ProbXiv
Explorersign in
Problem archiveProblem record

Statement

Let [n]={1,2,…,n}[n] = \{1, 2, \ldots, n\}. We define a dd-dimensional entropy power region as follows:

Γd(n)≜{[N(∑i∈SXi)]:X1,…,Xn∈Rd independent, random vectors ,S⊆[n],S≠∅}.\Gamma_d(n) \triangleq \left\{ \begin{bmatrix} N(\sum_{i \in S} X_i) \\ \end{bmatrix}: X_1, \ldots, X_n \in \mathbb{R}^d \text{ independent, random vectors }, S \subseteq [n], S \neq \emptyset \right\}.

Then we let the entropy power region be

Γ(n)=∪d≥1Γd(n)⊆R+2n−1.\Gamma(n) = \cup_{d \geq 1} \Gamma_d(n) \subseteq \mathbb{R}_{+}^{2^n - 1}.

Let G\mathcal{G} be a collection of nonempty subsets of [n][n]. Then given G\mathcal{G}, we call a function β:G→R+\beta: \mathcal{G} \rightarrow \mathbb{R}_{+} a \emph{fractional partition} if for each i∈[n]i \in [n], ∑s∈G:i∈sβS=1\sum_{s \in \mathcal{G}: i \in s} \beta_S = 1. A set function ν:2[n]→R+\nu: 2^{[n]} \rightarrow \mathbb{R}_{+} is called \emph{fractionally superadditive} if

ν([n])≥∑s∈Gβsν(s).\nu([n]) \geq \sum_{ s \in \mathcal{G}} \beta_s \nu(s).

Using these definitions, we define the class of all fractionally superadditive set functions ν\nu with ν(∅)=0\nu(\emptyset) = 0 as follows

ΓFSA(n)≜{[ν(s)]:ν is fractionally superadditive ∀ fractional partitions β,S⊆[n],S≠∅}.\Gamma_{FSA}(n) \triangleq \left\{ \begin{bmatrix} \nu(s) \\ \end{bmatrix}: \nu \text{ is fractionally superadditive } \forall \text{ fractional partitions } \beta, S \subseteq [n], S \neq \emptyset \right\}.

We know that Γ(n)⊆ΓFSA(n)\Gamma(n) \subseteq \Gamma_{FSA}(n) due to Ghassemi and Madiman of which the subset sum EPI of Artstein, Ball, Barthe, and Naor is a special case. Further, we define a set function ν\nu to be supermodular if

ν(s∪t)+ν(s∩t)≥ν(s)+ν(t),\nu(s \cup t) + \nu(s \cap t) \geq \nu(s) + \nu(t),

for all sets s,t⊂[n]s, t \subset [n]. If we let ΓSM\Gamma_{SM} be the class of all supermodular set functions ν\nu with ν(∅)=0\nu(\emptyset) = 0, then it is known that ΓSM(n)⊆ΓFSA(n)\Gamma_{SM}(n) \subseteq \Gamma_{FSA}(n) and that Γ(n)⊈ΓSM(n)\Gamma(n) \nsubseteq \Gamma_{SM}(n) for n=3n=3 or more.

The open problems are the following. First, is the closure of Γ(n)\Gamma(n) convex? It is known that the closure of Γ(n)\Gamma(n) is a log-convex cone. Second, how does one characterize the closure of Γ(3)\Gamma(3)? It is known, for example, that the closure of Γ(2)\Gamma(2) equals ΓFSA(2)\Gamma_{FSA}(2).

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.