The Overlap Gap Property in Principal Submatrix Recovery

We study support recovery for a principal submatrix with elevated mean , hidden in an symmetric mean zero Gaussian matrix. Here is a universal constant, and we assume for some constant . We establish that {there exists a constant such that} the MLE recovers a constant proportion of the hidden submatrix if , {while such recovery is information theoretically impossible if }. The MLE is computationally intractable in general, and in fact, for sufficiently small, this problem is conjectured to exhibit a \emph{statistical-computational gap}. To provide rigorous evidence for this, we study the likelihood landscape for this problem, and establish that for some and , the problem exhibits a variant of the \emph{Overlap-Gap-Property (OGP)}. As a direct consequence, we establish that a family of local MCMC based algorithms do not achieve optimal recovery. Finally, we establish that for , a simple spectral method recovers a constant proportion of the hidden submatrix.
View on arXiv