OpenAI reports stable blowup for defocusing Schrödinger flow on a torus

For arbitrarily large odd nonlinear powers, some Sobolev index k>8 admits a nonempty open set of Hᵏ initial data on the twelve-dimensional torus whose defocusing nonlinear Schrödinger solutions undergo classical finite-time self-similar blowup.

Sources for this record

What was already known

Defocusing energy conservation does not control the supercritical regularity needed for continuation. The manuscript addresses the possibility of stable blowup from high-regularity data and a related Gaussian global-existence question.

AI involvement AI-led

OpenAI credits itself as manuscript author and states that its internal frontier model produced the results in this release. Individual human input is not detailed enough to establish task-only autonomy.

OpenAI is named as author on the manuscript, and its release attributes the results to an internal frontier model. The individual human input is not detailed enough to establish autonomy.

Importance Notable

Gives stable finite-time blowup in a defocusing energy-supercritical nonlinear Schrödinger equation.

Limitations

The pinned Lean quantifier selects an odd power above each prescribed lower bound; it does not prove the assertion for every sufficiently large odd power. The Gaussian-data consequence is separately formalized but outside this selected claim.

Technical verification Formal (Lean 4)

OpenAI reports a Lean proof matching the selected claim.

Formal checking

The pinned Lean note proves stable self-similar blowup from a nonempty open Hᵏ data set for arbitrarily large odd powers on the twelve-dimensional torus.

Checker
OpenAI-reported Lean 4 / Comparator
Scope
For arbitrarily large odd nonlinear powers, some Sobolev index k>8 admits a nonempty open set of Hᵏ initial data on the twelve-dimensional torus whose defocusing nonlinear Schrödinger solutions undergo classical finite-time self-similar blowup.
Basis
Reported by the linked checker
Evidence source

Checking a formal statement does not establish that it matches the original problem.

Human understanding Few

Few is the default when evidence for a higher level is insufficient. No named person is documented as explaining the central argument for this assessed contribution.

Correction and publication history

2026-10-18 · publication

Primary manuscript, attribution, prior-work, and scoped checking review for non-notifying local inclusion.

Read this published revision