Abstract:We consider an Erdős-Rényi graph $\mathbb{G}(n,p)$ on $n$ vertices with edge probability $p$ such that \[ \sqrt{\frac{\log n}{\log \log n}} \ll np \le n^{1/2-o(1)}, \label{eq:abs} \tag{$\dagger$} \] and derive the upper tail large deviations of $\lambda(\mathbb{G}(n,p))$, the largest eigenvalue of its adjacency matrix. Within this regime we show that, for $p \gg n^{-2/3}$ the $\log$-probability of the upper tail event of $\lambda(\mathbb{G}(n,p))$ equals to that of planting a clique of an appropriate size (upon ignoring smaller order terms), while for $p \ll n^{-2/3}$ the same is given by that of the existence of a high degree vertex. This, in particular, shows an emergence of {\em non-planted localized structure} in the latter regime. We also confirm that in the entire regime \eqref{eq:abs} the large deviation probability is asymptotically approximated by the solution of the mean-field variational problem, and further identify the typical structure of $\mathbb{G}(n,p)$ conditioned on the upper tail event of $\lambda(\mathbb{G}(n,p))$ in a certain sub-regime of $p$. For $p$ such that $\log(np) \gtrsim \log n$ the large deviations of $\lambda(\mathbb{G}(n,p))$ is deduced from those of ${\rm Hom}(C_{2t}, \mathbb{G}(n,p))$, the homomorphism counts of cycle of length $2t$, for $t \ge 3$ and $p$ such that $n^{1/2-o(1)} \ge np \gg n^{1/t}$. In this latter regime the typical structure of $\mathbb{G}(n,p)$ conditioned on the upper tail of ${\rm Hom}(C_{2t}, \mathbb{G}(n,p))$ is identified and asymptotic tightness of the mean-field approximation is also established.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the large - deviation problem of the upper tail of the largest eigenvalue of sparse Erdős - Rényi graphs. Specifically, the paper considers the largest eigenvalue $\lambda(G(n, p))$ of the Erdős - Rényi graph $G(n, p)$, where $n$ is the number of vertices and $p$ is the probability of an edge, and satisfies the following condition: \[ \sqrt{\frac{\log n}{\log \log n}} \ll np \leq n^{1/2 - o(1)} \] Within this range, the paper studies the large - deviation behavior of the upper tail of the largest eigenvalue $\lambda(G(n, p))$. The main results include: 1. **Large - deviation behavior under different sparsities**: - When $p \gg n^{-2/3}$, the large - deviation probability of the upper tail of the largest eigenvalue is the same as the probability of implanting a complete subgraph (clique) of an appropriate size in the graph (ignoring lower - order terms). - When $p \ll n^{-2/3}$, the large - deviation probability of the upper tail of the largest eigenvalue is the same as the probability of the existence of high - degree vertices in the graph. 2. **Solution to the mean - field variational problem**: - Throughout the sparsity range, the large - deviation probability of the upper tail of the largest eigenvalue can be approximated by the solution to the mean - field variational problem. - The paper also determines the typical structure of the graph $G(n, p)$ conditional on the upper - tail event of the largest eigenvalue in certain sub - ranges. 3. **Large - deviation of homomorphism counts in cycle graphs**: - For the case where $p$ satisfies $\log(np) \gtrsim \log n$, the large - deviation of the largest eigenvalue can be derived from the large - deviation of homomorphism counts in cycle graphs $C_{2t}$ of length $2t$. - Within this range, the paper determines the typical structure of the graph $G(n, p)$ conditional on the upper - tail event of homomorphism counts in $C_{2t}$ and establishes the asymptotic tightness of the mean - field approximation. 4. **Rigorousness of the mean - field approximation**: - As a corollary of the main results, the paper proves that throughout the sparsity range, the large - deviation probability of the upper tail of the largest eigenvalue can be asymptotically approximated by the solution to the mean - field variational problem. 5. **Occurrence of non - implanted optimizer selection**: - In certain sub - ranges, the optimizer is selected non - implanted, which is different from previous research results. The main contribution of the paper lies in the detailed analysis of the large - deviation behavior of the upper tail of the largest eigenvalue of sparse Erdős - Rényi graphs, and provides the corresponding theoretical framework and proof methods. These results are of great significance for understanding large - deviation phenomena in random graph models.

Upper tail of the spectral radius of sparse Erdős-Rényi graphs

Spectra of adjacency and Laplacian matrices of Erdős-Rényi hypergraphs

Edge universality of sparse Erdős-Rényi digraphs

Largest eigenvalue statistics of sparse random adjacency matrices

Large deviation principle for the norm of the Laplacian matrix of inhomogeneous Erdős-Rényi random graphs

Eigenvalues outside the bulk of inhomogeneous Erdős-Rënyi random graphs

Random matrices with row constraints and eigenvalue distributions of graph Laplacians

Spectral radii of sparse random matrices

Large deviations for the largest singular value of sparse non-Hermitian matrices

The spectral gap of dense random regular graphs

The Spectral Edge of Constant Degree Erdős-Rényi Graphs

Large deviations of the empirical spectral measure of supercritical sparse Wigner matrices

Limiting Spectra of inhomogeneous random graphs

Local law and Tracy–Widom limit for sparse random matrices

Concentration and regularization of random graphs

Phase transition of degenerate Turán problems in $p$-norms

Large deviations of the largest eigenvalue of supercritical sparse Wigner matrices

Moderate deviations of triangle counts in sparse Erdős-Rényi random graphs $G(n,m)$ and $G(n,p)$

Twin-width of random graphs

On the rank, Kernel, and core of sparse random graphs

On the Smallest Eigenvalue of Grounded Laplacian Matrices