OctoInformBeta

OpenAI preprint, September 2026 · Section 2. An extremal weight and its cross-mass bound

Weighted graph masses and the variational functional: Entropy and edge interactions in a global maximization problem over nonnegative weights

Vertex weights and their variational potential.

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 G=(V,E)G=(V,E) be a finite simple graph and let w:V→ℝw:V\to\mathbb R be a vertex weight. Write N(u)N(u) for the neighborhood of a vertex u∈Vu\in V, the set of vertices adjacent to uu.✦ The vertex sets S,D⊆VS,D\subseteq V need not be disjoint.

Neighborhood Contributions

For a real vertex weight ww on GG, the common neighborhood mass cw(u,v)c_w(u,v) of vertices u,v∈Vu,v\in V is the weighted sum in (1). cw(u,v)=∑z∈N(u)∩N(v)wz.c_w(u,v)=\sum_{z\in N(u)\cap N(v)}w_z.(1)✦ For a real vertex weight ww on GG, the neighborhood triangle contribution hw(u)h_w(u) at a vertex u∈Vu\in V is the weighted sum in (2). hw(u)=∑v∈N(u)wvcw(u,v).h_w(u)=\sum_{v\in N(u)}w_v c_w(u,v).(2)✦

The Definition

Definition 1. For a real vertex weight ww on GG, the mass w(S)w(S) of a vertex set S⊆VS\subseteq V is the sum in (3). w(S)=∑v∈Swv.w(S)=\sum_{v\in S}w_v.(3)✦ For a real vertex weight ww on GG, the cross mass ew(S,D)e_w(S,D) between vertex sets S,D⊆VS,D\subseteq V is the sum over adjacent pairs directed from SS to DD in (4). ew(S,D)=∑u∈S∑v∈D∩N(u)wuwv.e_w(S,D)=\sum_{u\in S}\ \sum_{v\in D\cap N(u)}w_u w_v.(4)✦ The unordered edge mass MG(w)M_G(w) of a real vertex weight ww on GG is the quadratic expression in (5). MG(w)=12ew(V,V)=∑{u,v}∈Ewuwv.M_G(w)=\frac12 e_w(V,V)=\sum_{\{u,v\}\in E}w_u w_v.(5)✦ Let 𝒯(G)\mathcal T(G) denote the set of unordered triangles of GG. The unordered triangle mass TG(w)T_G(w) of a real vertex weight ww on GG is the cubic sum in (6). TG(w)=16∑u∈Vwuhw(u)=∑{u,v,z}∈𝒯(G)wuwvwz.T_G(w)=\frac16\sum_{u\in V}w_u h_w(u) =\sum_{\{u,v,z\}\in\mathcal T(G)}w_u w_v w_z.(6)✦ The vertex potential φ:ℝ→ℝ\varphi:\mathbb R\to\mathbb R is the function in (7), with natural logarithms and the convention 0log⁡0=00\log0=0. φ(t)={t(1−log⁡|t|),t≠0,0,t=0.\varphi(t)= \begin{cases} t(1-\log|t|),&t\ne0,\\ 0,&t=0. \end{cases}(7)✦ Subtracting the unordered edge mass from the sum of the vertex potentials gives the graph potential FG(w)F_G(w) of a real vertex weight ww, as in (8). FG(w)=∑v∈Vφ(wv)−MG(w).F_G(w)=\sum_{v\in V}\varphi(w_v)-M_G(w).(8)✦ A global optimizer of FGF_G is a weight ww belonging to the set 𝒪G\mathcal O_G defined by (9). 𝒪G={w∈ℝ≥0V:FG(q)≤FG(w) for every q∈ℝ≥0V}.\mathcal O_G=\left\{w\in\mathbb R_{\ge0}^{V}: F_G(q)\le F_G(w)\text{ for every }q\in\mathbb R_{\ge0}^{V}\right\}.(9)✦

In the paper, the definitions read: w(S)=∑v∈Swvw(S)=\sum_{v\in S}w_v, MG(w)=∑{u,v}∈EwuwvM_G(w)=\sum_{\{u,v\}\in E}w_u w_v, TG(w)=∑{u,v,z}∈𝒯(G)wuwvwzT_G(w)=\sum_{\{u,v,z\}\in\mathcal T(G)}w_u w_v w_z, and ew(S,D)=∑u∈S∑v∈D∩N(u)wuwve_w(S,D)=\sum_{u\in S}\sum_{v\in D\cap N(u)}w_u w_v, with MG(a)=∑{u,v}∈EauavM_G(a)=\sum_{\{u,v\}\in E}a_u a_v also defined for every real vector a∈ℝVa\in\mathbb R^V, while on nonnegative weights FG(w)=∑v∈Vwv(1−log⁡wv)−MG(w)F_G(w)=\sum_{v\in V}w_v(1-\log w_v)-M_G(w) and FG*=max⁡w∈ℝ≥0VFG(w)F_G^*=\max_{w\in\mathbb R_{\ge0}^{V}}F_G(w), using 0log⁡0=00\log0=0.

What Is Checked

These definitions are stated over finite simple graphs, real vertex weights and vertex subsets, with global optimization taken over nonnegative weights.