive noise. The main result shows that
    although exact recovery of the sparse solutions is unrealistic in noisy
    scenario, recovery of the support set of the solutions is guaranteed under
    suitable conditions.

  2. Performance Analysis of Orthogonal Matching Pursuit under General Perturbations.

    Пнд, 06/20/2011 - 15:01 — SomNambul
    Authors: Yuantao Gu, Jie Ding, Laming Chen
    Subjects: Information Theory
    Abstract

    Orthogonal Matching Pursuit (OMP) is the canonical greedy algorithm for
    sparse approximation. Previous studies have mainly considered non-perturbed
    observations $\bm y=\bm \Phi \bm x$, and focused on the exact recovery of $\bm
    x$ through $\bm y$ and $\bm \Phi$. Here, $\bm \Phi$ is a matrix with more
    columns than rows, and $\bm x$ is a sparse signal one wants to recover. In this
    paper, performance of OMP under general perturbations---from both $\bm y$ and
    $\bm \Phi$---is studied, using the Restricted Isometry Property (RIP).

  3. Regularized Least-Mean-Square Algorithms.

    Чт, 12/23/2010 - 16:01 — SomNambul
    Authors: Alfred O. Hero, Yilun Chen, Yuantao Gu
    Subjects: Methodology
    Abstract

    We consider adaptive system identification problems with convex constraints
    and propose a family of regularized Least-Mean-Square (LMS) algorithms. We show
    that with a properly selected regularization parameter the regularized LMS
    provably dominates its conventional counterpart in terms of mean square
    deviations. We establish simple and closed-form expressions for choosing this
    regularization parameter. For identifying an unknown sparse system we propose
    sparse and group-sparse LMS algorithms, which are special examples of the
    regularized LMS family.

Math-arch

RSS-материал