ProbXiv
Explorersign in
Problem archiveProblem record

Statement

On the Boolean hypercube G={−1,1}nG = \{-1,1\}^n with uniform measure λ\lambda, let Tμf(x)=∫Gf(x⊙y) dμ(y)T_\mu f(x) = \int_G f(x \odot y)\,d\mu(y) be convolution by a finite positive measure μ\mu, and set ψμ(u)=sup⁡{u λ({Tμf≥u}):f≥0, ∥f∥1=1},\psi_\mu(u) = \sup\{u\,\lambda(\{T_\mu f \ge u\}) : f \ge 0,\ \|f\|_1 = 1\}, which measures how much better than Markov's inequality convolution makes the tail. In 1989 Talagrand conjectured that for the biased-coin product measure μa=(1+a2δ1+1−a2δ−1)⊗n\mu_a = (\tfrac{1+a}{2}\delta_1 + \tfrac{1-a}{2}\delta_{-1})^{\otimes n} with 0<a<10 < a < 1, ψμa(u)≤Calog⁡u(u>1),\psi_{\mu_a}(u) \le \frac{C_a}{\sqrt{\log u}} \qquad (u > 1), with CaC_a depending on aa alone and not on the dimension nn. He offered a $1000 prize for a proof. The Gaussian analogue was settled by Eldan and Lee; the hypercube case, the original, stayed open.

This paper claims the conjectured bound.

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.