The sparse parity matrix

TL;DR

This paper investigates the overlap of solutions to sparse random parity matrices over GF(2), revealing a phase transition at d=e where the overlap behavior shifts from concentration to bifurcation, impacting theories of random CSPs and inference.

Contribution

It establishes the first rigorous analysis of the overlap distribution for solutions of sparse random parity matrices, identifying a phase transition at d=e with distinct behaviors.

Findings

For d<e, the overlap concentrates on a single value.

For d>e, the overlap conditioned on A concentrates, but unconditionally it bifurcates.

The results reveal a phase transition at d=e affecting solution structure.

Abstract

Let $\mathbf{A}$ be an $n\times n$-matrix over $\mathbb{F}_2$ whose every entry equals $1$ with probability $d/n$ independently for a fixed $d>0$. Draw a vector $\mathbf{y}$ randomly from the column space of $\mathbf{A}$. It is a simple observation that the entries of a random solution $\mathbf{x}$ to $\mathbf{A} x=\mathbf{y}$ are asymptotically pairwise independent, i.e., $\sum_{i<j}\mathbb{E}|\mathbb{P}[\mathbf{x}_i=s,\,\mathbf{x}_j=t\mid\mathbf{A}]-\mathbb{P}[\mathbf{x}_i=s\mid\mathbf{A}]\mathbb{P}[\mathbf{x}_j=t\mid\mathbf{A}]|=o(n^2)$ for $s,t\in\mathbb{F}_2$. But what can we say about the {\em overlap} of two random solutions $\mathbf{x},\mathbf{x}'$, defined as $n^{-1}\sum_{i=1}^n\mathbf{1}\{\mathbf{x}_i=\mathbf{x}_i'\}$? We prove that for $d<\mathrm{e}$ the overlap concentrates on a single deterministic value $\alpha_*(d)$. By contrast, for $d>\mathrm{e}$ the overlap concentrates on a single value once we condition on the matrix $\mathbf{A}$, while over the probability space of $\mathbf{A}$ its conditional expectation vacillates between two different values $\alpha_*(d)<\alpha^*(d)$, either of which occurs with probability $1/2+o(1)$. This bifurcated non-concentration result provides an instructive contribution to both the theory of random constraint satisfaction problems and of inference problems on random structures.