Low-Degree Hardness of Detection for Correlated Erdős-Rényi Graphs

Abstract: Given two Erd\H{o}s-R\'enyi graphs with $n$ vertices whose edges are correlated through a latent vertex correspondence, we study complexity lower bounds for the associated correlation detection problem for the class of low-degree polynomial algorithms. We provide evidence that any degree-$O(\rho<sup>{-1})$ polynomial algorithm fails for detection, where $\rho$ is the edge correlation. Furthermore, in the sparse regime where the edge density $q=n<sup>{-1+o(1)}$, we provide evidence that any degree-$d$ polynomial algorithm fails for detection, as long as $$\log d=o\big( \frac{\log n}{\log nq} \wedge \sqrt{\log n} \big)$$ and the correlation $\rho&lt;\sqrt{\alpha}$ where $\alpha\approx 0.338$ is the Otter's constant. Our result suggests that several state-of-the-art algorithms on correlation detection and exact matching recovery may be essentially the best possible.