23
84

Improved Sum-of-Squares Lower Bounds for Hidden Clique and Hidden Submatrix Problems

Abstract

Given a large data matrix ARn×nA\in\mathbb{R}^{n\times n}, we consider the problem of determining whether its entries are i.i.d. with some known marginal distribution AijP0A_{ij}\sim P_0, or instead AA contains a principal submatrix AQ,QA_{{\sf Q},{\sf Q}} whose entries have marginal distribution AijP1P0A_{ij}\sim P_1\neq P_0. As a special case, the hidden (or planted) clique problem requires to find a planted clique in an otherwise uniformly random graph. Assuming unbounded computational resources, this hypothesis testing problem is statistically solvable provided QClogn|{\sf Q}|\ge C \log n for a suitable constant CC. However, despite substantial effort, no polynomial time algorithm is known that succeeds with high probability when Q=o(n)|{\sf Q}| = o(\sqrt{n}). Recently Meka and Wigderson \cite{meka2013association}, proposed a method to establish lower bounds within the Sum of Squares (SOS) semidefinite hierarchy. Here we consider the degree-44 SOS relaxation, and study the construction of \cite{meka2013association} to prove that SOS fails unless kCn1/3/lognk\ge C\, n^{1/3}/\log n. An argument presented by Barak implies that this lower bound cannot be substantially improved unless the witness construction is changed in the proof. Our proof uses the moments method to bound the spectrum of a certain random association scheme, i.e. a symmetric random matrix whose rows and columns are indexed by the edges of an Erd\"os-Renyi random graph.

View on arXiv
Comments on this paper