arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2210.14263eess.SP
收藏

基于Gershgorin圆盘对齐的高效有向图采样

Efficient Directed Graph Sampling via Gershgorin Disc Alignment

Yuejiang Li, Hong Vicky Zhao, Gene Cheung

摘要:

图采样是通过采样矩阵$\mathbf{H} \in \{0,1\}^{K \times N}$选择一个节点子集来收集样本$\mathbf{y} = \mathbf{H} \mathbf{x} \in \mathbb{R}^K$、$K < N$,从而可以高保真地重建目标信号$\mathbf{x} \in \mathbb{R}^N$的问题。虽然无向图上的采样得到了很好的研究,我们提出了第一个专门为有向图定制的采样方案,利用以前的无向图采样方法的基础上Gershgorin盘对齐(GDAS)。具体地,给定由随机游走图拉普拉斯矩阵$\mathbf{L}_{rw}$指定的有向正图$\mathcal{G}^d$,我们首先定义使用图移位变化(GSV)$\|\mathbf{L}_{rw} \mathbf{x}\|^2_2$作为信号先验从样本$\mathbf{y}$重构平滑信号$\mathbf{x}^*$。为了最小化具有对称系数矩阵$\mathbf{C} = \mathbf{H}^\top \mathbf{H} + μ\mathbf{L}_{rw}^\top \mathbf{L}_{rw}$的线性系统解$\mathbf{x}^* = \mathbf{C}^{-1} \mathbf{H}^\top \mathbf{y}$的最坏情况重构误差,采样目标是选择$\mathbf{H}$以最大化$\mathbf{C}$的最小特征值$λ_{\min}(\mathbf{C})$。为了规避特征分解完全,我们最大化,而不是下限$λ^-_{\min}(\mathbf{S}\mathbf{C}\mathbf{S}^{-1})$的$λ_{\min}(\mathbf{C})$-最小的Gershgorin盘的$\mathbf{C}$的相似性变换的左端-通过一个变种的GDAS的基础上Gershgorin圆定理(GCT)。实验结果表明,我们的采样方法产生更小的信号重建误差在更快的速度相比,竞争计划。

英文摘要:

Graph sampling is the problem of choosing a node subset via sampling matrix $\mathbf{H} \in \{0,1\}^{K \times N}$ to collect samples $\mathbf{y} = \mathbf{H} \mathbf{x} \in \mathbb{R}^K$, $K < N$, so that the target signal $\mathbf{x} \in \mathbb{R}^N$ can be reconstructed in high fidelity. While sampling on undirected graphs is well studied, we propose the first sampling scheme tailored specifically for directed graphs, leveraging a previous undirected graph sampling method based on Gershgorin disc alignment (GDAS). Concretely, given a directed positive graph $\mathcal{G}^d$ specified by random-walk graph Laplacian matrix $\mathbf{L}_{rw}$, we first define reconstruction of a smooth signal $\mathbf{x}^*$ from samples $\mathbf{y}$ using graph shift variation (GSV) $\|\mathbf{L}_{rw} \mathbf{x}\|^2_2$ as a signal prior. To minimize worst-case reconstruction error of the linear system solution $\mathbf{x}^* = \mathbf{C}^{-1} \mathbf{H}^\top \mathbf{y}$ with symmetric coefficient matrix $\mathbf{C} = \mathbf{H}^\top \mathbf{H} + μ\mathbf{L}_{rw}^\top \mathbf{L}_{rw}$, the sampling objective is to choose $\mathbf{H}$ to maximize the smallest eigenvalue $λ_{\min}(\mathbf{C})$ of $\mathbf{C}$. To circumvent eigen-decomposition entirely, we maximize instead a lower bound $λ^-_{\min}(\mathbf{S}\mathbf{C}\mathbf{S}^{-1})$ of $λ_{\min}(\mathbf{C})$ -- smallest Gershgorin disc left-end of a similarity transform of $\mathbf{C}$ -- via a variant of GDAS based on Gershgorin circle theorem (GCT). Experimental results show that our sampling method yields smaller signal reconstruction errors at a faster speed compared to competing schemes.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑