Distributed recovery of jointly sparse signals under communication constraints
Fosson, Matamoros, Anton-Haro, Magli · math.OC,cs.MA · 2016-11-08 · 原文
The problem of the distributed recovery of jointly sparse signals has attracted much attention recently. Let us assume that the nodes of a network observe different sparse signals with common support; starting from linear, compressed measurements, and exploiting network communication, each node aims at reconstructing the support and the non-zero values of its observed signal. In the literature, distributed greedy algorithms have been proposed to tackle this problem, among which the most reliable ones require a large amount of transmitted data, which barely adapts to realistic network communication constraints. In this work, we address the problem through a reweighted \ell_1 soft thresholding technique, in which the threshold is iteratively tuned based on the current estimate of the support. The proposed method adapts to constrained networks, as it requires only local communication among neighbors, and the transmitted messages are indices from a finite set. We analytically prove the convergence of the proposed algorithm and we show that it outperforms the state-of-the-art greedy methods in terms of balance between recovery accuracy and communication load.
讲义
讲义·推断 依据「原文」自动生成的结构化摘要（推断），非原文表述；以原文为准。
1. 人话版
The problem of the distributed recovery of jointly sparse signals has attracted much attention recently.
Let us assume that the nodes of a network observe different sparse signals with common support; starting from linear, compressed measurements, and exploiting network communication, each node aims at reconstructing the support and the non-zero values of its observed signal.
2. 领域脉络
本文类目：math.OC、cs.MA，属于其所在研究脉络的最新进展。
3. 机制拆解
摘要未展开方法细节——精读时重点看方法/模型部分。
4. 证据与数字
In this work, we address the problem through a reweighted \ell_1 soft thresholding technique, in which the threshold is iteratively tuned based on the current estimate of the support.
5. 反例与边界
In the literature, distributed greedy algorithms have been proposed to tackle this problem, among which the most reliable ones require a large amount of transmitted data, which barely adapts to realistic network communication constraints.
The proposed method adapts to constrained networks, as it requires only local communication among neighbors, and the transmitted messages are indices from a finite set.
6. 跨领域连接与意外收获
横跨 2 个类目（math.OC、cs.MA），关注其在你兴趣板块间的迁移面。
7. 可复用方法
把本文机制与你手头项目对照，找一个两周内能验证的最小实验。
8. 术语表
精读时把不熟的术语记入此处，作为下次回忆的锚点。