Non-ergodicity for the noisy majority vote process on trees
Jian Ding, Fenglin Huang
TL;DR
The paper analyzes the noisy majority vote process on infinite regular trees $\mathbb T_d$ with $d\ge 3$ and proves non-ergodicity (existence of multiple equilibrium measures) for small noise $\epsilon$, extending prior results that were restricted to $d\ge 5$. The authors develop an induction framework based on backward update histories and cluster decompositions, bounding the probability that a set of odd vertices becomes $-1$ by a product over non-adjacent odd clusters with penalties for trifurcations. Central to the argument are new combinatorial lemmas that bound the enumeration of induced configurations (via notions like single/double trifurcations and odd/even clusters) and an extension to general trees through a minus-biased vote process that is dominated by the original dynamics. The results thus establish non-ergodicity for all $d\ge 3$ and small $\epsilon$, highlighting how disorder persists and prevents unique long-term behavior on trees even under modest stochastic perturbations. These insights advance the understanding of ergodicity-breaking phenomena in noisy spin systems on hierarchical graphs and extend Burris–Gray-type results beyond the previously known regimes.
Abstract
We consider the noisy majority vote process on infinite regular trees with degree $d\geq 3$, and we prove the non-ergodicity, i.e., there exist multiple equilibrium measures. Our work extends a result of Bramson and Gray (2021) for $d\geq 5$.
