VibeMathedMath problems solved with AI

Whiteley's vertex splitting conjecture for generic global rigidity

A graph GG is generically globally rigid in Rd\mathbb{R}^d if, for a generic placement of its vertices in Rd\mathbb{R}^d, the edge lengths determine the configuration up to isometry. A dd-dimensional vertex split replaces a vertex vv by two adjacent vertices v1,v2v_1, v_2 that share exactly d−1d-1 of the neighbours of vv and divide the remaining neighbours between them. Whiteley proved that vertex splitting preserves generic rigidity in Rd\mathbb{R}^d, and conjectured that it also preserves generic global rigidity in Rd\mathbb{R}^d whenever v1v_1 and v2v_2 both have degree at least d+1d+1. Is this true in every dimension dd?

Result
Proved(see note)
Status
Candidate (review pending)
AI contribution
AI-discovered
Method
Argument
Field
Discrete geometry, rigidity theory
Posed by
W. Whiteley; stated as Conjecture 16 in R. Connelly and W. J. Whiteley, Global Rigidity: The Effect of Coning, Discrete Comput. Geom. 43 (2010), 717-735
Year posed
2010
Years open
16y
Solved
2026-11-07
Model
ChatGPT 6.1 Sol
Vendor
OpenAI
Collaborators
James Cruickshank
Verification
Unreviewed
Publication
Preprint
Significance
28 / 100
Disclosed cost
—
Wikipedia
No dedicated article

What was actually shown

Theorem 1.2: if uv∈E(G)uv\in E(G), deg⁡G(u),deg⁡G(v)≥d+1\deg_G(u),\deg_G(v)\ge d+1, ∣NG(u)∩NG(v)∣=d−1|N_G(u)\cap N_G(v)|=d-1, and H=G/uvH=G/uv is generically globally rigid in Rd\mathbb{R}^d, then so is GG, for every d≥1d\ge 1. Corollary 1.5 allows at least d−1d-1 common neighbours. Theorem 1.3 gives the stronger statement that GG is infinitesimally rigid at every generic realisation in which uu and vv coincide, which yields global rigidity through the Cruickshank-Jackson-Tanigawa coincident-rigidity criterion; two proofs, both via maximum-rank equilibrium stress matrices. Theorem 1.4 also proves the Jordan-Tanigawa conjecture (JCTB 2019, Conjecture 2) that every generically globally rigid graph has a nondegenerate equilibrium stress of maximum rank at every generic realisation.

What the AI did

The paper's "Statement on the use of artificial intelligence": "The proofs of all the new results in this paper were discovered by ChatGPT 6.1 Sol, which also produced the first draft of the paper. The author directed ChatGPT to investigate the problem and contributed to the revision of the manuscript to its current form. This included requesting clarifications and rewrites of the proofs, and making decisions about the organisation and presentation of the paper. The author has checked the arguments presented here and takes full responsibility for the paper."

Verification

Read by this site on 15 November 2026. No independent mathematician has checked this yet; the author states that he checked the arguments. Read: the paper's abstract, AI statement, introduction and main results (Theorems 1.2-1.4, Corollary 1.5 with its proof). The conjecture's wording was compared with its restatement in Cruickshank, Jackson and Tanigawa (arXiv 2002.08680), which attributes it to Whiteley via Connelly-Whiteley 2010 and Jordan-Tanigawa; the 2010 paper itself was not opened. Independent numerical check written here: 1,393 random vertex splits in R2\mathbb{R}^2 and R3\mathbb{R}^3 of graphs on up to 10 vertices that pass a randomized Gortler-Healy-Thurston stress test for generic global rigidity; every split graph was infinitesimally rigid at a random realisation with the two split vertices coincident, as Theorem 1.3 predicts. The proofs were not checked.

Sources

Submitted by MellowGander562 on

Changelog2 changes

Discussion