ProbXiv
Explorersign in
Problem archiveProblem record

Statement

Let X\mathcal{X} be a finite alphabet with ∣X∣>2|\mathcal{X}|>2. For an information-distillation function IDM(∣X∣,β)\mathrm{ID}_M(|\mathcal{X}|,\beta), where MM is the number of quantization levels and β\beta is the mutual-information scale, the conjecture concerns maintaining a fraction η(∣X∣)\eta(|\mathcal{X}|) of that scale as β\beta tends to zero. Information-distillation lower-bound tightness conjecture. For any ∣X∣>2|\mathcal{X}|>2, there exist η(∣X∣)\eta(|\mathcal{X}|), β(∣X∣)>0\beta(|\mathcal{X}|)>0, and c(∣X∣)>0c(|\mathcal{X}|)>0 such that

∣X∣−2∣X∣−1<η(∣X∣)<1,\frac{|\mathcal{X}|-2}{|\mathcal{X}|-1}<\eta(|\mathcal{X}|)<1,

and, for all 0<β<β(∣X∣)0<\beta<\beta(|\mathcal{X}|) and M<c(∣X∣)(log⁡(1/β))∣X∣−1M<c(|\mathcal{X}|)(\log(1/\beta))^{|\mathcal{X}|-1},

IDM(∣X∣,β)<η(∣X∣)⋅β.\mathrm{ID}_M(|\mathcal{X}|,\beta)<\eta(|\mathcal{X}|)\cdot\beta.

The preceding bounds establish tight quantization-level growth for fractions below 1/(∣X∣−1)1/(|\mathcal{X}|-1), while the conjecture would extend a matching lower-bound characterization to a fraction strictly larger than (∣X∣−2)/(∣X∣−1)(|\mathcal{X}|-2)/(|\mathcal{X}|-1) and arbitrarily close to 11. The statement remains unproved in the source.

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.