ProbXiv
Explorersign in
Problem archiveProblem record

Statement

For every n∈Nn\in\mathbb{N} and every random vector X=(X1,…,Xn)∈{±1}nX=(X_1,\dots,X_n)\in\{\pm1\}^n, there is a universal constant C>0C>0 such that, for every integer t≥1t\ge1, one can condition on a set S⊆[n]S\subseteq[n] with ∣S∣≤Ct|S|\le Ct so that the remaining conditional correlations satisfy a bound of order O(1/t)O(1/t); equivalently, the conditioning complexity required to make the relevant aggregate conditional-covariance quantity at most ε\varepsilon is O(1/ε)O(1/\varepsilon) rather than the previously known O(1/ε2)O(1/\varepsilon^2). The conjecture is proved in the cited preprint for signed multivariate totally positive (MTP2\mathrm{MTP}_2) laws, but remains open in general.

Record

Added
Links

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.