VibeMathedMath problems solved with AI

The sharp lower bound for the canonical volume of threefolds of general type with positive geometric genus

For a complex projective threefold XX of general type with canonical singularities, the canonical volume is Vol⁡(X)=lim sup⁡m→∞6Pm(X)m3,\operatorname{Vol}(X)=\limsup_{m\to\infty}\frac{6P_m(X)}{m^3}, which equals KY3K_Y^3 on a minimal model YY and can be a small rational number. J. A. Chen and M. Chen asked for the sharp lower bound of the volume among all threefolds of general type. For threefolds with geometric genus pg(X)>0p_g(X)>0 the bound 1/301/30 was proved by them when χ(OX)≤0\chi(\mathcal{O}_X)\le 0, and it is attained by the hypersurface X28⊂P(1,3,4,5,14)X_{28}\subset\mathbb{P}(1,3,4,5,14); without that assumption the best general bound was 1/751/75 (Chen-Chen 2015). What is the minimal canonical volume of a threefold of general type with pg>0p_g>0: is it 1/301/30?

Result
Proved(see note)
Status
Candidate (review pending)
AI contribution
AI-assisted
Method
Argument
Field
Algebraic geometry; explicit birational geometry of threefolds
Posed by
J. A. Chen and M. Chen (2008; Compos. Math. 2015, Open problem 1.1): find the sharp volume bound for threefolds of general type; here the case of positive geometric genus
Year posed
2008
Years open
18y
Solved
2026-11-10
Model
ChatGPT (version unspecified)
Vendor
OpenAI
Collaborators
Pinxian Bie, Peien Du, Zhengjie Yu
Verification
Unreviewed
Publication
Preprint
Significance
20 / 100
Disclosed cost
—
Wikipedia
No dedicated article

What was actually shown

Theorem 1.1 of arXiv:2611.14402 (Bie, Du, Yu): a complex projective threefold XX of general type with canonical singularities and pg(X)>0p_g(X)>0 has Vol⁡(X)≥1/30\operatorname{Vol}(X)\ge 1/30; the bound is sharp, and strict if χ(OX)>0\chi(\mathcal{O}_X)>0. Theorem 1.2: at equality, pg=1p_g=1, q=h2,0=χ(OX)=0q=h^{2,0}=\chi(\mathcal{O}_X)=0 and the canonical model is a hypersurface X28⊂P(1,3,4,5,14)X_{28}\subset\mathbb{P}(1,3,4,5,14). Theorems 1.3 and 1.4 describe these models and their moduli (irreducible, rational, dimension 103). The proof relies on two preprints posted by the same authors the day before: Yu's fourteenth-canonical birationality theorem (arXiv:2611.12717) and the companion claiming the general minimum 1/4201/420 with no hypothesis on pgp_g (arXiv:2611.12786), which is a separate and larger claim not recorded by this entry.

What the AI did

From the paper's paragraph "AI assistance": the authors' principal contributions are the birational arguments (canonical restriction, divisor markings, local indices), the framework using holomorphic two-forms and global polar obstructions, the trace and fixed-part estimates, and the deformation and moduli analysis. "During this research, ChatGPT (OpenAI) provided a differential-geometric perspective and suggested several key ideas for the analysis of canonical subrings and function fields, the local differential and Darboux arguments, and the treatment of equality cases, including the route to applying the cited hypersurface recognition theorem." The model version is not given. Two Python programs for the exact basket computations are shipped as ancillary files.

Verification

Read by this site on 19 November 2026 in the TeX source of arXiv:2611.14402v1 (67 pages, posted 10 November 2026): the introduction with Theorems 1.1 to 1.4, the account of prior bounds, the AI assistance paragraph and the bibliography. The prior bounds were checked against their sources on arXiv: Chen-Chen 2008 (1/30 for Euler characteristic at most 0, with the example of degree 28) and Chen-Chen 2015 (1/75 for geometric genus 1, and Open problem 1.1). The equality example was recomputed here: degree 28 in weights 1, 3, 4, 5, 14 gives volume 28/840 = 1/30 and canonical degree 1. The proof was not refereed, the ancillary programs were not run, and there is no formal proof. The argument cites two preprints by the same authors posted on 9 November, neither checked here. An arXiv search found no competing claim; expert reaction outside arXiv was not searched for.

Sources

Submitted by MellowGander562 on

Changelog2 changes

Discussion