ProbXiv
Explorersign in
Problem archiveProblem record

Statement

Let S\mathcal{S} be the set of pairs of random variables (U,W)∈(0,1)2(U,W)\in(0,1)^2 with prescribed values of E[U]\mathbb{E}[U], E[W]\mathbb{E}[W], E[H2(U)]\mathbb{E}[H_2(U)], and E[H2(W)]\mathbb{E}[H_2(W)]. Define

ζ(mu,mw,eu,ew)=inf⁡(U,W)∈S12E[(U−W)(J(W)−J(U))].\zeta(m_u,m_w,e_u,e_w)=\inf_{(U,W)\in\mathcal{S}}\frac12\mathbb{E}[(U-W)(J(W)-J(U))].

Let Ψ\Psi be the class of functions satisfying the functional inequality defining the paper's finite-dimensional formulation, and define η\eta and ϕ\phi by

η(x)=(1−2H2−1(x))J(H2−1(x)),\eta(x)=(1-2H_2^{-1}(x))J(H_2^{-1}(x)),

with rr determined as in the paper, and

ϕ(x,y)={η(y)−yrη(r),H2(x)>y,0,H2(x)≤y.\phi(x,y)=\begin{cases}\eta(y)-\dfrac{y}{r}\eta(r),&H_2(x)>y,\\0,&H_2(x)\leq y.\end{cases}

The candidate function conjecture. The function ϕ\phi belongs to the class Ψ\Psi. If true, this would establish the most-informative Boolean function conjecture for balanced functions. The claim is a finite-dimensional inequality supported by numerical evidence; its general proof remains open.

Record

Comments

Nothing has been published against this problem yet, and nothing has been said about it.

Sign in with an institutional address to take part in the discussion. Reading every thread stays open to everyone.

Sign in

Solve with an agent

Open the statement in a chat, with the problem and the ground rules already written into the prompt.

This opens a third-party site. Nothing is posted back to ProbXiv and nothing you write there is recorded here — what a model gives you is an attempt, which a person still has to check.