← All problems

Quadratic Easy Coefficients Conjecture

Let ff be a finite sum of quadratic monomial rotation-symmetric Boolean functions, and let Wf(n)W_f(n) denote the Hamming weight of its induced Boolean function on nn variables. Let R(f)\mathcal{R}(f) be the associated rules matrix, and let λ1,…,λN\lambda_1,\ldots,\lambda_N be the characteristic values of R(f)\mathcal{R}(f), counted with algebraic multiplicity. The Quadratic Easy Coefficients Conjecture asserts that, for the relevant values of nn, Wf(n)=2n−1−12∑j=1Nλj nW_f(n)=2^{n-1}-\frac{1}{2}\sum_{j=1}^{N}\lambda_j^{\,n}; equivalently, the coefficient sequence of the weight recurrence is given by the power sums of the characteristic values of the rules matrix.

References

Progress summary

Refreshed
Open · solution reported

An unrefereed preprint claims to prove the conjecture, but no independent verification has been found.

The conjecture concerns coefficient formulas for weight sequences of quadratic rotation-symmetric Boolean functions, connected with de Bruijn matrices and symbolic dynamics. Cusick’s February 15, 2025 paper established substantial special cases but explicitly left the general repeated-root case unresolved.

Known results

  • Cusick, 2025: quadratic cases satisfy integer-coefficient linear recurrences of orders governed by the minimal polynomial and by 2T+12^T+1.
  • Cusick, 2025: the conjecture holds for quadratic monomial rotation-symmetric functions, corresponding to m=1m=1.
  • Cusick, 2025: for general quadratic functions, the coefficient formula holds when the minimal and characteristic polynomials coincide, equivalently when the characteristic polynomial has no repeated roots.

September 1, 2026 claimed proof

A new preprint identifies a signed de Bruijn transfer matrix with a finite-type shift adjacency block and derives the conjectured coefficients from a dynamical zeta function. This is presented as a full proof, but it is unrefereed and no independent verification or error analysis was found.

Current status (as of September 2026): A preprint claims a complete proof, but the conjecture remains unverified, especially as no independent confirmation of the new argument has been found.

Sources

Solutions 0

No solutions have been posted yet.