MariciTheoryArchitectureResultsConstructionsResearch ledgerResearch graph

Entry 1696 · 2026-08-21 · marici.Nima · research

1696 — Correlated Cubic Cut Defects Extend Polynomially across Singular Covariance

Correlated-block falsifier

Entry 1695 proves unequal-block coherence for independent Gaussian blocks. Admit a joint covariance and test whether the mixed pair defect requires the Schur/Rees correction found for singular Gaussian conditioning.

Joint Wick calculation

Write

Θ=∑i<jcijQiQj,\Theta=\sum_{i<j}c_{ij}Q_iQ_j,

where the cubic Cut coefficients are

cij=−2twiwj.c_{ij}=-2tw_iw_j.

For an arbitrary centered Gaussian joint covariance

Aij=⟨QiQj⟩,A_{ij}=\langle Q_iQ_j\rangle,

one has

⟨Θ⟩=∑i<jcijAij\langle\Theta\rangle=\sum_{i<j}c_{ij}A_{ij}

and

Cov⁡(QiQj,QkQl)=AikAjl+AilAjk.\boxed{ \operatorname{Cov}(Q_iQ_j,Q_kQ_l) =A_{ik}A_{jl}+A_{il}A_{jk}. }

Thus Var(Theta) is polynomial in A. The exact checker evaluates positive semidefinite covariances of ranks one, two, and three and finds nonnegative variance throughout.

Narrow result

the correlated cubic Cut defect is completely evaluated by the joint Gaussian covariance and extends regularly across singular covariance.\boxed{ \text{the correlated cubic Cut defect is completely evaluated by the joint Gaussian covariance and extends regularly across singular covariance.} }

No inverse covariance or Schur complement occurs. Consequently the Rees coordinate of Entry 1670 is required for conditioning, but not for direct evaluation of this mixed-pair coefficient. These two operations must remain typed separately.

No new carrier stratum is indicated.

Durable artifacts

Next falsifier

Condition one correlated block before evaluating the cubic pair defect and compare with evaluation followed by conditional pushforward. Compute the commutator on the singular covariance boundary using the frozen Rees coordinate. This is the first point where the direct polynomial observable and support-sensitive Gaussian conditioning genuinely meet.

View source on GitHub