VibeMathedMath problems solved with AI

Sparse domination implies convex body domination

Nazarov, Petermichl, Treil and Volberg conjectured that scalar sparse domination should imply convex body domination for the corresponding coordinate-wise vector-valued extension. More precisely, if a bilinear form Λ\Lambda admits an (r,s)(r,s)-sparse bound, then its extension to Cn\mathbb C^n-valued functions should admit an (r,s)(r,s)-convex body sparse bound. Laukkarinen and Lorist prove this implication for 1r,s<1\leq r,s<\infty with 1/r+1/s>11/r+1/s>1, in particular for the classical (1,1)(1,1)-sparse setting.

Result
Proved(see note)
Status
Resolved
AI contribution
AI co-developed
Method
Argument
Field
Harmonic analysis
Posed by
Fedor Nazarov, Stefanie Petermichl, Sergei Treil, Alexander Volberg
Year posed
2017
Years open
9y
Solved
2026-08-25
Model
GPT-5.6 Sol Pro
Vendor
OpenAI
Collaborators
Aapo Laukkarinen, Emiel Lorist
Verification
Unreviewed
Publication
Preprint
Significance
27 / 100
Disclosed cost
Wikipedia
No dedicated article

What was actually shown

If a bilinear form admits an (r,s)(r,s)-sparse bound, its coordinate-wise extension to Cn\mathbb C^n-valued functions admits an (r,s)(r,s)-convex body sparse bound, for 1r,s<1\le r,s<\infty with 1r+1s>1\tfrac1r+\tfrac1s>1. It holds both in a fixed dyadic lattice (Theorem 2.6, constants independent of the ambient dimension) and for arbitrary cubes (Theorem 3.4), via a randomization of the good part of the form.

On scope, two things separate. What Nazarov, Petermichl, Treil and Volberg proposed was the principle for the classical sparse setting, and that is settled outright, the hypothesis holding comfortably there. The (r,s)(r,s) range is the authors' own generalization beyond what was asked, and the restriction bites only inside it, leaving 1r+1s1\tfrac1r+\tfrac1s\le1 open. Their summary: "we provide a proof of the full general statement".

It also gives sparse domination for iterated commutators directly from that of the underlying form. The dependence on nn is not claimed optimal.

What the AI did

GPT-5.6 Sol Pro was used to prototype proof strategies. In an initial extended dialogue, the authors and ChatGPT iteratively developed a proof that sparse domination for commutators [b,T][b,T] implies convex body domination for TT. The authors then suggested replacing the role of bb by Rademacher random variables; from this idea, a rough version of the dyadic proof of the main theorem was developed in collaboration with ChatGPT. The final proof, as well as its extension from a fixed dyadic lattice to arbitrary cubes, was fully developed, verified and written by the authors. ChatGPT was subsequently also used to check the manuscript for typos and errors.

Verification

Unreviewed: an arXiv preprint (v1, 25 August 2026, math.CA), unrefereed, with no formalization and no computational certificate, so there was nothing mechanical to re-run and none of the analysis was checked here. What was verified on 29 August 2026: the paper exists at arXiv:2608.24802 with this title and both authors; Theorem A and Corollary B are as the entry describes them; the AI disclosure is a dedicated section, quoted in the AI-role note; and the attribution is sound - arXiv:1701.01907 is Nazarov, Petermichl, Treil and Volberg's 2017 paper, and its Section 3.1 states the principle and says the authors could not prove it, in the words quoted in the age note.

One discrepancy a reader may hit, flagged rather than silently smoothed: the abstract states the implication with no hypothesis on (r,s)(r,s), while Theorem A in the body requires 1r,s<1\le r,s<\infty with 1r+1s>1\tfrac1r+\tfrac1s>1. This entry follows Theorem A.

Sources

Submitted by RustyKestrel290 on

Changelog2 changes

Discussion