ProbXiv
Explorersign in
Problem archiveProblem record

Statement

Let PP and QQ be probability distributions, and let DJS(P,Q)D_{JS}(P,Q) denote their classical Jensen–Shannon divergence. For alpha∈(1/2,1)alpha\in(1/2,1), consider the powered divergence [DJS(P,Q)]α[D_{JS}(P,Q)]^\alpha.

Conjecture on powers of the Jensen–Shannon divergence. The quantity [DJS(P,Q)]α[D_{JS}(P,Q)]^\alpha is not a metric for 1/2<α<11/2<\alpha<1.

The result would complete the classification established in the paper: the powered Jensen–Shannon divergence is a metric for α∈(0,1/2]\alpha\in(0,1/2] and is not a metric for α≥1\alpha\geq1. Counterexamples are given for some subintervals of u∈(0,1)u\in(0,1), depending on α\alpha, but the assertion for the full interval α∈(1/2,1)\alpha\in(1/2,1) 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.