Upper tail of the spectral radius of sparse Erdős-Rényi graphs
arXiv:2109.06242
Abstract
We consider an Erdős-Rényi graph on vertices with edge probability such that \[ \sqrt{\frac{\log n}{\log \log n}} \ll np \le n^{1/2-o(1)}, \label{eq:abs} \tag{} \] and derive the upper tail large deviations of , the largest eigenvalue of its adjacency matrix. Within this regime we show that, for the -probability of the upper tail event of equals to that of planting a clique of an appropriate size (upon ignoring smaller order terms), while for 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 conditioned on the upper tail event of in a certain sub-regime of . For such that the large deviations of is deduced from those of , the homomorphism counts of cycle of length , for and such that . In this latter regime the typical structure of conditioned on the upper tail of is identified and asymptotic tightness of the mean-field approximation is also established.
arXiv commentMinor changes in presentation