Eigenvalues of block structured asymmetric random matrices

Abstract

We study the spectrum of an asymmetric random matrix with block structured variances. The rows and columns of the random square matrix are divided into $D$ partitions with arbitrary size (linear in $N$). The parameters of the model are the variances of elements in each block, summarized in $g\in {\mathbb{R}}_{+}^{D\times D}$. Using the Hermitization approach and by studying the matrix-valued Stieltjes transform we show that these matrices have a circularly symmetric spectrum, we give an explicit formula for their spectral radius and a set of implicit equations for the full density function. We discuss applications of this model to neural networks.

References in corpus (3)

• Transition to chaos in random networks with cell-type-specific connectivity
• Products of independent elliptic random matrices
• Eigenvalue spectra of asymmetric random matrices for multi-component neural networks

Cited by in corpus (9)

• Correlations between synapses in pairs of neurons slow down dynamics in randomly connected neural networks
• On the low dimensional dynamics of structured random networks
• Non-Hermitian random matrices with a variance profile (I): Deterministic equivalents and limiting ESDs
• Eigenvalue spectra of large correlated random matrices
• Consequences of Dale's law on the stability-complexity relationship of random neural networks
• The circular law for random regular digraphs with random edge weights
• The Spectral Boundary of Block Structured Random Matrices
• Circular Law for Random Block Band Matrices with Genuinely Sublinear Bandwidth
• Singular value distribution of dense random matrices with block Markovian dependence