ProbXiv
Explorersign in
Problem archiveProblem record

Statement

Let M(W)\mathsf{M}(\bm W) denote the Sherrington–Kirkpatrick Hamiltonian instance with random matrix W\bm W, and let c(W)c(\bm W) be the approximation ratio certified by an algorithm. A certification algorithm runs in polynomial time if its running time is polynomial in the input size. Conjectural hardness of certification. For any ϵ>0\epsilon > 0, there does not exist a polynomial-time certification algorithm for M(W)\mathsf{M}(\bm W) such that c(W)≤2−ϵc(\bm W) \leq 2 - \epsilon with high probability. This conjecture predicts that no efficient certification method can asymptotically improve on the spectral threshold for the Sherrington–Kirkpatrick model. The paper presents evidence for this prediction via the degree-4 sum-of-squares relaxation and a low-degree likelihood-ratio argument, conditional on a broader conjecture concerning the validity of low-degree likelihood-ratio analyses for a large class of hypothesis-testing problems.

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.