arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1707.08461math.PR
收藏

随机矩阵特征向量的离域化。课堂讲稿

Delocalization of eigenvectors of random matrices. Lecture notes

Mark Rudelson

摘要:

令$x \in S^{n-1}$是$n \times n$随机矩阵的单位特征向量。如果这个向量在实球或复球上大致均匀分布,则它是离域的。这个直观的概念可以用各种方式量化。在这些讲座中,我们将集中在无隙离域。这种类型的离域意味着在高概率情况下,$x$的支撑物的任何不可忽略的子集携带不可忽略的质量。证明无间隙离域性需要建立随机向量投影的小球概率界。使用傅里叶变换,我们将证明这样的界限在一个简单的情况下,随机向量具有独立的坐标的有界密度。这将使我们能够推导出具有有界密度的随机项的矩阵的无间隙离域化。在最后一节中,我们将讨论离域化在Erdens-Rényi随机图的谱性质中的应用。

英文摘要:

Let $x \in S^{n-1}$ be a unit eigenvector of an $n \times n$ random matrix. This vector is delocalized if it is distributed roughly uniformly over the real or complex sphere. This intuitive notion can be quantified in various ways. In these lectures, we will concentrate on the no-gaps delocalization. This type of delocalization means that with high probability, any non-negligible subset of the support of $x$ carries a non-negligible mass. Proving the no-gaps delocalization requires establishing small ball probability bounds for the projections of random vector. Using Fourier transform, we will prove such bounds in a simpler case of a random vector having independent coordinates of a bounded density. This will allow us to derive the no-gaps delocalization for matrices with random entries having a bounded density. In the last section, we will discuss the applications of delocalization to the spectral properties of Erdős-Rényi random graphs.

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。

↑