OctoInformBeta

OpenAI preprint, September 2026 · §7. Characteristic transfer and smooth linear sections

Proposition 7.3: Effective section sizes and degree parameters for diagonal permanent coefficients

Critical dimensions constrain border determinantal representations.

Hover over a blue or green term in the text, or a node in the graph, to find it in the other. Select it to read its declaration both in Lean and in words.

The Setting

Adjoining two variables along the diagonal of a matrix and extracting a coefficient produces a polynomial in the matrix entries. The question is how the dimension of its critical locus constrains the size of a border determinantal representation of the permanent.

For a finite variable set σ\sigma, write ℂ[𝒙σ]\mathbb C[\mathbf x_\sigma] for the polynomial ring over ℂ\mathbb C.✦ For a nonnegative integer mm, put Im={0,…,m−1}I_m=\{0,\ldots,m-1\}, so that the matrix variables of size mm are indexed by Im×ImI_m\times I_m.✦ The permanent per⁡m∈ℂ[𝒙Im×Im]\operatorname{per}_m\in\mathbb C[\mathbf x_{I_m\times I_m}] of size mm is the polynomial defined in (1), where SImS_{I_m} is the permutation group of ImI_m. per⁡m(𝒙)=∑τ∈SIm∏i∈Imxi,τ(i).\operatorname{per}_m(\mathbf x) =\sum_{\tau\in S_{I_m}}\prod_{i\in I_m}x_{i,\tau(i)}.(1)✦

Given complex n×nn\times n matrices A0A_0 and AvA_v for v∈σv\in\sigma, the determinant DA0,A∈ℂ[𝒙σ]D_{A_0,A}\in\mathbb C[\mathbf x_\sigma] of their affine pencil is the polynomial in (2). DA0,A(𝒙)=det⁡(A0+∑v∈σxvAv).D_{A_0,A}(\mathbf x) =\det\!\left(A_0+\sum_{v\in\sigma}x_vA_v\right).(2)✦ A border determinantal representation of p∈ℂ[𝒙σ]p\in\mathbb C[\mathbf x_\sigma] of size n>0n>0 is a sequence of affine n×nn\times n matrix pencils in the variables indexed by σ\sigma, with unrestricted complex matrix entries, whose determinants satisfy (3) for every exponent vector α∈ℕσ\alpha\in\mathbb N^\sigma; its existence is denoted HasBorderDeterminant⁡(n,p)\operatorname{HasBorderDeterminant}(n,p). limj→∞[𝒙α]DA0,j,Aj=[𝒙α]p(α∈ℕσ).\lim_{j\to\infty}[\mathbf x^\alpha]D_{A_{0,j},A_j} =[\mathbf x^\alpha]p \qquad(\alpha\in\mathbb N^\sigma).(3)✦

Coefficient and Critical Dimension

For a commutative ring RR, a finite index set II, a square matrix M∈Mat⁡I(R)M\in\operatorname{Mat}_I(R), weights θ∈RI\theta\in R^I, and q∈ℕq\in\mathbb N, the diagonal permanent coefficient 𝒞q(M,θ)\mathcal C_q(M,\theta) is the element of RR defined by (4). Here s,ts,t are independent variables, and [sqtq][s^qt^q] extracts the coefficient of bidegree (q,q)(q,q). 𝒞q(M,θ)=[sqtq]per⁡(M+diag⁡i∈I(s+θit)).\mathcal C_q(M,\theta) =[s^qt^q]\operatorname{per}\!\left( M+\operatorname{diag}_{i\in I}(s+\theta_i t)\right).(4)✦

Let KK be a field and kk a positive integer, let X=(xab)a,b∈IkX=(x_{ab})_{a,b\in I_k} be the matrix of independent variables over KK, and choose θ∈KIk\theta\in K^{I_k}. Put fθ,q=𝒞q(X,θ)f_{\theta,q}=\mathcal C_q(X,\theta). The critical ideal Jθ,qJ_{\theta,q} of fθ,qf_{\theta,q} is the ideal of K[xab:a,b∈Ik]K[x_{ab}:a,b\in I_k] generated by all first partial derivatives of fθ,qf_{\theta,q}, and the critical dimension δθ,q\delta_{\theta,q} of fθ,qf_{\theta,q} is the Krull dimension of the quotient in (5). Jθ,q=(∂fθ,q∂xab:a,b∈Ik),δθ,q=dim⁡Krull(K[xab:a,b∈Ik]/Jθ,q).J_{\theta,q} =\left(\frac{\partial f_{\theta,q}}{\partial x_{ab}}: a,b\in I_k\right), \qquad \delta_{\theta,q} =\dim_{\mathrm{Krull}}\bigl(K[x_{ab}:a,b\in I_k]/J_{\theta,q}\bigr).(5)

For a positive integer kk, the selector index qkq_k, degree parameter rkr_k, and linear section size dkd_k are defined in (6). qk=⌊k−13⌋,rk=k−2qk,dk=⌊(qk+22)18⌋.q_k=\left\lfloor\frac{k-1}{3}\right\rfloor, \qquad r_k=k-2q_k, \qquad d_k=\left\lfloor\frac{\binom{q_k+2}{2}}{18}\right\rfloor.(6)

A homogeneous polynomial pp in a polynomial ring over a field is projectively smooth when pp and all its first partial derivatives have only the origin as a common affine zero.

The Claim

Theorem 1. Let KK be a field, and let k,q,d,B,m,nk,q,d,B,m,n be nonnegative integers with k>0k>0, q≤kq\le k, k−2q≥2k-2q\ge2, and d≥2d\ge2. Choose θ∈KIk\theta\in K^{I_k} and suppose that m≥11k−2qm\ge11k-2q, δθ,q≤B\delta_{\theta,q}\le B, and B+d−1<k2B+d-1<k^2. If HasBorderDeterminant⁡(n,per⁡m)\operatorname{HasBorderDeterminant}(n,\operatorname{per}_m), then (7) holds. (k−2q−1)(d−1)4exp⁡(1)≤n.\frac{(k-2q-1)(d-1)}{4\exp(1)}\le n.(7)✦ For every integer k≥128k\ge128, the effective parameters in (6) satisfy the section size bounds (8) and the degree bound (9). 2≤dk≤k2,k2330≤dk−1.2\le d_k\le k^2, \qquad \frac{k^2}{330}\le d_k-1.(8)✦ 33k100≤rk−1.\frac{33k}{100}\le r_k-1.(9)✦

In the paper, the result reads: for k≥128k\ge128, q=⌊(k−1)/3⌋q=\lfloor(k-1)/3\rfloor, r=k−2qr=k-2q, and d=⌊k2/100⌋d=\lfloor k^2/100\rfloor, there is a nonzero homogeneous polynomial G∈ℂ[y1,…,yd]G\in\mathbb C[y_1,\ldots,y_d] of degree rr with smooth projective zero set, and for every m≥11k−2qm\ge11k-2q there is an affine map Φm:ℂd→ℂIm×Im\Phi_m:\mathbb C^d\to\mathbb C^{I_m\times I_m} with G=per⁡m∘ΦmG=\operatorname{per}_m\circ\Phi_m, while r−1≥k/4r-1\ge k/4 and d−1≥k2/200d-1\ge k^2/200.

What Is Checked

The border bound is stated for the complex representations in (3) and the critical quotient over an arbitrary field KK in (5).