ProbXiv
Explorersign in
Problem archiveProblem record

Statement

Let p=qp=q denote the common crossover probability of the teacher's and student's binary symmetric channels, and let D(a∥b)D(a\|b) denote the binary relative entropy. The optimal learning rate is the best exponential rate at which the student's error probability decays under admissible teaching and learning strategies.

Huleihel et al.'s conjecture. In the regime p→1/2p\to 1/2, the optimal learning rate of the student is

D(1/2∥p)(1+o(1));D(1/2\|p)(1+o(1));

i.e., it agrees with the teacher's learning rate up to first-order error terms.

This conjecture concerns whether the student can asymptotically match the teacher's learning rate when both communication channels have the same noise level near maximal noise. The source attributes it to Huleihel et al.; its resolution is not given here.

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.