English

Broadcasting on trees near criticality

Information Theory 2020-05-19 v1 math.IT

Abstract

We revisit the problem of broadcasting on dd-ary trees: starting from a Bernoulli(1/2)(1/2) random variable X0X_0 at a root vertex, each vertex forwards its value across binary symmetric channels BSCδ\mathrm{BSC}_\delta to dd descendants. The goal is to reconstruct X0X_0 given the vector XLhX_{L_h} of values of all variables at depth hh. It is well known that reconstruction (better than a random guess) is possible as hh\to \infty if and only if δ<δc(d)\delta < \delta_c(d). In this paper, we study the behavior of the mutual information and the probability of error when δ\delta is slightly subcritical. The innovation of our work is application of the recently introduced "less-noisy" channel comparison techniques. For example, we are able to derive the positive part of the phase transition (reconstructability when δ<δc\delta<\delta_c) using purely information-theoretic ideas. This is in contrast with previous derivations, which explicitly analyze distribution of the Hamming weight of XLhX_{L_h} (a so-called Kesten-Stigum bound).

Keywords

Cite

@article{arxiv.2005.07801,
  title  = {Broadcasting on trees near criticality},
  author = {Yuzhou Gu and Hajir Roozbehani and Yury Polyanskiy},
  journal= {arXiv preprint arXiv:2005.07801},
  year   = {2020}
}

Comments

Accepted to 2020 IEEE International Symposium on Information Theory (ISIT 2020)

R2 v1 2026-06-23T15:35:03.580Z