OctoInformBeta

OpenAI preprint, September 2026 · Section 2. An invariant-kernel obstruction

Section 2: fibre-valued translations, blocks, rows, columns and energy: Coordinate operators and unnormalized Hilbert-Schmidt energy on square-summable functions with finite Euclidean fibres

Translation invariance makes blocks depend on differences.

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

Let GG be a discrete group with identity 11, and let KK be a finite set. The coordinate blocks of a bounded operator record its action from one group fibre to another.

The fibre ℂK\mathbb C^K is the Euclidean space of complex families indexed by KK, of dimension |K||K|.✦ The Hilbert space HH in (1) consists of square-summable functions from GG to this fibre: H=ℓ2(G;ℂK)={ξ:G→ℂK:∑x∈G∥ξ(x)∥2<∞}.H=\ell^2(G;\mathbb C^K) =\left\{\xi:G\longrightarrow\mathbb C^K: \sum_{x\in G}\|\xi(x)\|^2<\infty\right\}.(1)✦ Write ℬ(H)\mathcal B(H) for the bounded complex-linear operators from HH to itself.✦

For x∈Gx\in G, the coordinate inclusion ιx:ℂK→H\iota_x:\mathbb C^K\to H is the bounded complex-linear map supported at xx, defined by (2): (ιxv)(y)={v,y=x,0,y≠x.(\iota_x v)(y)= \begin{cases} v,&y=x,\\ 0,&y\ne x. \end{cases}(2)✦ The coordinate evaluation πx:H→ℂK\pi_x:H\to\mathbb C^K is the bounded complex-linear map given by (3): πxξ=ξ(x).\pi_x\xi=\xi(x).(3)✦

For F∈ℬ(H)F\in\mathcal B(H), its operator row at xx is the bounded map row⁡x(F):H→ℂK\operatorname{row}_x(F):H\to\mathbb C^K defined in (4): row⁡x(F)=πxF.\operatorname{row}_x(F)=\pi_xF.(4)✦ The operator column at yy is the bounded map col⁡y(F):ℂK→H\operatorname{col}_y(F):\mathbb C^K\to H defined in (5): col⁡y(F)=Fιy.\operatorname{col}_y(F)=F\iota_y.(5)✦ The matrix block Fx,y∈ℬ(ℂK)F_{x,y}\in\mathcal B(\mathbb C^K) is the operator in (6): Fx,y=πxFιy.F_{x,y}=\pi_xF\iota_y.(6)✦

Let (ea)a∈K(e_a)_{a\in K} be the standard orthonormal basis of ℂK\mathbb C^K. The coefficient vector c(B)∈ℂK×Kc(B)\in\mathbb C^{K\times K} of an operator B∈ℬ(ℂK)B\in\mathcal B(\mathbb C^K) is the Euclidean array defined by (7): c(B)a,b=(Beb)a(a,b∈K).c(B)_{a,b}=(Be_b)_a\qquad(a,b\in K).(7)✦ For an array W∈ℂK×KW\in\mathbb C^{K\times K}, its coefficient matrix M(W)M(W) is the complex matrix indexed by KK whose (a,b)(a,b) entry is Wa,bW_{a,b}.✦

The fibre translation representation U:G→ℬ(H)U:G\to\mathcal B(H) is the group representation defined by (8): (U(g)ξ)(x)=ξ(g−1x).(U(g)\xi)(x)=\xi(g^{-1}x).(8)✦

The Claim

Definition 1. The translation commutant 𝒞⊆ℬ(H)\mathcal C\subseteq\mathcal B(H) is the set of operators defined in (9): 𝒞={F∈ℬ(H):FU(g)=U(g)F for every g∈G}.\mathcal C=\{F\in\mathcal B(H):FU(g)=U(g)F\text{ for every }g\in G\}.(9)✦ For A∈ℬ(H)A\in\mathcal B(H), its kernel energy E(A)E(A) is the sum in (10): E(A)=∑s∈G∥c(A1,s)∥2.E(A)=\sum_{s\in G}\|c(A_{1,s})\|^2.(10)✦

Theorem 1. Let FF belong to the translation commutant 𝒞\mathcal C. Its blocks satisfy (11) for every x,y∈Gx,y\in G: Fx,y=F1,x−1y.F_{x,y}=F_{1,x^{-1}y}.(11)✦ For every coefficient array W∈ℂK×KW\in\mathbb C^{K\times K}, the Euclidean norm satisfies the unnormalized Hilbert-Schmidt identity (12), where tr\operatorname{tr} is the ordinary matrix trace: ∥W∥2=Re⁡tr⁡(M(W)*M(W)).\|W\|^2=\operatorname{Re}\operatorname{tr}\bigl(M(W)^*M(W)\bigr).(12)✦

In the paper, the result reads: for a countable group GG and a positive integer kk, Hk=ℓ2(G;ℂk)H_k=\ell^2(G;\mathbb C^k) carries (Uk(g)ξ)(x)=ξ(g−1x)(U_k(g)\xi)(x)=\xi(g^{-1}x); the coordinate inclusions ιx\iota_x give Fx,y=ιx*FιyF_{x,y}=\iota_x^*F\iota_y, row⁡1(F)=ι1*F\operatorname{row}_1(F)=\iota_1^*F and col⁡1(F)=Fι1\operatorname{col}_1(F)=F\iota_1; the translation commutant consists of the bounded operators commuting with every Uk(g)U_k(g), its members satisfy Fx,y=F1,x−1yF_{x,y}=F_{1,x^{-1}y}, and the squared Hilbert-Schmidt norm uses the unnormalized trace.

What Is Checked

The constructions are stated over an arbitrary discrete group GG, a finite set KK, and bounded complex-linear operators on the space (1). For every permutation σ:G→G\sigma:G\to G, the coordinate permutation Pσ:H→HP_\sigma:H\to H is a complex-linear isometry given by (13): (Pσξ)(x)=ξ(σ(x)).(P_\sigma\xi)(x)=\xi(\sigma(x)).(13)✦