Saved in:
Bibliographic Details
Main Author: Mendo, Luis
Format: Preprint
Published: 2016
Subjects:
Online Access:https://arxiv.org/abs/1612.08923
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913583962521600
author Mendo, Luis
author_facet Mendo, Luis
contents Given a sequence of independent Bernoulli variables with unknown parameter $p$, and a function $f$ expressed as a power series with non-negative coefficients that sum to at most $1$, an algorithm is presented that produces a Bernoulli variable with parameter $f(p)$. In particular, the algorithm can simulate $f(p)=p^a$, $a\in(0,1)$. For functions with a derivative growing at least as $f(p)/p$ for $p\rightarrow 0$, the average number of inputs required by the algorithm is asymptotically optimal among all simulations that are fast in the sense of Nacu and Peres. A non-randomized version of the algorithm is also given. Some extensions are discussed.
format Preprint
id arxiv_https___arxiv_org_abs_1612_08923
institution arXiv
publishDate 2016
record_format arxiv
spellingShingle An asymptotically optimal Bernoulli factory for certain functions that can be expressed as power series
Mendo, Luis
Statistics Theory
65C50, 62L
G.3
Given a sequence of independent Bernoulli variables with unknown parameter $p$, and a function $f$ expressed as a power series with non-negative coefficients that sum to at most $1$, an algorithm is presented that produces a Bernoulli variable with parameter $f(p)$. In particular, the algorithm can simulate $f(p)=p^a$, $a\in(0,1)$. For functions with a derivative growing at least as $f(p)/p$ for $p\rightarrow 0$, the average number of inputs required by the algorithm is asymptotically optimal among all simulations that are fast in the sense of Nacu and Peres. A non-randomized version of the algorithm is also given. Some extensions are discussed.
title An asymptotically optimal Bernoulli factory for certain functions that can be expressed as power series
topic Statistics Theory
65C50, 62L
G.3
url https://arxiv.org/abs/1612.08923