English

Comparing Poisson and Gaussian channels (extended)

Information Theory 2023-06-30 v1 math.IT

Abstract

Consider a pair of input distributions which after passing through a Poisson channel become ϵ\epsilon-close in total variation. We show that they must necessarily then be ϵ0.5+o(1)\epsilon^{0.5+o(1)}-close after passing through a Gaussian channel as well. In the opposite direction, we show that distributions inducing ϵ\epsilon-close outputs over the Gaussian channel must induce ϵ1+o(1)\epsilon^{1+o(1)}-close outputs over the Poisson. This quantifies a well-known intuition that ''smoothing'' induced by Poissonization and Gaussian convolution are similar. As an application, we improve a recent upper bound of Han-Miao-Shen'2021 for estimating mixing distribution of a Poisson mixture in Gaussian optimal transport distance from n0.1+o(1)n^{-0.1 + o(1)} to n0.25+o(1)n^{-0.25 + o(1)}.

Cite

@article{arxiv.2306.16735,
  title  = {Comparing Poisson and Gaussian channels (extended)},
  author = {Anzo Teh and Yury Polyanskiy},
  journal= {arXiv preprint arXiv:2306.16735},
  year   = {2023}
}

Comments

7 pages, 2 figures, to appear in the 2023 IEEE International Symposium on Information Theory (ISIT)

R2 v1 2026-06-28T11:17:37.752Z