Alphabeta Math
LemmaStatement: AI-adaptedProof: AI-adaptedPipeline-generatedjudge pass (gpt-5.6-terra)audited 2026-09-09
How statement and proof provenance work

The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.

  • Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
  • AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
  • AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.

These labels describe origin, not correctness: citations and verification chips remain separate evidence.

Qid maximal blowup trichotomy

Statement

For a nonempty finite graph H, g∈V(H) and α>0, there exist β,γ>0 such that for every finite graph G with n=∣G∣≥2 and 0<x≤1/(8h), where h=∣H∣, at least one of the following holds:

  • Some A⊆V(G) has ∣A∣≥xβn and ind⁡H−g(G[A])<xα∣A∣h−1.
  • ind⁡H(G)≥xγnh.
  • Disjoint A,B⊆V(G) have ∣A∣≥xβn, ∣B∣≥n/(2h), and B is x-sparse to A in G or G‾.

Facts & Assumptions

Given: A fixed nonempty H, g∈V(H), α>0, and arbitrary G,x in the stated ranges.

[F1]

For every real y, its unique integer part satisfies ⌊y⌋≤y<⌊y⌋+1. (Integer part: for every real x there is exactly one integer m with m≤x<m+1).

[F2]

From Few induced copies exclude a fixed labelled blowup: If ind⁡J(G)<(t/j)j, no (t,1/j)-blowup of J exists in G.

[F3]

For a nonempty pattern with distinguished vertex g, parameters b,c>0, a=b+(1+c)∣H∣, and disjoint sets A,B with at least x∣B∣ nonneighbors in B for every vertex of A, 0<x≤1/2, the local trichotomy gives a few-(H−g) subset of B of relative size at least x, at least xa∣A∣∣B∣∣H∣−1 induced H embeddings, or a pair of relative sizes at least xa with cross density at most 2xc. (Local special copy trichotomy).

[F4]

From Qid fixed size density selection: Some m-subset C⊆A satisfies eG(C,B)≤meG(A,B)/N.

[F5]

From Qid bipartite density trimming: If eG(A,B)≤c∣A∣∣B∣, then some A′⊆A with ∣A′∣≥∣A∣/2 has ∣NG(v)∩B∣≤2c∣B∣ for every v∈A′.

Proof

1.1F1given

If h=1, take β=γ=1; the count is n≥xn. For h≥2, first prove the assertion when 1/x is an integer. Enlarge α to a positive integer at least h(h+1), using [F1]; proving the smaller copy threshold for this enlarged exponent implies the original first alternative. Set rh=0, ri=α+2h+1+(h+1)ri+1 for i<h, β0=r1+3, and γ0=2r1+hβ0. These are positive integers, ri≥1 for i<h, and r1≥krk for 1≤k<h: iterating ri≥(h+1)ri+1 gives r1≥(h+1)k−1rk≥krk.

2.1step 1.1given

If xβ0n≤1, choose a vertex v. Of its neighbors and nonneighbors outside v, one has size at least (n−1)/2≥n/(2h). That set is anticomplete to A={v} in one of the two graphs, proving the third alternative. Hence assume xβ0n>1 and all three alternatives fail for β0,γ0.

3.1F1F2step 1.1step 2.1

Let t=⌊xβ0−1n⌋. The argument exceeds 1/x≥2, so [F1] gives t≥xβ0−1n/2≥xβ0n>1. Put ti=x−rit, qi=xri/h; the reciprocal-integer assumption makes every ti an integer, and t1≤x2n<n. Since 2r1/h≥1, we have γ0/h≥β0+1, whence t≥xβ0n≥hxγ0/hn. Failure of the count alternative and [F2] exclude a (th,qh)=(t,1/h)-blowup of H.

4.1step 2.1step 3.1algebra

A one-label subgraph of H has a (t1,q1)-blowup by taking any t1 vertices. Among finitely many vertex subsets of H admitting the specified blowup, take one of greatest size, say J of size 1≤k<h, with blocks Aj. Choose w∈V(H)∖V(J) and let L=⋃jAj. Outside L, let Mj contain vertices with at most x∣Aj∣ neighbors if wj is an edge of H, or at most x∣Aj∣ nonneighbors otherwise. If ∣Mj∣≥n/(2h), then Aj,Mj give the third alternative. Thus every ∣Mj∣<n/(2h). Also ∣L∣=ktk≤hx2n≤n/(2h). Consequently Z=V(G)∖(L∪⋃jMj) has size at least n/2.

5.1F3step 4.1algebra

Write R=rk+1, Δ=rk−R, and s=xΔ. Then tk+1=stk, qk+1=qk/s, and a∗=α+(R+2)h=Δ−1. Fix j∈V(J) and Y⊆Z of size at least ∣Z∣sk−1. If wj is a nonedge, apply [F3] to H,g,G,Y,Aj with its parameters b=α, c=R+1. Its nonneighbor hypothesis holds by the definition of Z. If wj is an edge, apply it to H‾,g,G‾,Y,Aj. Complementing both graphs preserves every induced embedding and the count of H−g; thus the same first two contradictions below apply in this case too.

6.1F3step 3.1step 5.1algebra

The first outcome of [F3] would give a set of size at least xtk=x1−rkt≥t≥xβ0n with the forbidden few-copy property. For the second, ∣Y∣≥nx(k−1)Δ/2≥nxr1: indeed (k−1)Δ+1≤krk≤r1 and 1/2≥x. Also ∣Aj∣≥xβ0n and a∗≤rk. Therefore its lower count is at least xrk+r1+β0(h−1)nh≥xγ0nh. Both outcomes are excluded. The third gives E⊆Aj, D⊆Y of relative sizes at least xΔ−1=s/x, with cross density at most 2xR+1 in the graph chosen in the previous step.

7.1F4F5step 6.1algebra

Because ∣E∣≥(s/x)tk=tk+1/x≥2tk+1, [F4] chooses C⊆E of exactly 2tk+1 vertices without increasing its cross density to D. Apply [F5] to D,C to get X⊆D of size at least ∣D∣/2≥s∣Y∣, with degrees into C at most 4xR+1∣C∣≤(qk+1/2)∣C∣. This is the required one-block processing step.

8.1step 1.1step 7.1algebra

Process the k labels in any fixed order, starting with Y=Z and replacing Y by X at each step. Before the last step its size is at least ∣Z∣sk−1, so the preceding construction applies every time. Degree bounds into previously chosen Cj survive restriction of their source set. The final X has size at least nsk/2≥nxβ0−1−R≥tk+1: use kΔ+1+R≤krk+1≤r1+1<β0−1 and 1/2≥x. Take Dw⊆X of size tk+1.

9.1F5step 4.1step 5.1step 8.1

For each j, the bound from Dw into Cj implies cross density at most qk+1/2. Applying [F5] to Cj,Dw supplies at least ∣Cj∣/2=tk+1 vertices with degree into Dw at most qk+1∣Dw∣; take exactly that many as Dj. The reverse degree bound into Dj follows from the old bound into Cj: it is at most (qk+1/2) 2tk+1=qk+1∣Dj∣. For two old labels, restricting the target from Aj to Dj changes the bound by a factor at most 1/s, giving qk/s=qk+1 in both directions. Thus these disjoint sets form a (tk+1,qk+1)-blowup on J∪{w}, contrary to maximality. This proves the reciprocal-integer case.

10.1F1step 1.1step 9.1algebra∎

For arbitrary allowed x, put N=−⌊−1/x⌋ and y=1/N. By [F1], 1/x≤N<1/x+1, so x/(1+x)<y≤x and in particular x2≤y≤x. Apply the proved case at y and take final β=2β0, γ=2γ0. Then yβ0≥xβ, yγ0≥xγ, yα≤xα, and y-sparsity implies x-sparsity. Each of the three alternatives therefore implies its required counterpart at x, including the original unenlarged α.

Source notes

Proof/convention locator: Bucic, Nguyen, Scott and Seymour, Induced subgraph density I, 4.3 complete proof.

Depends on

Used by

Dependency tree · two levels

38 results within two dependency steps of this one, each drawn at its shortest distance from it. An arrow runs from a result to what uses it, so the chart reads left to right and ends at this result, which carries a heavier outline. Every node is a link to that result. Click elsewhere on the chart to enlarge it.

Sources